Brenner and Rector's The Kidney, 8th ed.

CHAPTER 58. Hemodialysis

Jonathan Himmelfarb   Peale Chuang   Gerald Schulman



The Hemodialysis Population, 1957



Transition from Chronic Kidney Disease to Hemodialysis, 1960



Management of Chronic Kidney Disease, 1960



Vascular Access, 1961



Artificial Physiology: General Principles of Hemodialysis, 1965



Components of the Extracorporeal Circuit, 1969



The Dialysis Prescription, 1974



Hemodialysis Adequacy, 1980



Alternative Chronic Hemodialysis Prescriptions, 1983



Alternative Approaches to Quantification of Dialysis, 1984



Management of the Maintenance Hemodialysis Patient, 1985



Complications During Hemodialysis, 1994



The Future of Renal Replacement Therapy, 1998



For more than 300,000 patients in the United States who have reached end-stage renal disease (ESRD), hemodialysis (HD) has become a routine therapy. However, it should not be forgotten that this life-saving treatment has only been routinely applied for ESRD for the past 35 years. Historically, beginning with the successful experiments by Abel and co-workers,[1] it was shown that when blood was circulated through numerous collodion tubes surrounded by a jacket containing dialysis fluid, diffusion of substances from the blood to the dialysate was shown and the investigators demonstrated that the composition of the dialysate fluid would be a major determinant of what was removed or retained during the procedure. Based on these findings Kolff and Berek[2] developed the rotating drum artificial kidney that rapidly found use in treating patients with acute kidney injury (AKI). Pioneering physicians such as Merrill, Scribner and Schreiner[2a] successfully supported patients through the oligoanuric phase of AKI. Teschan introduced HD to the battlefield during the Korean War.

Successful application of labor-intensive and technically demanding HD to AKF was accomplished in the acute care setting. However, the adaptation of the acute procedure to the management of permanent irreversible kidney failure required the intersection of a number of medical and social developments that led to the creation of an infrastructure capable of supporting the current ESRD program. A major issue was finding a method of reliably and repeatedly entering the patient's circulation to perform the treatment. The use of an external bridge or shunt, connected to an artery and vein in the wrist championed by Scribner[2a] for use in AKI, could be used in treating chronic kidney failure (CKF), although with great difficulty owing to infection and repeated episodes of clotting. It was then demonstrated that an internal connection between the radial artery and the distal cephalic vein in the wrist (Cimino fistula) could be hemodynamically tolerated. The vein returning blood under high pressure would eventually become “arterialized” and allow repeated access to the circulation.

The last of the necessary pieces required to allow the infrastructure for delivery of dialysis to develop was the will of the nation to allow a sufficient portion of its treasure devoted to health care to be dedicated to treating for an indefinite period a chronic, unremitting process such as ESRD. Until such a decision was made, HD for ESRD was an expensive and inefficient procedure. Often, it was made available by the decision of a selection committee who chose only the young who were free of comorbid conditions. Employment and level of education were important criteria for selection. However, after much debate in both the lay and the medical communities, in 1973, the U.S. Congress passed the law entitling both dialysis and transplantation treatments by Medicare.[3] Perhaps as a legacy of the restrictive acceptance policies before entitlement, the initial estimate of the numbers of patients who would enter the ESRD program was underestimated. This legislation has eventually evolved to provide care for patients with ESRD irrespective of means, education, employment, or other medical conditions. Although the ESRD entitlement legislation represents a landmark providing life-sustaining therapy, no similar program has been proposed for any other chronic disorder.

Incidence and Prevalence of Hemodialysis Patients in the United States

The U.S. Renal Data Survey (USRDS) collects, analyzes, and distributes information regarding ESRD in the United States based upon records provided by Medicare and other sources and reports their findings in an annual data report (ADR). The most recent ADR reveals in 2003, the United States had 102,567 new ESRD patients of which 93,276 (91%) were started on HD.[4] This is equal to 307 per million population being initiated on HD that year.Figure 58-1 demonstrates that the incidence is not uniform across the nation.

FIGURE 58-1  Incidence of hemodialysis patients in the United States.  (From USRDS 2005 Annual Data Report.) (United States Renal Data Services (USRDS). Accessed April 2006.)



The highest incidences appear in the southeastern and western states, but HD is also becoming increasingly frequent in the eastern states. These areas in general correspond to regions of the nation having a high incidence of patients with diabetes, the most common cause of ESRD. In addition, these regions have large populations of African Americans and Hispanics who are at increased risk of developing ESRD, with incidences of 995.7 and 495.9 per million population, respectively. Figure 58-2 shows that, since the mid-1980s, increasing numbers of patients were reaching ESRD owing to diabetes and, to a lesser extent, hypertension. The rates due to diabetes appear to have started to plateau over the last few years though, changing from only 144.3 to 149.9 cases per million population from 2000 to 2003. The incidence of ESRD owing to glomerulonephritis and cystic disease has remained essentially constant over the last 25 years.

FIGURE 58-2  Causes of end-stage renal disease in the United States.  (USRDS 2005 Annual Data Report.) (United States Renal Data Services (USRDS). Accessed April 2006.)



International Comparisons

The highest prevalence rate for ESRD is found in Japan at 2045 per million population, followed then by the United States with a rate of 1509 per million popula-tion. [5] [6] These high numbers reflect the policies of these two nations to provide open access to chronic dialysis therapy and nearly universal health care for patients with ESRD. The worldwide prevalence rates of ESRD varies greatly, with less than 10 per million population reported by several countries. Currently, 52% of the global dialysis population resides in just four countries (United States, Japan, Germany, and Brazil) that make up only 11% of the world population. Wide differences in access to treatment, the availability of specific modalities of therapy, and methods of reimbursement between nations likely accounts for the broad ranges of reported ESRD prevalence.

Outcomes: Morbidity and Mortality of Hemodialysis Patients in the United States

Figure 58-3 shows that from 1995 to 1999, the number of comorbid diseases in patients at initiation had increased. From 1999 to 2003, the prevalence of ischemic heart disease (ISHD) and congestive heart failure (CHF) has remained stable while cancer and chronic obstructive pulmonary disease (COPD) have become more prevalent. The increased prevalence of the latter two conditions may reflect the older age of the average incident HD patient, which has increased from 54.2 years in 1993 to 57.1 years in 2003. Even though CHF and ISHD prevalence at dialysis initiation has not increased, hospitalizations for these conditions within the first year of initiation of dialysis continue to rise.

FIGURE 58-3  Increasing comorbid conditions in the end-stage renal disease population.  (USRDS 2001 Annual Data Report.) (United States Renal Data Services (USRDS). Accessed April 2006.)



Patients initiated on HD generally have a greater number of comorbidities and are older than patients who are initiated on peritoneal dialysis. HD patients are more likely to have cardiovascular disease (CHF or ISHD), cerebral vascular disease, peripheral vascular disease, and COPD. An important exception is that a high percentage of insulin-dependent diabetics are initiated on peritoneal dialysis. These differences in the population must be taken into account when comparing the two modalities.

The number of hospitalizations and hospital days per patient-year serve as indices of patient morbidity. A total of 2.00 admissions per each patient-year occurred in HD patients in 2003. Hospital days per patient-year for HD patients averaged 14.1 days. These rates have remained fairly stable since the mid-1990s, although the indication for hospitalization has changed, with increased hospitalizations for cardiovascular disease and infections and decreased hospitalizations for vascular access. For all incident and prevalent HD patients, admission rates are highest for those with less than 1 year on dialysis.

In 2003, over 69,000 dialysis patients enrolled in the ESRD program died. This translates to an annual adjusted mortality rate of 210.7 per 1000 patient-years at risk for the dialysis population, which represents an 14% decrease since peaking at 244.5 per 1000 patient-years in 1988. For HD patients and peritoneal dialysis patients, the adjusted mortality rates were 211.3 and 229 per 1000 patient-years at risk, respectively. The highest mortality rate is within the first 6 months of initiating dialysis, which then tends to improve over the next 6 months before increasing gradually over the next 4 years. Since the mid-1990s, a modest decline in the mortality rate has occurred except in the oldest (>80 yr) and youngest (0–19 yr) age groups. Cardiovascular disease is by far the leading cause of death in HD patients, followed by other causes and then infections ( Figure 58-4 ). Indeed, patients on dialysis have a markedly increased risk of morbid cardiovascular events ( Figure 58-5 ) when compared with patients with pre-ESRD chronic kidney disease (CKD), patients who have received a renal transplant, and the general population. The mortality rates associated with HD are striking and indicate that the life expectancy of patients entering into HD is markedly shortened. Figure 58-6 demonstrates that, at every age, patients with ESRD on dialysis are at significantly increased mortality when compared with nondialysis patients and individuals without kidney disease. At age 60, a healthy person can expect to live for more than 20 years, whereas the life expectancy of a 60-year-old patient starting HD is closer to 4 years.

FIGURE 58-4  Adjusted cause-specific mortality, by modality.  (USRDS 2005 Annual Data Report.) (United States Renal Data Services (USRDS). Accessed April 2006.)



FIGURE 58-5  Adjusted rates of cardiovascular complications by age. (Inpatient events per 1000 patient-years at risk.)  (USRDS 2005 Annual Data Report.) (United States Renal Data Services (USRDS). Accessed April 2006.)



FIGURE 58-6  Adjusted rates of mortality by age.  (USRDS 2005 Annual Data Report.) (United States Renal Data Services (USRDS). Accessed April 2006.)



A recurrent controversy involves the question of whether peritoneal or HD represents a superior form of depuration. This question is difficult to answer with certainty because selection for the two therapies is not random and is subject to selection bias. In general, patients who start peritoneal dialysis have fewer comorbid conditions independent of other factors that may influence modality selection. [7] [8] [9] [10] Studies addressing this issue have had conflicting results, with either minimal difference in mortality [7] [11] or a trend with improved mortality with either vintage, [7] [12] [13] [14] [15] [16] depending on the population studied and statistical model used. Since the early 2000s, the number of patients in the United States receiving peritoneal dialysis has remained fairly stable at approximately 25,000, whereas the HD population has continued to increase. In 2003, less than 10% of the dialysis population was treated by peritoneal dialysis.[4] Figure 58-7 demonstrates that, in the first 2 years of treatment for ESRD, survival of peritoneal dialysis patients is superior to that of HD patients. However, in subsequent years, the survival rate is greater for those patients receiving HD.

FIGURE 58-7  Survival of period prevalent hemodialysis and peritoneal dialysis patients adjusted for age, gender, race, and primary diagnosis.  (USRDS 2005 Annual Data Report.) (United States Renal Data Services (USRDS). Accessed April 2006.)



The reasons for this finding have not been established. Loss of residual (native) kidney function, which has been demon strated to correlate with mortality in peritoneal dialysis patients, has been suggested as a factor in this phenomenon. [17] [18] In contrast to the general trend of improved mortality of peritoneal dialysis patient over the first 2 years reported by the USRDS, a prospective cohort study of U.S. incident dialysis patients found that there was no difference in mortality over the 1st year but increased mortality for peritoneal dialysis during the 2nd year.[18] The major strengths of this study were that the population included only incident dialysis patients, these patients had up to 7 years of follow-up, included patients switching modalities, and was an intention-to-treat analysis. However, the study population included only 1041 patients, most of whom were cared for by a single large dialysis organization, and the peritoneal membrane transport characteristics and the rates of residual renal decline were not described.[19]

HD and peritoneal dialysis techniques will change over time. Hence, survival is a moving target. Overall, many issues such as lifestyle, quality of life, and patient preference outweigh current survival statistics in choosing one modality over the other


The knowledge that many patients enter into treatment for ESRD with comorbid conditions, particularly cardiac and vascular complications, and have a strikingly higher mortality rate than that of the general population, has led to the realization that these conditions often begin much earlier, at a time when CKD has just been identified. Therefore, in patients who are at risk of having a progressive, inexorable decline in glomerular filtration rate (GFR), measures should be undertaken as early as possible in the course to correct the common abnormalities associated with CKD, to reduce comorbidity, and to allow a felicitous transition to HD. To this end, the National Kidney Foundation has created the Kidney Disease Outcomes Quality Initiative (K/DOQI) clinical practice guidelines for CKD, a term chosen to be readily understood by patients as well as by physicians.[20]

Cardiovascular risk factors should be assessed as early as possible and treated to prevent complications. Education should be provided sufficiently early in the course whenever possible so that a permanent access can be ready for use if the patient reaches ESRD. Whenever possible, in patients with documented progression, a working permanent HD access should be in place by the time the GFR falls below 20 mL/min.


Reducing Comorbidity

In K/DOQI, guidelines have been promulgated regarding the evaluation, classification, and management of CKD. These guidelines are aimed at correcting or modifying the common abnormalities that are found as renal disease progresses in order to reduce comorbidity in the ESRD population. Optimal pre-ESRD includes strategies aimed at preventing or slowing progression by identifying those who have CKD and initiating appropriate care with dietary management, blood pressure, and glycemic control blockade of the renin-angiotensin-aldosterone system; preventing complications of uremia such as anemia, renal osteodystrophy, and malnutrition; and preparing the patient for the advent of ESRD with education concerning the available treatment modalities, planning for the creation of a permanent access for HD in order avoid the use of temporary catheters, and planning for initiation before major symptoms of uremia arise.

In addition, the high degree of comorbid conditions that coexist in patients suffering from renal disease must also be treated. Indeed, the measures undertaken to treat the complications of CKD will often also modify the course of these comorbid conditions. Examples of this phenomenon include regression of left ventricular hypertrophy when erythropoietin is used to treat the anemia associated with renal failure and the suggestion that the calcium content in the vasculature is partly dependent upon the type of phosphate binder used in the treatment of osteodystrophy.

Multiple lines of evidence suggest that the patient's health status at the time of initiation of dialysis will affect subsequent morbidity and mortality. Protein intake often falls spontaneously as renal function deteriorates.[21] This can lead to a loss of muscle mass and lower serum albumin levels. The serum albumin level at initiation of is a powerful predictor of subsequent mortality: Patients beginning HD with an albumin level 3.0 of 3.5 g/dL have a 20% greater annual mortality rate than patients with an albumin of 3.5 to 4.0 g/dL. Cardiovascular diseases, the cause of the greatest mortality in the HD population, most often have their origin long before dialysis is initiated. Untreated or inadequately treated anemia may lead to worsened left ventricular hypertrophy and exacerbation of angina as the patient approaches ESRD. [23] [24] [25] [26]

Initiation of Hemodialysis

The common indications for initiation of HD in AKI include uncontrolled hypertension, pulmonary edema, acidosis, hyperkalemia, pericarditis, encephalopathy, and elevated blood urea nitrogen (BUN) and creatinine. These indications should never be reasons for initiating chronic maintenance HD. There will be the occasional patient who presents with manifestations of the acute uremic syndrome, owing to lack of prior medical attention or denial. However, the goal for the patients is a smooth transition from CKD to ESRD.

Referral to Nephrologists

Input from a nephrologist should be obtained once CKD has been identified. The frequency of visits will increase as renal function declines. There is emerging evidence that referral to nephrology influences the time at which dialysis is initiated and subsequent outcome. In patients who are followed by nephrologists, dialysis is initiated at a creatinine concentration as much as 4 mg/dL lower than that in patients who have had no medical care. Multiple lines of evidence support the finding that the timing of referral to the nephrologist influences outcome and cost of dialytic therapy. [27] [28] [29] [30] [31]

Starting Hemodialysis

HD should be initiated at a level of residual renal function above which the major symptoms of uremia usually supervene. Among the accepted criteria for initiating dialysis in the United States are residual creatinine clearances of 15 mL/min and 10 mL/min for diabetics and nondiabetics, respectively. Clinical practice guidelines suggest that dialysis be initiated at a creatinine clearance between 9 and 14 mL/min.[20] The determination of an adequate dose of HD is discussed subsequently. At best, with current HD technology, thrice-weekly sessions of 5 hours each will achieve the equivalent urea clearance of approximately 20 mL/min in a 70-kg individual.

The current guidelines of creatinine clearances of 15 mL/min and 10 mL/min or of serum creatinine concentrations of 6 mg/dL and 8 mg/dL for diabetics and nondiabetics, respectively, are reasonable criteria for initiating HD. It may be necessary to initiate dialysis earlier if there are otherwise uncorrectable symptoms or signs of renal failure such as nausea and vomiting, weight loss, intractable CHF or hyperkalemia. Once the creatinine clearance falls below 20 mL/min, patients should be periodically questioned regarding symptoms related to nutrition: loss of appetite, nausea and/or vomiting, especially in the morning owing to overnight retention of uremic toxins in the gut, and unintended weight loss. These are very often the earliest markers of uremia. Asterixis, restless leg syndrome, and a reversal of the sleep-wake cycle are early neurologic manifestations of uremia. If alternative explanations for these symptoms and signs cannot be discerned, they should be indications for initiating dialysis.



The use of HD for the treatment of patients with AKI was introduced by Kolff and Berk[31] in 1943 with temporary access to the circulation. However, the development of maintenance HD therapy for the treatment of ESRD requires repeated access to the circulation. This was not feasible until the introduction of the external arteriovenous (AV) Quinton-Scribner shunt in 1960.[32] The Quinton-Scribner shunt was made of Silastic tubing connected to a Teflon cannula. The Quinton-Scribner shunt developed frequent problems with thrombosis and infection and typically functioned for a period of months. In 1966, Brescia-Cimino and colleagues[33] developed the endogenous AV fistula, which remains the access of choice for maintenance HD today. The infeasibility of developing functional native AV fistulas in all patients led to the development of interpositional bridge grafts in the late 1960s and 1970s. Initial graft biomaterials consisted of autogenous saphenous veins, bovine carotid arteries, and human umbilical veins. In the late 1970s, synthetic bridge grafts made of expanded polytetrafluoroethylene (ePTFE) were introduced. [35] [36] ePTFE grafts can be placed in the majority of patients, are usable within weeks of surgical placement, and are relatively easy to cannulate. ePTFE grafts remain the most frequently utilized graft biomaterial today and indeed continue to be the type of permanent dialysis access most frequently placed in the United States.

Whereas the relative advantages and disadvantages of each type of permanent dialysis access are discussed further, it is clear that native vein AV fistulas are preferable to all other currently available vascular access options. Current clinical practice guidelines recommend the radiocephalic primary AV fistula as the access of choice, followed by the brachial-cephalic primary AV fistula. Patients with CKD should be referred for a surgical attempt to create a primary AV fistula when the creatinine clearance falls below 25 mL/min, when the serum creatinine level is greater than 4 mg/dL, or within 1 year of the anticipated need for maintenance dialysis therapy.

The use of catheters for HD access also parallels the history of dialysis itself. In 1961, Shaldon and associates[36] first described femoral artery catheterization for HD access. In 1979, Uldall and co-workers[37] first reported the use of guidewire exchange techniques and subclavian vein puncture for placement of temporary dialysis catheters. In the late 1980s, the use of surgically implanted tunneled, cuffed, double-lumen catheters was introduced.[38] Recently, subcutaneous vascular ports have been introduced as an alternative to the cuffed tunneled catheter.[39] Whereas the major use of catheters for HD access is as a bridging device to allow time for maturation of a more permanent access or for patients who need only temporary vascular access, catheter use as a permanent vascular access in patients for whom all other options have been exhausted is increasing in frequency.[40]

Whereas there have been impressive technical improvements over time in the development of HD vascular access, it should be noted that none of the currently available types of vascular access meets the performance characteristics required of the “ideal” vascular access ( Table 58-1 ). Indeed, vascular access continues to be referred to as the “Achilles heel” of the HD procedure.[41] As is discussed later, vascular access complications remain responsible for considerable cost, morbidity, and mortality in the maintenance HD patient population.

TABLE 58-1   -- Ideal Vascular Access

High blood flow rates

Instant usability

No needles

Long survival

Low thrombosis rates

Low infection rates

Patient comfort

Minimal cosmetic effect





The rapid growth of ESRD programs in the United States and worldwide has been accompanied by a tremendous increase in dialysis vascular access-associated morbidity and cost.[41] The creation, maintenance, and replacement of vascular access in HD patients is recognized as a major source of morbidity and cost within the U.S. ESRD program, with recent estimates that exceed $1 billion dollars within the Medicare program on an annual basis.

There is now compelling evidence that there are large differences in patterns of vascular access usage between Europe and the United States. The Dialysis Outcomes and Practice Patterns Study (DOPPS) compared vascular access use and survival in Europe and the United States.[42] Native AV fistulas were used by 80% of European patients compared with 24% of prevalent dialysis patients in the United States. AV fistula use was significantly associated with male gender, younger age, lower body mass index, absence of diabetes mellitus, and a lack of peripheral vascular disease. However, even after adjusting for these risk factors, the odds ratio for AV fistula use in Europe versus the United States was 21%. Enormous facility variation has also been noted in the United States, with the prevalence of AV fistulas ranging from 0% to 87%.[42] A follow-up study from DOPPS suggests that predialysis care by a nephrologist does not account for substantial variations in the proportion of patients commencing dialysis with an AV fistula; also, the time to fistula cannulation after creation varies greatly between countries.[43] Thus, practice pattern variations in vascular access care are not entirely due to patient-related factors, but also to processes of care. The late creation of fistulas strongly affects the risk for sepsis.[44]

The importance of vascular access care has been emphasized by data from the USRDS demonstrating that adjusted relative mortality risk is substantially higher for patients with a central venous catheter compared with an AV fistula in both diabetic and nondiabetic patient populations.[45] For diabetic patients, the use of AV grafts is also associated with significantly higher mortality risk compared with AV fistulas ( Figure 58-8 ). Mortality due to both infectious and cardiovascular causes is implicated.

FIGURE 58-8  Adjusted patient survival based on dialysis access type. AVF, arteriovenous fistula; AVG, arteriovenous graft; CVC, cuffed venous catheter.  (Reproduced with permission from Dhingra RK, Young WE, Hulbert-Shearon TE, et al: Type of vascular access and mortality in US hemodialysis patients. Kidney Int 60:1443–1451, 2002.)



Data from Medicare and the USRDS indicate that the prevalence of AV fistula use is increasing in the United States. Medicare has launched a “fistula first” quality initiative that is showing signs of success in increasing placement of AV fistulas.[46] The increase in fistula placement coincides with the publication of Dialysis Outcome Quality Initiative (DOQI) guidelines in relevant clinical practice guidelines. However, the K/DOQI clinical practice guideline recommending that an autologous fistula be placed in 40% of prevalent HD patients is currently not being met. Furthermore, the use of tunneled catheters as the primary means of HD access appears to be rising.[47] Thus, considerable challenges remain in attempting to optimize vascular access practice patterns in the future.

Arteriovenous Fistula

Creation of a native vein fistula requires that an anastomosis be made between an adequate artery and an adequate vein in close proximity to each other ( Figure 58-9 ). The most commonly used site is at the wrist, where the cephalic vein is connected to the radial artery (Brescia-Cimino fistula). The original operation described by Brescia and colleagues[33] was a side-to-side artery-to-vein anastomosis. However, the end-to-side configuration is preferred by many surgeons today as there is a lower incidence of venous hypertension in the hand.[48] A radiocephalic fistula can also be fashioned more distally in the anatomic snuff box.[49] Whereas radiocephalic fistulas are preferred when feasible, several reports suggest that when an aggressive approach is undertaken, a high per centage of radiocephalic fistulas either fail to mature or develop thrombosis. Failure of radiocephalic fistula maturation is especially prevalent in female, diabetic, and older HD patients.[50]

FIGURE 58-9  Types of autogenous fistulas.  (Reproduced with permission from Pereira B, Sayegh M, Blake P [eds]: Chronic Kidney Disease, Dialysis, and Transplantation, 2nd ed. Philadelphia, Elsevier Saunders, 2005, p 344.)

An alternative to the radiocephalic fistula is to use the upper arm veins to create either a brachiocephalic fistula or a brachiobasilic fistula. Because the basilic vein in the upper arm generally lies under deep fascia, use of this vein for an AV fistula requires that the vein be dissected and transposed into a more convenient subcutaneous position.[51] The brachial artery can also be anastomosed to a perforating vein, joining the superficial and deep venous system just below the elbow crease (the Gracz fistula).[52] Compared with brachiocephalic fistulas, transposed brachiobasilic fistulas are more likely to mature but have a higher long-term thrombosis rate.[53] Better patency for upper arm fistulas compared with either lower arm fistulas or AV grafts have been reported, supporting the trend toward creation of upper arm AV fistulas as the access of choice in patients with poor forearm venous anatomy. [54] [55] Age should not be a factor in planning for fistula placement.[55]

Several strategies are evolving in an attempt to increase the prevalence of functioning AV fistulas in the U.S. dialysis population.[56] The surgeon and surgical center characteristics can clearly affect the choice of fistula placement.[57] It has been suggested that systematic use of preoperative ultrasonographic imaging can increase the success rate in the surgical placement of AV fistulas. [59] [60] Both arterial and venous anatomy should be examined, and the cephalic vein should be examined from the wrist to the cephalic-subclavian junction. The vein should be at least 2.5 mm in diameter at the point of anastamosis to increase AV fistula success. An additional successful strategy for increasing AV fistula prevalence involves ligation of tributary veins when AV fistulas fail to mature promptly. Both surgical and endovascular techniques have been used. Whether the short-term use of antiplatelet agents after AV fistula creation can reduce early fistula thrombosis and increase the number of functioning AV fistulas is currently under study in National Institutes of Health (NIH)–sponsored clinical trials. [61] [62]

Arteriovenous Grafts

Dialysis AV grafts made of ePTFE continue to be the most frequently placed type of permanent dialysis access in the United States, accounting for up to 80% of all accesses, depending on geographic region ( Figure 58-10 ). Compared with bovine carotid artery graft biomaterial, ePTFE grafts appear to have fewer complications and to allow easier management of infection by a potential surgical excision of the infected graft segment. ePTFE grafts have the advantage of low early thrombosis rate, surgical ease of placement, and a relatively short time between access creation and successful cannulation. However, short-term advantages are more than outweighed by the long-term increased risk of infection and thrombosis. [42] [63] Unfortunately, the current 1- and 2-year primary patency rate for ePTFE grafts are a dismal 50% and 25%, respectively.[63] Graft thrombosis accounts for 80% of all vascular access dysfunction, and in over 90% of thrombosed grafts, venous stenosis is detected at or distal to the graft vein anastamosis ( Figure 58-11 ).[64] The underlying pathology for the development of venous stenosis is venous neointimal hyperplasia with exuberant vascular smooth muscle cell proliferation, neoangiogenesis within the neointima and adventitia, and an inflammatory macrophage cell layer lining the ePTFE graft material. [66] [67] [68]Immunohistochemical studies have revealed the presence of vascular growth factors, cytokines, byproducts of oxidative stress, and inflammatory proteins within the intimal hyperplastic lesions obtained from HD patients. [66] [69]These studies suggest that specific pathophysiologic processes lead to venous stenoses in HD patients with AV grafts that may be amenable to pharmacologic inhibitory approaches, including systemic and local drug delivery approaches. [70] [71]

FIGURE 58-10  Configuration for typical forearm loop graft.  (Modified from Kapoian T, Kaufman JL, Nosher J, Sherman RA: Dialysis access and recirculation. In Henrich WL [ed]: Dialysis as Treatment of End-Stage Renal Disease, on-line ed. Vol 5. Philadelphia, Current Medicine & Blackwell Science, 1999.)



FIGURE 58-11  Location of stenosis by angiography in failing grafts.  (From Kanterman RY, Vesely TM, Pilgram TK, et al: Dialysis access grafts: Anatomic location of venous stenosis and results of angioplasty. Radiology 195:135–139, 1995.)



Vascular Access Monitoring and Surveillance

A significant advance in the care of HD patients with AV fistulas and grafts is the recognition that physiologic monitoring of access function can frequently identify incipient access failure prior to thrombosis. Vascular access monitoring is based on the premise that identification of high-risk patients for access thrombosis, coupled with elective correction of stenotic lesions, can decrease the incidence of vascular access failure and improve patient outcomes. Schwab and co-workers[71] initially demonstrated that intradialytic dynamic venous pressure monitoring of AV grafts had utility in the detection of graft-associated stenosis. Currently available HD vascular access monitoring techniques include physical examination,[72] static and dynamic venous pressure monitoring, [72] [74] vascular access blood flow monitoring, [75] [76] [77] [78] [79] [80] [81] [82] vascular access imaging, [83] [84] [85] and measurement of access recirculation. The implementation of comprehensive vascular access surveillance programs has allowed achievement of access thrombosis rates lower than those targeted by current clinical practice guidelines. In a nonrandomized study, instituting a vascular access blood flow monitoring program not only substantially decreased graft thrombosis rates but also reduced hospital days, missed outpatient dialysis treatments, and the use of dialysis catheters because of thrombotic events.[74] Although venous stenoses developed less frequently and at a slower rate in patients with native AV fistulas than in patients with AV grafts, it has been suggested that vascular access blood flow monitoring has utility in this patient population as well.[85] Most studies demonstrate that vascular access monitoring and surveillance has efficacy in detecting AV fistula and graft stenosis prior to thrombosis. However, it is currently unclear whether prophylactic angioplasty is successful in prolonging access patency.[86] Whereas nonrandomized studies have suggested a lowering of access thrombosis rate, there is a high degree of restenosis, especially with AV grafts, which exhibit exuberant neointimal hyperplasia.[70] Several relatively small, single-center randomized clinical trials have not shown benefit in the correction of early AV graft stenosis by angioplasty with respect to improving overall graft patency compared with observation and correction at the time of thrombosis. [88] [89] [90] [91] Data with respect to AV fistula patency are also mixed with some studies showing prolongation of fistula patency and others not demonstrating prolongation of fistula patency after prophylactic angioplasty.[91] Larger multicenter randomized clinical trials will be necessary to resolve these controversies. The use of stents may also prolong patency after angioplasty.[92] Although almost all available data suggest considerable utility for vascular access monitoring and surveillance, there are no published randomized trials comparing results of monitoring with no monitoring.

Cuffed Venous Catheters

It has been suggested that the use of cuffed venous catheters is “a conundrum,” and patients “hate living with them, but can't live without them.”[93] Cuffed tunneled dialysis catheters have the advantage of immediate usability and relatively easy placement. In addition, cuffed tunneled catheters can be used as a permanent vascular access for those patients who have exhausted all options for placement of an AV fistula or graft.[40] However, the high infections and thrombotic complications associated with catheter use and the epidemiologic data suggesting higher mortality in patients using catheters make the current trend toward increased prevalence of catheter use in the U.S. dialysis population a disconcerting one.

Noncuffed temporary catheters are suitable for acute vascular access for less than 3 weeks' duration. They should be inserted immediately prior to use, and real-time ultrasound-guided venous puncture is recommended for catheter insertion. Temporary catheters are most suitable for patients with AKI, for the treatment of poisoning, in the intensive care unit setting for continuous renal replacement therapy, and as a short-term bridge until more permanent access can be placed. Tunneled cuffed catheters are employed when more permanent vascular access will not be available for at least 3 weeks, in patients with such substantial comorbidity that there is a short life expectancy, and in patients with no remaining sites for permanent dialysis access.

The most frequent serious complication of venous catheter use is infection. The importance of sepsis as a cause of mortality in ESRD patients is emphasized by data demonstrating a 100- to 300-fold higher sepsis mortality in all dialysis patients compared with the general population.[94] A recent study clearly demonstrates that late permanent access creation (within 4 months of initiating dialysis) is associated with higher sepsis and all-cause mortality rates.[44] In many studies, the frequency of catheter-associated bloodstream infection is approximately two to four episodes per 1000 patient-catheter days.[95] In contrast, the frequency of bacteremias associated with the use of AV fistulas is approximately 0.05 per patient-year. Recent data strongly emphasize the serious morbidity and mortality associated with catheter-related bacteremia in HD patients. In general, attempting to treat catheter-related bacteremia with antibiotics without catheter removal is unsuccessful in the majority of patients. It has been suggested that catheter salvage can be obtained by installation of an antibiotic lock solution into the catheter lumen to eradicate luminal biofilms in addition to a 3-week course of appropriate systemic antibiotics. However, even this approach is successful only approximately 50% of the time.[96] A currently accepted strategy for the treatment of catheter-related bacteremia involves removal of the catheter with delayed replacement until defervescence in patients with severe clinical symptoms, catheter exchange by guidewire in patients with minimal symptoms and normal-appearing tunnel and exit site, and catheter replacement by guidewire with creation of a new tunnel in patients with exit site or tunnel infection. [98] [99] [100] Using this strategy, cure rates of over 80% have been reported.[98] In all cases, patients should be treated with at least 3 weeks of systemic antibiotic therapy. [94] [99] However, even employing these conservative strategies, 15 to 20% of patients with catheter-related bacteremias will experience complicated infections including osteomyelitis, discitis, endocarditis, and septic arthritis. Catheter-related bacteremia due to Staphylococcus aureus is particularly associated with metastatic infection. Several preliminary studies suggest that the use of mupirocin ointment at the catheter exit site may decrease the incidence of S. aureus–associated bacteremias.

Venous catheters are also subject to frequent episodes of thrombosis requiring either thrombolytic therapy or replacement of the catheter. A prospective, randomized, placebo-controlled trial of minidose warfarin for the prevention of dialysis catheter malfunction did not demonstrate a significant effective of warfarin on thrombosis-free catheter survival.[100] The long-term use of cuffed venous catheters may also lead to the development of right atrial thrombi. In a concerning report, intravascular ultrasound prospectively identified the presence of right atrial thrombi in 22% of HD patients with indwelling venous catheters.[101] The use of cuffed venous catheters also predisposes patients to the development of central venous stenosis. Because subclavian vein stenosis may preclude the subsequent successful placement of ipsilateral AV fistulas or grafts, the use of subclavian venous catheters is generally contraindicated in dialysis patients except as a last resort. [103] [104] [105] [106] [107] [108]

Treatment of Vascular Access Dysfunction

When an AV fistula or AV graft thromboses either thrombectomy must be performed or a new dialysis access site must be created. Thrombectomy can be performed via endovascular or surgical techniques and must also be accompanied by correction of the underlying pathophysiology leading to thrombosis (frequently a venous stenosis). When native AV fistula thrombose early, technical factors are often responsible. Successful surgical revision has been reported to occur in 14% to 90% of cases. Late thrombotic occlusion of autologous fistulas is usually due to outflow stenosis, and surgical approaches are less successful. Endovascular treatment of thrombosed autologous fistulas is generally reported to have a higher success rate than surgical techniques but should not be attempted when there is suspected infection in the access. When AV grafts thrombose, similar success is reported with surgical and endovascular techniques. In all cases, attention must be directed to searching for and dilating or bypassing stenoses, including central venous lesions. Unfortunately, primary patency rates after graft thrombosis has occurred are dismal, generally in the 20% to 40% range at 1 year. [109] [110] [111] [112]

Interventional Nephrology

With the advent of an endovascular approach to the dysfunctional dialysis vascular access, it has become clear that vascular access care works best in the context of a comprehensive management program. Vascular access management programs are by nature multidisciplinary and generally involve nephrologists, dialysis unit staff, interventionalists, and surgeons. Increasingly, nephrologists are becoming involved in the direct performance of endovascular procedures, and interventional nephrology is a relatively new area of competency for nephrologists.[112] In the United States, growth of interventional nephrology has been fostered by the development of the American Society of Diagnostic and Inter-ventional Nephrology (ASDIN). Recently published data document that procedures performed by interventional nephrologists can be associated with excellent outcomes, in terms of both safety and efficacy.[113]

Pharmacologic Prevention of Vascular Access Failure

Given the high cost of morbidity associated with vascular access failure, effective pharmacologic prevention of vascular access dysfunction and failure would likely have clinical utility and be cost effective.[69] Currently, there are few randomized clinical trials of drug therapy to prevent HD vascular access dysfunction. Two separate pilot double-blind, randomized, prospective trials have suggested that either fish oils or dipyridamole may be effective in reducing ePTFE graft thrombosis. [115] [116] A multicenter Veterans' Administration Cooperative Trial comparing a combination of clopidogrel plus aspirin with placebo was discontinued prematurely owing to an increase in bleeding complications in the active treatment group.[116] Several studies have suggested that short-term use of antiplatelet agents may reduce early thrombosis after native AV fistula placement. Retrospective analysis has suggested that the use of angiotensin-converting enzyme inhibitors and calcium entry blockers may prolong survival of dialysis grafts. [118] [119] Because of the importance of this clinical and biologic problem, the NIH has recently developed a Dialysis Access Consortium that is conducting multicenter, prospective, randomized, clinical trials of pharmacologic agents designed to reduce the vascular access failure rate for both AV grafts and fistulas.


It is of interest to reflect upon the differences between native and artificial kidneys. There are 1 to 2 million functioning elements or nephrons in the two native kidneys. The artificial kidney, in its hollow-fiber format, contains 8000 to 10,000 fibers and provides a surface area for exchange as high as 1.8 to 2 m2. The diameter of proximal tubule of the nephrons is 40 mm and 14 mm in length, and the diameter of each hollow fiber is approximately 200 mm and longer than 25 cm. In addition to being influenced by diffusive and convective forces, the tubules perform a myriad of biochemical processes on the fluid that is filtered at the glomerulus, whereas depuration in the artificial kidney is solely dependent on the physical forces of diffusion and convection across a semipermeable membrane.


The physiologist uses the concept of clearance to describe the net result of the transport functions of the kidney. The clearance of a substance is the amount removed from plasma divided by the average plasma concentration over the time of measurement. It can be thought of as the volume of plasma that can be completely cleared of the substance in a unit of time. Clearance is also a useful concept when describing the process of dialysis.

The goal of dialysis is straightforward: to remove accumulated fluid and toxins. With respect to toxins, the goal is to maintain their concentrations below the levels at which they produce uremic symptoms. However, the toxic levels of retained substances are not used as performance measures for dialysis because their identities are unknown. Instead, performance of dialysis is judged by clearance. If the generation of a substance is fixed, its clearance (removal rate from plasma/plasma concentration) then becomes a measure of its concentration levels in the patient. The principles and the calculation of clearance are the same for all substances removed by dialysis.

Dialysis relies on the mass transfer across semipermeable membranes. The HD membranes separate the blood and dialysate compartments. Diffusion, convection, and ultrafiltration (UF) across the membrane are properties integral to the dialysis procedure. Diffusion describes the movement of solutes from one compartment to another, relying on a concentration gradient between the two compartments. This is the principal mechanism for toxin removal during HD. Convective transport involves the movement of solutes by bulk flow in association with fluid removal. Convective clearance is the mechanism of toxin removal by the depurative process known as hemofiltration. It is not dependent on concentration gradients, and the magnitude of its contribution to clearance is directly related to the UF rate. Mass solute removal across the dialyzer is a function of effective blood flow (QB) and differences between the afferent and the efferent concentrations of solute, traditionally labeled as arterial and venous (CA and CV). Thus, the definition of diffusive dialyzer clearance (K), similar to creatinine clearance in the normal kidney, is calculated as

K = (QB)(CA - CV)/CA

where (QB)(CA - CV) represents the amount of solute removal and CA is the driving force. This formula describes diffusive clearance when single-pass HD circuits, wherein the blood side is always in contact with fresh dialysate that is continuously being generated, are employed. Older systems employed discrete batches of dialysate that were recirculated. The result of these configurations is that concentrations of the removed toxins rise in the dialysate. The consequence is that the driving force to diffusion decreases. In this case, the driving force is described by CA - DI, where DI is the concentration of the removed substance in the dialysate. Instead of clearance, the term dialysance is used:

D = (QB)(CA - CV)/(CA - DI).

The efficiency of recirculated HD systems are inferior to single-pass systems because the driving force for diffusion is continuously being dissipated by the appearance in the dialysate of the substances being removed. Systems employing single-pass dialysate circuits are now almost universally employed.

These equations, however, neglect the contribution of convection. This phenomenon is directly related to UF and involves the bulk movement of fluid across dialyzer membranes. The driving force for UF is the hydrostatic pressure gradient across the membrane, the transmembrane pressure (TMP). With UF, blood flow leaving the dialyzer (QBo) is less than blood flow entering the dialyzer (QBi). The difference between these values represents UF (Quf). This can be incorporated into the previous equation to yield a more precise definition of clearance:



True clearance should be calculated by using the concentration in the aqueous compartment of blood and the concentration of solute in that compartment. Because solutes diffusing out of blood will appear in the dialysate, it is possible to calculate clearance for solutes not present in the incoming dialysate (e.g., urea) as:



K = QDo(CDo)/CA

where CDo and QDo are the concentrations of solute in the dialysate outlet and the effluent dialysate flow, respectively. Although this equation provides a simple concept for determining clearance, the necessity of measuring low concentrations of any substance in the dialysate increases the error of measurement. Indeed, the most accurate measurement of dialyzer clearance is achieved when both “blood-side” and “dialysate-side” clearances are obtained, thus assuring mass balance.

Factors Influencing Clearance

A number of variables affect clearance by the dialyzer ( Table 58-2 ). The physical and chemical properties of the substance to be removed and its distribution in the body are toxin-related variables. The procedure-related variables include the permeability of the membrane to toxins of various sizes (flux), its hydraulic permeability, the membrane surface area, dialysis time, blood and dialysate flow rates, and dialysate composition.

TABLE 58-2   -- Factors Influencing Clearance by the Dialyzer


Procedure-Related (In Order of Importance)


Low-Molecular-Weight Toxin

High-Molecular-Weight Toxin


Dialysate composition[*]



Blood flow


Protein binding

Dialysate flow

Membrane surface area

Volume of distribution

Membrane surface area

Blood flow



Dialysate flow



Dialysate composition[*]



For potassium, calcium, sodium, and total CO2, this factor is nonapplicable.


The size and charge of the molecule are important intrinsic physical features governing its removal. If the molecule is charged, its behavior will be governed by the Donnan equilibrium. The cation concentration on the blood-side of the membrane will be higher owing to the presence of plasma proteins that are negatively charged.

In the case of sodium, the dialysate-side concentration of sodium should be approximately 3 mEq/L less than the blood-side concentration in order to prevent net transfer of sodium from dialysate to the patient. The lower the molecular weight of the substance, the greater its rate of movement across the membrane or flux (J). Other factors such as binding of the toxin to plasma proteins, a large volume of distribution, or delayed transfer of the substance from the intracellular pool to the intravascular space will result in decreased clearance. An important example of the latter principle involves phosphorus. Phosphorus is rapidly removed from the intravascular compartment, and the patients actually become hypophosphatemic during the HD treatment. However, there is a rebound in the postdialysis interval with phosphorus levels rapidly returning to predialysis levels. Although phosphorus is a relatively small molecule, HD alone is not sufficient to control its level.

As the molecular weight of the substance increases, the properties of the dialysis membrane become increasingly important factors with regard to clearance. The relationship relating J to the clearance of the toxin isJ = AΔC/R

where A is the surface area of the membrane in the dialyzer, ΔC is the concentration gradient between blood and dialysate, and R represents resistance to diffusivity of the substance across the membrane and the thickness of the membrane.

Removal of low-molecular-weight (LMW) substances follows first-order kinetics. Thus, the efficiency of the removal of substances depends on the concentration of the toxin in the blood. Blood and dialysate flow rates are the most important variables for clearance of the LMW substances. Conversely, the clearance of higher-molecular-weight substances depends on membrane porosity, surface area, and dialysis time.

In order to increase the clearance of LMW substances, mere engineering issues must be addressed. Thus, dialyzers with larger surface areas or hollow-fiber geometry that permit higher blood flows to be used without promoting turbulence can be constructed. Two large dialyzers can be connected in series or in parallel configurations to increase the clearance of LMW toxins. High-efficiency dialysis can be accomplished in patients with low to average body surface areas (BSAs; i.e., BSA bears a direct relationship to the volume of distribution of substances such as urea or creatinine) without major increases in dialysis time.

The Hemodialysis Membrane: Effects and Characteristics

In place of glomeruli and tubules that have the ability to perform active transport, the point of exchange for HD is the membrane in the dialyzer. As noted previously, the surface area, surface charge, and pore size are properties of the membrane that directly govern the molecules that can diffuse from blood to the dialysate. These variables determine mass transfer coefficient (KoA) of the membrane for a given solute. The relationship between KoA and clearance (K) of a solute is a function of the blood (QB) and dialysate (QD) flow rates and described by the equationKoA = {QB/1 - QB/QD}Log{(1 - K/QD)/(1 - K/QB)}

The classic view of membranes as inert structures providing solely fluid, ion, and molecular transport is now obsolete. Modern dialysis membranes display numerous physical and adsorptive properties that also contribute to the degree to which blood components are activated by the membranes. Membranes that produce little interaction with blood components such as white blood cells and the humoral components of plasma are described as being biocompatible.

The structural compounds composing the dialysis membranes may be their simplest distinguishing feature, dividing dialyzers into cellulosic, semisynthetic, and synthetic membranes. The use of cellulose and its derivatives, cuprophane and cellulose acetate, is in decline. The side groups of the cellulosic membranes activate the complement system via the alternate pathway, resulting in the repeated generation of the anaphylatoxins C3a and C5a.[119]Synthetic membranes differ from cellulose-based dialyzers in several ways. All cellulose membranes have hydroxyl radicals at the surface that increase their hydrophilicity (membrane wetability). Techniques that mask hydroxyl radicals enhance hydrophobicity and increase protein adsorption.[120] Most synthetic membranes are thicker than less permeable cellulosic membranes. Membrane permeability is inversely proportional to membrane thickness and directly proportional to the membrane's intrinsic diffusion coefficient. However, synthetic membranes also display greater intrinsic diffusion coefficients and maintain their thickness when wet.

Membranes can be symmetrical or asymetrical. The smooth “skin” side interacts with blood for asymetrical membranes. Assymetry, obtained by altering membrane precipitation during manufacturing, allows for greater diffusive permeability.[121] Hence, asymetrical membranes are very useful for hemofiltration. Polyacrylonitrile (PAN) and polysulfone (PS) membranes are commonly used asymmetrical membranes. Polymethylmethacrylate (PMMA) membranes also manifest many of these characteristics. PAN, polyamide (PA), and PMMA membranes have low hydrophilicity and appreciable protein adsorption. Surface charge of the membranes also differs, which affects the sieving of charged solutes.[122]

Biocompatibility of Dialysis Membranes

During HD, patients may experience a number of reactions that are a direct consequence of establishing the extracorporeal circuit. The number and severity of these reactions define the degree of dialysis biocompatibility. In the broadest sense, all aspects of the treatment, such as dialysate composition and temperature, the nature of the anticoagulant, or whether clearance is achieved predominantly by diffusion or convection, affect biocompatibility. However, the biocompatibility of the membrane surface itself is most important and has been most closely studied. It is also important to remember that because HD is a repetitive process, even low-grade or minor membrane-induced reactions at each treatment can eventually lead to important clinical manifestations. The initial observation that transient neutropenia occurred during HD was made about 1968.[123] It was later determined that complement activation via the alternative pathway, with generation of the anaphylatoxins C3a and C5a, was responsible for the decline in neutrophils shortly after initiation of HD.[124]

The effect of the format of the membrane is also significant. Format refers to the structure of the dialyzer rather than the nature of membrane. Dialyzers may be formatted as hollow-fiber or parallel-plate devices. The same membrane type may produce alterations in neutrophils (e.g., release of proteolytic enzymes) that differ depending on whether the membrane in the artificial kidney is formatted as hollow fibers or parallel plates.[125]

Finally, the adsorptive capacity of a membrane can affect its biocompatibility. PAN membranes bind vasoactive materials including C3a, C5a, and bradykinin to a greater extent than cuprophane. [127] [128] The ability to adsorb potentially harmful substances may confer distinct advantages to certain membranes. Conversely, beneficial substances may be removed and protein adsorption may alter the characteristics of the membrane.

Blood-Membrane Interactions

All humoral and cellular components in blood can potentially interact with the dialysis membrane ( Figure 58-12 ). The extent to which any given membrane has been examined with respect to activation of each of the blood components differs ( Figure 58-13 and Table 58-3 ). Operationally, the lack of complement activation and early neutropenia during HD should serve as useful indices of biocompatible membranes.

FIGURE 58-12  Pathways of blood and dialysis membrane interactions. LTB4, leukotriene B4; NK, natural killer; PMMA, polymethylmethacrylate; TNFa, tumor necrosis factor-a.


FIGURE 58-13  Range of changes in parameters of various dialysis membranes. WBC, white blood cell.


TABLE 58-3   -- Comparative Biocompatibility


Relative Order

Heparin consumption

Cuprophane > PS = PMMA

Complement activation

Cuprophane > Cellulose acetate > PC > PAN


Cuprophane > Cellulose acetate > PC > PAN = PMMA

Granulocyte elastase

Cuprophane = PMMA > PS = PAN

Granulocyte adherence

Cuprophane > PS = PAN

Adapted from Mujaijsk R, Ivanovich P: Membranes for extracorporeal therapy. In Maher JF (ed): Replacement of Renal Function by Dialysis. Dordrecht, Kluwer Academic Publishers, 1989, pp 181–188.

PAN, polyacrylonitrile; PC, polyetherpolycarbonate; PMMA, polymethylmethacrylate; PS, polysulfone.




Membrane Permeability

Just as the glomeruli restrict the passage of substances based on surface charge and size, the dialysis membrane also limits the passage of material based on these same physical properties. The nephrologist has the option of regulating what is removed by selecting the membrane based on its permeability to small and larger substances. Dialyzers are classified as conventional, high-efficiency, and high-flux. There is some imprecision surrounding these definitions. The blood flow and the length of treatment employed when using these dialyzers should not be part of the definition. Nor should urea clearance be used in the definition because clearance or dialysance varies with blood flow. Instead, the dialyzer should be defined by the KoAurea (urea mass transfer coefficient), the UF coefficient, and the degree of hydrophobicity/hydrophilicity of its membrane ( Table 58-4 ). The latter parameter governs the permeability of the membrane to high-molecular-weight substances, its degree of biocompatibility, and its ability to adsorb plasma proteins and peptides to its surface.

TABLE 58-4   -- Comparison of Different Dialyzers


KoAurea (mL/min)

Ultrafiltration Coefficient


Membrane Structure



<10 mL/mm Hg/hr





10–19 mL/mm Hg/hr





>15 mL/mm Hg/hr






The conventional dialyzer has a membrane that is homogenous and permits effective small solute clearance, but its clearance of middle molecules is relatively low. Urea clearance at a blood flow of 300 mL/min is less than 200 mL/min The relatively low hydraulic permeability of many of the membranes permits treatments with a dialysis machine that does not have a UF controller. These membranes are cellulose based and contain nucleophilic groups that permit complement activation unless they have been chemically modified. The blood flow and membrane structural limitations on urea mass transfer preclude their use in high-efficiency HD. Both high-efficiency and high-flux dialyzers have membranes with a KoAurea of more than 450 mL/min. Under standard operating conditions of a blood flow of 400 mL/min, the urea clearance can be in excess of 250 mL/min.

In addition to having a high KoAurea, the high-flux membranes are constructed with pores that permit the passage of molecules exceeding 10,000 Da or more with a clearance as high as 40 mL/min. In addition, significant binding of protein and peptides from the blood may occur with these membranes. When the high-flux membrane is chemically modified, hydraulic permeability as well as the permeability to high-molecular-weight substances is reduced, creating a high-efficiency membrane. The final result is that with respect to LMW substances, high-flux and high-efficiency dialyzers have similar performance characteristics. They differ in their ability to remove high-molecular-weight substances.

There are several reasons to use high-efficiency and high-flux dialyzers. Both of these dialyzers have LMW substance clearance rates far in excess of conventional dialyzers. They are useful in large patients with high urea volumes to ensure delivery of an adequate level of therapy. In addition, the high-flux dialyzers also clear higher-molecular-weight substances including substances proven to produce toxicity such as β2-microglobulin (molecular weight 11,800 Da). The surfaces of these membranes are more biocompatible and cause less activation of complement and less neutropenia and immune cell dysfunction during dialysis. However, the primary motivation behind the use of the efficient dialyzers is often the facilitation of shorter dialysis times.


The point of exchange between blood and dialysate in the HD circuit is the artificial kidney. The machine is designed to deliver blood and properly constituted dialysate to the artificial kidney where diffusiom and convection may occur. This requires that blood and dialysate be delivered at accurate rates. In addition, the machine is also invested with a number of on-line monitors to ensure that acceptable ranges of chemical content, temperature, and circuit pressures are continuously present.

Blood Circuit

Blood in the extracorporeal circuit is contained within tubing that is connected to the venous and arterial sides of a patient's access.

Needles are inserted into the patient's blood access, and blood tubing is connected to the needle hubs. Blood is withdrawn from the arterial segment by the blood pump and pumped through the dialyzer back to the patient via the venous segment of tubing. Inadvertent entry of air into the dialysis circuit, air embolism, is a potentially lethal complication and is likeliest to occur between the vascular access site and the blood pump. Air can enter the dialysis circuit from areas around the arterial needle, through leaky or broken tubing or tubing connections, and through the saline infusion set. Air traps are located in the blood tubing to trap air and prevent air from entering the patient's circulation. Air detectors are linked to a relay switch that automatically clamps the venous blood line and shuts off the blood pump if air is detected. If air embolism is suspected, the venous line leading back to the patient should be clamped and the patient should immediately be placed on the left side in the Trendelenburg position. This will tend to sequester the air in the right ventricle of the heart, preventing its propagation into the pulmonary circulation and allowing it to be reabsorbed.

Blood pumps used for HD are roller pumps that use the principles of peristaltic pumping to move blood through tubing. A compressible part of the tubing (the pump segment) is occluded between rollers and a curved rigid track. Elastic recoil refills the pump tubing after the roller has passed over it. The flow rate of the blood pump is dependent on the stroke volume, the speed of rotation of the rollers, and the volume of the pump segment. The blood flow rate displayed on the dialysis machine is based on these three parameters, rather than an actual value from a blood flow probe. This can lead to significantly higher values for the displayed blood flow compared with the true blood flow rate. Incomplete occlusion of the pump segment owing to a pump maladjustment leads to a reduced volume of blood with each pump rotation. This is a common cause of overestimation of blood flow and, hence, clearance. Careful maintenance of the pump is essential to ensure that the prescribed dialysis dose is actually delivered to the patient.

Pressure monitors are usually located proximal to the blood pump and immediately distal to the dialyzer. The proximal monitor, the “arterial monitor,” guards against excessive suction on the vascular access site by the blood pump, and the distal monitor, the “venous monitor,” gauges the resistance to blood return to the venous side of the vascular access. Some machines place the arterial monitor distal to the blood pump and proximal to the dialyzer to detect clotting in the dialyzer and more precisely estimate pressure in the dialyzer blood compartment. To prevent blood clotting in the dialyzer, an anticoagulant such as heparin is often infused into the circuit. A peristaltic pump or syringe pump delivers the anticoagulant into blood in the circuit via a T-tube or T-fitting usually located between the blood pump and the dialyzer.

A blood leak detector is usually placed in the dialysis circuit in the dialysate outflow line. If a blood leak develops through the dialysis membrane, then blood leaking into the dialysate is sensed by the blood leak detector and the appropriate alarm is activated.

Dialysate Circuit

Another major function of the HD machine is to deliver dialysate to the circuit. In some large dialysis units, dialysate is made as a batch and stored in tanks. The dialysate is then simply delivered to each dialysis station and connected to the machine. In other cases, water that has been treated to remove most elements is sent to the HD machine and then mixed with a dialysate concentrate. The water and concentrated dialysate are then properly proportioned by the machine so that dialysate with proper concentrations of electrolytes enters the dialyzer.

Two properties of the dialysate require constant monitoring: conductivity and temperature. A proportioning system dilutes a concentrated dialysate with water. If this system malfunctions, patient blood can be exposed to a hyperosmolar dialysate, resulting in hypernatremia, or a hypoosmolar dialysate, leading to hyponatremia and hemolysis. The primary solutes in the dialysate are electrolytes. Therefore, the concentration of the dialysate is reflected by the concentration of electrolytes and their electrical conductivity. Appropriate proportioning of water and the dialysate is monitored by a meter measuring conductivity of the product dialysate fed into the dialyzer. A temperature monitor prevents complications related to overheated dialysate. Cool dialysate can be uncomfortable for the patient and is dangerous when the patient is unconscious but otherwise may have therapeutic value in preventing hypotension. Overheated dialysate (>42°C), however, can lead to hemolysis. If the conductivity or the temperature is outside of the normal range, then a bypass valve diverts the dialysate around the dialyzer and directly to a drain.

Variations in the Extracorporeal Circuit

Continuous Renal Replacement Therapy

Continuous renal replacement therapy (CRRT) is a form of extracorporeal clearance therapy used in the treatment of AKF. It is used in place of standard HD and requires an intensive care setting for its implementation. There are several techniques that fall under CRRT. They are similar to one another with respect to the extracorporeal circuit and to the general principles of exchange:



SCUF: Slow continuous UF



CAVH or CVVH: Continuous AV/venovenous hemofiltration



CAVHD or CVVHD: Continuous AV/venovenous HD



CAVHDF or CVVHDF: Continuous AV/venovenous hemodiafiltration.

The AV configuration indicates that an artery and a vein are entered to remove and return blood in the extracorporeal circuit, whereas the venovenous configuration indicates that blood is both removed and returned via venous access, such as a catheter. AV procedures can be performed using an external shunt, connecting an artery to a vein. The use of AV access allows arterial pressure to propel blood through the extracorporeal circuit without the absolute need of a blood pump, although a blood pump can be used. The venovenous procedures require a blood pump to be used. The disadvantage of AV procedures is that a large artery must be cannulated for a long period of time. An occasional patient with an existing AV fistula or graft will undergo CRRT. Although the blood in an AV graft is under arterial pressure, a blood pump is still required because the blood must also be returned to the high-pressure system. In general, the LMW substance clearnace rates of the continuous therapies approximate the volume of fluid removed or the dialysate flow rate ( Table 58-5 ).

TABLE 58-5   -- Reasons to Use High-Efficiency Dialyzers

LMW clearance

Ensure adequate dialysis in large patients

HMW clearance

Clearance of HMW substances such as β2-microglobulin (high-flux)


Reduced complement activation, less morbidity and mortality

Short dialysis

Improved lifestyle while receiving adequate therapy


HMW, high-molecular-weight; LMW, low-molecular-weight.




CAVH, CVVH, and SCUF are all variations on the depurative process of hemofiltration (see Artificial Physiology). In hemofiltration, removal of fluid and waste occur entirely by convection or bulk flow. Transmembrane pressure governs the amount of fluid and dissolved waste being ultrafiltered across the membrane. Blood flow through the extracorporeal circuit ranges between 150 and 200 mL/min. With the usual parameters for blood flow and transmembrane pressure, the UF rate, as much as 40 L/day can be removed. The fluid removed is replaced with a balanced electrolyte solution at a rate that is less the rate fluid removed and governed by the desired amount of excess volume that is to be removed from the patient. For instance, a patient with AKI after surgery may be volume expanded, receiving parenteral nutrition and medications amounting to an intake of more than 4 L/day. If 40 L/day is ultrafiltered, one may want to set the replacement rate at only 35 L/day to begin to correct volume expansion.

At high UF rates, replacement fluid must be infused before the filter (predilution) to prevent excessive hemoconcentration in the filter. Excessive hemoconcentration poses a risk of clotting of the filter. Whereas infusing replacement fluid prefilter will prevent clotting, it reduces the efficiency of the clearance because the urea concentration in the blood flowing into the filter will be diluted by the replacement fluid. Thus, with respect to LMW substances, continuous hemofiltration is less efficient than continuous HD or continuous hemodiafiltration. Conversely, for substances of higher molecular weight, convective clearance is superior to HD.

SCUF is a form of limited hemofiltration that provides slow continuous ultrafitration in volume-overloaded patients who are resistant to diuretics or who are hemodynamically unstable. The desired volume to be removed is calculated and removed over the entire day. Patients with massive volume overload and marginal blood pressures can be treated. No replacement fluid is required, but clearances are very low with this technique (∼3.5 mL/min) and SCUF cannot be used to provide total renal replacement therapy.

CAVHD/CVVHD employs diffusion to replace renal function. Blood flow rate through the extracorporeal circuit is similar to that for continuous hemofiltration. Dialysate is run through the filter at 1 to 2 L/hr. This results in a urea clearance of 17 to 34 mL/min, because the outflow dialysate is over 90% saturated with urea at the blood flow rates employed. One can futher increase the clearance of urea by combining hemofiltration with the continuous HD procedure.

Current technology permits any of these treatments to be implemented with the same machine. The tubing that directs the circuit path of blood, dialysate, and/or replacement fluid is merely confiured to provide the desired form of therapy. The current machines provide pressure monitors and pumps for the fluid pathway. However, unlike the machines used for intermittent HD that can take a dialysate concentrate and proportionate it with ultrapure water, the dialysate and the replacement fluid are not generated by the machine. Prepackaged dialysate and/or hemofiltration replacement fluid or fluid provided by the hospital pharmacy must be used. Some of the newer HD machines have the software that permits their use for CAVHD/CVVHD. The software permits dialysate to be proportioned so that the lower dialysate flow rates associated with the continuous therapies may be applied to the treatment.

Extended Daily Dialysis

An alternative form of therapy has been suggested for the management of the unstable patient with AKI.[128] A conventional HD machine is used to perform HD six to seven times a week at low blood and dialysate flows (blood flow of 100-200 mL/min and dialysate flow of 300 mL/min). Each treatment lasts from 6 to 8 hours. The urea clearance of extended daily dialysis (EDD) can exceed that of the continuous therapies ( Table 58-6 ).

TABLE 58-6   -- Clearance Comparisons of Extracorporeal Therapies










Filtrate (L/d)









Diaysate flow


















KUREA (mL/min)









From Conger J: Dialysis and related therapies. Semin Nephrol 11:533–540, 1998.

CAVH, continuous arteriovenous hemofiltration; CAVHD, continuous arteriovenous hemodialysis; CAVHDF, continuous arteriovenous hemodiafiltration; CVVH, continuous venovenous hemofiltration; CVVHD, continuous venovenous hemodialysis; CVVHF, continuous venovenous hemofiltration; IHD, intermittent hemodialysis; PD, peritoneal dialysis.





EDD is less labor intensive than the continuous therapies. The continuous therapies often require a 1 : 1 nurse-to-patient ratio owing to the need to monitor fluid balance and to change the replacement fluid. EDD employs the proportioning system of the HD machine and does not require the continuous presence of dialysis peronnel at the bedside. EDD also allows the patient to more easily be available for diagnostic test and therapeutic procedures, because the treatment lasts for only 6 to 8 hours. Indeed, delivered therapy via EDD may be superior to continuous therapy owing to fewer interruptions and clotting of the extracorporeal circuit associated with the former therapy.[128] EDD generally requires less heparin than the continuous therapies. In addition, because it is easliy implemented and similar to conventional intermittent HD, it is accepted by nursing staff and remains under control of nephrologists.

On-line Monitoring

Dialysis machines function as more than just dialysis delivery systems. Built-in monitors assess the physical and chemical characteristics of the dialysate, as noted previously, and record and store data ranging from patient blood pressure and heart rate to treatment-related parameters (blood and dialysate flows, arterial and venous pressures, temperature), medication data, measures of delivered dialysis dose, plasma volume, thermal energy loss, and even dialysis access recirculation. Computerized medical information systems have been linked with dialysis delivery systems to provide information networks that can control treatments at individual patient stations while maintaining information and treatment records for future use. Real-time information regarding treatment parameters and patient information can also be recorded with other monitors during dialysis treatments. It is now possible to integrate data, such as comparing present and past dialysis treatments, into a real-time display to help gauge therapy, change prescription and UF goals, and generate better immediate assessment of a patient's and a unit's overall status. [130] [131]Such on-line monitoring that allows sensors from the machine to change treatment parameters has been termed a biofeedback system. Automatic biofeedback systems have the potential to reduce adverse events such as hypotension, to monitor the state of the HD access, and to increase the efficiency of the HD treatment ( Table 58-7 ).

TABLE 58-7   -- On-Line Features of the Hemodialysis Machine



Blood pressure

Changes in ultrafiltration rate, sodium modeling

Plasma volume by hemoglobin

Changes in ultrafiltration rate, sodium modeling

Thermal energy loss/gain

Change in dialysate temperature

Transient change in dialysate sodium

Measurement of Kt/V, an index of dialysis adequacy

Transient change in hemoglobin

Hypotension, access blood flow, recirculation, cardiac output

Transient change in temperature

Access recirculation




The on-line measurement of changes in hemoglobin concentrations is an example of this application to measure access flow and function and to make accurate determinations of circulating blood volume during the dialytic procedure. Single- and dual-sensor systems using saline injections and sound velocity dilution calibration have been investigated as a method for accurately determining access flow during HD.[131] Similar efforts have led to noninvasive optical hematocrit monitoring to continually measure hematocrit during dialysis to better determine circulating blood volume.[132] The measurement is based upon the principle that as blood volume changes in one direction, hematocrit changes in the opposite direction. Thus, a rise in hematocrit is related to a decrease in plasma volume.

The measurement of on-line blood volume with these devices can identify patients who are not near their estimated dry weight. In 18% of HD patients, a less than 5% decrease in blood volume was noted during routine HD sessions. In subsequent treatments, increased volume was successfully removed without hypotensive episodes.[133] The patients were able to have intradialytic fluid removal intentionally increased by 47% (average 0.8 L). The change in blood volume can be determined noninvasively during HD with these devices.[134] With these devices, a critical hematocrit can be determined above which hypotension can be reliably predicted in 75% of patients.[135] Thus, these devices provide an added on-line safety measure to the treatment.

Dialysate Composition

One of the major aims of HD is the restoration of normal ion concentrations. As such, the levels of individual ions in the dialysate can be set to their desired plasma levels; however, in some instances dialysate levels are set for the diffusible fraction of the ion found in plasma. Dialysis solutions have undergone substantial changes since the inception of HD and are discussed in detail in a later section.

Water Treatment

HD patients are exposed to as much as 600 L of dialysis water a week—and to all its potential contaminants. Although water treatment systems used by dialysis centers produce high-quality water for safe dialysis, water treatment systems are susceptible to malfunction or to user error. Technical advances such as high-flux and high-efficiency dialysis, reuse, and bicarbonate dialysate have heightened awareness about water safety.

Hazards Associated with Dialysis Water

Numerous reports of patient injury or death have been linked to improperly treated or inadequately monitored water used for HD. High levels of aluminum sulfate in dialysate water have been linked to bone disease (osteomalacia and aplastic bone disease) and dialysis-associated encephalopathy (dialysis dementia). [137] [138] [139] Limiting aluminum levels in dialysate water to 10 mg/L has resulted in a continued decline in the incidence and case-fatality rate of dialysis dementia. [140] [141]

Chloramines, used as bactericidal agents in treatment of municipal water, denature hemoglobin by oxidation and inhibition of the hexose monophosphate shunt. Chloramine exposure during dialysis has been associated with hemoly-sis, Heinz-body hemolytic anemia, and methemoglobinemia. [142] [143] [144] Other compounds have adverse effects in dialysis patients. Sodium azide, used frequently with glycerine as a preservative for water treatment system ultrafilters, has been associated with hypotension.[144] Fluoride, even at the recommended level of 1 mg/L, can cause osteomalacia and bone disease as well as cardiac death. [146] [147] Excess calcium and magnesium in dialysate water have been linked to the “hard water syndrome”—a constellation of symptoms including nausea, vomiting, weakness, flushing, and fluctuations in blood pressure. [148] [149] Untoward effects have also been reported with nitrates (methemoglobinemia with cyanosis and hypertension), copper (hemolytic anemia), and zinc in excess concentrations in dialysate water. [150] [151] [152] [153] Formaldehyde toxicity, secondary to improper disinfectant use and leaching from sediment filters, has caused hemolytic anemia and death. [154] [155]

Essential Components of Water Purification

The efficiency of water treatment systems depends on the capacity of the system, the nature of the water supply, variations in quality of municipal water, and the quality of product water. Temperature-blending valves mix incoming hot and cold water to provide an optimum water temperature for downstream components. Most reverse osmosis (RO) membranes work with greatest efficacy at 77°F. Water temperatures below 77°F reduce the flow rate of the RO system, and water hotter than 100°F may damage RO membranes. Filters remove particulate matter from the water. Sand filters remove particles of 25 to 100 mm, cartridge filters extract particles of 1 to 100 mm, and submicron filters remove particles as small as 0.25 mm. In general, 5-mm filters are usually accepted as adequate protection for equipment and water treatment.

Water softeners, often sodium-containing cation-exchange resins, can remove calcium, magnesium, and other polyvalent cations from the feedwater. Because calcium and magnesium are removed from water in exchange for sodium, the amount of sodium released can be problematic. Removing calcium and magnesium prevents these ions from depositing on the RO system with resulting malfunction. Granular activated-carbon filters (GAC) absorb chlorine, chloramines, and other organic substances from the water. Carbon filters are porous with a high affinity for organic material. GAC can be contaminated with bacteria if they are not serviced properly or exchanged frequently.

RO applies high hydrostatic pressure to a solution across a semipermeable membrane to prepare a purified solvent. RO rejects 90% to 99% of monovalent and divalent ions and microbiologic contaminants, producing water safe for dialysis. An RO device is often used as pretreatment to deionization (DI), as an economic measure to provide longer service life for the DI system. Subsequent deionization of permeate (product) RO water is usually unnecessary. DI systems remove all types of cations and anions. The cation exchange resin exchanges hydrogen ions (H+) for other cations; the anion exchange resin exchanges hydroxyl ions (OH-) for other anions. DI efficacy is determined by measuring the resistivity of the effluent. Resistivity varies with temperature; therefore, resistivity monitors must be temperature compensated. When the DI system is exhausted, previously adsorbed ions can elute into the effluent, causing ion-related toxicities.[155]

Microbiology of Hemodialysis Systems

Water used by HD centers is usually obtained from the community water supply. Community water treatment can reduce bacteria and the concentration of endotoxins in the water; yet the dialysis water treatment system (apart from ultraviolet light) can still become contaminated with bacteria and endotoxins. [157] [158] The primary microbial contaminants in dialysis fluids are water bacteria, gram-negative bacteria, and nontuberculous mycobacteria.

Nontuberculous mycobacteria in particular are problematic. They do not produce endotoxins, but they are more resistant to germicides than gram-negative bacteria and they are infectious, especially in the setting of inadequately disinfected dialyzers. [159] [160] [161] [162] They can survive and multiply in RO-treated-water or DI water that contains little organic matter. Indeed, the Centers for Disease Control and Prevention (CDC) documented that nontuberculous mycobacteria were present in the water of 83% of dialysis centers surveyed in 1984.[159]

Sterilization destroys microorganisms, including highly resistant bacterial spores. Disinfection, in contrast, eliminates all but the highly resistant microorganisms. [163] [164] Disinfection can be high-level, intermediate, or low, depending on the germicidal activity. High-level disinfection inactivates all microorganisms except bacterial spores. Low-level disinfection reduces the bacterial population to a “safe” level. Water treatment system disinfection generally utilizes low-level disinfection. High-level disinfection is more often used for dialyzer reprocessing.

Pyrogenic Reactions During Hemodialysis

Pyrogenic reactions (PRs) often develop during or after dialysis treatment, with an incident rate of 0.5% to 12%. [151] [152] A PR can be defined as chills (or rigors) and/or fever (oral temperature >37.8°C [100°F]) in a previously afebrile patient with no recorded signs or symptoms of infection before dialysis. [154] [165] Hypotension is sometimes also included in the definition. Other signs of a PR are headache, myalgia, nausea, and vomiting. The symptoms usually begin 30 to 60 minutes into the dialysis treatment and stop shortly after, unless they are extreme. There appears to be little difference in PR rates between different HD modalities. [152] [165]

Three lines of evidence implicate endotoxin in the pathogenesis of PR: (1) antiendotoxin antibodies in dialysis patients; (2) Limulus lysate reactivity in plasma from patients experiencing PR; and (3) an association of PR with fluids contaminated with gram-negative bacteria. [166] [167] [168] [169] [170] It is unlikely that microorganisms cross intact dialyzer membranes because of the diameter of the pores. Rather, the endotoxins or other pyrogenic substances probably gain access to the patient's bloodstream across the dialysis membrane. [169] [171] Some of these substances are bacterial pyrogens released by gram-negative bacteria ( Table 58-8 ),[164] including lipopolysaccharides (LPS), its subunit, layer A, other LPS fragments, peptidoglycans, muramylpeptides, exotoxins, and exotoxin fragments.[171]

TABLE 58-8   -- Naturally Occurring Water Bacteria Commonly Found in Hemodialysis Systems

Gram-Negative Bacteria

Nontuberculous Mycobacteria


Mycobacterium chelonae


Mycobacterium fortuitum


Mycobacterium gordonae


Mycobacterium scrofulaceum


Mycobacterium avium


Mycobacterium abscessus


Mycobacterium intracellularis






Assays for determining the permeability of pyrogens include the Limulus amoebocyte lysate (LAL) assay, the mononuclear cell (MNC) assay, radiolabeled LPS fragments, and neutrophil activation. Many bacterial substances, like endotoxin fragments, are small enough to penetrate tight cellulosic membranes. These fragments go undetected in the LAL assay. Thus, measuring in vitro cytokine production by MNCs may be more sensitive and specific, allowing detection of these LMW substances. [173] [174] [175]

Severe PRs in HD patients appear to correlate with the extent of bacterial contamination in the dialysate. Recent studies have suggested that up to 35% of all water samples and 19% of all dialysate samples in the United States do not comply with Association for the Advancement of Medical Instrumentation (AAMI) standards (<200 colony-forming units [CFU]/ml water, 2000 CFU/mL in dialysate). Presumably, bacteria adhere to and grow in the dialysis tubing, releasing endotoxin and endotoxin fragments into the dialysate.

Changing dialysis practices have had an impact on PR. PRs have been reported with higher frequency in association with dialyzer reuse. Theoretically, use of RO and membrane integrity monitoring should lead to a decrease in the incidence of PR.[175] The use of bicarbonate and high-flux dialysis have been linked with a higher risk of PR. In dialysis units that used bicarbonate dialysis, a higher frequency of PR occurred only in centers that also performed high-flux dialysis. Centers that prepared their own bicarbonate dialysate were also more likely to report PR than centers that used commercially prepared bicarbonate dialysate. The method for preparing bicarbonate dialysate entails potential contamination.[176] Bicarbonate concentrates can support halotolerant endotoxin-producing, gram-negative organisms. As many as 105 to 106 CFU/mL can develop in liquid bicarbonate in as few as 10 days after dialysate preparation. Because of this, active quality assurance should be exercised to use liquid bicarbonate concentrate (LBC) as soon as possible after manufacture or receipt by the dialysis center. Tanks and distribution lines containing stored LBC should be disinfected at least twice weekly.

Dialyzer reuse practices have been associated with PR independent of high-flux dialyzer use. Manual dialyzer reprocessing has been associated with higher incidence of PR compared with automated reprocessing. Manual reprocessing can allow defects in dialyzer membranes to go undetected because testing for integrity of the membrane is generally not performed with this technique.

Several outbreaks of patient infection and PR have been reported in HD patients. [178] [179] [180] [181] Many of these involved substandard reprocessing or poor water quality. Inadequate mixing of germicide, or the use of a new germicide (e.g., chlorine dioxide) have been implicated in several of these outbreaks. [182] [183] Errors in the design and maintenance of a water treatment system were responsible for PR and gram-negative bacteremia in another center.[182] Damage to RO membranes contributed to this outbreak, leading to the recommendation of a thorough inspection for RO damage whenever the RO system removes less than 90% to 95% of total dissolved solids. Finally, though HD has been safely conducted outside the hospital or dialysis center setting, fatal endotoxemia occurred in dialysis patients at summer camp, illustrating the importance of dialysis water treatment systems in different environmental conditions.[183]

The formaldehyde content used for disinfection may also be important for PR frequency. Formaldehyde 2% does not effectively or reproducibly eradicate mycobacterial organisms within 36 hours.[184] If the concentration of formaldehyde is increased to 4%, mycobacteria cannot survive at room temperature beyond 24 hours.[185] However, increasing evidence indicates that lower concentrations of formaldehyde (e.g., 1%) can be effective if the dialyzers are kept at a temperature of 37°C to 40°C.[186]


The goal of HD in patients with ESRD is to restore the body's intracellular and extracellular fluid environment toward the body composition of healthy individuals with functioning kidneys to the extent possible. On a biophysical level, the use of HD as renal replacement therapy is accomplished via solute removal from the blood into the dialysate, as exemplified by intradialytic removal of potassium, urea, and phosphorous, as well as the addition of solute from the dialysate into the blood, as is exemplified by bicarbonate and calcium. An additional goal of the dialysis procedure is the elimination of excess water volume from the patient via UF. Thus, the prescription for an individual HD session must take in to account an examination and physiologic assessment of individual patient needs in order to achieve these goals. The variables in the HD prescription that may be manipulated by the physician on the basis of clinical assessment are listed in Table 58-9 .

TABLE 58-9   -- Components of the Dialysis Prescription

Dialyzer (membrane, configuration, surface area)


Blood flow rate

Dialysate flow rate

Ultrafiltration rate

Dialysate composition

Dialysate temperature


Intradialytic medications

Dialysis frequency




The separate components of the dialysis prescription are interrelated and must be integrated to meet the unique nature of the patient and clinical circumstances.

The principles that underlie the HD procedure are simple in practice. Blood and dialysate are circulated on opposite sides of a semipermeable membrane, thereby permitting the passage of solutes elevated as a consequence of renal failure but restricting the transfer of blood proteins and cellular elements. The device containing the semipermeable membrane is the hemodialyzer. Removal of water occurs by control of the hydrostatic pressure gradient across the semipermeable membrane and may be augmented by increasing the osmolality of the dialysate fluid.

Dialyzer Choice

In making a decision about the choice of dialyzer, the three most critical determinants are its capacity for solute clearance, capacity for UF or fluid removal, and the nature of dialyzer membrane interactions with components of the blood and their potential clinical sequelae (referred to as biocompatibility). The ideal HD membrane would have high clearance of LMW and middle-molecular-weight uremic toxins, negligible loss of vital solutes, and adequate UF in an effort to maximize efficiency and reduce adverse metabolic effects from the HD procedure. Additional characteristics of the ideal dialyzer would be a low blood volume compartment, beneficial biocompatibility effects, high reliability, and low cost. In evaluating dialyzer solute clearance characteristics, urea is the solute most often used owing to its relevance to kinetic models of dialysis adequacy. A detailed examination of the relationship of urea kinetic modeling to the delivery of adequate dialytic therapy is discussed later. In clinical practice, physicians typically rely on industry-derived determinations of in vitro dialyzer clearance of LMW and middle-molecular-weight solutes. Frequently Gibbs-Donnan effects, membrane adsorption of solute, protein binding of solute, and solute aggregation are not taken into account in determining in vitro dialyzer clearances.

A further complication in evaluating solute clearance by different dialyzers is the variable relationship between the diffusive and the convective clearance of a solute. As described mathematically in an earlier section, solutes that are larger than 300 Da will have relatively lower diffusive clearance values in comparison to smaller solutes such as urea and potassium. Clearance of larger solutes from the blood depends primarily on convective rather than diffusive clearance. Thus, in clinical circumstances in which large volumes of ultrafiltrate are generated, simple comparisons of diffusive solute characteristics (K values) of high-molecular-weight solutes can be misleading.

As discussed previously, the capacity for fluid removal by a dialyzer is described by its UF coefficient. Similar to the information provided for the diffusive clearance of a particular solute by a specific dialyzer, each dialyzer model also has a determined UF coefficient. Because these values are also typically derived in vitro, similar limitations describe their application in the in vivo situation. Therefore, it is not unusual for the UF coefficient in vivo to vary by 10% to 20% in either direction.

Virtually all of the commercial dialyzers available in the United States are configured as large cylinders packed with hollow fibers, known as hollow-fiber dialyzers. Blood flows within the hollow fibers while dialysate flows outside and around these fibers, generally in a countercurrent fashion whereby dialysate flow is in opposite direction to blood flow. Hollow-fiber dialyzers are generally noncompliant with fixed blood volumes. In parallel-plate dialyzers, a less frequently used physical configuration, multiple sheets of flat dialysis membranes are stacked in a layered configuration, with separation of blood and dialysate compartments. The blood compartment of parallel-plate dialyzers is more compliant and, therefore, varies more with transmembrane pressure. Parallel-plate dialyzers generally require a larger blood volume compartment than do hollow-fiber dialyzers. Hollow-fiber dialyzers range in blood volume compartment from about 50 to 150 mL.

The biomaterials of the hollow-fiber dialzyer will dictate its clearance characteristics, UF characteristics, and biocompatibility. Coupled with the development of dialysis delivery systems with UF controllers, the dialysis industry has developed dialyzers with a wide array of solute clearance and UF coefficients. Dialyzers with urea clearances of 50 to 250 mL/min (usually calculated for a blood flow of 200-300 mL/min), and UF coefficients of approximately 2 to 65 mL/mm Hg/hr are readily available.

As the HD procedure constitutes an extracorporeal circulation, the blood, out of necessity, must be removed from the vascular space and come into contact with dialysis membrane biomaterials. The interaction of soluble and cellular elements of the blood with the dialysis membrane can result in activation of protein-mediated humoral cascades, as well as activation of several types of circulating cells, many of which have been described previously.

An obvious humoral pathway frequently activated by the dialysis membrane is the clotting cascade. Thrombus formation from extracorporeal thrombogenesis occurs when thrombin adsorbs to the dialyzer membrane in an enzymatically active form, thereby becoming a site for platelet adhesion and further thrombin deposition.[187] The use of anticoagulants to reduce active thrombin deposition on dialysis membranes is discussed later. Attempts to limit dialyzer membrane thrombogenicity by ionic bonding of heparin to the dialyzer membrane surface have been complicated by a resulting tendency for more pronounced complement activation.[188]

An additional consideration in the prescription of the dialysis membrane is whether the intent is for the membrane to be reused. Dialyzer reprocessing for reuse of disposable dialyzers has been widely practiced in the United States, largely owing to financial constraints. Dialyzer reuse may be performed manually or with an automated rinsing device. Sterility during reprocessing is maintained by either the use of a chemical disinfectant (such as paracetic acid, glutaraldehyde, or formaldehyde) or via heat sterilization. After reprocessing, dialyzer adequacy is assessed indirectly by measuring the volume of the dialysis fiber bundle in the blood compartment (fiber bundle volume) and by pressurizing the dialyzer to evaluate the structural integrity of the fibers (pressure test). For a dialyzer to have acceptable reuse parameters, the fiber bundle volume must be greater than 80% of the initial value, the in vitro UF rate must be greater than 20% of the manufacturer's stated value, and the dialyzer should not leak at a pressure that is within 20% of the maximal operating pressure. The safety of dialyzer reuse practices has been closely scrutinized, and data concerning the practice have been controversial.[189] Data suggest that overall, facilities that reuse dialyzers have a risk-adjusted mortality similar to that of facilities that do not reuse dialyzers.[190] By 1997, reuse was practiced by as many as 82% of HD facilities in the United States.[191] Until recently, many nephrologists believed that patients are not placed at increased risk if strict infection control precautions and quality assurance measures are implemented in dialyzer reprocessing.[192] However, a large study involving over 70,000 HD patients who were initially using reprocessed dialyzers suggested a lower annual mortality rate when patients were switched to single use of dialyzers.[193] Reuse has been a much less popular practice in Europe and Japan. In addition, it is now argued that the medical justification for the reuse of dialyzers has been eliminated with the widespread use of dialyzers constructed with the synthetic membranes.[194] These dialyzers have improved biocompatibility and superior performance, characteristics that led to the virtual disappearance of the dialyzers containing cellulosic membranes in the United States. It is suggested that the best practice may be associated with the single use of dialyzers. [194] [195] A marked decline in the number of facilities practicing reuse in the United States has been noted in recent years.

Anticoagulation for Hemodialysis

Interaction of plasma with the dialysis membrane produces activation of the clotting cascade, characterized by the development of thrombosis in the extracorporeal circuit, thrombin deposition in dialyzer hollow fibers, and resulting dialyzer dysfunction.[195] Dialyzer thrombogenicity is determined by dialysis membrane composition, surface charge, surface area, and configuration. In addition, the rate of blood flow through the dailyzer, the UF rate prescribed (owing to hemoconcentration), and the length, diameter, and composition of blood lines all affect thrombogenicity. In addition, a number of patient-specific variables influence thrombogenicity and determine anticoagulation requirements during HD. These include acquired and inherited coagulopathies, neoplasia, malnutrition, hemoglobin concentration, and the presence or absence of CHF.

By far the most widely used anticoagulant for dialysis is heparin.[196] Heparin is easy to administer, has a low cost, and has a relatively short biologic half-life. For most patients, heparin is administered systemically during the dialysis procedure, either as a single bolus or incrementally during the dialysis treatment. For patients at high risk of bleeding, heparin is occasionally administered as regional anticoagulation, in which only the exracorporeal dialyzer circuit is anticoagulated by administering heparin into the arterial line and protamine into the venous line. [198] [199] [200] From a practical standpoint, the time constraints of dialysis limit the ability to monitor heparin efficacy through the partial thromboplastin time (PTT). In routine HD practice, the intensity of anticoagulation is not measured, whereas under some circumstances, the activated clotting time (ACT) is utilized. In this assay, whole blood is mixed with an activator of the extrinsic clotting cascade, and the time necessary for blood to first congeal is measured.

A simple method of heparin use is the systemic administration of 50 to 100 U/kg of heparin at the initiation of dialysis, frequently followed by a bolus of 100 U/hr. When ACT is being measured, the target ACT is approximately 50% above baseline values. In fractional anticoagulation, a smaller initial bolus of heparin is administered (10–50 U/kg), followed by an infusion of 500 to 1000 U/hr. Fractional heparinization can be utilized to achieve less intensive anticoagulation when the target ACT is maintained at 25% (fractional) or 15% (tight fractional), above the baseline value. These approaches are generally reserved for patients with a higher risk of bleeding complications.

Regional heparinization can be effective in preventing extracorporeal thrombogenesis with minimal systemic anticoagulation, but it is labor-intensive and prone to error when utilized by less experienced personnel. [199] [200] In regional anticoagulation, the extracorporeal circuit alone is anticoagulated by administering 500 to 750 U/hr into the arterial line (often with an initial 500-U bolus at the initiation of dialysis) and by the parallel administration of protamine into the venous line. The use of regional anticoagulation requires frequent checks of the ACT from the arterial and venous lines with adjustments of the heparin and protamine infusion rates to maintain the ACT for the patient at baseline while the ACT in the dialysis circuit is prolonged 10 seconds or longer. Because heparin has a longer half-life than protamine, additional protamine should be given at the end of the dialysis procedure to prevent a rebound heparin bleeding risk. [201] [202]

A variant of regional anticoagulation uses sodium citrate with dialysate containing no calcium, administered in the arterial line to bind calcium, as an important co-factor in the coagulation cascade. Coagulation of the circuit is thus inhibited. Before the blood is returned to the patient, a calcium infusion is administered via the venous line and the ability of the blood to clot is restored. Because of these difficulties, region anticoagulation is now rarely employed in the chronic dialysis unit. As an alternative to regional anticoagulation for patients at high bleeding risk, dialysis may be performed without any anticoagulation. Using the saline flush technique, HD is initiated at a high blood flow rate to reduce thrombogenicity, and the dialyzer is flushed every 15 to 60 minutes with 50 mL of saline. [203] [204] This technique is not likely to be successful in hypercoagulable patients, in circumstances where high blood flow rates are not attainable, and in dialysis where a high UF rate is required. Recently dialysate with citrate, a chelator of calcium, a co-factor required in multiple phases of the coagulation cascade, has been used to reduce the chance of clotting.

For patients at high risk for serious adverse events from hemorrhage, guidelines for anticoagulation must be based on comorbid conditions, and the risk of thrombosis of the extracorporeal circuit becomes a secondary consideration. Under these circumstances, the following guidelines are recommended:



Patients who are bleeding, are at significant risk for bleeding, have a baseline major thrombostatic defect, are within 7 days of a major operative procedure, or are within 14 days of intracranial surgery should undergo dialysis without heparin or by regional anticoagulation.



Patients who are within 72 hours of a biopsy of a visceral organ should undergo dialysis without heparin or by regional anticoagulation.



Patients who are more than 7 days past a major surgery or 72 hours past a biopsy can have dialysis by fractional heparinization. If they have previously received fractional heparinization, they can now be considered for systemic anticoagulation.



Patients with pericarditis should have dialysis without heparin or by regional anticoagulation.



Patients who have undergone minor surgical procedures within the previous 72 hours should have dialysis by fractional anticoagulation.



Patients anticipated to undergo a major surgical procedure within 8 hours of HD should undergo dialysis without heparin or with tight fractional anticoagulation. If they are within 8 hours of a minor procedure, fractional anticoagulation is appropriate.

Blood and Dialysate Flow

The clearance of a solute during the HD procedure may functionally be defined as the volumetric removal of the solute from the patient, as expressed mathematically earlier. Prescriptions of the blood flow and dialysate flow rates are critical elements of the dialysis prescription that can be altered to modify solute clearance. However, as blood and dialysate flow rates increase, resistance and turbulence within the dialyzer also increase. As a result, increases in nonlinear flow within hollow fibers occur, leading to a decline in the clearance per unit flow of blood or dialysate.[204] The resulting flow-limited mass transfer indicates that solute clearance will approach an asymptotic rate as blood flow or dialysate flow increases. The flow-limited mass transfer and membrane-limited mass transfer (defined by the specific dialyzer and the solute being measured) together determine clearance characteristics. A similar relationship is obtained for solute clearance and dialysate flow rate. Studies by Sigdell and Tersteegen[205] have demonstrated that the practical upper limit of effective dialysate flow is twice the blood flow rate, beyond which the gain in solute removal is minimal. As a consequence, the use of high dialysate flow rates to enhance solute removal should be confined to clinical circumstances in which the achievable blood flow rate is in excess of 300 mL/min. In addition to the prescribed blood and dialysate flow rates, convective mass transport that occurs with UF will also affect solute removal, thereby adding complexity to predictions of actual solute clearance. [207] [208] [209] [210]

In clinical practice, the efficacy of angioaccess may affect solute clearance obtained at a given prescribed blood flow rate. Access blood flow is a function of pressure and resistance. When blood is pumped out of the access into the dialyzer, a lower resistance circuit is created, which generally results in an increase in total access blood flow. The increased blood flow increases pressure in the venous drainage of the access during dialysis. Should venous outflow be restricted, there is an increased likelihood of backflow (termed recirculation) from the venous to the arterial side of the access. Backflow, or recirculation, is also facilitated by greater negative pressure at the arterial needle at higher blood pump speeds when there is impaired arterial flow. During recirculation, “dialyzed” blood reenters the dialytic circuit, thereby decreasing the efficiency of solute clearance.[210] Recirculation will also increase when dialysis needles are placed in close approximation within the dialysis access.

Recirculation is diagnosed when the concentration of a dialyzable solute in arterial line blood is lower than that of systemic blood, indicating that there has been mixing of dialyzed “venous” blood with blood entering the dialyzer. The fractional recirculation (R) is calculated using the equation: R = Cs - Ca/Cs - Cv. In this equation, Cs, Ca, and Cv are the concentrations (C) of the measured solute in systemic (s), arterial line (a), and venous (v) blood.

Recirculation has traditionally been measured by simultaneous measurement of a solute (usually blood urea nitrogen [BUN]) from the arterial line and from a peripheral blood source during the dialysis procedure. In traditional recirculation measurements, the source for the “systemic blood” sample was to draw blood from a vein in the contralateral arm. However, it is now recognized that this approach is inaccurate and tends to overestimate recirculation.[211] A major problem with this approach is that there may be disequilibrium between a peripheral venous sample and the arterial predialyzer sample. Whereas normally, there is a minimal difference in urea concentration between arterial and venous blood, during the dialysis procedure, blood leaving the dialyzer returns to the central veins, gets oxygenated in the pulmonary circulation, and is then delivered to the sys-temic circulation. Urea rapidly equilibrates across a gradient between the intracellular compartments and the arterial blood with a lower BUN, leading to a higher peripheral BUN concentration compared with the arterial BUN concentration. The resulting AV disequilibrium has also been termed cardiopulmonary recirculation and will result in an overestimation of recirculation.[212] Cardiopulmonary recirculation is increased with high efficiency dialysis (owing to more blood with lower BUN concentration entering the circuit) and in low cardiac output states (in which dialyzed blood constitutes a greater proportion of the cardiac output). Another form of urea compartmentalization termed venovenous disequilibrium occurs when peripheral vasoconstriction leads to a decrease in blood flow and, thus, a lower amount of total urea removal from the tissue bed. [214] [215] In principal, obtaining a peripheral arterial sample for measuring systemic urea concentration would overcome the inherent problems of AV and venovenous disequilibrium but is not practical during the HD procedure setting.

To more accurately measure recirculation, newer techniques have been developed that do not rely on withdrawal of blood for sampling solute concentrations. Instead, an indicator is infused, and measurement of disappearance of the indicator and lack of reappearance on the arterial side can be used to estimate recirculation.[215] Saline has been used as an indicator with monitoring of hematocrit changes. Similarly, a bolus of cold fluid can be used with an accurate blood temperature monitor. Studies using indicator techniques demonstrate that recirculation is rare unless the blood pump is set at a higher rate of speed than vascular access blood flow.[216]

Dialysis Time

The clearance of any of a solute, such as urea, can be increased by lengthening the dialysis treatment. Because the typical dialysis prescription often emphasizes optimal blood and dialysate flows and the selection of dialyzers with large mass transfer coefficient characteristics, the duration of dialysis is often the sole variable that can be used to augment solute clearance during an individual dialysis session. However, because diffusive solute clearance depends on solute concentration on the blood side, the efficiency of solute removal declines over the course of the dialysis procedure. Thus, from the standpoint of total solute removal, there are frequently “diminishing returns” in increasing the length of the dialysis procedure. The relationship of dialysis time to dialysis adequacy is further discussed in the section on dialysis adequacy.

The duration of the dialysis procedure may also be important in achieving adequate volume homeostasis. A longer duration of the dialysis procedure allows for a lower net UF rate per hour for a given targeted UF goal over the course of the procedure. This, in turn, may result in fewer intradialytic symptoms such as hypotension and cramping (see section on complications of dialysis). Data from Charra and colleagues in Tassin, France,[217] indicate that the use of long dialysis with slow rates of UF coupled with the use of a low sodium diet allow for an excellent volume homeostasis and blood pressure control, frequently without the use of antihypertensive medications. Patients undergoing long HD treatments with slow UF rates are reported to have excellent long-term survival.[218]

Dialysate Composition

When blood comes into contact with dialysate across a semipermeable dialysis membrane, a bidirectional diffusive process takes place by which solutes tend to reach similar concentrations on both sides of the dialysis membrane. In HD, a countercurrent flow configuration is utilized for blood and dialysate flow in order to maintain concentration gradient as a driving force for solute transport throughout the length of the dialysis membrane. Thus, the composition of dialysate is crucial to attaining the desired blood purification in order to achieve body fluid and electrolyte homeostasis. Physical and microbiologic characteristics of dialysate are also critical to the dialysis procedure. In addition to influencing the final concentration of blood solute at the end of the dialysis procedure, dialysate composition can influence intermediary protein, carbohydrate, and lipid metabolism and affect systemic vasomotor tone, cardiac contractility and rhythm, pulmonary gas exchange, and bone turnover. Thus, the selection of dialysis solute concentrations is a critical component of the dialysis procedure.


Sodium concentration is the major determinant of tonicity of extracellular fluids. Because sodium readily crosses dialysis membranes, sodium dialysate concentration plays a crucial role in determining cardiovascular stability during HD. Historically, the dialysate sodium concentration was maintained at hyponatric levels (130–135 mEq/L) in order to favor diffusive sodium loss during the dialysis procedure. [220] [221] [222] The use of hyponatric dialysate prevented interdialytic hypertension, exaggerated thirst, and excessive interdialytic weight gain.[222] However, the use of hyponatric dialysate, by inducing an interdialytic decline in plasma osmolality, favors fluid shifts from the extracellular to the intracellular space, thereby exacerbating the plasma volume depleting effects of HD. As a result, the intravascular space becomes dehydrated while the intracellular space became overhydrated. The osomolar changes and fluid shifts result in a high incidence of dialysis disequilibrium, characterized by headaches, nausea, vomiting, and in severe cases, seizures. [224] [225] In addition, interdialytic hypotension and cramps frequently result from the decline in intravascular volume. As a consequence, hyponatric dialysate has largely been abandoned with the result being a reduction in interdialytic hemodynamic instability and an improvement in interdialytic patient well-being. Thus, today it is “standard” to have a dialysate sodium concentration similar to plasma sodium concentration.

In an effort to further reduce dialysis disequilibrium, intradialytic hypotension, and dialysis-related symptomatology, particularly after the introduction of high-efficiency HD, the prescription of high-sodium dialysate has become a common practice. [226] [227] [228] Unfortunately, an increase in the dialy-sate sodium concentration frequently results in polydipsia, increased interdialytic weight gain, and increased interdialytic hypertension, thereby offsetting the beneficial effects of increased intradialytic hemodynamic stability.[228] Currently, the influence of interdialytic weight gain on blood pressure determination is controversial. In some studies, a positive correlation has been identified between intradialytic weight gain and blood pressure, whereas other studies have shown no relationship.

In an effort to achieve greater intradialytic hemodynamic stability using higher dialysate sodium activity while minimizing potential effects on interdialytic hypertension and weight gain, a number of studies have evaluated the strategy of varying the dialysate sodium concentration over the course of a dialysis session using variable dilution proportioning systems. These techniques, described as sodium modeling or sodium ramping, have been espoused as a means of individualizing the HD session to optimize blood pressure support. [230] [231] Sodium modeling is often performed either in a step fashion, in which the initial dialysate sodium concentration is greater than or equal to 145 mEq/L and during the second half of the dialysis session is abruptly reduced, or reduced as a linear gradient over the course of the dialysis session (linear sodium modeling).

Although sodium modeling can reduce the frequency of hypotension during dialysis, it is unclear whether this technique offers any advantage over a fixed dialysate sodium concentration of 140 to 145 mEq/L. [231] [232] Furthermore, it has been suggested that plasma sodium concentrations may be high in patients who undergo sodium modeling immediately postdialysis, resulting in increased interdialytic weight gains.[232]


Unlike sodium, which is largely distributed in the extracellular space, only 1% to 2% of the 3000 to 3500 mEq of potassium is present in the extracellular space. In patients with ESRD, potassium tends to accumulate in the plasma during the interdialytic interval, and life-threatening plasma potassium concentrations can result. Thus, dialysis is important in the maintenance of a normal or near-normal serum potassium concentration ESRD patient, and the removal of excess potassium is achieved by use of a dialysate potassium concentration lower than that plasma concentration.

The rate of potassium removal during dialysis is largely a function of the predialysis potassium concentration. However, the flux of potassium from the intracellular compartment to the extracellular space is frequently not in equilibrium with the mass transfer of potassium across the dialysis membrane to the dialysate compartment.[233] After the completion of a standard HD session, there is an increase in the plasma potassium concentration of approximately 30% over 4 to 5 hours owing to ongoing movement of potassium from the intracellular to the extracellular space.[234]

The efficacy of intradialytic potassium removal is highly variable, difficult to predict, and influenced by dialysis-specific and patient-specific factors. In a study that controlled dialysis-specific components of the dialysis procedure (dialysate composition, dialyzer type and surface area, blood and dialysate flow rate, and duration of dialysis), interpatient potassium removal varied by approximately 70%, whereas intrapatient potassium removal varied by 20% with the same dialysis prescription.[235] During HD, approximately 70% of the removed potassium is derived from the intracellular compartment. However, the volume of distribution of potassium is not constant. Paradoxically, the greater the total body potassium content, the lower the volume of distribution.[234] As a consequence, the fractional decline in the plasma potassium concentration during HD will be proportionately greater when there is a higher predialysis potassium concentration. Furthermore, most of the cardiac morbidity that arises from dialysate potassium concentration occurs during the first half of the dialysis session.[236] It is the rapidity of the fall in plasma potassium concentration rather than the absolute plasma potassium level that determines the risk of interdialytic cardiac arrhythmias.

As a further complication in attempting to model potassium removal during dialysis, the movement of potassium between the intracellular and the extracellular space is partially controlled by a number of factors that are simultaneously modified during the dialysis procedure. The transcellular distribution of potassium is influenced by plasma hydrogen ion concentration, as extracellular alkalosis shifts potassium into cells, whereas acidosis causes potassium to efflux from cells. Hyperinsulinemia promotes cellular potassium uptake, as do β-adrenergic catecholamines. Changes in plasma tonicity can also affect the transcellular distribution of potassium.

From the previous discussion, it is apparent that the selection of a dialysate potassium concentration is empirical and guided by patient-specific factors. Generally, a dialysate potassium concentration of 1 to 3 mEq/L is used in most patients. Patients who have excessive potassium loads from diet, medications, hemolysis, tissue breakdown, or gastrointestinal bleeding may require a lower dialysate potassium. Low dialysate potassium concentrations should be used with caution, however, as an analysis has found an association between the use of low dialysate potassium and sudden cardiac death in outpatient HD patients.[237]


Patients with ESRD are susceptible to hyperphosphatemia, hypocalcemia, secondary and tertiary hyperparathyroidism, and hypovitaminosis D. As a consequence, positive intradialytic calcium balance may be desired as an adjunct therapy for control of metabolic bone disease. In the setting of renal failure, over 60% of plasma calcium is not protein-bound and is thus capable of diffusible equilibrium during dialysis.[238] Assuming an additional conductance of calcium across the dialysis membrane owing to convective losses, a dialysis calcium concentration of 3.5 mEq/L (7.0 mg/dL) is necessary to prevent intradialytic calcium losses.[239] Thus, for many years, a dialysate calcium concentration of 3.5 mEq/L was relatively standard. However, with the use of calcium-containing salts and phosphorous binders and the aggressive use of vitamin D analogs, many dialysis facilities now employ a standard dialysate calcium concentration of 2.5 to 3.0 mEq/L in an effort to prevent interdialytic hypercalcemia.

The dialysate calcium concentration may also affect hemodynamic stability during the HD procedure.[240] Left ventricular contractility has been demonstrated to be proportional to serum ionized calcium concentration during HD.[225] Several prospective crossover trials have demonstrated higher interdialytic mean arterial pressure with higher dialysate calcium concentrations. [242] [243] An increase in intradialytic electrocardiographic QT dispersion can also be reduced by increasing the dialysate calcium concentration.[243]


Similar to potassium, the serum magnesium concentration is a poor determinant of total body magnesium stores. Only approximately 1% of total body magnesium content is present in the extracellular fluid, and only 60% of extracellular magnesium is free and diffusible.[244] Magnesium flux that occurs during HD is difficult to predict, despite knowledge of the serum and dialysate magnesium concentrations. Thus, the ideal serum magnesium concentration in patients with ESRD and the appropriate dialysate magnesium concentration are unresolved. Many centers use a dialysate magnesium concentration of 1 mEq/L.


Correction of uremic metabolic acidosis is a goal of the HD procedure. HD therapy cannot remove large quantities of free hydrogen ion, because the low concentration in the blood prevents the development of a large gradient to support mass transfer. As hydrogen ions are produced, they are rapidly buffered by plasma bicarbonate and other body buffers. Thus, in HD, correction of acidosis is largely achieved by using a dialysate with a higher concentration of alkaline equivalents than are present in the blood, promoting flux of base from the dialysate into the blood. Base transfer across the dialysis membrane has been achieved using either bicarbonate- or acetate-containing dialysate.

Historically, Kolff and colleagues[245] and later Scribner[2a] used dialysate bicarbonate as a buffer, bubbling CO2 through the dialysate to lower pH in order to prevent precipitation of calcium and magnesium salts. Because of the instability of bicarbonate in aqueous solution at neutral pH in the presence of divalent cations, Mion and associates in 1964[246] introduced acetate as a base equivalent in dialysate. Acetate-containing dialysate was biochemically more stable and also avoided frequent bacterial contamination problems that result when bicarbonate is used as a buffer. Acetate-containing dialysate solutions became the clinical standard of practice world-wide for more than 20 years. However, in the 1980s, the introduction of high-efficiency dialysis led to reports of cardiovascular instability and intradialytic hypotension owing to the slow conversion of acetate into bicarbonate with acetate dialysate. Under these circumstances, acetate accumulation contributes to nausea, vomiting, headache, fatigue, decreased myocardial contractility, peripheral vasodilatation, and arterial hypoxemia. [223] [230] [231] [232] [248] Adverse metabolic and hemodynamic effects associated with acetate dialysate have led to its virtual replacement as a buffer with bicarbonate.

The introduction of single-patient proportioning systems that permit mixing of two separate dialysate concentrations (separately containing sodium bicarbonate and an acid concentrate with divalent cations) close to the entry point of the final dialysate into the dialyzer allowed the widespread reintroduction of bicarbonate as a dialysate buffer in the late 1970s and 1980s. A small amount of acetic acid (3–6 mEq/L) is present in the acid concentrate and serves to titrate some of the bicarbonate to carbonic acid and carbon dioxide, thereby controlling the pH of the final dialysate. Dialysate bicarbonate concentrations of 30 to 35 mEq/L are now commonly used.


Chloride is the major anion in dialysate. Dialysate chloride concentration is defined by the prescribed concentration of cations and anionic buffers in the dialysate so as to maintain electrical neutrality.


In the early 1960s, high glucose concentrations in dialysis fluid were used to provide osmotic pressure for water removal. However, advances in hydraulic UF and the demonstration that high dialysate glucose (>320 mg/dL) increased the risk for hyperosmolar syndrome, postdialysis hyperglycemia, and hyponatremia rendered the use of high dialysate glucose obsolete.[248] Contemporary dialysis fluids range from glucose-free to slightly hyperglycemic (up to 200 mg/dL).[249] Most non-insulin-dependent diabetic patients tolerate dialysis with glucose-free dialysate well, despite losing 25 to 30 g of glucose across the dialyzer. However, this glucose loss may potentiate hypoglycemia and adversely affect HD catabolism, raising levels of free amino acids during dialysis, and increasing the intradialytic protein catabolic rate. [83] [84] [85] [86] [94] Ketogenesis and gluconeogenesis are usually sufficient to maintain serum glucose in the physiologic range despite reductions in plasma insulin, lactate, and pyruvate. By contrast, physiologic dialysate glucose (200 mg/dL) has few adverse effects, aside from aggravating hypertriglyceridemia.[94]Dialysate glucose can affect potassium removal, the risk of dialysis disequilibrium syndrome, and postdialysis fatigue. In general, an optimal dialysate glucose concentration is 100 to 200 mg/dL for most patients. However, in diabetic patients, insulin doses may require adjustment to account for this dialysis-imposed “glucose clamp” in which levels of plasma glucose may be kept constant during dialysis owing to the concentration in the dialysate.

Dialysate Temperature

Although not strictly reflecting dialysate composition, prescription of the dialysate temperature can be an important component of the dialysis prescription. Dialysate temperature is generally maintained between 36.5°C and 38°C at the inlet of the dialyzer. Heat exchangers and temperature monitors are built into dialysis machines. In cases in which temperature monitor failure has occurred, severe hemolysis has been reported. [251] [252]

Body temperature is determined by the balance between heat production through metabolism and heat losses from the body surface and through respiration. Although HD does not lead to direct heat transfer from the extracorporeal circulation to the patient, patient temperatures do increase during HD with conventional dialysate temperatures of 37°C. An increase in body temperature during dialysis can occur owing to an increase in heat generation as a response to pyrogens in the dialysate. In addition, UF-induced volume contraction during HD results in peripheral vasoconstriction, which limits peripheral heat loss and raises core body temperature. Eventually, there is a reflex dilatation of peripheral blood vessels, which allows heat escape but also reduces peripheral vascular resistance, resulting in an intradialytic fall in blood pressure. This has led to the suggestion (originally by Maggiore) that lowering dialysate solution temperature may permit increased hemodynamic stability in hypotension-prone dialysis patients.[252] Numerous studies now confirm that dialysate temperature is an important determinant of intradialytic blood pressure.[254] [255] [256] [257] [258] [259] [260] Lower dialysate temperature may also increase cardiac contractility, improve oxygenation, increase venous tone, and reduce complement activation during dialysis.

Ultrafiltration Rate

In addition to its use in comparing the UF performance of individual dialyzers, the UF coefficient is used to calculate the quantity of pressure that must be exerted across the dialysis membrane (transmembrane pressure) to generate a given volume of ultrafiltrate per unit time during a single dialysis session. The net pressure across the dialyzer membrane is determined by combining the hydraulic pressure, osmotic pressure, and oncotic pressure across the membrane. Because the hydraulic pressure is substantially higher than either the osmotic or the oncotic pressure, the net pressure gradient is approximated by the difference between blood and dialysate hydraulic pressure. For most currently available dialyzers, the hydraulic pressure can be calculated from the arithmetic mean of the inlet and outlet pressures. The effective pressure required for achieving a particular fluid loss during dialysis is described as the TMP and is calculated by TMP = desired weight loss/(UF coefficient × dialysis time). The performance of UF during HD has been greatly simplified with the development of dialysis machines that possess UF control systems. UF with these machines is remarkably precise. Use of UF control systems is critical when HD is performed with high-efficiency or high-flux dialyzers owing to their massive UF capacities.

During HD, UF and diffusive clearance are typically performed simultaneously. However, it is possible to temporally segregate the two procedures via a modification of the dialysis procedure that has been described as sequential UF/clearance. [261] [262] Sequential UF/clearance is accomplished by first ultrafiltering to the desired ultrafiltrate volume removal, followed by the performance of diffusive clearance without UF. During the initial UF phase, dialysate is not circulated through the dialyzer, thus preventing diffusive clearance. During the second phase, net obligate UF losses are balanced by the infusion of saline. Sequential UF/clearance is designed to reduce intradialytic hemodynamic instability by separating intravascular volume losses due to UF from intravascular volume losses due to movement of fluid into the interstitium and intracellular spaces after a decrease in plasma osmolality. Sequential UF/clearance has the disadvantage of decreasing net diffusive clearance unless dialysis time is substantially increased. Recent data also suggest that the strategy of sequential UF/clearance is less effective in maintaining intradialytic hemodynamic stability than other maneuvers such as sodium modeling or the use of low-temperature dialysate.[262]

Prescription of the UF rate in maintenance HD patients is generally based on an assessment of estimated dry weight. Dry weight is currently defined as the lowest weight a patient can tolerate without the development of signs or symptoms of intravascular hypovolemia.[263] Tolerance of the UF rate during HD is largely determined by the rate of vascular refilling to ensure that intravascular volume depletion does not occur.[264] Monitoring of blood volume changes during dialysis may provide further insight into prescribing an UF rate, and blood volume changes have been associated with blood pressure changes, hypotensive events, and hydration status during HD.[265] Relative changes in circulating blood volume can be estimated in real time by continuously monitoring the hematocrit during dialysis.[266] The observed real-time changes in relative blood volume can be used to adjust UF rates in an effort to reduce interdialytic hypotensive events. However, to date there are limited data suggesting that this approach in and of itself can enhance the achievement of estimated dry weight while reducing intradialytic symptomatology.

An additional approach to balancing the need for intradialytic hemodynamic stability with the achievement of estimated dry weight involves the use of variable UF rates during the HD procedure, known as UF modeling. UF modeling has been introduced as an available component in some commercially available dialysis machines. At present, there are little or no data suggesting that the use of UF modeling can significantly reduce intradialytic complications.[267] However, several studies suggest that combining sodium and UF profiling may reduce the slope of the blood volume curve during dialysis and thereby significantly reduce HD-related symptoms.[268] Eventually, the development of biofeedback loops, whereby continuous variation in sodium and UF modeling is combined with assessment of blood volume and cardiac output changes, may significantly improve hemodynamic stability during dialysis.[269]


Dialysis is a life-saving treatment for more than 300,000 patients with ESRD in the United States and well over a million patients worldwide. Despite this enviable record extending over more than a quarter of a century of treatment, the annual mortality rate of dialysis patients in the United States since the mid-1990s remains in excess of 20%.[270] Multiple lines of evidence implicate inadequate dialysis prescriptions and the underdelivery of the prescribed dose of dialysis as central factors responsible for the high mortality. [272] [273] [274] [275] [276] [277] This section examines the evidence supporting the use of quantitative assessment of dialysis adequacy both as an outcome measure and as a method of improving the care of patients with ESRD.

Historical Perspective

It is instructive to review the evolution of the methods that have been employed to determine the adequacy of the dose of HD. One might pose a question regarding adequate dialytic therapy that is similar to the one asked by Dr. Henry Stubb centuries ago concerning the practice of blood transfusions: “What regulation shall we have for the operation? Shall a man transfuse [dialyze] he knows not what, to correct he knows not what, God knows how?”[277]

The maintenance of homeostasis and of fluid balance and the elimination of toxins generated from dietary protein catabolism and other sources are chief functions of the kidneys. The accumulations of toxins result in the manifestations of the uremic syndrome. In CKF, uremic symptoms are worsened by excessive protein intake and may be ameliorated by restriction of protein intake. This clearly indicates that nitrogenous compounds are central in the pathogenesis of uremia. However, a quandary exists as to which specific substance or combinations of compounds produce symptoms. This makes the task of assessing treatment adequacy by indexing the dose of dialysis to a simple plasma level or removal rate of particular compound difficult. Easily measured substances, such as urea or creatinine, are themselves not major toxins. Furthermore, the plasma levels of these substances are influenced by many factors beyond clearance by the artificial kidney. Thus, generation rates and removal rates and volume of distribution must be used to describe the fate of the substance being used to describe the efficiency of the treatment.

Low-Molecular-Weight Substances and Middle Molecules

The use of compounds like urea to judge adequacy rests on the assumption that the clearance rate of LMW solutes correlates with well-being. Conversely, the observation that the severity of peripheral neuropathy could be mitigated by long treatment sessions with the Kiil dialyzer, a relatively inefficient device with respect to urea but one with a large-surface-area membrane. Long dialysis time and large-surface-area membrane are characteristics that enhance the removal of higher-molecular-weight substances. The finding that these features improved neuropathy led to the square meter-hour hypothesis.[278] This hypothesis holds that solutes with molecular weights in the range of 300 to 12,000 Da, termed middle molecules, play a role in the pathogenesis of the uremic syndrome.[279]

The subject of middle molecules as uremic toxins has been reviewed.[280] Compounds resulting from protein catabolism as well as peptides such as parathormone (PTH) and β2-microglobulin are among the larger solutes that are retained in kidney failure. The latter two high-molecular-weight substances have been given roles as mediators of uremic toxicity. [282] [283] [284] [285] [286] [287] Vanholder[280] has suggested that LMW substances may also behave as middle molecules by virtue of their physical properties such as charge, steric configurations, or ability to bind to plasma proteins or because of their high generation rates. These properties result in a reduction in their clearance rate by dialysis that would not have been predicted by size alone. Candidates for this designation include methylguanidine, indoxyl sulfate, hippuric acid, and inorganic phosphate. Conversely, Teschan and the results of the National Cooperative Dialysis Study (NCDS) have suggested that an LMW compound, urea, can serve as a legitimate surrogate for uremic toxins. [288] [289]

The attributes of an ideal marker of the adequacy of dialysis that have been suggested by Vanholder[280] and are listed in Table 58-10 . No single marker meets all of the requirements. There are unresolved issues that preclude the use of middle molecules to determine adequacy.

TABLE 58-10   -- The Ideal Marker of Dialysis Adequacy

Retained in renal failure

Eliminated by dialysis

Proven dose-related toxicity

Generation and elimination representative of other toxins

Easily measured




The National Cooperative Dialysis Study

The NCDS was initiated in 1976. The study applied pharmacokinetic principles to urea concentrations as they varied during the intra- and interdialytic periods of the HD session.[289] For the purposes of the analysis, a single pool volume of distribution for urea was assumed. Developed by Gotch and Sargent, changes in serum urea concentrations are measured over time, so that “average” concentration of urea for the treatment session can be expressed: TACurea (timed average urea concentration). From the intradialytic curve, the index related to the elements of the dialysis treatment and the size of the patient or Kt/V can be calculated and from the interdialytic curve urea generation can be determined, as seen in Figure 58-14 .

FIGURE 58-14  The hemodialysis cycle and elements of kinetic modeling.


The NCDS was a multicenter prospective, randomized two-by-two factorial trial ( Table 58-11 ). The study participants were HD patients randomized to one of four groups based on short or long dialysis treatment times and high and low TACurea. Based on the design, groups I and III received a higher level of dialysis delivered over a longer or shorter time, respectively; groups II and IV received a lower level of dialysis delivered over a longer or shorter time, respectively. The goals were achieved by manipulation of dialyzer size and TD. If groups I and III have a good outcome, this would suggest that the dose of HD could be indexed to LMW compounds such as urea.

TABLE 58-11   -- Design of the National Cooperative Dialysis Study



Less Intensive

Long duration

Group I

Group II


TAC = 51.3 ± 1.1 mg/dL

TAC = 87 ± 1.4 mg/dL


TD = 269 min

TD = 271 min

Short duration

Group III

Group IV


TAC = 54.1 ± 1.1 mg/dL

TAC = 89.6 ± 1.2 mg/dL


TD = 199 min

TD = 194 min


TAC, timed average concentration; TD, time of dialysis.




Two measures of outcome were analyzed: subjects who withdrew from the study for medical reasons or death (F1), and those who withdrew from the study for hospitalization within the first 6 months of the experimental phase (F2). Of 160 randomized patients, approximately 50% completed the study protocol. Importantly, group IV (high TAC, short dialysis time) was discontinued before the study was completed because of excessive hospitalizations and medical withdrawal.

The TACurea, the index of dialysis adequacy used in the analysis of the primary outcome of the NCDS, was the best predictor of failure. A much weaker but statistically significant relation was also found for dialysis time: Short time was associated with a greater incidence of F2 failure. Not all adverse medical events or hospitalizations were those commonly associated with too little dialysis, such as episodes of volume overload, hyperkalemia, or pericarditis. Although only 3 patients died during the actual study, an additional 13 died during a 12-month follow-up after withdrawal from the study. Ten were assigned to groups II and IV. In many instances, these patients were returned to a higher level of therapy at the completion of the study, yet the adverse effects of what was shown to be an inadequate level of therapy were difficult to correct. The NCDS suggested that removal of small molecules strongly predicted morbidity and that urea kinetic modeling (determination of TAC) could be used to index the level of therapy delivered, despite the fact that urea does not fulfill all the criteria listed in Table 58-11 .

Subsequent analysis of the NCDS suggested that it would be informative to separately analyze the components of the dialysis cycle (see Figure 58-14 ). The dimensionless term, Kt/V, describes aspects directly related to the HD treatment factored by the volume of urea distribution in the patient.[290] Morbidity could be indexed to this term. The advantage of using Kt/V as a marker of adequacy is that it allows one to focus on the elements of the intradialytic period. This is the part of the dialysis cycle that is amenable to manipulation of the prescription: blood and dialysate flow, UF rate, size of the artificial kidney and dialysis time, in the case of HD, or dialysate volume in the case of peritoneal dialysis. The initial analysis of data from the NCDS indicated that a Kt/V greater than 0.8 was associated with a good outcome.

By design, the prescriptions in the NCDS were manipulated to achieve high or low TACurea goals for the study. The TACurea covers the entire dialysis cycle, and thus, it will be influenced by urea generation, an index of dietary protein intake in the stable HD patient (see Figure 58-1 ). Thus, in order for the TAC goal to be reached, the dialysis dose (Kt/V) was partially determined by the subjects' protein intake (urea generation). The implication of this design is that adequacy in the NCDS has been defined by Kt/V levels that have been interpreted in the context of protein intake of the subject ( Figure 58-15 ).

FIGURE 58-15  Relationship between Kt/V and protein intake in the National Cooperative Dialysis Study. K, urea clearance; t, treatment time; V, volume of urea distribution.


Although the initial interpretation from the NCDS suggested that there was little to be gained by increasing Kt/V to values beyond 0.8 to 1.0, subsequent analysis of the data indicates that the relationship between morbidity and Kt/V may be continuous, with improved outcome found at higher doses of Kt/V.[291]

Assessment of Dialysis Adequacy and Derivatives of the National Cooperative Dialysis Study Outcome Measures

Since the conclusion of the NCDS, the principles of urea kinetic modeling have been applied to the assessment of the adequacy of both HD and peritoneal dialysis. The impetus behind this practice has suggested that improving the clearance of LMW substances would favorably affect the unacceptably high mortality rate experienced by dialysis patients in the United States. Subsequently, three retrospective and observational studies provided further evidence that patient outcome correlated with the dose of HD as measured by Kt/V. [272] [276] [277] A 5% and 7% decrease in relative risk of mortality can be demonstrated for each 0.1 increase in Kt/V in nondiabetics and diabetics, respectively.[277] In all of these studies, patient survival improved as Kt/V was increased.

The benefits of high levels of HD delivered by conventional cellulosic membranes over extremely long sessions of 6 to 8 hours has resulted in remarkable survival statistics.[292] In patients achieving a mean Kt/V of 1.67 delivered in this fashion, the 10-year survival ranges from 88% for patients initiating HD at 35 years of age or younger to 64% for patients older than 65 years of age. The 15-year survival for all patients in this dialysis center is 55%. These studies along with data gathered from other dialysis registries provide strong circumstantial evidence that Kt/Vurea, an index of removal of LMW substances, is a predictor of mortality in HD.

Dialysis Adequacy: Applicability and Limitations of the National Cooperative Dialysis Study

The subjects eligible to enter the NCDS differed substantially from the current HD population. They were younger, compliant, and free of comorbid conditions. Only 20% of the current HD population would meet the NCDS entry criteria.[293] The follow-up period of the study was short. Thus, the influence of factors such as age and comorbid conditions that occur over a longer period of time could not be assessed by the NCDS. The NCDS was able to define a dose of dialysis below which an unacceptable number of complications occur, but it was not designed to define an optimal level of dialysis beyond which no further improvement was realized.

Technical advances in dialysis delivery have also occurred since the NCDS was completed. Thus, the methods employed in that study may not be generalizable to our current practice. Thus, biocompatible, synthetic high-efficiency or high-flux dialyzers have largely replaced low-flux cellulosic dialyzers used in the NCDS. Acetate-buffered dialysate has been universally replaced with bicarbonate-buffered dialysate. HD treatments have been more stable with the use of dialysate with higher sodium concentration and with volumetric machines capable of changing UF rate and sodium concentration during the treatment.

Increasing evidence points to nutrition as an important independent determinant of survival in dialysis patients.[294] In the NCDS, the level of protein intake influenced the dialysis dose: By design, the levels of dialysis determined by Kt/V and protein intake were interdependent rather than independent variables.

The dose of dialysis in the study was adjusted by alterations in blood flow, the dialyzer size, and length of time. Indeed, time was an independent variable in the design. Dialyzer surface area and time will affect middle molecule clearance as well as the clearance of LMW substances. Longer dialysis time and dialyzers of greater surface areas introduced a confounding variable, enhanced middle molecule clearance, when these techniques were used to increase Kt/Vurea. Group IV (short time, low Kt/V) fared most poorly of all groups, and it remains possible that reduced middle molecule clearance played a role in this outcome.

The effect of the length of the dialysis treatment on outcome was not determined in the NCDS, and it is yet to be included as an independent variable in any study of the dialysis prescription to date. As mentioned previously, the issue remains unsettled because time was partially confounded with the effects of Kt/Vurea: An increase or a decrease in time was one of the methods used to change the TAC to meet the goals to which the study participants were assigned. The current practice of using large-surface-area dialyzers at blood flows higher than 400 mL/min and at dialysate flows of 600 to 800 mL/min allows a Kt/V to be reached far in excess of values in NCDS group I or III in many patients without increasing time. Indeed, many small- to medium-sized individuals reach high levels of dialysis in times that are shorter than those employed in group III.

The HEMO Study

Although multiple lines of evidence indicate that kinetic modeling is an important index of dialysis adequacy and that the degree of removal of LMW substances correlates with survival, these relationships have not been proved by prospective studies. It is important to confirm the impression of the observational studies that high doses of dialysis will have a favorable impact on patient outcome. The time, effort, and costs associated with providing high doses of dialysis are substantial. The DOQI guidelines have already made recommendations regarding a dose of dialysis below which poor patient outcomes are likely to occur. What is completely lacking is prospective data regarding the effects of increasing Kt/V to very high levels. Consequently, the NIH initiated a multicenter, prospective, randomized trial to assess the impact of the dialysis prescription on morbidity and mortality of HD patients. [296] [297]

The study consisted of a two-by-two factorial design that assessed the effect of HD dose and membrane flux on outcome. In this study, an equilibrated Kt/V of 1.05 was compared with an equilibrated Kt/V of 1.45, comparable on average to single-pool Kt/V of 1.25 and 1.65, respectively. In addition, the effect on mortality and morbidity of high-flux versus low-flux dialyzers was compared. All-cause mortality is the primary outcome and morbidity assessed from hospitalization; time to hospitalization for cardiovascular and infectious causes and time to a decline in serum albumin concentration are secondary outcome measures. The design called for a concurrent sample size of 900 patients from 15 clinical centers with replacement of those participants who died or dropped out.

In the group of control subjects randomized to usual HD dose arm, the achieved single-pool Kt/V was 1.32 ± 0.09, and the achieved equilibrated Kt/V was 1.16 ± 0.08; in the subjects randomized to the high-dose arm, the achieved values were 1.71 ± 0.11, and 1.53 ± 0.09, respectively. Dialyzer flux, based on the clearance of β2-microglobulin clearance, was 3 ± 7 mL/min in the low-flux group and 34 ± 11 mL/min in the high-flux group. The primary outcome, death from any cause, was not influenced by the dialysis dose or the dialyzer flux assignment: The relative risk of death in the high-dose group compared with the usual-dose group was 0.96 (95% confidence interval, 0.84–1.10; P = .53), and the relative risk of death in the high-flux group compared with the low-flux group was 0.92 (95% confidence interval, 0.81–1.06; P = .23). The main secondary outcomes including first hospitalization for cardiac causes or infection or all-cause mortality, decline in albumin or all-cause mortality, and all hospitalization not related to vascular access problems also did not differ between either the dose or the flux groups.[295]

The effects of HD dose and flux intervention were also adjusted for a series of prespecified baseline factors of age, gender, race, years on dialysis, presence or absence of diabetes, score for coexisting conditions excluding diabetes, and albumin level. For the entire study population of 1846 randomized patients, all of the prespecified covariates were independent predictors of death. Thus, older age (per 10-year increment), male gender, white race, presence of diabetes, longer time on dialysis (per 1-year increment), higher baseline Index of Coexisting Disease, and lower baseline albumin level were associated with higher mortality in all patients, independently of their randomization.

When subgroup analysis was performed, based on these prespecified baseline factors, interactions with the primary treatment interventions were detected. Females randomized to the high-dose HD group had a lower risk of mortality. Subjects with a longer length of time on HD at entry into the study had a lower mortality rate if they were allocated to the high-flux arm of the study. The reasons for these subgroup outcomes are not completely clear. For instance, lower body weight versus dose does not explain the improved outcome among female subjects. When length of time in the study is added to the baseline length of time on dialysis, the benefit of flux disappears.

Although the data from the subgroup analyses are interesting and suggest further avenues for inquiry, the primary results indicate that within the conventional schedule of thrice-weekly HD, neither an increased dose of dialysis nor the use of a high-flux membrane improves survival, reduces the hospitalization rate, or maintains a higher serum albumin level compared with a standard HD dose and the use of high-flux membranes. The results from the HEMO Study should be reassuring to nephrologists that if the current DOQI guidelines are achieved, adequate therapy is being delivered to their patients who are receiving thrice-weekly therapy. However, the study should not be interpreted as sanctioning a minimal dose of HD. It is prudent to provide a margin above a minimal dose in order to protect the patient from receiving less dialysis than intended owing to factors that result in lower than intended blood flows, poor blood pump calibration, poor access function, or premature treatment termination. In particular, the HEMO Study should not be used as a justification to reduce HD time. Time was not an independent intervention in this study. Dialysis time itself is an important factor in blood pressure control and in avoiding hypotension in patients.[217] Thus, one cannot conclude from the HEMO Study that minimizing time while maintaining an “acceptable” Kt/V is justified. Furthermore, in the HEMO Study, the dialysis times for the majority of patients ranged between 3 and 5 hours. This is a relatively narrow range of time. As noted in Table 58-2 , dialysis time is a major determinant of the clearance of high-molecular-weight substances. Thus, the time interval may not have been sufficiently long enough to allow a difference between high- and low-flux dialyzers to be clearly realized. It is also apparent from the HEMO Study that the 22% gross mortality rate currently experienced by HD patients in the United States will not be affected by changing the dose of thrice-weekly treatments


The HEMO Study can be applied only to thrice-weekly HD, as currently practiced in the United States. There is increasing interest in different treatments times and frequency designed to improve outcome. Although the removal of LMW substances has been validated as a method to index dose of dialysis, it is also possible that improved clearance of larger substances can affect mortality and morbidity. [279] [280] [281] The failure of the HEMO Study to show a benefit of high-flux dialyzers does not conclusively eliminate the potential benefits of the removal of high-molecular-weight substances. The removal of high-molecular-weight substances is dependent on both porosity and the length of the dialysis treatment. Thus, one can argue that for the full benefit of these membranes to be realized, longer treatment times than those employed in the HEMO Study are required.

Factors that influence the HD treatment include patient acceptance, the need for delivery of an adequate treatment (dialysis time, blood and dialysate flow, dialyzer size, and frequency), and economics.[297] The evolution of current reimbursement has been a major stimulus for the movement away from earlier HD regimens consisting of greater than thrice-weekly treatments. However, observational data have suggested a beneficial effect on survival of increasing LMW substances, as measured by Kt/V or urea reduction ratio (URR). [272] [276] [277] In addition, the remarkable patient survival of patients undergoing conventional, thrice-weekly, low-flux HD but for 8 hours per session in Tassin, France, has been reported.[298] Efforts to improve survival and rehabilitation of patients with ESRD have led to a renewed interest in alternative HD schedules. The alternative schedules result in treatments that are longer and/or more frequent than standard 2.5 to 5 hours per session, thrice-weekly intermittent HD (IHD) that is widely practiced.

Definition of Daily Dialysis

There are several alternative methods to conventional IHD, defined previously. In Tassin, slow, long HD is also thrice-weekly, with blood flow rates of 200 to 250 mL/min and TD of 6 to 8 hours. Short daily HD (DHD) is characterized by five to sev en treatments per week, each lasting 1.5 to 2.5 hours, and using high-flux biocompatible membranes at blood flow rates greater than 400 mL/min and dialysate flow rates of 500 to 800 mL/min. Nocturnal HD (NHD) is also performed five to seven times per week, each treatment lasting 6 to 8 hours, and using biocompatible membranes at blood flow rates of 200 to 300 mL/min and dialysate flows of 200 to 300 mL/min. The single-pool Kt/V values are 1.2 to 1.8 with conventional IHD, 1.6 to 1.8 with IHD as practiced in Tassin, 0.2 to 0.8 with DHD, and 0.9 to 1.2 with NHD.[299]

Daily Dialysis and Outcome

The experience of groups who have practiced long IHD (6–8 hr) or who have markedly increased the intensity of standard-length IHD by the introduction of high-efficiency dialyzers is an improvement in the nutritional status of the patients. Dietary protein intake (DPI) is reported to be as high 1.3 ± 0.42 g/kg/day and serum albumin is 4.2 ± 0.5 g/dL. In contrast, the DPI is 1.0 ± 0.3 g/kg/day and the serum albumin averages 3.8 ± 0.3 in patients undergoing conventional IHD.[298] Against this backdrop, the nutritional status of patients undergoing DHD can be compared.

Nocturnal Hemodialysis

Although the number of patients studied has been rather limited, multiple lines of evidence that compare the status of patients on conventional IHD with their status on NHD have been remarkably consistent. Neutron activation analysis, an extremely accurate method to measure total body nitrogen (TBN), demonstrates a significant increase in TBN in 18 of 24 patients after they were switched from IHD to NHD. The observation period spanned 12 to 30 months. The change in TBN was from 1.43 ± 0.38 kg to 1.89 ± 0.60 kg.[300]

In one study, initially involving five patients, 8 weeks after changing from IHD to NHD, significant increases in nitrogen intake, caloric intake, and sodium intake were noted. Protein catabolic rate (PCR) increased from 1.07 ± 0.12 g/kg/day to 1.27 ± 0.20 g/kg/day.[301] In studies of a small number of patients from two different groups, no differences were noted in albumin levels. The abnormal plasma and intracellular amino acid profiles found in patients receiving IHD were altered upon changing to NHD. After 1 year on NHD, total amino acid, essential and nonessential amino acid, and branched-chain amino acid levels increased significantly.[302] However, a number of aberrations persisted such as abnormal ratios of essential/nonessential amino acids, tyrosine/phenylalanine, and valine/glycine.

A remarkable and unprecedented feature of NHD is the change in the management of renal osteodystrophy and phosphate control.[303] NHD results in the removal of over 160 mmol of phosphate each week. This is more than double the removal seen with conventional IHD. This results in a serum phosphate of 6.0 mg/dL with IHD falling to 3.9 mg/dL, despite an increase in phosphate intake. All patients were able to discontinue the use of phosphate binders entirely. Indeed, some patients actually require the addition of phosphate to the dialysate.

Short Daily Hemodialysis

Experience with 10 patients currently enrolled in the DHD arm of an ongoing clinical trial, the London, Ontario Daily/Nocturnal Hemodialysis Study, indicates that this form of therapy is associated with a significant improvement in both PCR and serum albumin levels.[297] Thus, PCR increased from 1.0 g/kg/day to 1.7 g/kg/day, and albumin levels in-creased from 38.6 g/L to 40.8 g/L at the end of 18 months. No significant change was noted in a control group of patients undergoing IHD.

In another study of five patients, nutritional parameters were compared before and after a switch to DHD. Significant improvement was noted in urea, creatinine, total CO2 levels, albumin, and dry weight.[304] The weekly sum of Kt/V was unchanged, but the authors did not express dose as standard Kt/V (stdKt/V).

Outcome Studies

Nocturnal Hemodialysis

Much of what has been described concerning daily dialysis modalities involves changes in biochemical parameters, quality of life, response to erythropoietin, and dose of dialysis. To date, because these modalities are new and the number of patients enrolled is relatively few, there have been very few outcome studies of daily dialysis. One report describes the use of NHD in four young patients with growth retardation and failure to thrive on peritoneal dialysis with treatments lasting 7 to 8 hours, six times weekly.[305] NHD was performed from 5 to 55 months. Treatment was accepted by the patients and resulted in improved nutritional status and increased bone length and mineralization. Improved quality of life and the chance for catch-up growth were features of the treatment.

Nocturnal Hemodialysis and Short Daily Dialysis: Comparative Effects on Nutrition

The two daily dialysis regimens are not equivalent. NHD requires long treatment times of 7 to 8 hours at relatively low blood and dialysate flows, whereas DHD employs short treatment times of 1.5 to 2.5 hours at high blood flow rates. Single-treatment Kt/V with NHD is greater than with DHD, but stdKt/V based on LMW substance removal is similar. The removal of high-molecular-weight substances is greater with NHD. Whereas estimated dry weight and PCR increase with both modalities, an increased albumin levels is more often seen with DHD. The control of phosphate is better with NHD. NHD is better adapted to the home, whereas DHD could easily be performed at home or in-center. The clinical and economic practicality of applying each of these modalities to large groups of patients awaits a randomized clinical trial.


The use of Kt/Vurea in determining adequacy of dialysis is based on mathematical models and is supported by clinical experience. However, a number of paradoxical observations have led some to question the validity of Kt/Vurea as the best index of judging adequacy. One paradox is that the curve relating dialysis dose and survival is J shaped.[306] Low dialysis dose is associated with high mortality, and mortality declines with increased doses of dialysis, but mortality again trends upward at the highest levels of dialysis. A second observation is that survival of African Americans on dialysis is better than that of white Americans, despite the finding that the latter group generally receives a higher dose of dialysis. [308] [309] [310] It is important to note that these observations do not necessarily invalidate the practice of indexing adequacy against LMW substances. Rather, the issue is whether Kt/Vurea is the best measure of LMW solute removal.

A common feature that may explain these observations relates to patient size.[310] At the same K × t, the smaller individuals are more likely to receive higher Kt/Vurea than larger individuals because their urea volume is smaller. African Americans tend to have a body mass than greater than that of whites. [310] [312] [313] A low body mass is an independent risk factor for death in dialysis patients. [277] [313] [314] [315] [316] V, the urea volume, may be an independent variable of survival because it tends to vary directly with body mass. Thus, when the work of dialysis, K × t, is divided by V, a parameter that may also correlate with survival, in the computation of Kt/V, “these elements may offset each other, producing a complex quantity that does not reflect a true relationship between dialysis exposure and clinical outcome.”[310] Proponents of this concept have demonstrated that when patient survival is examined as a function of Kt, the J-shaped curve disappears and mortality declines over the entire range of Kt.

Further analysis along these lines provides clarification of this complex relationship. From the USRDS database, 9165 prevalent patients treated between 1990 and 1995 were studied. A Cox proportional hazards model, adjusting for patient characteristics, was used to calculate the relative risk for death. HD dose (equilibrated Kt/V) and various indices of body size (body mass index, body weight, and volume) were found to be independently inversely related to mortality ( Figure 58-16 ).

FIGURE 58-16  Risk of death as a function of urea volume and dialysis dose.


Thus, mortality was lower in patients with larger body size or volume and decreased as a function of HD dose. The relationship between Kt/V and declining mortality is valid but patient size must also be considered. The implication of this analysis is that the indices of body size may be surrogates for nutritional status. Nutritional status is clearly an important predictor of survival.

Urea Reduction Ratio and Solute Removal Index

There are alternative methods to the quantification of the dose of HD. URR (predialysis BUN - postdialysis BUN/predialysis BUN) and solute removal index (SRI), based on dialysate urea measurements (total dialysate urea in grams × 100/predialysis BUN × V). URR depends exclusively on the changes that occur in urea levels during IHD. The urea removed by convection is not accounted for by URR. Although URR has been shown to correlate with survival in a fashion similar to Kt/V and is recognized by DOQI guidelines as a valid index of HD adequacy, Kt/V is more precise. Unlike URR, Kt/V also permits rational adjustments to the dialysis prescription. Furthermore, URR cannot be used to judge the adequacy of peritoneal dialysis because urea levels are essentially in a steady state (URR ∼ 0).

The SRI measures the amount of urea removed rather than the fractional change in urea. It is not influenced by compartmental distribution of urea. However, the measurement of dialysate urea requires special techniques and is not routinely done. Few studies have been done validating SRI as an index of adequacy.

The NCDS was designed to prospectively determine which determinants of the dialysis prescription had an impact on patient outcome. The study was able to validate urea removal, a surrogate for LMW substances, as an index of morbidity. Based on urea kinetics, minimum level of HD below which increased morbidity resulted was a key finding that has stood the test of time. More importantly, the NCDS provided the stimulus for a large number of observational studies that has resulted in the recommendations of the DOQI guidelines. The validity of these recommendations is now supported by the findings of the HEMO Study.

Standard Kt/V

There are no established criteria for evaluating HD dose for intensive HD therapies such as DHD or NHD; two dose measures based on serum urea concentrations have been proposed: equivalent renal clearance (EKR) and stdKt/V. The EKR concept equalizes the TACurea for different therapies; weekly EKR normalized by urea distribution volume (EKRt/V) is approximately equivalent to simply summing the eKt/V for each individual HD treatment during the week. This dose measure has been criticized by some because it does not explain why patients treated by continuous ambulatory peritoneal dialysis (CAPD) and conventional HD have similar outcomes but different EKR values for urea. Gotch[316] has argued that EKR is not an appropriate dose measure because it does not account for the first-order nature of HD therapy. Although its true that the time-averaged weekly urea clearance or Kt/V will be doubled by either doubling the clearance of a set number of treatments per week or doubling the number of treatments per week while leaving the clearance/treatment unchanged, total solute removal is not the same. As the frequency of HD increases, the peak levels of urea decline and solute removal per session declines because urea is cleared by first-order kinetics, which are concentration dependent.

Therefore, it has been alternatively proposed that dialysis dose be better expressed as stdKt/V, a dose measure that combines treatment dose with treatment frequency and allows for various intermittent therapies to be compared with continuous therapy. stdKt/V can be defined as the continuous removal rate divided by the average peak concentration. In the steady state, the continuous removal rate for urea is equal to its generation rate (Gurea). Thus, stdKt/V = Gurea/average peak urea concentration. As treatment frequency and/or length increases, peak concentration approaches the mean concentration found in continuous therapies. In this calculation, regimens that result in the same mean pre-HD serum urea concentrations during the week would have equivalent stdKt/V values ( Figure 58-17 ).

FIGURE 58-17  Relationship between intermittent hemodialysis (spKt/V) performed one to seven times each week and standard Kt/V/wk. A thrice-weekly treatment providing a single-pool Kt/V of 1.2/treatment is equivalent to a regimen providing an single-pool Kt/V of 0.5 five times each week. K, urea clearance; t, treatment time; V, volume of urea distribution.


Because the weekly stdKt/V during intermittent therapy is based on peak urea concentration rather than mean urea concentration found during continuous therapy, it will be lower than the sum of the intermittent single-pool Kt/V treatments per week.


The Uremic Syndrome

The uremic syndrome is a complex phenomenon involving dysfunction of a number of organ systems in the body attributable to the retention of a myriad of solutes normally secreted by healthy kidneys. Increased attention is currently being focused on identification of specific uremic toxins and on characterization of their biochemical and pathophysiologic processes that contribute to the uremic syndrome.[317] Many of the effects of uremia on organ system function as well as a description of how organ systems adapt to compromised renal function are the topic of discussion of other chapters in this textbook. In this section, the focus is on describing abnormalities attributable to uremia that persist in the dialysis patient or that are potentially accentuated by the HD procedure. It must be emphasized that whereas the effects of uremia on specific organ systems are generally considered isolated phenomena, there are clearly multiple interactions between organ systems in the adaptation to uremic toxicity. Among the many examples that could be cited is that the influence of uremic toxicity on gastrointestinal tract function may exacerbate malnutrition by decreasing nutrient intake. Worsening nutritional status may in turn affect endocrine function and increase the risk for infectious and cardiovascular complications. Hormonal abnormalities associated with uremia may similarly affect intermediary metabolism hematopoietic function. Alterations in lipid and carbohydrate metabolism may also adversely affect the development of atherosclerosis and cardiovascular complications.

Whereas many of the organ system abnormalities attributable to uremia may be improved with the institution of maintenance HD therapy, it must also be recognized that HD may actually potentiate or worsen other uremic complications. Gastrointestinal bleeding may be potentiated by the use of anticoagulation during HD. The HD procedure may induce a more catabolic state through amino acid losses and contribute to malnutrition. Similarly, treatments designed to prevent one uremic complication may potentiate another, such as suppression of secondary hyperparathyroidism leading to adynamic bone disease.

An additional important consideration in evaluating the effects of HD on the uremic syndrome is that many of the solutes retained in uremia are not well cleared by the HD procedure. Because the major mode of clearance during HD is via diffusion, LMW solutes are preferentially removed. Higher-molecular-weight solutes greater than 500 Da (middle molecules), protein-bound solutes, and proteins modified by the uremic milieu are minimally cleared during HD. Several middle-molecular-weight solutes, such as β2-microglobulin, are associated with specific uremic syndromes (β2-microglobulin amyloidosis), whereas highly protein-bound solutes such as para-cresol are increasingly recognized as contributors to vascular toxicity.[318] An effort is currently under way, led by the The European Uremic Toxicity Working Group (EUTox) to catalog the biochemistry and biology of solutes retained in excess in the uremic syndrome.[319] However, associating individual uremic toxins with specific in vivo toxicities on a molecular and cellular level remains a challenge to uremic toxicity investigators.

Owing to the complexities of the interrelationships between uremia, renal replacement therapy, and various organ functions, it is important that the nephrologist who takes care of maintenance dialysis patients be knowledgeable and familiar with the numerous ramifications that result from the loss of renal function. Indeed, the nephrologist caring for the maintenance HD patient is practicing “anephric” internal medicine. The following section describes frequently occurring medical management problems in the maintenance HD patient.


In healthy individuals, the kidney produces up to 80% of circulating erythropoietin, accounting for the pivotal role of the kidney in regulating erythropoiesis.[320] The pathogenesis of anemia in renal failure is described in detail in Chapters 49 and 55 . In this section, emphasis is placed on describing factors specific to the management of anemia in maintenance HD patients and the interactions of HD and erythropoietic therapy. The first clinical trials of recombinant human erythropoietin in maintenance HD patients began in 1985, and erythropoietin received U.S. Food and Drug Administration (FDA) approval in June of 1989.[321] The target maintenance hematocrit for patients in the initial phase I and II clinical trials was 35% to 40%, that is, at the lower range of normal values. A target hematocrit of 35% was used in the subsequent phase III multicenter clinical trials. From the outset of these clinical trials, it was apparent that erythropoietin would be a highly effective agent in the maintenance HD population, with a resulting dramatic decrease in requirement of erythrocyte transfusions.[288] In the early trials, exacerbation of hypertension and, on occasion, the development of seizures were observed with the use of erythropoietin, likely as a consequence of the rapidity with which hematocrit was increased. In subsequent studies with a slower rate of rise of increase in hematocrit, only modest increases in the level of hypertension have been observed.[322]

The use of erythropoietin leads to improvement in cardiac hemodynamics, with studies demonstrating a decrease in left ventricular hypertrophy, improvement in oxygen-releasing capacity, enhanced exercise capability, and reduction in intradialytic hypotension. [324] [325] [326] [327] [328] [329] [330] Improvements of physical performance, work capacity, quality of life, sexual function, and cognitive capacity have also been demonstrated. [331] [332] [333] [334] [335] [336]The decrease in erythrocyte transfusions associated with the use of erythropoietin has led to a beneficial decrease in the frequency of transfusion-associated hepatitis and a decreased human leukocyte antigen (HLA) presensitization for renal transplantation.

In the years prior to the introduction of erythropoietin, erythrocyte transfusions were used to replace blood losses and treat the anemia of CKD. In many dialysis patients, repeated erythrocyte transfusions led to a high frequency of iron overload manifested as hemochromatosis, with subsequent hepatic, cardiac, and muscle dysfunction. [337] [338] Deferoxamine was often administered for treatment of iron overload, but it was associated with severe complications including propensity to increased infection.[338] Treatment of anemia with erythropoietin has virtually eliminated this syndrome by minimizing transfusion requirements in the dialysis population.

The appropriate dose and route of administration of erythropoietin remain somewhat controversial. Erythropoietin therapy is indicated for all patients undergoing maintenance HD who present with or develop a predialysis hemoglobin below 11 g/dL. In many studies, subcutaneous administration of erythropoietin has more favorable pharmacodynamic characteristics than intravenous administration despite incomplete bioavailability of erythropoietin after subcutaneous administration.[339] When erythropoietin is administered subcutaneously, it is recommended that the site of injection be rotated with each administration. Greater potency is achieved owing to a prolonged half-life after subcutaneous versus intravenous erythropoietin administration. At least 36 published studies involving over 2000 patients comparing subcutaneous with intravenous administration reveal that, on average, the dose of erythropoietin required to maintain a stable hematocrit is approximately 30% lower (range 0%–68%).[340] However, whereas, in the aggregate, patients respond better to subcutaneous than intravenous erythropoietin, in some studies, up to 23% of patients require more erythropoietin on switching from intravenous to subcutaneous route. Enthusiasm for administering erythropoietin subcutaneously, however, has been diminished by the discovery that, occasionally, patients treated with recombinant erythropoietin develop anti-erythropoietin antibodies, clinically manifested as pure red cell aplasia. In virtually all cases, patients who develop pure red cell aplasia have had erythropoietin administered through the subcutaneous route. At the present time, the intravenous route remains the predominant mode of erythropoietin administration in HD patients owing to patient preference, convenience, and lack of patient discomfort.

Optimal target hemoglobin concentrations for maintenance HD patients remain controversial. The original target hematocrit recommended in 1989 by the FDA was 33%, which in 1994, was widened to a target range of 30% to 36%. An extensive review as part of the DOQI and K-DOQI Clinical Practice Guidelines recommended a target hemoglobin of 11 to 12 g/dL. However, many of the initial physiologic and quality of life studies in HD patients used erythropoietin to achieve target hematocrit values above 36%. Virtually all of these studies demonstrated that with increased hematocrit, there were marked improvements in physiologic measures, including oxygen utilization, cardiac function, cognitive and brain electrophysiologic function, and sexual function. These studies led some investigators to advocate that normalization of hematocrit (or to a hemoglobin 14 g/dL) might be beneficial in maintenance HD patients.

Several published studies have examined the consequences of targeting a normalized hemoglobin and hematocrit in HD patients. In the largest study, 1233 patients known to have cardiovascular disease (CHF or ischemic heart disease) were randomized to a target hematocrit of 42% versus 30%.[341] The study was terminated prematurely owing to a trend for a higher incidence of nonfatal myocardial infarction or death in the normal hematocrit group. A randomized prospective Canadian trial examined the effect of normalizing hematocrit on left ventricular geometry in HD patients with known cardiomyopathy. This study, although relatively underpowered in design, did not demonstrate significant benefits in left ventricular remodeling associated with normalizing hematocrit.[342] Quality of life did improve in this study with no apparent excess mortality in the normal hematocrit group. A Spanish cooperative group conducted a 6-month prospective trial on the effect of normalized hematocrit on patient quality of life in nondiabetic HD patients without known symptomatic cardiovascular disease.[343] In this study, quality of life scores and functional status improved significantly, hospitalization rates declined significantly, and no deaths were reported among the 156 patients. In an effort to shed further light on the divergent results of published randomized clinical trials, observational data on incident U.S. patients with Medicare insurance between 1996 and 1998 have been examined.[344] The all-cause relative risk of death was higher in patients with hematocrits less than 33, but was similar in the 33 to 36, 36 to 39, and greater-than-39 hematocrit groups. The relative risks for hospitalization were lower in the 36 to 39 and greater-than-39 hematocrit groups compared with the reference 33 to 36 hematocrit group. However, a recent analysis incorporating time dependence suggests no additional survival benefit above a hemoglobin level of 12 g/dL.[345]

Independent of administration of exogenous erythropoietin, the initiation of HD has been demonstrated to partially correct the anemia of CKF and to improve erythrocyte survival. [347] [348] This effect may well be masked if erythropoietin is used in CKD prior to the initiation of dialysis. Studies have also demonstrated that inadequate dialysis may contribute to erythropoietin resistance. Conversely, several components of the HD procedure may contribute to higher erythropoietin requirements. Blood loss occurring through the dialyzer or dialyzer lines, or as a consequence of repeated phlebotomy, can contribute to lower hemoglobin level.[348] Hemolysis can rarely occur as a complication of the HD procedure owing to kinking of dialysis tubing, other mechanical factors, thermal erythrocyte injury, dialysate contamination, or changes in osmolarity from technical errors.[349] The two most important factors that may contribute to erythropoietin resistance in the maintenance dialysis population are chronic inflammation and iron deficiency, which are discussed later.

Aluminum intoxication, now relatively rare in the maintenance dialysis population, can lead to microcytic hypochromic anemia, in addition to inducing osteomalacia and dementia. Although unusual, deficiencies of folic acid and/or vitamin B12 can cause macrocytic anemia in the dialysis population.[350] It is noteworthy that macrocytosis is often observed in patients who are responding vigorously to erythropoietin as reticulocytes have a larger mean corpuscular volume than do mature erythrocytes.

Endogenous erythropoietin is a heavily glycosylated protein, and the glycosylation is essential for its biologic activity. Commercially available recombinant erythropoietin (known as Epoetin) is produced by recombinant methods in Chinese hamster-ovary cells. There are differences in the glycosylation of recombinant Epoetin compared with native erythropoietin, primarily involving the sialic acid composition of the oligsaccharide groups. Until 2002, there were only three cases in which antierythropoietin antibodies were reported to have developed after the administration of Epoetin. In 2002, 13 cases in which patients with CKF developed severe transfusion-dependent anemia after an initial hematologic response to Epoetin were reported.[195] In all cases, this was due to the development of pure red cell aplasia in association with neutralizing antierythropoietin antibodies. Since this publication, numerous other cases, predominantly from Europe, of pure red cell aplasia developing after Epoetin therapy have been reported.

Darbepoetin alfa, a novel erythrocyte-stimulating protein, has also been developed for the treatment of anemia associated with CKF and has recently received FDA approval. Darbepoetin alfa is a glycoprotein with a threefold longer terminal half-life than recombinant human erythropoietin in HD patients, thereby allowing it to be administered less frequently with similar clinical efficacy.[351] In clinical trials, an equivalent number of patients receiving darbepoetin alfa and recombinant human erythropoietin achieved a targeted increase in hemoglobin. [353] [354] A similar frequency of adverse events, withdrawals, and deaths have been reported in clinical trials between patients receiving darbepoetin alfa and recombinant human erythropoietin. Darbepoetin alfa is now utilized in a substantial number of HD patients because of its longer half-life and less frequent need for administration.

With the introduction of erythropoietin therapy, there was concern that an increased hematocrit would cause a reduction in relative plasma volume, thereby having an adverse effect on solute clearance during HD. Clinical studies have not validated this concern. The slight decrease in dialyzer clearance associated with increasing the hematocrit has at most a minor effect on dialysis efficiency, which can be overcome by an appropriate change in the dialysis prescription. An additional concern with increasing the hematocrit is that improved platelet function and correction of the bleeding time that occur with correction of anemia would increase the potential for vascular access thrombosis. Several studies have suggested that there may be an increase in the frequency of AV fistula or graft thrombosis with higher hemoglobin levels.[354] This effect was notably seen in the large U.S. Normalization of Hematocrit Trial.[341] Current clinical practice guidelines do not recommend an increased heparin dose or increased surveillance of HD accesses when patients are treated with erythropoietin.

The routine use of erythropoietic agents has increased the extent to which iron deficiency, rather than iron overload, is the norm for maintenance HD patients. Iron deficiency is among the most common causes of erythropoietin resistance. Iron deficiency likely develops due to ongoing blood losses and to an increased rate of erythrocyte turnover in maintenance HD patients. Whereas oral iron can be utilized, generally speaking, oral iron preparations are relatively ineffective and poorly tolerated in the HD population.[355] Gastrointestinal distress and impaired iron absorption can adversely affect oral intake and also potentially increase the propensity to malnutrition. As a consequence, the provision of intravenous iron has become a common practice in the maintenance HD population. [357] [358] [359] However, concern has been raised that administration of large doses of parenteral iron may contribute to morbidity and mortality in HD patients due to infectious or cardiovascular causes. Concern has arisen because of observational data suggesting that the administration of higher doses of iron is associated with high rates of hospitalization and from post hoc analysis of the U.S. Normalization of Hematocrit results. [342] [360]

Iron is known to be a growth factor for bacteria, and furthermore, staphylococci express a cell wall transferring-binding protein that can function as a transferrin receptor thereby facilitating bacterial iron uptake. Iron administration can inhibit in vitro phagocytic cell function. However, in evaluating the potential infectious risks of iron therapy, the largest prospective multicenter study evaluating risk factors for bacteremia in HD patients did not find that neither the serum ferritin level nor the extent of iron administration was a significant risk factor for bacteremia. At the present time, there are no prospective clinical data to support theoretical concerns that intravenous iron administration may contribute to infection.

In the Normalization of Hematocrit Trial, patients randomized to the normal hematocrit group received a higher amount of intravenous iron than the lower hematocrit group, which led to the post hoc hypothesis that iron administration could have been responsible for excess cardiovascular mortality. Four hundred and sixty-four of 615 patients in the low hematocrit group received intravenous iron compared with 526 of 618 patients in the normal hematocrit group. Free iron in the plasma has the potential to increase oxidative stress and thereby contribute to cardiovascular mortality. However, serum transferrin and ferritin are potent binders of free iron, such that it rarely appears in the plasma except when excessive doses of iron are given over rapid time intervals. Lipid peroxidation in response to intravenous iron has been demonstrated with measurement of plasma esterified F2-isoprostane levels, which are a sensitive measure of lipid peroxidation.[360] Investigators have also demonstrated that administration of high dose D tocopherol before HD can attenuate increases in lipid peroxidation induced by the administration of intravenous iron.[361] Thus, whereas there is speculation that iron administration may be cardiotoxic, perhaps via increased oxidative stress, there are no prospective clinical trials that have compared differences in intravenous iron administration strategies that have shown any difference in cardiovascular or other clinical outcomes.[362] In a recent analysis of a large observational database that used time-dependent modeling and adjusted for nutrition and inflammation, doses of intravenous iron up to 400 mg/mo were associated with improved survival compared with no intravenous iron administration.[363]

Until recently, the only FDA-approved source of parenteral iron in the United States was iron dextran. Dextran binds relatively tightly to iron, which allows slower dissociation and less potential for free iron toxicity. However, iron dextran administration is associated with anaphylactoid reactions in some patients owing to the development of antidextran antibodies. Life-threatening anaphylactoid reactions have been reported to occur in up to 0.7% of iron dextran-treated HD patients, and at least 30 deaths have been attributed to iron dextran reactions. In the United States, two forms of parenteral iron, iron sucrose and iron gluconate complex, have received FDA approval. The use of iron sucrose and iron gluconate complex is associated with substantially lower rates of adverse drug events than those of iron dextran. [365] [366] Iron sucrose and iron gluconate have both been safely administered to patients who have had adverse drug events when receiving iron dextran. [367] [368] The newer iron preparations are now in widespread clinical use in the U.S. HD patient population. Ferumoxytol, an investigational carbohydrate-coated magnetic iron oxide preparation, may also have utility in dialysis patients.[368]

Cardiovascular Disease (see Chapter 48 )

Morbidity and mortality from cardiovascular disease is greatly increased in patients on maintenance HD therapy. Recognition that cardiovascular mortality is greater than 100-fold higher in HD patients under 45 years of age compared with the general population and at least fivefold higher at every age group has led to a recent resurgence of interest in this seemingly intractable problem.[369] Cardiovascular mortality continues to account for more than 50% of deaths in HD patients. Cardiac mortality can arise from arrhythmia, cardiomyopathy, and ischemic heart disease. It is clear that in dealing with the high prevalence of cardiovascular disease in the maintenance HD population, nephrologists will rely heavily on involvement of cardiologists. However, there are also important differences in pathophysiology, pharmacology, and prognosis between HD patients and patients in the general population with cardiovascular disease. Therefore, it is important for nephrologists to remain aware of disease-specific factors that influence the care of cardiac problems in the HD patient population.

The first reports of accelerated atherosclerosis in dialysis patients originated in the 1970s from Scribner and colleagues [371] [372] based on their Seattle experience. The presence of angiographically confirmed atherosclerotic coronary arterial disease in the HD population varies with the study population reported. Overall, reports vary from 24% prevalence in young, nondiabetic patients on HD undergoing evaluation for renal transplantation to as high as 85% in diabetic HD patients over 45 years of age.[372] Although the nature and distribution of coronary atherosclerotic lesions in the dialysis population have not been intensely studied, there is a much greater frequency of complex calcified atheroma similar to the disease process seen in diabetic patients with coronary disease.[373] Postmortem examination of coronary atherosclerotic disease in dialysis and control patients confirms more calcified plaques in patients with ESRD. Coronary plaques in dialysis patients are also characterized by increased medial thickness.[374]

Traditional atherogenic factors that are highly prevalent in the HD population include dyslipidemia and hypertension.[375] Lipoprotein (a) (Lp[a]), an atherogenic lipoprotein, consists of a low-density lipoprotein (LDL)-cholesterol particle that is covalently bonded to apolipoprotein (a) (Apo[a]), an glycoprotein with genetic size polymorphism. Lp(a) level is negatively associated Apo(a) isoform size and is elevated in dialysis patients. Prospective studies of Lp(a) in atherosclerotic coronary vascular disease have yielded conflicting results. [377] [378] [379] However, an inception cohort study of incident dialysis patients followed prospectively demonstrated that small Apo(a) size predicts mortality, even with multiple adjustments for demographics, comorbidity, cause of renal failure, and congestive heart failure.[379] The association of Apo(a) size with cardiovascular outcomes was greater in African American than in white patients. Whereas both LMW Apo(a) size and high Lp(a) level predict cardiovascular risk in dialysis patients, the association with LMW isoforms is stronger than the association with high Lp(a) concentration.[380]

An area of considerable controversy in evaluating atherosclerotic risk in HD patients concerns homocysteine metabolism. Hyperhomocysteinemia is highly prevalent in HD patients, and serum homocysteine concentrations have been correlated with cardiovascular risk in the general population.[381] Hyperhomocysteinemia may contribute to a prothrombotic state by reducing endothelial dysfunction. Early studies in the HD population also suggested that plasma homocysteine concentration is associated with cardiovascular risk. However, increased plasma homocysteine concentration was actually associated with a reduced overall mortality in HD patients in two prospectively published cohort studies.[382] Because plasma homocysteine levels in renal failure may be affected by dietary protein intake, it has been suggested that higher plasma homocysteine levels may be a surrogate for enhanced nutrition in HD patients. In recent trials, the use of neither folic acid nor folinic acid has been successful in decreasing plasma homocysteine levels to normal levels. Moreover, recent trials of homocysteine-lowering therapies for cardiovascular risk protection in the general population have supported the null hypothesis. [384] [385] At present, the contribution of hyperhomocysteinemia to cardiovascular risk in HD patients is speculative, and in the absence of B vitamin deficiency, no effective therapy exists to significantly lower plasma levels. It is also possible that the simultaneous elevation of plasma fibrinogen, Lp(a), and homocysteine levels synergistically contribute to a prothrombotic state in HD patients.

There has been a surge of interest in how the development of a persistent microinflammatory state, which is a frequent accompaniment to renal failure, can contribute to atherosclerotic complications.[385] Plasma levels of C-reactive protein (CRP) as a marker of acute-phase inflammatory reactions have been demonstrated in prospective cross-sectional studies and in cohort studies to be powerful predictors of cardiovascular and all-cause mortality in HD patients.[387] [388] [389] Hypoalbuminemia is also associated with elevated CRP levels, likely as a negative acute-phase reactant. The strong association between atherosclerosis, elevated CRP levels, and hypoalbuminemia has led to the term malnutrition, inflammation, and atherosclerosis (MIA) syndrome to describe this microinflammatory process.[389] The inverse relationship between serum albumin and prealbumin with CRP levels has also been noted for other acute-phase proteins including serum amyloid(a), fibrinogen, and proinflammatory cytokines.[385] More recently, plasma levels of the proinflammatory cytokine interleukin-6 (IL-6) have similarly been associated with the presence and subsequent rapid progression of underlying atherosclerosis in HD patients.[390] High plasma IL-6 levels also predict subsequent cardiovascular and all-cause mortality.[391] This has led to the suggestion that atherosclerosis in the HD population may be driven by phagocytic cell activation with increased production of proinflammatory cytokines.[392] Endothelial dysfunction, measured by circulating endothelial cells or circulating endothelial microparticles, are associated with future vascular events in HD patients. [394] [395]

Oxidative stress is highly prevalent in HD patients and may contribute to an acceleration of atherogenesis.[395] Oxidation of LDL, particularly in the subendothelial space, leads to uptake of oxidized LDL by monocytes/macrophages and conversion into foam cells, the earliest stage in the atherosclerotic process. Numerous studies using multiple separate biomarkers of oxidative stress status have identified CKD in ESRD as states of increased oxidative stress. Oxidative stress may occur in HD patients directly as a result of loss of renal clearance of oxidants and by increased production through activated phagocytic cells. In particular, myeloperoxidase-catalyzed oxidants including hypochlorous acid have been postulated to contribute to excess oxidant production in dialysis patients.[396] Further larger trials are required to confirm or refute the hypothesis that antioxidants may lower cardiovascular risk in dialysis patients. The potential importance of increased oxidative stress as a contributor to cardiovascular disease is emphasized by the SPACE Study (Secondary Prevention with Antioxidants of Cardiovascular Disease in End-Stage Renal Disease).[397] In this study, administration of the antioxidant D-tocopherol as a secondary prevention agent reduced cardiovascular events in a cohort of HD patients by approximately 50% compared with placebo, although there was no difference in all-cause mortality. A similar result has recently been reported with the use of the thiol-containing antioxidant N-acetylcysteine.[398]

Several additional cardiovascular correlates have been implicated in mortality in HD patients. The pulse pressure, or the difference between systolic and diastolic blood pressure, appears to have considerably more predictive power than either systolic or diastolic blood pressures alone in HD patients for adverse events. [400] [401] [402] The aortic pulse wave velocity index is a strong independent predictor of overall and cardiovascular mortality in HD patients.[402] QT dispersion, which is defined as the difference in duration between the shortest and the longest QT intervals on an electrocardiogram, is a measure of regional heterogeneity in myocardial repolarization. Corrected QT interval dispersion can predict adverse cardiovascular outcomes in HD patients. [244] [404] Nocturnal hypoxemia may also be an important predictor of cardiovascular complications in HD patients.[404] Asymmetrical dimethylarginine, an endogenous nitric oxide inhibitor that accumulates in uremia, correlates with carotid atherosclerosis and has been associated with subsequent cardiovascular mortality in HD patients.[405]

Thus, there are a myriad of traditional and nontraditional risk factors that appear to have an importance in predicting cardiovascular events in the HD population and the enormous atherosclerotic cardiovascular disease burden. It is surprising that there has been a paucity of secondary cardiovascular prevention trials in HD patients. Recently, the results of a randomized clinical trial were reported in which 1255 diabetic HD patients were randomized to receive either atorvastatin or placebo. The results of the study somewhat surprisingly supported the null hypothesis in that the primary endpoint (a composite of cardiovascular events and mortality) did not differ between the treatment groups. There was a trend toward lower cardiac event rate in the atorvastatin-treated group, but a statistically significantly higher rate of stroke in the atorvastatin-treated group.[406] Two additional large, well-powered, randomized clinical trials examining the effect of other statins and cholesterol lowering therapies in HD patients are under way at this time. There have been no prospective, randomized, controlled trials examining the efficacy of antiplatelet agents reported in the HD population. Similarly, prospective randomized studies using angiotensin-converting enzyme inhibitors or angiotensin-receptor blockers are under way but have not been reported. A small randomized, placebo-controlled trial demonstrated that carvedilol improved mortality in HD patients with dilated cardiomyopathy.[407] Adequately powered, randomized clinical trials for secondary prevention of cardiovascular disease in HD patients remain a high priority for improving the outcomes of these patients.

Treatment of underlying atherosclerotic coronary artery disease in the HD population is complicated by the occasional discordance between symptoms of angina pectoris and active coronary artery disease. Several studies suggest that the prevalence of angina significantly exceeds the prevalence of large-vessel coronary artery disease. Angina in the absence of large-vessel coronary artery disease may occur owing to left ventricular hypertrophy, intracardiac fibrosis, intramyocardial coronary artery disease, abnormal endothelial vasomotor function, and cardiac autonomic neuropathy. [409] [410] [411] [412] [413] Subendocardial myocardial perfusion in diastole may also be adversely affected by decreased aortic compliance in HD patients. Conversely, the presence of severe coronary artery disease without symptomatology is also common in HD patients. In one report, 75% of diabetic patients with angiographically significant coronary disease had no symptoms.[413] Similar findings have also been identified in nondiabetic HD patients. [415] [416] The lack of anginal symptoms in nondiabetic patients may be explained by uremic autonomic neuropathy or a low level of physical activity in many HD patients.[416] Of note, plasma levels of troponin T and I have prognostic utility even in asymptomatic HD patients.[417]

There are conflicting reports as to what screening tests have the best sensitivity and specificity in HD patients. [373] [419] [420] Ambulatory electrocardiographic recordings have not been well studied in HD patients. Exercise electrocardiography is often limited by abnormal resting electrocardiograms and by markedly reduced exercise tolerance. There are conflicting reports on the usefulness of nuclear medicine studies in comparison to stress echocardiography. Relatively high sensitivity and specificity have been reported for dipyridamole exercise thalium imaging. Similar positive and negative predictive values have been reported in several studies using dobutamine stress echocardiography.[420] Myocardial scintography using fatty acid analogs can detect coronary artery disease with very high sensitivity and reasonable specificity.[421]

Determining optimal therapeutic options, including revascularization strategies, in HD patients must be based on observational data as controlled long-term studies have not been performed. However, data from U.S. HD patients strongly suggest that patients do not receive therapies that are routine in the management of acute coronary syndromes in patients without renal disease. Fewer HD patients with acute myocardial infarction in the United States receive thrombolytic therapy. [423] [424] This is despite the overall 1-year mortality rate being extraordinarily high after acute myocardial infarction in dialysis patients. [425] [426] Evidence suggests that patients undergoing either coronary artery bypass graft (CABG) surgery or percutaneous coronary interventions (PCI) do better than patients not undergoing revascularization.[426] A number of studies have suggested that coronary artery bypass surgery may be associated with improved survival compared with percutaneous transluminal coronary angioplasty (PTCA) in dialysis patients.[427] Unfavorable outcomes in HD patients after PTCA compared with CABG surgery include increased coronary restenosis, as well as an increase in cardiovascular events.[428] Using USRDS data, it has been projected that long-term survival of dialysis patients is more favorable after CABG than after PTCA, a result that has been independently confirmed. In a recent comparison of PTCA, coronary artery stenting, and CABG, 2-year all-cause survival was 56% for CABG patients, 48% for PTCA patients, and 48% for stent patients. In particular, diabetic patients receiving a CABG had a better outcome than those receiving stents or PTCA.[427] In nondiabetic ESRD patients, there was an advantage for stents in comparison to PTCA. Whether drug-eluting stents, brachytherapy, or other new revascularization approaches will have increased clinical utility in HD patients has yet to be determined. Cardiac rehabilitation may also improve survival after CABG.[429]

In addition to atherosclerosis, alterations in left ventricular geometry are common in HD patients. Left ventricular hypertrophy and left ventricular dilatation are found in 75% of patients at the start of dialysis and have been found to be independent risk factors for mortality. [431] [432] The pathologic characteristics of the heart in HD patients with dilated cardiomyopathy (determined by endomyocardial biopsy) are interstitial fibrosis and severe myocyte hypertrophy with occasional disarray.[432] Alterations in left ventricular geometry result from chronic volume and pressure overload, often in association with metabolic and neurohumoral abnormalities. Anemia, hypoalbuminemia, and an increase in systolic blood pressure have all been found to contribute to the development of left ventricular hypertrophy.[433] Increased cardiac output and decreased peripheral vascular resistance associated with AV fistulas may also contribute to the development of cardiomyopathy.[398] Of importance in HD patients, left ventricular alterations tend to progress over time in the majority of patients. In some studies, left ventricular hypertrophy can undergo regression with vigorous treatment of hypertension using angiotensin-converting enzyme inhibitors and/or with treatment of anemia with erythropoietin. However, the finding that correction of anemia can induce partial regression of left ventricular hypertrophy was challenged in a prospective, randomized trial involving normalization of hemoglobin in patients with asymptomatic cardiomyopathy.[342] HD patients who respond to aggressive treatment of hypertension and anemia with a regression of left ventricular hypertrophy have been shown to have better long-term survival than nonresponders. Nonresponders to medical management tend to have increased aortic stiffness and may also have a higher degree of microinflammation as evidenced by elevated CRP levels. [432] [435]

Vascular Calcification

In addition to a high prevalence of atherosclerosis, it is now becoming appreciated that HD patients are also subject to excess vascular calcification and cardiac valvular calcification. In the general population, coronary artery calcification correlates with and predicts cardiovascular mortality. In HD patients, histologic and radiographic evidence of vascular calcification is more striking than in the general population, even being observed in young patients.[436] [437] However, whether the findings of excessive arterial calcification using imaging techniques such as electron beam computed tomography in dialysis patients represents evidence of atherosclerosis or a different process involving medial calcification is not yet entirely clear. Examination of the epigastric arteries of dialysis patients undergoing renal transplantation demonstrate excessive medial artery calcification that is associated with bone matrix protein deposition and disorganization of vascular smooth muscle cells.[437] Elevated serum phosphorus and calcium × phosphorus product are demonstrated to be risk factors for excessive vascular calcification in addition to increased mortality in HD patients. [439] [440] [441] Cardiac valvular calcification is also associated with high calcium phosphorus product but is additionally associated with biomarkers of inflammation including CRP, low serum albumin, and high fibrinogen levels.[441] Alterations in plasma concentrations of pyrophosphate and osteopontin, which are physiologic inhibitors of vascular calcification, may also contribute to medial calcification in uremia. [443] [444]

The association of a high calcium × phosphorus product with progressive vascular calcification would suggest that the use of non-calcium-containing phosphorus binders might reduce propensity for aortic and other vascular calcification.[444] In a randomized, placebo-controlled study, the use of sevelamer, a non-calcium-containing phosphorus binder, was found to be less likely to cause progressive coronary and aortic calcification than calcium-containing phosphorus binders.[445] However, a larger similarly randomized controlled trial with clinical event primary endpoints failed to show a beneficial effect of sevelamer (unpublished results).


The pathogenesis and treatment of uremic metabolic bone disease is discussed in detail in Chapters 52 and 56 , and therefore are not discussed further in this chapter. Calciphylaxis, a small vessel vasculopathy with some clinical and histopathologic features suggestive of hyperparathyroidism and vascular calcification, is discussed briefly. Calciphylaxis is a vasculopathy largely confined to patients with renal insufficiency. Ischemia of the skin and subcutaneous tissues are the most common clinical presentation, leading to necrotizing skin ulcers that heal poorly, subcutaneous nodules of infarction, and areas of poor wound healing. The most common sites of involvement are subcutaneous tissue with increased adipose content, including the breast, abdominal wall, and thighs. [447] [448] [449] [450] [451] The overall prognosis of calciphylaxis is poor, with death owing to sepsis extremely common.[451]

Risk factors for the development of calciphylaxis appear to include obesity, type 2 diabetes mellitus, and white ethnic background.[452] Much attention has also focused on the purported role of elevated serum calcium, serum phosphorous, and PTH level in patients who develop calciphylaxis.[437] In particular, sustained hyperphosphatemia may be a risk factor for calciphylaxis. However, phosphorus and calcium × phosphorus products in patients who develop calciphylaxis often overlap those of patients without calciphylaxis. Controversy exists as to whether hypercalcemia alone can contribute to the development of calciphylaxis. A case report identifies a patient with hypercalcemia due to primary hyperparathyroidism who rapidly developed calciphylaxis skin lesions.[453] However, in two case series, the time average serum calcium concentrations in HD patients who develop calciphylaxis were no different than those in patients who did not develop calciphylaxis. [452] [455]

Confusion exists in distinguishing vascular calcification associated with a high calcium × phosphorus product from calciphylaxis. At the risk of overgeneralizing, the relatively common vascular calcification associated with an elevated calcium × phosphorus product in HD patients largely involves medial calcification, whereas calciphylaxis appears to be an endovascular calcification involving the intima. The role of PTH in the pathogenesis of calciphylaxis also remains controversial. In early animal experiments, Selye and co-workers [456] [457] suggested that PTH may be a sensitizing agent for the development of calciphylaxis. In most, but not all, reported cases of calciphylaxis, serum PTH levels are elevated. Vitamin D analogs were similarly used as sensitizers in Selye's model of calciphylaxis, and some evidence suggests that high levels of 1-25-O(OH)2-vitamin D3 can cause media wall calcification.

Several other risk factors for the development of calciphylaxis have been hypothesized. Protein calorie malnutrition, use of warfarin, vitamin K deficiency, and protein C and/or protein S deficiency have all be postulated as potentially having an etiologic role in the development of calciphylaxis. Calciphylaxis is generally diagnosed clinically by the presence of subcutaneous and cutaneous nodules, necrotic lesions, and eschar with hyperesthesia of the skin. Tissue biopsy can confirm the diagnosis of calciphylaxis by demonstrating calcification of small arterioles or venules. However, skin biopsies in calciphylaxis are not always recommended owing to a high likelihood of poor healing at the biopsy site. Low transcutaneous oxygen can be helpful identifying calciphylaxis skin lesions.[454]

Unfortunately, treatment outcomes for patients with calciphylaxis are poor. Because of the sporadic nature of the disease, only anecdotal case reports of successful treatment modalities exist. Strategies that have been suggested include the discontinuation of calcium-containing phosphorus binders, subtotal parathyroidectomy, and hyperbaric oxygen therapy. For patients on warfarin, it may be appropriate to attempt to utilize a different type of anticoagulant, and a search for hypercoagulable states that may mimic calciphylaxis should be undertaken in appropriate cases. Careful attention to analgesia, avoidance of trauma, and wound care are essential to promote survival in patients with calciphylaxis. Several recent case reports suggest that treat-ment with sodium thiosulfate may have efficacy in this syndrome. [458] [459]

Nephrogenic Systemic Fibrosis

This adverse response to gadolinium is discussed in detail in Chapter 27 .


Protein calorie malnutrition has been shown to be highly prevalent and associated with increased morbidity and mortality in maintenance HD patients. [64] [460] Alterations in nutritional status owing to uremia per se as well as from the HD procedure may predispose the ESRD patient to multiple nutritional complications ( Table 58-12 ). A comprehensive review of nutritional therapy in renal disease is provided in Chapter 53 ; therefore, this section focuses specifically on nutritional evaluation and therapy in the HD unit.

TABLE 58-12   -- Factors Causing Malnutrition

Inadequate protein/calorie intake

Increased energy expenditure

Metabolic acidosis

Hormonal alterations


Dialytic nutrient losses

Dialysis-induced catabolism





Optimal monitoring of protein energy nutritional status for maintenance HD patients requires that several different parameters be measured for the assessment of different aspects of nutritional status. There is no single biochemical marker or clinical measure that in and of itself provides a complete overview of protein energy nutritional status. A combination of measures of protein and energy intake, biochemical measures of visceral protein pools, and anthropometric measures of body compositions are complementary in the assessment of nutritional status.

In HD patients, relatively simple biochemical measures reflecting the visceral protein stores, such as serum albumin, creatinine, and BUN, as well as more complex, less readily available measures such as transferrin, prealbumin, and insulin-like growth factor-1 (IGF-1) have been proposed as nutritional indices. The serum albumin is the most extensively examined nutritional index in virtually all patient populations owing to ready availability of its measurement and association with clinical outcomes. Serum albumin levels are closely affected by the level of dietary protein intake; however, it must be recognized that in HD patients, inflammation and dietary protein intake exert competing effects on serum albumin concentration.[386] The serum albumin is also a negative acute-phase reactant, and its serum concentration decreases abruptly and sharply in response to stress and inflammation. [386] [461] Thus, taken in isolation, the serum albumin may not necessarily reflect visceral nutritional status in acutely ill patients.

In addition to the serum albumin, the BUN and serum creatinine concentrations are also considered simple markers of nutritional status. Urea is the metabolic end product of dietary protein intake, and thus the BUN is a composite measure of protein intake, volume of distribution of urea, as well as renal and dialyzer urea clearance.

Similar to the serum albumin level, the serum prealbumin level can be used as a marker of nutritional status. Serum prealbumin has a shorter half-life (1–2 days) than serum albumin, and thus, it may be a better marker to determine early response to a nutritional therapeutic intervention. Prealbumin is a serum transport protein that has a smaller body pool than serum albumin; however, the major route of excretion of prealbumin is via the kidneys, and thus, serum concentrations of prealbumin may be falsely elevated in patients on HD. Serum prealbumin can also function as a negative acute-phase reactant. Some data have suggested that the prealbumin level may be an even better prognostic indicator than albumin levels in HD patients.[461]

IGF-1 can also be a nutritional index in HD patients.[462] IGF-1 is a growth factor that is structurally related to insulin and produced primarily within the liver. Ninety-five percent of plasma IGF-1 is protein-bound. Thus, there are limited daily fluctuations in serum concentration. Several studies suggest that serum IGF-1 concentrations may have a better correlation with body composition than serum albumin and transferrin.[463] However, at the present time, the level of IGF-1 at which malnutrition is significant has not been established in HD patients.

Analysis of body composition also provides important nutritional assessment information for HD patients. Anthropometric studies are easy to perform but, unfortunately, are unreliable in the HD setting. More sophisticated body composition tools such as prompt neutron activation analysis or dual energy x-ray absorptiometry (DEXA) have reported utility as nutritional measures in HD patients, but these require expensive equipment and are available only in specialized centers. [465] [466] Bioelectrical impedance analysis (BIA) has also been proposed as useful for body composition analysis in HD patients.[466] Although there was a good correlation of total body water and lean body mass between BIA and DEXA in healthy subjects, the variation in HD patients may be large. The true utility of BIA as a measure of body composition in HD patients remains to be clarified and established at this time.

Subjective global assessment is a recently proposed methodology for evaluating nutritional status of HD patients. Thus, subjective global assessment is a simple technique based on objective and subjective aspects of physical examination and the medical history. Subjective global assessment was initially developed for evaluation of nutritional status in patients undergoing elective surgery, but it has subsequently been applied to other patient populations. The use of subjective global assessment has been well studied in peritoneal dialysis patients, with less data currently available in HD patients.

Dietary protein intake assessment through dietary interviews and diaries can provide important information concerning protein, energy, and other nutrient intake. Three-day dietary record reviews, along with dietary interviews by qualified staff, are recommended for accurate information. The protein equivalent of total nitrogen appearance, or protein catabolic rate, can also be used to assess dietary protein intake.[467]

A number of potential causes for malnutrition in HD patients exist (see Table 58-12 ). Because of the propensity for protein calorie malnutrition in maintenance HD patients, the recommended dietary protein for stable patients is 1.2 g/kg of body weight per day. At least 50% of dietary protein should be of high biologic value. The recommended daily energy intake is 35 kcal/kg body weight per day for patients who are less than 60 years of age and 30 to 35 kcal/kg body weight per day for individuals 60 years of age or older, according to current clinical practice guidelines.

Institution of maintenance dialysis may have complex effects on protein and energy balance. [469] [470] Early studies using low-flux dialyzers have documented a loss of 5 to 8 g of free amino acids during each HD session. With the use of high-flux dialysis membranes, these losses further increase by approximately 30% owing to larger surface area and higher blood flows used.[470] These losses may be increased when a glucose-free dialysate is used owing to the stimulation of gluconeogenesis. There may be an increase in amino acid and albumin losses with the reprocessing of high-flux dialyzers by an increase in membrane porosity.[471] Amino acid losses during dialysis also stimulate a catabolic process that increases catabolism for hours after the dialysis procedure has ended. [473] [474] [475]

For HD patients who sustain inadequate nutrient intake for extended periods of time and/or develop indices of protein energy malnutrition, nutritional support may be indicated. For patients with an intact functional intestinal tract, the enteral route of nutritional support is favored. Oral nutritional supplementation during HD has been demonstrated to contribute to increases in serum albumin, serum prealbumin, and subjective global assessment scores. [476] [477] [478] [479] For patients unable to tolerate enteral feeding, or unable to utilize nutrients provided by the intestinal tract, either total parenteral nutrition or intradialytic parenteral nutrition can be considered. There have been few large-scale prospective studies of intradialytic parenteral nutrition for malnourished HD patients. Some studies have demonstrated efficacy in increasing serum albumin and other markers of nutritional status; however, interdialytic parenteral nutrition is expensive and provides insufficient calories and protein to support long-term needs owing to the limitations on its administration during HD sessions only.[424]

It has been proposed that many HD patients may have L-carnitine deficiency and benefit from carnitine supplementation.[479] L-Carnitine is a naturally occurring substance that shuttles fatty acids into the mitochondria for oxidation. L-Carnitine is critical for energy production in cardiac and skeletal muscle tissues that are dependent on fatty acid oxidation. Because carnitine is water-soluble and readily dialyzed, plasma concentrations of carnitine will decline by as much as 75% with HD. The decrease in plasma carnitine concentration is quickly corrected by transport of carnitine from muscle and other tissues, which may lead to tissue carnitine deficiency. It has been proposed that L-carnitine supplementation would be beneficial in HD patients to improve a variety of metabolic abnormalities including hypertriglyceridemia, hypercholesterolemia, anemia, and exercise tolerance.[480] At the present time, however, whether there is clinical benefit to be derived from L-carnitine administration to HD patients is controversial. L-Carnitine administration may have the greatest clinical utility in the treatment of erythropoietin-resistant anemia. L-Carnitine is currently not recommended for routine use in HD patients, albeit selected individuals who have not responded adequately to standard therapies may be appropriate candidates for L-carnitine administration.

The concentration in the body of many trace elements is dependent on the degree of renal failure. Thus, the serum concentrations of trace elements, with the exception of selenium and zinc, tend to be elevated in HD patients. Indeed, it has been proposed that trace element accumulation may contribute to uremic toxicity, and thus, only selenium and zinc should be considered for supplementation in dialysis patients. Selenium deficiency has been associated with cardiovascular disease as selenium functions as a co-factor in antioxidant enzyme function. Studies have documented decreased concentrations of selenium in HD patients, likely secondary to inadequate dietary intake. However, whether selenium supplementation would lead to clinical benefit in dialysis patients is not well defined. Similarly, low concentrations of zinc have been reported in HD patients. Zinc deficiency may be associated with anorexia and impotence; however, the potential beneficial effects of supplemental zinc therapy have not been confirmed in HD patients.[481]

Supplementation of water-soluble vitamins is recommended for HD patients. [483] [484] B vitamin and folic acid supplementation may be required for optimal metabolism of homocysteine. Because folic acid is abundant in foodstuffs that are frequently restricted in dialysis patients owing to the concomitant presence of high potassium, folic acid supplementation is recommended.[484] Several studies have suggested that there may be a role for vitamin C supplementation to prevent oxidative injury and to aid in iron metabolism. With the exception of vitamin D, the lipid-soluble vitamins have not been found to be depleted in HD patients. Vitamin A or beta-carotene supplements should be avoided owing to potential toxicity in HD patients.[485] Vitamin E may have clinical benefit in reducing cardiovascular complications in HD patients (see Cardiovascular Disease in this chapter).[397] The metabolism of vitamin D and its supplementation are also discussed in Chapter 56 .

Infection and Immunity

Infection is the second leading cause of death in HD patients after cardiovascular diseases. The mortality rate owing to infection in ESRD patients is approximately 12% to 22%.[486] Septicemia accounts for more than 75% of these infectious deaths. Overall, the annual percentage mortality that is secondary to sepsis in dialysis patients is approximately 100- to 300-fold higher than in the general population, a stark portrayal of how serious infectious risks are for these patients.[94] Dialysis access is an important component of sepsis risk, as patients using a catheter are at much higher sepsis risk than patients using a catheter or graft.[487]

Several clinical and treatment-related characteristics make HD patients particularly susceptible to septicemia. The repeated disruption of dermal integrity to gain vascular access for HD increases infection risk. In patients with catheters for vascular access, there was risk of infection within and around the indwelling catheter's lumen. Clinical data confirm that central venous catheters are a major source of bacterial colonization and infection in HD patients compared with patients using AV grafts or AV fistulas.[488] There is increasing recognition that dialysis catheter bacterial colonization leads to bacterial adherence and biofilm formation. Biofilms, which are microbial communities attached to surfaces, develop resistance to phagocytes and to antibiotics. Biofilms eventually become covered with a dense layer of matrix material, rendering resistance to antibiotic penetration.[489]

Most infections in HD patients are due to common catalase-producing bacteria such as Staphylococcus species, rather than to opportunistic infections. [489] [491] The pattern of infectious organisms in HD patients is similar to that seen in patients with chronic granulomatous disease, in which phagocytic cells lack capability to produce reactive oxygen species. In this context, a number of studies have documented alterations in granulocyte function in patients with uremia on HD. Although the number and morphology of granulocytes are generally normal, defects in granulocyte chemotaxis, adherence, phagocytic capability, and reactive oxygen species production have all been demonstrated. [492] [493]

There is increasing concern about the development of antibiotic resistance in HD patients. The frequent occurrence of methacillin-resistant Staphylococcus aureus infections over time has led to the frequent use of vancomycin to treat suspected and proven septicemia. A CDC survey in 1996 and 1997 demonstrated that approximately 5% of HD patients are administered vancomycin over the course of a month and 30% of dialysis facilities reported having patients with known vancomycin-resistant Enterococcus (VRE).[493] The percentage of dialysis centers reporting VRE colonization or infection is increasing over time.[191] A valence study at 49 hospitals revealed that receipt of HD and peritoneal dialysis was an independent risk factor for VRE bacteremia.[494] Strains of S. aureus with intermediate sensitivity to vancomycin are now being reported.

Viral infections are also common in HD patients. In the early days of dialysis, multiple blood product transfusions will predispose many patients to hepatitis B virus (HBV) infections. The availability and use of erythropoietin has reduced the number of transfusions in dialysis patients. Coupled with the availability of the HBV vaccine, the frequency of HBV has decreased markedly in HD patients. According to the CDC, during the period from 1976 to 2000, the incidence of HBV infection in HD patients decreased from 4.4% to 0.05%.[493] Similarly, during this time period, prevalence of HBV surface antigen positivity among HD patients had declined from 7.8% to 0.9%. The use of HBV vaccine for patients and staff has been increasing.[191] Currently, HD patients may acquire HBV infection from community sources, from transmission in HD centers owing to inadequate infection control precautions, or owing to accidental breaks in infection control technique.

The hepatitis C virus (HCV) was cloned in 1989 and identified as the leading cause of parentally transmitted, non-A, non-B hepatitis. HCV infection has become the most important cause of liver disease among HD patients and is additionally a major concern for HD staff. The prevalence of anti-HCV antibody positivity in HD units has varied from 5% to 40%. [496] [497] Data do suggest that the incidence of HCV infection in HD units is declining worldwide, including in the United States, since the mid-1990s. Many patients with anti-HCV antibodies do not exhibit hepatic enzyme abnormalities, and serum alanine aminotransferase (ALT) levels are elevated in only a minority of HD patients with anti-HCV antibodies, or with detectable HCV RNA.[497] The clinical course of HCV infection in HD patients tends to be chronic and indolent with a characteristic fluctuating course and multiple peaks and troughs in ALT levels.[498] Molecular genotyping of HCV by polymerase chain reaction and nucleotide sequencing of the viral genome have unequivocally demonstrated nosocomial transmission of HCV within the dialysis unit. Current recommendations are that HD patients be considered a high-risk population for HCV infection and should undergo periodic screening. Isolation of HCV carriers within the HD unit is not currently recommended beyond standard universal precautions. Although most HD patients with HCV infection remain asymptomatic over a long period of time, identified patients should be instructed to avoid additional hepatic toxins including alcohol consumption and potentially hepatotoxic medications. HCV-positive patients should undergo vaccination against hepatitis A and HBV. Currently, antiviral therapy with interferon-α is recommended for selected categories of HCV-infected HD patients.[499]

HIV infection appears to be increasing in HD patients. During the period of 1985 to 2000, the percentage of centers reporting that dialysis care is being provided for patients with HIV infection increased from 11% to 37%. Because only a minority of dialysis centers routinely test for HIV infection, these figures are likely underestimates of the true rate of HIV infection. According to the CDC, in 2000, 1.5% of HD patients were reported to have HIV infection and 0.4% to have acquired immunodeficiency syndrome.[493] HIV-positive HD patients are treated with highly active antiretroviral therapy.

Given the increased susceptibility to infections and the high associated mortality in HD patients, it is notable that many of the types of deaths are due to infections that are vaccine-preventable. The incidence of pneumonia in dialysis patients has been reported to be as high as 4.9 episodes per 1000 patient-months, and 53% of these are due to Streptococcus pneumoniae. Thus, it is currently recommended by the advisory committee on immunization practices that HBV, pneumococcal, and influenza vaccines be specifically recommended for HD patients.[500]

Patients on HD should receive three doses of recombinant HBV vaccine as early in the course of renal disease as possible. The recommended dosage for adults on HD is 40 mg of either Recombivax HB or Engerix-B given intramuscularly in the deltoid. Only 50% to 75% of adult patients develop protective antibody levels against HBV surface antigen after three doses of vaccine compared with more than 90% of healthy adults. Revaccination with up to three additional doses is recommended for those HD patients who do not develop protective antibody levels. Postvaccination anti-HBV surface antigens testing is recommended 1 to 2 months after vaccination in HD patients to demonstrate protective antibody levels. Additional postvaccination testing is recommended annually. A booster dose is recommended if the anti-HBV surface antigen titer falls below 10 mU/mL. The use of an adjuvant may increase the immunogenicity and response rate of the HBV vaccine in HD patients.[501]


HD has evolved into a relatively safe procedure, with an estimated 1 death in 75,000 treatments as a result of technical error. However, an extensive list of complications is related to this treatment, some of which are potentially life-threatening. In this discussion of the acute complications, it should be noted that the age of the patient; the presence of underlying medical conditions such as diabetes mellitus, coronary artery disease, or CHF; and the patient's degree of compliance with a complex medical regimen necessary in ESRD have a great influence on the frequency and severity of adverse events.


Hypotension is the most common acute complication of HD.[502] Several dialytic- and patient-related factors influence blood pressure during the treatment. The incidence of hypotension in the dialysis population is quite frequent, occurring with 20% to 50% of dialysis treatments. [504] [505] The frequency varies with the age and sex of the patient, with the greatest number of episodes being seen in older patients and in women. Intradialytic hypotension is associated with increased morbidity and mortality, especially when episodes occur frequently.[505]

The hemodynamic response to HD must be reviewed for an understanding of the pathogenesis of hypotension.[506] The dialysis procedure can be considered to be made up of two separable processes: convection and diffusion. Convection refers to the movement of fluid and solute brought about by pressure across the dialysis membrane (transmembrane pressure). The higher the transmembrane pressure, the greater the rate of convection. The process of removing fluid by hydraulic forces is termed ultrafiltration. During isolated UF, a progressive increase in total systemic vascular resistance maintains blood pressure as fluid is removed. When diffusion is added to UF (i.e., the usual dialysis treatment, which is a combination of UF and diffusion), thermal energy can be transferred from the heated dialysate to the patient, resulting in a stimulus for vasodilatation and increased blood flow to the skin. As a consequence, vasoconstriction is less effective and maintenance of central blood volume may be impaired during fluid removal. Cardiac output and blood pressure are maintained by an increase in heart rate and, in some instances, by an increase in myocardial contractility. However, if the intravascular volume is low, increased heart rate and improved contractility may not adequately maintain a tolerable blood pressure. In addition, the large burden of cardiovascular disease in the HD population often limits the ability of the heart and the vessels to response appropriately to the stress of fluid removal. These inherently different responses to UF and diffusion greatly influence maintenance of blood pressure during HD.

Besides factors related to the method of exchange, abnormalities in autonomic function are often present in dialysis patients. The afferent arm of the baroreceptor reflex is believed to be blunted in hypotension-prone HD patients, and such patients do not mount reflex vasoconstriction during hypotensive episodes. Although the efferent arm of this reflex pathway, which involves sympathetic output, is believed to be normal or even overactive in patients with CKF, this limb of the pathway has also been shown to fail in patients prone to hypotension during HD.[507]

Ultrafiltration Rate

During UF, a protein-free ultrafiltrate of plasma is removed from the intravascular space. The resultant rise in plasma oncotic pressure causes fluid to move from the interstitial and intracellular spaces to replenish plasma volume. Hypotension results when the rate of intravasuclar volume depletion exceeds the rate of refilling of this space, especially if total peripheral resistance cannot compensate for the loss of intravascular volume. As mentioned previously, whereas total peripheral resistance increases during isolated UF, this compensatory response is attenuated when diffusion is added to the process. Thus, during combined UF and diffusion when vasoconstriction is not evident, the ability to remove volume during HD is primarily dependent on the ability to refill the intravascular space.

Very large interdialytic weight gains cannot be easily removed during a typical treatment (usually lasting between 3.5 and 4 hr), even in the presence of volume overload, because the refilling of intravascular space is time dependent. Frequent hypotension in such individuals is probable with UF rates in excess of 1.5 L/hr, and the incidence of hypotension is, in general, an exponential function of the rate of fluid removal.

Finally, hypotension can occur when the weight of the patient is at or below her or his “estimated dry weight” when volume shifts no longer are able to compensate for intravascular depletion and maintain blood pressure. Indeed, the estimated dry weight of a patient may be defined as that weight below which the patient develops symptomatic hypotension, in the absence of edema and excessive interdialytic weight gains. The assessment of volume status by physical examination can be augmented by measuring inferior vena cava diameter by echocardiography[508] or by on-line monitoring of changes in hematocrit (see later). A narrow or collapsing inferior vena cava suggests volume depletion. A patient experiencing hypotension with these findings would benefit from an increase in dry weight.

Dialysate Composition

The composition of the dialysate can influence blood pressure in several ways. Sodium and calcium concentration, the nature of the buffer (either bicarbonate or acetate), and the temperature of the dialysis fluid are among the factors that have been demonstrated to influence the frequency of hypotension during dialysis, as discussed previously.

The process of diffusion also leads to a decline in plasma osmolality because of removal of solutes from the patient. The magnitude of the fall in effective plasma osmolality ranges between 10 and 25 mOsm/kg. The fall in plasma osmolality creates an osmotic gradient between the plasma and the interstitial and intracellular spaces. Fluid moves from the plasma into cells and the interstitium, resulting in a reduction in plasma volume. This intravascular volume loss is superimposed upon volume removed by UF, and its magnitude can be as much as 1 to 1.5 L during the treatment. This shift is opposed by an increase in oncotic pressure induced by UF. Increases in the concentration of sodium, the principal osmotic agent in the dialysate, will reduce this osmotic gradient. Indeed, the frequency of hypotension reported at a dialysate sodium concentration of 140 mEq/L is substantially lower than the frequency at 130 mEq/L. A similar effect can be seen with mannitol administration during HD. However, chronic mannitol administration can lead to its accumulation in HD patients and should be avoided.

Dialysate with a sodium concentration of 130 mEq/L has been used in the past, in part because of the fear that dialysate with higher tonicity would stimulate thirst, leading to greater interdialytic weight gains and hypertension. Indeed, because of the Gibbs-Donnan equilibrium, net positive sodium balance occurs at sodium concentrations of 140 mEq/L. Although weight gains may be somewhat greater with higher sodium concentrations, this additional fluid weight gain can usually be successfully removed. Nevertheless, an increase in interdialytic hypertension may be a problem in some individuals. Despite these potential problems, the use of dialysate with a sodium concentration of 140 mEq/L is common.

The calcium concentration of the dialysate has also been shown to affect myocardial contractility. Higher calcium concentrations in the dialysate, up to a concentration of 3.5 mEq/L, have been associated with improved contractility, independent of the nature of other factors in the dialysate. However, with the common use of calcium salts to prevent hyperphosphatemia, hypercalcemia is seen more often when the calcium concentration of dialysate is this high. In response to concerns of possible increased morbidity and mortality in dialysis patients with hypercalcemia, the recent trend in the United States is to reduce dialysate calcium to 2.5 mEq/L.

The buffer used to replenish bicarbonate lost during the interdialytic interval has also been clearly implicated in the pathogenesis of hypotension during HD. Historically, the principal buffer employed had been acetate. The use of acetate as a buffer was to reduce the potential of calcium precipitation (as calcium carbonate) when bicarbonate was used as the buffer. Acetate is a peripheral vasodilator and may also predispose to hypotension by reducing myocardial contractility. As the technology for delivery of bicarbonate dialysate has improved, the use of acetate-based dialysate has dramatically declined in recent years, and this has resulted in a significant improvement in the rates of hypotension during dialysis.

Dialysate temperature has been shown to affect blood pressure during HD. [510] [511] The majority of patients will have positive thermal balance unless the dialysate temperature is below the patient's core temperature, leading to mild temperature increases. Dialysated cooled to 35°C or 36°C reduces the frequency of hypotensive episodes, leading to more stable treatments, and is generally tolerated by patients. [510] [512]

Theoretically, vasoactive substances may be removed during the treatment. However, during HD, the changes in plasma norepinephrine levels or potassium concentration, for instance, have not been shown to play an important role in dialysis-induced hypotension.

Extracorporeal Volume

In the past, the volume of blood required to prime the extracorporeal circuit as well as the compliance of the dialyzer were two potential causes of hypotension. The blood volume of coil dialyzers, no longer in use, and to a lesser extent, of parallel-plate dialyzers, increases as transmembrane pressure increases. Hollow-fiber dialyzers are much less compliant, and their widespread use has reduced the potential for hypotension from this cause. Nevertheless, the blood volume needed to establish the extracorporeal circuit (including the dialyzer and blood lines) may still be as high as 200 mL.


Patients undergoing HD are often receiving antihypertensive agents or other medications that can interfere with the normal hemodynamic response to UF. β-Adrenergic receptor blockers reduce myocardial contractility and also exert a negative chronotropic effect. Such agents, by preventing a compensatory increase in the heart rate, interfere with a major defense supporting blood pressure during dialysis. Verapamil and diltiazem can be expected to exert a similar effect. Vasodilators can prevent the normal vasoconstriction response to UF.

The development of several antihypertensive agents formulated to be administered orally as a single daily dose or by a transdermal delivery system (e.g., nifedipine and clonidine, respectively) has reduced the incidence of drug-induced hypotension. These agents are generally well tolerated and may often be used on dialysis days, because high peak levels are avoided. Nitroglycerin ointment can often aggravate the propensity for hypotension by inducing peripheral vasodilatation.

Other Factors Producing Hypotension

The previous discussion relates to factors inherent in the treatment itself that can produce hypotension. The health of the patient is another important variable that directly influences the frequency of hypotension. Patients at increased risk for hypotension are those who have arrhythmias, which can often be exacerbated by HD; those with low hematocrits, leading to tissue ischemia; those with poor cardiac function or pericarditis; or those with autonomic dysfunction such as diabetic patients. The last two problems may prevent adequate changes in cardiac output or peripheral resistance to compensate for fluid removed during HD.


The first step in the approach to hypotension is to determine whether hypotension occurs early or late in the treatment period. In a previously stable patient, who is free of edema and signs of CHF, in whom hypotension occurs late in the treatment, the most common cause will be that the patient's dry weight has been underestimated. Reducing the amount of UF during HD, and thereby effectively raising the postdialysis dry weight, will correct the hypotension. In contrast, the patient with excessive intradialytic weight gains may become hypotensive before the dry weight is achieved because the rate at which fluid can be mobilized to refill the intravascular space is limited. In this instance, the patient should be counseled to limit his or her salt and fluid intake, and dialysis time or frequency may need to be increased for removal of all necessary fluid at a tolerable rate. More frequent dialysis in both the acute and the chronic settings has been associated with a decrease in frequency and severity of hypotension episodes and may be considered in patients with persistent refractory hypotension. [513] [514]

Whenever possible, doses of short-acting antihypertensive medication should not be administered within 4 hours of the HD treatment. Many of the long-acting blood pressure medications can be taken at bedtime to avoid peak concentrations during dialysis. In patients with frequent hypotension early into the treatment, pericarditis with tamponade must be suspected.

Often, a multifaceted approach will prevent hypotension. The use of bicarbonate dialysate with a sodium concentration of 140 mEq/L will be helpful. As mentioned elsewhere, sodium and UF modeling can be applied to the treatment.[108] The newer dialysis machines have programs that permit the dialysate sodium and/or the UF rate to be automatically changed during the treatment. Dialysate sodium can be gradually altered during the treatment from an initial concentration of 150 to 140 mEq/L. Fluid is more easily mobilized from the intracellular space with sodium modeling. Most dialysis machines allow dialysate temperature to be easily lowered to 35°C. At this dialysate temperature, thermal energy is transferred from the patient to the dialysate and the resultant vasoconstriction raises blood pressure.[514] Many membranes now produce less reaction during contact with the blood. The reuse of cuprophane membranes (with formaldehyde, but not bleach) also increased biocompatibility of the membrane. For patients with persistent hypotension or autonomic insufficiency, the oral α1-adrenergic agonist midodrine can be prescribed. A dose of 5 to 10 mg given 30 to 60 minutes before HD has been shown to be effective in reducing the incidence of hypotension. [515] [516] Biofeedback devices have been shown to be effective in treatment of hypotension-prone dialysis patients but are not currently approved for use in the United States. [517] [518] Devices that assess the blood volume during dialysis by monitoring the hematocrit have also been developed.[518] Hematocrit-based intradialytic monitoring may potentially be used to help prevent hypotensive episodes.

Hypotension, when it does occur, is treated by placement of the patient in the Trendelenburg position, administering a 100- to 200-mL normal saline bolus, and reducing the UF rate, at least temporarily. Alternatives to normal saline are mannitol, glucose, hypertonic saline, and albumin. Supplemental oxygen may also be useful to improve hypoxemia and cardiac contractility in some patients.

Muscle Cramps

Muscle cramps, the second most common reported complication associated with HD, occur in as many as 20% of dialysis treatments. Although their pathogenesis is uncertain, cramps are known to be more frequent when UF rates are high, during episodes of hypotension, and when dialysate with low sodium concentration is employed, an indication that cramps are caused by acute extracellular volume contraction.

Reducing UF rates will improve cramps. Bolus administration of normal saline (200 mL), small volumes (5 mL) of 23% hypertonic saline, or D50W (50% dextrose in water) is effective in treating cramps. In nondiabetic patients, D50W is especially useful, particularly toward the conclusion of the dialysis treatment, because as glucose is metabolized, hyperosmolality and intravascular volume expansion in the postdialysis period are avoided. The pain resulting from very severe cramps may be alleviated by administration of agents such as diazepam but at the risk of increased hypotension, which may aggravate the condition.

Quinine sulfate, an agent that increases the refractory period and excitability of skeletal muscle, is effective in preventing cramping if administered 1 to 2 hours before dialysis commences. Patients using quinine must be observed for the development of thrombocytopenia. Indeed, this agent was temporarily banned by the FDA in 1994 because of this complication. Alternatives to quinine in preventing cramps include vitamin E and carnitine. [520] [521] [522] In patients with excessive weight gains, dialysis time must be increased to prevent cramps during attempts to achieve the patient's dry weight.

Dialysis Disequilibrium Syndrome

Dialysis disequilibrium refers to a constellation of systemic and neurologic symptoms, many of which are nonspecific, including nausea and vomiting, restlessness, headaches, and fatigue during HD or in the immediate postdialysis period. Severe disequilibrium may result in life-threatening emergencies, including seizures, coma, and arrhythmias. The cause of this syndrome is uncertain. Most believe these symptoms arise from rapid rates of change in solute concentration and pH during HD in the central nervous system. A transient gradient may be created between plasma and cerebrospinal fluid (CSF) urea concentration, leading to increased concentration of water in the brain. A study found increased expression of aquaporins and decreased expression of urea transporters in the brain of uremic rats, which may be a possible mechanism contributing to the development cerebral edema.[522]

A fall in CSF pH may also contribute to cerebral edema. During dialysis, there is a rapid correction of arterial pH, an increase in plasma bicarbonate, and therefore, a rise in arterial pressure of carbon dioxide (Pco2). The rise in plasma Pco2 is accompanied by an increase in CSF Pco2 because carbon dioxide is freely diffusible. However, bicarbonate is slower to enter the CSF. The net result is a decrease in CSF pH. It is possible that an increase in hydrogen ion concentration contributes to the increase in brain cell osmolality (idiogenic osmoles), which in turn causes an increase in brain edema.

Dialysis disequilibrium is most commonly seen in situations in which the initial solute concentrations are very high and the rate at which they decline is rapid. This syndrome, therefore, is seen most commonly and in its severest form during the first few HD sessions experienced by the patient. Milder symptoms may occur in patients in chronic maintenance HD, particularly if noncompliant behavior has resulted in missed or shortened treatments. The overall incidence is reported to be in the range of 10% and 20% of treatments. In particular, the shorter treatment times that are possible because of dialyzers of high clearance (high-efficiency and high-flux dialyzers) may lead to symptoms in smaller individuals who have low urea volumes.


During the initiation of a new patient to HD, measures that reduce the rate of osmolar change are helpful. The use of smaller surface area dialyzers and reduced rates of blood flow and maintaining the direction of flow of dialysate in the same direction as blood flow (rather than the customary countercurrent configuration) are measures that can be employed to lower solute clearance rates and reduce symptoms. A high dialysate sodium (e.g., 145 mg/L) may also be helpful. For severe headache, seizures, or obtundation, the dialysis procedure should be immediately terminated. Intravenous administration of mannitol or diazepam is useful in treating seizures caused by disequilibrium.

Mild symptoms, when present, may be treated less intensively. In general, because patients are often seen before the onset of uremic symptoms, their initiation on dialysis may be planned to avoid these symptoms. Daily dialysis for 3 to 4 days with gradual increases in dialysis time and blood flow often prevents symptoms and signs of disequilibrium. An initial dose of dialysis equivalent to Kt/V of 0.3 on the first day, 0.6 on the second day, 0.9 on the third day, and 1.4 on the fourth day should be used during initiation.

Arrhythmias and Angina

Patients with ESRD frequently develop arrhythmias; there is a high prevalence during HD. Predisposing factors include the high prevalence of left ventricular hypertrophy and valvular sclerosis. Because of disordered calcium and phosphate metabolism, the conduction system may be affected by calcific deposits. In addition, coronary artery disease is common in the dialysis population, and pericardial effusions are frequently noted by echocardiography. Superimposed upon these organic problems are the rapid changes in electrolyte concentrations inherent in efficient HD. It is not surprising that HD may provoke cardiac arrhythmias. Ventricular ectopic activity, including nonsustained ventricular tachycardia, is seen most frequently in patients who are taking digoxin, particularly when dialysate potassium concentration is less than 2.0 mEq/L. Supraventricular tachycardia and atrial fibrillation are also precipitated by hypotension and coronary ischemia.[523]

One must attempt to strike a balance between the need to remove potassium that accumulates during the interdialytic period and the exigency to avoid low serum potassium levels, which will produce arrhythmias. In patients taking digoxin or who have myocardial dysfunction, the use of a dialysate with a potassium concentration of 3.0 mEq/L may reduce the frequency of arrhythmia. The acute therapy for arrhythmias during HD is similar to that for patients with normal renal function, but appropriate dose adjustments must be made for those drugs normally removed by the kidney. [525] [526] A reassessment of the need for digoxin should also be considered. Digoxin is often started before dialysis is started, in an attempt to improve cardiac contractility and lessen CHF. After initiation of dialysis, patients' vascular volume status can often be well controlled by adjustments to this estimated dry weight, and digoxin can be discontinued.

Occasional episodes of atrial fibrillation after dialysis also occur in some patients at the end of dialysis. In many cases, these are self-limited episodes that last 1 to 2 hours, with controlled ventricular rate and no signs or symptoms of ischemia. Neither digoxin nor anticoagulation is definitely indicated in these cases because the risk of subsequent more serious arrhythmias, with concomitant digoxin and hypokalemia, may be greater.

Angina frequently occurs during dialysis. As mentioned previously, coronary artery disease is common in the dialysis population. The anemia associated with CKF adds to the risk of episodes of angina. Also, increases in heart rate frequently accompany UF during diffusive clearance (see the discussion of hypotension), making angina a likely event in patients with coronary artery disease. There is often a need to withhold β-blockers immediately before the HD treatment. Tachyarrhythmias and hypotension can also precipitate angina. Supplemental oxygen should be administered if angina occurs. Decreasing blood flow and temporarily stopping UF may also be helpful. If hypotension is not present, sublingual nitroglycerin may be given, but the patient should be in the recumbent position as a fall in blood pressure can be expected. Preadministration of 2% nitroglycerin ointment applied 1 hour prior to dialysis or oral administration of nitrates or β-blockers may be of benefit to prevent angina, but this must be carefully done as the risk for intradialytic hypotension may be increased.


Carbohydrate metabolism is abnormal in patients with CKF. Although there is a peripheral resistance to the effects of insulin in uremia, the half-life of insulin is significantly prolonged when the GFR is less than 20 mL/min. Finally, the effect of a given dose of insulin is enhanced once dialysis is instituted because an improvement in peripheral responsiveness to insulin occurs once HD has been initiated.

The implication of the foregoing is that a diabetic patient who takes a usual dose of insulin may experience hypoglycemia when undergoing dialysis against a bath with a fixed glucose concentration (glucose clamp) and too low for the amount of insulin being administered. It is frequently necessary to decrease the dose of insulin on dialysis days to pre-vent hypoglycemic episodes. Furthermore, diabetic patients should not be dialyzed against a bath that has a glucose concentration of less than 100 mg/dL.


The uremic environment produces impaired platelet functioning, changes in capillary permeability, and anemia, all of which can impair hemostasis. In addition, there may also be increased blood loss from the gastrointestinal tract because of gastritis or angiodysplasia, lesions associated with renal failure.

The initiation of HD is reported to partially correct the defects responsible for the platelet dysfunction and capillary permeability that occur in uremia. However, patients undergoing HD still have a higher risk of hemorrhagic events because of repeated exposure to heparin. As discussed previously, heparin is commonly used to prevent clotting in the extracorporeal circuit. Although strategies have been developed to dialyze patients without systemic anticoagulation, these techniques are time consuming and require greater supervision than is practical in the setting of an outpatient chronic HD center.

In addition to acute bleeding episodes, patients undergoing HD are exposed to chronic low-grade episodes of blood loss with each dialysis treatment. Five to 10 mL of residual blood remains in the artificial kidney and tubing even after thorough rinsing. There may be blood loss as needles are inserted and removed and as repeated blood tests are performed on the patients. Estimates of loss of between 5 and 50 mL of blood per dialysis treatment have been made.

Blood-Membrane Interaction

The membrane interposed between the blood and the dialysate should not be considered an inert material. Numerous reactions, involving the activation of the complement pathway and the coagulation cascade, as well as with the formed elements of blood, occur during contact of the blood with the dialysis membrane.

First-Use Syndrome

The majority of chronic HD units reuse their dialyzers in the United States. The first-use syndrome refers to a symptom complex encountered when a new dialyzer made of Cuprophane, a cellulosic material, is employed. The symptoms associated with the first use of a dialyzer appear early, usually within the first half-hour after the commencement of the treatment. One group of symptoms resembles an anaphylactic reaction, with urticaria, angioedema, and wheezing. A severe reaction is associated with profound hypotension and cardiac arrest.

In many patients who have suffered from this reaction, elevated levels of immunoglobulin E (IgE) directed against serum proteins that have interacted with ethylene oxide, a sterilizing agent used in the manufacture of dialyzers, are found. The hollow fibers of the artificial kidney are embedded in a potting compound that may be a reservoir for residual ethylene oxide even if the dialyzer is flushed before its use.[526] Owing to improvements by dialyzer manufacturers to remove all residual ethylene oxide, in some cases changing potting compounds, these reactions are much less common.

Complement activation also has been implicated in some of these reactions. Very high rates of complement activation have been demonstrated in patients who have had reactions during the first use of cellulosic dialyzers. In addition, the response of their plasma to zymosan, an activator of complement via the alternative pathway, is also exaggerated.

The full anaphylactoid response occurs rarely (1 of 60,000 exposures to new dialyzers). Much more frequent is the host of nonspecific symptoms, such as coughing, sneezing, pruritus, and back pain that occur early in the treatment with a new dialyzer. Their frequency is reduced when dialyzers are reused. The attenuated reaction in this instance is likely to be attributable to coating of the membrane with serum proteins during the first use. Protein is fixed to the membrane by formaldehyde, which is used as a sterilant in many reuse procedures. Indeed, a reused dialyzer can be shown to produce symptoms if the protein coat is removed by bleach in the reuse process. Noncellulosic membranes, such as polyacrylonitrile, polysulfone, or poly(methyl methacrylate), do not cause large amounts of complement to be released into the circulation, and they appear to be better tolerated as well.

Anaphylactoid reactions have been reported when patients taking angiotensin-converting enzyme (ACE) inhibitors undergo HD using polyacrylonitrile (AN69) membranes or other reused membranes of various kinds. These reactions occur despite the fact that the biocompatibility profile of these membranes, at least with respect to complement activation, is superior to that of new cuprophane membranes. Evidence indicates that AN69, because of its negative surface charge, is capable of generating bradykinin by activation of Hageman factor and the kallikrein-kininogen pathway. [528] [529] ACE is a potent kinase responsible for degrading brady-kinin. Its inhibition by ACE inhibitors may lead to higher bradykinin levels and to the unopposed action of this substance, [530] [531] resulting in bradykinin-induced hypotension and bronchoconstriction. The role of reuse in these occurrences is uncertain, but possibly the character of several membrane surfaces is altered by the reuse procedure.

Treatment of mild forms of this syndrome is symptomatic, but anaphylactoid reactions need to be treated with epinephrine and steroids. Blood in the extracorporeal circuit should not be returned to the patient. The use of biocompatible membranes, reuse programs, the avoidance of ACE inhibitor when AN69 membranes are used, and dialyzers constructed without ethylene oxide-absorbing potting compounds can significantly lessen the frequency of these reactions.

Intradialytic Hemolysis

Acute hemolysis during dialysis is caused either by mechanical problems leading to traumatic fragmentation of the red blood cells or by direct toxic effects of the dialysis solution.[526e] Dialysate contaminants that can lead to hemolysis include formaldehyde, chloramine, bleach, copper, and nitrates, as discussed previously. Hypotonicity or overheating of the dialysate owing to a failure of the conductivity monitor or heating system, respectively, can also lead to hemolysis.

Traumatic fragmentation of the red blood cell can occur anywhere along the length of the tubing where the blood flows. Common sites of hemolysis include the dialyzer roller pump, dialysis catheter or tubing, and at the needle (high flow rates through a small-gauge needle). A defect in disposable tubing sets in 1998 led to an outbreak of hemolysis and two deaths and an urgent recall of these defective parts by their manufacturer.[527] More recently, in 2005, a recall was required because a kink in the blood tubing sets used during dialysis treatment caused hemolysis.[528] Subclinical hemolysis may also occur with high blood flow rates through small needle and catheter holes.[529]

Patients with acute hemolysis will present with chest tightness, shortness of breath, and back pain. Skin discoloration with a dramatic deepening of skin pigmentation has also been described in a patient with severe intradialytic hemolysis.[530] The blood in the venous line may have a “port-wine” appearance, and the plasma samples of centrifuged blood may have a pink color from free hemoglobin. Patients will often have a sharp hematocrit decline, have evidence of hemolysis on a peripheral blood smear, and be hyperkalemic.


The blood pump should be stopped and the hemolyzed blood should be discarded. Preparations for immediate blood transfusions and treatment of hyperkalemia should be made. A thorough analysis of the water used for the dialysate to evaluate for the presence of metal contaminants and chloramines should be performed, and the blood flow tubing should be carefully inspected for kinking and structural defects.


In 2002, Belding Scribner and Willem Kolff won the Albert Lasker Award for clinical and medical research “for the development of renal HD, which changed kidney failure from a fatal to a treatable disease, prolonging the useful lives of millions of patients.” Currently, HD is the mainstay of treatment for ESRD, as far more patients with ESRD receive HD therapy in the United States than either peritoneal dialysis or kidney transplantation combined. Furthermore, recent data suggest a significant survival advantage for patients receiving HD compared with that of peritoneal dialysis.[18] The ever-increasing numbers of patients in the United States and worldwide who are developing ESRD and are likely to receive HD therapy present a challenge to healthcare providers and systems to optimize treatment outcomes in the most cost-effective manner.

Despite these many achievements, the limitations of HD as a treatment for uremia are becoming more apparent, bringing dialysis therapy to a crossroads.[531] The results of the HEMO Study, in which increasing the delivered dose of dialysis or using high-flux dialyzer membranes did not improve mortality, suggests that new approaches will be required to improve overall mortality and morbidity rates for HD patients. Since the mid-1990s, there has been minimal improvement in either hospitalization or mortality rates in the HD population despite steady improvement in achieving clinical performance measures and adhering to clinical practice guidelines. Recent data from the USRDS suggests that long-term outcomes for HD patients may be getting worse, even after adjusting for changing demographics and comorbidities in the HD population. Thus, an examination of the future of renal replacement therapy is timely and indicated.

In the future, there will undoubtedly be further technical advances to support the delivery of HD. The recent past has seen the development and implementation of a number of on-line monitoring devices that provide physiologic data in real time during the HD session. In the future, integrated systems simultaneously measuring multiple physiologic parameters with built-in feedback loops controlling many features of the dialysis prescription and treatment will undoubtedly be developed. Artificial intelligence programs may also assist in the development of more patient-specific dialytic therapy.[532] There will likely be further technologic improvements in the biocompatibility of dialysis membranes, dialysate, and water use, for dialysis treatment. It is hoped for and anticipated that there will be technologic improvements in vascular access for HD such that the success rate for vascular access will be higher and the complication rate (especially thrombogenecity and infection) will be lower. The lack of significant technologic breakthroughs in vascular access care since the development of the autogenous fistula and prosthetic bridge graft over 30 years ago has limited improvement in this area.

Despite more than 30 years of the ESRD program in the United States, there has not been a published randomized clinical trial demonstrating that any pharmacologic agent can lower mortality in the dialysis population. Indeed, to date, despite the high morbidity, mortality, and cost associated with the care of HD patients, there have been few adequately powered pharmaceutical studies published. Targeting cardiovascular complications, which are extraordinarily high in this patient population, will undoubtedly happen and, hopefully, will result in identification of successful therapeutic agents. Currently, there is also a great deal of interest in performing clinical trials of more frequent dialysis than three times a week HD, which has been noted for decades to be inherently “unphysiologic.”[533] Uncontrolled observational cohorts of patients receiving more frequent HD appear to experience substantial clinical improvements. It is hoped and anticipated that randomized trials will demonstrate that these improvements in cohort studies are the result of the therapy and not due to patient selection bias or other factors.

As far back as the inception of maintenance HD therapy for the treatment of ESRD, it was recognized that HD attempts to recapitulate glomerular filtration but not replace renal tubular function.[371] Tubular processing of glomerulofiltrate via selective metabolism and transport may be essential in mitigating uremic toxicity. Cell-based therapies providing proximal tubular function are in development and indeed have entered phase I human clinical trials in the treatment of patients with AKI. Ultimately, partial or complete renal organogenesis may lead to successful renal replacement therapy without the allogenicity or xenogenicity associated with heterotopic transplantation.[534] One cannot but believe (and hope) that the future of renal replacement therapy is bright.


1. Abel JJ, Rowntree LG, Turner BB: On the removal of diffusable substances from the circulating blood by means of dialysis. Transactions of the Association of American Physicians, 1913.  Transfus Sci  1990; 11(2):164-165.

2. Kolff W, Berek H: De kunstmatige nier, een dialysator met groot oppervlak.  Ned Tijdschr Geneeskd  1943; 87:1684-1692.

2a. Grimsrud L, Cole JJ, Lehman GA, et al: A central system for the continuous preparation and distribution of hemodialysis fluid.  Trans Am Soc Artif Intern Organs  1964; 10:107-109.

3. Evans RW, Blagg CR, Bryan Jr FA: Implications for health care policy. A social and demographic profile of hemodialysis patients in the United States.  JAMA  1981; 245:487-491.

4.   United States Renal Data Services, 2005, Accessed April 2006.

5. Grassmann A, Gioberge S, Moeller S, Brown G: ESRD patients in 2004: Global overview of patient numbers, treatment modalities and associated trends.  Nephrol Dial Transplant  2005; 20(12):2587-2593.

6. Xue JL, Chen SC, Ebben JP, et al: Peritoneal and hemodialysis: I. Differences in patient characteristics at initiation.  Kidney Int  2002; 61(2):734-740.

7. Xue JL, Everson SE, Constantini EG, et al: Peritoneal and hemodialysis: II. Mortality risk associated with initial patient characteristics.  Kidney Int  2002; 61(2):741-746.

8. Miskulin DC, Meyer KB, Athienites NV, et al: Comorbidity and other factors associated with modality selection in incident dialysis patients: The CHOICE Study. Choices for Healthy Outcomes in Caring for End-Stage Renal Disease.  Am J Kidney Dis  2002; 39(2):324-336.

9. Jager KJ, Korevaar JC, Dekker FW, et al: The effect of contraindications and patient preference on dialysis modality selection in ESRD patients in The Netherlands.  Am J Kidney Dis  2004; 43(5):891-899.

10. Foley RN, Parfrey PS, Harnett JD, et al: Mode of dialysis therapy and mortality in end-stage renal disease.  J Am Soc Nephrol  1998; 9(2):267-276.

11. Bloembergen WE, Port FK, Mauger EA, Wolfe RA: A comparison of mortality between patients treated with hemodialysis and peritoneal dialysis.  J Am Soc Nephrol  1995; 6(2):177-183.

12. Fenton SS, Schaubel DE, Desmeules M, et al: Hemodialysis versus peritoneal dialysis: A comparison of adjusted mortality rates.  Am J Kidney Dis  1997; 30(3):334-342.

13. Collins AJ, Hao W, Xia H, et al: Mortality risks of peritoneal dialysis and hemodialysis.  Am J Kidney Dis  1999; 34(6):1065-1074.

14. Tanna MM, Vonesh EF, Korbet SM: Patient survival among incident peritoneal dialysis and hemodialysis patients in an urban setting.  Am J Kidney Dis  2000; 36(6):1175-1182.

15. Locatelli F, Marcelli D, Conte F, et al: Survival and development of cardiovascular disease by modality of treatment in patients with end-stage renal disease.  J Am Soc Nephrol  2001; 12(11):2411-2417.

16. Termorshuizen F, Korevaar JC, Dekker FW, et al: The relative importance of residual renal function compared with peritoneal clearance for patient survival and quality of life: An analysis of the Netherlands Cooperative Study on the Adequacy of Dialysis (NECOSAD)-2.  Am J Kidney Dis  2003; 41(6):1293-1302.

17. Bargman JM, Thorpe KE, Churchill DN, CANUSA Peritoneal Dialysis Study Group : Relative contribution of residual renal function and peritoneal clearance to adequacy of dialysis: A reanalysis of the CANUSA study.  J Am Soc Nephrol  2001; 12(10):2158-2162.

18. Jaar BG, Coresh J, Plantinga LC, et al: Comparing the risk for death with peritoneal dialysis and hemodialysis in a national cohort of patients with chronic kidney disease.  Ann Intern Med  2005; 143(3):174-183.

19. Schulman G: Mortality and treatment modality of end-stage renal disease.  Ann Intern Med  2005; 143(3):229-231.

20. National Kidney Foundation: K/DOQI Clinical Practice Guidelines for Chronic Kidney Disease: Evaluation, Classification and Stratification.  Am J Kidney Dis  2002; 39:S1-S266.

21. Kopple J, Greene T, Chumlea WC, et al: Relationship between nutritional status and the glomerular filtration rate: Results from the MDRD Study.  Kidney Int  2000; 57:1688-1703.

22. Levin A, Foley RN: Cardiovascular disease in chronic renal insufficiency.  Am J Kidney Dis  2000; 36(6 suppl 3):S24-S30.

23. Collins AJ, Ma JZ, Xia A, Ebben J: Trends in anemia treatment with erythropoietin usage and patient outcomes.  Am J Kidney Dis  1998; 32:133-141.

24. Parfrey PS, Foley RN, Harnett JD, et al: Outcome and risk factors for left ventricular disorders in chronic uraemia.  Nephrol Dial Transplant  1996; 11:1277-1285.

25. Locatelli F, Conte F, Marcelli D: The impact of hematocrit levels and erythropoietin treatment on overall and cardiovascular mortality and morbidity—The experience of the Lombardy Dialysis Registry.  Nephrol Dial Transplant  1998; 13:1642-1644.

26. Roubicek C, Brunet P, Huiart L, et al: Timing of nephrology referral: Influence on mortality and morbidity.  Am J Kidney Dis  2000; 36(1):35-41.

27. Ratcliffe PJ, Phillips RE, Oliver DO: Late referral for maintenance dialysis.  Br Med J (Clin Res Ed)  1984; 288:441-443.

28. Eadington DW: Delayed referral for dialysis.  Nephrol Dial Transplant  1996; 11:2124-2126.

29. Ifudu O, Dawood M, Homel P, Friedman EA: Excess morbidity in patients starting uremia therapy without prior care by a nephrologist.  Am J Kidney Dis  1996; 28:841-845.

30. Schmidt RJ, Domico JR, Sorkin MI, Hobbs G: Early referral and its impact on emergent first dialyses, health care costs, and outcome.  Am J Kidney Dis  1998; 32:278-283.

31. Kolff WJ, Berk HT: The artificial kidney: A dialyzer with a great area.  Acta Med Scand  1944; 117:121-129.

32. Quinton W, Dillard D, Scribner BH: Cannulation of blood vessels for prolonged hemodialysis.  ASAIO Trans  1960; 6:104-107.

33. Brescia MJ, Cimino JE, Appel K, Hurwich BJ: Chronic hemodialysis using venipuncture and a surgically recreated arteriovenous fistula.  N Engl J Med  1966; 275:1089-1092.

34. Baker Jr LD, Johnson JM, Goldfarb D: Expanded polytetrafluoroethylene (PTFE) subcutaneous arteriovenous conduit: An improved vascular access for chronic hemodialysis.  ASAIO Trans  1976; 22:382-387.

35. Kaplan MS, Mirahmadi KS, Winer RL, et al: Comparison of “PTFE” and bovine grafts for blood access in hemodialysis patients.  ASAIO Trans  1976; 22:388-392.

36. Shaldon S, Chianussi L, Higgs B: Hemodialysis by percutaneous catheterization of the femoral artery and vein with regional heparinization.  Lancet  1961; 2:75-81.

37. Uldall R, Dyke R, Woods F: A subclavian cannula for temporary vascular access for hemodialysis or plasmapheresis.  Dial Transplant  1979; 8:963-968.

38. Schwab SJ, Buller GL, McCann RL, et al: Prospective evaluation of a Dacron cuffed hemodialysis catheter for prolonged use.  Am J Kidney Dis  1988; 11:166-169.

39. Schwab SJ, Weiss MA, Rushton F, et al: Multicenter clinical trial results with the LifeSite hemodialysis access system.  Kidney Int  2002; 62(3):1026-1033.

40. Blake PG, Huraib S, Wu G, Uldall R: The use of dual-lumen jugular venous catheters as definitive long-term access for hemodialysis.  Int J Artif Organs  1990; 13:26-31.

41. Hakim RM, Himmelfarb J: Hemodialysis access failure: A call to action.  Kidney Int  1998; 54:1029-1040.

42. Pisoni RL, Young EW, Dykstra DM, et al: Vascular access use in Europe and the United States: Results from the DOPPS.  Kidney Int  2002; 61(1):305-316.

43. Rayner HC, Pisoni RL, Gillespie BW, et al: Creation, cannulation and survival of arteriovenous fistulae: Data from the Dialysis Outcomes and Practice Patterns Study.  Kidney Int  2003; 63(1):323-330.

44. Oliver MJ, Rothwell DM, Fung K, et al: Late creation of vascular access for hemodialysis and increased risk of sepsis.  J Am Soc Nephrol  2004; 15(7):1936-1942.

45. Dhingra RK, Young EW, Hulbert-Shearon TE, et al: Type of vascular access and mortality in U.S. hemodialysis patients.  Kidney Int  2001; 60(4):1443-1451.

46. Tonnessen BH, Money SR: Embracing the fistula first national vascular access improvement initiative.  J Vasc Surg  2005; 42(3):585-586.

47. ESRD Clinical Performance Measures Project: Vascular access.  Am J Kidney Dis  2002; 39(S):S23-S29.

48. Wedgwood KR, Wiggins PA, Guillou PJ: A prospective study of end-to-side vs. side-to-side arteriovenous fistulas for haemodialysis.  Br J Surg  1984; 71:640-642.

49. Bonalumi U, Civalleri D, Rovida S, et al: Nine years' experience with end-to-end arteriovenous fistula at the “anatomic snuff box” for maintenance haemodialysis.  Br J Surg  1982; 69:486-488.

50. Miller PE, Tolwani A, Luscy CP, et al: Predictors of adequacy of arteriovenous fistulas in hemodialysis patients.  Kidney Int  1999; 56:275-280.

51. Rivers SP, Scher LA, Scheehan E, et al: Basilic vein transposition: An under used autologous alternative to prosthetic dialysis angioaccess.  J Vasc Surg  1993; 18:391-396.

52. Gracz KC, Ing TS, Soung LS, et al: Proximal forearm fistula for maintenance hemodialysis.  Kidney Int  1997; 11:71-74.

53. Oliver MJ, McCann RL, Indridason OS, et al: Comparison of transposed brachiobasilic fistulas to upper arm grafts and brachiocephalic fistulas.  Kidney Int  2001; 60(4):1532-1539.

54. Dixon BS, Novak L, Fangman J: Hemodialysis vascular access survival: Upper-arm native arteriovenous fistula.  Am J Kidney Dis  2002; 39(1):92-101.

55. Lok CE, Oliver MJ, Su J, et al: Arteriovenous fistula outcomes in the era of the elderly dialysis population. 2005.  Kidney Int  2005; 67(6):2462-2469.

56. Gibson KD, Caps MT, Kohler TR, et al: Assessment of a policy to reduce placement of prosthetic hemodialysis access.  Kidney Int  2001; 59(6):2335-2345.

57. O'Hare AM, Dudley RA, Hynes DM, et al: Impact of surgeon and surgical center characteristics on choice of permanent vascular access.  Kidney Int  2003; 64(2):681-689.

58. Allon M, Lockhart ME, Lilly RZ, et al: Effect of preoperative sonographic mapping on vascular access outcomes in hemodialysis patients.  Kidney Int  2001; 60(5):2013-2020.

59. Silva MB, Hobson RW, Pappas PJ, et al: A strategy for increasing use of autologous hemodialysis access procedures: Impact of preoperative noninvasive evaluation.  J Vasc Surg  1998; 27:302-308.

60. Dember LM, Kaufman JS, Beck GJ, et al: Design of the Dialysis Access Consortium (DAC) Clopidogrel Prevention of Early AV Fistula Thrombosis Trial.  Clin Trials  2005; 2(5):413-422.

61. Dixon BS, Beck GJ, Dember LM, et al: Design of the Dialysis Access Consortium (DAC) Aggrenox Prevention Of Access Stenosis Trial.  Clin Trials  2005; 2(5):400-412.

62. Himmelfarb J, Saad T: Hemodialysis vascular access: Emerging concepts.  Curr Opin Nephrom Hypertens  1996; 5:485-491.

63. Woods JD, Tureene MN, Strawderman RL, et al: Vascular access survival among incident hemodialysis patients in the United States.  Am J Kidney Dis  1997; 30:50-57.

64. Kanterman RY, Vesely TM, Pilgram TK, et al: Dialysis access grafts: Anatomic location of venous stenosis and results of angioplasty.  Radiology  1995; 195:135-139.

65. Roy-Chaudhury P, Kelly BS, Miller MA, et al: Venous neointimal hyperplasia in polytetrafluoroethylene dialysis grafts.  Kidney Int  2001; 59(6):2325-2334.

66. Swedberg SH, Brown BG, Sigley R, et al: Intimal fibromuscular hyperplasia at the venous anastomosis of PTFE grafts in hemodialysis patients.  Circulation  1989; 80:1726-1736.

67. Roy-Chaudhury P, Kelly BS, Miller MA, et al: Venous intimal hyperplasia in polytetrafluoroethylene dialysis grafts.  Kindey Int  2002; 59:2325-2334.

68. Weiss MF, Scivittaro V, Anderson JM: Oxidative stress and increased expression of growth factors in lesions of failed hemodialysis access.  Am J Kidney Dis  2001; 37(5):970-980.

69. Himmelfarb J: Pharmacological prevention of vascular access stenosis.  Curr Opin Nephrol Hypertens  1999; 8:569-572.

70. Roy-Chaudhury P, Sukhatme VP, Cheung AK: Hemodialysis vascular access dysfunction: A cellular and molecular viewpoint.  J Am Soc Nephrol  2006; 17(4):1112-1127.

71. Schwab SJ, Raymond JR, Saeed M, et al: Prevention of hemodialysis fistula thrombosis: Early detection of various stenoses.  Kidney Int  1989; 36:707-711.

72. Beathard GA: Physical examination of the dialysis vascular access.  Semin Dial  1998; 11:231-236.

73. Besarab A, Sullivan KL, Ross RP, Moritz MJ: Utility of intra-access pressure monitoring in detecting and correcting venous outlet stenoses prior to thrombosis.  Kidney Int  1995; 47:1364-1373.

74. McCarley P, Wingard RL, Shyr Y, et al: Vascular access blood flow monitoring reduces access morbidity and costs.  Kidney Int  2001; 60(3):1164-1172.

75. Neyra NR, Ikizler TA, May RE, et al: Change in access blood flow over time predicts vascular access thrombosis.  Kindey Int  1998; 54:1714-1719.

76. Depner TA, Krivitsky NM: Clinical measurement of blood flow in hemodialysis access fistulae and grafts by ultrasound dilution.  ASAIO J  1995; 41:M745-M749.

77. Depner TA: Analysis of new methods for vascular access monitoring.  Semin Dial  1999; 12:376-381.

78. Krivitski NM: Theory and validation of access blood flow measurements by dilution technique during hemodialysis.  Kidney Int  1995; 46:244-250.

79. Lindsay RM, Blake PG, Malek P, et al: Accuracy and precision of acess recirculation measurements by the hemodynamic recirculation monitor.  Am J Kidney Dis  1998; 31:242-249.

80. Bosman PJ, Boereboom FT, Eikelboom BC, et al: Graft flow as a predictor of thrombosis in hemodialysis grafts.  Kidney Int  1998; 54:1726-1730.

81. Wang E, Schneditz D, Nepomuceno C, et al: Predictive value of access blood flow in detecting access thrombosis.  ASAIO J  1998; 44:M555-M558.

82. Sands J, Glidden D, Miranda C: Access flow measured during hemodialysis.  ASAIO J  1996; 42:M530-M532.

83. Bay WH, Henry ML, Lazerus JM, et al: Predicting hemodialysis access failure with color flow Doppler ultrasound.  Am J Nephrol  1998; 18:296-340.

84. Strauch BS, O'Connell RS, Geoly KL, et al: Forecasting thrombosis of vascular access with Doppler color flow imaging.  Am J Kidney Dis  1992; 19:554-557.

85. Tonelli M, Jindal K, Hirsch D, et al: Screening for subclinical stenosis in native vessel arteriovenous fistulae.  J Am Soc Nephrol  2001; 12(8):1729-1733.

86. White JJ, Bander SJ, Schwab SJ, et al: Is percutaneous transluminal angioplasty an effective intervention for arteriovenous graft stenosis?.  Semin Dial  2005; 18(3):190-202.

87. Robbin ML, Oser RF, Lee JY, et al: Randomized comparison of ultrasound surveillance and clinical monitoring on arteriovenous graft outcomes.  Kidney Int  2006; 69(4):730-735.

88. Dember LM, Holmberg EF, Kaufman JS: Randomized controlled trial of prophylactic repair of hemodialysis arteriovenous graft stenosis.  Kidney Int  2004; 66(1):390-398.

89. Moist LM, Churchill DN, House AA, et al: Regular monitoring of access flow compared with monitoring of venous pressure fails to improve graft survival.  J Am Soc Nephrol  2003; 14(10):2645-2653.

90. Ram SJ, Work J, Caldito GC, et al: A randomized controlled trial of blood flow and stenosis surveillance of hemodialysis grafts.  Kidney Int  2003; 64(1):272-280.

91. Shahin H, Reddy G, Sharafuddin M, et al: Monthly access flow monitoring with increased prophylactic angioplasty did not improve fistula patency.  Kidney Int  2005; 68(5):2352-2361.

92. Maya ID, Allon M: Outcomes of thrombosed arteriovenous grafts: Comparison of stents vs angioplasty.  Kidney Int  2006; 69(5):934-937.

93. Schwab SJ, Beathard GA: The hemodialysis catheter conundrum: Hate living with them, but can't live without them.  Kidney Int  1999; 56:1-17.

94. Sarnak MJ, Jaber BL: Mortality caused by sepsis in patients with end-stage renal disease compared with the general population.  Kidney Int  2000; 58(4):1758-1764.

95. Marr KA, Sexton DJ, Conlon PJ, et al: Catheter-related bacteremia and outcome of attempted catheter salvage in patients undergoing hemodialysis.  Ann Intern Med  1997; 127:275-280.

96. Krishnasami Z, Carlton D, Bimbo L, et al: Management of hemodialysis catheter-related bacteremia with an adjunctive antibiotic lock solution.  Kidney Int  2002; 61(3):1136-1142.

97. Tanriover B, Carlton D, Saddekni S, et al: Bacteremia associated with tunneled dialysis catheters: Comparison of two treatment strategies.  Kidney Int  2000; 57(5):2151-2155.

98. Beathard GA: Management of bacteremia associated with tunneled-cuffed hemodialysis catheters.  J Am Soc Nephrol  1999; 10:1045-1049.

99. Shaffer D: Catheter-related sepsis complicating long-term, tunneled central venous dialysis catheters: Management by guidewire exchange.  Am J Kidney Dis  1995; 25:593-596.

100. Mokrzycki MH, Jean-Jerome K, Rush H, et al: A randomized trial of minidose warfarin for the prevention of late malfunction in tunneled, cuffed hemodialysis catheters.  Kidney Int  2001; 59(5):1935-1942.

101. Bolz KD, Fjermeros G, Wideroe TE, Hatlinghus S: Catheter malfunction and thrombus formation on double-lumen hemodialysis catheters: An intravascular ultrasonographic study.  Am J Kidney Dis  1995; 25:597-602.

102. Barrett N, Spenser S, McIvor J, Brown EA: Subclavian stenosis: A major complication of subclavian dialysis catheters.  Nepohrol Dial Transplant  1988; 3:423-425.

103. Cheung AK, Gregory MC: Subclavian vein thrombus in hemodialysis patients.  Trans ASAIO  1985; 31:131.

104. Fant GF, Dennis VW, Quarles D: Late vascular complications of the subclavian dialysis catheter.  Am J Kidney Dis  1986; 7:225-228.

105. Hernandez D, Diaz F, Rufino M, et al: Subclavian vascular access stenosis in dialysis patients: Natural history and risk factors.  J Am Soc Nephrol  1998; 9:1507-1510.

106. Khanna S, Sniderman K, Simons M, et al: Superior vena cava stenosis associated with hemodialysis catheters.  Am J Kidney Dis  1993; 21:278-281.

107. Schwab SJ, Quarles D, Middleton JP, et al: Hemodialysis-associated subclavian vein stenosis.  Kidney Int  1988;1156-1159.

108. Beathard GA, Marston WA: Endovascular management of thrombosed dialysis grafts.  Am J Kidney Dis  1998; 32:172-175.

109. Beathard GA: Percutaneous transvenous angioplasty in the treatment of vascular access stenosis.  Kidney Int  1992; 42:1390-1397.

110. Beathard GA: Thrombolysis versus surgery for the treatment of thrombosed dialysis access grafts.  J Am Soc Nephrol  1995; 6:1619-1624.

111. Valji K, Bookstein JJ, Roberts AC, Davis GB: Pharmacomechanical thrombolysis and angioplasty in the management of clotted hemodialysis grafts: Early and late clinical trials.  Radiology  1991; 178:243-247.

112. Beathard GA, Trerotola SO, Vesely TM, et al: What is the current and future status of interventional nephrology?.  Semin Dial  2005; 18(5):370-379.

113. Beathard GA, Litchfield T: Effectiveness and safety of dialysis vascular access procedures performed by interventional nephrologists.  Kidney Int  2004; 66(4):1622-1632.

114. Schmitz PG, McCloud LK, Reikes ST, et al: Prophylaxis of hemodialysis graft thrombosis with fish oil: Double-blind, randomized, prospective trial.  J Am Soc Nephrol  2002; 13(1):184-190.

115. Sreedhara R, Himmelfarb J, Lazarus JM, Hakim RM: Antiplatelet therapy in graft thrombosis: Results of a prospective, randomized double-blind study.  Kidney Int  1994; 45:1477-1483.

116. Kaufman J, O'Connor T, Cronin R, et al: Combination aspirin plus clopidogrel in the prevention of hemodialysis access graft thrombosis. Abstract.  J Am Soc Nephrol  2001; 12:219A.

117. Saran R, Dykstra DM, Wolfe RA, et al: Association between vascular access failure and the use of specific drugs: The Dialysis Outcomes and Practice Patterns Study (DOPPS).  Am J Kidney Dis  2002; 40(6):1255-1263.

118. Gradzki R, Dhingra RK, Port FK, et al: Use of ACE inhibitors is associated with prolonged survival of arteriovenous grafts.  Am J Kidney Dis  2001; 38(6):1240-1244.

119. Kaplow LS, Goffinet JA: Profound neutropenia during the early phase of hemodialysis.  JAMA  1968; 203:1135-1137.

120. Mujais S, Schmidt B: Operating Characteristis of Hollow-Fiber Dialyzers,  3rd ed. Norwalk, CT: Appleton & Lange; 1995:8.

121. Gohl H, Konstantin P: Membranes and filters for hemofiltration.   In: Henderson L, Quellhorst E, Baldamus C, Lysaght M, ed. Hemofiltration,  Berlin: Springer-Verlag; 1986:41-56.

122. Leypoldt JK, Frigon RP, Henderson LW: Macromolecular charge affects hemofilter solute sieving.  Artif Intern Organs  1986; 32:384-387.

123. Kaplow LS, Goffinet JA: Profound neutropenia during the early phase of hemodialysis.  JAMA  1968; 203:1135-1137.

124. Akizawa T, Kitaoka T, Koshikawa S, et al: Development of a regenerated cellulose non-complement activating membrane for hemodialyisis.  ASAIO Trans  1986; 32:76-80.

125. Horl WH, Schaefer RM, Heidland A: Effect of different dialyzers on proteinases and proteinase inhibitors during hemodialysis.  Am J Nephrol  1985; 5:320-326.

126. Sawada K, Malchesky PS, Guidubaldi JM, et al: In vitro evaluation of a relationship between human serum- or plasma-material interaction and polymer bulk hydroxyl and surface oxygen content.  ASAIO J  1993; 39:910-917.

127. Cheung AK, Chenoweth DE, Otsuka D, Henderson LW: Compartmental distribution of complement activation products in artificial kidneys.  Kidney Int  1986; 30:74-80.

128. Depner TA: Benefits of more frequent dialysis: Lower TAC at the same Kt/V.  Nephrol Dial Transplant  1998; 13:20-24.

129. Ronco C, Brendolan A, Milan M, et al: Impact of biofeedback-induced cardiovascular stability on hemodialysis tolerance and efficiency.  Kidney Int  2000; 58(2):800-808.

130. Ronco C, Ghezzi PM, La Greca G: The role of technology in hemodialysis.  J Nephrol  1999; 12(suppl 2):68-81.

131. Krivitsky NM: Theory and validation of access flow measurement by dilution technique during hemodialysis.  Kidney Int  1995; 48:244-250.

132. Leypoldt JK, Cheung AK, Steuer RR, et al: Determination of circulating blood volume by continuously monitoring hematocrit during hemodialysis.  J Am Soc Nephrol  1995; 6:214-219.

133. Steuer RR, Germain MJ, Leypoldt JK, Cheung AK: Enhanced fluid removal guided by blood volume monitoring during chronic hemodialysis.  Artif Organs  1998; 22:627-632.

134. Steuer RR, Leypoldt JK, Cheung AK, et al: Reducing symptoms during hemo-dialysis by continuously monitoring the hematocrit.  Am J Kidney Dis  1996; 27:525-532.

135. Steuer RR, Leypoldt JK, Cheung AK, et al: Hematocrit as an indicator of blood volume and a predictor of intradialytic morbid events.  ASAIO J  1994; 40:691-696.

136. Dunea G, Mahurkar SD, Mamdani B, Smith EC: Role of aluminum in dialysis dementia.  Ann Intern Med  1978; 88:502-504.

137. Coburn JW, Norris KC, Sherrard DJ, et al: Toxic effects of aluminum in end-stage renal disease: Discussion of a case.  Am J Kidney Dis  1988; 12:171-184.

138. Llach F, Felsenfeld AJ, Coleman MD, et al: The natural course of dialysis osteomalacia.  Kidney Int Suppl  1986; 18:74-79.

139. Tokars JI, Alter MJ, Favero MS, et al: National surveillance of hemodialysis associated diseases in the United States, 1990.  ASAIO Trans  1993; 39:71-80.

140. Alter MJ, Favero MS, Moyer LA, Bland LA: National surveillance of hemodialysis associated diseases in the United States, 1989.  ASAIO Trans  1991; 37:97-109.

141. Topple M, Bland L, Favero M, et al: Investigation of hemolytic anemia after chloramine exposure in a dialysis center. Letter.  ASAIO J  1988; 34:1060.

142. Yawata Y, Kjellstrand C, Buselmeier T, et al: Hemolysis in dialyzed patietns: Tap water-induced red blood cell metabolic deficiency.  Trans Am Soc Artif Intern Organs  1972; 18:301-304.

143. Neilan BA, Ehlers SM, Kolpin CF, Eaton JW: Prevention of chloramine-induced hemolysis in dialyzed patients.  Clin Nephrol  1978; 10:105-108.

144. Gordon SM, Drachman J, Bland LA, et al: Epidemic hypotension in a dialysis center caused by sodium azide.  Kidney Int  1990; 37:110-115.

145. Lough J, Noonan R, Gagnon R, Kaye M: Effects of fluoride on bone in chronic renal failure.  Arch Pathol  1975; 99:484-487.

146. National News: Dialysis patients in Chicago die from fluoride poisoning: FDA issues safety alert.  Contemp Dial Nephrol  1993; 17:110-111.

147. Freeman RM, Lawton RL, Chamberlain MA: Hard-water syndrome.  N Engl J Med  1967; 276:1113-1118.

148. Evans DB, Slapak M: Pancreatitis in the hard water syndrome.  Br Med J  1975; 3:748-752.

149. Carlson DJ, Shapiro FL: Methemoglobinemia from well water nitrates: A complication of home dialysis.  Ann Intern Med  1970; 73:757-759.

150. Manzler AD, Schreiner AW: Copper-induced acute hemolytic anemia: A new complication of hemodialysis.  Ann Intern Med  1970; 73:409-412.

151. Matter BJ, Pederson J, Psimenos G, Lindeman RD: Lethal copper intoxication in hemodialysis.  Trans Am Soc Artif Intern Organs  1969; 15:309-315.

152. Gallery E, Blomfield J, Dixon S: Acute zinc toxicity in haemodialysis.  Br Med J  1973; 4:33.

153.   Centers for Disease Control and Prevention (CDC): Formaldehyde Intoxication Associated with Hemodialysis. California Epidemic Investigation Report, EPI 81-733-2. Atlanta, CDC, May 7, 1984.

154. Orringer EP, Mattern WD: Formaldehyde-induced hemolysis during chronic hemodialysis.  N Engl J Med  1976; 294:1416-1420.

155. Johnson WJ, Taves DR: Exposure to excessive fluoride during hemodialysis.  Kidney Int  1974; 5:451-454.

156. Favero MS, Petersen NJ, Carson LA, et al: Gram-negative water bacteria in hemodialysis systems.  Health Lab Sci  1975; 12:321-334.

157. Bland L, Favero ML: Microbiologic aspects of hemodialysis systems.  Association for the Advancement of Medical Instrumentation (eds): Dialysis, Vol 3. Arlington, VA: American National Standards; 1993:257-265.

158. Lowry PW, Beck-Sagne CM, Bland LA, et al: Mycobacterium chelonei infection among patients receiving high-flux dialysis in a hemodialysis clinic in California.  J Infect Dis  1990; 161:85-90.

159. Carson LA, Bland LA, Cusick LB, et al: Prevalence of nontuberculous mycobacteria in water supplies of hemodialysis centers.  Appl Environ Microbiol  1988; 54:3122-3125.

160. Carson LA, Petersen NJ, Favero ME, Aguero SM: Growth characteristics of atypical mycobacteria in water and their comparative resistence to disinfectants.  Appl Environ Microbiol  1978; 36:839-846.

161. Bolan G, Reingold AL, Carson LA, et al: Infections with Mycobacterium chelonei in patients receiving dialysis and using processed hemodialyzers.  J Infect Dis  1985; 152:1013-1019.

162. Favero M: Distinguishing between high-level disinfection, reprocessing and sterilization.  Association for the Advancement of Medical Instrumentation (eds): Reuse of Disposables: Implications for Quality Health Care and Cost Containment. Technical Assessment Report No. 6,  Arlington, VA: American National Standards; 1983:19-20.

163. Favero M, Bland L: Microbiologic principles applied to reprocessing hemodialyzers.   In: Deane N, Wineman R, Bemis J, ed. Guide to Reprocessing of Hemodyalizers,  Boston: Martinus Nijhoff; 1986:63-73.

164. Gordon SM, Oettinger CW, Bland LA, et al: Pyrogenic reactions in patients receiving conventional, high-efficiency, or high-flux hemodialysis treatments with bicarbonate dialysate containing high concentrations of bacteria and endotoxin.  J Am Soc Nephrol  1992; 2:1436-1444.

165. Jones DM, Tobin BM, Harlow GR, Ralston AJ: Antibody production in patients on regular haemodialysis to organisms present in dialysate.  Proc Eur Dial Transplant Assoc  1972; 9:575-576.

166. Hindman SH, Favero MS, Carson LA, et al: Pyrogenic reactions during haemodialysis caused by extramural endotoxin.  Lancet  1975; 2:732-734.

167. Raij L, Shapiro FL, Michael AF: Endotoxemia in febrile reactions during hemodialysis.  Kidney Int  1973; 4:57-60.

168. Passavanti G, Buongiorno E, De Fino G, et al: The permeability of dialytic membranes to endotoxins: Clinical and experimental findings.  Int J Artif Organs  1989; 12:505-508.

169. Laurence RA, Lapierre ST: Quality of hemodialysis water: A 7-year multicenter study.  Am J Kidney Dis  1995; 25:738-750.

170. Dinarello CA: Interleukin-1 and its biologically related cytokines.  Adv Immunol  1989; 44:153-205.

171. Loppnow H, Brade H, Durrbaum I, et al: IL-1 induction—Capacity of defined lipopolysaccharide partial structures.  J Immunol  1989; 142:3229.3238

172. Duff GW, Atkins E: The detection of endotoxin by in vitro production of endogenous pyrogen: Comparison with limulus amebocyte lysate gelation.  J Immunol Methods  1982; 52:323-331.

173. Lonnemann G, Bingel M, Floege J, et al: Detection of endotoxin-like interleukin-1 inducing activity during in vitro dialysis.  Kidney Int  1988; 33:29-35.

174. Evans RC, Holmes CJ: In vitro study of the transfer of cytokine-inducing substances across selected high-flux hemodialysis membranes.  Blood Purif  1991; 9:9-101.

175. Gault MH, Duffett AL, Murphy JF, Purchase LH: In search of sterile, endotoxin-free dialysate.  ASAIO J  1992; 38:431-435.

176. Urena P, Herbelin A, Zingraff J, et al: Permeability of cellulosic and non-cellulosic membranes to endotoxin subunits and cytokine production during in-vitro haemodialysis.  Nephrol Dial Transplant  1992; 7:16-28.

177.   Centers for Disease Control and Prevention (CDC): Clusters of Bacteremia and Pyrogenic Reactions in Hemodialysis Patients: Georgia. Epidemic Investigation Report, EPI 86-65-2. Atlanta, CDC, April 22, 1987.

178. Centers for Disease Control and Prevention (CDC) : Bacteremia associated with reuse of disposable hollow-fiber hemodialyzers.  MMWR Morb Mortal Wkly Rep  1986; 35:417-418.

179.   Centers for Disease Control and Prevention (CDC): Pyrogenic Reactions in Patients Undergoing High-Flux Hemodialysis, California. Epidemic Investigation Report, EPI 86-80-2. Atlanta, CDC, June 1, 1987, p 48.

180. Alter MJ, Favero M, Miller JK, et al: Reuse of hemodialyzers. Results of nationwide surveillance for adverse effects.  JAMA  1988; 260:2073-2076.

181. Bland LA, Favero M, Oxborrow GS, et al: Effect of chemical germicides on the integrity of hemodialyzer membranes.  ASAIO Trans  1988; 34:172-175.

182. Jenkins S, Lin F, Lin R, et al: Pyrogenic reactions and Pseudomonas bacteremias in a hemodialysis center.  Dial Transplant  1987; 16:192-197.

183. Oberle M, Favero M, Carson L, et al: Fatal endotoxemia in dialysis patients at a summer camp.  Dial Transplant  1980; 9:549-550.

184. Bingel M, Lonnemann G, Shaldon S, et al: Human interleukin-1 production during hemodialyzis.  Nephron  1986; 43:161-163.

185. Bland L, Favero M: Microbiologic and endotoxin considerations in hemodialyzer reprocessing.  Association for the Advancement of Medical Instrumentation (eds): Dialysis, Vol 3. Arlington, VA: American National Standards; 1993:45-52.

186. Gazenfield-Gazit E, Eliahou HE: Endotoxin antibodies in patients on maintenance hemodialysis.  Isr J Med  1969; 5:1032-1036.

187. Chuang HY, Mohammad SF, Sharma NC, Mason RG: Interaction of human alpha-thrombin with artificial surfaces and reactivity of adsorbed alpha-thrombin.  J Biomed Mater Res  1980; 14:467-476.

188. Chueng AK, Parker CJ, Janatova J, Brynda E: Modulation of complement activation on hemodialysis membranes by immobilized heparin.  J Am Soc Nephrol  1992; 8:1328-1337.

189. Feldman HI, Kinosian M, Bilker WB, et al: Effct of dialyzer reuse on survival of patients with hemodialysis.  JAMA  1996; 276:620-625.

190. Port FK, Wolfe RA, Hulbert-Shearon TE, et al: Mortality risk by hemodialyzer reuse practice and dialyzer membrane characteristics: Results from the USRDS dialysis morbidity and mortality study.  Am J Kidney Dis  2001; 37(2):276-286.

191. Finelli L, Miller JT, Tokars JI, et al: National surveillance of dialysis-associated diseases in the United States, 2002.  Semin Dial  2005; 18(1):52-61.

192. Baris E, McGregor M: The reuse of hemodialyzers: An assessment of safety and potential savings.  Can Med Assoc J  1993; 148:175-183.

193. Lowrie EG, Li Z, Ofsthun N, Lazarus JM: Reprocessing dialysers for multiple uses: Recent analysis of death risks for patients.  Nephrol Dial Transplant  2004; 19(11):2823-2830.

194. Lacson Jr E, Lazarus JM: Dialyzer best practice: Single use or reuse?.  Semin Dial  2006; 19(2):120-128.

195. Cazenave JP, Mulvihill J: Interaction of blood with surfaces: Hemocompatibility and thromboresistance of biomaterials.  Contrib Nephrol  1988; 62:118-127.

196. Ouseph R, Ward RA: Anticoagulation for intermittent hemodialysis.  Semin Dial  2000; 13(3):181-187.

197. Gordon LA, Simon ER, Richards JM: Studies in regional heparinization. II. Artificial kidney hemodialysis without systemic heparinization—Preliminary report of a method using simultaneous infusion of heparin and protamine.  N Engl J Med  1956; 255:1063-1068.

198. Spencer P, Cozzi E, Easterling RE, Penner JA: Regional heparinization with hollow fiber artificial kidney.  Proc Am Assoc Nephrol Nurses Technicians  1977; 4:69-72.

199. Lindholm DD, Murray JS: A simplified method of regional heparinization during hemodialysis according to a predetermined dosage formula.  Trans Am Soc Artif Intern Organs  1964; 10:92-97.

200. Hampers CL, Blaufox MD, Merrill JP: Anticoagulation rebound after hemodialysis.  N Engl J Med  1966; 275:776-781.

201. Blaufox MD, Hampers CL, Merrill JP: Heparinization for hemodialysis.  Trans Am Soc Artif Intern Organs  1966; 12:207-209.

202. Schwab SJ, Onorato JJ, Sharar LR, Dennis PA: Hemodialysis without anticoagulation: One-year prospective trial in hospitalized patients at risk for bleeding.  Am J Med  1987; 83:405-410.

203. Caruana R, Raja R, Bush J, et al: Heparin-free dialysis: Comparative data and results in high-risk patients.  Kidney Int  1987; 31:1351-1355.

204. Charm S, Kurland G: Viscometry of human blood for shear rates 0-100,000 sec-1.  Nature  1965; 206:617-618.

205. Sigdell JE, Tersteegen B: Clearance of a dialyzer under varying operating conditions.  Artif Organs  1986; 10:219-225.

206. Hootkins R, Bourgeois B: The effect of ultrafiltration on dialysance. Mathematical theory and experimental verification.  Trans Am Soc Artif Intern Organs  1991; 37:M375-M377.

207. Waniewski J, Lucjanek P, Werynski A: Alternative descriptions of combined diffusive and convective mass transport in hemodialyzer.  Artif Organs  1993; 17:3-7.

208. Tyagi VP, Abbas M: An exact analysis for solute transport, due to simultaneous dialysis and ultrafiltration in a hollow-fibre artificial kidney.  Bull Math Biol  1987; 49:697-717.

209. Gupta BB, Jaffrin MY: In vitro study of the combind convection-diffusion mass transfer in hemodialyzers.  Int J Artif Organs  1984; 7:263-268.

210. Sherman R, Levy SS: Rate-related recirculation: The effect of altering blood flow on dialyzer recirculation.  Am J Kidney Dis  1991; 17:170-173.

211. Sherman RA: Recirculation revisited.  Semin Dial  1991; 4:221-223.

212. Schenditz D, Polaschegg HD, Levin NW, et al: Cardiopulmonary recirculation in dialysis—An underrecognized phenomenon.  ASAIO J  1992; 38:M194-M196.

213. Schneditz D, Kaufman AM, Polaschegg HD, et al: Cardiopulmonary recirculation during hemodialysis.  Kidney Int  1992; 42:1450-1456.

214. Depner TA, Rizwan S, Cheer A, Wagner J: Peripheral urea disequilibrium during hemodialysis is temperature-dependent.  ASAIO Trans  1991; 37:141-143.

215. Wang E, Schneditz D, Kaufman AM, Levin NW: Sensitivity and specificity of the thermodilution technique in detection of access recirculation.  Nephron  2000; 85(2):134-141.

216. Magnasco A, Alloatti S, Bonfant G, et al: Glucose infusion test: A new screening test for vascular access recirculation.  Kidney Int  2000; 57(5):2123-2128.

217. Katzarski KS, Charra B, Luik AJ, et al: Fluid state and blood pressure control in patients treated with long and short haemodialysis.  Nephrol Dial Transplant  1999; 14:369-375.

218. Charra B, Calemard E, Ruffet M, et al: Survival as an index of adequacy of dialysis.  Kidney Int  1992; 41:1286-1291.

219. Boquin E, Parnell S, Grondin G, et al: Cross-over study of the effect of different dialysate sodium concentrations in large surface area short-term dialysis.  Proc Clin Dial Transplant Forum  1977; 7:48-52.

220. Comty C, Rottka H, Shaldon S: Blood pressure control in patients with end-stage renal failure treated by intermittent hemodialysis.  Proc Eur Dial Transplant Assoc  1964; 1:209-214.

221. Drukker W, Jungerius NA, Alberts C: Report on regular dialysis treatment in Europe.  Proc Eur Dial Transplant Assoc  1964; 4:3.

222. Van Stone JC, Bauer J, Carey J: The effect of dialysate sodium concentration on body fluid distribution during hemodialysis.  Trans Am Soc Artif Intern Organs  1980; 26:383-386.

223. Boal R, Shideman JR, Meyer RM, et al: Effects and complications of high efficiency dialysis.  Nephron  1975; 15:151-160.

224. Levine J, Falk B, Henriquez M, et al: Effects of varying dialysate sodium using large surface area dialyzers.  Trans Am Soc Artif Intern Organs  1978; 24:139-141.

225. Heinrich WL, Hunt JM, Nixon JV: Increased ionized calcium and left ventricular contractility during hemodialysis.  N Engl J Med  1984; 310:19-23.

226. Bihaphala S, Bell AJ, Bennett CA, et al: Comparison of high- and low-sodium bicarbonate and acetate dialysis in stable chronic hemodialysis patients.  Clin Nephrol  1985; 23:179-183.

227. Port FK, Johnson WJ, Klass DW: Prevention of dialysis disequilibrium syndrome by use of high sodium concentration in the dialysate.  Kidney Int  1973; 3:327-333.

228. Flanigan MJ, Khairullah QT, Lim VS: Dialysate sodium delivery can alter chronic blood pressure management.  Am J Kidney Dis  1997; 29:383-391.

229. Dumler F, Grondin G, Levin NW: Sequential high/low sodium hemodialysis: An alternative to ultrafiltration.  Trans Am Soc Artif Intern Organs  1979; 25:351-353.

230. Raja R, Kramer M, Barber K, Chin S: Sequential changes in dialysate sodium (DNa) during hemodialysis.  Trans Am Soc Artif Intern Organs  1983; 29:649-651.

231. Daugirdas JT, Al-Kudsi RR, Ing TS, Norusis MJ: A double-blind evaluation of sodium gradient hemodialysis.  Am J Nephrol  1985; 5:163-168.

232. Song JH, Lee SW, Suh CK, Kim MJ: Time-averaged concentration of dialysate sodium relates with sodium load and interdialytic weight gain during sodium-profiling hemodialysis.  Am J Kidney Dis  2002; 40(2):291-301.

233. Ketchersid TL, Van Stone JC: Dialysate potassium.  Semin Dial  1991; 4:46-49.

234. Feig PU, Shook A, Sterns RH: Effect of potassium removal by hemodialysis.  Nephron  1981; 27:25-30.

235. Sherman RA, Hwang ER, Bernholc AS, Eisinger RP: Variability in potassium removal by hemodialysis.  Am J Nephrol  1986; 6:284-288.

236. Hou S, McElroy PA, Nootes S, Beach M: Safety and efficacy of low potassium dialysis.  Am J Kidney Dis  1989; 13:137-143.

237. Karnik JA, Young BS, Lew NL, et al: Cardiac arrest and sudden death in dialysis units.  Kidney Int  2001; 60(1):350-357.

238. Wing AJ: Optimum calcium concentration of dialysis fluid for hemodialysis.  Br Med J  1968; 4:145-149.

239. Bone JM, Davison AM, Robson JS: Role of dialysate concentration on calcium mass transfer during maintenance hemodialysis.  Lancet  1972; 1:1047-1048.

240. Fellner SK, Lang RM, Neumann A: Physiologic mechanisms for calcium-induced changes in systemic arterial pressure in stable dialysis patients.  Hypertension  1989; 13:213-218.

241. Maynard JC, Cruz C, Kleerekoper M, Levin NW: Blood pressure response to changes in serum ionized calcium during hemodialysis.  Ann Intern Med  1986; 104:358-361.

242. Sherman RA, Bialy GB, Gazinski B, et al: The effect of dialysate calcium levels on blood pressure during hemodialysis.  Am J Kidney Dis  1986; 8:244-247.

243. Wang CL, Lee WL, Wu MJ, et al: Increased QTc dispersion and mortality in uremic patients with acute myocardial infarction.  Am J Kidney Dis  2002; 39(3):539-548.

244. Vaporean ML, Van Stone JC: Dialysate magnesium.  Semin Dial  1993; 6:46-49.

245. Kolff WJ, Berk HT: The artificial kidney: A dialyzer with a great area.  Acta Med Scand  1944; 117:121-129.

246. Mion CM, Hegstrom RM, Boen ST, Scribner BH: Substitution of sodium acetate for bicarbonate in the bath fluid for hemodialysis.  Trans Am Soc Artif Intern Organs  1964; 10:110-115.

247. Palmer BF: The effect of dialysate composition on systemic hemodynamics.  Semin Dial  1992; 5:30-33.

248. Mendelssohn S, Swartz CD, Yudis M, et al: High glucose concentration dialysate in chronic hemodialysis.  Trans Am Soc Artif Intern Organs  1967; 13:249-252.

249. Rosborough DC, Van Stone JC: Dialysis glucose.  Semin Dial  1993; 6:260-263.

250. Berkes SL, Khan SI, Chazan JA, Garella S: Prolonged hemodialysis from overheated dialysate.  Ann Intern Med  1975; 83:363-364.

251. Fortner RW, Nowakowski A, Carter CB, et al: Death due to overheated dialysate during dialysis.  Ann Intern Med  1970; 73:443-444.

252. Maggiore Q, Pizzarelli F, Sisca S, et al: Blood temperature and vascular stability during hemodialysis and hemofiltration.  ASAIO Trans  1982; 28:523-527.

253. Sherman RA, Faustino EF, Bernholc AS, Eisinger RP: Effects of variations in dialysate temperature on blood pressure during hemodialysis.  Am J Kidney Dis  1984; 4:66-68.

254. Coli U, Landinki S, Lucatello S, et al: Cold as cardiovascular stabilizing factor in hemodynamic evaluation.  ASAIO Trans  1983; 29:71-75.

255. Lindhold T, Thysell H, Yamamoto Y, et al: Temperature and vascular stability in hemodialysis.  Nephron  1985; 39:130-133.

256. Kerr PG, van Bakel C, Dawborn JK: Assessment of the asymptomatic benefits of cool dialysate.  Nephron  1989; 52:166-169.

257. Marcen R, Quereda C, Orfino L, et al: Hemodialysis with a low-temperature dialysate: A long-term experience.  Nephron  1988; 49:29-32.

258. Orfino L, Marcen R, Quereda C, et al: Epidemiology of symptomatic hypotension in hemodialysis: Is cool dialysate beneficial for all patients?.  Am J Nephrol  1990; 10:177-180.

259. Sherman RA, Rubin MP, Cody RP, Eisinger RP: Amelioration of hemodialysis-associated hypotension by the use of cool dialysate.  Am J Kidney Dis  1985; 5(2):124-127.

260. Shaldon S: Sequential ultrafiltration and dialysis.  Proc Eur Dial Transplant Assoc  1976; 13:300-304.

261. Asaba H, Bergstrom J, Furst P, et al: Sequential ultrafiltration and diffusion as alternatives to conventional hemodialysis.  Proc Clin Dial Transplant Forum  1976; 6:129-135.

262. Dheenan S, Henrich WL: Preventing dialysis hypotension: A comparison of usual protective maneuvers.  Kidney Int  2001; 59(3):1175-1181.

263. Jaeger JQ, Mehta RL: Assessment of dry weight in hemodialysis: An overview.  J Am Soc Nephrol  1999; 10:392-403.

264. Schneditz D, Roob J, Oswald M, et al: Nature and rate of vascular refilling during hemodialysis and ultrafiltration.  Kidney Int  1992; 42:1425-1433.

265. Schneditz D, Pogglitsch H, Horina J, Binswanger U: A blood protein monitor for the continuous measurement of blood volume changes during hemodialysis.  Kidney Int  1990; 38:342-346.

266. Leypoldt JK, Cheung AK, Steuer RR, et al: Determination of circulating blood volume by continuously monitoring hematocrit during hemodialysis.  J Am Soc Nephrol  1995; 6:214-219.

267. Donauer J, Kolblin D, Bek M, et al: Ultrafiltration profiling and measurement of relative blood volume as strategies to reduce hemodialysis-related side effects.  Am J Kidney Dis  2000; 36(1):115-123.

268. Oliver MJ, Edwards LJ, Churchill DN: Impact of sodium and ultrafiltration profiling on hemodialysis-related symptoms.  J Am Soc Nephrol  2001; 12(1):151-156.

269. Santoro A, Mancini E, Paolini F, et al: Blood volume regulation during hemodialysis.  Am J Kidney Dis  1988; 32:739-748.

270. USRDS : USRDS 1999 annual data report,  Bethesda, National Institutes of Health, 1999.

271. Parker III TP, Husni L, Huang W, et al: Survival of hemodialysis patients in the United States is improved with a greater quantity of dialysis.  Am J Kidney Dis  1994; 23:670-680.

272. Hakin RM: Assessing the adequacy of dialysis.  Kidney Int  1990; 37:822-832.

273. Gotch FA, Yarian S, Keen M: A kinetic survey of US hemodialysis prescriptions.  Am J Kidney Dis  1990; 15:511-515.

274. Sargent JA: Shortfalls in the delivery of dialysis.  Am J Kidney Dis  1990; 15:500-510.

275. Hakim RM, Breyer J, Ismail N, Schulman G: Effects of dose of dialysis on morbidity and mortality.  Am J Kidney Dis  1994; 23:661-669.

276. Collins AJ, Ma JZ, Umen A, Keshaviah P: Urea index and other predictors of hemodialysis patient survival.  Am J Kidney Dis  1994; 23:272-282.

277. Dau P: Plasmapheresis therapy in myasthenia gravis.  Muscle Nerve  1980; 3:468-482.

278. Babb AL, Farrell PC, Uvelli DA, Scribner BH: Hemodialyzer evaluation by examination of solute molecular spectra.  Trans Am Soc Artif Intern Organs  1972; 18:98-105.

279. Schoots A, Mikkers F, Cramers C, et al: Uremic toxins and the elusive middle molecules.  Nephron  1984; 38:1-8.

280. Vanholder R: Middle molecules as uremic toxins: Still a viable hypothesis.  Semin Dial  1994; 7:65-68.

281. Malachi T, Bogin E, Gafter U, Levi J: Parathyroid hormone effect on the fragility of human young and old red blood cells in uremia.  Nephron  1986; 42:52-57.

282. Bogin E, Massry SG, Harary I: Effect of parathyroid hormone on rat heart cells.  J Clin Invest  1981; 67:1215-1227.

283. Hajjar SM, Fadda GZ, Thanakitcharu P, et al: Reduced activity of Na(+)-K+ ATPase of pancreatic islets in chronic renal failure: Role of secondary hyperparathyroidism.  J Am Soc Nephrol  1992; 2:1355-1359.

284. Perna AF, Smogorzewski M, Massry SG: Effects of verapamil on the abnormalities in fatty acid oxidation of myocardium.  Kidney Int  1989; 36:453-457.

285. Fadda GZ, Akmal M, Premdas FH, et al: Insulin release from pancreatic islets: Effects of CRF and excess PTH.  Kidney Int  1988; 33:1066-1072.

286. Gejyo F, Brancaccio D: Beta-2-microglobulin: A possible new uremic toxin?.  Int J Artif Organs  1988; 11:3-5.

287. Lowrie EG, Laird NM, Parker TF, Sargent JA: Effect of the hemodialysis prescription of patient morbidity: Report from the National Cooperative Dialysis Study.  N Engl J Med  1981; 305(20):1176-1181.

288. Eschbach JW, Abdulhadi MH, Browne JK, et al: Recombinant human erythropoietin in anemic patients with end-stage renal disease.  Ann Intern Med  1989; 111:992-1000.

289. Sargent JA, Gotch FA: The analysis of concentration dependence of uremic lesions in clinical studies.  Kidney Int Suppl  1975;35-44.

290. Gotch FA, Sargent JA: A mechanistic analysis of the National Cooperative Dialysis Study (NCDS).  Kidney Int  1985; 28(3):526-534.

291. Keshavish P: Urea kinetic and middle molecule approaches to assessing the adequacy of hemodialysis and CAPD.  Kidney Int Suppl  1993; 40:28-38.

292. Charra B, Calemard E, Ruffet M, et al: Survival as an index of adequacy of dialysis.  Kidney Int  1992; 41:1286-1291.

293. Eggers PW: Mortality rates among dialysis patients in Medicare's End-Stage Renal Disease Program.  Am J Kidney Dis  1990; 15:414-421.

294. Owen WF, Lew NL, Liu Y, et al: The urea reduction ratio and serum albumin concentration as predictors of mortality in patients undergoing hemodialysis.  N Engl J Med  1993; 329:1001-1006.

295. Eknoyan G, Beck GJ, Cheung AK, et al: Effect of dialysis dose and membrane flux in maintenance hemodialysis.  N Engl J Med  2002; 347(25):2010-2019.

296. Eknoyan G, Levey A, Beck G, Schulman G: The Hemodialysis (HEMO) Study: Rationale for selection of interventions.  Semin Dial  1996; 9:24-33.

297. Lindsay RM, Kortas C: Hemeral (daily) hemodialysis.  Adv Ren Replace Ther  2001; 8(4):236-249.

298. Raj DS, Charra B, Pierratos A, Work J: In search of ideal hemodialysis: Is prolonged frequent dialysis the answer?.  Am J Kidney Dis  1999; 34:597-610.

299. Lacson Jr E, Diaz-Buxo JA: Daily and nocturnal hemodialysis: How do they stack up?.  Am J Kidney Dis  2001; 38(2):225-239.

300. Pierratos A, Ouwendyk M, Rassi M: Total body nitrogen increases on nocturnal dialysis. Abstract.  J Am Soc Nephrol  1999; 10:299A.

301. O'Sullivan DA, McCarthy JT, Kumar R, Williams AW: Improved biochemical variables, nutrient intake, and hormonal factors in slow nocturnal hemodialysis: A pilot study.  Mayo Clin Proc  1998; 73:1035-1045.

302. Raj DS, Ouwendyk M, Francoeur R, Pierratos A: Plasma amino acid profile on nocturnal hemodialysis.  Blood Purif  2000; 18(2):97-102.

303. Mucsi I, Hercz G, Uldall R, et al: Control of serum phosphate without any phosphate binders in patients treated with nocturnal hemodialysis.  Kidney Int  1998; 53:1399-1404.

304. Andre MB, Rembold SM, Pereira CM, Lugon JR: Prospective evaluation of an in-center daily hemodialysis program: Results of two years of treatment.  Am J Nephrol  2002; 22(5-6):473-479.

305. Simonsen O: Slow nocturnal dialysis (SND) as a rescue treatment for children and young patients with ESRD.  J Am Soc Nephrol  2002; 18:165A.

306. Chertow GM, Owen WF, Lazarus JM, et al: Exploring the reverse J-shaped curve between urea reduction ratio and mortality.  Kidney Int  1999; 56:1872-1878.

307. Frankenfield DL, McClellan WM, Helgerson SD, et al: Relationship between urea reduction ratio, demographic characteristics, and body weight for patients in the 1996 National ESRD Core Indicators Project.  Am J Kidney Dis  1999; 33:584-591.

308. Price DA, Owen Jr WF: African-Americans on maintenance dialysis: A review of racial differences in incidence, treatment, and survival.  Adv Ren Replace Ther  1997; 4:3-12.

309. Owen Jr WF, Chertow GM, Lazarus JM, Lowrie EG: Dose of hemodialysis and survival: Differences by race and sex.  JAMA  1998; 280:1764-1768.

310. Zhensheng L, Lew N, Lazarus M, et al: Comparing the urea reduction ratio and the urea product as outcome-based measures of hemodialysis dose.  Am J Kidney Dis  2000; 35:598-605.

311. Lowrie EG, Zhu X, Lew NL: Primary associates of mortality among dialysis patients: Trends and reassessment of Kt/V and urea reduction ratio as outcome-based measures of dialysis dose.  Am J Kidney Dis  1998; 32:16-31.

312. Kopple JD, Zhu X, Lew NL, Lowrie EG: Body weight-for-height relationships predict mortality in maintenance hemodialysis patients.  Kidney Int  1999; 56:1136-1148.

313. Wolfe RA, Ashby VB, Daugirdas JT, et al: Body size, dose of hemodialysis, and mortality.  Am J Kidney Dis  2000; 35(1):80-88.

314. Fleischmann E, Teal N, Dudley J, et al: Influence of excess weight on mortality and hospital stay in 1346 hemodialysis patients.  Kidney Int  1999; 55:1560-1567.

315. Lowrie EG, Chertow GM, Lew NL, et al: The urea (clearance × dialysis time) product (Kt) as an outcome-based measure of hemodialysis dose.  Kidney Int  1999; 56:729-737.

316. Gotch FA: Is Kt/V urea a satisfactory measure for dosing the newer dialysis regimens?.  Semin Dial  2001; 14(1):15-17.

317. Vanholder R, Argiles A, Baurmeister U, et al: Uremic toxicity: Present state of the art.  Int J Artif Organs  2001; 24(10):695-725.

318. Yavuz A, Tetta C, Ersoy FF, et al: Uremic toxins: A new focus on an old subject.  Semin Dial  2005; 18(3):203-211.

319. Vanholder R, De Smet R, Glorieux G, et al: Review on uremic toxins: Classification, concentration, and interindividual variability.  Kidney Int  2003; 63(5):1934-1943.

320. Erslev AJ: Humoral regulation of red cell production.  Blood  1953; 8:349-387.

321. Eschbach JW, Egrie JC, Downing MR, et al: Correction of anemia of end stage renal disease with recombinant human erythropoietin: Results of a combined phase I and II clinical trial.  N Engl J Med  1987; 316:73-78.

322. Eschbach W, Kelly MR, Haley NR, et al: Treatment of the anemia of progressive renal failure with recombinant human erythropoietin.  N Engl J Med  1989; 321:158-163.

323. Verbeelen D, Bossuyt A, Smitz J, et al: Hemodynamics of patients with renal failure treated with recombinant human erythropoietin.  Clin Nephrol  1989; 31:6-11.

324. Satoh K, Masuda T, Ikida Y, et al: Hemodynamic changes by recombinant erythropoietin therapy in hemodialyzed patients.  Hypertension  1990; 15:262-266.

325. Mayer G, Cada EV, Watzinger U, et al: Hemodynamic effects of partial correction of chronic anemia by recombinant human erythropoietin in patients on dialysis.  Am J Kidney Dis  1991; 17:286-289.

326. Low I, Grutzmacher P, Bergmann M, Schoeppe W: Echocardiographic findings in patients on maintenance hemodialysis substituted with recombinant human erythropoietin.  Clin Nephrol  1989; 31:26-30.

327. Cannella G, La Canna G, Sandrini M, et al: Reversal of left ventricular hypertrophy following recombinant human erythropoietin treatment of anaemic dialyzed uraemic patients.  Nephrol Dial Transplant  1991; 6:31-37.

328. Pascual J, Teruel JL, Moya JL, et al: Regression of left ventricular hypertrophy after partial correction of anemia with erythropoietin in patients on hemodialysis A prospective study.  Clin Nephrol  1991; 35:280-287.

329. Zehnder C, Zumer M, Sulzer M, et al: Influence of long-term amelioration of anemia and blood pressure control on left ventricular hypertrophy in hemodialyzed patients.  Nephron  1992; 61:21-25.

330. Bommer J, Kugel M, Schwobel B, et al: Improved sexual function during recombinant human erythropoietin therapy.  Nephrol Dial Transplant  1990; 5:204-207.

331. Shaefer RM, Kokot P, Wernze H, et al: Improved sexual function in hemodialysis patients on recombinant erythropoietin: A possible role for prolactin.  Clin Nephrol  1989; 31:1-5.

332. Braumann KM, Nonnast DB, Boning D, Bocker A: Improved physical performance after treatment of renal anemia with recombinant human erythropoietin.  Nephron  1991; 58:129-134.

333. Delano BG: Improvements in quality of life following treatment with r-HuEPO in anemic hemodialysis patients.  Am J Kidney Dis  1989; 14:14-18.

334. Evans RW, Rader B, Manninen DL, et al: The quality of life of hemodialysis recipients treated with recombinant human erythropoietin.  JAMA  1990; 263:825-830.

335. Evans RW: Recombinant human erythropoietin and the quality of life of end-stage renal disease patients: A comparative analysis.  Am J Kidney Dis  1991; 18:62-70.

336. Ali M, Fayemi AO, Rigolosi R, et al: Hemosiderosis in hemodialysis patients: An autopsy study of 50 cases.  JAMA  1980; 244:343-345.

337. Bregman H, Gelfand MC, Winchester JF, et al: HLA-linked iron overload and myopathy in maintenance hemodialysis patients.  Trans Am Soc Artif Intern Organs  1980; 26:366-368.

338. Hakim RM, Stivelman JC, Schulman G, et al: Iron overload and mobilization in long-term hemodialysis patients.  Am J Kidney Dis  1987; 10:293-299.

339. Kaufman JS, Reda DJ, Fye CL, et al: Subcutaneous compared with intravenous epoietin in patients receiving hemodialysis.  N Engl J Med  1998; 339:578-583.

340. Besarab A, Reyes CM, Hornberger J: Meta-analysis of subcutaneous versus intravenous epoetin in maintenance treatment of anemia in hemodialysis patients.  Am J Kidney Dis  2002; 40(3):439-446.

341. Besarab A, Bolton WK, Browne JK, et al: The effects of normal as compared with low hematocrit values in patients with cardiac diseae who are receiving hemodialysis and erythropoietin.  N Engl J Med  1998; 339:584-590.

342. Foley RN, Parfrey PS, Morgan J, et al: Effect of hemoglobin levels in hemodialysis patients with asymptomatic cardiomyopathy.  Kidney Int  2000; 58(3):1325-1335.

343. Moreno F, Sanz-Guajardo D, Lopez-Gomez JM, et al: Increasing the hematocrit has a beneficial effect on quality of life and is safe in selected hemodialysis patients. Spanish Cooperative Renal Patients Quality of Life Study Group of the Spanish Society of Nephrology.  J Am Soc Nephrol  2000; 11(2):335-342.

344. Collins AJ, Li S, St Peter W, et al: Death, hospitalization, and economic associations among incident hemodialysis patients with hematocrit values of 36 to 39%.  J Am Soc Nephrol  2001; 12(11):2465-2473.

345. Robinson BM, Joffe MM, Berns JS, et al: Anemia and mortality in hemodialysis patients: accounting for morbidity and treatment variables updated over time.  Kidney Int  2005; 68(5):2323-2330.

346. Radtke HW, Frei U, Erbes PM, et al: Improving anemia by hemodialysis: Effect on serum erythropoietin.  Kidney Int  1980; 17:382-387.

347. Santiago GC, Sreepada Rao TK, Laird NM: Effect of dialysis therapy on the hematopoietic system: The NCDS.  Kidney Int Suppl  1983; 23:S95-S100.

348. Lindsay RM, Burton JA, Dargie HJ, et al: Dialyzer blood loss.  Clin Nephrol  1973; 1:24-34.

349. Sweet SJ, McCarthy S, Steingart R, Callahan T: Hemolytic reactions mechanically induced by kinked hemodialysis lines.  Am J Kidney Dis  1996; 27:262-266.

350. Hampers CL, Streiff R, Nathan DG, et al: Megaloblastic hematopoiesis in uremia and in patients on long-term dialysis.  N Engl J Med  1967; 276:551-554.

351. Ibbotson T, Goa KL: Darbepoetin alfa.  Drugs  2001; 61(14):2097-2104.discussion 2105-2106

352. Locatelli F, Olivares J, Walker R, et al: Novel erythropoiesis stimulating protein for treatment of anemia in chronic renal insufficiency.  Kidney Int  2001; 60(2):741-747.

353. Macdougall IC, Gray SJ, Elston O, et al: Pharmacokinetics of novel erythropoiesis stimulating protein compared with epoietin alfa in dialysis patients.  J Am Soc Nephrol  1999; 10:2392-2395.

354. Tang IY, Vrahnos D, Valaitis D, Lau A: Vascular access thrombosis during recombinant human erythropoietin therapy.  ASAIO J  1992; 38:M528-M531.

355. Wingard RL, Parker RA, Ismail N, Hakim RM: Efficacy or oral iron therapy in patients receiving recombinant human erythropoietin.  Am J Kidney Dis  1995; 25:433-439.

356. Fishbane S, Shapiro W, Dutka P, et al: A randomized trial of iron deficiency testing strategies in hemodialysis patients.  Kidney Int  2001; 60(6):2406-2411.

357. Besarab A, Amin N, Ahsan M, et al: Optimization of epoetin therapy with intravenous iron therapy in hemodialysis patients.  J Am Soc Nephrol  2000; 11(3):530-538.

358. Macdougall IC, Chandler G, Elston O, Harchowal J: Beneficial effects of adopting an aggressive intravenous iron policy is a hemodialysis unit.  Am J Kidney Dis  1999; 23(suppl):S40-S46.

359. Feldman HI, Santanna J, Guo W, et al: Iron administration and clinical outcomes in hemodialysis patients.  J Am Soc Nephrol  2002; 13(3):734-744.

360. Salahudeen AK, Oliver B, Bower JD, Roberts 2nd LJ: Increase in plasma esterified F2-isoprostanes following intravenous iron infusion in patients on hemodialysis.  Kidney Int  2001; 60(4):1525-1531.

361. Roob JM, Khoschsorur G, Tiran A, et al: Vitamin E attenuates oxidative stress induced by intravenous iron in patients on hemodialysis.  J Am Soc Nephrol  2000; 11(3):539-549.

362. Fishbane S, Ungureanu VD, Maesaka JL, et al: The safety of intravenous iron dextran in hemodialysis patients.  Am J Kidney Dis  1996; 28:529-534.

363. Kalantar-Zadeh K, Regidor DL, McAllister CJ, et al: Time-dependent associations between iron and mortality in hemodialysis patients.  J Am Soc Nephrol  2005; 16(10):3070-3080.

364. Michael B, Coyne DW, Fishbane S, et al: Sodium ferric gluconate complex in hemodialysis patients: Adverse reactions compared to placebo and iron dextran.  Kidney Int  2002; 61(5):1830-1839.

365. Chandler G, Harchowal J, Macdougall IC: Intravenous iron sucrose: Establishing a safe dose.  Am J Kidney Dis  2001; 38(5):988-991.

366. Coyne DW, Adkinson NF, Nissenson AR, et al: Sodium ferric gluconate complex in hemodialysis patients. II. Adverse reactions in iron dextran-sensitive and dextran-tolerant patients.  Kidney Int  2003; 63(1):217-224.

367. Van Wyck DB, Cavallo G, Spinowitz BS, et al: Safety and efficacy of iron sucrose in patients sensitive to iron dextran: North American clinical trial.  Am J Kidney Dis  2000; 36(1):88-97.

368. Spinowitz BS, Schwenk MH, Jacobs PM, et al: The safety and efficacy of ferumoxytol therapy in anemic chronic kidney disease patients.  Kidney Int  2005; 68(4):1801-1807.

369. Foley RN, Parfrey PS, Sarnak MJ: Clinical epidemiology of cardiovascular disease in chronic renal disease.  Am J Kidney Dis  1998; 32:S112-S119.

370. Lindner A, Charra B, Sherrard DJ, Scribner BH: Accelerated atherosclerosis and prolonged maintenance hemodialysis.  N Engl J Med  1974; 290:697-701.

371. Scribner BH: The treatment of chronic uremia by means of intermittent hemodialysis: A preliminary report.  Trans Am Soc Artif Intern Organs  1960; 6:114-122.

372. Goldsmith DJ, Covic A: Coronary artery disease in uremia: Etiology, diagnosis, and therapy.  Kidney Int  2001; 60(6):2059-2078.

373. Sinha SK: Coronary angiography and coronary artery by-pass grafts in diabetics.  Diabetes Res Clin Pract  1996; 30:89-92.

374. Schwarz U, Buzello M, Ritz E, et al: Morphology of coronary atherosclerotic lesions in patients with end-stage renal failure.  Nephrol Dial Transplant  2000; 15(2):218-223.

375. Cheung AK, Wu LL, Kabitz C, Leypoldt JK: Atherogenic lipids and lipoproteins in hemodialysis patients.  Am J Kidney Dis  1993; 11:271-276.

376. Cressman MD, Aboud D, O'Neil J, et al: Lp(a) and premature mortality during chronic hemodialysis treatment.  Chem Phys Lipids  1994; 67-68:419-427.

377. Goldwasser P, Kaldas AI, Barth RH: Rise in serum albumin and creatinine in the first half year on hemodialysis.  Kidney Int  1999; 56:2260-2268.

378. Kronenberg F, Neyer U, Lhotta K, et al: The low molecular weight apo(a) phenotype is an independent predictor for coronary artery disease in hemodialysis patients: A prospective follow-up.  J Am Soc Nephrol  1999; 10:1027-1036.

379. Longenecker JC, Klag MJ, Marcovina SM, et al: Small apolipoprotein(a) size predicts mortality in end-stage renal disease: The CHOICE study.  Circulation  2002; 106(22):2812-2818.

380. Longenecker JC, Klag MJ, Marcovina SM, et al: High lipoprotein(a) levels and small apolipoprotein(a) size prospectively predict cardiovascular events in dialysis patients.  J Am Soc Nephrol  2005; 16(6):1794-1802.

381. Bostom AG, Lathrop L: Hyperhomocysteinemia in end-stage renal disease: Prevalence, etiology, and potential relationship to arteriosclerotic outcomes.  Kidney Int  1997; 52:10-20.

382. Suliman ME, Qureshi AR, Barany P, et al: Hyperhomocysteinemia, nutritional status, and cardiovascular disease in hemodialysis patients.  Kidney Int  2000; 57(4):1727-1735.

383. Lonn E, Yusuf S, Arnold MJ, et al: Homocysteine lowering with folic acid and B vitamins in vascular disease.  N Engl J Med  2006; 354(15):1567-1577.

384. Bonaa KH, Njolstad I, Ueland PM, et al: Homocysteine lowering and cardio-vascular events after acute myocardial infarction.  N Engl J Med  2006; 354(15):1578-1588.

385. Kaysen GA: The microinflammatory state in uremia: Causes and potential consequences.  J Am Soc Nephrol  2001; 12(7):1549-1557.

386. Kaysen GA, Chertow GM, Adhikarla R, et al: Inflammation and dietary protein intake exert competing effects on serum albumin and creatinine in hemodialysis patients.  Kidney Int  2001; 60(1):333-340.

387. Cheung AK, Sarnak MJ, Yan G, et al: Atherosclerotic cardiovascular disease risks in chronic hemodialysis patients.  Kidney Int  2000; 58(1):353-362.

388. Arici M, Walls J: End-stage renal disease, atherosclerosis, and cardiovascular mortality: Is C-reactive protein the missing link?.  Kidney Int  2001; 59(2):407-414.

389. Stenvinkel P, Heimburger O, Paultre F, et al: Strong association between malnutrition, inflammation, and atherosclerosis in chronic renal failure.  Kidney Int  1999; 51:1899-1911.

390. Kato A, Odamaki M, Takita T, et al: Association between interleukin-6 and carotid atherosclerosis in hemodialysis patients.  Kidney Int  2002; 61(3):1143-1152.

391. Pecoits-Filho R, Barany P, Lindholm B, et al: Interleukin-6 is an independent predictor of mortality in patients starting dialysis treatment.  Nephrol Dial Transplant  2002; 17(9):1684-1688.

392. Sela S, Shurtz-Swirski R, Cohen-Mazor M, et al: Primed peripheral polymorphonuclear leukocyte: A culprit underlying chronic low-grade inflammation and systemic oxidative stress in chronic kidney disease.  J Am Soc Nephrol  2005; 16(8):2431-2438.

393. Amabile N, Guerin AP, Leroyer A, et al: Circulating endothelial microparticles are associated with vascular dysfunction in patients with end-stage renal failure.  J Am Soc Nephrol  2005; 16(11):3381-3388.

394. Koc M, Richards HB, Bihorac A, et al: Circulating endothelial cells are associated with future vascular events in hemodialysis patients.  Kidney Int  2005; 67(3):1078-1083.

395. Himmelfarb J, Stenvinkel P, Ikizler TA, Hakim RM: The elephant in uremia: Oxidant stress as a unifying concept of cardiovascular disease in uremia.  Kidney Int  2002; 62(5):1524-1538.

396. Himmelfarb J, McMenamin ME, Loseto G, Heinecke JW: Myeloperoxidase-catalyzed 3-chlorotyrosine formation in dialysis patients.  Free Radic Biol Med  2001; 31(10):1163-1169.

397. Boaz M, Smetana S, Weinstein T, et al: Secondary prevention with antioxidants of cardiovascular disease in endstage renal disease (SPACE): Randomised placebo-controlled trial.  Lancet  2000; 356(9237):1213-1218.

398. Tepel M, van der Giet M, Statz M, et al: The antioxidant acetylcysteine reduces cardiovascular events in patients with end-stage renal failure: A randomized, controlled trial.  Circulation  2003; 107(7):992-995.

399. Tozawa M, Iseki K, Iseki C, Takishita S: Pulse pressure and risk of total mortality and cardiovascular events in patients on chronic hemodialysis.  Kidney Int  2002; 61(2):717-726.

400. Klassen PS, Lowrie EG, Reddan DN, et al: Association between pulse pressure and mortality in patients undergoing maintenance hemodialysis.  JAMA  2002; 287(12):1548-1555.

401. Guerin AP, London GM, Marchais SJ, Metivier F: Arterial stiffening and vascular calcifications in end-stage renal disease.  Nephrol Dial Transplant  2000; 15(7):1014-1021.

402. Blacher J, Safar ME, Guerin AP, et al: Aortic pulse wave velocity index and mortality in end-stage renal disease.  Kidney Int  2003; 63(5):1852-1860.

403. Beaubien ER, Pylypchuk GB, Akhtar J, Biem HJ: Value of corrected QT interval dispersion in identifying patients initiating dialysis at increased risk of total and cardiovascular mortality.  Am J Kidney Dis  2002; 39(4):834-842.

404. Zoccali C, Mallamaci F, Tripepi G: Nocturnal hypoxemia predicts incident cardiovascular complications in dialysis patients.  J Am Soc Nephrol  2002; 13(3):729-733.

405. Zoccali C, Benedetto FA, Maas R, et al: Asymmetric dimethylarginine, C-reactive protein, and carotid intima-media thickness in end-stage renal disease.  J Am Soc Nephrol  2002; 13(2):490-496.

406. Wanner C, Krane V, Marz W, et al: Atorvastatin in patients with type 2 diabetes mellitus undergoing hemodialysis.  N Engl J Med  2005; 353(3):238-248.

407. Cice G, Ferrara L, Di Benedetto A, et al: Dilated cardiomyopathy in dialysis patients—Beneficial effects of carvedilol: A double-blind, placebo-controlled trial.  J Am Coll Cardiol  2001; 37(2):407-411.

408. Tornig J, Amann K, Ritz E, et al: Arteriolar wall thickening, capillary rarefaction and interstitial fibrosis in the heart of rats with renal failure: The effects of ramipril, nifedipine, and moxonidine.  J Am Soc Nephrol  1996; 7:667-675.

409. van Guldener C, Lambert J, Janssen MJ, et al: Endothelium-dependent vasodilatation and distensibility of large arteries in chronic haemodialysis patients.  Nephrol Dial Transplant  1997; 12:14-18.

410. Narita M, Kurihara T, Sindoh T, et al: Characteristics of myocardial ischemia in patients with chronic renal failure and its relation to cardiac sympathetic activity.  Kaku Igaku  1999; 36:979-987.

411. Amann K, Mall G, Ritz E: Myocardial interstitial fibrosis in uraemia: Is it relevant?.  Neprhol Dial Transplant  1994; 9:127-128.

412. Raine AE, Seymour AM, Roberts AF, et al: Impairment of cardiac function and energetics in experimental renal failure.  J Clin Invest  1993; 92:2934-2940.

413. Pidgeon GB, Lynn KL, Bailey RR, Robson RA: Coronary angiography prior to renal transplantation.  Nephrology  1995; 1:59-64.

414. Braun WE, Philips DF, Vidt DG: Coronary artery disease in 100 diabetics with end-stage renal failure.  Transplant Proc  1984; 16:603-607.

415. Kremastinos D, Paraskevaidis I, Voudiklari S, et al: Painless myocardial ischemia in chronic hemodialysed patients: A real event?.  Nephron  1992; 60:164-170.

416. Robinson TG, Carr SJ: Cardiovascular autonomic dysfunction in uremia.  Kidney Int  2002; 62(6):1921-1932.

417. Khan NA, Hemmelgarn BR, Tonelli M, et al: Prognostic value of troponin T and I among asymptomatic patients with end-stage renal disease: A meta-analysis.  Circulation  2005; 112(20):3088-3096.

418. Murphy SW, Foley RN, Parfrey PS: Screening and treatment for cardiovascular disease in patients with chronic renal disease.  Am J Kidney Dis  1998; 32(suppl 3):S184-S199.

419. Greene WH, Ray C, Mauer SM, Quie PG: The effect of hemodialysis on neutrophil chemotactic responsiveness.  J Lab Clin Med  1976; 88:971-974.

420. Reis G, Marcovitz PA, Leichtman AB, et al: Usefulness of dobutamine stress-echocardiography in detecting coronary artery disease in end-stage renal disease.  Am J Cardiol  1995; 75:707-710.

421. Nishimura M, Hashimoto T, Kobayashi H, et al: Myocardial scintigraphy using a fatty acid analogue detects coronary artery disease in hemodialysis patients.  Kidney Int  2004; 66(2):811-819.

422. Herzog CA: Poor long-term survival of dialysis patients after acute myocardial infarction: Bad treatment or bad disease?.  Am J Kidney Dis  2000; 35(6):1217-1220.

423. Herzog CA: Acute myocardial infarction in patients with end-stage renal disease.  Kidney Int Suppl  1999; 56:S130-S133.

424. Chertow GM, Normand SL, Silva LR, McNeil BJ: Survival after acute myocardial infarction in patients with end-stage renal disease: Results from the Cooperative Cardiovascular Project.  Am J Kidney Dis  2000; 35(6):1044-1051.

425. Herzog MD, Ma JZ, Collins AJ: Poor long-term survival after acute myocardial infarction among patients on long-term dialysis.  N Engl J Med  1998; 339:799-805.

426. Hemmelgarn BR, Southern D, Culleton BF, et al: Survival after coronary revascularization among patients with kidney disease.  Circulation  2004; 110(14):1890-1895.

427. Herzog CA, Ma JZ, Collins AJ: Comparative survival of dialysis patients in the United States after coronary angioplasty, coronary artery stenting, and coronary artery bypass surgery and impact of diabetes.  Circulation  2002; 106(17):2207-2211.

428. Le Feuvre C: Angioplasty and stenting in patients with renal disease.  Heart  2000; 83(1):7-8.

429. Kutner NG, Zhang R, Huang Y, Herzog CA: Cardiac rehabilitation and survival of dialysis patients after coronary bypass.  J Am Soc Nephrol  2006; 17(4):1175-1180.

430. Silberberg JS, Barre PE, Prichard SS, Sniderman AD: Impact of left ventricular hypertrophy on survival in end-stage renal disease.  Kidney Int  1989; 36:286-290.

431. London GM, Pannier B, Guerin AP, et al: Alterations of left ventricular hypertrophy in and survival of patients receiving hemodialysis: Follow-up of an interventional study.  J Am Soc Nephrol  2001; 12(12):2759-2767.

432. Aoki J, Ikari Y, Nakajima H, et al: Clinical and pathologic characteristics of dilated cardiomyopathy in hemodialysis patients.  Kidney Int  2005; 67(1):333-340.

433. Foley RN, Parfrey PS, Harnett JD, et al: Hypoalbuminemia, cardiac morbidity and mortality in end-stage renal disease.  J Am Soc Nephrol  1996; 7:728-736.

434. Wanner C: C-reactive protein risk prediction: Adding cardiac hypertrophy to the list.  Am J Kidney Dis  2002; 40(6):1340-1341.

435. Goodman WG, Goldin J, Kuizon BD, et al: Coronary-artery calcification in young adults with end-stage renal disease who are undergoing dialysis.  N Engl J Med  2000; 342(20):1478-1483.

436. Braun J, Oldendorf M, Moshage W, et al: Electron beam computed tomography in the evaluation of cardiac calcifications in chronic dialysis patients.  Am J Kidney Dis  1996; 27:394-401.

437. Ahmed S, O'Neill KD, Hood AF, et al: Calciphylaxis is associated with hyperphosphatemia and increased osteopontin expression by vascular smooth muscle cells.  Am J Kidney Dis  2001; 37(6):1267-1276.

438. Ganesh SK, Stack AG, Levin NW, et al: Association of elevated serum PO(4), Ca × PO(4) product, and parathyroid hormone with cardiac mortality risk in chronic hemodialysis patients.  J Am Soc Nephrol  2001; 12(10):2131-2138.

439. Davies MR, Hruska KA: Pathophysiological mechanisms of vascular calcification in end-stage renal disease.  Kidney Int  2001; 60(2):472-479.

440. Block GA, Hulbert-Shearon TE, Levin NW, Port FK: Association of serum phosphorus and calcium × phosphate product with mortality risk in hemodialysis pateints.  Am J Kidney Dis  1998; 31:601-617.

441. Wang AY, Woo J, Wang M, et al: Association of inflammation and malnutrition with cardiac valve calcification in continuous ambulatory peritoneal dialysis patients.  J Am Soc Nephrol  2001; 12(9):1927-1936.

442. Lomashvili KA, Khawandi W, O'Neill WC: Reduced plasma pyrophosphate levels in hemodialysis patients.  J Am Soc Nephrol  2005; 16(8):2495-2500.

443. Lomashvili KA, Cobbs S, Hennigar RA, et al: Phosphate-induced vascular calcification: Role of pyrophosphate and osteopontin.  J Am Soc Nephrol  2004; 15(6):1392-1401.

444. Chertow GM, Burke SK, Lazarus JM, et al: Poly(allylamine hydrochloride) (Rena Gel): A noncalcemic phosphate binder for the treatment of hyperphosphatemia in chronic renal failure.  Am J Kidney Dis  1997; 29:66-71.

445. Chertow GM, Burke SK, Raggi P: Sevelamer attenuates the progression of coronary and aortic calcification in hemodialysis patients.  Kidney Int  2002; 62(1):245-252.

446. Patetsios P, Bernstein M, Kim S, et al: Severe necrotizing mastopathy caused by calciphylaxis alleviated by total parathyroidectomy.  Am Surg  2000; 66(11):1056-1058.

447. Ilkani R, Gardezi S, Hedayati H, Schein M: Necrotizing mastopathy caused by calciphylaxis: A case report.  Surgery  1997; 122:967-968.

448. Loynes JT: A rare cause of breast necrosis: Calciphylaxis.  Postgrad Med  1999; 106:161-162.

449. Morris DG, Fisher AH, Abboud J: Breast infarction after internal mammary artery harvest in a patient with calciphylaxis.  Ann Thorac Surg  1997; 64:1469-1471.

450. Ruggian JC, Maesaka JK, Fishbane S: Proximal calciphylaxis in four insulin-requiring diabetic hemodialysis patients.  Am J Kidney Dis  1996; 28:409-414.

451. Wilmer WA, Magro CM: Calciphylaxis: Emerging concepts in prevention, diagnosis, and treatment.  Semin Dial  2002; 15(3):172-186.

452. Mazhar AR, Johnson RJ, Gillen D, et al: Risk factors and mortality associated with calciphylaxis in end-stage renal disease.  Kidney Int  2001; 60(1):324-332.

453. Khafif RA, Delima C, Silverberg A, et al: Acute hyperparathyroidism with systemic calcinosis: Report of a case.  Arch Intern Med  1989; 149:681-684.

454. Wilmer WA, Voroshilova O, Singh I, et al: Transcutaneous oxygen tension in patients with calciphylaxis.  Am J Kidney Dis  2001; 37(4):797-806.

455. Seyle H, Golie I, Strebel R: Calciphylaxis in relation to calcification in periarticular tissues.  Clin Orthop  1963; 28:181-192.

456. Seyle H, Somogyi A, Mecs I: Inhibition of local calcergy by topical application of calciphylactic challengers.  Calcif Tissue Res  1968; 2:67-76.

457. Guerra G, Shah RC, Ross EA: Rapid resolution of calciphylaxis with intravenous sodium thiosulfate and continuous venovenous haemofiltration using low calcium replacement fluid: Case report.  Nephrol Dial Transplant  2005; 20(6):1260-1262.

458. Cicone JS, Petronis JB, Embert CD, Spector DA: Successful treatment of calciphylaxis with intravenous sodium thiosulfate.  Am J Kidney Dis  2004; 43(6):1104-1108.

459. Hakim RM, Levin NW: Malnutrition in hemodialysis patients.  Am J Kidney Dis  1993; 21:125-137.

460. Kaysen GA, Dubin JA, Muller HG, et al: The acute-phase response varies with time and predicts serum albumin levels in hemodialysis patients. The HEMO Study Group.  Kidney Int  2000; 58(1):346-352.

461. Sreedhara R, Avram MM, Blanco M, et al: Prealbumin in the best nutritional predictor of survival in hemodialysis and peritoneal dialysis.  Am J Kidney Dis  1996; 28:937-942.

462. Sanaka T, Shinobe M, Ando M, et al: IGF-I as an early indicator of malnutrition in patients with end-stage renal diseae.  Nephron  1994; 67:73-81.

463. Parker TF, Wingard RL, Husni L, et al: Effect of the membrane biocompatibility on nutritional parameters in chronic hemodialysis patients.  Kidney Int  1996; 49:551-556.

464. Pollock CA, Ibels LS, Allen BJ: Nutritional markers and survival in maintenance dialysis patients.  Nephron  1996; 74:625-641.

465. Stenver DI, Gotfredsen A, Hilsted J, Neilsen B: Body composition in hemodialysis patients measured by dual-energy x-ray absorptiometry.  Am J Nephrol  1995; 15:105-119.

466. Chertow GM, Lowrie EG, Wilmore DW, et al: Nutritional assessment with bioelectrical impedance analysis in maintenance hemodialysis patients.  J Am Soc Nephrol  1995; 6:75-81.

467. Borah MF, Schoenfeld PY, Gotch FA, et al: Nitrogen balance during intermittent dialysis therapy of uremia.  Kidney Int  1978; 14:491-500.

468. Schneeweiss B, Graninger W, Stockenhuber F, et al: Energy metabolism in acute and chronic renal failure.  Am J Clin Nutr  1990; 52:596-601.

469. Ikizler TA, Wingard RL, Sun M, et al: Increased energy expenditure in hemodialysis patients.  J Am Soc Nephrol  1996; 7:2646-2654.

470. Chanard J, Toupance O, Gillery P, Lavaud S: Evaluation of protein loss during hemofiltration.  Kidney Int Suppl  1988; 33:S114-S116.

471. Graeber CW, Halley SE, Lapkin RA, et al: Protein losses with reused dialyzer. Abstract.  J Am Soc Nephrol  1993; 4:349.

472. Caglar K, Peng Y, Pupim LB, et al: Inflammatory signals associated with hemodialysis.  Kidney Int  2002; 62(4):1408-1416.

473. Guitierrez A, Alvestrand A, Wahren J, Bergstrom J: Effect of in vivo contact between blood and dialysis membranes on protein catabolism in humans.  Kidney Int  1990; 38:487-494.

474. Guitierrez A: Protein catabolism in maintenance hemodialysis: The influence of the dialysis membrane.  Nephrol Dial Transplant  1996; 11(suppl 2):108-112.

475. Leon JB, Majerle AD, Soinski JA, et al: Can a nutrition intervention improve albumin levels among hemodialysis patients? A pilot study.  J Ren Nutr  2001; 11(1):9-15.

476. Eustace JA, Coresh J, Kutchey C, et al: Randomized double-blind trial of oral essential amino acids for dialysis-associated hypoalbuminemia.  Kidney Int  2000; 57(6):2527-2538.

477. Caglar K, Fedje L, Dimmitt R, et al: Therapeutic effects of oral nutritional supplementation during hemodialysis.  Kidney Int  2002; 62(3):1054-1059.

478. Heidland A, Kult J: Long-term effects of essential amino acids supplementation in patients on regular dialysis treatment.  Clin Nephrol  1975; 2:234-239.

479. Ahmad S: L-carnitine in dialysis patients.  Semin Dial  2001; 14(3):209-217.

480. Golper TA, Wolfson M, Ahmad S, et al: Multiecenter trial of L-carnitine in maintenance hemodialysis patients. I. Carnitine concentrations and lipid effects.  Kidney Int  1990; 38:904-911.

481. Rodger RSC, Sheldon WL, Watson MJ, et al: Zinc deficiency and hyperprolactinaemia are not reversible causes of sexual dysfunction in uraemia.  Nephrol Dial Transplant  1989; 4:888-892.

482. Stein G, Sperschneider H, Koppe S: Vitamin levels in chronic renal failure and need for supplementation.  Blood Purif  1985; 3:52-62.

483. Descombes E, Hanck AB, Fellay G: Water-soluble vitamins in chronic hemodialysis patients and need for supplementation.  Kidney Int  1993; 43:1319-1328.

484. Kopple JD, Swenseid ME: Vitamin nutrition in patients undergoing maintenance hemodialysis.  Kidney Int Suppl  1975; 7:79-84.

485. Werb R, Clark WF, Lindsay RM, et al: Serum vitamin A levels and associated abnormalities in patients on regular dialysis treatments.  Clin Nephrol  1979; 12:63-68.

486. Bloembergen BE, Port FK: Epidemiological perspective on infections in chronic dialysis patients.  Adv Renal Replac Ther  1996; 3:201-207.

487. Ishani A, Collins AJ, Herzog CA, Foley RN: Septicemia, access and cardiovascular disease in dialysis patients: The USRDS Wave 2 study.  Kidney Int  2005; 68(1):311-318.

488. Hoen B, Paul-Dauphin A, Hestin D, Kessler M: EPIBACDIAL: A multicenter prospective study of risk factors for bacteremia in chronic hemodialysis patients.  J Am Soc Nephrol  1998; 9:869-976.

489. Dasgupta MK: Biofilms and infection in dialysis patients.  Semin Dial  2002; 15(5):338-346.

490. Keane WF, Shapiro FL, Raij L: Incidence and type of infections occurring in 445 chronic hemodialysis patients.  Trans Am Soc Artif Intern Organs  1977; 23:41-47.

491. Lewis SL, Van Epps DE: Neutrophil and monocyte alterations in chronic dialysis patients.  Am J Kidney Dis  1987; 9:381-395.

492. Goldblum SE, Reed WP: Host defenses and immunologic alterations associated with chronic hemodialysis.  Ann Intern Med  1980; 93:597-613.

493. Tokars JI, Frank M, Alter MJ, Arduino MJ: National surveillance of dialysis-associated diseases in the United States, 2000.  Semin Dial  2002; 15(3):162-171.

494.   Edmond MB, Wallace SE, Pfeffer MA, et al: Surveillance of 49 US medical centers for VRE bacteremia. Paper presented at the 34th Annual Meeting of the Infectious Disease Society of America, September 1996.

495. Knudsen F, Wantzin P, Rasmussen K, et al: Hepatitis C in dialysis patients: Relationship to blood transfusions, dialysis, and liver disease.  Kidney Int  1993; 43:1353-1356.

496. Kuhns M, de Medina M, McNamara A, et al: Detection of hepatitis C virus RNA in hemodialysis.  J Am Soc Nephrol  1994; 4:1491-1497.

497. Caramelo C, Ortiz A, Aguilera B, et al: Liver disease patterns in hemodialysis patients with antibodies to hepatitis C virus.  Am J Kidney Dis  1993; 22:822-828.

498. Fabrizi F, Martin P, Dixit V, et al: Biological dynamics of viral load in hemodialysis patients with hepatitis C virus.  Am J Kidney Dis  2000; 35(1):122-129.

499. Koenig P, Vogel W, Umlauft F, et al: Interferon treatment for chronic hepatitis C virus infection in uremic patients.  Kidney Int  1994; 45:1507-1509.

500. Rangel MC, Coronado VG, Euler GL, Strikas RA: Vaccine recommendations for patients on chronic dialysis. The Advisory Committee on Immunization Practices and the American Academy of Pediatrics.  Semin Dial  2000; 13(2):101-107.

501. Tong NK, Beran J, Kee SA, et al: Immunogenicity and safety of an adjuvanted hepatitis B vaccine in pre-hemodialysis and hemodialysis patients.  Kidney Int  2005; 68(5):2298-2303.

502. Bregman H, Daugridas J, Ing T: Complictions during hemodialysis.   In: Daugridas J, Ing T, ed. Handbook of Dialysis,  2nd ed. Boston: Little, Brown; 1994:13-29.

503. De Simone G: Left ventricular geometry and hypotension in end-stage renal disease: A mechanical perspective.  J Am Soc Nephrol  2003; 14(10):2421-2427.

504. Schreiber Jr MJ: Clinical case-based approach to understanding intradialytic hypotension.  Am J Kidney Dis  2001; 38(4 suppl 4):S37-S47.

505. Schreiber BD: Congestive heart failure in patients with chronic kidney disease and on dialysis.  Am J Med Sci  2003; 325(4):179-193.

506. Wehle B: Factors affecting blood pressure in hemodialysis.  Scand J Urol Nephrol Suppl  1982; 69:1-66.

507. Converse Jr RL, Jacobsen TN, Toto RD, et al: Sympathetic overactivity in patients with chronic renal failure.  N Engl J Med  1992; 327:1912-1918.

508. Oe B, de Fijter CW, Oe PL, et al: Diameter of inferior caval vein (VCD) and bioelectrical impedance analysis (BIA) for the analysis of hydration status in patients on hemodialysis.  Clin Nephrol  1998; 50:38-43.

509. Santoro A, Mancini E, Canova C, Mambelli E: Thermal balance in convective therapies.  Nephrol Dial Transplant  2003; 18(Suppl 7):vii41-vvii5.discussion vii57

510. Schneditz D, Rosales L, Kaufman AM, et al: Heat accumulation with relative blood volume decrease.  Am J Kidney Dis  2002; 40(4):777-782.

511. Donauer J, Schweiger C, Rumberger B, et al: Reduction of hypotensive side effects during online-haemodiafiltration and low temperature haemodialysis.  Nephrol Dial Transplant  2003; 18(8):1616-1622.

512. Kumar VA, Craig M, Depner TA, Yeun JY: Extended daily dialysis: A new approach to renal replacement for acute renal failure in the intensive care unit.  Am J Kidney Dis  2000; 36(2):294-300.

513. Okada K, Abe M, Hagi C, et al: Prolonged protective effect of short daily hemodialysis against dialysis-induced hypotension.  Kidney Blood Press Res  2005; 28(2):68-76.

514. Cruz DN, Mahnensmith RL, Brickel HM, Perazella MA: Midodrine and cool dialysate are effective therapies for symptomatic intradialytic hypotension.  Am J Kidney Dis  1999; 33:920-926.

515. Lin YF, Wang JY, Denq JC, Lin SH: Midodrine improves chronic hypotension in hemodialysis patients.  Am J Med Sci  2003; 325(5):256-261.

516. McIntyre CW, Lambie SH, Fluck RJ: Biofeedback controlled hemodialysis (BF-HD) reduces symptoms and increases both hemodynamic tolerability and dialysis adequacy in non-hypotension prone stable patients.  Clin Nephrol  2003; 60(2):105-112.

517. Santoro A, Mancini E, Basile C, et al: Blood volume controlled hemodialysis in hypotension-prone patients: A randomized, multicenter controlled trial.  Kidney Int  2002; 62(3):1034-1045.

518. Schroeder KL, Sallustio JE, Ross EA: Continuous haematocrit monitoring during intradialytic hypotension: Precipitous decline in plasma refill rates.  Nephrol Dial Transplant  2004; 19(3):652-656.

519. Roca AO, Jarjoura D, Blend D, et al: Dialysis leg cramps: Efficacy of quinine versus vitamin E.  ASAIO J  1992; 38:481-485.

520. Jansen PH, Veenhuizen KC, Wesseling AI, et al: Randomised controlled trial of hydroquinine in muscle cramps.  Lancet  1997; 349:528-532.

521. Ahmad S, Robertson HT, Golper TA, et al: Multicenter trial of L-carnitine in maintenance hemodialysis patients. II. Clinical and biochemical effects.  Kideny Int  1990; 38:912-918.

522. Trinh-Trang-Tan MM, Cartron JP, Bankir L: Molecular basis for the dialysis disequilibrium syndrome: Altered aquaporin and urea transporter expression in the brain.  Nephrol Dial Transplant  2005; 20(9):1984-1988.

523. Zebe H: Atrial fibrillation in dialysis patients.  Nephrol Dial Transplant  2000; 15(6):765-768.

524. Gehr T, Sica D: Antiarrhythmic medications: Practical guidelines for drug therapy in dialysis.  Semin Dial  1990; 3:33-40.

525. Epstein A, Kay G, Plumb V: Considerations in the diagnosis and treatment of arrhythmias in patients with end-stage renal disease.  Semin Dial  1990; 2:31-37.

526. Hakim RM, Breillatt J, Lazarus JM, Port FK: Compelment activation and hypersensitivity reactions to dialysis membranes.  N Engl J Med  1984; 311:878-882.

526a. Lemke H, Fink E: Accumulation of bradykinin formed by the AN69 or PAN17 DX-membrane is due to the presence of an ACE-inhibitor in vitro (abstract).  J Am Soc Nephrol  1992; 3:376.

526b. Schulman G, Hakim R, Arias R, et al: Bradykinin generation by dialysis mem-branes: possible role in anaphylactic reaction.  J Am Soc Nephrol  1993; 3:1563-1569.

526c. Parnes EL, Shapiro WB: Anaphylactoid reactions in hemodialysis patients treated with the AN69 dialyzer.  Kidney Int  1991; 40:1148-1152.

526d. Pegues DA, Beck-Sague CM, Woollen SW, et al: Anaphylactoid reactions associated with reuse of hollow-fiber hemodialyzers and ACE inhibitors.  Kidney Int  1992; 42:1232-1237.

526e. Eaton JW, Leida MN: Hemolysis in chronic renal failure.  Semin Nephrol  1985; 5:133-139.

527. Centers for Disease Control and Prevention (CDC) : Multistate outbreak of hemolysis in hemodialysis patients—Nebraska and Maryland, 1998.  JAMA  1998; 280(15):1299-1300.

528.   United States Food and Drug Administration: FDA notifies health care professionals of problem with Baxter Healthcare's Meridian hemodialysis instrument: Kinks in device tubing pose serious risk to patients. Accessed April 2006.

529. De Wachter DS, Verdonck PR, De Vos JY, Hombrouckx RO: Blood trauma in plastic haemodialysis cannulae.  Int J Artif Organs  1997; 20(7):366-370.

530. Seukeran D, Fletcher S, Sellars L, Vestey JP: Sudden deepening of pigmentation during haemodialysis due to severe haemolysis.  Br J Dermatol  1997; 137(6):997-999.

531. Himmelfarb J: Dialysis at a crossroads: Reverse engineering renal replacement therapy.  Clin J Soc Nephrol  2006; 1:896-902.

532. Akl AI, Sobh MA, Enab YM, Tattersall J: Artificial intelligence: A new approach for prescription and monitoring of hemodialysis therapy.  Am J Kidney Dis  2001; 38(6):1277-1283.

533. Chertow GM: “Wishing don't make it so”—Why we need a randomized clinical trial of high-intensity hemodialysis.  J Am Soc Nephrol  2001; 12(12):2850-2853.

534. Fissell WH, Lou L, Abrishami S, et al: Bioartificial kidney ameliorates gram-negative bacteria-induced septic shock in uremic animals.  J Am Soc Nephrol  2003; 14(2):454-461.

If you find an error or have any questions, please email us at Thank you!