Iain M. Smith
Roger M. Lee
Spencer R. Adams
Samuel A. Skootsky
Because most patients with gynecological cancer are middle-aged to elderly, they have a high incidence of medical problems at the time of presentation. These need to be carefully assessed, and if necessary treated before the patient undergoes aggressive surgery, so that her medical status can be optimized. The early identification, evaluation, and management of emerging medical problems are also essential, especially in the perioperative period. This chapter discusses the most frequently encountered problems in patients with gynecologic cancers.
The cornerstone of all perioperative medical management is the anticipation of specific problems. It is always better to have a management plan than to react to complications. Careful assessment of risk and monitoring of patients in the perioperative period minimizes morbidity and mortality.
Surgery can represent a major cardiovascular stress because of depression in myocardial contractility, changes in sympathetic tone induced by general anesthetic agents, and rapid changes in intravascular volume that occur due to blood loss and “third spacing” of fluids. The magnitude of cardiovascular stress depends on patient characteristics, the nature and site of the operation, the duration of the operation, and whether it is elective or emergent.
Cardiovascular Risk Factors
Multiple studies have been published over the last 30 years assessing clinical risks for cardiac events during surgery (1,2,3,4,5). A commonly used “simple index” called the Revised Cardiac Risk Index (Table 18.1) has been incorporated in the 2007 revision of the American College of Cardiology (ACC) and the American Heart Association (AHA) guidelines on perioperative cardiovascular evaluation and care for noncardiac surgery (5,6).
The recently revised ACC/AHA 2007 guidelines offer a simplified yet comprehensive approach to the assessment of cardiac risks for patients undergoing surgery (6). An algorithm based on this approach is shown in Fig. 18.1. This approach represents a consensus view derived from a review of the literature to date and is periodically updated on the ACC Web site (http://www.acc.org). An important overriding theme of these guidelines is that cardiac stress testing and/or interventions (such as coronary stenting or coronary bypass graft surgery) are rarely necessary simply to lower the risk of surgery. In fact, such interventions are likely unnecessary unless they would have been performed even if the patient were not undergoing surgery. Additionally, no test should be performed unless it likely to influence patient treatment.
Table 18.1 Revised Cardiac Risk Index
If the patient has an active cardiac condition that indicates major clinical risk (Table 18.2), the surgery likely will be delayed or cancelled until the condition is stabilized unless the surgery is emergent (6,7). For stable patients, the algorithm can be used as follows: Obtain information from the patient and then use the Revised Cardiac Risk Index to determine how many of the clinical risk factors (from Table 18.1) that the patient has. One point is assigned for each of the six possible clinical risk factors. Based on how many points are assigned, patients can be characterized as high risk (3 or more points), intermediate risk (1-2 points), or low risk (0 points).
The first step is to determine the urgency of surgery. If emergent surgery is needed there is no time for any cardiac assessment beyond what history is available and the patient should proceed immediately to the operating room. For urgent or elective surgery, there is more time to assess a patient's cardiac risks. If the patient has an active cardiac condition(Table 18.2), all but emergency surgeries should be delayed until the cardiac condition is properly evaluated and treated.
Often patients need very little evaluation. In general, patients who are undergoing low-risk surgery (Table 18.3) do not need further evaluation. In addition, patients with good functional capacity (metabolic equivalents [METs] levels greater than or equal to 4) (Table 18.4) without symptoms and active conditions can proceed with typical gynecologic surgery without further evaluation (8) (Fig. 18.1).
Table 18.2 Major Active Cardiac Conditions
Table 18.3 Cardiac Risk Stratification for Noncardiac Surgical Procedures
If the patient has poor or unknown functional capacity, then the presence of clinical risk factors from the Revised Cardiac Risk Index help determine the need for further evaluation. For those patients undergoing intermediate risk surgeries (most gynecological surgeries), the presence of any clinical risk factor should prompt consideration of cardiology consultation and/or noninvasive stress testing if it will change management (9). Perioperative beta blockade will be discussed in more detail below, but given the results of recent studies, caution should be used when newly prescribing these drugs to patients with intermediate risk. Dipyridamolethallium imaging or dobutamine stress echocardiography can be considered for noninvasive testing in these patients (10,11). If noninvasive testing shows only minor abnormalities, the patient can likely proceed with surgery with appropriate medical management including beta blockade. If moderate or severe abnormalities are found, subsequent care should include cardiology consultation. The cardiologist may recommend cancellation or delay of surgery, coronary revascularization followed by noncardiac surgery, or intensified care in such patients (6).
Even with the best preoperative assessment and preparation, postoperative myocardial infarction (MI) can still occur after surgery under general anesthesia. The risk factors for perioperative MI are related to the underlying risk of ischemic heart disease. Before the advent of more modern management of ischemic heart disease, the risk of a second MI after anesthesia and general surgery was considered too high during the first few months post myocardial infarction (12). Current cardiologic practice, including revascularization, angioplasty, or very aggressive medical therapy with lipid-lowering agents and use of β-blockers, makes this rule less useful. It is now commonly believed that with proper treatment, patients can undergo surgery 6 weeks after myocardial infarction if necessary.
Traditionally, it was felt that coronary artery bypass surgery lowered the risk in patients with coronary artery disease (CAD) (13,14). Recent evidence, however, suggests that only certain patients with severe coronary artery disease benefit from coronary revascularization (significant left main stenosis, 3-vessel CAD with a decreased left ventricular ejection fraction [<50%], 2-vessel CAD with proximal left anterior descending artery stenosis and either EF <50% or ischemia on stress-testing, or active acute coronary syndromes such as acute MI) (6). If none of the factors listed above are present, aggressive medical therapy, including β-blockade, is likely as effective as revascularization at reducing surgical risk even in high-risk ischemic patients (e.g. patients with significantly abnormal preoperative dobutamine stress echocardiograms) (15).
Figure 18.1 Stepwise approach to preoperative cardiac assessment in gynecologic surgery. *Active cardiac conditions refers to Table 18.2; **Risk of surgery is shown in Table 18.3 ***MET, metabolic equivalent (see Table 18.4) ****Clinical Risk Factors refers to Table 18.1. From the following source: Fliesher LA, Beckman JA, Brown KA, Calkins H, Chaikof E, et al. ACC/AHA 2007 guidelines on perioperative cardiovascular evaluation and care for noncardiac surgery: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Revise the 2002 Guidelines on Perioperative Cardiovascular Evaluation for Noncardiac Surgery). Circulation 2007;116:e418-e499.
Table 18.4 Estimated Energy Requirements for Various Activities
Also, with the increasing use of percutaneous coronary intervention (PCI), elective or nonurgent surgery should be delayed for at least 14 days after balloon angioplasty. When a stent is placed, dual oral antiplatelet therapy with aspirin and clopridogrel is needed to reduce the risk of stent restenosis and stent thrombosis. Therefore, clopidigrel and aspirintherapy should be instituted following placement of a coronary stent and elective or nonurgent surgery should be delayed at least 30 days after the placement of a bare-metal coronary stent, and 365 days after the placement of a drug-eluting stent. When surgery is undertaken, perioperative aspirin should be continued in each of these situations (6).
Postoperative MI can be painless. The risk of MI is throughout the first week, but the incidence is thought to peak on the third postoperative day. Additionally, diagnosis of a perioperative MI has both short- and long-term prognostic significance. Perioperative MI has been associated with significant perioperative mortality (30% to 50%) and reduced long-term survival (16,17). Hence, surveillance for perioperative MI is likely prudent in high-risk patients.
Electrocardiography (ECG), beginning in the immediate postoperative period and continuing at least through postoperative day 3, is the most well-established method of surveillance (18,19). Several studies suggest that serum troponin assays may be helpful for surveillance of perioperative MI as well, but their use is not yet well established (6,20,21). Because the risk of perioperative MI is increased in patients who are subjected to intraoperative hypotension, measures must be taken to maintain high-risk patients in a normotensive state during surgery. If intraoperative hypotension occurs, the patient should be considered at high-risk of postoperative MI and monitored appropriately.
Theoretically, β-blockers would be expected to facilitate the development of intraoperative hypotension because of the additive myocardial depressive effect of these medications with general anesthesia. However, abrupt discontinuation of β-blocker medication can be associated with a dangerous rebound syndrome (i.e., acute hypertension and coronary ischemia), with the incidence of the syndrome peaking at 4 to 7 days after discontinuation of the drug (22,23). Patients tolerate general anesthesia in the presence of continued β-blocker treatment. While several studies in the 1990s suggested that perioperative β-blocker use reduced postoperative nonfatal MI and mortality in many patients including those with intermediate risk, subsequent research has cast doubt on this claim (24,25,26). In fact, the largest placebo-controlled trial of perioperative β-blocker use to date was recently published and showed an increase in mortality and stroke in those receiving β-blockers compared with placebo (26).
Given the current controversy surrounding β-blockers, their use can only be strongly recommended in high-risk patients (at least 3 points on the Revised Cardiac Risk Index) undergoing vascular surgery (6). In addition, as stated previously, β-blockers should be continued in patients already taking them before surgery, because ischemia can be precipitated if a β-blocker is abruptly discontinued (22,23).
If β-blockers are prescribed, the recommended target heart rate for effective β-blockade is below 65, but not lower than 50 beats per minute. The appropriate duration of β-blockade is also currently being debated, but should certainly begin before surgery and continue throughout the hospitalization. If possible, there may be additional benefit if β-blockers can be started one month before surgery to titrate the heart rate, and be continued after hospitalization for at least 30 days if adequate postoperative medical follow up can be arranged (25,27).
Congestive Heart Failure
Heart failure is well known to be associated with a poorer outcome after noncardiac surgery (1,2,5). The cause of heart failure should be identified if possible as this may have implications concerning perioperative risk (6). Also, it may be reasonable to perform a noninvasive evaluation of left ventricular function (e.g., echocardiogram) in symptomatic patients, but the utility of this is still questionable (28). Patients with moderate or severe congestive heart failure should be treated before surgery with appropriate medications to optimize their cardiovascular status.
Perioperative use of a pulmonary artery (Swan-Ganz) catheter was previously advocated to allow cardiac function to be optimized and to aid in the intraoperative management of fluids and cardiac medications. Recent studies, however, have not shown clear benefit to these devices in managing high-risk surgical patients, and this procedure is generally no longer recommended (29).
Cardiac arrhythmias that are hemodynamically significant or symptomatic should be treated and stabilized prior to elective or nonurgent surgery. Atrial fibrillation is the most frequently seen arrhythmia and may require electrical or pharmacological cardioversion. Alternatively, a rate-control strategy can be attempted with β-blockers, calcium channel blockers, or digoxin. Ventricular arrhythmias, such as simple premature ventricular contractions, complex ventricular ectopy, or nonsustained tachycardia usually require no therapy unless they are associated with hemodynamic compromise or occur in the presence of left ventricle dysfunction or ongoing cardiac ischemia (30,31). However, careful evaluation for underlying cardiopulmonary disease, drug toxicity, metabolic disturbances, and infection should be undertaken in patients who have any arrhythmia in the perioperative period.
High-Grade Conduction Abnormalities
Patients who do not have permanent pacemakers and who have third-degree heart block at the time of presentation are at substantial risk of cardiopulmonary arrest during surgery.Typically, they are unable to mount an appropriate pulse response to the vasodilatation and decreased myocardial contractility induced by general anesthesia or to the volume depletion induced by surgical blood loss. Patients with high-grade conduction abnormalities including complete heart block may require temporary or permanent transvenous pacing.
In patients with lower degrees of heart block, specifically bifascicular block (right heart block with left axis deviation), the risk of development of a higher degree of ventricular block during surgery is not significantly increased, provided there is no history of previous third-degree heart block or syncope. Such patients rarely require insertion of a temporary pacemaker (32). Patients with bifascicular block who have a history of third-degree heart block should be managed for complete heart block with preoperative cardiology evaluation and likely pacemaker insertion.
A new bifascicular block developing in the setting of acute MI carries a high risk of progression to complete heart block. Therefore, if this problem occurs after surgery, the patient should be considered at significant risk for the development of complete atrioventricular block. Such patients require a cardiology consultation and insertion of a temporary pacemaker.
Patients with permanently implanted pacemakers should have a preoperative pacemaker evaluation to allow examination of all pacemaker functions. This precaution ensures that backup demand pacemaker failure will not be uncovered unexpectedly with the vagotonic stimuli associated with general anesthesia in abdominal surgery. Patients with implanted defibrillators typically have their devices turned off shortly before surgery and then turned back on shortly afterward.
Even newer pacemakers and defibrillators can sense the electromagnetic impulses created by electrocautery, especially when the electrocautery plate is close to the pacemaker unit. It is prudent to place the indifferent electrocautery electrode as far as possible from the chest and to use electrocautery sparingly. An added precaution consists of keeping a magnet available in the operating room to convert a pacemaker rapidly from the demand to a fixed pacing mode. Inappropriate discharges from the implanted defibrillator are avoided by having the device turned off during the time of surgery (33). Those with permanent pacemakers should also have their device assessed for proper function after surgery (6).
Recently revised guidelines for infective endocarditis (IE) do not recommend administration of antibiotics solely to prevent endocarditis for patients who undergo a genitourinary (GU) or gastrointestinal (GI) tract procedure. As a result of a lack of published data demonstrating a benefit from prophylaxis, current guidelines differ substantially from previous guidelines and far fewer patients should be recommended for IE prophylaxis than previously thought (34). Very few data exist on the risk or prevention of IE with a GI or GU tract procedure. Enterococci are part of the normal flora of the GI tract and are the primary bacteria from this area likely to cause IE. In patients with the highest risk cardiac conditions (prosthetic cardiac valve, previous IE, or congenital heart disease) who are to receive antibiotic therapy to prevent wound infection, it may be reasonable to include an antibiotic that is active against enterococci, such as penicillin, ampicillin, or vancomycin. However, no published studies demonstrate that such therapy will prevent enterococcal IE (34).
The significance of mild to moderate hypertension (stage 1 or 2 with systolic blood pressure below 180 mm Hg and diastolic blood pressure below 110 mm Hg) in patients undergoing surgery remains controversial. This controversy stems from the difficulty in sorting out the risk of hypertension per se from the risk of hypertension in the setting of hypertensive or atherosclerotic heart disease.
Numerous studies have shown that uncomplicated mild to moderate hypertension (stage 1 or 2), regardless of treatment status, is not an independent risk factor for perioperative complications (1,2,35,36). However, the presence of hypertension may be of consequence because it has been reported that patients with preoperative hypertension may demonstrate marked intraoperative blood pressure lability and postoperative hypertensive episodes (37). Certain medications used for the treatment of chronic hypertension including angiotensin-converting enzyme (ACE) inhibitors and angiotensin II receptor antagonists (ARBs) seem to make intraoperative hypertension more likely (38,39). It is also generally agreed that severe hypertension (stage 3 with systolic pressure greater than 180 mm Hg, diastolic greater than 110 mm Hg) should be controlled with effective oral medications in the days to weeks before an elective operation. Another option for severe hypertension is the use of rapidly acting intravenous agents, which usually can bring blood pressure under control in a few hours. One randomized trial was unable to demonstrate a benefit to delaying surgery in a select patient group with severe hypertension (40).
The causes of perioperative hypertension are presented in Table 18.5. Patients with both hypertensive and atherosclerotic heart disease may be at greater risk than those with uncomplicated hypertension alone. As is the case for cardiac complications, the type of surgery is important in understanding the risk of hypertension. Hypotension due to any cause remains a concern in patients with coronary artery disease.
Table 18.5 Causes of Perioperative Hypertension
Hypertensive management begins with identification, followed by development of a plan for control. In general, most antihypertensive medications should be given on the morning of surgery. β-blockers and clonidine continue to be used in patients with hypertension in spite of many newer agents. It is especially important to continue β-blockers and clonidineto avoid withdrawal and potential heart rate or blood pressure rebound. Because of the possible problem of intraoperative hypotension, several authors have suggested holding ACE inhibitors and ARBs the morning of surgery (41,42). Most clinicians also hold diuretics to avoid volume depletion. Although diuretic use is associated with volume depletion and hypokalemia, the importance of correcting mild degrees of diuretic-induced hypokalemia in the absence of significant heart disease is controversial (43). Repletion should never be rapid and is safest by the oral route or by adjustment of medication.
In the postoperative period, many patients, especially the elderly, need less antihypertensive medication because of the salutary effects of bed rest and relative sodium restriction.If blood pressure is not elevated, it is wise to plan on reinstating drugs stepwise, beginning with the most active agent at approximately half the usual dose and finally adding the diuretic, if used, sometime later. An exception to this would be the use of β-blockers, which, because of their possible benefit in decreasing cardiac events and concern about rebound hypertensive effects and cardiac ischemia if they are stopped abruptly, should be continued in the postoperative setting. Likewise, clonidine should be continued in the perioperative period to avoid rebound hypertension. Patients whose only antihypertensive drugs are thiazide diuretics are best observed in the immediate postoperative period. In general, patients who need additional antihypertensive therapy in the immediate preoperative period should not be treated with diuretics because of the risk of associated hypovolemia and hypokalemia.
It has been estimated that pulmonary complications are at least as common as cardiac complications after noncardiac surgery (44). Atelectasis, postoperative pneumonia, respiratory failure, and exacerbation of underlying pulmonary condition can all develop after abdominal and pelvic surgery. In fact, pulmonary complications are often associated with the longest hospital stays after abdominal surgery (45). Clinicians should attempt to identify patients at increased risk for pulmonary complications from surgical procedures, and reduce these risks whenever possible (Fig. 18.2).
Figure 18.2 Pulmonary evaluation and postoperative care. *Refers to measures to reduce pulmonary complications (see Table 18.6). COPD, chronic obstructive pulmonary disease.
Pulmonary Risk Factors
Pulmonary risk factors are typically grouped as “procedure related” or “patient related.” For “procedure-related” risks, a large literature review recently confirmed the traditional teaching that procedures closest to the diaphragm increase perioperative pulmonary risks (46). This observation is presumably due to the higher likelihood of diaphragmatic dysfunction, or related postoperative pain and shallow inspiration. For this reason upper abdominal surgeries create more postoperative risk than lower abdominal procedures.
General anesthesia, emergency surgery, and prolonged surgeries (greater than 3 hours in duration) also increase the risk of postoperative pulmonary complications (46). There is some evidence that the use of shorter-acting neuromuscular blockers during an operation can reduce postoperative pulmonary complications (47). Most clinicians believe that epidural anesthetics and analgesia, and the use of laparoscopic rather than open surgeries should reduce postoperative pulmonary risks. Comparative data about these techniques and their association with pulmonary complications, though, is still limited (48,49).
A review of multiple studies over the last several years has confirmed several patient-related risk factors for postoperative pulmonary complications: advanced age; ASA (American Society of Anesthesiologists) class ≥2; heart failure; functional dependence; and chronic obstructive pulmonary disease (COPD). A serum albumin level <3.5 mg/dL and overt malnutrition were also associated with increased risk. Current smoking also seems to increase postoperative pulmonary risks (50).
Although, COPD is a known risk factor for postoperative pulmonary complications, there has been considerable debate about the routine use of spirometry to screen for this condition before nonthoracic surgery. The American College of Physicians (ACP) recommends spirometry only if the history and physical examination suggest an undefined lung condition (51). A history of prolonged cigarette smoking, dyspnea, chronic cough, sputum production, wheezing, or prolonged expiration and hyperinflation noted on examination would all be suggestive findings for COPD. Patients with these findings would typically receive pulmonary function testing with spirometry (and possibly a blood gas), whether an operation was intended or not.
Although COPD increases the risk of postoperative complications, it does not typically make these risks prohibitive. Even patients with severe COPD can tolerate abdominal surgery when properly prepared (52). When patients with COPD are identified before elective operations, every effort should be made to optimize their lung function and minimize any other perioperative risks.
Interestingly, studies have not identified asthmatic patients as having significant risks of serious postoperative complications when managed appropriately (53). Isolated obesity has also not been associated with increased pulmonary risks. However, many of the attendant comorbidities with obesity, such as obstructive sleep apnea, are known to increase perioperative risk (54). Sleep apnea patients are often prescribed positive pressure airway masks to assist their breathing at home during the night. This equipment should be available in the postoperative period to assist with any apneic breathing episodes. Obese patients (as well as other patients with unusual upper airway anatomy) may present difficulties for intubation, and fiber optic instruments may be needed in the operating room. Pulmonary hypertension, whether associated with sleep apnea or not, has also been associated with increased postoperative risks in noncardiac surgery (55).
Pulmonary Risk Reduction
Physicians should attempt to reduce perioperative risks whenever possible (Table 18.6). Some identified risks (such as location of surgery, type of anesthesia, age, poor general health status, and fixed airways obstruction) cannot be improved. Smoking cessation, however, should be encouraged. At least one study has shown that 6 weeks of smoking abstinence is needed to reduce smoking-related pulmonary risks (56). Patients with airway obstruction (COPD and asthma) should be optimized to their baseline pulmonary status before surgery. This may involve bronchodilator use, inhaled steroids, or possibly antibiotics and/or oral steroids. As an example, a one-week course of preoperative oral steroids in severe asthmatics has been shown to be safe in reducing the risks of postoperative bronchospasm (57).
Both COPD and asthmatic patients should continue their home medications during their postoperative course. Modern ambulatory therapy for asthma emphasizes the use of inhaled steroids as well as inhaled beta agonists. Exacerbations during the postoperative period can be treated with additional doses of inhaled bronchodilators, and intravenous steroids if necessary.
COPD management typically involves both inhaled β-agonists and anticholinergics. Exacerbations can also be treated with additional doses of inhaled medications or steroids.Noninvasive ventilation has been used successfully in postoperative patients with COPD exacerbations (58).
There is good evidence that lung expansion maneuvers decrease postoperative pulmonary complications (59). Because low lung volumes produced by anesthesia, operative site pain, and bowel distension all contribute to respiratory dysfunction in the postoperative period, clinicians have prescribed deep-breathing exercises, intermittent positive pressure breathing (IPPB), and simple incentive spirometry for years to attempt lung expansion in the postoperative period. Continuous positive airway pressure has also been used with success to decrease postoperative pulmonary complications after abdominal surgeries (60). Adequate pain control is also important to improve deep breathing and lung expansion after abdominal surgery. Recent reviews have also emphasized more judicious use of nasogastric tubes in the postoperative period to reduce postoperative pneumonia and atelectasis (60). Recent guidelines suggest that these tubes be used selectively for nausea and vomiting, inability to tolerate oral intake, or for symptomatic abdominal distension.
Table 18.6 Measures to Reduce Pulmonary Complications
Diabetes mellitus affects approximately 7.8% of the adult population in the United States. Type I diabetes is an autoimmune disease that attacks pancreatic beta cells. People with type I diabetes have a near-total lack of insulin due to pancreatic beta cell destruction and become ketoacidotic if insulin is withheld. Although common in juveniles, type I diabetes can occur in adults. People with type II diabetes are not insulin deficient in an absolute sense and thus are not generally prone to ketoacidosis. The problem in type II diabetes is usually one of relative insulin resistance. Insulin treatment is not limited to type I disease because many patients with type II do require some insulin therapy. Patients with type II diabetes are usually older and overweight. Both groups may experience the complications listed in Table 18.7. Many elderly patients have mild type II diabetes of recent onset related to obesity, are well controlled with diet or oral hypoglycemic drugs, and have few overt complications, but may have occult atherosclerotic vascular disease.
The management of diabetes begins with some understanding of the factors that influence perioperative glucose metabolism. Insulin is the principal glucose-lowering hormone; cortisol, glucagon, growth hormone, and catecholamines are the principal glucose-raising hormones. In the preoperative period, stress and the “dawn” phenomenon may elevate blood glucose. The dawn phenomenon is early-morning hyperglycemia resulting from nocturnal surges of growth hormone. During surgery, cortisol, epinephrine, and growth hormone levels rise. In this period, there is hyperglycemia in diabetic and nondiabetic patients alike. This is caused by glycogenolysis, inhibition of glucose uptake, and decreased insulin release. After surgery, in nondiabetic patients, the hyperglycemia is brought under control by increased endogenous insulin release over a period of 4 to 6 hours. Patients with diabetes may need additional exogenous insulin.
Table 18.7 Complications of Diabetes Mellitus
In addition to these hormonal factors, several other factors are important in modulating the blood glucose level in the perioperative period. Inactivity, stress, and intravenous glucose infusions tend to raise blood glucose. Decreased caloric intake and semistarvation tend to lower blood glucose. Because the net effect of these factors is sometimes difficult to anticipate, it is important frequently to monitor blood glucose levels.
There are many more oral agents being used to treat diabetes than in the past (Table 18.8). Sulfonylureas such as glyburide (Diabeta) remain the most popular. Most sulfonylureas are primarily excreted by the liver. These drugs are typically withheld 24 to 48 hours before surgery, depending on their half-life. They can be restarted when the patient starts eating. The biguanide metformin (Glucophage) is being used more frequently. However, metformin should not be used in the perioperative period and probably should be avoided altogether in systemically ill gynecologic oncology patients. There is a serious risk of lactic acidosis if renal function declines as a result of chemotherapy, dehydration, congestive heart failure, sepsis, radiologic contrast agents, or third spacing. It should not be used in patients with liver disease. Acarbose (Precose) is a complex oligosaccharide glucosidase inhibitor that delays the digestion of ingested carbohydrates. There is little use for this drug in the perioperative period. Repaglinide (Prandin) is a meglitinide that stimulates release of insulin from the pancreas. It is generally used to cover mealtime blood sugar and may have limited role in blood sugar management during the perioperative period when patients are not eating. Its safe use depends on stable renal and hepatic function. The thiazolidinedione rosiglitazone (Avandia) improves peripheral use of glucose by improvinginsulin sensitivity. This class of medication has a risk of fluid retention. It is generally not recommended for patients with NYHA class III or IV status. The use of thiazolidinediones around the perioperative period should be exercised with caution, especially in patients with cardiac disease and/or in those who have received more intravenous fluids during the perioperative period.
The newest oral diabetic medication is dipeptidyl peptidase-4 (DPP-4) enzyme inhibitor, Sitagliptin (Januvia). DDP-4 breaks down incretin hormones. Sitagliptin therefore increases the level of incretin hormones (glucose-dependent insulinotropic polypeptide [GIP] and glucadon-like peptide-1 [GLP-1]) by inhibiting their breakdown. Incretins enhance glucose-dependent insulin secretion, glucose-dependent suppression of inappropriately high glucagon secretion, slowing of gastric emptying, reduction of food intake, and promotion of beta-cell activity. This medication is generally not useful during the perioperative period if patients are nil per os (NPO). Dose adjustment is needed for renal insufficiency patients.
Insulin is the mainstay of in-hospital management of diabetes because it is easily titrated during management. In spite of newer therapies, insulin also remains an important tool in the ambulatory management of diabetes. There are various types of insulin available for treatment of diabetes (Table 18.9). The use of different insulin types depends on the goals of treatment. Long-acting insulins like glargine and detemir are used to cover basal blood sugar needs. Short-acting insulins like ultrashort-acting insulin or regular insulin are used to cover mealtime blood sugar or any elevated blood sugar not covered by the basal insulin.
Table 18.8 Characteristics of Oral Hypoglycemics
Hyperglycemia is known to impair neutrophil function, wound healing, and to increase the risk of wound infection (61,62,63,64,65,66). In addition, it also impairs cardiac ischemic preconditioning (a protective mechanism for ischemic insult), enhances neuronal damage following ischemia, decreases nitric oxide, increases platelet activation, increases inflammatory markers, and increases reactive oxygen species (67), all of which have significant impact on a patient's morbidity and mortality. Several observational studies have shown that hyperglycemic patients have a higher mortality, increased risk of infection, poorer functional recovery, and longer length of stay (68,69,70,71,72,73,74,75). Tight glycemic control has been shown to improve mortality, decrease risk of infection, and decrease length of ICU stay (76,77,78,79,80). In one study of critically ill surgical ICU patients (77), it was shown that tight glycemic control with blood glucose at or around 110 mg/dL reduced blood stream infections, acute renal failure requiring dialysis, blood transfusion, length of mechanical ventilation and critical care, and in-hospital mortality. Mortality at 12 months was reduced as well in the intensive insulin therapy patient (77). The American Diabetic Association and American College of Endocrinology recommend controlling preprandial blood sugar around 110 mg/dL in critical care patients, but differ in noncritical care patients, recommending 90 to 130 mg/dL and 110 mg/dL, respectively (Table 18.10). In considering the merits of tight glucose control, the increased risk of hypoglycemia also needs to be considered.
Table 18.9 Characteristics of Insulin
Details of the management of the diabetic patient who is taking an oral hypoglycemic agent are presented in Table 18.11. A patient with well-controlled diabetes who takes sulfonylureas is at risk of hypoglycemia if sulfonylureas are given while the caloric intake is reduced. Sulfonylureas should be held on the morning of surgery or longer if the medication has a long duration of action. Dextrose infusion should be given to prevent any hypoglycemia. However, glucose monitoring should be performed at regular intervals to ensure the blood sugar falls within acceptable range. Any hyperglycemia can be treated with supplemental insulin. When oral nutrition is reinstated, sulfonylureas can be resumed. Thiazolidinediones can be continued during the perioperative period, but should not be used if the patient has received excessive amount of fluids or developed NYHA Class III or IV heart failure. Glucosidase inhibitors, meglitinides, and DDP-4 inhibitor can be resumed only when the patient is on oral nutrition. Metformin is generally not used during the perioperative period for reasons mentioned above.
Table 18.10 Target Goals of Inpatient Blood Glucose Control
Table 18.11 Details of Perioperative Diabetes Management for Well-Controlled Patients Taking Oral Hypoglycemics
Alternatively, patients can be started on basal-bolus insulin regimen perioperatively instead of resuming on oral hypoglycemic medications. Patients' oral nutritional intake is often unpredictable because of postoperative nausea and vomiting or NPO status for various studies. Total insulin need is between 0.3 and 0.6 U/kg/day depending on patients' insulin resistance status (81,82). Half of that insulin dose should be given as basal insulin with glargine or detemir. The other half should be divided into four doses given every 6 hours or before each meal and at bedtime using ultra short-acting or short-acting insulin. Any additional hyperglycemia should be covered with supplemental insulin with corrective scale. Basal insulin dosage can be increased daily by 10% to 20% if blood sugar is not controlled, or it can be decreased daily by 10% to 20% if patient has episodes of hypoglycemia.
Patients on insulin treatment should not be on sliding scale insulin, as that strategy produces fluctuating high and low blood sugars levels and the blood sugar may be still not adequately controlled (81,83,84,85). In a randomized control trial comparing sliding-scale insulin to a basal-bolus insulin regimen, the latter resulted in significantly improved glycemic control, without significant risk of hypoglycemia (81).
The management of the diabetic patient who routinely takes insulin at home is presented in Table 18.12. Because caloric intake is reduced on the day of surgery, total daily insulindose should be reduced. Usually, half the dose of basal insulin is given the night before or the morning of surgery to cover endogenous glucose production. Patients should be started on a dextrose infusion to prevent hypoglycemia. Glucose monitoring should be performed to make sure the patient's blood sugar is within an acceptable range. Postoperatively, patients likely have increased stress hormone levels and are on a dextrose infusion; thus, hyperglycemia is often seen despite the patient being on nil orally. Frequent blood sugar monitoring is needed for management of these hyperglycemic episodes with corrective dose insulin. Once patients recover from the surgery and start to take oral nutrition, they can be resumed on their home insulin regimen. However, caution should be exercised with any changes in patients' nutritional status.
In critically ill patients or patients with uncontrolled diabetes, continuous insulin infusion is a better strategy for glycemic control (76,77,78,79,80). Continuous insulin infusion with glucose infusion maintains normal insulin sensitivity during the perioperative period and decreases blood cortisol, glucagons, fat oxidation, and free fatty acids when compared with controls (86). An insulin infusion can be started at 1 U per hour and the rate titrated by 0.5 U per hour increments to keep blood glucose levels below 140 mg/dL. Five percent dextrose infusion with or without potassium at rate of 50 to 100 mL/hour should be given as well to avoid any hypoglycemia. With recent data showing improved morbidity and mortality outcomes with tight blood sugar control, many hospitals have instituted insulin drip protocols. Using those protocols tailored to the particular institution is preferable to empirically adjusting insulin drip de novo.
Table 18.12 Details of Perioperative Insulin Management for Well-Controlled Patients Taking Insulin
Hypothyroidism Hypothyroidism is common and may go undetected in patients being prepared for surgery (87). Symptoms include cold intolerance, recent or progressive constipation, hoarseness, fatigability, and changes in cognition. Signs include associated goiter, skin dryness, and a delayed relaxation phase of peripheral reflexes (best demonstrated in the Achilles tendon). Studies have suggested that unrecognized mild to moderate hypothyroidism is clinically important, but fears of hyponatremia, prolonged respirator dependency, hypothermia, delayed recovery from anesthesia, or death are probably unwarranted (88,89). One retrospective study suggested that such patients have more intraoperative hypotension, postoperative ileus, and confusion and that infection is less often accompanied by fever (88).
For patients who are suspected before surgery of being hypothyroid, thyroid hormone levels should be measured. Hypothyroid patients should be treated with replacement hormone and rendered euthyroid before surgery. In urgent situations, patients who are not myxedematous should be given 1 or 2 days of oral replacement before surgery, with careful postoperative follow-up (90,91).
Hyperthyroidism can be a dramatic illness, with tachycardia, fever, and exophthalmos associated with goiter. Other common symptoms and signs include frequent weight loss, fatigue, diarrhea, heat intolerance, tremor, hyperreflexia, and muscle weakness. Hyperthyroidism may be occult in older patients. Unexplained tachycardia, weight loss, arrhythmias, or fever may be the only clinical indicators and always raise suspicions of unrecognized hyperthyroidism in surgical patients. With proper preparation (92), hyperthyroid patients undergoing thyroid surgery do well. However, there are scant data concerning the problems of the hyperthyroid patient undergoing nonthyroidal surgery, such as radical hysterectomy. Exacerbation of the illness into a “thyroid storm” is the usual concern. Because of this, when any patient is suspected before surgery of being hyperthyroid, thyroid hormone levels should be measured. If the diagnosis is confirmed, elective surgery should be delayed until treatment has produced a euthyroid state. (93) In the postoperative period, thyroid hormone levels should be measured when any patient has persistent unexplained tachycardia, fever, or tachyarrhythmias.
Patients taking corticosteroids or those who have taken them in the recent past should be evaluated for the need of supplemental corticosteroid coverage. In general, patients taking less than the equivalent of 5 mg prednisone daily should not have adrenal suppression (94,95,96). There is variability between patients in their response to suppression of the hypothalamic-pituitary-adrenal (HPA) axis by exogenous steroid. In a prospective cohort study, 75 patients were given short-term, high-dose glucocorticoid treatment of at least 25 mg prednisone daily for 5 to 30 days (97). Forty-five percent of the patients experienced HPA suppression. Of those patients, the majority recovered within 14 days. However, a couple of patients remained suppressed at 3 and 6 months.
In a retrospective study, 279 patients were taking prednisone or its equivalent steroid at doses of 5 to 30 mg per day for between 1 week and 15 years (98). Human corticotropin-releasing hormone (CRH) was used to assess HPA suppression. There was a trend toward an inverse correlation between dosage and duration of therapy and the plasma cortisol response to CRH. On the other hand, there were numerous patients taking high-dose steroids for more than 100 weeks who still had an intact HPA axis. Despite this variability,suppression of the HPA axis should be anticipated in patients taking more than 25 mg of hydrocortisone, 5 mg of prednisone, or 0.75 mg dexamethasone per day for more than 3 weeks (99).
To further clarify whether patients on chronic steroids have suppressed HPA axis, a cortrosyn (ACTH) stimulation test can be performed. Baseline cortisol and ACTH levels should be obtained. Immediately 250 µg of ACTH is given intramuscularly or intravenously. A cortisol level is obtained 30 minutes after ACTH is given. If the stimulated cortisol level does not rise above 18 µg/mL, the patient is suspected of having a suppressed HPA axis and stress dose steroid should be given perioperatively (97,100,101,102).
Corticosteroid supplementation for patients suspected of adrenal suppression will depend on the type of surgery performed. Patients having minor surgeries like hernia repair or colonoscopy should be able to take their usual dose of oral steroid on the day of surgery without additional supplementation. Patients undergoing moderate surgical stress, such as hysterectomy, should take their usual steroid dose on the morning of the surgery and be supplemented with 50 mg intravenous hydrocortisone on call to surgery followed by 25 mg intravenously every 8 hours for three doses and resume usual oral steroids on the following morning. Patients undergoing major surgery, such as primary cytoreduction for advanced ovarian cancer, should take their usual steroid dose on the morning of surgery and be supplemented with 100 mg intravenous hydrocortisone on call to surgery, followed by 50 mg intravenous every 8 hours, tapering the dose by half each day over the next 24 to 48 hours. The patient can then resume oral steroids in the morning after tapered off intravenous stress dose hydrocortisone (103,104). For those patients who continue to not be able to take oral medications, equivalent dose of intravenous hydrocortisone should be given in the mornings.
Thromboembolic Disease Prevention
Almost all hospitalized patients are at risk for venous thromboembolism (VTE), and should receive some type of prophylaxis. Surgical patients, in particular, are at increased risk for deep venous thrombosis and associated pulmonary embolism related to immobility and the operative stimulation of the coagulation cascade. One older analysis of historical data suggested nearly a third of hospitalized surgical patients might develop VTE, and perhaps 1% may develop fatal pulmonary embolism if no prophylaxis was given (105).
Although young, ambulatory patients without additional risks who undergo short (<30 minutes) surgeries may not need specific interventions other than early mobilization, almost all other postoperative patients should receive some type of thromboprophylaxis. Gynecologic surgery patients known to be at particularly high risk include those with malignant disease; open abdominal (versus vaginal or laparoscopic) surgeries; elderly patients; and those who have had previous venous thrombotic events. A more complete list of risk factors for venous thromboembolic disease in hospitalized patients is shown in Table 18.13.
Table 18.13 Factors Related to Increased Risk of Thromboembolic Disease
Preventive therapy for venous thrombosis includes mechanical compression devices placed on the lower extremities, and various subcutaneous anticoagulation regimens withunfractionated heparin, low molecular weight heparins, and newer agents such as fondaparinux. Higher doses of the latter agents may have associated risks for postoperative bleeding. There have been conflicting reports on the relative protective benefits of each of these treatments. At least one randomized trial has shown that proper use of compression stockings may be as effective as subcutaneous heparin in major surgeries for gynecologic malignancies (106). Other trials have suggested that higher doses of subcutaneous unfractionated heparin (three times a day) or low molecular weight heparin may be more protective than lower doses of unfractionated heparin (107). Some surgeons have even advocated the use of pneumatic compression devices and subcutaneous anticoagulants in their highest risk patients, although there is no clear data these treatments are additive in effectiveness. Thromboprophylaxis in hospitalized patients is typically continued until hospital discharge. A recent consensus statement summarizes general recommendations for prevention of venous thromboembolism in postoperative gynecology patients (see Table 18.14).
The question of how much preoperative laboratory testing is warranted has been the subject of considerable interest and debate (108,109). It is important for the surgeon to recognize that data from many studies confirm that unless clinical indicators are present, preoperative test results will usually be normal, falsely positive, or truly positive with no significant clinical outcome on perioperative complications (110,111,112,113,114,115).
Table 18.14 Recommendations for Venous Thromboembolic Prophylaxis in Gynecologic Surgery: Eighth (2008) American College of Chest Physicians Guidelines for Antithrombotic Therapy for Prevention and Treatment of Thrombosis
The National Institute for Clinical Excellence of the United Kingdom developed guidelines in 2003 in an attempt to give some directions for clinicians on preoperative testing for elective surgery. The guidelines incorporated as much evidence as possible, but unfortunately, the evidence base is often lacking so that recommendations are frequently based on experts' opinions and consensus (116). These guidelines categorize patients by age; surgical grade (minor, intermediate, major, major+); anesthetic grade as per American Society of Anesthesiologists (ASA); and co-morbidity (cardiac, respiratory, and renal) (Tables 18.15 and 18.16). Preoperative test recommendations cover chest x-ray; complete blood count; electrocardiogram (ECG); coagulation studies; chemistry; renal function; and glucose, urine analysis, blood gas, and lung function tests. The complete guideline is available online (116).
Patients who are younger (<40 years old), healthy, without co-morbidity, and undergoing minor surgery generally do not need any preoperative testing. However, gynecologic oncology patients are at least Surgical Grade 3—major surgery. Such patients often have clinical indicators that support additional testing, particularly when they are older (>60 years old), have higher ASA Class, multiple co-morbidities, or with cancer. A preoperative ECG should be obtained on all women over 60 years of age or younger when patients have cardiac disease. Complete blood count and chemistries are recommended. Glucose level is not recommended by the guidelines in many patients, but is usually warranted to exclude diabetes due to the increased morbidity and mortality associated with inpatient hyperglycemia. Chest x-ray is generally not recommended, but can be considered if the patient has respiratory symptoms or abnormal chest examination. Urine analysis is considered in all gynecologic patients. Interestingly, coagulation studies are not recommended in the majority of the patients, except for those patients with renal co-morbidity or cardiac co-morbidity in ASA Class 3 status. In addition, for those patients with metastatic liver cancer or who are significantly malnourished, coagulation studies prior to surgery are reasonable. Blood gas can be considered in patients with multiple co-morbidities. Pulmonary function tests are not recommended for gynecologic patients even if they have respiratory co-morbidity (Tables 18.17, 18.18, and 18.19).
Table 18.15 Surgical Grade
Table 18.16 American Society of Anesthesiologists (ASA) Class
The guidelines on preoperative tests are merely a general roadmap for the clinicians. Combining patients' medical conditions and symptoms with the guidelines will provide clinicians with a more focused approach to preoperative testing when evaluating patients for elective surgery.
Screening for Hemostatic Defects
A good history and physical examination are often most helpful in screening patients for hemostatic defects before operations. Some of the most important information involves the outcome of prior hemostatic stress and the family history. Minor surgical procedures should not have required transfusion, and a history of postoperative bleeding 2 or 3 days after surgery is also suspicious. Many patients have had tooth extractions. Bleeding should not last more than 24 hours and should not start again after stopping. A familial history of bleeding or suspected bleeding should be investigated. Patients should be questioned about nosebleeds, intestinal bleeding, and heavy menstrual bleeding. Large ecchymosis and mucosal bleeding on examination can be a cause for concern.
Like other laboratory tests, some have suggested that in otherwise healthy patients, screening for hemostatic defects may not be warranted (117). For cancer patients undergoing surgery for which bleeding is expected, some laboratory screening seems prudent. A platelet count, INR (international normalized ratio), and PTT (partial thromboplastin time) are the most commonly ordered screening tests for this purpose. A low platelet count can be caused by decreased production, sequestration into the spleen, or increased destruction. For platelet counts less than 100,000/cc, platelet transfusions may be necessary before the operation, depending on additional risks. Certain commonly prescribed drugs (such asaspirin and nonsteroidal antiinflammatory drugs [NSAIDs]) can inhibit platelet function and should be held for a week before the operation if possible. Renal dysfunction is another common cause of acquired platelet dysfunction.
Table 18.17 Grade 3 Surgery (Major)
Elevated INR and PTT values often reflect blood coagulation protein deficiencies (or inhibitors). Patients with elevated values may require plasma factor replacement before surgery to minimize their bleeding risks.
There are some patients with normal screening laboratory tests who nonetheless have suggestive histories and/or examinations for hemostatic defects. One possible culprit might beVon Willebrand's disease—an inheritable coagulation defect in platelet function. Identification and perioperative management of this and other more uncommon bleeding disorders may require further laboratory testing and the expertise of a hematologist.
Perioperative Antibiotics for Wound Infection Prophylaxis
It is reasonable to administer 1 g cefotetan intravenously or intramuscularly just before surgery and then every 6 hours for two additional doses in patients undergoing extensive gynecologic oncology surgery. One study demonstrated that preoperative antibiotics must be given within 2 hours of surgery to be effective (118). In addition, if bowel resection is anticipated, mechanical cleansing of the bowel on the day before surgery is prudent, with or without oral neomycin and erythromycin base.
Table 18.18 Grade 3 Surgery (Major)
High blood pressure, both labile and persistent, is a common problem for acutely ill patients. Perioperative hypertensive episodes occur commonly in hypertensive patients and occasionally in normotensive patients because of pain, anxiety, stress, medications, and other factors (see Table 18.5). Perioperative hypertension is most common during laryngoscopy and induction (primarily because of sympathetic stimulation) and immediately after surgery, often in the recovery room.
Patients with preexisting hypertension usually require some continuation of their daily antihypertensive medication when they are brought into the hospital or are critically ill, but the dosages may need titration, and diuretic use for blood pressure control is infrequent. Many agents can be converted to an intravenous form or administered with minimal fluid down a gastric tube if the patient is not eating or drinking. The use of β-blockers as antihypertensives in the acute care setting may have additional benefits by decreasing the risks of atrial fibrillation and myocardial ischemia in vulnerable patients, and these agents are often selected as first line agents for this reason. Some studies have suggested that routine use of these agents might decrease cardiovascular mortality after noncardiac surgery in patients at risk (119). At the minimum, care should be taken that β-blockers are not discontinued suddenly as this can cause rebound hypertension and associated problems. Sublingual, short-acting calcium channel blockers (e.g., nifedipine), on the other hand, should be avoided because their use can lead to reflex tachycardia and myocardial ischemia.
Table 18.19 Grade 3 Surgery (Major)
Corticosteroid medications can sometimes cause hypertension in susceptible patients. Mild antihypertensives may be necessary until the steroid dose is lowered or discontinued.
Many hypertensive episodes resolve spontaneously. Patients with pain and anxiety are best treated with appropriate analgesics and anxiolytics. When evaluating the hypertensive postoperative patient, adequacy of ventilation and stable cardiac status should be verified by examination, arterial blood gases, and ECG. Bladder distension can cause elevated blood pressure and should be relieved. Occasionally, a patient may require a continuous intravenous infusion to control severe hypertension. Intravenous drugs with short half-livesare chosen to allow safe titration (the vasodilator nitroprusside or short acting β-blockers are two popular choices), and the patient is changed over to longer-acting agents as their condition stabilizes (120).
Myocardial Injury and Ischemia
Patients undergoing oncologic treatment may have underlying coronary artery disease. The variable stresses in the postoperative period, such as inflammation, increased hypercoagulable state and hypoxemia, can lead to myocardial injury or ischemia. One study of unselected patients over age 50 years undergoing noncardiac surgery showed the risk of postoperative cardiac events was nearly 1.5% (5). Most postoperative myocardial infarcts (MI) occur during the first three days after surgery, at a time when patients may be receiving narcotics that may mask the symptoms of ischemia.
Patients experiencing postoperative MI have an increased hospital mortality. These patients need to be managed promptly by a surgeon and a consulting cardiologist, and observed closely in a monitored bed for any complicating features such as arrhythmia, pulmonary edema, and shock. In addition to correcting anemia, hypoxia, and starting β-blockers in those that can tolerate them, the treatment of such acute coronary syndromes involves the use of anticoagulants such as aspirin and heparin. The use of these agents must be balanced against the risk of postoperative bleeding. Myocardial infarcts are typically divided into ST segment or non-ST segment elevation injury depending on the ECG appearance. Myocardial infarction patients with ST elevation or with hemodynamic instability have improved outcomes if they can receive rapid angiography and angioplasty, while those in a stable condition with non-ST segment elevations can often be managed medically (at least initially). Unless contraindicated, all patients with postoperative myocardial infarction should be on aspirin, β-blockers, HMG co A reductase inhibitors, and ACE inhibitors by the time of discharge (121).
Every physician working in an acute care hospital should be familiar with the use of a defibrillator and the algorithms developed by the American Heart Association for Advanced Life Support (122). Fluid shifts, electrolyte changes, and myocardial ischemia can put the patient receiving treatment for gynecologic malignancy at increased risk for heart rhythm abnormalities.
The tachyarrhythmias, both ventricular and supraventricular, can be quite dangerous and should be electrically cardioverted immediately if the blood pressure is low or the patient is unstable. In less urgent situations, a variety of antiarrhythmic medications are available for chemical conversion and stabilization of these tachyarrhythmias. Many of these medications are proarrhythmic as well, and a search for an underlying cause of the rhythm disturbance is indicated to reduce the propensity for recurrence. Often, when the electrolyte imbalance or other precipitant is corrected the heart rhythm normalizes, and these agents can be discontinued. Persistent or unstable ventricular arrhythmias are often managed in the acute setting with intravenous amiodarone. Supraventricular tachycardias may respond to vagal maneuvers or a rapid bolus of adenosine.
Atrial fibrillation is the most common postoperative tachyarrhythmia and deserves special mention. Once the blood pressure is stabilized in a patient with atrial fibrillation, attempts should be made to control the heart rate. Popular drugs for rate control include β-blockers (if left ventricular function is preserved), diltiazem, or digoxin. If the rhythm persists once other precipitants have been corrected (e.g., hypokalemia, hypoxemia, fluid overload), many clinicians would attempt chemical or electrical conversion if the atrial fibrillation has not been present more than 48 hours. Restoration of normal sinus rhythm often improves cardiac output (CO) and mitigates the risk of stroke from left atrial thrombus forming in the fibrillating chamber. Patients with atrial fibrillation lasting longer than a few days, and who have no contraindication, should be considered for anticoagulation, to decrease their risk of a stroke (123).
Bradyarrhythmias often arise from excessive vagal stimulation. Nausea, bladder distension, pain, and endotracheal tube manipulation can all stimulate excessive vagal tone. As with tachyarrhythmias, attention to blood pressure is paramount. Those patients who develop hypotension should receive atropine and/or catecholamines. Patients not responding to these agents may need urgent transvenous pacemaker placement. Transcutaneous pacing can also be attempted if available at the bedside.
Shock is defined as a clinical syndrome in which the patient shows signs of decreased perfusion of vital organs. Typical findings include alterations in mental status, cold and clammy skin, oliguria, and metabolic acidosis. In general, patients with shock have a substantial decrease in blood pressure, but no absolute value is used to define shock.
The therapeutic approach to these patients is facilitated by a functional classification of shock states. Each class of shock has its own pathophysiologic process and requires a different management strategy. Traditionally, four varieties of shock are described:
Common causes of shock in the perioperative management of gynecologic malignancy include:
A careful physical examination, review of laboratory and test results, as well as consideration of the clinical situation often suggests the etiology of a particular shock syndrome. Therapy should begin promptly, as information is being gathered. In the past, invasive hemodynamic monitoring with pulmonary artery catheters was often used to diagnose and manage these conditions. There use, however, has declined as recent studies have shown they do not clearly improve patient outcomes (124). There has been increased interest in recent years in using central venous catheters and echocardiograms to gather this information noninvasively. The pulmonary catheter may still be useful in certain clinical situations where the diagnosis and treatment for shock remains uncertain.
Regardless of whether measured directly with a pulmonary catheter, or deduced from less invasive data, each of the shock states has an expected hemodynamic profile. Therapy is therefore targeted to the underlying defect in cardiovascular performance:
An algorithm for the management of shock syndromes is shown in Fig. 18.3, although critically ill patients may often develop mixed shock states.
Respiratory failure can be defined as a failure of gas exchange, that is, failure of the respiratory system to accomplish the exchange of oxygen and carbon dioxide between ambient air and red blood cells in amounts required to meet the body's metabolic needs. Respiratory syndromes characterized by difficulty in oxygenation of the blood are grouped under the umbrella term, hypoxic respiratory failure, and those with difficulty removing carbon dioxide from the blood are described as ventilatory failure. It is often helpful for assessment and therapy to consider these as separate entities, although in reality they are closely connected. The arterial blood gas is used to determine the degree and type of gas exchange failure and should be performed as part of the initial evaluation. Patients with respiratory failure commonly have abnormal mental status (agitation, somnolence, and disorientation), and physical findings may include tachycardia, hypertension, and occasionally cyanosis and sweating (Fig. 18.4).
Figure 18.3 Algorithm for the management of hypotension. UOP, urine output; PE, pulmanory embolism.
Common causes of respiratory failure in the perioperative management of gynecologic cancer patients include:
Hypoxic Respiratory Failure
Hypoxic respiratory failure is usually caused by a mismatch between inhaled gas and blood circulation in the lung parenchyma. Blood circulating in areas of mismatch is relatively deoxygenated. The degree of hypoxic respiratory failure can be characterized by the alveolar-arterial oxygen gradient. This value is determined by measuring the arterial oxygen tension with a blood gas, and then calculating the alveolar oxygen tension using known values for the fraction of inspired air that consists of oxygen (dependent on ambient barometric pressure and amount of oxygen supplementation) and the amount of carbon dioxide tension in the alveolus (calculated by measuring the arterial carbon dioxide tension on blood gas and adjusting for the expected exchange into the alveolus to maintain metabolic processes). This calculation is frequently performed at the bedside using the following alveolar gas equation:
Figure 18.4 Management of respiratory failure. ABG, arterial blood gases; CPAP, continuous positive airway pressure; PEEP, positive end-expiratory pressure; SIMV, synchronized intermittent mandatory ventilation; AC, assist control; FIO2, fraction of inspired oxygen; A-a gradient, alveolar-arterial gradient.
Alveolar Po2 = Inspired O2 concentration — alveolar CO2 concentration
Alveolar Po2 = (Fio2 × [barometric pressure - water vapor pressure* ])
- ([Paco2 arterial] × 1.25)
Without oxygen supplementation at sea level, the inspired oxygen concentration is approximately 150 mm Hg. Normal Paco2 is 40 mm Hg in arterial blood. Therefore, the alveolar oxygen concentration in a healthy patient breathing ambient air by the preceding equation is approximately 100 mm Hg. This would then be compared with the measured arterial oxygen concentration on a blood gas sample to describe the alveolar-arterial gradient (alveolar Po2 - arterial Po2). The alveolar-arterial difference in oxygen concentration increases with age, but typically does not exceed 20 mm Hg. A gradient wider than 20 mm Hg is the hallmark of hypoxic respiratory failure.
Treatment of hypoxic respiratory failure involves improving oxygenation of arterial blood as well as attempting correction of the underlying mismatch in lung function. Interventions to improve oxygenation include supplemental oxygen by nasal cannula or mask, or by positive-pressure breathing for refractory hypoxemia. Positive airway pressure serves in part to inflate partially or totally collapsed regions of the lung, often with a dramatic improvement in oxygenation. On a mechanical ventilator, different manipulations can be made to increase airway pressures. Most commonly, this is done by increasing positive end-expiratory pressure (PEEP). Additional measures to help reverse underlying ventilation-perfusion mismatch in hypoxic respiratory failure are as follows:
Ventilatory failure occurs in patients who fail to “excrete” adequate carbon dioxide from their lungs. These problems typically do not arise from mismatch at the alveolar-capillary level, but more likely from failure of the lungs to effectively pump gas out of the respiratory circuit. As Pco2 builds up in the alveoli, the arterial Pco2 begins to rise as well.Hypercarbia on the blood gas measurement is the hallmark of ventilatory failure. “Pump” dysfunction can occur anywhere from the medulla, to the diaphragm, to the thickened or destroyed airways of the patient with COPD. Typical scenarios for the oncologic patient include the oversedated patient with an inadequate respiratory rate, or the weakened patient unable to pump air adequately through diseased lungs.
Treatment of ventilatory failure consists of reversing any precipitants, and if these are not readily correctable, providing an adequate tidal volume and respiratory rate with a mechanical ventilator. This is typically administered through an endotracheal tube with a mechanical ventilator, although there is increasing interest in the use of noninvasive masks to administer continuous or phasic positive airway pressure in certain situations (58).
Chronic Respiratory Failure
Some patients with gynecologic malignancy may have adapted to chronic respiratory insufficiency. These patients with chronic lung disease may have abnormal alveolar-arterial oxygen gradients or carbon dioxide tensions as their baseline equilibrium. Chronic hypoxemia leads to elevated hemoglobin and improved oxygen delivery chemistry, and these patients are not in acute distress unless their Po2 dips into the 50 mm Hg range. Chronic lung disease can lead to carbon dioxide retention, which is compensated by a metabolic alkalosis. Increasing oxygen supplementation beyond that necessary to maintain hemoglobin saturations at the patient's baseline can sometimes lead to worsening pump function in patients with chronic CO2 retention. Likewise, improving ventilation by mechanical means to a “normal” Pco2 on blood gas measurement may lead to dangerous alkalemia in a patient with chronic lung disease who was in acid-base balance at a higher Pco2. The goal of oxygenation and ventilation in patients with chronic respiratory insufficiency should be to maintain their baseline status.
Adult Respiratory Distress Syndrome
One pattern of severe respiratory failure that deserves special mention is the adult respiratory distress syndrome (ARDS). This is a pattern of lung injury that can be precipitated by direct damage (aspiration) or can occur as part of a septic syndrome and resulting lung inflammation. It is characterized by severe hypoxemia and decreased lung compliance thought to be secondary to diffuse capillary leakage into the lung parenchyma. The chest radiograph has the appearance of pulmonary edema, although direct measurement with a pulmonary catheter typically shows low or normal left sided filling pressures. Management typically involves mechanical ventilation with lower tidal volumes to minimize lung distension and to avoid fluid overload (127,128). Despite aggressive support, the mortality rate from this syndrome remains high (129).
Physicians working with critically ill patients need to understand the principles of mechanical ventilation. Postoperative patients sometimes remain on mechanical ventilation until they are stabilized. Even apparently stable oncologic patients on the wards are often at risk for hypoxic and ventilatory failure that can progress to the need for mechanical ventilation.
Mechanical ventilation is typically performed by placement of an endotracheal tube, although tight-fitting masks are sometimes used in patients who are awake enough to protect their airways (noninvasive positive-pressure ventilation) (58). There has been a proliferation in both the types and terminology for mechanical ventilation over the years, often leading to some confusion. Despite the many modalities, little is known about improved benefits of one ventilator setting versus another in terms of long-term patient outcome.
A basic understanding of ventilator management can be divided into two realms (much like the understanding of respiratory failure)—ventilation and oxygenation. Management of ventilation requires adjusting when and how often the machine delivers a breath (with every patient effort or on a timer), and how it delivers that breath (either as a preset volume or applying a preset pressure). These settings are chosen to help the clinician accomplish two goals: full or partial support of the patient's breathing efforts, and adequate ventilation without excessive airway pressures. Management of oxygenation requires adjustment of the fraction of inspired oxygen delivered into the lungs and the end-expiratory airway pressure settings. These values are also set to achieve adequate blood oxygen saturation without damaging the lungs.
When the machine is set to deliver a full mechanical breath with each patient effort, the patient is receiving fully supported ventilation. Typically, a backup respiratory rate is set, but the patient can breathe as often as she wants and receive a fully supported tidal breath each time. This is typically called assist control (AC) ventilation. When the machine is set to deliver only a certain number of breaths each minute, the patient needs to breathe without full machine support for any additional respirations above the set rate. This is considered partially supported ventilation and is most typically set as synchronized intermittent mandatory ventilation (SIMV).
The mechanical breath itself can be delivered as a preset volume with each breath, so-called volume control ventilation. This ensures an adequate tidal volume but risks increased airway pressures if the lungs become difficult to inflate because of increased airway resistance or lung stiffness. High airway pressures can cause barotrauma, such as pneumothorax, and it is recommended to keep peak airway pressures less than 35 cm H2O if possible (130). Instead of volume control, the mechanical breath can be administered as a preset pressure; this is usually termed pressure control (or in a slightly different mode, pressure support). This avoids the risks of increased airway pressures but may provide smaller (or larger) tidal volumes with each breath if lung mechanics (or patient effort) changes. Some of the more sophisticated ventilators can adjust pressures with each breath to meet a targeted volume in a mode called pressure regulated volume control. Regardless of which mode is chosen, arterial blood gases are typically followed for patients on mechanical ventilation, and adjustments are made in the aforementioned settings to keep the patient's arterial carbon dioxide level near her baseline value.
When adjusting oxygenation settings on the ventilator, most critical care physicians attempt to lower the fraction of inspired oxygen (FIO2) to below 65%. Values above this for prolonged periods are believed to be damaging to lung parenchyma (131). The addition of PEEP often increases the functional reserve capacity of the diseased lung and allows FIO2reductions. PEEP should be titrated to maximize oxygenation in respiratory failure, although some caution is needed because higher values can begin to precipitate barotrauma from increased peak pressures. In some forms of acute respiratory failure, such as ARDS, additional measures may be tried for refractory hypoxemia, including lengthening the inspiratory time on the ventilator cycle or changing patients to the prone position. Unfortunately, these interventions have not been shown in any prospective trials to improve long-term outcomes (128).
Once the cause of respiratory failure is improved or improving, and the patient is judged hemodynamically stable, attempts to remove the patient from mechanical ventilation should begin. This process has become known as weaning, although it does not need to be as slow as this appellation suggests. Although clinicians have looked at many screening methods for identifying patients who are ready to come off mechanical ventilation, none offers perfect sensitivity or specificity. Traditional weaning criteria such as negative inspiratory force less than 25 cm and minute ventilation less than 10 L/minute have shown poor predictive value. Many physicians think the bedside test with best predictive accuracy may be the rapid shallow breathing index, which divides respiratory rate (breaths per minute) by tidal volume (liters) measured with the patient removed briefly from ventilatory support. A rapid shallow breathing index value less than 100 breaths/minute/L showed the best predictive value for successful extubation in a recent review of multiple different predictors (132).
In the end, the most useful test to determine a patient's readiness for extubation is probably a trial of spontaneous breathing with little or no support from the ventilator. This is often performed with minimal pressure support (7 cm of H2O per breath or less) or by having the patient breathe through the endotracheal tube without any ventilator support at all. Patients who can tolerate 30 minutes to 2 hours of such unsupported breathing should be considered for extubation if they can protect their airways, manage their secretions, and maintain oxygenation and ventilation (133). Early extubation can help avoid nosocomial pneumonia as well as other complications associated with the mechanical ventilator and a prolonged ICU stay.
Renal Insufficiency, Fluids, and Electrolytes
Acute Renal Failure
Acute renal failure (ARF) or acute renal injury remains a serious postoperative complication in surgical patients. Surgical patients are particularly predisposed to ARF due to the physiologic insult induced by the surgical procedure, preexisting comorbidity, and sepsis. The overall incidence of ARF in surgical patients has been estimated at 1% to 2%, although it is higher in at-risk groups (Table 18.20), and mortality rates remain high despite advances in dialysis and supportive care (134). The most consistent preoperative factor contributing to ARF is preexisting renal impairment (135). At present the best form of treatment is prevention.
While a number of definitions of acute renal failure exist in the literature, the first sign of new or worsening renal dysfunction is a rising serum creatinine concentration or a low urinary output. Oliguria is defined as a urine output of less than 400 mL/day. Acute renal failure typically reflects an abrupt loss and sustained decline in the glomerular filtration rate, which manifests as an increasing serum creatinine to twice its baseline value (136). In the hospital setting, acute oliguria or renal failure is usually caused by hypovolemia, decreased cardiac output, postoperative kidney injury, or the use of nephrotoxic drugs (137). Prerenal causes (see below) are responsible for the majority of cases (up to 90%) of ARF in surgical patients (135). Patients are also at increased risk for acute kidney injury from etiologies arising from cancer treatment such as severe vomiting or diarrhea, nephrotoxic chemotherapy, or tumor lysis syndrome. The cancer itself may also cause ureteral obstruction (Table 18.21) (138).
Table 18.20 Risk Factors for Developing Perioperative Acute Renal Failure
Table 18.21 Common Causes of Acute Renal Failure in Surgical Cancer Patients*
The causes of acute renal failure are grouped into three categories: prerenal, intrinsic, and postrenal (Table 18.21). Initial evaluation attempts to group the patient into one of these classes. Prerenal failure and intrinsic renal failure due to ischemia and nephrotoxins (acute tubular necrosis) are responsible for most cases (135). The physical examination should exclude orthostatic blood pressure changes, evidence of liver disease, a palpable bladder, and an elevated postvoid residual urine volume as determined by bladder catheterization. Laboratory evaluation should include determinations of urinary and serum sodium and creatinine concentrations (to calculate the fractional excretion of sodium [FexNa]; see below), urine osmolality, and microscopic urinalysis. Unfortunately, the results of urinary electrolytes may be unreliable and should be interpreted with caution in patients with glycosuria, preexisting renal disease, and those receiving diuretics.
Postrenal causes of oliguria or acute renal failure arise from obstruction of the urinary tract. If this remains a suspicion even after a urethral catheter is placed, a renal ultrasound can be ordered, which may show characteristic dilation of the collection system (hydronephrosis) above the obstruction. Although quite specific, this finding may not be present in all cases of ureteral obstruction, and additional radiographic tests may be needed (139). Percutaneous or cystoscopic stenting is often performed for cases of acute ureteral obstruction.
Prerenal azotemia can often be diagnosed with urinary indices. It tends to be associated with high urine osmolality, low urinary sodium, and a high urine-to-plasma creatinine ratio. It was shown many years ago that the best discriminator between prerenal and other causes of acute renal failure is the FexNa (140). This can be easily calculated from serum and urine sodium and creatinine concentrations as follows:
Prerenal azotemia is associated with a FexNa of less than 1%, whereas obstructive uropathy (postrenal) and most forms of intrinsic renal failure (except pigment- or radiocontrast-induced acute tubular necrosis) are associated with FexNa levels greater than 2%. Patients with a history or physical examination suggestive of volume depletion, or urinary indices consistent with prerenal azotemia, should be treated with aggressive fluid administration and frequent examination for evidence of volume overload. Characteristic urinary indices for prerenal and other causes of acute renal failure are listed in Table 18.22.
If prerenal and postrenal causes of renal dysfunction are excluded, the patient likely has intrinsic disease of the kidney. Most of these cases are caused by acute tubular necrosis (ATN), although a small percentage of patients may have interstitial nephritis or a form of glomerulonephritis (136). Urinalysis in conjunction with urine microscopy may suggest the etiology. Red blood cell casts in the sediment are diagnostic of glomerulonephritis, “muddy” and cellular casts are suggestive of ATN, and eosinophils in the urine (stained with Wright's stain) are often indicative of interstitial nephritis induced by drugs.
Acute tubular necrosis is usually caused by either ischemia or nephrotoxic agents. Many drugs are known nephrotoxins, including radiocontrast agents and many chemotherapeutic drugs. Clinicians should attempt to avoid nephrotoxicity by careful dosing and avoidance of hypovolemia. Administration of intravenous fluids (isotonic saline or sodium bicarbonate) shortly before administering radiocontrast dye, for instance, has been shown to decrease the risk of nephropathy from contrast agents. While still controversial, studies have suggested a possible reduction in nephrotoxicity of contrast agents with the use of N-acetylcysteine (141,142).
Table 18.22 Urinary Diagnostic Indicesa
One cause of intrinsic renal failure that should be considered in any oncologic patient being treated with chemotherapy is tumor lysis syndrome. This condition is caused by the rapid destruction of large numbers of tumor cells (often just after the initiation of chemotherapy) and results in the sudden release of intracellular phosphate and other intracellular ions (138). Metabolic abnormalities including hyperphosphatemia, hypocalcemia, hyperkalemia, hyperuricemia, and increased serum creatinine can be seen. Acute renal failure develops due to uric acid crystal formation in the renal tubules. The syndrome is most commonly seen in high-grade hematological malignancies such as lymphoma, but it can occur in solid malignancies that respond dramatically to chemotherapeutic agents. Prophylactic measures include hydration and drugs that reduce uric acid levels such as allopurinol and urate oxidases. Urgent hemodialysis is often indicated (143).
Management of patients with acute renal failure includes discontinuation of any nephrotoxic agents and appropriate adjustment of continuing medications to the patient's new creatinine clearance. Adequate fluid resuscitation to ensure euvolemia is critical. Careful attention to volume status, serum electrolytes, and acid-base status is also warranted. Nephrology consultation is recommended for all patients in whom the diagnosis is uncertain or the acute renal failure persists. Many clinicians use diuretics to maintain urine output (144). This strategy may make volume management easier, but is controversial as several studies suggest that diuretics are potentially detrimental and do not improve mortality or make it less likely that the patient will require renal replacement therapy with dialysis (145).
Sometimes the use of renal replacement therapy (hemodialysis) is needed, as there are no effective pharmacological agents for the treatment of established acute renal failure. Indications for hemodialysis include hyperkalemia; pulmonary edema and volume overload that cannot be corrected; acidemia; and uremic symptoms (encephalopathy, bleeding from platelet dysfunction, or pericarditis). The selection of modality of renal replacement therapy (intermittent versus continuous) and the optimal timing of initiation and dose of therapy remain unclear (146). Patients who are hemodynamically unstable might benefit from continuous renal replacement techniques (continuous venovenous hemodialysis or ultrafiltration) rather than intermittent therapy (147). An algorithm for managing oliguria and/or rising serum creatinine (ARF) is shown in Fig. 18.5.
Prevention of acute renal failure in the surgical setting is key. Several preventative measures are essential for high-risk patients (especially those with preexisting impairment). These include (i) optimizing volume status; (ii) keeping the mean arterial pressure >80 mmHg; (iii) reducing the risk of nosocomial infections by rapid removal of intravascular and bladder catheters; (iv) appropriate use of antibiotics; (v) aggressive treatment of any sepsis; and (vi) restricted use or avoidance of potentially nephrotoxic agents (134).
Disorders of acid-base homeostasis are common in critical care medicine, and accurate interpretation of these disorders is important for successful management. For an exhaustive review of acid-base disorders, the reader is referred to several excellent summaries (148,149,150).
The human body requires tight regulation of acid-base balance despite ongoing metabolic processes that produce substantial acid loads. It does this with several buffering systems, all of which are in balance and correct any disturbances. The most important (and most easily measured) is the bicarbonate-carbonic acid equilibrium. This buffering system is reflected in serum bicarbonate levels and carbon dioxide tension (which is in equilibrium with serum carbonic acid). Changes in these measurements from baseline reflect changes in acid or base balance.
Changes in serum carbon dioxide tension (Pco2) reflect either primary lung disorders (hyperventilation or hypoventilation) with resulting respiratory disturbances in acid-base balance, or attempts by the lungs to compensate for metabolic disturbances causing changes in bicarbonate concentration in the blood. Hyperventilation as a primary disturbance results in a low Pco2 in the blood and resulting respiratory alkalemia. Hypoventilation as a primary disturbance raises Pco2 in the blood and causes a respiratory acidemia. On the other hand, when the primary acid-base disturbance is caused by a metabolic disturbance, the patient's ventilatory system attempts to keep the pH balanced by hyperventilation or hypoventilation, thereby creating compensatory changes in the Pco2. By studying the serum pH and comparing changes in the serum Pco2 to changes in the serum bicarbonate concentration, the clinician is often able to distinguish primary respiratory alkalemia and acidemia (as well as their duration) from secondary compensation (150,151). Bedside nomograms have been designed for this purpose as well.
Figure 18.5 Management of a rising serum creatinine.
Table 18.23 Causes of Metabolic Acidosis
Primary changes in the serum bicarbonate concentration often reflect metabolic processes initially less obvious than primary lung disturbances. Metabolic acidosis is defined as a decrease in serum bicarbonate level and occurs as a primary disorder or as a compensation for a respiratory disturbance. Typically, the first step in evaluating a patient with a primary metabolic acidosis is to measure serum electrolytes and calculate the anion gap. A formula for the anion gap using serum electrolyte concentrations is:
Anion gap = (Na+) - (Cl- + HCO3-)
A “normal” anion gap is 10 to 14 mEq/L. Selected causes of metabolic acidosis with elevated and normal anion gaps are presented in Table 18.23.
The second step in evaluating a metabolic acidosis is assessment of the adequacy of the patient's ventilatory response. The normal mechanism of compensation for decreased serum bicarbonate is hyperventilation, which lowers the Pco2 and offsets the impact of the decreased bicarbonate on serum pH. The expected response to a primary metabolic acidosis can be estimated by the following equation (151):
Expected Pco2 = 1.5 (measured HCO3-) + 8 (range ± 2)
Patients who have metabolic acidosis and whose measured Pco2 levels fall below those expected on the basis of this equation should be suspected of having a second disturbance (i.e., an additional respiratory alkalosis). In patients with a Pco2 higher than this expected level, additional respiratory acidosis should be suspected as complicating their metabolic disturbance (151).
The treatment of metabolic acidosis depends on its severity. In most cases, identification and treatment of the underlying cause is the only direct therapy necessary. In patients who have profound disturbances and bicarbonate levels less than 10 or pH less than 7.2, especially if there is associated hypotension or if the underlying disease is expected to worsen, bicarbonate therapy can be considered. Bicarbonate therapy is controversial and should be undertaken with caution, because there is a theoretical risk of causing a transient worsening of the cerebrospinal fluid pH or of inducing fluid overload and rebound metabolic alkalosis. Some researchers have suggested the administration of exogenous bicarbonate may even worsen the outcome in lactic acidosis (152,153).
Metabolic alkalosis can also occur in hospitalized patients. Perhaps most commonly, metabolic alkalosis is associated with volume contraction. In such conditions, sodium reabsorption by the kidney is linked to bicarbonate resorption. Metabolic alkalosis does not resolve until the patient regains intravascular volume. To determine the primary precipitant of metabolic alkalosis and the appropriate treatment, the clinician can use urinary chloride measurements to divide patients into two groups (provided the patient has not received recent diuretic therapy): (i) patients with very low urinary chlorides. These include those who have received nasogastric drainage, diuretic therapy, have been vomiting, or have lingering alkalosis after hypercapnic lung failure. These “chloride-responsive” patients are treated with normal saline solution; (ii) patients with higher urinary chloride concentrations. These patients do not respond to sodium chloride and must be managed by treatment of the underlying disease (149). Table 18.24 lists the causes of metabolic alkalosis.
Table 18.24 Differential Diagnosis to Metabolic Alkalosis in Gynecologic Oncology Patients
Proper management of water and electrolyte therapy is an integral component in the care of surgical and oncologic patients, particularly those who are not taking oral hydration or nourishment. In the average adult who is taking fluids orally, the average daily loss of water is approximately 3 L (2 L as urine and 1 L as insensible losses from perspiration, respiration, and feces). The condition of critically ill patients may be complicated by additional ongoing losses, derangements in renal function, increased insensible losses, and disturbances in free water metabolism induced by the underlying disease. Successful management of these patients requires frequent monitoring of volume status and serum electrolytes. Predictable losses of fluids and electrolytes must be replaced, particularly those from nasogastric suctioning and the increased insensible losses associated with fever and diarrhea (154).
Several simple guidelines can be kept in mind when managing fluid replacement in the hospitalized patient. In a patient with no preexisting renal disease and no disorder of water or electrolyte metabolism, a reasonable maintenance fluid regimen is 3 L daily of a halfnormal saline solution with 20 mEq of potassium chloride in each liter. In the presence of significant renal impairment (glomerular filtration rate <25 mL/minute), potassium therapy should not be given routinely, replacement being based on serial determinations of serum potassium. In patients suspected of having a defect in free water excretion (see below), it is prudent to decrease the free-water content of the initial maintenance fluids (typically by giving normal saline at half the rate). Gastric fluid is composed of hypotonic saline solution (one-fourth to one-half normal saline) with 5 to 10 mEq/L of potassium. Gastric fluid losses should be replaced with replacement fluids in addition to the maintenance prescription.
Hyponatremia and Hypernatremia
Hyponatremia is a common disorder in gynecologic oncology patients. Serum sodium concentration reflects total body water content. Total body sodium content is reflected in extracellular fluid volume. Hyponatremia represents a relative water excess. Disorders of sodium excess or deficit are expressed as either extracellular volume overload or depletion. Hyponatremia is common, affecting up to 15% of hospitalized patients. The most common causes of hyponatremia in hospitalized patients are hypovolemia and the syndrome of inappropriate diuretic hormone secretion (SIADH) (155). Hyponatremic conditions are best grouped into three different categories:
The treatment of hyponatremia is tailored to its pathophysiology. Immediate treatment depends on the patient's symptoms, the rate at which the hyponatremia has developed, and the absolute serum sodium concentration. Patients with diminished extracellular volume are treated with an infusion of normal saline. Patients with normal or increased extracellular volume can be managed initially with free water restriction (157). Those with persisting or worsening hyponatremia and adequate extracellular volume can be managed acutely with furosemide to induce a hypotonic diuresis, and then with replacement of urine output with normal saline infusion. Therapy with hypertonic saline is rarely necessary,and reserved for patients with profound hyponatremia typically associated with seizures and/or markedly diminished mental status. An algorithm detailing an approach to the patient with hyponatremia is presented in Fig. 18.6.
Hypernatremia, less commonly encountered in hospitalized patients, represents relative total-body water deficit (158). Usually it is the result of inadequate water replacement in a patient unable to take fluids spontaneously. This might be exaggerated or precipitated by failure of the kidneys to adequately reabsorb water (concentrate urine), a condition termed nephrogenic diabetes insipidus. Hypercalcemia affects the kidney's ability to concentrate. Hyperglycemia can also worsen water losses by causing an osmotic diuresis. Treatment of hypernatremia is directed at providing adequate hypotonic fluids (often as “free water” or fluids with very minimal solute) and treating hypercalcemia and/or hyperglycemia. Rare patients may have disorders of antidiuretic hormone manufacture and secretion in the hypothalamus and posterior pituitary (central diabetes insipidus) and must receive exogenous hormone to maintain water balance.
Hypokalemia and Hyperkalemia
Disturbances in serum potassium concentration are common and important because of the pivotal role played by this ion in maintaining transmembrane potentials of the heart. Because 98% of total-body potassium is intracellular, small changes in serum potassium concentration may reflect very large excesses or deficits in total-body potassium content. For instance, a decrease in the plasma potassium concentration to 3 mEq/L can reflect a 100- to 200-mEq deficit in total-body potassium content; a decrease to 2 mEq/L can reflect a total-body deficit of 300 to 500 mEq of potassium.
Changes in hydrogen ion concentration can have an impact on the distribution of potassium between the intracellular and extracellular spaces. In acidemic patients, there is a shift in potassium from intracellular to extracellular sites. In a patient who is acidemic and hypokalemic, the plasma potassium concentration is not appropriately diminished, and the total-body potassium deficit will be underestimated.
Possible causes of hypokalemia include decreased dietary intake or insufficient replacement in maintenance fluids, often worsened by diarrhea, nasogastric suction, or diuretic therapy. Hypokalemia can present as weakness, ileus, and muscular cramps. Of most concern, hypokalemia increases myocardial irritability and can precipitate dangerous arrhythmias (159). Treatment involves reversal of the underlying cause and repletion of the potassium deficit. Potassium is replaced relatively slowly in most circumstances to allow cell membranes to equilibrate. In general, patients should not receive more than 10 mEq/hour intravenously. Patients undergoing potassium therapy in the presence of renal failure have a diminished capacity to excrete potassium, and therefore added caution is needed to avoid the dangers of hyperkalemia (discussed below).
Figure 18.6 Algorithm for the evaluation of hyponatremia. ADH, antidiuretic hormone.
Common causes of hyperkalemia include renal insufficiency and decreased ability to excrete daily potassium load, cellular breakdown (including hemolysis) and increased potassium release into extracellular fluids, and redistribution from the intracellular to the extracellular compartment associated with acidemia. Occasionally, high measured serum potassium results from hemolysis of the drawn blood sample in the test tube and does not accurately reflect the serum concentration (159).
Patients with elevated serum potassium typically have no symptoms. The condition is usually noted on screening laboratory tests or when changes are noted on an ECG. Although variable, ECG changes associated with elevated serum potassium include peaked T waves, prolonged PR interval, and widening of the QRS complex. These changes often herald dangerous serum potassium levels that need correction to avert cardiac arrest.
The initial approach to hyperkalemia is to identify and remove the precipitating cause and rapidly assess any ECG changes. In patients with mild hyperkalemia (serum K+ < 6 mEq/L) and minimal ECG changes, treatment of the underlying cause and careful monitoring of the serum potassium levels may be the only therapy necessary. In patients with potassium levels greater than 6.5 mEq/L and evidence of QRS widening, rapid steps should be taken to decrease serum potassium levels with the use of oral or rectal potassium exchange resins such as sodium polystyrene sulfonate (Kayexalate) and a loop diuretic such as furosemide. If there is associated renal failure, urgent arrangements for dialysis are indicated. In patients with prolonged QRS duration approaching sine-wave configuration, or in patients who are hypotensive, the following treatments can “temporize” until more definitive treatment is arranged: calcium gluconate to stabilize myocardial cell membranes, intravenous glucose and insulin (one unit of insulin for each gram of glucose in an ampule of glucose), and sodium bicarbonate (156).
Hypercalcemia is associated with malignant disease and deserves mention. There are several mechanisms for the development of this disorder, the most common in gynecologic oncology patients being increased osteoclastic bone resorption. This is believed to result from tumor secretion of humoral factors. Clear cell and small cell tumors of the ovary are commonly associated with this syndrome.
The clinical presentation of hypercalcemia includes lethargy, confusion, psychiatric disturbances, polyuria (caused by a concentrating defect in the kidney), constipation, and occasionally abdominal pain and nausea. Acute management includes hydration with normal saline and administration of a loop diuretic to increase urinary calcium excretion. Subsequent treatment is targeted to the underlying cause (treatment of the tumor) and also may include the use of intravenous bisphosphonates to control the elevated calcium level (159).
Patients who cannot eat for several days should be considered for nutritional support (beyond the minimal calories available in glucose-based maintenance fluids). Although the criteria for this intervention are not well defined, and clinical trials have shown variable results, most clinicians would consider this intervention after several days without adequate calories and several more anticipated. Enteral feeding is preferred because it may protect patients from gastrointestinal bleeding and infectious complications (160).Parenteral feeding usually requires central line placement and carries additional risks. One large trial in postsurgical patients has shown these risks are outweighed by benefits only when the patients required parenteral feeding for longer than 14 days (161). Various enteral and parenteral feeding formulas are available, and adjustments in constituents are often needed to avoid many of the fluid and electrolyte problems described previously.
Anemia/Red Blood Cells
Anemia can be an additional stress on the cardiovascular system and may worsen myocardial ischemia or heart failure. Appropriate preoperative or perioperative transfusion may reduce cardiac morbidity in patients with significant coronary artery disease or heart failure (6). Some studies suggest that even mild preoperative anemia is associated with increased risk of postoperative cardiac events and mortality in elderly patients undergoing noncardiac surgery (162).
The evidence base for what threshold to use and which patients benefit from transfusion is poor (163). One study suggested that restricting red cell transfusion to patients with hemoglobin levels less than 7.0 g/dL (and who were not actively losing blood) did not worsen outcomes (164). Other investigators have demonstrated harm in transfused patients including increased rates of infection, myocardial ischemia, and postoperative morbidity and mortality (165). In addition, there are risks of transfusion reactions and transmission of communicable diseases. Due to these factors, a conservative or “restrictive” approach with respect to transfusion may be at least as effective and possibly safer than a more liberal transfusion policy (166). Patients with unstable angina and risks for cardiac ischemia or other evidence of inadequate perfusion of vital organs (see below) may require more liberal blood transfusion. Some studies suggest that patients with acute coronary syndromes do better with hemoglobin levels kept closer to 10.0 g/dL (167). Current guidelines on perioperative transfusion practices have been recently published (163).
Red blood cells can be transfused in the form of whole blood or packed red blood cells. Whole blood contains red cells as well as platelets and plasma. Packed red blood cells are red cell concentrates that are prepared by removal of most platelets and all but approximately 100 mL of plasma from a unit of whole blood. In addition, red cells can be washed to remove leukocytes and contaminating plasma proteins for transfusion to selected patients who have had febrile reactions to them.
With the possible exception of the patient with massive exsanguination, there is little advantage to transfusion of whole blood and, as a practical rule, packed red blood cells should be used when red cell therapy is indicated. The possible indications for transfusion of red blood cells include:
Acute Hemolytic and Nonhemolytic Reactions
Acute hemolytic and nonhemolytic transfusion reactions are also possible. Nonhemolytic transfusion reactions are often characterized by fever, chills, or urticaria. Hemolytic transfusion reactions can be life threatening because of associated hypotension, disseminated intravascular coagulation (DIC), and renal failure. It is often manifested by fever, chest pain, back pain, hypotension, tachycardia, and red urine indicating hemoglobinuria (163). When a patient undergoing a red cell transfusion has any signs or symptoms suggestive of a hemolytic transfusion reaction, the transfusion should be stopped immediately and the remaining aliquot of blood sent to the blood bank, along with a sample of the patient's blood for culture and repeat cross-matching. A screening for DIC, a urinalysis for hemoglobin, and a blood sample for bilirubin also should be obtained. In patients with symptoms of a hemolytic transfusion reaction who, on analysis, show no evidence of hemolysis, a hypersensitivity reaction to transfused leukocytes or plasma proteins contaminating the red cells should be suspected. The incidence of hemolytic transfusion reactions can be minimized by careful attention to clerical information, ensuring that the patient is receiving blood cross-matched to her blood sample, and careful cross-matching in the blood bank.
Platelet concentrates are prepared by removal of platelets from whole-blood fractions. Platelets can be stored for 5 days, and each 50-mL platelet “unit” from a whole blood fraction contains approximately 6 × 1010 platelets. Platelets may also be obtained by pheresis from a single donor, and each “unit” of single donor platelet pheresis is equivalent to 5 or 6 whole blood fraction units.
Platelet transfusions are indicated in patients with:
Each whole blood fraction unit of platelets should be expected to raise the recipient's platelet count approximately 5,000 to 10,000/mL3. In general, prophylactic platelet transfusions are indicated only in patients whose platelet count is expected to recover in the future because platelets express human leukocyte antigens and hence induce antibodies in the recipient. After prolonged platelet therapy, most patients become resistant to platelet transfusions, presumably because of immune destruction of all transfused platelets.
Several plasma fractions are available for transfusion. The two most commonly used fractions are fresh frozen plasma and cryoprecipitate.
Fresh Frozen Plasma (FFP) All the blood-clotting proteins present in the original unit of blood are contained in fresh frozen plasma, and it is an adequate source of all coagulation factors for the treatment of mild coagulation factor deficiencies. FFP may be used to rapidly reverse warfarin toxicity if bleeding is present. FFP should also be given in bleeding patients if the international normalized ratio (INR) or activated partial thromboplastin time (aPTT) is elevated (163). Plasma may also be used in reversing the coagulopathy associated with massive blood loss and red cell replacement by restoring the lost coagulation factors. In patients undergoing transfusion of multiple units of red blood cells, the INR and aPTT should be monitored for evidence of coagulopathy and FFP given as needed to correct this. It may require several “units” of fresh frozen plasma to restore clotting factors to adequate levels. The half-life of transfused clotting factors is measured in hours, and bleeding risks can recur when these factors are consumed.
Cryoprecipitate Cryoprecipitate is produced by freezing of plasma, followed by thawing, and produces a precipitate rich in factor VIII and fibrinogen. This cryoprecipitate fraction contains approximately 250 mg of fibrinogen per unit and 80 clotting units of factor VIII. Cryoprecipitate units are smaller in volume than fresh frozen plasma and can be useful adjuncts in treating certain bleeding conditions such as DIC (see below). Cryoprecipitate should be given to bleeding patients when fibrinogen concentrations are less than 80 mg/dL (163).
Massive Blood Transfusion
Pelvic surgeons may at times encounter unexpected and dramatic intraoperative bleeding. This may require large amounts of transfused blood, typically given as packed red blood cell (RBC) units. “Massive transfusion” is typically defined as replacing the entire blood volume (5-6 L in a 70-kg patient) over a 24-hour period, or half the blood volume over a 3-hour period. Such “washout” coagulopathy has been shown in some studies to begin when greater than 10 units of packed red cells have been transfused (168,169). Such rapid blood transfusion can result in a dilutional coagulopathy, as these transfused units do not contain adequate amounts of blood proteins or platelets. This can be worsened by the hypothermia and acidosis brought on by hypovolemic shock.
The INR, aPTT, fibrinogen concentration, and platelet counts should be followed closely in these cases. In general, INR and aPTT levels should be kept less than 1.5 times control values, fibrinogen maintained greater than 100 mg/dL, and platelet count supported to greater than 50,000.
Disseminated Intravascular Coagulation
Disseminated intravascular coagulation is a syndrome that complicates the course of a variety of disease states and is characterized by the pathologic activation of the coagulation cascade and the fibrinolytic system. It occurs most commonly in critically ill patients with sepsis and/or liver disease. In gynecologic oncology patients, it might also be associated with certain mucin-producing adenocarcinomas (170).
In its acute form, DIC appears rapidly and is manifested by bleeding from multiple sites, including venipunctures, surgical wounds, gingiva, gastrointestinal tract, and skin. More rarely, DIC can take a chronic course over a period of months, with thrombotic complications more common than bleeding complications. This form is more typical for the syndrome associated with adenocarcinomas.
Direct evidence of DIC requires demonstration of intravascular fibrin deposition. The laboratory diagnosis of DIC depends on indirect evidence of coagulation activity. The most common laboratory abnormalities in this disorder are decreased platelet count; elevated INR and aPTT; elevated D-dimer (a breakdown product of fibrin); and decreased fibrinogen concentration (170,171).
The primary treatment of DIC is aimed at controlling the underlying cause. Typical treatment would include empiric antibiotic therapy when sepsis is suspected, treatment of other conditions adversely affecting coagulation, and replacement with appropriate clotting factors in patients with active bleeding. Some of the controversial options, such as anticoagulation, factor replacement, epsilon aminocaproic acid with heparin, and antiplatelet drugs, might also be considered (170).
Adverse Effects of Blood Transfusions
Some adverse effects of transfusions include bacterial contamination, transfusion-related lung injury (TRALI), transmission of infectious diseases, and transfusion reactions. Bacterial contamination, most commonly from platelets, is the leading cause of death from blood transfusion and often presents with a fever within 6 hours of the transfusion. TRALI is noncardiogenic pulmonary edema from an immune response of the recipient to leukocyte antibodies in the transfused blood. Hypoxia, fever, and dyspnea typically appear a few hours after transfusion. No specific therapy is available and treatment is limited to stopping the transfusion and supportive measures (163).
Blood products have become much safer with the development of better screening techniques. Transmission of infectious diseases, including hepatitis and human immunodeficiency virus, is rare but still occurs. While not zero, the risk of tranfusion-transmitted viral infections, Hepatitis B and C and human immunodeficiency virus (HIV), is very low. Currently, bacterial contamination and sepsis are the most common infectious complications of blood transfusions (172). Cytomegalovirus can still be a problem, although seroconverters are typically asymptomatic. The use of pooled products, such as platelets and plasma, increases the risks because multiple donors are required. Autologous blood donation may be considered for elective surgery in certain patients.
Detection of Deep Venous Thrombosis
Clinical signs and symptoms of venous thromboembolism (VTE) are notoriously unreliable, and detection of thromboembolic disease once it occurs is a challenging problem in acute care medicine. Deep venous thrombosis (DVT) of the lower extremity classically presents as a swollen, painful leg, but this may be completely absent in some patients. Lower-extremity venous clot is now typically screened with compression ultrasonography of the femoral, popliteal, and calf vein trifurcation. This method is greater than 90% sensitive for proximal thrombosis, but much less sensitive for calf thrombosis. Clinicians should perform serial testing with this method or use contrast venography (gold standard test) if clinical suspicion remains high (173).
Detection of Pulmonary Embolus
Pulmonary embolism (PE) can be a life-threatening emergency. Autopsy studies suggest that pulmonary embolism may account for as many as 5% of unexpected deaths in hospitalized patients (174). Presenting symptoms and signs of pulmonary embolism can include pleuritic chest pain, shortness of breath, cough with or without hemoptysis, and, in the case of a very large embolus, acute clinical decompensation or syncope, and hypotension. Patients are often tachycardic and hypoxic, the ECG may show evidence of right heart strain, and the chest radiograph may show an infarct with effusion. Unfortunately, these same symptoms and signs may be present in other conditions, such as pneumonia and myocardial ischemia. Conversely, a pulmonary embolism can occur with only a few, or even none of these clinical features. Because of this variable presentation, clinicians must be vigilant, particularly in their relatively immobile postoperative patients (175).
Once they suspect it, most clinicians attempt to confirm pulmonary embolism with radiographic testing. The reference standard for diagnosis of pulmonary embolism remains direct angiography, although this test is rarely performed today. The two most commonly performed tests for pulmonary embolism are the ventilation-perfusion (VQ) scan of the lung and the CT angiogram of the chest. These tests are most valuable when combined with pretest suspicion. Unfortunately, in the case of VQ scanning, a large study in hospitalized patients as compared to conventional pulmonary angiogram showed that VQ test results were often not conclusive. An entirely negative test was helpful in excluding a pulmonary embolus (risk of pulmonary embolism <1%), and a classic positive test obviated further testing, but, unfortunately, the test results were more typically intermediate or low probability. Combining results of the ventilation-perfusion scan with pretest clinical probability did improve its accuracy if clinical suspicion correlated with the scan results (176). One advantage of VQ scanning is the absence of the contrast nephropathy risk associated with CT angiography; VQ scanning is often preferred in patients with marginal renal function.
CT angiography performed with multi-detector CT scanners and delayed venography of the thighs has been shown to be both sensitive and specific for VTE detection in recent trials.If the patient's renal function can tolerate angiography dye, this test has largely supplanted VQ scanning at many centers. The predictive value of this test (like the VQ scan) does decrease if the test result conflicts with pretest suspicion (177).
In the outpatient or emergency room setting, assays for increased blood levels of D-dimer (one of the byproducts of the coagulation cascade) are sensitive tests for VTE, and can be used to screen for DVT or PE. Unfortunately, this test is relatively nonspecific in the hospitalized, postoperative, and cancer patients who often have many other reasons for elevated d-dimer levels (178).
Once documented, DVT or PE requires immediate treatment to decrease the risk of complications (pulmonary embolism or recurrent embolism). Patients who do not have an overriding contraindication to anticoagulation should be treated immediately with heparin even before the diagnosis is confirmed. Traditionally this has been treated with intravenous unfractionated heparin in a large initial bolus, followed by a continuous heparin infusion adjusted to achieve an aPTT approximately two to three times that of the control value. More recently, several trials have shown that treatment with low-molecular-weight heparin is equal to, if not better than, treatment with standard unfractionated heparin for pulmonary embolism (179). Postsurgical and intensive care unit patients, however, are more often treated with unfractionated heparin infusions that are more easily reversed because of shorter drug half-lives.
Thrombolytic therapy can be considered in cases of acute pulmonary embolism with hypotension. Many postoperative patients, however, are at high risk for bleeding from lytic therapy, and this treatment should be carefully considered.
If patients with pulmonary embolism have active bleeding, or a high risk of bleeding and cannot be anticoagulated, a filter can be inserted into the inferior vena cava (IVC) to prevent recurrent pulmonary emboli. A percutaneously placed filter into the IVC can also be used for those patients who develop recurrent pulmonary emboli despite adequate anticoagulation (179). Temporary IVC filters have been developed that can be removed several months after placement, and their role in the treatment of pulmonary embolism is likely to increase (180).
Once a patient with pulmonary embolism is stabilized, treatment with intravenous heparin is often transitioned to oral vitamin K antagonists (such as Coumadin) for continued outpatient treatment. The Coumadin oral dosage is titrated to an INR of 2.0 to 3.0. Alternatively a patient may go home on subcutaneous injections of low molecular weight heparinor fondaparinux.
The optimal duration of anticoagulation for deep venous thrombosis and/or pulmonary embolism remains controversial, but in general, these medications are continued for at least 3 to 6 months. Recent studies have suggested the risk of recurrent thrombosis can persist indefinitely, particularly in patients with underlying malignancy (181). For select patients, clinicians now consider lengthy or even lifelong treatment with anticoagulation.
In general, infections in patients with gynecologic malignancies should be managed as in all hospitalized patients—that is, with antibiotics chosen initially on the basis of possible infecting organisms and changed if necessary when the results of culture and sensitivity testing are known (182). Clinicians should be particularly alert to possible central venous catheter infections because many oncologic patients require these devices for treatment. Central venous catheter infection rates range up to 5% depending on the type of catheter inserted (183,184).
Fever in the Neutropenic Patient
Febrile neutropenia is one of the most serious adverse events related to antineoplastic chemotherapy and causes significant morbidity and mortality. When the absolute granulocyte count falls below 1,000/mm3, the incidence of infection rises and infected patients frequently decompensate rapidly. Many clinicians use prophylactic antibiotics in an attempt to decrease the frequency and mortality of this condition. While the most recent guidelines advise against the use of prophylactic antibiotics (185), several more recent studies have provided evidence for the benefit of prophylactic levofloxacin (186,187). In patients with solid tumors receiving moderately myelosuppressive chemotherapy, prophylactic levaquinhas been shown to reduce the incidence of febrile neutropenia and all-cause mortality in the first cycle of treatment (188). Controversy about issues such as cost, toxicity, and antibiotic resistance remain.
Patients with established neutropenia who become febrile can decompensate rapidly and should be treated immediately with empiric broad-spectrum intravenous antibiotics, regardless of whether focal signs of infection or a positive culture result is present. Frequently encountered pathogens in the neutropenic patient include Staphylococcus aureusand gram-negative enteric organisms such as Escherichia coli, Klebsiella, and Pseudomonas.
There are many combinations of empiric antibiotics that achieve the goal of covering likely pathogens in the febrile neutropenic patient (185,189,190). For hospitalized, low-risk patients who have fever and neutropenia during cancer chemotherapy, empiric therapy with oral ciprofloxacin and amoxicillin-clavulanate appears to be safe and effective (189). This approach is as effective as intravenous therapy and may therefore be the preferred approach (190). If intravenous antibiotics are considered necessary, a broad-spectrum semisynthetic penicillin, sometimes combined with an aminoglycoside and vancomycin, may be used. If febrile neutropenia persists despite several days of broad-spectrum antibiotics, guidelines suggest the empiric addition of antifungal agents (185). Meticulous daily followup and adjustment of antibiotic coverage is critical and likely more important than the precise initial combination chosen.
Use of growth factor rescue for febrile neutropenia has been studied, and the use of granulocyte colony-stimulating factor (G-CSF) can shorten the duration of neutropenia. G-CSF is sometimes prescribed prophylactically during courses of chemotherapy, although this usage remains controversial in many treatment regimens (191,192).
Fungemia is a life-threatening postoperative complication of surgery and severe medical illness. Typical patients at risk include those receiving multiple antibiotics and hyperalimentation. Central venous access lines and Foley catheters can provide entry sites for fungal organisms. Additional important risk factors are cancer, chemotherapy, corticosteroids, and hyperglycemia (193,194). The clinical presentation of disseminated disease is identical to that of gram-negative sepsis. These patients may have signs of local fungal disease, such as oral thrush. The principal organisms found are Candida species. There are now several antifungal agents available to treat invasive disease (195). Patients with localized fungal infections should be aggressively treated with topical agents.