Trauma, 7th Ed.

CHAPTER 55. Principles of Critical Care

Raul Coimbra, Jay Doucet, and Vishal Bansal


The last two decades have seen great advances in the care of the injured patient, both in prehospital triage and transport and in the intensive care unit (ICU). More than ever, the outcome of these patients, in terms of both morbidity and mortality, is very dependent on a solid understanding of the pathophysiology and also of the evolution of certain injuries. Attention to detail is critical and an awareness of the pitfalls is essential, if one has to be successful in avoiding preventable morbidity and mortality.

Excluding early deaths in the operating room, most complications and subsequent deaths following injury will occur in the ICU. Care in the ICU is designed to reestablish normal homeostasis and minimize complications of primary, secondary, and iatrogenic injury. Surgical critical care is inherently different from medical intensive care insofar as surgical patients, and particularly trauma patients, require intensive care as the result of an acute surgical intervention or injury and not as part of the (often inexorable) progression or exacerbation of chronic disease. This fundamental difference affects a multitude of patient management practices and decisions.

In the last several years, increasing emphasis has been placed on quality of care indicators and physician staffing models for ICUs. Therefore, a modern surgical ICU in the 21st century has to provide evidence-based care using algorithms, clinical management guidelines (CMGs), and checklists, use cutting-edge technology for physiologic monitoring, and has to have a robust continuous quality improvement process to constantly evaluate its outcomes and to identify opportunities for improvement.

This chapter focuses on elements of critical care essential to the management of the acutely injured patient, reviews some recent advancements in the monitoring of the critically ill patient, and lists some of the common complications and pitfalls observed in the ICU.


Given the wide variety of clinical expertise and patient populations, several patterns of ICU physician organization have developed. The first is the “closed” unit that relies almost exclusively on a critical care team (or attending intensivist) for primary patient management. Under this scheme, comprehensive management is assumed by the ICU team along with responsibility for all orders and procedures, with other services providing care as consultants on an as-needed basis. Most medical ICUs are staffed in this manner along with some surgical ICUs where the ICU team is directed by another surgeon.

In an alternative model, the “open” unit, there may or may not be a designated ICU director, a separate ICU team, or even an intensivist immediately available to the ICU. Under this system, individual physicians manage and direct intensive care for their respective patients, depending on their institutional privileges, with or without house staff. Consultative involvement of a board-certified intensivist is at the discretion of each primary attending physician, and is neither required nor necessarily expected.

Many larger surgical ICUs have a “semi-open” or transitional unit plan of practice whereby the ICU is staffed, 24 hours a day, 7 days a week, with dedicated on-site ICU physicians. On-call physicians are often attending-supervised house staff in larger teaching centers, and responsibility for care is shared between the ICU team and the primary specialty service. With the 24/7 in-unit staffing in the semi-open model, critical care team involvement in each patient is typically either mandatory or expected. There are often specific areas of designated critical care autonomy, such as the management of mechanical ventilators, invasive hemodynamic monitoring, pain management, and conscious sedation. In these units, the ultimate responsibility for the patient remains with the primary team, but patient care is a collaborative effort. This semi-open model combines the advantages of maintaining a separate in-house critical care team 24 hours per day while maintaining primary surgical service responsibility for overall patient management. This arrangement is consistent with Accreditation Council for Graduate Medical Education (ACGME) program requirements for general surgery training programs as well as guidelines for the optimal care of the injured patient suggested by the American College of Surgeons Committee on Trauma for Level 1 Trauma Centers.1 There is now a growing body of work examining the relationship between ICU staffing models and patient outcome. For these purposes, rather than trying to compare various models of care, a distinction has been made between “high-intensity” and “low-intensity” physician staffing models. Loosely translated, a “high-intensity” model involves 24/7 dedicated ICU physician staffing and mandatory ICU team involvement with patient management. This includes all closed units and most semi-open units (as previously defined). The remaining “open” units typically involve “low-intensity” ICU physician staffing. The principal hypothesis is that dedicated, higher-intensity intensivist staffing for ICU patients will ultimately improve outcome from a variety of conditions and illnesses.

In a meta-analysis of 26 pooled studies, Pronovost et al. found a relative risk of 0.71 (95% CI = 0.62–0.82) for hospital mortality and 0.61 (95% CI = 0.5–0.75) for ICU mortality associated with high-intensity ICU staffing for adults and children.2 This validated previous work done by the same author.3,4 Similar results were found by Nathens et al. who specifically examined the effect of high-intensity staffing on outcome from major trauma.5 Utilizing prospective cohort data from 68 trauma centers, the authors reported a relative risk reduction of 0.78 (95% CI = 0.58–1.04) for ICUs whose patients were either managed (closed unit) or comanaged (semi-open) by board-certified intensivists. The effect of dedicated intensivist involvement seems to extend also to neurology and neurosurgical patients, with reports demonstrating improved overall mortality and length of stay.6,7

These and other observations have led to attempts to improve patient safety in the ICU. In 2000, a group of Fortune 500 companies and large public and private health care purchasers formed the Leapfrog Group for the purpose of identifying and creating incentives for sustained patient safety measures in acute care hospitals. This group has identified standards for physician staffing the ICU that involves the presence of experienced intensivists providing daytime care and response requirements for off-hours care.8 The intent of the Leapfrog initiative is to provide market incentives designed to stimulate preferential use of institutions adhering to these types of guidelines. However, the current shortage of critical care specialists makes it impossible to maintain care in “closed” units.


Quality assurance and performance improvement in a surgical CCU are complex processes requiring the ongoing identification of outcome measures or performance indicators, data collection and analysis, and the development of action plans to correct deficiencies and subsequent monitoring of the performance (outcome) measures. The underlying goal of delivering high-quality care also depends, to a significant degree, on specialization (critical care specialists), provider education and training, and good communication and collaborative interaction between specialty and ancillary services. Existing critical care quality assurance programs have used a variety of clinical indicators or “filters” as a measure of the quality of care. These indicators may reflect process measures (e.g., percentage of eligible patients receiving deep venous thrombosis prophylaxis in a timely manner), and outcome measures (complicaton rate [%] for central venous line sepsis). Illness severity indices have been developed for the purpose of predicting outcome of critically ill patients and are being increasingly used as benchmarking tools, allowing participating units, through the use of statistical comparisons, to compare their outcomes with those predicted by the various indices. These performance measurement systems will become increasingly important in allowing managed care organizations to assess program performance. In addition to improved instruments for assessing critical care outcome, the development and implementation of clinical protocols and CMGs9,10 directed at reducing undesirable treatment variability will be linked to CCU performance improvement. Protocols and guidelines, once developed and implemented, can later be analyzed in terms of their clinical efficacy and cost-effectiveness and can be subsequently modified and further developed as part of the overall programs in performance improvement and cost control.

The effectiveness of protocols and CMGs has been demonstrated for a variety of problems and conditions including ventilator weaning, pneumonia, nutrition, and sedation.1117 The difficulties that most institutions experience, however, relate more to the implementation of CMGs rather than to their development. Current methods of improving implementation of and compliance with CMGs include ongoing education, standing (preprinted) order forms, the assignment of some management responsibilities to specialized teams (e.g., nutrition, respiratory therapy), and the use of advance practice staff (nurse practitioners, physician’s assistants).

More recently, several fields in medicine, motivated by an increased awareness about errors in medical care delivery and mandatory reporting of quality measures, have incorporated lessons learned from the aviation industry to clinical practice in an attempt to improve quality and patient safety. A checklist is just one of the tools used in the aviation industry that have been tested in many ICUs to potentially improve safety, quality, and consistency as part of a continuous quality process.18,19 A detailed description of how to develop and implement a quality improvement program in the ICU is beyond the scope of this chapter but important information can be found in the excellent reviews by Curtis et al.,15 and McMillan and Hyzy.20

Inspired by an article by Vincent21 where he described a checklist using the mnemonic FASTHUG (feeding, analgesia, sedation, thromboembolic prevention, head of the bed elevation, stress ulcer prophylaxis, and glucose control), we implemented a modified checklist (mFASTrHUGS) by adding mouth care (m), restraints (r), and skin care (s) (Table 55-1). We use the checklist as a template for daily multidisciplinary rounds and the nursing staff use it at the end of their shift for turnover. Each item on the checklist is reviewed during rounds assessing its implementation and need. If any item has not been implemented for a particular patient, then the team has to identify a contraindication (e.g., pharmacologic DVT prophylaxis during the first day following a significant traumatic brain injury), otherwise the measure has to be implemented. mFASTrHUGS is a package implemented to all patients in our surgical ICU. Other variations of the original mnemonic have been reported recently.22

TABLE 55-1 “mFASTrHUGS” for SICU TEAM Rounding Tool



One of the characteristics of the ICU is the utilization of intensive physiologic monitoring. In past years there were a limited number of types of ICU monitoring available. There has been a dramatic increase in recent years in the ability to monitor a wider variety of physiologic parameters. Today significant issues for trauma surgeons deciding which monitoring to use include lack of familiarity and required training, maintenance, acquisition costs, safety, and efficacy.

Patients admitted to the ICU will have an anticipated and possibly optimal clinical “trajectory” that will be the expectation of the experienced clinician. A considerable effort is expended by the critical care physician in determining if the patient is following the anticipated trajectory, and in anticipating and detecting complications. Monitoring is used to determine that the patient has stable organ function, and detect anomalies that may indicate organ dysfunction and incipient or actual complications. Monitors can be useful in determining if a patient trajectory is deviating from the anticipated or optimal course; however, any monitor’s output requires interpretation and integration with the clinical picture and clinical experience. Monitoring itself is associated with serious and lethal complications, including complications associated with invasive monitoring, and errors in interpretation of results that can lead to error in clinical decision making. The questions that must be addressed with any form of monitoring include “when?” (indication and timing), “what?” or “how much?” (which specific monitoring tools or techniques), and “why?” (an analysis of the associated risks and benefits).


Hemodynamic monitoring is directed at assessing the results of resuscitation and as a guide to reestablishing and maintaining tissue and organ perfusion (see Chapter 56). The restoration of normal arterial blood pressure, central venous pressure, pulmonary arterial (PA) pressure, and cardiac output provides some reassurance that there is adequate organ perfusion. This reduces the likelihood of death and serious morbidity due to complications of hemorrhagic shock, including postinjury inflammatory response syndrome and multiple organ failure (see Chapter 61). However, the phenomenon of regional circulation and inadequate perfusion of some organs, particularly the gut, in the presence of normal arterial pressures remains a concern. Direct monitoring of tissue perfusion, particularly in potentially comprised capillary beds such as the intestine, is more difficult and not commonly performed. New types of monitors are becoming available that have some promise to determine perfusion at the tissue level in organs of interest.

Image Arterial Pressure Monitoring

Noninvasive blood pressure monitoring is available throughout the Medical Center by the use of the automated blood pressure cuff. These devices can be set to do repetitive blood pressure measurements as frequently as every minute. However, these are usually considered insufficient in patients who are having significant hypotension due to the lack of sensitivity of the cuff at low blood pressures as well as the intermittent nature of the readings. Insertion of an arterial intraluminal catheter or “arterial line” allows instantaneous measurement and display of arterial blood pressure, even at very low blood pressures. Arterial lines are the preferred method of arterial blood pressure management in patients receiving continuous drips of vasoactive drugs as they allow better drug titration and have a wide variety of indications (Table 55-2). They also allow for the ready sampling of arterial blood for laboratory testing including arterial blood gas and lactate. Arterial lines are typically placed in the radial artery, although they can also be placed in the femoral artery or brachial artery. Infection of arterial lines is less common than that of central venous lines; however, sterile technique should still be utilized to avoid the risk of infectious complications. Thrombotic complications can occur at any arterial line site; these may be as common as 30–40% and may lead to tissue loss, including loss of digits. This is particularly common in patients who had prolonged hypotension or sepsis or who have been treated with pressors. Although Allen tests are frequently performed before the insertion of a radial arterial line, these do not exclude the possibility of this complication.

TABLE 55-2 Indications (Suggested) for Hemodynamic Monitoring



Minimally invasive continuous arterial pulse waveform analysis devices are available that look at variation in area under the arterial pressure waveform as a surrogate for stroke volume variation with respiration. These devices can provide an estimate of cardiac output without need of a PA catheter, although they are subject to error.

Image Central Venous Pressure Monitoring

Monitoring central venous pressure is most useful in patients at high risk of overresuscitation or underresuscitation (see Chapter 56). These include patients with limited cardiac reserve, including elderly patients and patients with cardiac disease. Patients with traumatic brain injury are another group in whom central venous pressure monitoring is frequently performed to avoid hypotension associated with underresuscitation or overresuscitation resulting in increased cerebral edema. Patients with pulmonary contusions can suffer from overresuscitation, which can lead to increased lung water and worsening respiratory status. Interpretation of the central venous pressure in a ventilated patient should optimally be done while the patient has been removed from positive end-expiratory pressure (PEEP). Central venous pressure can also be approximated by subtracting the level of PEEP in excess of 5 mm Hg from the CVP measurement. The central venous catheter can also be used to obtain a central venous blood sample; this can be used to obtain a mixed venous blood gas measurement. Since the mixed venous oxygen saturation (MVO2) is usually slightly lower in the SVC as compared with that in the IVC due to the relatively low oxygen consumption (VO2) of the kidneys, the blood oxygen saturation obtained from a central venous catheter is usually slightly higher than that obtained from a PA catheter. MVO2 is sometimes used as a surrogate for global perfusion; falling MVO2 is frequently associated with hypovolemic shock and worsening tissue perfusion.23 A rise in MVO2 above baseline may represent decreased VO2 such as in sepsis or poisoning. Central venous catheters that have an oximetric tip can measure the saturation of blood at the tip of the central venous catheter or ScvO2. This allows another surrogate measure of global perfusion on a continuous basis. Combination of the oximetric CVP catheter data and the stroke volume variation from arterial line catheter waveform analysis data in a commercially available monitor can allow continuous estimates of cardiac output, SVR, and stroke volume without the use of a PA catheter.24 Initial validation studies of one system (FloTrac/Vigileo) were disappointing, but better results may be obtained with improved software.25

Image Pulmonary Artery Catheters

The introduction of the Swan Ganz PA catheter in 1970 was followed by wide adoption in critical care units. However, in the mid-1990s there were a number of reports questioning their efficacy and safety. A study by Connors et al. concluded that use of PA catheters increased mortality and led to increased utilization of resources.26 A more recent meta-analysis has not demonstrated increased major adverse sequelae associated with the PA catheter27; however, its utilization in ICUs has decreased significantly. ICUs typically restrict use of the PA catheter to patients who have known or possible myocardial dysfunction, or in those patients who are thought to be sensitive to changes in preload, contractility, or afterload.28 In 2000 an NHLBI workshop report on PA catheters and clinical outcomes produced a consensus statement that indicated that a clinician’s ability to obtain and interpret PA catheter data and intervene appropriately is an important determinant of outcome.29 An assumption is that if the PA catheter use changes therapy and is used to correct physiologic deficits, mortality and morbidity can be reduced. In surgical patients it appears that monitoring data obtained through the PA catheter results in therapy changes in 30–60% of cases. Of these, estimated at 25–30%, are major therapeutic changes. Since there is less experience available in interpretation of PA catheter data for trainees in critical care, courses and online training are available that allow interpretation of simulated data to improve understanding of PA catheter data. A useful tool that can help validate the data obtained with the PA catheter at insertion and at intervals thereafter is the use of echocardiography.

Image Echocardiography

Echocardiography has been available in the ICU for years, usually as a service performed by the hospital’s cardiology service. The formal transthoracic echocardiogram (TTE) can be invaluable in determining the function and structure of the heart as well as the health of its components including valves and chambers. Patients in the operating room undergoing major procedures such as coronary artery bypass grafting or liver transplantation and patients with major trauma procedures are frequently monitored using transesophageal echocardiography (TEE). TEE has been used in the ICU to provide excellent visualization of the heart and its functions as well as detection of such injuries as blunt aortic rupture (see Chapter 26).30,31 However, frequent use of TEE in the ICU is usually limited due to the availability of the devices, cleaning requirements, and the skill required by the operator. Repeated measurements with the TEE are usually impractical for these reasons. However, there are now FDA-approved TEE devices using a disposable probe that can be left in the esophagus and stomach for up to 72 hours that allow for repeated measurements. This can permit observation of ventricular function and ejection fraction over time, and may have a safety advantage over PA catheters. The increasing utilization of ultrasonography by critical care clinicians now means many North American ICUs have clinician-performed ultrasound available. This has led to a number of studies and protocols utilizing a limited TTE by noncardiologists in trauma and critical care patients who may be too hemodynamically unstable to wait for formal TTE. In a patient who is hypotensive with an unclear etiology, visualization of ventricular size and function can be very useful in determining whether shock has a hypovolemic or cardiogenic origin. Limited TTE can also detect such conditions as pneumothorax and pleural effusions. However, in about 50% of trauma patients it is difficult or impossible to perform a complete TTE exam due to issues such as chest trauma, subcutaneous air, obesity, and dressings. Despite this, clinicians often attempt limited TTE in patients with hypotension or during cardiac arrest, due the utility of visualizing cardiac motion in successful examinations. Simplified protocols such as the “focus assessed transthoracic echo” (FATE) or the “bedside echocardiographic assessment in trauma/critical care (BEAT)” can be utilized by trauma and critical care physician to rapidly determine the cause of hemodynamic instability.

Image Other Monitoring Modalities

Gastric Tonometry

Gastric tonometry is one of the few modalities utilizing a direct measurement of gut perfusion as end point in resuscitation. The rationale for measuring the intramucosal pH (pHi) is based on the observation that splanchnic perfusion can be impaired in the setting of adequate blood pressure and cardiac output, leading to impaired GI mucosal barrier function that may induce multiple organ dysfunction (see Chapter 61). The catheter used resembles an NG tube with a saline-filled silicone balloon at the tip. When the balloon is in contact with the stomach wall, oxygen and carbon dioxide in the saline-filled balloon will equilibrate with oxygen and carbon dioxide in the stomach mucosa. Measurement of the pCO2 of the saline in the balloon allows determination of pHi via the Henderson–Hasselbalch equation. Gastric tonometry has not demonstrated a clear advantage compared with conventional resuscitation end points and its use has not been widely adopted.32,33


Transthoracic impedance, also known as bioimpedance cardiography, as a method to continuously measure cardiac output has been evaluated for decades. A high-frequency, low-alternating electrical current is applied to the thorax, and changes in bioimpedance to this current are related to cardiac events and blood flow in the thorax. The technique has been refined by looking at differing algorithms and different parts of the signal such as electrical velocimetry, phase, or frequency. However, results of studies have led to conflicting and inconclusive results, which may lead to inappropriate clinical interventions. As a result, transthoracic impedance has not yet been widely adopted as a method to measure cardiac output in noninvestigational settings. A new device is available that utilizes impedance electrodes on the outside surface of an endotracheal tube.34 This allows impedance measurements in the trachea immediately adjacent to the aorta. This has had promising results in initial studies in cardiac surgery patients.


Near-infrared spectroscopy (NIRS) is a noninvasive technique in which a probe utilizing near-infrared light is applied to the thenar eminence. This allows detection of the tissue oxygen saturation (StO2) of the muscle of the thenar eminence. Decreased StO2 correlates with increased mortality and organ failure in patients with hemorrhagic shock or sepsis.35

Image Hemoglobin Therapies


Blood transfusions independent of shock or injury/disease severity are associated with worse outcomes. Increased infection, multiple organ dysfunction, and mortality are correlated with the amount of blood transfused and pH of transfused blood. Blood transfusions are common in the ICU with 40–45% of critically ill patients receiving an average of 5 U of packed red blood cells. Lowering target hemoglobin from 10–12 to 7–9 g/L was associated with improved outcomes in ICU patients in the Transfusion Requirements in Critical Care (TRiCC) study.36 In a medical population, the trial by Rivers et al. found a 30% relative survival advantage for a bundle of interventions, including inotropes and transfusion to maintain a central venous oxygenation of 70%.37 Based on this, the “Surviving Sepsis” guidelines currently recommend to “transfuse packed red cells if necessary to hematocrit of >30%.”38 However, the TRiCC guidelines have been more recently reaffirmed by a guideline from the Eastern Association for the Surgery of Trauma and Society of Critical Care Medicine (EAST-SCCM), referring to multiple studies that failed to find physiologic benefit from red cell transfusion in septic patients.39 The EAST-SCCM guidelines do support aggressive, empirical transfusion in trauma and other surgical patients with uncontrolled hemorrhage. A general approach to transfusion is that patients with hemorrhage need transfusion, while those who are not bleeding infrequently require immediate transfusion. In the trauma bay, operating room, and ICU, red blood cells are often the most available and effective initial resuscitation fluid, and prudence suggests a liberal transfusion strategy until anatomic hemorrhage control is achieved and laboratory values stabilize. Coagulopathy associated with trauma and massive transfusion may require the early ministration of plasma at a 1:1 ratio with packed red cells (see Chapter 13).

Erythropoiesis-Stimulating Agents (ESAs)

The anemia of critical illness resembles that of chronic inflammatory disease; low circulating erythropoietin levels are found in critical illness and are thought to be one of the causative factors. Although epoetin alfa and darbepoetin alfa, ESAs, have been widely accepted for the indication of anemia due to chronic kidney disease, these agents are also attractive for anemia of acute critical illness as they may allow more rapid restoration of hemoglobin levels without transfusion. However, the optimal hemoglobin targets and dosing have not been established and clinical trials have used target hemoglobins higher than those suggested by TRiCC. The Normal Hematocrit Study used larger doses of epoetin alfa to increase hematocrit to 42 ± 3% or to continue epoetin alfa therapy to maintain a hematocrit value of 30 ± 3%; the trial was stopped early due to 1.3 times increased mortality and nonfatal myocardial infarction in the 42 ± 3% hematocrit group.40 Increased adverse outcomes have been seen in subsequent ESA trials, the Correction of Hemoglobin and Outcomes in Renal Insufficiency (CHOIR) trial, using a target of 13.5 g/dL, and the Trial to Reduce Cardiovascular Events with Aranesp Therapy (TREAT) at a 13.0 g/dL target.41,42 CHOIR was halted at an interim analysis when mortality reached 17.5% of patients in the high-hemoglobin group and in 13.5% of patients in the low-hemoglobin group (hazard ratio, 1.34; 95% CI, 1.03–1.74; P = .03). TREAT showed no increase in mortality but there were significant increased risks of fatal and nonfatal stroke and thromboembolic complications. Studies of ESAs in critical illness have sometimes shown a reduction in need for transfusions. The EPO-1 pilot study (n = 160) demonstrated a reduction in red blood cell transfusion and a rise in hemoglobin with epoetin alfa treatment using a dose of 300 U/kg per day for 5 days and then every other day.43 The follow-up study (EPO-2, 1,302 patients), using a significantly lower dose of epoetin alfa (40,000 U per week), confirmed the transfusion findings.44 A third randomized study (EPO-3, 1,460 patients) was performed using a weekly epoetin alfa dose of 40,000 U in which the primary outcome was again transfusion reduction.45 No transfusion reduction was identified with epoetin alfa treatment in this third trial, although hemoglobin concentration did rise. A subsequent subset analysis of the trauma patients in the EPO-2 and EPO-3 studies suggested there was approximately a 50% reduction in 29-day mortality.46 There is no clear consensus on ESA dosing in trauma patients. A multicenter study in anemic (Hb <12 g/dL) ICU patients (n = 60) examined six different dosing regimens, intravenous or subcutaneous for epoetin alfa administered for a duration of 15 days.47 Only 30 patients were evaluable. Erythropoietin serum concentrations were 10–45 times greater for intravenous compared with subcutaneous dosing. Mean absolute reticulocyte count peaked on day 11 or 15, and absolute reticulocyte count was greater for subcutaneous dosing, but the pharmacokinetics did not predict pharmacodynamic response in these anemic critically ill patients. More frequent administration of lower doses of epoetin alfa was not superior to less frequent administration of larger doses, but the total cumulative doses of epoetin alfa were similar in all groups (120,000–170,000 IU). This would suggest that subcutaneous, weekly dosing is as effective as other regimens, although the optimal dosing regimen and route of administration of ESAs in critically ill patients for the treatment of anemia are yet to be determined. Optimal response to ESAs may not be achieved in the presence of relative iron deficiency, which is common in ICU patients. Concomitant iron administration has been recommended with ESAs. Overall the benefits of ESAs appear to be related to a reduced need for transfusion rather than an improvement in other outcomes. A retrospective study of a revised trauma practice guideline for anemia after deletion of ESAs showed a significant cost reduction and no increase in blood product utilization.48


One of the principal indications for admission to the ICU for trauma patients is in need for frequent neurologic assessments in those with known or suspected traumatic brain injury (see Chapter 19). There are two goals for monitoring traumatic brain injury: one, the avoidance of secondary brain injury by avoidance of hypoxia and hypotension and maintenance of cerebral perfusion; and second, the detection of increased intracranial pressure (ICP) due to posttraumatic phenomenon such as cerebral edema or expanding intracranial hematomas. The neurologic assessments are composed of the Glasgow Coma Scale for assessment of the verbal, motor, and ocular responses in association with pupillary exam. Patients with coma and patients requiring sedation and ventilation require frequent assessment of brainstem reflexes. Reassessment is done frequently as even subtle changes may herald increases in ICP or cerebral ischemia.

Image Intracranial Pressure Monitoring

ICP monitoring with a ventriculostomy catheter or through a subdural bolt is a common practice in most major centers for patients who have neurologic exam that is unavailable or unreliable following traumatic brain injury. The Brain Trauma Foundation Guidelines recommend early and aggressive monitoring of ICP and the calculated cerebral perfusion pressure (CPP), which is the difference between the mean arterial pressure (MAP) and ICP.49Monitoring of the ICP and CPP can be used to guide therapy such as vasopressors or opening the ventriculostomy catheter to drainage and to provide surveillance for increasing cerebral edema or intracranial hemorrhage. ICP monitoring in a large population of traumatic brain injury patients was found to improve survival versus those who did not receive ICP monitoring despite a higher severity of injury in the ICP-monitored group, 51% versus 35%.50Investigational noninvasive monitoring systems for ICP include transocular ultrasound monitoring of optic nerve sheath diameter (ONSD).51,52 The optic nerve sheath is an extension of the intracranial dura matter and has been shown to become distended with elevated ICP. Continuous transcranial Doppler ultrasound monitoring of cerebral vessels is another investigational approach for ICU monitoring of traumatic brain injury.


Current guidelines indicate the need for nutritional assessment and monitoring for patients admitted to the ICU (see Chapter 60). Enteral nutrition, which may help preserve gut mucosal barrier function, is the preferred approach. Nutritional monitoring should begin with an assessment of the degree of preinjury malnutrition as well as current requirements. Weekly assessments in the ICU should include a calculation of nitrogen balance; this calculation takes a difference between excreted nitrogen from a 24-hour urine urea nitrogen and the ingested nitrogen. An overall goal is to have the patient had a 2–4 g positive nitrogen balance if possible. Indirect calorimetry (metabolic cart) provides an assessment of caloric requirements and metabolic rate. It may be significantly more accurate than predictions based on formulae such as the Harris–Benedict equation.53 This technique uses the rate at which gases are produced in the intubated patient to estimate caloric expenditure. The device calculates VO2 and carbon dioxide production (VCO2). Standard metabolic carts requiring an integral mechanical ventilator were intended for intermittent use; however, newer devices can be used with standard ventilators with an adapter that can then continuously sample end-tidal oxygen and carbon dioxide. CO2 and pCO2 can be used to then determine the respiratory quotient (RQ) and caloric needs can be estimated. The RQ can also be used to determine the prominent nutritional substrate being used, excess carbohydrates leading to RQs of 1.0 or more, with RQs below 0.8 indicating possible excess lipid utilization.


Hyperglycemia with or without insulin resistance appears to be a common phenomenon in the critically ill patients. Based on literature that indicated outcomes were improved if blood sugars were maintained below 215 mg/dL, van den Berghe et al. hypothesized that hyperglycemia leads to worse outcomes in a critically ill population.54 In their prospective randomized controlled trial, their group demonstrated that tightly controlled blood glucose levels (at or below 110 mg/dL) with intensive insulin therapy improved mortality and reduced complications such as infection rate, multiorgan failure rate, ventilator days, and morbidity. The adverse consequences of hyperglycemia may be associated with even worse mortality in trauma patients. A retrospective analysis of 12 years of data on trauma and on trauma patients showed that hyperglycemia in trauma patients correlated with higher mortality rates than in other surgical patients.55 The intensive insulin regimen initially proposed by van den Berghe et al. has been associated with increased hypoglycemic complications and lack of benefit in subsequent multi-institutional studies such as VISEP, GLUCONTROL, and Normoglycemia in Intensive Care Evaluation Survival Using Glucose Algorithm Regulation (NICE-SUGAR).5658 The NICE-SUGAR trial showed the primary outcome variable of 90-day mortality was actually increased in patients randomly assigned to intensive insulin therapy, as compared with an intermediate target range for blood glucose. Improved glucose control is still a guideline used in most major centers; however, a less intensive insulin protocol than that suggested by van den Berghe et al. may be prudent in avoiding the likelihood of severe hypoglycemia.


The normal response to physiologic stress leads to increased levels of tissue corticosteroids; this is also seen in the patient’s response to critical illness. Failure to recognize and treat adrenal insufficiency has been associated with increased mortality. Normal corticosteroid response can be impaired in a variety of conditions including sepsis, systemic inflammatory response syndrome (SIRS), and traumatic brain injury. The use of steroid therapy in sepsis and critical illness has been a topic of an extensive literature and pendulum has swung between apparent benefit and apparent risk. Identification of patients with acute adrenal insufficiency can be difficult; random cortisol sampling (<15 μg/dL) and corticotropin stimulation tests (15–34 mg/dL) were often used to identify patients suitable for low-dose steroids (100–300 mg per day hydrocortisone). Patients with abnormal stimulation tests (increases <9 mg/dL) were felt to require corticosteroid supplementation. While a 2002 study by Annane et al. suggested benefit to low-dose steroids in sepsis, the 2008 multi-institutional CORTICUS trial did not show any 28-day mortality benefit with 300 mg per day of hydrocortisone in patients with septic shock, either overall or in patients who did not have a response to corticotropin.59,60 However, hydrocortisone did hasten reversal of shock in CORTICUS study patients. Lacking adequately powered positive studies, low-dose steroid therapy should probably be limited to patients with septic shock whose blood pressure is poorly responsive to fluid resuscitation and vasopressor therapy. Corticotropin stimulation probably has no role in determining steroid use in patients with septic shock. Steroids should be discontinued if the patient does not respond to treatment given the potential risks of infection, hyperglycemia, and critical illness polyneuropathy (CIP).


Mortality after trauma in the ICU is often attributed to infection. Infections are thought to contribute to more than 88,000 ICU deaths annually in the United States.61 Infectious complications of major injury (see Chapter 18) may be due to the results of the injury itself (e.g., open fractures), as a result of complications of treatment (e.g., anastomotic leak), or as iatrogenic complications of critical care management (e.g., ventilator-associated pneumonia [VAP], central line infection). Specific monitoring for infections will depend on injury type and severity, types of interventions performed, and the duration of postinjury critical illness. Regulatory authorities have recently mandated surveillance cultures on admission to the ICU for certain health care–associated infections (HAIs) such as nasal swabs for methicillin-resistant Staphylococcus aureus (MRSA), which is required in some states. In a study using multiple regression analysis, the most common variables for infection were found to be central venous catheters, mechanical ventilation, chest tubes, and trauma with open fractures.62 Patients at risk for infectious sequelae should be routinely tested by culture and examined for clinical indications such as fever, leukocytosis, change in physical examination, pyuria, and development of purulent sputum or new infiltrate on chest x-ray. HAIs such as central line–associated bloodstream infection (CLABSI) may have standardized definitions in some jurisdictions and regulatory requirements for surveillance. Other HAIs such as VAP may require adoption of a local or institutional standard definition and therapy in the lack of a national consensus. The decision to start presumptive (empirical) antibiotics should be based on risk factors, the expected sequelae of injury, and prior infections. Presumptive treatment should be started, if indicated, at defined end point for culture-negative patients and applied and reviewed with sensitivity- and spectrum-based adjustment when final culture information is available. Importantly, all critically injured trauma patients with unexplained fevers do not require antibiotics. Persistently febrile, culture-negative patients often present a difficult diagnostic and therapeutic challenge. Fever in these patients may have a noninfectious origin (Table 55-3) or be due to occult infection that has not been considered, successfully cultured, or has no identifiable signs or symptoms. Fungal sepsis should be considered in patients with prolonged ICU stays, multiple prior antibiotic therapies, and immuno-suppression.

TABLE 55-3 Noninfectious Sources of Fever in Critical Illness



The need for mechanical ventilation is the most common indication for admission to the ICU. Familiarity and advanced understanding of mechanical ventilation principles is the lynchpin to managing severely injured patients in the ICU. Ventilator management can change rapidly; therefore, a nucleus of critical care expertise is needed not only to optimize respiratory support but also to interpret acute changes in pulmonary mechanics that can often be a harbinger of systemic physiologic pathology.

Providing a secure airway via endotracheal intubation generally happens early during the initial resuscitation. Even though intubation for airway security can be secondary to ventilatory insufficiency, the two can be mutually exclusive. The general principles for intubation and mechanical ventilation are the following:

1. Secure and establish an airway

2. Decrease the work of breathing (WOB)

3. Improve oxygenation

4. Improve ventilation (CO2 gas exchange) and maintain control of PaCO2 especially in acute brain injury

5. Anticipate worsening respiratory status or airway patency such as the need for large-volume resuscitation, severe neck/thoracic trauma, upper torso/facial burns, and inhalation injury

Since the development of modern ventilatory support in the 1960s, significant technological advancements have been made to improve the ventilators themselves. Using contemporary ventilator technologies, new modes of mechanical ventilation are often introduced, each with an acronym and specific terminology. Despite these advancements, the essential physical principle of mechanical ventilation remains constant, which, simply put, is pushing oxygen-rich air into the lungs by positive pressure and removal of waste CO2 by negative pressure. Understanding this fundamental principle simplifies even the most “advanced” modes of ventilation and therefore allows the surgeon a streamlined method to manage complex ICU patients with clarity and an evidence-based approach.


Rapid assessment of arterial oxygen tension (PaO2) is essential for both evaluating and managing the adequacy of alveolar–arterial oxygen gas exchange. Oxygen is driven from the alveolar airspace into the pulmonary capillaries primarily due to the difference in the O2 diffusion gradient between the respective tissue beds. Measuring PaO2 quickly and reliably from arterial blood gas samples may be taken for granted in modern ICU care. The physical basis for measuring PaO2 is a result from the development of the oxygen electrode and measuring resultant electrical current that is directly proportional to oxygen concentration.63 Calculating the efficiency of pulmonary oxygen exchange can be cumbersome since the equation for the A–a gradient requires alveolar and arterial CO2 concentrations, shunt fraction, water vapor pressure, and body temperature. A more convenient and simple bedside index of oxygen exchange is the PaO2/FiO2 (P/F) ratio that adjusts for a fluctuating FiO2 and helps in defining lung injury (see below). The use of pulse oximetry to determine arterial oxygen saturation (SaO2) has replaced continuous arterial blood gas measurements as a real-time, noninvasive method to assess arterial oxygenation.64There are limitations to SaO2 and its nuances are important in interpreting the significance of an absolute SaO2 percentage. First, because of the kinetics of oxygen–hemoglobin binding, SaO2 and the oxygen dissociation curve is sigmoidal and not linear. Therefore, small changes in SaO2 may reflect a much larger drop in PaO2. Second, under instances of carbon monoxide poisoning or severe circulatory shock, oxygen delivery will be abnormally low despite a near-normal SaO2. Finally, SaO2 is not indicative of ventilation. A rising PaCO2 and subsequent acidosis may not initially affect SaO2 especially if a patient is on high FiO2 settings. For these reasons, it is important to interpret SaO2 in combination with appropriate PaO2 measurements to individually determine a patient’s ability for oxygen exchange.


Adequate pulmonary exchange, exhalation of CO2, and resultant arterial CO2 tension (PCO2) have important physiologic and clinical implications. The immediate detection of CO2 through disposable endotracheal capnography relies on the lower pH of EtCO2-rich air changing the color of pH-sensitive filter paper in the capnograph. As a rule, a disposable capnograph remains purple if EtCO2 is < 0.5%, whereas a yellow color change is equivalent to EtCO2 of >2.0%.65 Normal EtCO2 is >4%; therefore, capnography turns yellow when the endotracheal tube is positioned correctly. The accuracy of this device is very sensitive unless the patient is in circulatory arrest and adequate pulmonary perfusion is compromised. The presence of a large volume of acidic gastric contents may also give a false impression of successful endotracheal intubation even though it is the esophagus that has been intubated. In these circumstances, initial detection of EtCO2 decreases rapidly with subsequent tidal volumes. Clogging of the capnograph with mucous and enteric contents can also give false readings.

The measurement of PCO2, through blood gas analysis, remains the most accurate measure of assessing ventilation. Continuous capnography analyzes EtCO2 and, depending on the alveolar–arterial gradient, can give an indication of the arterial PCO2. After anatomic dead space has been cleared, EtCO2 rises until the end of exhalation. This rise is generally steep but plateaus during the alveolar phase of ventilation. In patients without preexisting pulmonary disease, the normal alveolar–arterial gradient is between 1 and 3 mm Hg. Therefore, using continuous capnography in a ventilated patient estimates the patient’s PCO2especially during the initial resuscitation phase. Utilizing capnography can be helpful particularly in brain injury where hyperventilation may curtail rising ICP. However, EtCO2 measurements are affected by pulmonary dead space fraction and pulmonary perfusion that can be greatly altered in cases of hemorrhagic shock, thoracic trauma, and increased airway resistance. Warner et al. performed a prospective study of 180 freshly intubated trauma patients and correlated EtCO2 measurements with PCO2 from blood gas analysis.66 Using regression analysis, the authors found a direct correlation between EtCO2 and PCO2; however, patients ventilated with an EtCO2 of 35–40 mm Hg were likely to have a PCO2 >40 mm Hg 80% of the time, and a PCO2 >50 mm Hg 30% of the time. In severely injured patients, where an increase in pulmonary shunt may exist, appropriate confirmation of EtCO2 by blood gas sampling is important.

Causes of hypoventilation can be multifactorial. Severe thoracic injury resulting in pulmonary contusion or flail chest can be a factor warranting expedient intubation. Increased dead space from pulmonary contusion, ALI, acute pulmonary embolism, and oversedation can be contributing factors causing hypercapnia especially during the weaning phase of mechanical ventilation.


The mechanically ventilated lung is an active continuum that reflects real-time physiologic complexities of an ICU patient. Assessing specific pulmonary mechanics can significantly aid in diagnosing and treating a patient’s sudden or progressive pulmonary insufficiency. One of the most important parameters to measure is compliance as is the ability to distinguish between static and dynamic compliance. Compliance, the change in volume produced by a change in pressure, is calculated based on measurements taken from the ventilatory circuit itself. In general, a sudden decrease in compliance will mean a concomitant rise in both plateau pressure and peak inspiratory pressures (PIPs) for a given tidal volume as seen in pneumothorax or in abdominal compartment syndrome. However, measuring the individual changes of either static or dynamic compliance may indicate specific pulmonary pathology.

Static Compliance

If the respiratory system is reduced to a simple model of a straw and a balloon, the straw would represent the airway, and its resistance, and the balloon would represent the elastic recoil of the lung and the chest wall. In static compliance, the change of volume produced is measured from the inspiratory hold pressures or plateau pressure using the following formula: Compliancestatic = Volumetidal/Pressureplateau – PEEP. Thereby static compliance measures the pressure resistance of the alveoli and chest wall, not of the airways. Cases of decreased static compliance (normal = 50–100 mL/cm H2O) with normal or elevated PIP generally indicate intra-alveolar pathology such as acute respiratory distress syndrome (ARDS), ALI, and pulmonary edema.

Dynamic Compliance

Dynamic compliance measures compliance by equating the pressure, PIP, needed to overcome the resistance of the airways or the straw using the balloon analogy (Compliancedynamic = Volumetidal/PIP – PEEP). Cases of a sudden decrease in dynamic compliance or a measured difference between static and dynamic compliance are a reflection of increased airway resistance during bronchospasm, mucous plugging, kinked endotracheal tube, or foreign body aspiration.

Pressure–Volume Curves

A well-accepted principle in mechanical ventilation is the importance of increasing alveolar recruitment to improve oxygenation. Assessment of the pressure–volume curve in mechanical ventilation has been utilized to determine the mean pressure needed for the inflection point, or Pflex (Fig. 55-1). This point, corresponding to the transition between the curve’s flat portion and linear portion during inhalation, represents the critical pressure needed to open collapsed alveoli and corresponds to the zone of optimal alveolar recruitment and gas exchange. Initial driving pressure therefore precedes alveolar filling and corresponds to gas flow within the airways. The pressure reading at the inflection point may reflect an optimal PEEP setting. In healthy lungs with normal compliance, oxygenation and ventilation is achieved with modest PEEP (5 cm H2O) and plateau pressures (Pmax) less than 30 cm H2O. In ARDS, ALI, and pulmonary edema the normal pressure–volume curve is shifted (Fig. 55-1) leading to derecruitment of alveolar units and increasing pulmonary shunt. During these instances, increasing driving pressure and increasing PEEP are often employed to maximize gas exchange in the zone of optimal ventilation. However, using pressure–volume curves alone in clinical decision making has been neither shown to be efficacious nor shown to improve outcomes. There are significant inconsistencies in curve interpretation, which leads to significant operator variability. A standardized method of pressure curve interpretation may aid in predicting the development of ALI and ARDS.


FIGURE 55-1 Pressure–volume curves.


Assessing ventilatory capacity in the mechanically ventilated patient can be an important physiologic indicator to monitor clinical improvement and predict successful extubation. Its use is particularly applicable in patients who have been intubated for prolonged periods of time, have spinal cord injury or injury to the thorax such as flail chest or rib fractures, or patients with neuropathic disease (i.e., Guillain–Barré syndrome, myasthenia gravis). A forced vital capacity (FVC) and a negative inspiratory force (NIF) are the most frequent measured indices of ventilator capacity. A normal vital capacity, which consists of both inspiratory capacity and expiratory reserve, is between 65 and 75 cm3/kg depending on ideal body weight and sex. An FVC of 10–15 cm3/kg is an acceptable value to predict successful extubation.67

The NIF, in conjunction with FVC and the rapid shallow breathing index (RSBI; see below), is also an important variable during weaning. Despite the need for a cooperative and awake patient, the NIF estimates inspiratory muscle strength using a one-way valve manometer attached to the airway. The inspiratory force is then calculated based on the negative force generated by the patient during occlusion of the valve. A value of –25 to –30 cm H2O is indicative of adequate inspiratory force to maintain airflow after extubation.67

Despite estimating ventilatory capacity and other respiratory parameters, there exists great variability in practice in when and how parameters are determined. In a survey of nine Los Angeles area ICUs, Soo Hoo and Park found significant variability in the method and frequency of both NIF and FVC measurements. This study only further underscores the notion that current weaning practice is just as much art as science.68


It is well established that high airway pressures during positive pressure ventilation can cause lung injury secondary to overdistension and alveoli rupture (see Chapter 57).69 Ventilator-induced lung injury can be delineated into mechanical and inflammatory categories since these processes differ in their pathophysiology and treatment.

Repetitive distention and collapse of lung tissue from mechanical ventilation is the basic mechanical pathophysiology that is thought to cause ventilator-induced lung injury from inflammatory changes. Pathologic features of ventilator-induced lung injury include edema and an upregulation of inflammatory cytokines and cells, the combination of which can further damage pulmonary tissue and progressively lead to pneumonia and sepsis. Webb and Tierney conducted the first comprehensive study in rats demonstrating that 1 hour of positive pressure ventilation, with peak pressure at 30 cm H2O or higher, caused pulmonary edema.70 Several other experiments also show pulmonary edema with a progressive increase in inflammatory cytokines, such as IL-1, TNF-α, and IL-8, following mechanical ventilation, although larger animals and humans require longer periods of intubation for these inflammatory changes to occur.71 Given these initial experimental observations, several human studies were initiated with the intent to demonstrate that decreasing alveolar stretch by lower tidal volumes will prevent ongoing lung injury and improve ICU outcomes. A detailed description of current ventilating modalities used in the management of ARDS is beyond the scope of this chapter (see Chapter 57).


Spontaneous breathing utilizes negative intrathoracic pressure to fill alveoli (Fig. 55-2). In contrast, mechanical ventilation utilizes the principles of external insufflation generating positive pressure to fill alveoli and negative pressure for exhalation. Positive pressure mechanical ventilation can improve gas exchange by recruiting atelectatic alveoli, increasing functional residual capacity, and improving areas of ventilation/perfusion mismatch, and thereby decreases pulmonary shunt fraction. Negative effects of mechanical ventilation vary according to ventilatory mode; however, adverse effects common to all positive pressure modes include barotrauma, ventilator-induced lung injury, and impairment of cardiac output from decreased venous return. As previously mentioned, the physical simplicity of mechanical ventilation is often lost within the technical jargon of individual ventilator modes. It is important to assess the physical mechanics behind each mode of ventilation since correct management of ventilated patients will decrease total ventilation days and improve patient outcome. In choosing which mode of ventilation is best suited for a particular patient, it is helpful to evaluate the patient’s current oxygenation, ventilation, pulmonary compliance, muscular strength, and mental status.


FIGURE 55-2 Spontaneous breathing mode of ventilation.


Both continuous mandatory ventilation (CMV) and assist-control (AC) mode ventilation are preset volume-cycle modes that deliver a fixed tidal volume at a specific respiratory frequency. The individual names of these modes are commonly referenced interchangeably; however, CMV delivers a set tidal volume exclusive of a patient’s ventilatory effort, whereas AC, in addition to a set frequency, synchronizes a set tidal volume with each patient-initiated breath. During a patient-initiated breath in CMV mode, the ventilator will not provide support or, conversely, a full tidal volume may be delivered if the patient is spontaneously exhaling. Consequently, an awake and breathing patient will have episodes of respiratory dyssynchrony causing significant discomfort, breath stacking, and barotrauma.72,73 Given these reasons, CMV mode should rarely be used.

In AC, the ventilator recognizes a patient-triggered breath (either flow or pressure mediated) and immediately delivers a synchronous, identical preset tidal volume accordingly (Fig. 55-3). To assure appropriate synchrony, the ventilator must sense the patient’s spontaneous effort to trigger tidal volume delivery. In older ventilators, a “pressure trigger” mechanism relied on pressure sensors that were activated by negative inspiratory pressure generated by a patient breath. Most modern ventilators are now “flow triggered” where a continuous flow of gas passes around the breathing system. Patient inspiration deflects this flow, and triggers the ventilator to deliver a full tidal volume. AC modes have a background set rate to assure ventilation in an apneic, paralyzed, or heavily sedated patient. If under these circumstances patients do not initiate breathing, AC becomes indistinguishable from CMV since there is complete absence of patient-triggered breathing. In an awake or lightly sedated patient, full synchronous respiratory support will be achieved with each initiated breath, thereby decreasing WOB and assuring adequate ventilation. It is important that there is consistent evaluation of patients on AC. Awake or agitated patients may trigger several full tidal volume breaths causing high minute ventilation and respiratory alkalosis. In a prospective analysis of intubated patients, Sternberg and Sahebjami showed AC to have significantly higher peak airway pressures compared with the same patient on a pressure support mode, although during AC there was a decreased incidence of hypoxia and tachypnea.74 It is important to frequently assess a patient’s ability to transition from AC to a pressure support mode with the goal of eventually weaning from mechanical ventilation.


FIGURE 55-3 Assist-control mode ventilation.


Excessive airway pressures (>50 cm H2O) can increase the incidence of barotrauma and lung injury especially in states of decreased pulmonary compliance. Pressure-regulated volume control (PRVC) ventilation is a relatively new ventilatory mode that modulates the inspiratory flow pattern during a volume-controlled breath, thus attenuating and limiting PIP. This flow pattern, coined “decelerating inspiratory flow,” is in contrast to pure AC/CMV modes where there is constant inspiratory flow resulting in a continuous rise in airway pressure. Despite the theoretical advantage, few studies have shown any clinical benefit with PRVC in the adult population. In a small study of ARDS patients, Guldager et al. showed that PRVC mode had mean PIPs of 20 cm H2O compared with 24 cm H2O under volume control; however, there was no difference in mortality or days of ventilation between each group.75


SIMV allows for spontaneous ventilation with the addition of intermittent volume-controlled breaths determined by a preset rate and tidal volume (Fig. 55-4). The “synchronous” component allows the ventilator to detect patient-initiated negative inspiratory flow and accordingly times delivery of a preset tidal volume, thus preventing breath stacking and patient discomfort. In contrast to AC ventilation, where each patient-initiated breath receives a full tidal volume, SIMV mode allows for respiratory work between the preset tidal volumes. The degree of ventilatory support can be increased or decreased by adjusting the set rate.


FIGURE 55-4 Synchronous intermittent mandatory ventilation.

This mode was originally developed to wean patients from mechanical ventilation and continues to be a widely used method of stepwise weaning despite evidence that this method has decreased efficacy in comparison to pressure support trials (see Section “Weaning from Mechanical Ventilation”). However, SIMV can be useful to reduce minute ventilation in patients who cannot be transitioned to full pressure support mode yet are significantly overbreathing during AC ventilation causing respiratory alkalosis.


PSV, originally termed pressure assist, provides a baseline level of inspiratory airway pressure and decreases the WOB by augmenting spontaneous respiration (Fig. 55-5).76 The patient, therefore, is aided in overcoming the resistance of the ventilatory circuit and has complete control over the rate, and tidal volume. Each PSV breath is supported by a specific flow limited by a preset pressure that is triggered by patient inspiration. Termination of pressure support occurs when the patient’s inspiratory tidal volume nears the preset pressure limit (usually 25% of preset peak) allowing spontaneous expiration. Thus, each tidal volume varies according to the patient-derived inspiratory effort. Because this mode requires spontaneous breathing, PSV can only be utilized in lightly sedated or awake patients without paralytic therapy or neuromuscular disease. Although PSV can be used as a sole ventilation mode, perhaps the single most important use of PSV is in the process of weaning (see Section “Weaning from Mechanical Ventilation”). Several trials, including the Esteban Spanish Collaborative study, have shown weaning protocols incorporating lower-pressure PSV mode are advantageous to IMV or high-pressure PSV weaning.77 Intermittent use of PSV with higher set pressures may improve muscle conditioning in chronically ventilated patients. However, data supporting this approach to hasten mechanical ventilation weaning have not been conclusive.


FIGURE 55-5 Pressure support ventilation.


As the correlation between higher plateau pressures, excessive peak airway pressures, and ALI/ARDS is increasingly recognized, the limitations of volume-cycled ventilation become apparent in the setting of poor pulmonary compliance. To achieve a specific preset tidal volume, progressively higher airway pressure must be delivered as lung compliance worsens increasing the likelihood of barotrauma and alveolar damage. In contrast to volume control, TCPV delivers a breath at a fixed flow rate dictated by a preset pressure (driving pressure). Regardless of pulmonary or chest wall elasticity, the PIP is fixed and will not exceed the set driving pressure; however, the delivered tidal volume will vary as a function of changing pulmonary compliance. TCPV is useful in patients with progressive pulmonary insufficiency or ARDS, while also meeting acute respiratory distress syndrome network (ARDSNet) criteria so long as the set driving pressure and PIPs are below 30 cm H2O. Importantly, TCPV is not well tolerated by awake patients and normally requires deep sedation. Under deep sedation, inverse ratio ventilation (IRV) can also be used with the goal of increasing oxygenation. In conventional AC/CMV mode the inspiratory to expiratory ratio is approximately 1:2, thereby minimizing airway pressures and allowing adequate ventilation. Typically IRV allows I:E ratios 1.5:1 to 2:1 achieved by inspiratory flow rates and decelerating flow patterns dictated by the set diving pressure. In principle, IRV allows maximal recruitment of alveoli by expanding marginal airspaces. Beyond the theoretical short-term advantage, few existing studies show any benefit in mortality or long-term ventilatory improvement versus conventional ventilation. When using IRV, it also is important to be cognizant of rising intrinsic PEEP that can lead to barotrauma or a progressive decrease in cardiovascular return effecting cardiac output.


APRV, sometimes referred to as bilevel ventilation, is essentially continuous positive airway pressure ventilation with the exception that the pressure setting is generally higher (Phigh) for a longer duration (Thigh) than typical CPAP and there is a short “release” of high pressure (Tlow and Plow) allowing for ventilation and CO2 exhalation. The long duration of Phigh (as long as 6 seconds) is thought to improve alveoli recruitment and gas exchange, similar in concept to TCPV, and decrease overall minute ventilation.78 Perhaps APRV’s greatest benefit is due to the presence of a floating release valve that allows for patients to breathe spontaneously during the prolonged Phigh phase. This, in principle, provides greater patient comfort and decreases the need for sedatives. Few studies have been able to show sufficient evidence supporting this theory. Initially, APRV was heralded as a rescue ventilation mode for severe ARDS/ALI; however, subsequent studies have not shown APRV to be superior to standard mechanical ventilation protocols. However, if the Phigh setting is appropriate (<30 cm H2O in ARDS) and oxygenation/ventilation goals are met, an APRV mode can still relatively adhere to guidelines established by ARDSNet and may benefit individual patients.


The largest experience with HFOV has been in neonatal patients who present with severe hypoplastic lung disease and a variety of other congenital or acquired pulmonary pathology. The principle behind HFOV is to optimally utilize open airways and literally “oscillate” air within open alveoli to provide gas exchange. This process is theorized to prevent lung injury by reducing the repetitive opening and closing of airways and alveoli. The mechanics of the HFOV ventilator utilizes a piston assembly structure and an electronic control magnetic motor that provides the rapid frequency for oscillation. The frequency of oscillation (1 Hz = 1 breath/s = 60 breaths/min) can be adjusted to achieve a patient’s pulmonary requirement. Adjusting oscillation varies inspiratory time, mean airway pressure, and ultimately gas exchange. It is recommended that the lung be optimally recruited before beginning HFOV to maximize open alveoli. Beyond the neonatal population, several studies have shown moderate efficacy of HFOV improving oxygenation in severe cases of ARDS. Fort et al. initially demonstrated an efficacy and safety profile in 17 patients with severe ARDS.79 In a recent comparison of HFOV to conventional ventilation modes, Bollen et al. failed to show any difference in mortality between the two strategies, although there was slight improvement in initial oxygenation on first receiving HFOV.80 Future investigations will be needed to determine the utility and practicality of HFOV in ALI and ARDS.


Severe respiratory failure can be a daunting task to the ICU team and difficulty maintaining adequate oxygenation, despite attempting various ventilation modes, can at times seem futile. In the past several years adjunct therapies, both mechanical and pharmaceutical, have been evolving, although with variable efficacy.

Image Prone Positioning

The concept of prone positioning to improve oxygenation was demonstrated by Douglas et al. in a prospective study of six patients with severe ARDS.81 Poorly or nonaerated lung units localize in the dependent lung zones while in the supine position. Prone positioning may improve gas exchange and ventilation/perfusion mismatch by expanding atelectatic portions of the lung akin to the zones of West. One of the largest prospective trials evaluating the efficacy of prone positioning to date randomized 342 patients with moderate and severe ARDS.82 The authors reported that prone positioning did not improve 28-day or 6-month mortality compared with nonprone patients. The prone cohort also had a significantly higher incidence of complications such as hypoxia, need for neuromuscular blockade, loss of vascular access, and endotracheal tube dislodgment. However, some reports suggest that the maximal benefits of prone positioning are only realized when implemented in early ARDS and should be utilized before refractory hypoxia ensues.

Image Pharmacologic Agents

Based on promising animal data, inhaled nitric oxide (NO) was quickly introduced to improve hypoxia in severe ARDS. The mechanism is likely related to the relaxation of vascular smooth muscle in the lung, thereby improving shunt fraction and optimizing the ventilation/perfusion ratio.83 Clinical trials, however, have shown NO therapy to have limited improvement in oxygenation and no significant difference in mortality except possibly in the pediatric population. Other early promising pharmacologic therapies for ARDS, including sildenafil, corticosteroids, and exogenous surfactant, have not been shown to be consistently efficacious.


Managing and weaning a patient from mechanical ventilation is highly variable where the duration and success of weaning is completely dependent on each specific clinical scenario. Some patients are weaned and extubated within hours; others may take weeks to months. It is additionally important to separate the principles of weaning from extubation despite the fact that the terminology is used interchangeably. Weaning is best described as the physical transition and liberation from mechanically assisted ventilation to patient physical autonomy responsible for moving air into and out of the lungs. The process of weaning is not standardized, and various ICUs approach weaning through a variety of methods that often times lack empirical support. As a set of general principles: (1) weaning should be considered early in the course of a ventilated patient, (2) pressure support weaning or AC mode transitional weaning seems to be the most tolerated, and (3) spontaneous breathing trial is the best diagnostic test to assess progress of weaning.

Several factors influence successful weaning including the type of injury sustained, preexisting medical problems, patient toxicology, comorbid or preinjury conditions, and the general state of health. When ventilatory reserve and requirements are exceeded by WOB requirements, patients are considered to be in ventilatory failure and thus mechanical ventilation becomes necessary. It is important to delineate contributing factors leading to respiratory failure since efforts at weaning should be targeted to these conditions to decrease WOB and improve ventilatory capacity (Table 55-4).

TABLE 55-4 Factors Contributing to Respiratory Failure




The concept of WOB is a function of the force required to expand the airways, lungs, and chest wall to an appropriate volume in order to meet oxygen and physiologic demands (pressure × volume/time). Excess WOB may be secondary to inappropriate ventilator settings (i.e., ineffective triggering, inappropriate pressure support, decelerating flow rates), increased airway resistance (mucous plugging, bronchospasm), and decreased pulmonary or thoracic wall compliance (pulmonary edema, ARDS, or hematoma). Additionally, increased WOB may also reflect higher CO2 production from sepsis, fever, or overfeeding. These factors increase respiratory rate since a higher minute ventilation is needed to compensate for acidosis. In concept, overfeeding can be monitored and prevented by a metabolic cart that can assess the RQ calculated from CO2 generated divided by O2 consumed. However, studies supporting the fact that frequent metabolic monitoring improves weaning have been lacking.


Ventilator capacity encompasses the ability of a patient to generate muscular forces needed to meet the ventilatory work required for physiologic demand. To that end, muscle strength (diaphragm, intercostals, and abdominal musculature) combined with other conditions (spinal injury, CIP, nutritional state, metabolic and endocrine disorders) plays a role in defining a patient’s ventilatory capacity.

Significant thoracic trauma, such as severe rib fractures or flail chest, can also limit a patient’s ventilatory capacity. It is also important to consider the effects of pain, improper analgesia, and recent abdominal surgery when beginning the weaning process as overall ventilatory capacity may transiently be decreased.


Image The Weaning Process

The process of weaning can be long and variable and in some instances may account for over 40% of total ventilator time.84 The fact that after several decades of mechanical ventilation therapy, a uniform weaning protocol has not been adopted is testimony to great variability in practice and highlights the paucity of empirical data available to standardize weaning management. Nevertheless, weaning should utilize the concept of conditioning and target strengthening the muscles of breathing (diaphragm, intercostals, and abdominal) that respond similarly to other skeletal muscle conditioning. However, testing strength of conditioning is often difficult and varies greatly depending on injury patterns and overall respiratory health including underlying pneumonia, ALI, and previous medical comorbidities such as COPD.

The methodology of weaning has evolved in the last three decades from a physician-driven approach to a modern protocol-based approach. Under a “stepwise” methodology, patients’ ventilatory requirements are decreased in a stepwise manner, often using a progressive lower set frequency through IMV mode. This method is safe and is highly successful in patients with low WOB and high ventilatory capacity and, during the 1980s, replaced spontaneous breathing trials. However, in the stepwise approach there is no set period of unassisted respiratory exercise. Thus, in a subset of patients with high WOB requirements, a stepwise weaning methodology will prolong mechanical ventilation and may lead to chronic fatigue and deconditioning.84 This method is in contrast to the “sprint” approach (spontaneous breathing trials) where intermittent trials of spontaneous breathing are used once or twice daily and an assessment of WOB and ventilatory capacity is determined accordingly.

In an effort to empirically evaluate optimal weaning techniques, the Spanish Lung Failure Collaborative Group analyzed four weaning methods in an important, large (n = 546), multicenter, prospective trial.77In this study, Esteban et al. randomized patients deemed ready to wean into a stepwise progressive mode: (1) IMV where patients had a set mean frequency of 10 breaths/min and decreased, (2) PS where support was initially a mean of 18 cm H2O and decreased, or into an intermittent spontaneous breathing method, (3) two or times a day with PS <5 cm H2O, or (4) once daily T-piece for 2 hours only. This study reported that both spontaneous breathing methods had the highest rate of successful weaning. Spontaneous breathing led to extubation three times more quickly than IMV and about twice as quickly as high PS ventilation. There was no statistical difference in the rate of success between a once-daily trial and multiple daily trials of spontaneous breathing. Among the three most popular weaning modes, T-piece wean, PS wean, and IMV wean, current data suggest a trend that IMV weaning leads to a longer duration of mechanical ventilation process than spontaneous trials; however, it is important to note that little data have shown any link between weaning method and mortality.


Despite the controversy that exists in method of weaning, the benefit of a formal ICU-wide weaning protocol to hasten weaning is clear.85 Implementing an aggressive weaning protocol should incorporate the Screen, Trial, Exercise, Evaluate, and Report (STEER) Principle. Using this protocol, a team of highly trained respiratory therapists, in conjunction with the ICU team, implement the following daily regimen on each mechanically ventilated patient:

Screen for weaning contraindications (hypoxia, fevers, hemodynamic instability, high WOB, uncontrollable ICP, sedation).

Trial of minimum support breathing (i.e., PS <5 cm H2O).

Exercise according to set protocol (i.e., a 2-hour PS trial, or T-piece trial).

Evaluate progress.

Report the results to the ICU team.

These protocols are designed to reduce physician practice variability and have been shown, in several studies, to reduce ventilator days, ICU days, and weaning duration.

Image Extubation

The weaning process, in principle, begins shortly after mechanical ventilation is initiated. However, extubation, the final removal of the artificial airway, should not be equated or confused with weaning. In a collective national task force consensus paper, discontinuation of mechanical ventilation should be considered if the following physiologic parameters were met86:

1. Reversal of underlying cause of respiratory failure

2. Adequate oxygenation (P/F ratio >150–200 with PEEP <5–8, FiO2 <0.4–0.5, and pH <7.25)

3. Hemodynamic stability

4. The capability to initiate a respiratory effort

As in weaning, prediction of successful extubation varies greatly and traditional parameters such as tidal volume, NIF, and minute ventilation are not well standardized. In a prospective study of 100 weaning patients, Yang and Tobin studied the predictive value of the RSBI calculated by frequency/tidal volume (f/Vt) during spontaneous respiration and the CROP index—a more complicated equation including compliance, rate, oxygenation, and maximal inspiratory pressure.84 Success was defined as patients who remained extubated greater than 24 hours from initial extubation and subsequent isopleths and ROC curves were generated accordingly to predict likelihood of success. The authors concluded that an RSBI >100 breaths/(min L) was 95% likely to fail; conversely, an RSBI <100 breaths/(min L) was 80% likely to succeed. Although these authors have demonstrated the importance of an RSBI, other criteria should be utilized when deciding extubation, including ability for airway protection, secretion control, and sustainability (Table 55-5). Head injury patients can be particularly difficult to manage, since neurologic symptoms can change making criteria for airway and secretion control difficult to evaluate. A GCS of >8 is generally associated with the greatest extubation success. In these patients, early tracheostomy may be a consideration to establish definitive airway control (see Section “Tracheostomy”).

TABLE 55-5 Criteria for Extubation



Tracheostomy in a chronically ventilated patient can offer several advantages including improved oral care, improved pulmonary toilet, decreased oral and vocal cord ulcerations, decreased airway resistance, and improved ability to aggressively wean ventilation while maintaining a secure airway. However, tracheostomy is also associated with such complications as stomal infection, subcutaneous emphysema, hemorrhage, tracheal stenosis, innominate artery fistula, tracheomalacia, and excessive granulation tissue. In a 10-year retrospective analysis, Goettler et al. found that higher injury severity, low admission GCS, higher age (>50 years), and refractory intracranial hypertension all predicted the need for tracheostomy.86 Nevertheless, significant controversy still exists regarding optimal timing for tracheostomy and has broadly resulted in an “early” or “delayed” tracheostomy management approach. Despite several prospective and retrospective studies, definitive and consistent evidence supporting improved outcomes from early tracheostomy has been lacking. Defining what constitutes early tracheostomy further complicates interpreting existing literature since an “early tracheostomy” can vary from 48 hours to 8 days from initiating mechanical ventilation. In one of the few prospective trials addressing this question, Rumbak et al. randomized 120 medical ICU patients into an early percutaneous tracheostomy group (within 48 hours of ICU admission) and a delayed tracheostomy group (14–16 days).87 Patients in the early tracheostomy cohort had 50% decreased mortality and an 80% decrease in pneumonia compared with the delayed group. Conversely, in another study, Sugerman et al. found no difference in ICU days, ventilation days, pneumonia, or mortality between patients randomized to either early (3–5 days) or late (10–14 days) tracheostomies, although the earlier cohort did have a decrease in vocal cord ulceration and subglottic inflammation.88 Perhaps the greatest benefit of early tracheostomy may be seen in the traumatic brain injury population. One prospective study and several retrospective analyses have shown early tracheostomy in severe head injury decreases ventilation days and decreases the incidence of pneumonia, although no benefit in mortality was conferred.89 Ultimately, timing of tracheostomy will be guided by the individual patient, although the need for tracheostomy seems to be directly proportional to the severity of injury.

Image Unplanned Extubations

Despite rigorous patient observation and appropriate sedation, unplanned extubations remain a serious complication with a reported incidence of 3–8% in both the medical and surgical ICU.90 The mortality for these events is near 2% and is mostly a result of sudden hypoxia and ensuing cardiopulmonary arrest. It has also been reported that unplanned extubations requiring reintubation increase the risk of VAP.90In a prospective study analyzing the risk for unplanned extubations, Chevron et al. found that agitation is the highest risk factor followed by oral intubation.91 The authors also measured nursing workload and noted that higher nursing workloads were not predictive of higher unplanned extubations. Interestingly, unplanned extubations were relatively well tolerated in this group and only 37% required reintubation. Predictors of reintubation include a GCS <11, accidental versus patient-initiated extubations, hypoxia, and PaO2/FiO2 ratio of <200. Preventing unplanned extubations requires attention to endotracheal tube security with particular attentiveness during patient transfers, bedside procedures, or prone positioning. Adequate sedation and analgesia are paramount and, although their use should be limited, physical restraints, especially in patients with high sedative tolerance, may be temporarily necessary. An emergency protocol should be readied in the event of an unplanned extubation. Reintubation may be difficult and a team of experts with appropriate experience that can provide a surgical airway should be immediately present.


Patients suffering traumatic injury may be suffering severe pain and may require directed and effective analgesic therapy. Trauma of significant severity may lead to prolonged critical illness, mechanical ventilation, and long-term conscious sedation. Many trauma patients have a history of substance abuse and are at risk for acute withdrawal symptoms and opiate and benzodiazepine tolerance. Inadequate control of pain, agitation, anxiety, and delirium can lead to patient suffering, additional complications, and unnecessarily prolonged intubation, mechanical ventilation, and ICU stay. The need for adequate analgesia and sedation must be balanced against the risks of oversedation. Oversedation may result in hemodynamic instability, increased length of stay and costs, increased respiratory complications including VAP, and possibly long-term decreases in cognitive function. Oversedation may also increase the risks of delirium and possibly posttraumatic stress disorder (PTSD). Delirium is related to increased length of hospital stay, increased health care costs, and higher mortality. ICU care may lead to pain and discomfort due to catheters, drains, endotracheal tubes, and performance of routine nursing care such as airway suctioning, physical therapy dressing changes, and an enforced mobilization. There is practice variation between caregivers and institutions; therefore, several organizations have proposed guidelines for ICU analgesia, sedation, and delirium.92,93 Most academic ICUs employ algorithms for assessment and management of pain, sedation, and delirium. An example of an ICU patient assessment algorithm is shown in Fig. 55-6.


FIGURE 55-6 San Diego Patient Safety Council ICU Sedation Guidelines of Care.


Pain is almost universal in trauma patients; it has been described as an unpleasant sensory or emotional experience that is associated with tissue damage or described in terms of tissue damage. Many patients later recall unrelieved pain when interviewed about their ICU stays. It is a standard of practice to regularly and frequently assess pain; this is readily accomplished by the use of validated pain scores. The Joint Commission has labeled the pain score “the fifth vital sign.” The patient’s sedation should be assessed together with pain as the patient’s comfort is affected by both variables. Parenteral opioids have a long history of efficacy and safety for pain and ICU patient. Since bioavailability and efficacy is variable in this patient population, the intravenous route is to be used whenever possible. Intramuscular and subcutaneous roots should be avoided due to irregular absorption and highly variable serum levels. Patient-controlled analgesia (PCA) may be of great utility in patients able to use it due to the self-control and immediate administration of agent given to the patient. Recent guidelines have recommended preferred parenteral likely agents in the treatment of pain and adult ICU patient. Table 55-6 lists characteristics of the preferred agents in the guidelines: morphine, fentanyl, and hydromorphone (Dilaudid®).92

TABLE 55-6 Suggested Intravenous Opioid Doses for Adult ICU Patientsa (San Diego Patient Safety Council)


Image Morphine

Morphine remains the most commonly used opioid analgesic in the ICU setting due to clinicians’ familiarity in dosing and pharmacokinetics and low cost. It penetrates the blood brain barrier slowly to its poor lipid solubility. This results in a delayed peak effect and a long-acting effect as compared with more lipid-soluble agents such as fentanyl. Morphine does not have a direct negative inotropic effect; however, it has predictable effects of arterial and venous dilation. Morphine also can increase venous capacitance to a greater extent than arteriolar capacitance and also decreases heart rate through a central sympatholytic effect and also direct effect on the spinal atrial node. These effects suggest morphine may be advantageous in the hemodynamically stable or hypertensive patient with myocardial ischemia or cardiogenic pulmonary edema; morphine is recommended as the first-line open agent for parenteral use in the ICU. Disadvantages of morphine and other opioid agonists are the opioid receptor–associated side effects, the most serious of which is centrally mediated respiratory depression. Other effects include sedation, nausea, and sphincter of Oddi spasm. Adverse effects can be reversed if necessary with careful titration of the opioid receptor antagonist naloxone. Careful titration of naloxone may allow reversal of side effects without complete reversal of analgesia. Naloxone must be administered with care as access effect is associated with undesirable hemodynamic effects such as hypertension, tachycardia, and myocardial ischemia as well as acute pulmonary edema. Morphine is also associated with a non-receptor-associated effect of histamine release, and this can result in undesired hypotension, tachycardia, and possibly exacerbating bronchospasm in patients with reactive airway disease. Prolonged use of morphine, especially in renal impairment, can result in the accumulation of a metabolite, morphine-6 glucuronide, and may result in a prolonged sedative effect.

Image Fentanyl

Fentanyl is a lipid-soluble synthetic opioid having a more rapid onset of action than morphine. It is also inexpensive with costs similar to morphine. With small doses the duration of action is short due to redistribution. When large doses are administered, especially by continuous infusion, termination of effect requires elimination, probably due to the saturation of poorly perfused adipose tissue. The pharmacokinetics of fentanyl is not much altered by the presence of cirrhosis, and clearance appears to remain normal in renal failure. Hemodynamically fentanyl maintains cardiovascular stability and does not have significant negative inotropic effect. In the presence of high sympathetic tone, fentanyl may decrease blood pressure indirectly by decreasing central sympathetic output. Fentanyl also predictably causes a decrease in heart rate by a central vagotonic effect. The receptor-associated side effects of fentanyl and their management are the same as described for morphine. Fentanyl is 80–100 times more potent than morphine and has a more rapid onset of action, requiring vigilance with its use. Unlike morphine, fentanyl does not release histamine and therefore may be a better choice in patients who are hemodynamically unstable or who have reactive airway disease. For ICU patients, fentanyl has been recommended as an alternative to morphine in situations of hemodynamic instability, known allergy to morphine, or previous histamine release with morphine.92

Image Hydromorphone (Dilaudid®)

Hydromorphone is a semisynthetic opioid agent that is lipophilic and 5–10 times more potent than morphine. Time to onset of action and duration of action are similar to those of morphine, and hydromorphone’s terminal half-life is 184 minutes. Hydromorphone appears to have minimal hemodynamic effect and does not result in release of histamine. Like morphine and fentanyl, hydromorphone is inexpensive, and is recommended as a third-line agent after morphine and fentanyl.92


Critical care practice guidelines for pain, sedation, and delirium, such as those developed by the SCCM, recommend regular assessment and response to therapy.92 The appropriate target level of sedation is a calm patient who can be easily aroused with maintenance of the normal sleep–wake cycle. Some patients require deeper levels of sedation to facilitate mechanical ventilation or reduce ICP. Use of a sedation scale such as the Richmond Agitation Sedation Scale (RASS) or Riker Sedation Agitation Scale (SAS) is common in most ICUs.94 The choice of intermittent or continuous sedation is important; intermittent sedation relies on recognition of anxiety or agitation and subsequent preparation and administration of the sedative. Continuous sedation reduces the likelihood of delay in administration of sedatives; however, continuous sedative infusions for ICU patients have been shown to increase the duration of mechanical ventilation and length of ICU stay. Weaning of patients from mechanical ventilation is often hampered by oversedation. The benzodiazepines (midazolam and lorazepam) are the most commonly used agents in the ICU for sedation. Propofol and dexmedetomidine have expanded the number of agents available for ICU sedation. The provision of anxiolysis and amnesia are of major importance for critical care patients undergoing intermittently painful procedures or mechanical ventilation. It has been demonstrated that the majority of patients surviving prolonged mechanical ventilation found memory of the experience to be unpleasant. Recent recommendations suggest the use of propofol or midazolam for short-term ICU sedation and lorazepam for long-term sedation.92 Due to its potential to form active metabolites with prolonged duration of action, diazepam is not recommended for a routine use in critical care patients. Characteristics of commonly used benzodiazepines, dexmedetomidine, and propofol are found in Table 55-7.

TABLE 55-7 Suggested Intravenous Sedation Agents for Adult ICU Patientsa (San Diego Patient Safety Council)


Image Midazolam

Midazolam has a rapid onset of action and relatively short elimination half-life (2–2.5 hours). In ICU patients, particularly those with impaired hepatic metabolism, its elimination half-life may be prolonged (4–12 hours). Additionally, midazolam can cause hypotension, particularly in the presence of hypovolemia. Respiratory depression is an expected side effect, but infusions of midazolam need not be completely withdrawn to the patient successfully from mechanical ventilation. Midazolam is metabolized to compounds known as hydroxymidazolam, which have minimal intrinsic benzodiazepine activity. A critical care task force has recommended midazolam as a first-line benzodiazepine for short-term (less than 24-hour) sedation and anxiolysis than in the critical care setting.92

Image Lorazepam

Lorazepam is a potent benzodiazepine having a lower lipid solubility than either diazepam or midazolam, but higher receptor specificity. The lower lipid solubility delays penetration of the blood brain barrier, resulting in longer time to take effect despite a short distribution half-life. The duration of clinical effect from a single bolus dose is also longer than that obtained with a single bolus of midazolam or diazepam. Lorazepam has an intermediate terminal half-life of 10–20 hours and is metabolized into compounds with no intrinsic benzodiazepine activity. Theoretically, lorazepam’s lower lipid solubility and lack of active metabolites decrease the likelihood of prolonged sedation after large doses, relative to midazolam and diazepam. There is no randomized controlled trial comparing the benefits of the benzodiazepines as long-term sedatives in the ICU; however, lorazepam has been recommended as the first-line benzodiazepine for sedation and anxiolysis in the ICU patients requiring sedation for the medium term (>24 hours). Due to its low water solubility, lorazepam is provided in a polyethylene glycol–containing solution that may cause phlebitis or intravascular catheter infiltration on injection. This carrier may cause myocardial depression in large doses. Lorazepam is inexpensive in comparison to midazolam.

Image Propofol

Propofol is a lipid-soluble alkyl phenol intravenous anesthetic that is insoluble in water and formulated in a lipid emulsion. It has hypnotic, amnestic, and antiemetic properties but is devoid of analgesic effect. When used at low doses, it produces sedation. Propofol has minimal active metabolites, even with repeated administration. In critically ill patients sedated with continuous infusions of propofol for a mean of 86 hours, serum levels declined by 50%, 10 minutes after termination of the infusion. Controlled trials of propofol for sedation have shown that doses of 17–52 g/(kg min) of propofol effectively sedate most critically ill patients.95 The effect is rapid and predictable and recovery occurs quickly when the drug is terminated. In a study comparing propofol with midazolam for sedation of ICU patients, time to wake up and tracheal extubation were significantly shorter in the group sedated with propofol.96 Propofol has predictable hemodynamic effects, including arterial and venous dilation, decreased inotropic effect, and decrease of systolic blood pressure of 20–30%. Propofol given as a loading dose will cause profound ventilatory depression. For this reason it is typically limited to mechanically ventilated patients only. Propofol has been recommended as an agent for short-term (less than 24 hours) sedation in the ICU. A study of ventilated patients receiving long-term infusions of propofol as compared with midazolam showed a shorter time to awakening (1 hour vs. 37 hours) and a shorter time to extubation (2 hours vs. 55 hours), suggesting that propofol may be useful for sedation of critically ill patients for longer than 24 hours. SCCM guidelines recommend the use of propofol for up to 48 hours, except in traumatic brain injury patients in whom it may be considered for use up to 5 days.92 Patients on propofol for longer than 24 hours need to be monitored carefully for propofol infusion syndrome. Hypertriglyceridemia and lipid-induced pancreatitis have been reported during prolonged infusion of propofol in the ICU, suggesting that serum triglyceride levels should be monitored in these patients. A lipid solution of propofol also supports rapid bacterial growth at room temperature, and a number of postoperative bacteremias had been linked to poor administration technique.

Image Dexmedetomidine

Dexmedetomidine is an imidazole compound that is a highly selective agonist of the alpha-2 adrenergic receptor with eight times greater affinity than clonidine. It is also shorter acting than clonidine that allows use as an intravenous infusion. Initially evaluated as an anesthetic, it was found to be associated with excess bradycardia and hypotension. Lower doses produce reliable sedation and dexmedetomidine is approved in the adult ICU as a sedative infusion. Initially it was approved for 24 hours or less but has been used in clinical trials for up to 5 days and longer. Patient selection and proper drug infusion are needed to avoid significant hemodynamic effects. Avoiding the initial bolus dose can reduce the incidence of significant bradycardia and hypotension on administration. Dexmedetomidine produces sedation but easy arousal, analgesic-sparing effect, and minimal depression of the respiratory drive. These characteristics are unique in that patients appear to be sedated but are readily roused and interactive and can follow commands. This makes the drug highly suited for patients who are being weaned from the ventilator, especially those who become agitated when other sedation is reduced. Dexmedetomidine is more expensive than benzodiazepines or propofol; however, studies suggest it is cost-effective as it reduces patient ventilator days, delirium, and ICU length of stay as compared with midazolam.97


Monitoring of sedation within the ICU can be performed by the Ramsey, RASS, or Riker (SAS) scales (Table 55-8). The scales have midrange scores that indicate a calm, cooperative patient with high and low scores indicating excess anxiety or agitation or oversedation. A common sedation scale should be used throughout the institution for consistent and effective use. Daily interruption of continuous IV sedation until awakening of mechanically ventilated patients decreases duration of mechanical ventilation and ICU length of stay. The daily awakenings allow for titration of sedation, making the dosing of sedation intermittent. Combination of daily awakening with spontaneous breathing trials results in patients spending less time on mechanical ventilation, less time in a sedative coma, and less time in the ICU and in the hospital. Daily awakening trials are most effective when following a standardized nurse-driven consistent protocol. Patients who should be excluded from a daily awakening trial include those with increased ICP or neuromuscular blockade, those requiring high levels of ventilatory support, and those who are postcoronary artery bypass grafting.

TABLE 55-8 Richmond Agitation Sedation Scale (RASS)2


The daily awakening period should be synchronized with spontaneous breathing trial; in a study of 336 mechanically ventilated ICU patients, doing so achieved significant reduction in ventilator-free days and ICU and hospital length of stay.98 The daily awakening also provides an opportunity for early exercise and mobilization (physical and occupational therapy) (Table 55-9). A recent study of early mobilization during daily awakening in 104 mechanically ventilated patients demonstrated increased independent functional status at hospital discharge, shorter duration of delirium, and fewer ventilator-free days during the 28-day follow-up period than those observed in controls.99 An example of the recent trends toward less use of continuous sedation includes a recent trial of “no sedation” in mechanically ventilated patients.100 In that study, the intervention group patients receiving no sedative medications were compared with control patients on continuous midazolam drips. The majority of the eligible ICU patients could not be enrolled as they had medical indications for deep sedation. Patients in the “no sedation” group were allowed morphine boluses and had frequent nonpharmacologic antianxiety interventions such as reassurance by nurses and reassessment by physicians of pain orders and the need for tubes and catheters. Patients in the “no sedation” group had significantly shorter ventilator days and hospital and ICU length of stay, but did have increased delirium and increased use of haloperidol.

TABLE 55-9 Daily Awakening Trials—Summary Recommendations



Monitoring of sedation with an objective device may be possible with the bispectral index (BIS) monitor. Raw electroencephalographic (EEG) information is obtained through a sensor placed on the forehead. The system processes the EEG information and calculates a number between 0 and 100, which provides a direct measure of the patient’s level of consciousness and response to sedation. A BIS at the time of 100 indicates that patient is fully awake, and a base level of 0 indicates the absence of electrical brain activity. Sedation may be titrated to a specific index depending on the goals for each patient. BIS correlated directly with commonly used sedation scales.

Image Delirium

Delirium is defined as fluctuation in mental status such as inattention, disorganized thinking, hallucinations, disorientation, and an altered level of consciousness. It occurs in up to 65% of hospitalized patients, and up to 87% of patients admitted to the ICU. The outcomes of delirium can be serious for the patient and should be considered as another organ failure that affects patient’s outcome. Delirium can increase the hospital stay and increase health care cost. There has been discussion of ICU delirium rates as being a possible quality measure. Delirium must be considered when assessing pain and sedation in the ICU. It can be further defined as (1) hyperactive delirium: previously referred to as ICU psychosis, includes such symptoms as hypervigilance, restlessness, anger, irritability, and uncooperativeness, and is associated with better overall outcomes; (2) hypoactive delirium: the more common and deleterious, characterized by a lack of awareness, decreased alertness, sparse or slow speech, lethargy, decreased motor activity, and apathy; (3) mixed delirium: apparent in patients with a mixed clinical picture and may occur in up to 54% of patients. Delirium occurs in patients typically 24–72 hours after admission to the ICU. Risk factors existing before hospitalization placing patients at risk for delirium include cognitive impairment, chronic illness (including hypertension), age over 65 years, depression, smoking, alcoholism, and severity of illness. Risk factors arising during hospitalization include congestive heart failure, sepsis, prolonged restraint use, immobility, withdrawal from substance abuse, seizures, dehydration, hyperthermia, head trauma, intracranial mass lesions, and the use of lorazepam, midazolam, morphine, fentanyl, and propofol. Assessment of delirium should be carried out in association with pain and sedation assessments and can be facilitated by the use of a validated delirium score such as the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU) or the Intensive Care Delirium Screening Checklist (ICDSC); these tools may be most usefully administered during a daily awakening.101,102 In addition, the patient should be assessed for QTc prolongation on EKG if therapy is being considered as many antidelirium agents are associated with QTc prolongation and risk of torsade de pointes. CAM-ICU and ISDSC allow rapid consistent assessment for altered level of consciousness, disorganized thinking, inattention, and other delirium features. Nonpharmacologic measures that can be used to manage delirium include daily awakening trials, continuous reorientation of the patient by nursing staff, clocks and calendars visible to the patient, promoting effective sleep/awake cycles, timely removal of restraints and catheters, ensuring use of glasses and hearing aids, minimizing ICU noise and stimulation, and avoiding excessive use of benzodiazepines. Pharmacologic agents used in the management of ICU delirium have not been supported by large clinical trials and agent selection should be based on patient-specific factors (Table 55-10). Intravenous drugs include dexmedetomidine, which may be especially useful in patients scaling spontaneous breathing trials secondary to agitation, and intravenous haloperidol. Intravenous haloperidol achieves peak levels 11 minutes after injection and the half-life of the drug is 10–24 hours. The dose required to control agitation and critically ill patients varies widely; a suggested maximum daily dose in the ICU is 35 mg per day; use of larger doses has been reported. Haloperidol has been associated with a reduced seizure threshold, extrapyramidal reactions (dyskinesia), and laryngeal dystonia. It has also been associated with malignant neuroleptic syndrome. Large doses of intravenous health care should only be administered in a monitored critical care setting and serial EKGs obtained to monitor QTc interval. In patients with hypoactive delirium, the dose of haloperidol may need to be reduced. Oral agents are longer acting and include aripiprazole, risperidone, or oral haloperidol. These agents are associated with QTc prolongation and should only be administered if the measured EKG QTc is less than 440 milliseconds. Another oral delirium agent with greater sedative properties is quetiapine; it is not associated with QTc prolongation.

TABLE 55-10 Pharmacology of Recommended ICU Antidelirium Agents (San Diego Patient Safety Council)



There is a respiratory deficit following upper bowel and thoracic surgery or blunt thoracic trauma (see Chapter 24). Opioid analgesics may exacerbate this deficit by causing respiratory depression, which can be undesirable in ICU patients. The use of regional analgesic techniques typically reduces opioid-related respiratory depression, but may fail to completely restore respiratory function to pretrauma levels. Multiple studies have demonstrated effectiveness of regional analgesia and improving lung volumes and oxygenation and decreasing point complications following blunt thoracic trauma. Epidural analgesia has been well except modality for the treatment of rib fracture pain and the restoration of ventilatory function; however, not all studies have consistently showed positive results. Even with complete relief of pain, epidural methods may only partially restore FVC and forced expiratory volume, minimally improve functional residual capacity, and modestly increase oxygenation. Thus, the pain is only one of several factors leading to a restrictive pulmonary pattern postoperatively. Diaphragmatic dysfunction has been associated as anterior to the decreased lung volumes observed after blunt thoracic trauma. This diaphragmatic dysfunction has been treated to visceral or somatic reflexes decreasing phrenic nerve activity. Spasm and splinting of the abdominal and intercostal muscles may also be involved. Thoracic epidural local anesthesia decreases diaphragmatic dysfunction, increases tidal volume, and decreases respiratory frequency while spinal opiates do not. These effects of epidural local anesthesia may be indirect or direct. Another option for regional control of pain includes paraspinal or regional administration of local anesthetics, particularly with the use of an indwelling catheter and pump system. Such catheters can be placed close to surgical incisions or rib fractures and continuously dilute local anesthetic agents.

In summary, pain is a significant problem after trauma; its relief is necessary not only for patient comfort but also for efficient respiratory therapy. However, pain is not the only mechanism underlying respiratory muscle dysfunction. Inhibitory reflexes of the phrenic nerve are involved and they do not appear to be blocked by spinal opiates; they are however partially blocked by thoracic epidural local anesthetics. No less the analgesia provided by spinal opiates or local anesthetic solutions improves respiratory function more than systemically to the narcotics and can be expected to reduce morbidity and some situations.


Critical care complications interpose a substantial cost burden to the health care institution and society and worsen patient outcomes. There is increasing public awareness of this issue; preventable complications are often reportable to regulatory authorities and many insurers, including Medicare, may refuse to pay hospital or physician charges in case of patients with certain preventable critical care complications. Health care consumers can readily obtain ICU complication rates for their health care institutions from a variety of media. The need to reduce preventable complications of critical care has resulted in a number of guidelines released by professional and regulatory organizations and local hospitals.11,38 These are typically focused on prophylaxis for common ICU complications. Other measures that may be adopted include use of an ICU database to track outcomes and complications and perform risk adjustment, use of robust quality and process improvement programs to detect complications and change practices, and a continuing education program for ICU staff that includes measures to identify and reduce complications. A complete discussion of critical illness complications is beyond the scope of this chapter; however, selected complications that are unique to critical care or are particularly significant to morbidity are discussed. Specific organ system complications are discussed more completely in other chapters.

Image Infectious Complications

Two particularly troublesome and expensive nosocomial infections that occur during critical care after major trauma are CLABSI and VAP. One of the principal issues regarding VAP is its effect on outcome after major injury. There is evidence that pneumonia may exacerbate multiple organ failure and ARDS and should presumably be associated with increased mortality, as has been found in other studies. About 25% of patients who have CVCs placed will develop catheter colonization and 5% will develop a CLABSI after an average of 8 catheter days.103 Both CLABSI and VAP are discussed in detail elsewhere.


Prolonged neuromuscular weakness associated with critical illness was reported as early as the 1950s. Found to be associated with sepsis, hypotension, or multiple organ dysfunction, CIP may prolong weaning from mechanical ventilation, delay return to ambulation, and significantly affect overall recovery from post-traumatic critical illness. The syndrome is characterized by the development of diffuse neurogenic muscle weakness, developing typically over a several-week course of severe critical illness. The neurologic manifestations may include unexplained failure to wean from mechanical ventilation, decreased/absent deep tendon reflexes, tetraparesis, muscle atrophy, decreased fibrillations and compound muscle action potentials, and axonal damage for which evidence is found on electrophysiologic testing. Nerve conduction velocities are near normal, and histologic evaluation of peripheral nerves has shown acute diffuse neurogenic atrophy in muscles and axonal degeneration in nerve tissue. CIP may be far more common than is currently recognized and may commonly affect ventilator weaning and recovery.104 It should be considered as a cause of weaning failures or generalized weakness in the setting of critical illness. Electrophysiologic evaluation of muscle and nerve function is important for the diagnosis. Although not conclusive, available data suggest that the avoidance of long-term agents, particularly in combination with corticosteroids or aminoglycoside antibiotics, may be an important preventative measure. Several factors contributing to CIP and the long-term use of neuromuscular blocking agents (e.g., pancuronium, vecuronium) and amino glycoside antibiotics have been identified. The combination of neuromuscular blocking agents and high-dose steroids may be a particularly important cause. Recovery from CIP, although may be prolonged, may be nearly complete from a clinical standpoint. Follow-up electromyographic studies have shown changes with chronic neurogenic damage, however.

Image The Geriatric Trauma Patient in the ICU

Mortality in elderly trauma patients has been associated with cardiovascular and septic complications.28 Therefore, aggressive monitoring is warranted as it may help in diagnosing physiologic deterioration and in assessing the effectiveness of a number of therapies deployed in an attempt to improve outcomes. In addition, aggressive monitoring may allow clinicians early identification of complications.

Sepsis and subsequent multiple organ system failure cause most late deaths following trauma in the elderly. Urosepsis and pneumonia are common in elderly trauma patients and what would be otherwise a relatively simple problem to treat in the young healthy individual may be the trigger to a cascade of events in the elderly patient, which may culminate with multiple organ dysfunction and death. For these reasons, aggressive and early treatment of these infections, initially with broad-spectrum antibiotics followed by culture-based de-escalation or adjustment of therapy, is a critical determinant of good outcome.

One of the most common causes of death in the geriatric trauma population is pneumonia following blunt chest trauma and rib fractures. Due to decreased pulmonary reserve and associated comorbidities, the elderly trauma patient is generally more susceptible to the development of pneumonia due to an inability to effectively clear secretions and take deep breaths. Two aspects in the early initial care in the ICU in these patients are important: avoidance of fluid overload and adequate analgesia. To this end, patient-controlled narcotic analgesia and/or epidural administration of opiate analgesics or local anesthetics may be helpful in appropriately selected trauma patients.

Geriatric patients are also at increased risk for thromboembolic complications following trauma. Patients in the high-risk group for thromboembolic events (traumatic brain injury, spinal cord injury, complex pelvic fractures, bilateral lower extremity fractures, prolonged immobilization, or a previous history of deep venous thrombosis or pulmonary embolism) should receive pharmacologic prophylaxis with low-molecular-weight heparin, and/or inferior vena caval filter placement when indicated.

More recently, trauma centers in the United States have experienced an epidemic of elderly falls. Many of these patients are taking antiplatelet agents such as aspirin or clopidogrel and warfarin. Even patients with mild to moderate traumatic brain injury are at a much higher risk of developing fatal intracranial hemorrhage due to the use of these drugs and because of “increased space” for hematoma expansion due to brain atrophy. Rapidly obtaining a brain CT scan and quickly assessing coagulation parameters for reversal of anticoagulation are critical in early management. Those individuals who need an immediate craniotomy may benefit from rapid reversal with recombinant factor VIIa or prothrombin complex.

Image Missed Injuries

Missed injuries are the most common cause of preventable death; however, the true incidence of missed injury is difficult to determine. An autopsy review of trauma patients compared their findings with discharge diagnoses and demonstrated that 34% of the patients had missed injuries and in 5%, the missed injury was the cause of death.105 The trauma victim who manifests delayed hypotension, that is, after resuscitation, may be bleeding from an occult injury. The injury can be in the thoracic cavity, the abdominal cavity, the pelvis, or the thigh.

The initial evaluation of the trauma patient centers around recognizing abnormal physiology and the pattern of injury. Surgical intensivists caring for trauma patients must recognize potential injuries likely to occur given a particular mechanism. The challenge becomes the rapid identification of occult injuries before the clinical condition of the patient deteriorates. Missed injury is a major pitfall in the care of trauma and an unexpected deterioration in the condition of the patient in the ICU should prompt a reexamination for possible missed injury, among other causes.

Image Postresuscitation Hypotension in the ICU

Any patient developing hypotension postresuscitation is bleeding until proven otherwise and aggressive investigation of the etiology is mandatory. Such episodes should prompt a reevaluation of the patient’s workup and raise the specter of a missed intra-abdominal injury or a solid organ injury that rebled due to overresuscitation and dislodgment of the initial hemostatic clot. It is a common pitfall to ignore such an episode as an aberration or, perhaps, to explain it away in an attempt to reassure oneself. A common mistake is to assume that resuscitation has been completed and hypotension is due to other causes such as traumatic brain injury or bleeding due to long bone fractures.

There are certain injuries that may require ongoing resuscitation with blood products, most prominent among them a vertical shear-type posterior element pelvic fracture. However, in the absence of any such known injury, a diligent attempt must be made to exclude a missed injury in the abdomen or perhaps an injury whose magnitude was underestimated.

The traditional end points of resuscitation include an adequate urinary output, the trend of the base deficit on the arterial blood gas, and normalization of arterial lactate levels. If one has a PA catheter in place (rarely used nowadays), the oxygen extraction ratio can be used as a gauge of resuscitation. The widening a-v O2 difference or an increasing oxygen extraction ratio is an indicator of inadequate flow, given a certain metabolic demand.

There are situations in which the urinary output may be misleading. The intoxicated patient will have good urinary output even in the face of hypovolemia, because alcohol inhibits the release of ADH from the posterior pituitary; in addition, it is hypertonic and leads to peripheral arterial vasodilation. Another situation in which the urinary output will be misleadingly elevated is in the setting of hyperglycemia. Whether the patient is a diabetic or he or she has received high-dose steroids for a spinal cord injury, the resultant hyperglycemia will cause a misleadingly comforting urinary output. A serum blood sugar over approximately 180–190 mg/dL results in glycosuria and this pitfall must be recognized.

Similarly, the base deficit can also be misinterpreted. The etiology of metabolic acidosis in the injured patient is, until proven otherwise, due to hypoperfusion from hemorrhagic shock; it must be understood, however, that the base deficit can be due to ketosis, nonanion gap acidosis, and sepsis.

Image Hypoxemia

Parenchymal disease is the most common cause of hypoxemia. The causes include aspiration pneumonia, hospital-acquired pneumonia, pulmonary contusion, or the ARDS. Pneumothorax and/or hemothorax may also manifest as hypoxemia but generally occur during the initial phase of the resuscitation and less often in the ICU.

Image Pulmonary Contusion

Pulmonary contusion consists of a direct injury to the lung followed by alveolar hemorrhage and edema from the injury. Experimental evidence has demonstrated clearly that the contusion evolves over the first 24 hours following injury, such that the po2 progressively decreases during that time period.106

The contused lung has leaky capillaries and aggressive fluid resuscitation may result in further deterioration of pulmonary function. In addition, administration of colloid can result in deposition of molecules such as albumin in the interstitial space, where such large molecules are not cleared by the lymphatics. Therefore, judicious administration of crystalloid is likely the best therapy in this circumstance. The pitfall in the management of pulmonary contusion centers around not anticipating the progression of the injury. Above all, the patient with major chest wall trauma, even with the normal chest x-ray on initial evaluation, must be closely observed for the development of a pulmonary contusion.

Image SIRS, Sepsis, Severe Sepsis, and Septic Shock

Injury leads to the activation of the inflammatory response, which may be aggravated by the severity of shock, degree of tissue injury, and secondary insults. To some extent, severely injured patients invariably develop an SIRS. SIRS is defined by manifestation of two or more of the following conditions: (1) temperature >38°C or <36°C; (2) heart rate >90 beats/min, (3) respiratory rate >20 breaths/min or PaCO2 <32 mm Hg, and (4) white blood cell count >12,000/mm3, <4,000/mm3, or >10% immature forms. The treatment of SIRS is supportive.

Sepsis is defined as the presence of a proven infection in a patient with SIRS. Severe sepsis includes sepsis and organ dysfunction, while septic shock encompasses severe sepsis accompanied by hypotension and hypoperfusion, refractory to volume replacement and requiring inotropes (Table 55-11).

TABLE 55-11 Vasopressors and Inotropes


Persistent SIRS and sepsis may ultimately lead to multiple organ dysfunction (discussed in another chapter). Typically, the first organ to fail is the lungs.

The hemodynamic derangements in sepsis, or, more properly, severe sepsis, are classically hypotension, hyperdynamic cardiac index, a low systemic vascular resistance, and metabolic acidosis.

The source of sepsis in the injured patient relates to the type of injuries. It is extremely rare for a patient to have septic shock early after injury, unless there is an obvious infection, such as an aspiration pneumonia or perforated colon. The patient who has leukocytosis with bandemia, fever, and clinical deterioration must be investigated closely for a source of infection. The diagnosis of an infection following major trauma is the biggest pitfall since the cardinal signs of infection such as fever, leukocytosis, and hyperdynamic hemodynamic state can and frequently are the result of the inflammatory cascade, in response to tissue trauma. The pitfall lies in the differentiation between the two entities, sepsis and SIRS. The consequences of liberal use of antibiotics to broadly cover for presumptive sepsis are real, including drug resistance, antibiotic-related colitis, and the risk of fungemia. The consequences of not treating a patient with fever, hyperdynamic state, and signs and symptoms of infection, in the absence of positive cultures or a clear source, are equally daunting, as the patient may indeed be harboring an infection, but the yield of blood cultures and the other surveillance tests are poor. This pitfall is real and common, and, unfortunately, there is no reliable method to clinically distinguish between the entities of sepsis and SIRS until a clear source of infection is identified. The usual sources of infection in the ICU are the lungs, indwelling vascular catheters, the urinary tract, and the wound. Each of these sites must be surveyed for infection.

Image Acalculous Cholecystitis

Acalculous cholecystitis is a “hidden” cause of fever in the ICU. Changes in gastrointestinal motility, characterized by increased gastric residuals and intolerance to enteral nutrition, imply the onset of an infection and one potential source of such infection is the gallbladder.

Acalculous cholecystitis is a disease of the critically ill patient; most of these patients have had major trauma or extensive burns, or are recovering from major surgery. Patients with acquired immune deficiency syndrome (AIDS) are at a special risk for acalculous cholecystitis. The diagnosis can be difficult because patients who develop acalculous cholecystitis tend to be critically ill or severely injured and are frequently unable to react to physical exam. An ultrasound will typically show a thickened gallbladder wall without stones. The diagnostic test of choice is the HIDA scan, which will demonstrate a lack of emptying of the gallbladder in response to administration of cholecystokinin. The initial treatment for acalculous cholecystitis is conservative (n.p.o., antibiotics, intravenous fluids), although some patients will require operative intervention. Alternatively, for those patients who are too sick to tolerate an operation, a percutaneous cholecystostomy tube placement is the best option.

Image Worsening Intracranial Hypertension

The patient who does not have an initial surgically treatable lesion may develop elevated ICPs. The CT finding that correlates best with intracranial hypertension is compression or obliteration of the basilar cisterns and midline shift.107 The level of intracranial hypertension at 72 hours following injury is the primary predictor of outcome in patients with these CT scan findings. It is necessary to anticipate this deterioration and plan on optimization of control of ICP, CPP, and cerebral blood flow. The CPP is calculated by subtracting ICP from the MAP; recent work has shown that the CPP should be maintained at least 60 mm Hg. However, the most important intervention revolves around management of ICP. Aggressive therapy should aim to keep the ICP below 20 mm Hg. The usual adjuncts are mannitol, hypertonic saline, ventriculostomy drainage, and sometimes barbiturate administration. Extreme cases of intractable intracranial hypertension may require surgical decompression (decompressive craniectomy). The techniques and the theories underlying the management of intracranial hypertension are beyond the scope of this chapter.


Ethics may be regarded, in part, as a series of societal formulations predicated on moral values designed to produce a theoretical “optimal good.” With the ability of modern ICUs to maintain physiologic support and homeostasis well beyond the bounds of reasonable sentient existence, physicians are increasingly called on to participate in sometimes difficult live-or-die, withdraw/withhold support decisions. While ICU ethics also include other considerations, such as consent issues and donor organ procurement (see Chapter 50), this section focuses on the withdrawal/withholding of physiologic support including the application of do not resuscitate (DNR) orders in the surgical ICU.


Ethical decision making should involve the careful, considered application of established principles. The temptation to apply one’s individual value system to the decision-making process is strong, but the personal values of the patient are paramount, and the personal judgments of both family members and physicians must be considered in the context of the ethical principles involved. With the possible exception of the principle of distributive justice, each of the following principles should be applied in any given decision-making process:

Beneficence: The principle of “doing good” as applies to a particular patient, individual, or situation.

Nonmaleficence: The principle of avoiding harm or wrongdoing, as applied to a particular situation or individual.

Autonomy: Perhaps the most important operative principle in critical care ethics is the right of self-determination—the inherent right of individuals to make decisions regarding their own treatment options (or withholding thereof) and ultimately make decisions that will impact their survival.

Full disclosure: The principle of accurate communication of information that will allow individuals (or their surrogates) to exercise autonomy or surrogate decision making.

Social (distributive) justice: The principle whereby benefits to an individual, if associated with burdens to another individual or individuals, must be weighed in terms of the most “good” done to the society or group as a whole. This last principle typically involves decisions regarding the distribution of scarce resources to allow the “optimum” treatment of not a single individual, but a population of individuals (society). Such a principle may apply during wartime casualty triage or civilian mass-casualty triage. With respect to ICUs in the United States, it is uncommon, within a given municipality or region, that critical resources are sufficiently scarce so as to result in a denial of treatment opportunities, regardless of the medical conditions involved.


Modern ICUs have developed the increasing technical capability to support homeostasis, even in the face of what would otherwise be overwhelming disease. Increasingly in the trauma population of ICU patients, patient demise is the result of a decision to withdraw or withhold life support. The consequences and irreversibility of such withdraw/withhold decisions in the CCU mandate that they be considered very carefully and predicted on the careful application of ethical and legal principles. Physicians often assume that withdrawing and withholding support are fundamentally different actions that require different thresholds and criteria and potentially different legal consequences. Case law has served to clarify this issue, beginning with Barber v Santa Barbara Superior Court.108 In this case, the court considered withdrawal of support to be the equivalent of withholding, equating each milligram of drug infused or each ventilator breath as an active intervention. That withdrawal of this type of intervention required an additional order was immaterial and still considered a passive (vs. active) act equivalent to withholding. This concept has been further supported by the President’s Commission and critical care consensus panels.109 This critical legal precedent, generally upheld since that time, had the effect of making the withdrawal of support for critical illness the equivalent of withholding support, and thereby not subjecting health care providers to subsequent legal action regarding active (vs. passive) interventions acting to hasten demise.

The importance and impact of withdraw/withhold decisions is considerable. In a review of such decision, Smedira et al. found that in a mixed population of surgical and medical ICU patients, support was withheld in 1% and withdrawn in 5%. The resultant deaths accounted for 45% of all ICU deaths in the two institutions under study.110 The general setting under which withdraw/withhold decisions are made is poor prognosis, which would include a low chance of survival and high likelihood of poor cognitive function. Specific criteria that may allow withdrawal/withholding include the following:

• Provision of further treatment is considered medically futile with respect to achieving well-defined therapeutic goals.

• The patient with decision-making capacity (DMC), in exercising his or her autonomy, chooses to have support withdrawn/withheld.

• A legal surrogate decision maker (legal guardian or durable power of attorney) chooses to have support withdrawn/withheld under medically appropriate circumstances.

• The decision is made on the basis of a prior written medical directive stipulating circumstances under which support may be withdrawn/withheld.

• The physician, acting as a surrogate decision maker in conjunction with next of kin, other consultants, and possibly an ethics committee, elects to withdraw/withhold support. The decision should be based on the following considerations:

Image Information regarding the patient’s wishes: What the patient would be likely to decide if he or she had DMC. This may be obtained from family, friends, or other providers with a history of relevant contact with the patient.

Image The benefits/burdens test: The benefits of continued treatment are outweighed by the burdens to the patient of such treatment in accomplishing the defined therapeutic goals.

Image Substituted judgment: The “reasonable person test.” In the absence of other available information regarding a patient’s wishes, a withdraw/withhold decision should be regarded as one likely to be made by any “reasonable person.”

Image Best interests test: The decision to withdraw/withhold must be medically appropriate and should not significantly diminish opportunity for recovery within the bounds of benefits/burdens.


The concept of futility, derived from the Greek word futilis, meaning “to leak” (as in Greek mythology), refers to the inability to accomplish a therapeutic goal through medical interventions regardless of the duration of such interventions or the frequency with which they are repeated. The concept of futility is central to initial withdraw/withhold support decisions. Determination of futility requires two elements: (a) the establishment of an agreed-upon therapeutic goal and (b) determination of the probability, given the application of medical therapy, of reaching these therapeutic goals. Therapeutic goals should be outlined as specifically as possible and may incorporate elements of both physical and cognitive functions. Patients with DMC must be allowed to define their own therapeutic goals, regardless of how “unreasonable” or stringent these goals may seem. The determination of the probability of medical therapy in achieving those goals should be left to the team of treating physicians. Schneiderman et al. have suggested a numerical definition of futility corresponding to a 1% probability.111 This would imply that medical care was futile if, in the experience of a provider or that reported in the literature, the intervention failed to achieve therapeutic goals in the last 100 cases in which it was applied. Although not all providers may adhere to such rigorous quantitative definitions, the general concept of futility is well recognized and broadly applied in such settings. Futility must always be judged, however, based on defined therapeutic goals. Once a given treatment or treatments are deemed to be futile, based on established therapeutic goals, the physician is under no legal or ethical obligation to provide such treatment. Current writing in medical ethics, public policy, and more recently by case law has supported this perspective, although some ethicists do not support this medical prerogative. Individual circumstances, however, particularly in regard to family considerations and consistent with the principle of beneficence, may occasionally be indications for short-term delivery of what otherwise would be considered medically futile care.

During the early ICU phase of care, determination of futility often is achieved in patients with severe nonsurvivable brain injuries (transaxial gunshot injuries or open skull fractures with loss of brain tissue and GCS = 3).

Patients with significant comorbidities, particularly those with end-stage liver or renal disease, and those with malignancies are not candidates for extreme resuscitation efforts or massive transfusion. To the same extent, blunt trauma victims with significant intra-abdominal bleeding and associated nonsurvivable brain injury, >60% TBSA burns, and cervical spinal cord transactions should not be aggressively resuscitated. Even if some of these patients survive, they will do so with a very poor quality of life and it is the trauma surgeon’s responsibility to provide the family with such information.

Patients who received massive blood transfusions in the OR without complete hemostasis usually present to the ICU as hypothermic, acidotic, and coagulopathic. Although difficult to estimate, the rapidity of bleeding and transfusion requirements greater than two blood volumes have traditionally been used as markers of irreversibility and most surgeons would agree that in these circumstances, if bleeding is nonmechanical and hypothermia and coagulopathy have not improved over a period of 6–8 hours, further resuscitation efforts are probably futile and will serve only to consume precious resources.112

Late in the course of ICU care, defining futility in patients with neurologic injury is difficult, as it is almost impossible to predict those who will not achieve a meaningful recovery. Age and comorbidities should be considered when presenting the facts to family members, and they may facilitate the family’s decision to withdraw care.

In patients with multiple organ dysfunction, determination of futility is often achieved in those with three or more organ failures as mortality in these circumstances approaches 100%.


DMC, which has different implications than the legal term competent, requires the following:

• The patient must be able to comprehend and communicate information relevant to making a decision.

• The patient must be able to comprehend his or her alternatives and the benefits and burdens (risks) associated with each.

• The patient must be able to reason and deliberate about these alternatives against a background of stable personal values.

The determination of DMC may generally be made by the primary physician or team, but occasionally, particularly in the presence of underlying psychiatric illness, psychiatric consultation may be of benefit in making this determination. The exercise of autonomy is best accomplished by allowing any patient with DMC to participate in withdraw/withhold or DNR decisions. In many cases, a patient’s specific wishes may not be known and legal surrogates or medical directives may be unavailable. With careful adjustment (lightening to the extent possible) of any conscious sedation and directed efforts made in communicating the alternatives (full disclosure), many patients can make decisions or at least provide valuable information allowing surrogates to do so.

The written medical directive provides an alternative means by which a patient may exercise autonomy. Medical directives may be structured in a variety of ways—depending on known medical conditions, age of the patient, and so forth—but they typically contain language stipulating preferences (or not) for life-sustaining treatment as a function of expected physical and cognitive outcomes.


In situations where the patient does not have DMC and no specific medical directive is available, a surrogate decision-making process is utilized. Participants in the surrogate decision-making process may include family members, friends, other health care providers, clergy, and members of the institutional ethics committee. Family members invested with durable power of attorney may make clinically appropriate decisions independently on behalf of the patient. (Durable refers to power of attorney, including instances in which the patient is incapacitated.) Surrogate decision making should involve close collaboration between health care providers and the family to the extent possible. The specific role of the family, however, in the absence of durable power, is not to actually make withdraw/withhold decisions but to provide information allowing providers to best formulate decisions consistent with what the patient’s desires would be. This information may include knowledge of the patient’s values, goals, religious or philosophical beliefs, or previously expressed wishes with respect to medical care. The importance of this information should not be underestimated; in the absence of durable power of attorney, family members have no recognized legal authority to dictate care or to make decisions that are medically or ethically inappropriate. Health care professionals also have a primary responsibility to act in what they perceive to be the best interests of the patient, which may occasionally be in conflict with wishes expressed by the family or, on rare occasions, the wishes expressed by a durable power agent.

In the absence of any information pertaining to the presumptive wishes of the patient, substituted judgment may be applied based on several considerations:

• What would a “reasonable” person want under similar circumstances? Physicians should rarely be making this judgment independently, and the involvement of colleagues or a formal ethics committee consultation may be appropriate.

• What would the likely benefits of continued treatment be (treatment outcome), weighed against the burdens imposed by both the treatment and the treatment outcome? The anticipated degree of functional recovery and resultant quality of life are important factors in this consideration. Therapeutic intervention designed to prolong life in a setting in which the quality of that life—because of severity of pain, lack of cognitive function, and so forth—is regarded as being excessively burdensome to an individual may not be medically appropriate.

• What would, overall, be in the patient’s best interests? In many cases, the outcome from a critical illness and specific degree of disability, at a given point in time, cannot be predicted with any degree of certainty. The best interests test prevents premature withdraw/withhold decisions from being made that might deprive a patient of an opportunity for satisfactory recovery.

Substituted judgment involves subjective analysis and is associated with a significant amount of uncertainty. Although the strong tendency may be to continue treatment when therapeutic goals or the patient’s wishes are not known, this course of action may also be inappropriate. Additional professional consultation, including the institutional ethics committee, may help resolve the more difficult cases.


With good ongoing communication, full disclosure of relevant information and prognosis, and sensitivity to patient/family dynamics, withdraw/withhold support decisions can generally be made smoothly. The liberal use of consultants, particularly those involving the neurosciences, may be a valuable adjunct under such circumstances. Ethics consultation services have also been shown to affect management outcome when they are made available. Conflicts may arise, however, over issues of futility, therapeutic goals, and the conduct of surrogate decision making. These conflicts are often based on misunderstanding or mistrust and exacerbated by the lack of good communication.

In situations in which conflict between care providers and family members appears to be irreconcilable, even with ethics committee and institutional risk management/legal consultation, a number of alternatives should be offered to the family. These alternatives may include the following: (a) procurement of additional intramural or extramural medical, religious, or ethical consultations; (b) transferring the care of the patient to another provider within the institution; (c) transfer of the patient to another institution, under the care of another provider; and (d) procurement of a court order mandating a course of treatment.

In most cases, careful collaboration with next of kin and extensive consultation with subspecialty services and/or ethics committee usually will serve to clarify issues surrounding the withdrawal/withholding of support and allow reasonable decisions to be made.


The institution of DNR medical directives has been primarily based on the desire to avoid the indignity of futile cardiopulmonary resuscitation in the setting of inexorable progression of underlying known disease processes. Such orders are frequently applied to patients with long-standing terminal disease (e.g., cancer, end-stage AIDS, irreversible multiple organ failure). DNR orders are implicitly coupled to futility judgments about the benefit of cardiopulmonary resuscitation to achieve prospectively defined therapeutic goals. Under such circumstances, these orders are entirely appropriate and desirable in the variety of clinical disease states. The practice of DNR suspension has occurred most frequently in the operating room, where the intensity of therapeutic intervention frequently results in iatrogenic complications. Most physicians and ethicists believe these complications should be treated, since they generally do not constitute inexorable progression of a known disease process, for which the DNR order was originally designed. In the ICU, particularly the surgical ICU, similar situations may exist. The intensity of critical care interventions with mechanical ventilators and the use of complex drug regimens, including vasoactive drugs, may lead to unexpected (iatrogenic) complications unrelated directly to the underlying disease process for which a DNR order may have been originally placed. In addition, the critical illness being treated in the surgical ICU, particularly for trauma patients, results from a discreet event (surgery or injury) as opposed to complications or exacerbations of chronic disease, as is often the case for medical patients. The presumed reversibility of the physiologic sequelae constitutes a basis for ICU treatment of the trauma patient. As such, DNR for the trauma patient should be applied very carefully to expected complications, or physiologic changes due to inexorable progression of known underlying disease, as opposed to unexpected, very reversible iatrogenic complications.

DNR orders for trauma patients may be more appropriately linked to withdraw/withhold support conditions than more generally to patients with a poor but potentially reversible prognosis. The variability with which DNR orders are interpreted and the “message” such orders send to providers also raises concerns about their applicability for trauma patients. A study by Clemency and Thompson suggests a wide variability in the perception of DNR orders in the perioperative period.113 This variability conceivably would apply to the “peri-injury” period as well. In addition to the potential for the inappropriate application of DNR orders in an ICU, application may result in less aggressive care on the part of ancillary providers.

DNR orders constitute the prospective application of a specific “withhold support” decision and should be made with the same care and consideration, as described in the above section, as any decision involving the limitation of critical care.


ICUs have the expectation of being able to provide innovative, highly complex care to help save the lives and limbs of severe trauma patients. However, with the aging of the population there are increased numbers of geriatric trauma victims, in addition to other cases where medical technology can sustain life but not necessarily return quality of life to the severely injured patient. Elderly patients may have preexisting multiple comorbidities to which trauma may be a final, irrecoverable insult. Patients may have multiple, complex problems and may require multiple ICU admissions over the course of their illness. There are several barriers to providing effective palliative care in ICU setting. A 2006 survey of critical care providers noted that these barriers include:

• Insufficient communication skills about end-of-life issues among health care providers

• Inability of patients to participate in discussions about their treatment

• Unrealistic expectations on the part of patients and families about the prognosis of patients or the effectiveness of ICU treatment

• Lack of advance directives from patients about how they wish their care to be handled at the end of life114

Palliative care programs have been instituted at ICUs that focus on effective pain and symptom management whether life-prolonging or curative care is being pursued or is being withheld or withdrawn. They may also ease case management burdens on primary physician and staff and provide assistance with care coordination and time-intensive patient–family communication. Members of the American College of Surgeons Surgical Palliative Care Task Force have identified common criteria where a palliative care consult may be indicated (Table 55-12).115

TABLE 55-12 Top 10 Criteria for Consultation Identified by Members of American College of Surgeons Surgical Palliative Care Task Force



Family meetings facilitate communication between families of patients and critical care providers. Effective communication improves clinical decision making, patient and family satisfaction, and the psychological well-being of family members (Table 55-13). As a result, research agendas and lists of quality indicators are increasingly recognizing the importance of good communication with families. Some practical guidelines for family meetings are shown in Table 55-14.

TABLE 55-13 Talking with ICU Families


TABLE 55-14 Guide to ICU Family Meetings



1. American College of Surgeons. Resources for Optimal Care of the Injured Patient. Chicago, IL: American College of Surgeons; 2006.

2. Pronovost PJ, Angus DC, Dorman T, et al. Physician staffing patterns and clinical outcomes in critically ill patients: a systematic review. JAMA. 2002;288:2151.

3. Pronovost PJ, Jenckes MW, Dorman T, et al. Organizational characteristics of intensive care units related to outcomes of abdominal aortic surgery. JAMA. 1999;281:1310.

4. Pronovost PJ, Young T, Dorman T, et al. Association between CCU physician staffing and outcomes: a systematic review. Crit Care Med. 1999;27:A43.

5. Nathens AB, Rivara FP, MacKenzie EJ, et al. The impact of an intensivist-model ICU on trauma-related mortality. Ann Surg. 2006;244:545.

6. Suarez JI, Zaidat OO, Suri MF, et al. Length of stay and mortality in neurocritically ill patients: impact of a specialized neurocritical care team. Crit Care Med. 2004;32:2311.

7. Varelas PN, Conti MM, Spanaki MV. The impact of a neurointensivist-led team on a semiclosed neurosciences intensive care unit. Crit Care Med. 2004;32:2191.

8. The Leapfrog Group. ICU Physician Staffing. Washington, DC: The Leapfrog Group. Available at: Accessed May 2010.

9. Agency for Health Care Policy and Research. Interim Manual Clinical Practice Guideline Development. Rockville, MD: Agency for Health Care Policy and Research; May 1991.

10. Hammond JJ. Protocols and guidelines in critical care: development and implementation. Curr Opin Crit Care. 2001;7:464.

11. Cocanour CS, Peninger M, Domonoske BD, et al. Decreasing ventilator-associated pneumonia in a trauma ICU. J Trauma. 2006;61:122.

12. McLean SE, Jensen LA, Schroeder DG, et al. Improving adherence to a mechanical ventilation weaning protocol for critically ill adults: outcomes after an implementation program. Am J Crit Care. 2006;15:299.

13. Aboutanos SZ, Duane TM, Malhotra AK, et al. Prospective evaluation of an extubation protocol in a trauma intensive care unit population. Am Surg. 2006;72:393.

14. Keroack MA, Cerese J, Cuny J, et al. The relationship between evidence-based practices and survival in patients requiring prolonged mechanical ventilation in academic medical centers. Am J Med Qual. 2006; 21:91.

15. Curtis JR, Cook DJ, Wall RJ, et al. Intensive care unit quality improvement: a “how-to” guide for the interdisciplinary team. Crit Care Med. 2006;34:211.

16. Kennedy JF, Nightingale JM. Cost savings of an adult hospital nutrition support team. Nutrition. 2005;1:1127.

17. Pun BT, Gordon SM, Peterson JF, et al. Large-scale implementation of sedation and delirium monitoring in the intensive care unit: a report from two medical centers. Crit Care Med. 2005;3:1199.

18. Winters BD, Gurses AP, Lehmann H, et al. Clinical review: checklists—translating evidence into practice. Crit Care. 2009;13(6):210.

19. Byrnes MC, Schuerer DJE, Schallom ME, et al. Implementation of a mandatory checklist of protocols and objectives improves compliance with a wide range of evidence-based intensive care unit practices. Crit Care Med. 2009;37(10):2775–2781.

20. McMillan TR, Hyzy RC. Bringing quality improvement into the intensive care unit. Crit Care Med. 2007;35(2 suppl):S59–S65.

21. Vincent JL. Give your patient a fast hug (at least) once a day. Crit Care Med. 2005;33(6):1225–1229.

22. Vincent RW. Critically ill patients need “FAST HUGS BID” (an updated mnemonic). Crit Care Med. 2009;37(7):2326–2327.

23. Shepherd SJ, Pearse RM. Role of central and mixed venous oxygen saturation measurement in perioperative care. Anesthesiology. 2009;111: 649–656.

24. de Wilde RB, Geerts BF, Cui J, van den Berg PC, Jansen JR. Performance of three minimally invasive cardiac output monitoring systems. Anaesthesia. 2009;64:762–769.

25. Mayer J, Suttner S. Cardiac output derived from arterial pressure waveform. Curr Opin Anaesthesiol. 2009;22:804–808.

26. Connors AF Jr, Speroff T, Dawson NV, et al. The effectiveness of right heart catheterization in the initial care of critically ill patients. SUPPORT Investigators. JAMA. 1996;276:889–897.

27. Ivanov R, Allen J, Calvin JE. The incidence of major morbidity in critically ill patients managed with pulmonary artery catheters: a meta-analysis. Crit Care Med. 2000;28:615–619.

28. Scalea TM, Simon HM, Duncan AO, et al. Geriatric blunt multiple trauma: improved survival with early invasive monitoring. J Trauma. 1990;30:129–134.

29. Bernard GR, Sopko G, Cerra F, et al. Pulmonary artery catheterization and clinical outcomes: National Heart, Lung, and Blood Institute and Food and Drug Administration Workshop Report. Consensus Statement. JAMA. 2000;283:2568–2572.

30. Burns JM, Sing RF, Mostafa G, et al. The role of transesophageal echocardiography in optimizing resuscitation in acutely injured patients. J Trauma. 2005;59:36–40.

31. Sohn DW, Shin GJ, Oh JK, et al. Role of transesophageal echocardiography in hemodynamically unstable patients. Mayo Clin Proc. 1995;70: 925–931.

32. Miami Clinical Trials Group. Splanchnic hypoperfusion-directed therapies in trauma: a prospective, randomized trial. Am Surg. 2005;71: 252–260.

33. Gomersall CD, Joynt GM, Freebairn RC, Hung V, Buckley TA, Oh TE. Resuscitation of critically ill patients based on the results of gastric tonometry: a prospective, randomized, controlled trial. Crit Care Med. 2000;28:607–614.

34. Wallace AW, Salahieh A, Lawrence A, Spector K, Owens C, Alonso D. Endotracheal cardiac output monitor. Anesthesiology. 2000;92: 178–189.

35. Cohn SM, Nathens AB, Moore FA, et al. Tissue oxygen saturation predicts the development of organ dysfunction during traumatic shock resuscitation. J Trauma. 2007;62:44–54.

36. Hebert PC, Wells G, Blajchman MA, et al. A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. Transfusion Requirements in Critical Care Investigators, Canadian Critical Care Trials Group. N Engl J Med. 1999;340:409–417.

37. Rivers E, Nguyen B, Havstad S, et al. Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med. 2001;345: 1368–1377.

38. Dellinger RP, Levy MM, Carlet JM, et al. Surviving Sepsis Campaign: international guidelines for management of severe sepsis and septic shock: 2008. Crit Care Med. 2008;36:296–327.

39. Napolitano LM, Kurek S, Luchette FA, et al. Clinical practice guideline: red blood cell transfusion in adult trauma and critical care. Crit Care Med. 2009;37:3124–3157.

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

41. Singh AK, Szczech L, Tang KL, et al. Correction of anemia with epoetin alfa in chronic kidney disease. N Engl J Med. 2006;355: 2085–2098.

42. Pfeffer MA, Burdmann EA, Chen CY, et al. A trial of darbepoetin alfa in type 2 diabetes and chronic kidney disease. N Engl J Med. 2009;361: 2019–2032.

43. Corwin HL, Gettinger A, Rodriguez RM, et al. Efficacy of recombinant human erythropoietin in the critically ill patient: a randomized, double-blind, placebo-controlled trial. Crit Care Med. 1999;27:2346–2350.

44. Corwin HL, Gettinger A, Pearl RG, et al. Efficacy of recombinant human erythropoietin in critically ill patients: a randomized controlled trial. JAMA. 2002;288:2827–2835.

45. Corwin HL, Gettinger A, Fabian TC, et al. Efficacy and safety of epoetin alfa in critically ill patients. N Engl J Med. 2007;357:965–976.

46. Napolitano LM, Fabian TC, Kelly KM, et al. Improved survival of critically ill trauma patients treated with recombinant human erythropoietin. J Trauma. 2008;65:285–297.

47. Arroliga AC, Guntupalli KK, Beaver JS, Langholff W, Marino K, Kelly K. Pharmacokinetics and pharmacodynamics of six epoetin alfa dosing regimens in anemic critically ill patients without acute blood loss. Crit Care Med. 2009;37:1299–1307.

48. Christmas AB, Camp SM, Barrett MC, et al. Removal of erythropoietin from anaemia trauma practice guideline does not increase red blood cell transfusions and decreases hospital utilization costs. Injury. 2009;40: 1330–1335.

49. Bratton SL, Chestnut RM, Ghajar J, et al. Guidelines for the management of severe traumatic brain injury. VI. Indications for intracranial pressure monitoring. J Neurotrauma. 2007;24(suppl 1):S37–S44.

50. Thomason MH, Rutledge R. A statewide analysis of the use of intracranial pressure monitoring with the outcome of severe head injury in 1,563 patients. J Trauma. 1998;45(1):192.

51. Goel RS, Goyal NK, Dharap SB, Kumar M, Gore MA. Utility of optic nerve ultrasonography in head injury. Injury. 2008;39:519–524.

52. Kimberly HH, Shah S, Marill K, Noble V. Correlation of optic nerve sheath diameter with direct measurement of intracranial pressure. Acad Emerg Med. 2008;15:201–204.

53. Reid CL. Poor agreement between continuous measurements of energy expenditure and routinely used prediction equations in intensive care unit patients. Clin Nutr. 2007;26:649–657.

54. van den Berghe G, Wouters P, Weekers F, et al. Intensive insulin therapy in the critically ill patients. N Engl J Med. 2001;345:1359–1367.

55. Vogelzang M, Nijboer JM, van der Horst IC, Zijlstra F, ten Duis HJ, Nijsten MW. Hyperglycemia has a stronger relation with outcome in trauma patients than in other critically ill patients. J Trauma. 2006;60: 873–877.

56. Brunkhorst FM, Engel C, Bloos F, et al. Intensive insulin therapy and pentastarch resuscitation in severe sepsis. N Engl J Med. 2008;358: 125–139.

57. Preiser JC, Devos P, Ruiz-Santana S, et al. A prospective randomised multi-centre controlled trial on tight glucose control by intensive insulin therapy in adult intensive care units: the Glucontrol study. Intensive Care Med. 2009;35:1738–1748.

58. Finfer S, Chittock DR, Su SY, et al. Intensive versus conventional glucose control in critically ill patients. N Engl J Med. 2009;360: 1283–1297.

59. Annane D, Sebille V, Charpentier C, et al. Effect of treatment with low doses of hydrocortisone and fludrocortisone on mortality in patients with septic shock. JAMA. 2002;288:862–871.

60. Sprung CL, Annane D, Keh D, et al. Hydrocortisone therapy for patients with septic shock. N Engl J Med. 2008;358:111–124.

61. Reed RL. Contemporary issues with bacterial infection in the intensive care unit. Surg Clin North Am. 2000;80:895–909.

62. Appelgren P, Hellstrom I, Weitzberg E, Soderlund V, Bindslev L, Ransjo U. Risk factors for nosocomial intensive care infection: a long-term prospective analysis. Acta Anaesthesiol Scand. 2001;45:710–719.

63. Severinghaus JW, Bradley AF. Electrodes for blood pO2 and pCO2 determination. J Appl Physiol. 1958;13:515–520.

64. Soubani AO. Noninvasive monitoring of oxygen and carbon dioxide. Am J Emerg Med. 2001;19:141–146.

65. Vukmir RB, Heller MB, Stein KL. Confirmation of endotracheal tube placement: a miniaturized infrared qualitative CO2 detector. Ann Emerg Med. 1991;20:726–729.

66. Warner KJ, Cuschieri J, Garland B, et al. The utility of early end-tidal capnography in monitoring ventilation status after severe injury. J Trauma. 2009;66:26–31.

67. Meade M, Guyatt G, Cook D, et al. Predicting success in weaning from mechanical ventilation. Chest. 2001;120:400S–424S.

68. Soo Hoo GW, Park L. Variations in the measurement of weaning parameters: a survey of respiratory therapists. Chest. 2002;121:1947–1955.

69. Petersen GW, Baier H. Incidence of pulmonary barotrauma in a medical ICU. Crit Care Med. 1983;11:67–69.

70. Webb HH, Tierney DF. Experimental pulmonary edema due to intermittent positive pressure ventilation with high inflation pressures. Protection by positive end-expiratory pressure. Am Rev Respir Dis. 1974;110:556–565.

71. Ricard JD, Dreyfuss D, Saumon G. Production of inflammatory cytokines in ventilator-induced lung injury: a reappraisal. Am J Respir Crit Care Med. 2001;163:1176–1180.

72. Tejeda M, Boix JH, Alvarez F, et al. Comparison of pressure support ventilation and assist-control ventilation in the treatment of respiratory failure. Chest. 1997;111:1322–1325.

73. Betensley AD, Khalid I, Crawford J, et al. Patient comfort during pressure support and volume controlled-continuous mandatory ventilation. Respir Care. 2008;53:897–902.

74. Sternberg R, Sahebjami H. Hemodynamic and oxygen transport characteristics of common ventilatory modes. Chest. 1994;105:1798–1803.

75. Guldager H, Nielsen SL, Carl P, et al. A comparison of volume control and pressure-regulated volume control ventilation in acute respiratory failure. Crit Care. 1997;1:75–77.

76. MacIntyre NR. Respiratory function during pressure support ventilation. Chest. 1986;89:677–683.

77. Esteban A, Frutos F, Tobin MJ, et al. A comparison of four methods of weaning patients from mechanical ventilation. Spanish Lung Failure Collaborative Group. New Engl J Med. 1995;332:345–350.

78. Garner W, Downs JB, Stock MC, et al. Airway pressure release ventilation (APRV). A human trial. Chest. 1988;94:779–781.

79. Fort P, Farmer C, Westerman J, et al. High-frequency oscillatory ventilation for adult respiratory distress syndrome—a pilot study. Crit Care Med. 1997;25:937–947.

80. Bollen CW, van Well GT, Sherry T, et al. High frequency oscillatory ventilation compared with conventional mechanical ventilation in adult respiratory distress syndrome: a randomized controlled trial [ISRCTN 24242669]. Crit Care. 2005;9:R430–R439.

81. Douglas WW, Rehder K, Beynen FM, et al. Improved oxygenation in patients with acute respiratory failure: the prone position. Am Rev Respir Dis. 1977;115:559–566.

82. Taccone P, Pesenti A, Latini R, et al. Prone positioning in patients with moderate and severe acute respiratory distress syndrome: a randomized controlled trial. JAMA. 2009;302:1977–1984.

83. Michael JR, Barton RG, Saffle JR, et al. Inhaled nitric oxide versus conventional therapy: effect on oxygenation in ARDS. Am J Respir Crit Care Med. 1998;157:1372–1380.

84. Yang KL, Tobin MJ. A prospective study of indexes predicting the outcome of trials of weaning from mechanical ventilation. New Engl J Med. 1991;324:1445–1450.

85. Kollef MH, Shapiro SD, Silver P, et al. A randomized, controlled trial of protocol-directed versus physician-directed weaning from mechanical ventilation. Crit Care Med. 1997;25:567–574.

86. Goettler CE, Fugo JR, Bard MR, et al. Predicting the need for early tracheostomy: a multifactorial analysis of 992 intubated trauma patients. J Trauma. 2006 (5):991–996.

87. Rumbak MJ, Newton M, Truncale T, et al. A prospective, randomized, study comparing early percutaneous dilational tracheotomy to prolonged translaryngeal intubation (delayed tracheotomy) in critically ill medical patients. Crit Care Med. 2004;32:1689–1694.

88. Sugerman HJ, Wolfe L, Pasquale MD, et al. Multicenter, randomized, prospective trial of early tracheostomy. J Trauma. 1997;43:741–747.

89. Ahmed N, Kuo YH. Early versus late tracheostomy in patients with severe traumatic head injury. Surg Infect. 2007;8:343–347.

90. de Lassence A, Alberti C, Azoulay E, et al. Impact of unplanned extubation and reintubation after weaning on nosocomial pneumonia risk in the intensive care unit: a prospective multicenter study. Anesthesiology. 2002;97:148–156.

91. Chevron V, Menard JF, Richard JC, et al. Unplanned extubation: risk factors of development and predictive criteria for reintubation. Crit Care Med. 1998;26:1049–1053.

92. Jacobi J, Fraser GL, Coursin DB, et al. Clinical practice guidelines for the sustained use of sedatives and analgesics in the critically ill adult. Crit Care Med. 2002;30:119–141.

93. Vender JS, Szokol JW, Murphy GS, Nitsun M. Sedation, analgesia, and neuromuscular blockade in sepsis: an evidence-based review. Crit Care Med. 2004;32:S554–S561.

94. Sessler CN, Gosnell MS, Grap MJ, et al. The Richmond Agitation-Sedation Scale: validity and reliability in adult intensive care unit patients. Am J Respir Crit Care Med. 2002;166:1338–1344.

95. Barrientos-Vega R, Mar Sanchez-Soria M, Morales-Garcia C, Robas-Gomez A, Cuena-Boy R, Ayensa-Rincon A. Prolonged sedation of critically ill patients with midazolam or propofol: impact on weaning and costs. Crit Care Med. 1997;25:33–40.

96. Aitkenhead AR, Pepperman ML, Willatts SM, et al. Comparison of propofol and midazolam for sedation in critically ill patients. Lancet. 1989;2:704–709.

97. Riker RR, Shehabi Y, Bokesch PM, et al. Dexmedetomidine vs midazolam for sedation of critically ill patients: a randomized trial. JAMA. 2009;301:489–499.

98. Girard TD, Kress JP, Fuchs BD, et al. Efficacy and safety of a paired sedation and ventilator weaning protocol for mechanically ventilated patients in intensive care (Awakening and Breathing Controlled trial): a randomised controlled trial. Lancet. 2008;371:126–134.

99. Schweickert WD, Pohlman MC, Pohlman AS, et al. Early physical and occupational therapy in mechanically ventilated, critically ill patients: a randomised controlled trial. Lancet. 2009;373:1874–1882.

100. Strom T, Martinussen T, Toft P. A protocol of no sedation for critically ill patients receiving mechanical ventilation: a randomised trial. Lancet. 2010;375:475–480.

101. Ely EW, Margolin R, Francis J, et al. Evaluation of delirium in critically ill patients: validation of the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU). Crit Care Med. 2001;29: 1370–1379.

102. Plaschke K, von Haken R, Scholz M, et al. Comparison of the confusion assessment method for the intensive care unit (CAM-ICU) with the Intensive Care Delirium Screening Checklist (ICDSC) for delirium in critical care patients gives high agreement rate(s). Intensive Care Med. 2008;34:431–436.

103. Saint S, Veenstra DL, Lipsky BA. The clinical and economic consequences of nosocomial central venous catheter-related infection: are antimicrobial catheters useful? Infect Control Hosp Epidemiol. 2000;21:375–380.

104. Hund EF, Fogel W, Krieger D, DeGeorgia M, Hacke W. Critical illness polyneuropathy: clinical findings and outcomes of a frequent cause of neuromuscular weaning failure. Crit Care Med. 1996;24:1328–1333.

105. Albreksten S, Thomsen J. Detection of injuries in traumatic deaths: the significance of the medicolegal autopsy. Forensic Sci. 1989;42:135.

106. Fulton RL, Peter ET. Compositional and histologic effects of fluid therapy following pulmonary contusion. J Trauma. 1974;14:783.

107. Patel NY, Hoyt DB, Nakaji P, et al. Traumatic brain injury: patterns of failure of nonoperative management. J Trauma. 2000;48(3): 367–375.

108. Luce JM. Withholding and withdrawing life support from the critically ill: how does it work in clinical practice? Respir Care. 1991;36:417–426.

109. Ethical and moral guidelines for the initiation, continuation, and withdrawal of intensive care. American College of Chest Physicians/Society of Critical Care Medicine Consensus Panel. Chest. 1990;97: 949–958.

110. Smedira NG, Evans BH, Grais LS, et al. Withholding and withdrawal of life support from the critically ill. N Engl J Med. 1990;322:309–315.

111. Schneiderman LJ, Jecker NS, Jonsen AR. Medical futility: its meaning and ethical implications. Ann Intern Med. 1990;112:949–954.

112. Coimbra R, Lee J, Bansal V, Hollingsworth-Fridlund P. Recognizing/accepting futility: prehospital, emergency center, operating room and intensive care unit. J Trauma Nurs. 2007;14(2):1–4.

113. Clemency MV, Thompson NJ. “Do not resuscitate” (DNR) orders in the perioperative period—a comparison of the perspectives of anesthesiologists, internists, and surgeons. Anesth Analg. 1994;78: 651–658.

114. Nelson JE, Angus DC, Weissfeld LA, et al. End-of-life care for the critically ill: a national intensive care unit survey. Crit Care Med. 2006; 34:2547–2553.

115. Bradley CT, Brasel KJ. Developing guidelines that identify patients who would benefit from palliative care services in the surgical intensive care unit. Crit Care Med. 2009;37:946–950.

If you find an error or have any questions, please email us at Thank you!