Chapter 293: Approach to the Patient with Critical Illness

The care of critically ill patients requires a thorough understanding of pathophysiology and centers initially on the resuscitation of patients at the extremes of physiologic deterioration. This resuscitation is often fast-paced and occurs early, without a detailed awareness of the patient’s chronic medical problems. While physiologic stabilization is taking place, intensivists attempt to gather important background medical information to supplement the real-time assessment of the patient’s current physiologic conditions. Numerous tools are available to assist intensivists in the accurate assessment of pathophysiology and management of incipient organ failure, offering a window of opportunity for diagnosing and treating underlying disease(s) in a stabilized patient. Indeed, the use of invasive interventions such as mechanical ventilation and renal replacement therapy is commonplace in the intensive care unit (ICU). An appreciation of the risks and benefits of such aggressive and often invasive interventions is vital to ensure an optimal outcome. Nonetheless, intensivists must recognize when a patient’s chances for recovery are remote or nonexistent and must counsel and comfort dying patients and their significant others. Critical care physicians often must redirect the goals of care from resuscitation and cure to comfort when the resolution of an underlying illness is not possible.

In the ICU, illnesses are frequently categorized by degree of severity. Numerous severity-of-illness (SOI) scoring systems have been developed and validated over the past three decades. Although these scoring systems have been validated as tools to assess populations of critically ill patients, their utility in predicting individual patient outcomes is not clear. SOI scoring systems are important for defining populations of critically ill patients. Such systematic scoring allows effective comparison of groups of patients enrolled in clinical trials. In verifying a purported benefit of therapy, investigators must be confident that different groups involved in a clinical trial have similar illness severities. SOI scores are also useful in guiding hospital administrative policies, directing the allocation of resources such as nursing and ancillary care and assisting in assessments of quality of ICU care over time. Scoring system validations are based on the premise that age, chronic medical illnesses, and derangements from normal physiology are associated with increased mortality rates. All existing SOI scoring systems are derived from patients who have already been admitted to the ICU.

SOI scoring systems cannot be used to predict survival in individual patients. No established scoring systems that purport to direct clinicians’ decision-making regarding criteria for admission to an ICU are available. Thus the use of SOI scoring systems to direct therapy and clinical decision-making cannot be recommended. Instead, these tools should be used as a source of important data to complement clinical bedside decision-making.

The most commonly utilized scoring systems are the SOFA (Sequential Organ Failure Assessment), the APACHE (Acute Physiology and Chronic Health Evaluation), and the SAPS (Simplified Acute Physiology Score) systems.

THE SOFA SCORING SYSTEM

The SOFA scoring system is composed of scores from six organ systems, graded from 0 to 4 according to the degree of dysfunction (Table 293-1). The score accounts for clinical interventions; it can be measured repeatedly (i.e., each day), and rising scores correlate well with increasing mortality. Patients with suspected infection can be predicted to have poor outcomes typical of sepsis if they have at least two of the following clinical criteria: respiratory rate >22, altered mental status, or systolic blood pressure <100 mmHg. Recently, a new bedside clinical score using two or more of the above clinical criteria has emerged and is termed quickSOFA (qSOFA). qSOFA is intended to screen patients for ICU admission from out-of-hospital, emergency department, and hospital ward settings.

THE APACHE II SCORING SYSTEM

The APACHE II system is the most commonly used SOI scoring system in North America. Age, type of ICU admission (after elective surgery vs nonsurgical or after emergency surgery), chronic health problems, and 12 physiologic variables (the worst values for each in the first 24 h after ICU admission) are used to derive a score. The predicted hospital mortality rate is derived from a formula that takes into account the APACHE II score, the need for emergency surgery, and a weighted, disease-specific diagnostic category (Table 293-2). The relationship between APACHE II score and mortality risk is illustrated in Fig. 293-1. Updated versions of the APACHE scoring system (APACHE III and APACHE IV) have been published.

THE SAPS SCORING SYSTEM

The SAPS II score, used more frequently in Europe than in the United States, was derived in a manner similar to the APACHE score. This score is not disease-specific but rather incorporates three underlying disease variables: AIDS, metastatic cancer, and hematologic malignancy. SAPS 3, which utilizes a 1-h rather than a 24-h window for measuring physiologic derangement scores, was developed in 2005.

See also Chap. 296.

INITIAL EVALUATION

Shock, a common condition necessitating ICU admission or occurring in the course of critical care, is defined by the presence of multisystem end-organ hypoperfusion. Clinical indicators include reduced mean arterial pressure (MAP), tachycardia, tachypnea, cool skin and extremities, acute altered mental status, and oliguria. Hypotension is usually, though not always, present. The end result of multiorgan hypoperfusion is tissue hypoxia, often with accompanying lactic acidosis. Since the MAP is the product of cardiac output and systemic vascular resistance (SVR), reductions in blood pressure can be caused by decreases in cardiac output and/or SVR. Accordingly, once shock is contemplated, the initial evaluation of a hypotensive patient should include an early bedside assessment of the adequacy of cardiac output (Fig. 293-2). Clinical evidence of diminished cardiac output includes a narrow pulse pressure (systolic BP minus diastolic BP)—a marker that correlates with stroke volume—and cool extremities with delayed capillary refill. Signs of increased cardiac output include a widened pulse pressure (particularly with a reduced diastolic pressure), warm extremities with bounding pulses, and rapid capillary refill. If a hypotensive patient has clinical signs of increased cardiac output, it can be inferred that the reduced blood pressure is from decreased SVR.

In hypotensive patients with signs of reduced cardiac output, an assessment of intravascular volume status is appropriate. A hypotensive patient with decreased intravascular volume status may have a history suggesting hemorrhage or other volume losses (e.g., vomiting, diarrhea, polyuria). Although evidence of a reduced jugular venous pressure (JVP) is often sought, static measures of right atrial pressure do not predict fluid responsiveness reliably; the change in right atrial pressure as a function of spontaneous respiration is a better predictor of fluid responsiveness (Fig. 293-3). Patients with fluid-responsive (i.e., hypovolemic) shock also may manifest large changes in pulse pressure as a function of respiration during mechanical ventilation (Fig. 293-4). A hypotensive patient with increased intravascular volume and cardiac dysfunction may have S₃ and/or S₄ gallops on examination, increased JVP, extremity edema, and crackles on lung auscultation. The chest x-ray may show cardiomegaly, widening of the vascular pedicle, Kerley B lines, and pulmonary edema. Chest pain and electrocardiographic changes consistent with ischemia may be noted (Chap. 298).

In hypotensive patients with clinical evidence of increased cardiac output, a search for causes of decreased SVR is appropriate. The most common cause of high-cardiac-output hypotension is sepsis (Chap. 297). Other causes include liver failure, severe pancreatitis, burns, trauma, anaphylaxis, thyrotoxicosis, and peripheral arteriovenous shunts.

In summary, the most common categories of shock are hypovolemic, cardiogenic, and high-cardiac-output with decreased SVR (high-output hypotension). Certainly more than one category can occur simultaneously (e.g., hypovolemic and septic shock).

The initial assessment of a patient in shock should take only a few minutes. It is important that aggressive resuscitation is instituted on the basis of the initial assessment, particularly since early resuscitation from septic and cardiogenic shock may improve survival (see below). If the initial bedside assessment yields equivocal or confounding data, more objective assessments such as ultrasound/echocardiography may be useful. In spontaneously breathing patients, inferior vena cava collapse seen on ultrasound predicts a fluid responsive state. Increasingly, ultrasound of the thorax and abdomen is used by intensivists as an extension of the physical examination to assess rapidly imputed filling volumes, adequacy of cardiac performance, and for indices of other specific conditions (e.g., pericardial tamponade, pulmonary embolus, pulmonary edema, pneumothorax). The goal of aggressive resuscitation is to reestablish adequate tissue perfusion and thus to prevent or minimize end-organ injury.

MECHANICAL VENTILATORY SUPPORT

(See also Chap. 295) During the initial resuscitation of patients in shock, principles of advanced cardiac life support should be followed. As such patients may be obtunded and unable to protect the airway, an early assessment of the airway is mandatory. Early intubation and mechanical ventilation often are required. Reasons for the institution of endotracheal intubation and mechanical ventilation include acute hypoxemic respiratory failure and ventilatory failure, which frequently accompany shock. Acute hypoxemic respiratory failure may occur in patients with cardiogenic shock and pulmonary edema (Chap. 298) as well as in those who are in septic shock with pneumonia or acute respiratory distress syndrome (ARDS) (Chaps. 294 and 297). Ventilatory failure often occurs as a consequence of an increased load on the respiratory system in the form of acute metabolic (often lactic) acidosis or decreased lung compliance due to pulmonary edema. Inadequate perfusion to respiratory muscles in the setting of shock may be another reason for early intubation and mechanical ventilation. Normally, the respiratory muscles receive a very small percentage of the cardiac output. However, in patients who are in shock with respiratory distress, the percentage of cardiac output dedicated to respiratory muscles may increase by tenfold or more. Lactic acid production from inefficient respiratory muscle activity presents an additional ventilatory load.

Mechanical ventilation may relieve the work of breathing and allow redistribution of a limited cardiac output to other vital organs. Patients demonstrate respiratory distress by an inability to speak full sentences, accessory use of respiratory muscles, paradoxical abdominal muscle activity, extreme tachypnea (>40 breaths/min), and decreasing respiratory rate despite an increasing drive to breathe due to exhaustion. When patients with shock are treated with mechanical ventilation, a major goal is for the ventilator to assume all or the majority of the work of breathing, facilitating a state of minimal respiratory muscle work. With the institution of mechanical ventilation for shock, further declines in MAP are frequently seen. The reasons include impeded venous return from positive-pressure ventilation, reduced endogenous catecholamine secretion once the stress associated with respiratory failure abates, and the actions of drugs used to facilitate endotracheal intubation (e.g., propofol, opiates). Patients with right heart dysfunction or preexiting pulmonary hypertension may also have diminished cardiac output related to the increases in right ventricular afterload resulting from positive pressure ventilation. Accordingly, hypotension should be anticipated during and following endotracheal intubation. Because many of these patients may be fluid responsive, IV volume administration should be considered. Figure 293-2 summarizes the diagnosis and treatment of different types of shock. For further discussion of individual forms of shock, see Chaps. 296, 297, and 298.

Respiratory failure is one of the most common reasons for ICU admission. In some ICUs, ≥75% of patients require mechanical ventilation during their stay. Respiratory failure can be categorized mechanistically on the basis of pathophysiologic derangements in respiratory function.

TYPE I: ACUTE HYPOXEMIC RESPIRATORY FAILURE

This type of respiratory failure occurs with alveolar flooding and subsequent intrapulmonary shunt physiology. Alveolar flooding may be a consequence of pulmonary edema, lung injury, pneumonia, or alveolar hemorrhage. Pulmonary edema can be further categorized as occurring due to elevated pulmonary microvascular pressures, as seen in heart failure and intravascular volume overload or ARDS (“low-pressure pulmonary edema,” Chap. 294). This syndrome is defined by acute onset (≤1 week) of bilateral opacities on chest imaging that are not fully explained by cardiac failure or fluid overload and of shunt physiology requiring positive end-expiratory pressure (PEEP). Type I respiratory failure occurs in clinical settings such as sepsis, gastric aspiration, pneumonia, near-drowning, multiple blood transfusions, and pancreatitis. The mortality rate among patients with ARDS was traditionally very high (50–70%), although changes in patient care have led to mortality rates closer to 30% (see below).

It is well established that mechanical ventilation of patients with ARDS may propagate lung injury. As seen in Fig. 293-5, the pressure-volume relationship of the lung in ARDS is not linear. Alveoli may collapse at very low lung volumes. Animal studies have suggested that stretching and overdistention of injured alveoli during mechanical ventilation can further injure the lung. Concern over this alveolar overdistention, termed ventilator-induced “volutrauma,” led to a multicenter, randomized, prospective trial comparing traditional ventilator strategies for ARDS (large tidal volume: 12 mL/kg of ideal body weight) with a low tidal volume (6 mL/kg of ideal body weight). This study showed a dramatic reduction in mortality rate in the low-tidal-volume group from that in the high-tidal-volume group (31 versus 39.8%). Other studies have shown that large tidal volumes may lead to ARDS in patients who initially do not have this problem. Neuromuscular blockade and prone positioning have been shown to improve survival in those with severe ARDS. In addition, a “fluid-conservative” management strategy (maintaining a low central venous pressure [CVP] or pulmonary capillary wedge pressure [PCWP]) is associated with fewer days of mechanical ventilation than a “fluid-liberal” strategy (maintaining a relatively high CVP or PCWP) in ARDS. There is growing interest in avoiding intubation in patients with ARDS by the use of a variety of devices, such as masks, high flow oxygen delivery systems, and helmets for respiratory support; however, this is tempered by concern that higher tidal volumes during spontaneous breathing with these devices could result in progression of preexisting lung injury.

TYPE II RESPIRATORY FAILURE

This type of respiratory failure is a consequence of alveolar hypoventilation and results from the inability to eliminate carbon dioxide effectively. Mechanisms are categorized by impaired central nervous system (CNS) drive to breathe, impaired strength with failure of neuromuscular function in the respiratory system, and increased load(s) on the respiratory system. Reasons for diminished CNS drive to breathe include drug overdose, brainstem injury, sleep-disordered breathing, and severe hypothyroidism. Reduced strength can be due to impaired neuromuscular transmission (e.g., myasthenia gravis, Guillain-Barré syndrome, amyotrophic lateral sclerosis) or respiratory muscle weakness (e.g., myopathy, electrolyte derangements, fatigue).

The overall load on the respiratory system can be subclassified into resistive loads (e.g., bronchospasm), loads due to reduced lung compliance (e.g., alveolar edema, atelectasis, intrinsic positive end-expiratory pressure [auto-PEEP]—see below), loads due to reduced chest wall compliance (e.g., pneumothorax, pleural effusion, abdominal distention), and loads due to increased minute ventilation requirements (e.g., pulmonary embolus with increased dead-space fraction, sepsis).

The mainstays of therapy for type II respiratory failure are directed at reversing the underlying cause(s) of ventilatory failure. Noninvasive positive-pressure ventilation with a tight-fitting facial or nasal mask, with avoidance of endotracheal intubation, often stabilizes these patients. This approach has been shown to be beneficial in treating patients with exacerbations of chronic obstructive pulmonary disease; it has been tested less extensively in other kinds of respiratory failure but may be attempted nonetheless in the absence of contraindications (hemodynamic instability, inability to protect the airway, respiratory arrest).

TYPE III RESPIRATORY FAILURE

This form of respiratory failure results from lung atelectasis. Because atelectasis occurs so commonly in the perioperative period, this form is also called perioperative respiratory failure. After general anesthesia, decreases in functional residual capacity lead to collapse of dependent lung units. Such atelectasis can be treated by frequent changes in position, chest physiotherapy, upright positioning, and control of incisional and/or abdominal pain. Noninvasive positive-pressure ventilation may also be used to reverse regional atelectasis.

TYPE IV RESPIRATORY FAILURE

This form results from hypoperfusion of respiratory muscles in patients in shock. Normally, respiratory muscles consume <5% of total cardiac output and oxygen delivery. Patients in shock often experience respiratory distress due to pulmonary edema (e.g., in cardiogenic shock), lactic acidosis, and anemia. In this setting, up to 40% of cardiac output may be distributed to the respiratory muscles. Intubation and mechanical ventilation can allow redistribution of the cardiac output away from the respiratory muscles and back to vital organs while the shock is treated.

(See also Chap. 295) Whereas a thorough understanding of the pathophysiology of respiratory failure is essential for optimal patient care, recognition of a patient’s readiness to be liberated from mechanical ventilation is likewise important. Several studies have shown that daily spontaneous breathing trials can identify patients who are ready for extubation. Accordingly, all intubated, mechanically ventilated patients should undergo daily screening of respiratory function. If oxygenation is stable (i.e., Pao₂/FIo₂ [partial pressure of oxygen/fraction of inspired oxygen] >200 and PEEP ≤5 cmH₂O), cough and airway reflexes are intact, and no vasopressor agents or sedatives are being administered, the patient has passed the screening test and should undergo a spontaneous breathing trial. This trial consists of a period of breathing through the endotracheal tube without ventilator support (continuous positive airway pressure [CPAP] of 5 cmH₂O with or without low level pressure support [e.g., 5 cmH₂O] and an open T-piece breathing system have all been validated) for 30–120 min. The spontaneous breathing trial is declared a failure and stopped if any of the following occur: (1) respiratory rate >35/min for >5 min, (2) O₂ saturation <90%, (3) heart rate >140/min or a 20% increase or decrease from baseline, (4) systolic blood pressure <90 mmHg or >180 mmHg, or (5) increased anxiety or diaphoresis. If, at the end of the spontaneous breathing trial, none of the above events has occurred and the ratio of the respiratory rate and tidal volume in liters (f/V_T) is <105, the patient can be extubated. Such protocol-driven approaches to patient care can have an important impact on the duration of mechanical ventilation and ICU stay. In spite of such a careful approach to liberation from mechanical ventilation, up to 10% of patients develop respiratory distress after extubation and may require resumption of mechanical ventilation. Many of these patients will require reintubation. The use of noninvasive ventilation in patients in whom extubation fails may be associated with worse outcomes than are obtained with immediate reintubation.

Mechanically ventilated patients frequently require sedatives and analgesics. Opiates are the mainstay of therapy for analgesia in mechanically ventilated patients. After adequate pain control has been ensured, additional indications for sedation include anxiolysis; treatment of subjective dyspnea; reduction of autonomic hyperactivity, which may precipitate myocardial ischemia; and reduction of total O₂ consumption (Vo₂). Non-benzodiazepine sedatives are preferred since benzodiazepines are associated with worse patient outcomes.

The neuromuscular blocking agent cisatracurium is occasionally used to facilitate mechanical ventilation in patients with profound ventilator dyssynchrony despite optimal sedation, particularly in the setting of severe ARDS. Use of these agents may result in prolonged weakness—a myopathy known as the postparalytic syndrome. For this reason, neuromuscular blocking agents typically are used as a last resort when aggressive sedation fails to achieve patient-ventilator synchrony. Because neuromuscular blocking agents result in pharmacologic paralysis without altering mental status, sedative-induced amnesia is mandatory when these agents are administered.

Amnesia can be achieved reliably with propofol and benzodiazepines such as lorazepam and midazolam. Outside the setting of pharmacologic paralysis, few data support the idea that amnesia is mandatory in all patients who require intubation and mechanical ventilation. Since many of these critical patients have impaired hepatic and renal function, sedatives and opiates may accumulate when given for prolonged periods. A nursing protocol–driven approach to sedation of mechanically ventilated patients or daily interruption of sedative infusions paired with daily spontaneous breathing trials has been shown to prevent excessive drug accumulation and shorten the duration of both mechanical ventilation and ICU stay.

Multiorgan system failure, which is commonly associated with critical illness, is defined by the simultaneous presence of physiologic dysfunction and/or failure of two or more organs. Typically, this syndrome occurs in the setting of severe sepsis, shock of any kind, severe inflammatory conditions such as pancreatitis, and trauma. The fact that multiorgan system failure occurs commonly in the ICU is a testament to our current ability to stabilize and support single-organ failure. The ability to support single-organ failure aggressively (e.g., by mechanical ventilation or by renal replacement therapy) has reduced rates of early mortality in critical illness. As a result, it is uncommon for critically ill patients to die in the initial stages of resuscitation. Instead, many patients succumb to critical illness later in the ICU stay, after the initial presenting problem has been stabilized.

Although there is debate regarding specific definitions of organ failure, several general principles governing the syndrome of multiorgan system failure apply. First, organ failure, no matter how it is defined, must persist beyond 24 h. Second, mortality risk increases with the accrual of failing organs. Third, the prognosis worsens with increased duration of organ failure. These observations remain true across various critical care settings (e.g., medical versus surgical).

Because respiratory failure and circulatory failure are common in critically ill patients, monitoring of the respiratory and cardiovascular systems is undertaken frequently. Evaluation of respiratory gas exchange is routine in critical illness. The “gold standard” remains arterial blood-gas analysis, in which pH, Pao₂, partial pressure of carbon dioxide (Pco₂), and O₂ saturation are measured directly. With arterial blood-gas analysis, the two main functions of the lung—oxygenation of arterial blood and elimination of CO₂—can be assessed directly. In fact, the blood pH, which has a profound effect on the drive to breathe, can be assessed only by such sampling. Although sampling of arterial blood is generally safe, it may be painful and cannot provide continuous information. In light of these limitations, noninvasive monitoring of respiratory function is often employed.

PULSE OXIMETRY

The most commonly utilized noninvasive technique for monitoring respiratory function, pulse oximetry takes advantage of differences in the absorptive properties of oxygenated and deoxygenated hemoglobin. At wavelengths of 660 nm, oxyhemoglobin reflects light more effectively than does deoxyhemoglobin, whereas the reverse is true in the infrared spectrum (940 nm). A pulse oximeter passes both wavelengths of light through a perfused digit such as a finger, and the relative intensity of light transmission at these two wavelengths is recorded. From this information, the relative percentage of oxyhemoglobin is derived. Since arterial pulsations produce phasic changes in the intensity of transmitted light, the pulse oximeter is designed to detect only light of alternating intensity. This feature allows distinction of arterial and venous blood O₂ saturations.

RESPIRATORY SYSTEM MECHANICS

Respiratory system mechanics can be measured in patients during mechanical ventilation (Chap. 295). When volume-controlled modes of mechanical ventilation are used, accompanying airway pressures can easily be measured as long as the patient is passive. The peak airway pressure is determined by two variables: airway resistance and respiratory system compliance. At the end of inspiration, inspiratory flow can be stopped transiently. This end-inspiratory pause (plateau pressure) is a static measurement, affected only by respiratory system compliance and not by airway resistance. Therefore, during volume-controlled ventilation, the difference between the peak (airway resistance + respiratory system compliance) and plateau (respiratory system compliance only) airway pressures provides a quantitative assessment of airway resistance. Accordingly, during volume-controlled ventilation, patients with increases in airway resistance typically have increased peak airway pressures as well as abnormally high gradients between peak and plateau airway pressures (typically >15 cmH₂O) at a constant inspiratory flow rate of 1 L/sec. The compliance of the respiratory system is defined by the change in volume of the respiratory system per unit change in pressure.

The respiratory system can be divided into two components: the lungs and the chest wall. Normally, respiratory system compliance is ~100 mL/cmH₂O. Pathophysiologic processes such as pleural effusions, pneumothorax, and increased abdominal girth all reduce chest wall compliance. Lung compliance may be reduced by pneumonia, pulmonary edema, interstitial lung disease, or auto-PEEP. Accordingly, patients with abnormalities in compliance of the respiratory system (lungs and/or chest wall) typically have elevated peak and plateau airway pressures but a normal gradient between these two pressures. Auto-PEEP occurs when there is insufficient time for emptying of alveoli before the next inspiratory cycle. Since the alveoli have not decompressed completely, alveolar pressure remains positive at the end of exhalation (functional residual capacity). This phenomenon results most commonly from obstruction of distal airways in disease processes such as asthma and COPD. Auto-PEEP with resulting alveolar overdistention may result in diminished lung compliance, reflected by abnormally increased plateau airway pressures. Modern mechanical ventilators allow breath-to-breath display of pressure and flow, permitting detection of problems such as patient-ventilator dyssynchrony, airflow obstruction, and auto-PEEP (Fig. 293-6).

CIRCULATORY STATUS

Oxygen delivery (Qo₂) is a function of cardiac output and the content of O₂ in the arterial blood (Cao₂). The Cao₂ is determined by the hemoglobin concentration, the arterial hemoglobin saturation, and dissolved O₂ not bound to hemoglobin. For normal adults:

It is apparent that nearly all of the O₂ delivered to tissues is bound to hemoglobin and that the dissolved O₂ (Pao₂) contributes very little to O₂ content in arterial blood or to O₂ delivery. Normally, the content of O₂ in mixed venous blood (C^–vo₂) is 15.76 mL/dL since the mixed venous blood is 75% saturated. Therefore, the normal tissue extraction ratio for O₂ is Cao₂ – C^–vo₂/Cao₂ ([21.16 – 15.76]/21.16) or ~25%. A pulmonary artery catheter allows measurements of O₂ delivery and the O₂ extraction ratio.

Information on the venous O₂ saturation allows assessment of global tissue perfusion. A reduced venous O₂ saturation may be caused by inadequate cardiac output, reduced hemoglobin concentration, and/or reduced arterial O₂ saturation. An abnormally high Vo₂ may also lead to a reduced venous O₂ saturation if O₂ delivery is not concomitantly increased. Abnormally increased Vo₂ in peripheral tissues may be caused by problems such as fever, agitation, shivering, and thyrotoxicosis.

The pulmonary artery catheter originally was designed as a tool to guide therapy for acute myocardial infarction but has been used in the ICU for evaluation and treatment of a variety of other conditions, such as ARDS, septic shock, congestive heart failure, and acute renal failure. This device has never been validated as a tool associated with reduction in morbidity and mortality rates. Indeed, despite numerous prospective studies, mortality or morbidity rate benefits associated with use of the pulmonary artery catheter have never been reported in any setting. Accordingly, it appears that routine pulmonary artery catheterization is not indicated as a means of monitoring and characterizing circulatory status in most critically ill patients.

Static measurements of circulatory parameters (e.g., CVP, PCWP) do not provide reliable information on the circulatory status of critically ill patients. In contrast, dynamic assessments measuring the impact of breathing on the circulation are more reliable predictors of responsiveness to IV fluid administration. A decrease in CVP of >1 mmHg during inspiration in a spontaneously breathing patient may predict an increase in cardiac output after IV fluid administration. Similarly, a changing pulse pressure during mechanical ventilation of a passive patient has been shown to predict an increase in cardiac output after IV fluid administration, assuming the R-R interval is stable.

SEPSIS IN THE CRITICAL CARE UNIT

(See also Chap. 297) Sepsis, is defined as life-threatening organ dysfunction (i.e., an increase in Sequential Organ Failure Assessment [SOFA] of 2 points or more) caused by a dysregulated response to infection. Poor outcomes can be anticipated in patients with 2 or more of the following: respiratory rate >22 per min, altered mentation, systolic blood pressure <100 mmHg. Sepsis is a leading cause of death in noncoronary ICUs in the United States, with case rates expected to increase as the population ages and a higher percentage of people are vulnerable to infection.

NOSOCOMIAL INFECTIONS IN THE ICU

Many therapeutic interventions in the ICU are invasive and predispose patients to infectious complications. These interventions include endotracheal intubation, indwelling vascular catheters, transurethral bladder catheters, and other catheters placed into sterile body cavities (e.g., tube thoracostomy, percutaneous intraabdominal drainage catheterization). The longer such devices remain in place, the more prone to these infections patients become. For example, ventilator-associated events such as ventilator-associated pneumonia correlate strongly with the duration of intubation and mechanical ventilation. Therefore, an important aspect of preventive care is the timely removal of invasive devices as soon as they are no longer needed. Moreover, multidrug-resistant organisms are commonplace in the ICU.

Infection control is critical in the ICU. Care bundles, which include measures such as frequent hand washing, are effective but underutilized strategies. Other components of care bundles, such as protective isolation of patients colonized or infected by drug-resistant organisms, are also commonly used. Silver-coated endotracheal tubes reportedly reduce the incidence of ventilator-associated pneumonia. Studies evaluating multifaceted, evidence-based strategies to decrease catheter-related bloodstream infections have shown improved outcomes with strict adherence to measures such as hand washing, full-barrier precautions during catheter insertion, chlorhexidine skin preparation, avoidance of the femoral site, and timely catheter removal.

DEEP-VENOUS THROMBOSIS (DVT)

(See also Chap. 273) All ICU patients are at high risk for this complication because of their predilection for immobility. Therefore, all should receive some form of prophylaxis against DVT. The most commonly employed forms of prophylaxis are subcutaneous low-dose heparin injections and sequential compression devices for the lower extremities. Observational studies report an alarming incidence of DVTs despite the use of these standard prophylactic regimens. Furthermore, heparin prophylaxis may result in heparin-induced thrombocytopenia, another nosocomial complication in critically ill patients.

Low-molecular-weight heparins such as enoxaparin are more effective than unfractionated heparin for DVT prophylaxis in high-risk patients (e.g., those undergoing orthopedic surgery) and are associated with a lower incidence of heparin-induced thrombocytopenia. Fondaparinux, a selective factor Xa inhibitor, is even more effective than enoxaparin in high-risk orthopedic patients.

STRESS ULCERS

Prophylaxis against stress ulcers is not necessary for all ICU patients. It should only be administered to high-risk patients, such as those with coagulopathy or respiratory failure. Histamine receptor-2 antagonists are preferred over proton pump inhibitors because the latter are associated with increased incidence of C. difficile colitis and pneumonia.

NUTRITION AND GLYCEMIC CONTROL

These are important issues that may be associated with respiratory failure, impaired wound healing, and dysfunctional immune response in critically ill patients. Early enteral feeding is reasonable, with some data suggesting that permissive underfeeding of nonprotein calories is not inferior to full goal feeding. Certainly, enteral feeding, if possible, is preferred over parenteral nutrition, which is associated with numerous complications, including hyperglycemia, fatty liver, cholestasis, and sepsis. When parenteral feeding is necessary to supplement enteral nutrition, delaying this intervention until day 8 in the ICU results in better recovery and fewer ICU-related complications. Tight glucose control is an area of controversy in critical care. Although one study showed a significant mortality benefit when glucose levels were aggressively normalized in a large group of surgical ICU patients, other studies of both medical and surgical ICU patients suggested that tight glucose control resulted in increased rates of mortality.

ICU-ACQUIRED WEAKNESS

ICU-acquired weakness occurs frequently in patients who survive critical illness, particularly those with SIRS and/or sepsis. Both neuropathies and myopathies have been described, most commonly after ~1 week in the ICU. The mechanisms behind ICU-acquired weakness syndromes are poorly understood, they are known to present with heterogeneous muscle pathophysiology. Intensive insulin therapy may reduce polyneuropathy in critical illness. Very early physical and occupational therapy in mechanically ventilated patients reportedly results in significant improvements in functional independence at hospital discharge as well as in reduced durations of mechanical ventilation and delirium.

ANEMIA

Studies have shown that most ICU patients are anemic as a result of chronic inflammation. Phlebotomy also contributes to ICU anemia. A large multicenter study involving patients in many different ICU settings challenged the conventional notion that a hemoglobin level of 100 g/L (10 g/dL) is needed in critically ill patients, with similar outcomes noted in those whose transfusion trigger was 7 g/dL. Red blood cell transfusion is associated with impairment of immune function and increased risk of infections as well as of ARDS and volume overload, all of which may explain the findings in this study. A conservative transfusion strategy has shown similar outcomes in septic shock, post-cardiac surgery, and post-hip surgery patients. A conservative transfusion strategy has been shown to enhance survival among patients with active upper gastrointestinal hemorrhage.

ACUTE KIDNEY FAILURE

(See also Chap. 304) Acute kidney failure occurs in a significant percentage of critically ill patients. The most common underlying etiology is acute tubular necrosis, usually precipitated by hypoperfusion and/or nephrotoxic agents. Currently, no pharmacologic agents are available for prevention of kidney injury in critical illness. Studies have shown convincingly that neither low-dose dopamine, fenoldapam nor vasopressin are not effective in protecting the kidneys from acute injury.

DELIRIUM

(See also Chaps. 24 and 300) This state is defined by (1) an acute onset of changes or fluctuations in mental status, (2) inattention, (3) disorganized thinking, and (4) an altered level of consciousness (i.e., a state other than alertness). Delirium is reported to occur in a wide range of mechanically ventilated ICU patients and can be detected by the Confusion Assessment Method (CAM)-ICU or the Intensive Care Delirium Screening Checklist. These tools are used to ask patients to answer simple questions and perform simple tasks and can be used readily at the bedside. The differential diagnosis of delirium in ICU patients is broad and includes infectious etiologies (including sepsis), medications (particularly sedatives and analgesics), drug withdrawal, metabolic/electrolyte derangements, intracranial pathology (e.g., stroke, intracranial hemorrhage), seizures, hypoxia, hypertensive crisis, shock, and vitamin deficiencies (particularly thiamine). The etiology of a patient’s ICU delirium impacts the prognosis. Those with persistent ICU delirium not related to sedatives have increases in length of hospital stay, time on mechanical ventilation, cognitive impairment at hospital discharge, and 6-month mortality rate. Interventions to reduce ICU delirium are limited. The sedative dexmedetomidine has been less strongly associated with ICU delirium than midazolam. In addition, very early physical and occupational therapy in mechanically ventilated patients has been demonstrated to reduce delirium.

ANOXIC CEREBRAL INJURY

(See also Chap. 301) This condition is common after cardiac arrest and often results in severe and permanent brain injury in survivors. Active cooling of patients after cardiac arrest is controversial, with some studies showing improved neurologic outcomes and others showing no such improvement. Certainly patients suffering cardiac arrest should have a temperature targeted to no higher than 36°C.

STROKE

(See also Chap. 419) Stroke is a common cause of neurologic critical illness. Hypertension must be managed carefully, since abrupt reductions in blood pressure may be associated with further brain ischemia and injury. Acute ischemic stroke treated with tissue plasminogen activator (tPA) has an improved neurologic outcome when treatment is given within 4.5 h of onset of symptoms. The mortality rate is not reduced when tPA is compared with placebo, despite the improved neurologic outcome. The risk of cerebral hemorrhage is significantly higher in patients given tPA. No benefit is seen when tPA therapy is given beyond 4.5 h after symptom onset. Heparin has not been convincingly shown to improve outcomes in patients with acute ischemic stroke. Decompressive craniectomy is a surgical procedure that relieves increased intracranial pressure in the setting of space-occupying brain lesions or brain swelling from stroke; available evidence suggests that this procedure may improve survival among select patients (≤55 years or age), albeit at a cost of increased disability for some.

SUBARACHNOID HEMORRHAGE

(See also Chap. 419) Subarachnoid hemorrhage may occur secondary to aneurysm rupture and is often complicated by cerebral vasospasm, re-bleeding, and hydrocephalus. Vasospasm can be detected by either transcranial Doppler assessment or cerebral angiography; it is typically treated with the calcium channel blocker nimodipine, aggressive IV fluid administration, and therapy aimed at increasing blood pressure, typically with vasoactive drugs such as phenylephrine. The IV fluids and vasoactive drugs (hypertensive hypervolemic therapy) are used to overcome the cerebral vasospasm. Early surgical clipping or endovascular coiling of aneurysms is advocated to prevent complications related to re-bleeding. Hydrocephalus, typically heralded by a decreased level of consciousness, may require ventriculostomy drainage.

STATUS EPILEPTICUS

(See also Chap. 418) Recurrent or relentless seizure activity is a medical emergency. Cessation of seizure activity is required to prevent irreversible neurologic injury. Lorazepam is the most effective benzodiazepine for treating status epilepticus and is the treatment of choice for controlling seizures acutely. Phenytoin or fosphenytoin should be given concomitantly since lorazepam has a short half-life. Other drugs, such as gabapentin, carbamazepine, and phenobarbital, should be reserved for patients with contraindications to phenytoin (e.g., allergy or pregnancy) or ongoing seizures despite phenytoin.

BRAIN DEATH

(See also Chap. 301) Although deaths of critically ill patients usually are attributable to irreversible cessation of circulatory and respiratory function, a diagnosis of death also may be established by irreversible cessation of all functions of the entire brain, including the brainstem, even if circulatory and respiratory functions remain intact on artificial life support. Such a diagnosis requires demonstration of the absence of cerebral function (no response to any external stimulus) and brainstem functions (e.g., unreactive pupils, lack of ocular movement in response to head turning or ice-water irrigation of ear canals, positive apnea test [no drive to breathe]). Absence of brain function must have an established cause and be permanent without possibility of recovery; a sedative effect, hypothermia, hypoxemia, neuromuscular paralysis, and severe hypotension must be ruled out. If there is uncertainty about the cause of coma, studies of cerebral blood flow and electroencephalography should be performed.

WITHHOLDING OR WITHDRAWING CARE

(See also Chap. 9) Withholding or withdrawal of care occurs commonly in the ICU setting. The Task Force on Ethics of the Society of Critical Care Medicine reported that it is ethically sound to withhold or withdraw care if a patient or the patient’s surrogate makes such a request or if the physician judges that the goals of therapy are not achievable. Since all medical treatments are justified by their expected benefits, the loss of such an expectation justifies the act of withdrawing or withholding such treatment; these two actions are judged to be fundamentally similar. An underlying stipulation derived from this report is that an informed patient should have his or her wishes respected with regard to life-sustaining therapy. Implicit in this stipulation is the need to ensure that patients are thoroughly and accurately informed regarding the plausibility and expected results of various therapies.

The act of informing patients and/or surrogate decision-makers is the responsibility of the physician and other health care providers. If a patient or surrogate desires therapy deemed futile by the treating physician, the physician is not obligated ethically to provide such treatment. Rather, arrangements may be made to transfer the patient’s care to another care provider. Whether the decision to withdraw life support should be initiated by the physician or left to surrogate decision-makers alone is not clear. One study reported that slightly more than half of surrogate decision-makers preferred to receive such a recommendation, whereas the rest did not. Critical care providers should meet regularly with patients and/or surrogates to discuss prognosis when the withholding or withdrawal of care is being considered. After a consensus among caregivers has been reached, this information should be relayed to the patient and/or surrogate decision-maker. If a decision to withhold or withdraw life-sustaining care for a patient has been made, aggressive attention to analgesia and anxiolysis is needed.