Admission to the Intensive Care Unit and High-Dependency Step-Down Unit

The need for invasive hemodynamic monitoring and observation remains the most common criterion for admission of selected patients to an ICU or step-down unit. The main distinction between these two environments in most institutions is the severity of critical illness and need for invasive positive pressure ventilation. These environments can provide patients the following: decreased myocardial oxygen demand through expeditious rewarming, effective fluid resuscitation, effective analgesia, meticulous control of hemodynamics, and careful monitoring and appropriate nursing care to aid in the early diagnosis and treatment of complications. In a cohort study of patients after elective aortic surgery, Lawlor et al² demonstrated that care in an ICU decreased mortality in patients with preoperative severe coronary artery disease defined as an ejection fraction of less than 40%, congestive heart failure, or New York Heart Association class III or IV angina. Consideration of direct admission to the ICU should also be given to patients with significant chronic obstructive pulmonary disease (COPD, defined as forced expiratory volume in 1 second of less than 1 L) and those who are dialysis dependent or have a need for continuous renal replacement therapy. Hemodynamically unstable patients (systolic blood pressure <90 mm Hg or inotropic support at the end of the operation), intubated patients with ongoing need for ventilatory support, those with clinically significant perioperative cardiac ischemia, patients with pulmonary artery catheters, patients with spinal drains, patients with hypothermia (temperature <35° C), and those who required massive transfusion of blood products (>3 L) during the procedure should be admitted to the ICU.

Selective use of high-dependency step-down units and rationing of ICU resources can be safe and cost-effective with appropriate patient risk stratification.^3,4 Each patient must be individualized according to preoperative comorbid diseases, perioperative hemodynamic stability, and surgical procedure performed. Patients undergoing less invasive vascular procedures associated with fewer fluid shifts and lower surgical morbidity can be safely managed in a step-down unit. This is the case with patients undergoing peripheral arterial bypass, elective repair of infrarenal AAAs, and carotid endarterectomy. Several scoring models predicting risk in vascular patients have been studied and validated as predictors of high-risk patients undergoing vascular surgery.⁵

Organizational Structure of the Intensive Care Unit

The organizational structure of the ICU, independent of a patient’s medical factors, has a direct impact on outcomes.^6–11 In looking at outcomes after abdominal aortic surgery, Pronovost et al^7,12 have shown a relationship between increased in-hospital mortality and the absence of a full-time ICU director, less than 50% of ICU attending physicians certified in critical care, no daily rounds by an ICU physician, and decreased ICU nurse-patient ratio in the evening (<1 : 2). In patients undergoing abdominal aortic surgery, intensivists not making routine ICU rounds is associated with an increase in hospital mortality (odds ratio [OR], 3.0; 95% confidence interval [CI], 1.9-4.9), cardiac arrest (OR, 2.9; 95% CI, 1.2-7.0), acute renal failure (OR, 2.2; 95% CI, 1.3-3.9), septicemia (OR, 1.8; 95% CI, 1.2-2.6), platelet transfusion (OR, 6.4; 95% CI, 3.2-12.4), and reintubation (OR, 2.0; 95% CI, 1.0-4.1). Dang et al⁶ also demonstrated that ICUs with decreased nurse staffing had increased rates of cardiac and respiratory complications after abdominal aortic surgery.

Central Venous Catheters and Central Venous Pressure

Central venous catheters are used primarily to infuse fluids, to administer vasoactive drugs, and to assess intravascular volume. The most common sites for placement of central venous catheters are the internal jugular veins, subclavian veins, and femoral veins. To measure CVP, internal jugular or subclavian central venous catheters must be positioned with the catheter tip in the distal segment of the superior vena cava. More distal placement in the right atrium is associated with potential risk for erosion, perforation, and cardiac tamponade. CVP measures right atrial pressure, providing an estimation of preload. This measurement, however, is inaccurate in patients with COPD or in those with valvular heart disease. Femoral lines are good for resuscitation but not for CVP measurements because of the effect of intra-abdominal pressure on measurements.

The transducer must be zeroed at the level of the midaxillary line, and normal values of CVP are between 6 and 12 mm Hg. In patients being managed with positive pressure ventilation, CVP should be measured at end-expiration, when pleural pressure in the chest is approximately zero.¹⁸ The addition of positive end-expiratory pressure (PEEP) in a ventilated patient is another important consideration in measuring pressure in the chest. In normal lungs and with low levels of PEEP (between 5 and 8 cm H₂O), intrapleural pressure, which affects CVP, is only slightly affected.¹⁸ The effect on CVP of high levels of PEEP (>12 to 15 cm H₂O) when the lungs are probably not compliant is unclear.¹⁹ High levels of PEEP prevent backflow of fluid outside the chest to the heart, thus decreasing venous return. If sudden changes in CVP occur, one should consider a cardiorespiratory cause, such as pneumothorax or cardiac tamponade. Further discussion of CVP monitoring can be found in Chapter 32.

Peripheral Arterial Lines

Peripheral arterial catheters are the “gold standard” for direct assessment of systolic blood pressure, and they allow direct vascular access for blood sampling. Complications include bleeding, hematoma, line-related infection, thrombosis, and, rarely, limb ischemia.²⁰ These complications occur to varying degrees, depending on the cannulation site. The most common site for cannulation is the radial artery, followed by the femoral artery, axillary artery, brachial artery, and, less commonly, dorsalis pedis and ulnar arteries. The radial artery and the femoral artery have similar cannulation complication rates,²¹ and some prefer the femoral artery as the primary cannulation site because of ease of insertion. The most common complication of radial artery cannulation is temporary arterial occlusion, which occurs approximately 20% (1.5% to 35%) of the time, but the occlusion is temporary in most and rarely causes acute ischemia.²⁰ In contrast to the radial artery, cannulation of the femoral artery is associated with a higher risk of bleeding (1.58% vs 0.53%) and pseudoaneurysm (0.3% vs 0.09%) but a lower risk of thromboembolism.²⁰ Cannulation of the axillary artery has a complication rate similar to that at other sites, and it may be the only site available for arterial access in some patients. In the ICU, however, it is not commonly used because of lack of familiarity with the approach and the potential, albeit unproven, risk for carotid embolization.²⁰ The level of the transducer relative to the pressure being measured will determine the measured value of the arterial pressure, thus making its position critical for obtaining an accurate measurement. The standard level is the midpoint of the right atrium, approximately 5 cm below the sternal angle in the midaxillary line. This is the area where the preload pressure of the heart is determined.²²

The distance of the measuring catheter from the heart, the length and compliance of the tubing, and the presence of air bubbles affect systolic and diastolic measurements in the system. These variables affect systolic and diastolic pressure proportionately in opposite directions and have no net effect on the measurement of mean arterial pressure (MAP). For this reason, measurement of MAP is a more accurate reflection of mean aortic pressure. MAP is manually calculated with the formula MAP = [(2 × diastolic blood pressure) + systolic blood pressure]. Electronic monitoring systems, however, measure MAP as the area under the pulse wave, often averaged over three or more cycles.

Pulmonary Artery Catheters

Pulmonary artery catheters provide central hemodynamic measurements, waveform tracings, and specific blood gas calculations such as mixed venous oxygen and carbon dioxide saturation (also see discussion of pulmonary artery catheter use in Chapter 32). Pulmonary artery catheters were initially used to assess and to treat patients with acute myocardial infarction (MI); however, the use of pulmonary artery catheters has declined significantly because several randomized trials have failed to demonstrate a reduction in mortality.^23–25 Pulmonary artery catheters are typically inserted in the ICU in situations in which the clinician believes that measuring cardiac output, cardiac pressure, and mixed venous oxygen saturation will aid in guiding therapy and for direct assessment of the effect of and response to various treatments, such as fluids and vasopressors.²⁶ Included in this category are patients in cardiogenic or refractory shock and patients in whom right ventricular dysfunction, such as right ventricular infarction, is suspected.^18,27 Pulmonary artery catheters may induce arrhythmias, which are usually transient and not life-threatening.²⁸ Left bundle branch block is a relative contraindication to the use of pulmonary artery catheterization because it can induce a temporary right bundle branch block and the potential for subsequent complete heart block.²⁸ Other catheter-related complications include catheter knotting, pulmonary artery rupture, catheter fragmentation, and cardiac rupture secondary to forceful insertion. Long-term placement, more than 72 to 96 hours, can also lead to thrombosis and infection.

Echocardiography

Echocardiography is used to assess cardiac function by measuring ventricular contractility, chamber size, valve function, and flow.^29,30 Global assessment of ventricular function with transthoracic echocardiography can be helpful in guiding hemodynamic management because one can easily differentiate between left and right ventricular dysfunction and adjust treatment appropriately.³¹ For instance, a patient with a poorly functioning left ventricle may benefit from an inotropic agent, whereas the presence of a hyperdynamic left ventricle and evidence of tissue hypoxia may indicate the need for fluid challenge and vasopressors.²⁹ Echocardiography is helpful in the case of normotensive shock in a patient who has been adequately fluid resuscitated where it may be necessary to use afterload-reducing agents such as nitroglycerin.³² Calculation of ejection fraction, pulmonary artery pressure, and cardiac output is also possible, but specific expertise is required, and such measurements may not be reliable in ICU settings.³¹

Transesophageal Doppler ultrasound is a noninvasive modality for measuring cardiac output.³³ Cardiac output is calculated by using the diameter of the descending aorta, the distribution of flow to the descending aorta, and the measured flow velocity of blood.¹⁵ The accuracy of the calculated cardiac output may be affected by the position of the ultrasound probe, which is inserted blindly in the esophagus. Poor positioning of the probe most commonly results in an underestimation of the cardiac output.¹⁵

Intra-Abdominal Pressure

Abdominal compartment syndrome is defined as clinically relevant organ dysfunction caused by intra-abdominal hypertension. Intra-abdominal pressure (IAP) is the steady-state pressure concealed within the abdominal cavity. For most critically ill patients, an IAP of 5 to 7 mm Hg is considered normal.³⁴ Morbidly obese and pregnant individuals may have chronically elevated IAP without adverse consequences. Intra-abdominal hypertension is defined as IAP in excess of 12 to 20 mm Hg.^35,36 Abdominal perfusion pressure is calculated as MAP minus IAP. Elevated IAP may reduce blood flow to the abdominal viscera. Abdominal compartment syndrome is diagnosed by the combination of intra-abdominal hypertension and evidence of organ malperfusion. Specifically, it is a constellation of symptoms consisting of new attributable organ dysfunction, such as cardiovascular (decreased cardiac output), respiratory (high peak airway pressure), and renal (oliguria) malperfusion, in the context of sustained IAP of 20 mm Hg or higher.³⁵ Patients at increased risk for abdominal compartment syndrome are those who received large volumes of fluid for resuscitation (>10 L of crystalloid, >5 L of colloid) or transfusion of more than 10 units of packed red blood cells (PRBCs) during a 24-hour period.³⁶

IAP is most commonly measured by the intravesicular technique.³⁷ An indwelling bladder catheter is connected to a pressure transducer. A defined volume of fluid (25 to 50 mL of normal saline) is instilled into the catheter and used to distend the bladder and eliminate bladder wall coaptation.³⁶ Volumes greater than 50 mL falsely elevate IAP values. IAP is then measured at end-expiration with the patient in the supine position and the manometer or other pressure measuring system at the level of the symphysis pubis. Another technique of IAP measurement uses a nasogastric or orogastric tube.³⁸ This technique involves a lower risk for sepsis secondary to iatrogenic urinary tract infection as a result of catheter manipulation, but it is a more cumbersome method and rarely used. The timing of IAP monitoring and the frequency of monitoring are not standardized, but the highest risk for the development of abdominal compartment syndrome is during and shortly after resuscitation. The main reason for monitoring of IAP is that early recognition and treatment of abdominal compartment syndrome appear to improve survival.³⁶

Hypertension

Hypertension is common after surgery and may be caused by hypoxia, hypercapnia, hypervolemia, hypothermia, gastric or bladder distention, agitation, and pain. Furthermore, a common cause of postoperative hypertension is simply neglecting to restart preoperative antihypertensive medications. Sedation, analgesia, and rewarming often resolve most cases of mildly elevated blood pressure. However, hypertension increases myocardial oxygen consumption and can potentially cause myocardial ischemia. In addition, in the immediate postoperative period, it may contribute to increased bleeding from raw surfaces or through vascular anastomoses. After all possible causes have been evaluated and addressed, drug treatment may be considered. In general, blood pressure treatment targets should center on systolic blood pressure values 20 mm Hg above or below preoperative pressure.³⁹ Deviations in MAP greater than 20% from preoperative values should also be treated.^40,41 Several agents may be used, and the surgeon should become familiar with a few for an effective and safe approach to hypertensive patients in the postoperative period. Nitrates, beta blockers, angiotensin-converting enzyme inhibitors, calcium channel blockers, and vasodilators are the most common classes of drugs used (see Chapter 30).⁴²

Hypertensive crisis is acute end-organ damage (heart, brain, retina, and kidney) associated with systolic blood pressure higher than 179 mm Hg or diastolic blood pressure higher than 109 mm Hg. In the absence of evidence of end-organ damage, these elevations in blood pressure constitute what are called hypertensive emergencies.⁴³ When hypertensive emergencies are diagnosed, short-acting intravenous antihypertensive agents should be used; the reduction in blood pressure should occur within 1 hour in hypertensive crisis and within 24 hours in hypertensive emergencies.⁹ The goal is to reduce the patient’s diastolic blood pressure to less than 110 mm Hg during a period of 30 to 60 minutes.⁴⁴ Hypertensive patients in these circumstances will manifest natriuresis, which causes intravascular fluid depletion requiring fluid administered along with antihypertensive medications. Several agents can be used in hypertensive emergencies; one of the most common and effective is labetalol because it is a combined α₁-adrenergic and nonselective β-adrenergic receptor blocker.⁴⁵

Hypotension

Hypotension is dangerous and is associated with an increased risk for end-organ dysfunction (MI, cerebrovascular events, and renal failure) and possibly bypass graft thrombosis. Common causes of hypotension after surgery are hypovolemia, cardiac dysfunction, and a diffuse vasodilatory state with or without sepsis. The two management strategies for correction of hypotension involve administration of fluids and vasoactive agents. A fluid challenge should be the first line of treatment of postoperative hypotension. Surgical bleeding should be excluded as a cause to prevent delays in returning the patient to the operating room.⁴⁶

When administering fluids for resuscitation, one must determine whether the patient is on the ascending portion of the Frank-Starling curve and will benefit from a fluid challenge. If the patient is on the flat portion of the curve, fluid loading will only increase tissue edema and worsen tissue dysoxia with little effect on cardiac output.¹⁵ A large pulse pressure–stroke volume ratio (>10%:15%) is indicative of hypovolemia and predictive of volume responsiveness.¹⁵ A fluid challenge in these circumstances will increase preload and subsequently increase stroke volume.⁴⁷ Similarly, in patients with a pulse pressure variation greater than 12%, fluid challenge is likely to have a positive effect on cardiac output.⁴⁸ If the hypotension does not respond to fluid resuscitation, cardiac output must be optimized through pharmacologic means. Short-acting peripheral vasoconstrictors, such as phenylephrine and dopamine, are readily available on crash carts and can be given peripherally and titrated to effect. Once patients have adequate central intravenous access, dopamine should be switched to more definably titratable vasopressors.

In cardiogenic shock, peripheral vasoconstrictors may increase myocardial workload and dysfunction. In this setting, inotropic agents may be necessary to increase cardiac output. Patients in cardiogenic shock secondary to acute MI have mortality rates higher than 50%.⁴⁹ In these situations, volume must be optimized, rhythm must be controlled, and beta agonists should be administered to enhance contractility.⁵⁰ Finally, intra-aortic balloon pump counterpulsations may be necessary to aid the heart in generating cardiac output when pharmacologic agents fail. Beta blockers and calcium channel blockers have a negative inotropic effect and are relatively contraindicated in this setting.

Arrhythmias

Patients with preexisting structural heart disease are at highest risk for postoperative arrhythmias.⁵¹ Common triggers for postoperative arrhythmias are hypoxia, hypercapnia, acid-base imbalances, electrolyte abnormalities, and myocardial ischemia. Treatment must focus on predisposing factors, and the goal should be hemodynamic stabilization, control of the ventricular response, and restoration of rhythm.⁵¹

Postoperative Myocardial Infarction

Ventilatory Support

Tracheostomy

Tracheostomy relieves the patient of the portion of the work of breathing required to overcome the resistance of the endotracheal tube and the upper airway. It also allows better oral hygiene and secretion management and improves patient comfort. The major disadvantages of a tracheostomy are procedure-related complications, stomal complications, formation of a tracheoinnominate fistula, and formation of a tracheoesophageal fistula. Early tracheostomy within the first week in patients who are thought to require ventilation for more than 10 days facilitates weaning and is associated with shorter length of stay in the hospital.⁸⁶ Jubran et al⁷⁹ found that in patients who required prolonged mechanical ventilation, unassisted breathing through a tracheostomy resulted in shorter median weaning time compared with pressure support. However, in a retrospective observational study after abdominal and thoracoabdominal repair, Diedrich et al⁸⁷ found that overall, tracheostomy correlated with worse outcomes but that patients with preexisting COPD who received a tracheostomy had improved survival.

Tracheostomy can be done either by an open surgical procedure or percutaneously. Percutaneous tracheostomy is as safe as open surgical tracheostomy, with the advantage that the patient does not require transport to the operating room, and operating room resources are thus conserved.⁸⁸ Furthermore, no difference in long-term outcomes between percutaneous and open surgical tracheostomy has been shown.⁸⁸ However, percutaneous tracheostomy is relatively contraindicated in patients who are hypoxic and have high PEEP requirements, in patients who are obese with short necks, in patients who are coagulopathic, and in those with recent (<10 days) anterior cervical spine fixation.⁸⁹

Evaluation of Bleeding Patients

Patients with greater than expected fluid requirements postoperatively must be urgently evaluated for bleeding.⁴⁶ The physiologic response to hypovolemia includes tachycardia, hypotension, low CVP, decreased urine output, and signs of peripheral vasoconstriction.⁹⁰

Infusion of Fluids

Fluid administration in the postoperative period needs to be tailored to the individual patient. A balance must be reached between adequate tissue perfusion by replacing fluid loss at surgery and avoidance of fluid overloading. Hypervolemia is as detrimental to the patient as hypovolemia because it causes unwanted tissue edema and impedes tissue oxygenation. If large volumes of crystalloid solution are administered for resuscitation, it is recommended that a balanced salt solution such as lactated Ringer’s solution be given instead of normal saline to reduce the development of iatrogenic hyperchloremic metabolic acidosis. Restricted fluid regimens based on physiologic response versus standard fluid administration based on fluid algorithms and fixed rates have been associated with improved outcomes and decreased length of hospital stay after major gastrointestinal surgery and decreased pulmonary morbidity after thoracic surgery.^94,95 I favor the administration of intravenous fluids in the postoperative period at low maintenance rates of 75 to 100 mL/h and small boluses of crystalloid (500 to 1000 mL) as needed, with titration to low blood pressure, increased heart rate, or low urine output (Fig. 33-2). Colloids are commonly used for fluid resuscitation because they have the theoretical benefit of expanding intravascular volume and because their osmotic activity reduces third spacing. However, there appears to be no survival benefit when albumin is used in severely ill patients as well as no difference among the different types of colloid used.⁹⁶

Transfusion of Blood Products

For a summary of blood product transfusion, see Table 33-2. The evidence for guiding transfusion practices is sparse with respect to the timing and volume of product administration and the ideal ratio of red blood cells (RBCs), platelets, and FFP. It is accepted that giving a disproportionate amount of one product over another can exacerbate dilutional coagulopathy. In addition, relying on laboratory tests to guide transfusion may be misleading because of the delay in obtaining results and because normal levels do not equate to normal function. Additional discussion of the following can be found in Chapters 32 and 37.

Adjuncts in Hemostasis

Several adjuncts may be used with the aim of improving hemostasis. Recombinant activated factor VII (VIIa) has been used to control bleeding in patients with hemophilia type A and type B. Recombinant factor VIIa also generates thrombin at the site of vessel injury and induces coagulation.⁹⁸ It has been used in the treatment of patients with hemorrhage from blunt trauma, thrombocytopenia, warfarin overdose, obstetric bleeding, and intracranial bleeding.^98,99 It is no longer recommended for off-label use in patients without hemophilia because of increased thromboembolic arterial events and lack of mortality benefit.¹⁰⁰

Antifibrinolytic agents such as tranexamic acid and aminocaproic acid have been successfully used to treat post­operative hemorrhage.¹⁰¹ These synthetic lysine analogues inhibit plasminogen and plasmin-mediated fibrinolysis.⁹¹ Similarly, serine proteases such as aprotinin, which neutralize trypsin, plasmin, and kallikrein, have been used when fibrinolysis contributes to bleeding.¹⁰² These agents have been shown to decrease transfusion of blood products in cardiac, hepatic, and orthopedic surgery.¹⁰² Aprotinin has also been used prophylactically in coronary artery bypass surgery, but there has been concern because of an increased incidence of renal failure and increased risk for MI in these patients that led to removal from the market for a time.⁹⁸ Finally, DDAVP, a synthetic vasopressin analogue that stimulates endothelial release of factor VIII and von Willebrand factor, has been used to enhance platelet aggregation.¹⁰³ It is used mainly in patients with hemophilia and those with both congenital and acquired platelet disorders, but its value in surgical hemostasis is unproven. Thus, its use in post­operative patients who do not have bleeding disorders is not recommended.⁹⁸

Complications of Massive Transfusion

The definition of massive transfusion is replacement of 50% of a patient’s blood volume in 3 hours or one blood volume in 24 hours.¹⁰⁴ The major complication of massive transfusion is the development of coagulopathy from multiple causes, including hemodilution, hypothermia, acidosis secondary to tissue hypoxia, and DIC.¹⁰⁴ Hemodynamically unstable patients are initially resuscitated with crystalloids, which are acellular and dilute the blood components, thereby aggravating the coagulopathy. Hypothermia associated with the infusion of large volumes of cold solution also inhibits the enzymatic coagulation cascade, synthesis of coagulation factors, and platelet function.¹⁰⁵ Administration of large volumes of fractionated blood products also causes coagulopathy by complex mechanisms, some known and some unknown: dilution of coagulation factors (particularly factors V and VIII : C, which have a limited shelf life), hypocalcemia because of the citrate used as an anticoagulant, and platelet dysfunction as a result of aging and metabolic alterations caused by storage.⁹⁰ Furthermore, when RBCs are stored, their extracellular potassium levels increase over time, so patients receiving large volumes of RBCs are at risk for hyperkalemia.

One of the most common sequelae of massive transfusion is acute lung injury secondary to congestion caused by overload, alveolar inflammation, and increased permeability.¹⁰⁶ Passive transfer of leukocyte antibodies in donated blood resulting in immune-mediated transfusion-associated lung injury can also occur.¹⁰⁷ Finally, massive transfusion has been described as an immunosuppressant leading to sepsis and sepsis-induced lung injury.¹⁰⁶ (See Chapter 32 for further discussion of transfusion-associated lung injury.)

Gastrointestinal Ischemia

Postoperative gastrointestinal ischemia is a rare but potentially lethal postoperative complication. Colonic ischemia occurs in 2% of patients after elective AAA repair and carries an overall mortality of 40% to 60% if it is transmural.¹¹³ The incidence of ischemic colitis in patients undergoing repair of a ruptured AAA is reported to be 15% to 65%.¹¹⁴ A high index of suspicion for colonic ischemia is necessary after aortic repair in patients with persistent acidosis and shock with no other obvious cause, greater than expected fluid requirements, new onset of gastrointestinal bleeding, or simply bloody bowel movements. Laboratory investigations often demonstrate anion gap acidosis and leukocytosis. In advanced cases, a plain abdominal radiograph may demonstrate free air in the peritoneal cavity or in the portal venous system. Abdominal computed tomography may demonstrate evidence of pneumatosis intestinalis. Bedside flexible sigmoidoscopy or colonoscopy can be diagnostic of colon ischemia. Stable patients with colonoscopic results suggesting ischemia limited to the mucosa may be treated conservatively with hydration and broad-spectrum intravenous antibiotics. However, immediate exploration is warranted in such patients if the condition worsens or hemodynamic instability ensues. Unstable patients with evidence of bowel ischemia should undergo urgent surgical exploration to reduce mortality. In selected cases of colon ischemia that has not progressed to gangrene, reimplantation of the inferior mesenteric artery can be considered. In the setting of frank gangrene, the colon should be resected and a colostomy created.

Nutrition

The nutritional state of critically ill patients has a direct effect on outcome. This is due to the catabolic stress state that critical illness creates. Malnutrition in this patient population is associated with increased morbidity and mortality as well as increased length of stay. In the general medical and surgical ICU population, it has been shown that commencing enteral nutrition within 24 to 48 hours of ICU admission is associated with improved outcomes.¹¹⁵ It has been associated with reduced infections, increased protection against stress ulceration as well as intestinal ischemia, maintenance of gut flora, decreased catabolism, and decreased mortality.¹¹⁶ Nutritional support in the ICU has had the goals of maintaining lean body mass to preserve immune function and to avert metabolic complications.¹¹⁷ This includes delivery of both micronutrients and macronutrients and meticulous glycemic control.¹¹⁷ Early enteral nutrition has been shown to be safe in patients admitted to the ICU after major abdominal surgery.¹¹⁸

Several studies have also shown that patients undergoing major vascular surgery have a significant incidence of preoperative and postoperative malnutrition.¹¹⁹ Patients undergoing surgery for AAA have higher nutritional depletion and increased rates of postoperative infection compared with patients having carotid or peripheral vascular surgery.¹²⁰ Interestingly, after AAA surgery, it has also been demonstrated that small bowel function returns to normal by 18 hours and starting enteral nutrition within 3 hours of surgery can promote small intestinal activity compared with fasting. This suggests that enteral nutrition may in fact aid the return of normal bowel function.¹²¹ As suggested by others, it is my practice to initiate oral or enteral feeding within the first 48 hours after AAA surgery, even in the absence of traditional clinical markers for readiness to feed (presence of bowel sounds, flatus, bowel movement, or volume of gastric aspirate).¹²² Most ICUs use enteral nutrition protocols and guidelines for critically ill patients for this reason.¹¹⁶ Although there is a paucity of studies of the applicability of these protocols specifically to vascular patients, both our vascular surgeons and intensivists think that these guidelines are safe and beneficial in this patient population.¹²³

There are several newly updated North American and European guidelines available to help guide assessment of nutrition and support therapy in critically ill adult patients.^116,117 These patients represent a heterogeneous population, and the nutritional needs of patients should be individualized. These guidelines are intended for patients who are expected to require more than 2 or 3 days of ICU care and have had significant metabolic or surgical stress. They are not intended for those patients who require only temporary monitoring or who have had minimal surgical stress. Some of the general highlights of the guidelines are reviewed here.

Nutrition in the ICU can be subdivided between dosing and route of administration. Several predictive equations exist that are used to estimate calorie requirements for patients in the ICU. However, these predictive equations are often less accurate than indirect calorimetry. These predictive equations are also more problematic in obese patients in making accurate assessments of nutritional goals. Once goal calories are determined, patients should be given 50% to 65% of goal calories in their first week of ICU admission, thereafter increasing to goal rate. In obese patients (body mass index >30) with critical illness, permissive underfeeding or hypocalorie feeding with enteral nutrition is recommended. The target in obese patients should remain at 60% to 70% of ideal weight.

The two main routes of administering nutrition in critically ill patients are the enteral route and the parenteral route. As a general rule, the enteral route has the most benefit for the patient. In patients who are unable to eat, enteral nutrition is preferred to parenteral. As mentioned before, feeding should be started within the first 24 to 48 hours of admission to the ICU and advanced toward goal during the next 48 to 72 hours. Enteral feeding can be started through either a gastric tube or small bowel tube in the absence of traditional markers of bowel function (bowel sounds, flatus, or passage of stool). Patients are often challenged with trickle feeding at 10 mL/h to ensure tolerance before increasing rapidly toward goal.

It is recommended that either parenteral or enteral nutrition be provided to critically ill patients who are unable to voluntarily maintain their nutritional needs. In general, enteral feeding is preferred to parenteral feeding because of superior nutritional efficacy and better safety profile. Enteral feeding through a nasogastric tube or small bowel feeding tube should be started within the first 24 to 48 hours after admission to the ICU and continued until adequate oral intake is firmly established. Placement of a feeding tube in the proximal small bowel should be considered in patients with inadequate gastric emptying as evidenced by residual gastric volumes in excess of 500 mL. If early enteral feeding is not feasible, parenteral nutrition should be considered if the anticipated duration of nutritional support will exceed 7 days. Shorter periods of parenteral nutrition (e.g., <5-7 days) expose the patient to a variety of vascular access, metabolic, and infectious risks without proven benefit. Even in patients stabilized on parenteral nutrition, periodic efforts should be made to start enteral feeding. Parenteral nutrition should not be stopped until enteral nutrition is providing at least 60% of target requirements. Recommended nutrition doses based on guidelines are up to 20 to 25 kcal/kg actual body weight per day in the acute phase of critical illness (72-96 hours) and 25 to 30 kcal/kg body weight per day in the recovery phase.¹¹⁶

Pain Management

Effective postoperative analgesia is important and may be associated with a reduction in the incidence of myocardial ischemia after surgery.⁶² This beneficial effect does not appear to be related to changes in heart rate, blood pressure, or requirement for postoperative vasoactive agents compared with patients with inadequate pain control.⁶² In nonintubated patients who are able, patient-controlled analgesia is preferred to intermittent on-demand narcotic dosing because patients can be in a significant amount of pain while waiting for the next narcotic dose. Epidural analgesia can also be effective in the postoperative period, and it offers the potential benefit of reducing the dose of opiates, which improves gastric emptying and ileus and increases patient mobility.¹²⁴ Patients who are mobilized earlier can benefit from reduced thrombotic events and pulmonary complications. However, care must be taken to prevent significant hypotension, which can be caused by the administration of epidural anesthesia. The balance between analgesia and sedation is discussed in the section on sedation and delirium in the ICU.

Deep Venous Thrombosis Prophylaxis

The etiology, treatment, and prophylaxis of deep venous thrombosis (DVT) are reviewed in depth in Chapter 51. Thus, only issues relevant to the prophylaxis and treatment of DVT after vascular surgical procedures are discussed here. Because most vascular surgery involves the intraoperative administration of heparin, it has been suggested that these patients do not carry the same postoperative risk for DVT as other surgical patients do. However, the incidence of DVT after aortic surgery has been reported to be between 2% and 18%,¹²⁵ and therefore all patients undergoing vascular surgical procedures should receive postoperative DVT prophylaxis.

General measures in postoperative care, such as elevation of the lower extremities, early ambulation, and leg exercises, may reduce the incidence of DVT.¹²⁶ Graduated compression stockings, intermittent pneumatic compression, and venous foot pumps have also been shown to decrease the incidence of DVT, but they have not been proved to prevent fatal pulmonary embolism when used as monotherapy.¹²⁶ It is also generally accepted that aspirin is not effective in the prevention of DVT.¹²⁷ In contrast, treatment with either low doses of subcutaneous unfractionated heparin (5000 units subcutaneously two or three times per day) or subcutaneous low-molecular-weight heparin at prophylactic doses is effective and safe in the prevention of postoperative DVT.¹²⁶ The potential benefit to prophylaxis with low-molecular-weight heparin is a reduced rate of heparin-induced thrombocytopenia. Prophylaxis is generally given until patients become ambulatory and mobile. This topic is extensively reviewed in Chapter 51.

Bridging Postoperative Oral Anticoagulation

Patients taking oral anticoagulants in the preoperative period require specific considerations. The most common strategy is to interrupt the oral anticoagulation and to use either subcutaneous low-molecular-weight heparin or intravenous unfractionated heparin.¹²⁸

The American College of Chest Physicians has published guidelines with regard to patient risk for thromboembolic events.¹²⁹ Patients are categorized as being at low, moderate, or high risk. Bridging with full-dose anticoagulation is recommended only for patients at high risk. Such patients include those with mechanical mitral valves or old-model mechanical aortic valves. However, in patients with low stroke and thromboembolic risk, such as those with AF and no other risk factors for stroke such as rheumatic valve disease, bridging may be unnecessary and potentially harmful.¹³⁰ In these patients, oral anticoagulation can be stopped several days before the procedure and restarted several days after it. Although limited data regarding benefit are available, prophylaxis-dose heparin may be used for these patients at low risk.

In patients taking oral anticoagulants who are at high risk for perioperative stroke, such as those with mechanical mitral valves or old-model mechanical aortic valves, oral anticoagulants should be stopped 4 days before and heparin started 2 days before surgery, when the oral agent is assumed to become subtherapeutic.¹²⁸ In the absence of postoperative bleeding, bridging doses of heparin can be restarted 12 to 24 hours after the surgical procedure. If unfractionated heparin is used, it should be restarted without a bolus dose. However, the risk of bleeding caused by restarting anticoagulation after major vascular surgery must be carefully considered. If anticoagulation is delayed because the risk of bleeding is deemed too high, prophylactic doses of heparin may be given.¹²⁸ After minor surgical procedures, oral anticoagulation can be restarted the evening of surgery. Regardless, clinicians must decide on a strategy that balances stroke risk with bleeding risk in each individual patient.

Alcohol Withdrawal Syndrome

Alcohol is the most abused drug worldwide.¹³¹ Autonomic hyperactivity, including tremulousness, sweating, nausea, vomiting, anxiety, and agitation, appears within the first 24 to 48 hours after withdrawal in patients who abuse alcohol.¹³¹ Symptoms can peak between the third and fifth postoperative days. Serious complications that may develop are delirium tremens and grand mal epileptiform seizures. Delirium tremens, which occurs in less than 5% of patients abusing alcohol, is associated with an increase in postoperative mortality¹³² and is characterized by fever, tachycardia, hypertension, tremors, diaphoresis, hallucinations, disorientation, agitation, and urinary incontinence. Physicians must exclude other life-threatening causes of confusion, such as myocardial ischemia, congestive heart failure, hypotension, pulmonary embolism, bleeding, stroke, metabolic and electrolyte disorders, and sepsis. Before sedation is initiated for presumed alcohol withdrawal, hypoxia must be ruled out because hypoxia is the most common cause of confusion and agitation in the postoperative period.

Patients can be identified preoperatively to be at risk by the quantity and frequency of their alcohol consumption. Preoperative detoxification or abstinence before elective surgery may prevent adverse postoperative complications, but data in this regard are lacking. Patients known to be at risk should be administered prophylactic doses of benzodiazepines throughout the entire perioperative period to prevent the development of alcohol withdrawal syndrome.¹³² Patients who regularly consume alcohol at dependency levels (determined by how alcohol consumption affects relationships, jobs, educational goals, and family life and not by the specific quantity consumed) have been shown to have increased postoperative morbidity and mortality.¹³³

First-line treatment of patients with alcohol withdrawal syndrome is benzodiazepines.¹³⁴ They mimic the central nervous system effects of alcohol on γ-aminobutyric acid receptors and suppress the withdrawal response.¹³⁵ Alcohol replacement therapy with oral or intravenous alcohol is still used but is controversial and has not been shown to be superior to benzodiazepine treatment.¹³⁶ Treatment of severe acute alcohol withdrawal may require intubation and admission to the ICU. Beta blockers may also be used to reduce sympathetic overactivity and cardiac complications. Seizures should be treated with intravenous benzodiazepines. All patients should also be given daily thiamine and multivitamins to prevent Wernicke-Korsakoff syndrome.

Sedation and Delirium in the Intensive Care Unit

Emphasis on sedation and analgesia in critically ill patients has recently come to the forefront of patient care in the ICU. The primary goal is achieving adequate pain control and anxiolysis with appropriate sedation and analgesia. Unfortunately, in critically ill patients on mechanical ventilation, this is very challenging and can lead to oversedation, delirium, and harm to the patient.¹³⁷ It is very important to titrate the medications used for analgesia and sedation in a goal-directed manner to safely provide comfort without impeding recovery, affecting weaning from the ventilator, and prolonging length of stay. Several sedation-monitoring systems have been developed in the ICU, such as the Sedation Agitation Scale, the Motor Activity Assessment Scale, and the Richmond Agitation-Sedation Scale (RASS).¹³⁷ A multidisciplinary team at Virginia Commonwealth University in Richmond has developed the RASS, and it is a validated method used to avoid oversedation in the ICU. It is unique from other methods of evaluation in that it separates verbal from physical stimulation in generating the score, thereby assessing the patient’s arousal with varying potency of stimulus.¹³⁸ The RASS assesses consciousness by use of a 10-point scale from −5 to +4 (−5, unarousable, to 0, alert and calm, to +4, combative) with three defined steps that incorporate the duration of eye contact to determine arousal and content of thought.¹³⁹ The entire assessment takes approximately 20 seconds.¹³⁷ The first step is simply to observe the patients; if they are alert, restless, or agitated, they receive a score of 0 to +4 accordingly. If they are not spontaneously alert, they are called to look at the rater, with duration of eye contact rated (more or less than 10 seconds) and scored from −1 to −3 accordingly. If patients do not respond to verbal stimulation, they are stimulated physically and scored −4 or −5.¹³⁹ However, if patients are calm but not alert before verbal and physical stimulation, they are rated from −1 to −5 even if they become agitated on stimulation.

Ely et al¹³⁷ compared the RASS with the Glasgow Coma Scale, which is a trauma-specific consciousness assessment tool, and demonstrated excellent correlation and discrimination between the scores in assessing levels of consciousness. They also compared RASS with cumulative lorazepam, propofol, fentanyl, and morphine doses of the prior 8-hour and 24-hour periods before RASS assessment. Interestingly, they found significantly different correlations with RASS between opiates (analgesics) and benzodiazepines (sedatives), with higher likelihood of reintubation with deeper sedation measured by RASS. There was further agreement in this study that RASS provided a consensus target for goal-directed sedation and improved team communication regarding sedation management.

Consciousness is considered a combination of level of arousal plus the content of that consciousness, such as delirium.¹⁴⁰ Unaddressed pain and anxiety in critically ill patients has been shown to lead to increased morbidity and mortality as well as posttraumatic stress disorder long term.¹⁴¹ RASS is used to assess the level of arousal, but another key to patient outcome is delirium, a key feature of which is the presence or absence of inattention. Because the RASS incorporates eye contact as part of its assessment, it also correlates with the onset of inattention and delirium. The Confusion Assessment Method for the Intensive Care Unit (CAM-ICU) is a delirium-monitoring tool that complements the RASS and is used to assess acute brain dysfunction.¹⁴² Whereas RASS measures the level of consciousness, CAM-ICU determines the content of that consciousness and the presence or absence of delirium. Delirium is a common but underdiagnosed form of organ dysfunction in the ICU. It was once thought to be a normal part of the process as patients arose from coma to consciousness, but it also occurs in the absence of this process or preexisting coma. Critically ill patients at risk for the development of delirium are those with preexisting cognitive impairment, advanced age, use of psychoactive drugs, mechanical ventilation, untreated pain, and sleep deprivation.¹⁴²

Three subtypes of delirium are described: hyperactive, hypoactive, and mixed. Hyperactive delirium is represented by the patient who is agitated and attempting to pull out lines and tubes; patients with hypoactive delirium typically are characterized by a flat or withdrawn affect; and those who fluctuate between the two have mixed delirium. ICU physicians have concentrated on treating multiorgan dysfunction, often dismissing delirium as an unavoidable, often iatrogenic consequence of critical illness with little long-standing ill effect. However, the development of delirium has been shown to be an independent predictor of higher 6-month mortality and longer hospital stay in mechanically ventilated patients.¹⁴³ In this study, the investigators demonstrated that the development of delirium, and the number of days spent in a delirious state, was associated with a threefold increase in mortality in mechanically ventilated patients compared with those who did not develop delirium. Current guidelines suggest in fact that critically ill patients be monitored daily for delirium.¹⁴⁴

The assessment for delirium combines the RASS score with inattention and either disorganized thinking or an altered level of consciousness.¹⁴² If patients respond to verbal stimulation (RASS score of −3 to +4) and have a positive CAM-ICU assessment, they are found to be delirious. Monitoring delirium with the CAM-ICU daily takes approximately 1 to 2 minutes, and nurses can easily be trained to perform it with their routine assessments.¹⁴³ One important aspect in daily assessment for delirium is identifying patients with the hypoactive delirium subtype. As mentioned before, this subtype is characterized by a flat affect present in a calm, seemingly alert patient.¹⁴⁵ These patients have normal or nearly normal arousal compared with the agitated patients with delirium, which is likely why they are underdiagnosed and may have a worse outcome. Outcomes can be improved by reducing unnecessary use of sedatives and analgesics.

Treatment of delirium can be divided into nonpharmacologic and pharmacologic agents. Nonpharmacologic methods concentrate on minimizing risk factors with repeated reorientation of patients, early mobilization and activity, nonpharmacologic sleep protocol, and timely removal of catheters and lines; unnecessary noise and stimuli are minimized, and patients are given their normal aids, such as glasses and hearing aids. Pharmacologic strategies first involve reduction or removal of unnecessary sedatives and analgesics. Many drugs, such as benzodiazepines, used to treat confusion often exacerbate it; lorazepam is a well-recognized offender. Pharmacologic agents with procognitive advantage, such as haloperidol, have also been recommended for the treatment of delirium, although this is based on sparse data in the ICU population.¹⁴⁴ Other antipsychotic and neuroleptic agents, such as risperidone, quetiapine, and olanzapine, are often used in both hyperactive and hypoactive delirium, but patients receiving these agents must be closely monitored for adverse cardiac and extrapyramidal events. An important resource for physicians is available at www.icudelirium.org. Figure 33-3 gives an overall approach in critically ill patients to balance both pharmacologic and nonpharmacologic approaches. It attempts to improve patient outcomes by addressing oversedation, immobilization, and delirium at the same time in high-risk critically ill patients.

Outside the ICU, postoperative delirium requires a hypoxic, metabolic, and septic workup. Medications and medication interactions have to be ruled out, and older patients such as our vascular patients suffer from multisensory deprivation, which makes them extremely confused at night.