Chapter 88. Preoperative Assessment of the High-Risk Surgical Patient

• Perioperative risk assessment by careful history, physical examination, and selective investigation is essential for directing therapy in the high-risk surgical patient.
• To decrease mortality and morbidity, cardiac, pulmonary, and endocrine abnormalities must be identified and appropriately managed.
• Age >70 years, current or prior angina pectoris, prior myocardial infarction, Q wave on resting electrocardiogram, and cardiac failure are significant clinical risk factors for perioperative cardiac complications.
• Perioperative cardiac morbidity can be minimized with pre-emptive medical management which includes the perioperative administration of β-blockers.
• Postoperative pulmonary complications can be reduced by aggressive pre- and postoperative care.
• Diabetes mellitus, abnormalities in thyroid function, and steroid dependence can significantly influence perioperative morbidity and mortality.

As indicated in Chap. 87, surgery and anesthesia trigger a host of physiologic responses. In the average otherwise healthy patient these responses result in no major untoward postoperative sequelae. However, in the medically compromised patient the additional burden of surgical stress can prove to be very challenging and sometimes insurmountable. Such patients frequently require detailed evaluation and monitoring in the preoperative as well as postoperative periods in the intensive care unit (ICU) setting. Careful planning of the preoperative assessment and management of identified abnormalities in these patients are crucial to optimize chances of a good postoperative outcome. A major component of this planning involves the assessment of risks for intraoperative and postoperative morbidity. Patients with cardiac, respiratory and endocrine abnormalities pose special risks for postoperative complications. In this chapter we present guidelines for identifying and managing patients at risk of developing postoperative cardiac-, respiratory-, and endocrine-related postoperative morbidity.

Clinical Assessment of Cardiac Risk

Over 25% of patients presenting for noncardiac surgery have diagnosed or clinically suspected ischemic heart disease with increased risk for perioperative complications.1 The American Society of Anesthesiologists' (ASA) classification of physical status (Table 88-1) is still commonly used as an index of surgical risk.2 Using multivariate analysis of 1001 consecutive patients presenting for noncardiac surgery, Goldman and associates developed an index for perioperative risk (Cardiac Risk Index; CRI) based on clinical, electrocardiographic (ECG), and routine biochemical parameters (Table 88-2).3 The strongest predictors of cardiac morbidity were the severity of coronary artery disease, a recent myocardial infarction (MI) and perioperative heart failure. Detsky and coworkers developed a more elaborate scoring system.4 Such studies suggest an overall perioperative cardiac risk for consecutive patients presenting for noncardiac surgery of 1.4%, and the risk of cardiac death at approximately 1%. However, in patients with previously documented or clinically suspected ischemic heart disease, the rates of perioperative MI (4.1%) and death (1.7%) were markedly higher. The retrospective analysis of 1001 patients from whom the CRI was derived is shown in Table 88-3. Those patients in class IV had a 22% incidence of major cardiac complications and a 56% mortality rate.

Table 88-4 compares the predictive values of the CRI with the ASA scoring system and other noninvasive tests used to assess cardiac risk. Class IV patients were still a high-risk group, but the positive predictive value (PPV) was lower at 0.5.5 The ASA has a low PPV because of the high number of false-positive results (i.e., although it is sensitive, it is not specific). Although a patient falling into CRI class IV is at high risk, the utility of the test is reduced by its low sensitivity. Only 55% of the patients at risk were in classes III and IV. The sensitivity may be even less when dealing with patients undergoing major vascular surgery.6

In summary, generally accepted independent risk factors for perioperative cardiac morbidity include: age >70 years, current or prior angina pectoris, prior MI, Q wave on resting ECG, and cardiac failure. It is widely accepted that patients presenting for elective surgery with zero, one to two, or three or more of these clinical markers are at low-, intermediate-, or high-risk respectively, for cardiac complications in the perioperative period. It is clear that patients in the low-risk category may proceed to surgery with an expectation of a low rate of cardiac complications. Similarly, patients deemed to be in the high-risk group may need to proceed to intensive management of their cardiac disease before initiation of the surgery. It is the group of patients in the intermediate-risk category who will benefit most from the investigations described below, in an effort to further elucidate the extent of their underlying cardiac disease and to attempt to quantify the perioperative risks before the commencement of the surgical procedure.

Exercise Electrocardiogram

Exercise ECG (ExECG) is a sensitive marker for subsequent myocardial infarction.7 Treadmill ExECG, utilized to calculate ventricular response to increasing oxygen demand, has been shown to be a sensitive noninvasive indicator of ventricular performance and predictor of perioperative ventricular failure.8 The prospective comparative value of ExECG was investigated in over 200 patients deemed to be at intermediate risk of adverse cardiac outcome following noncardiac surgery. Results showed odds ratios (OR) for ST-segment depression of 0.1 mV or greater in the ExECG (OR = 5.2) to be less than a known definite diagnosis of coronary artery disease (OR = 8.8), but greater than echocardiographic evidence of reduced left ventricular function (OR = 2.0). The combination of clinical variables and ExECG data improved preoperative risk stratification.9 However, ExECG is rarely practical in critically ill patients.

Ambulatory Electrocardiogram

Ambulatory electrocardiographic monitoring (AEM) using Holter monitors to detect silent preoperative and perioperative ischemia has been shown to be predictive of postoperative unstable angina, pulmonary edema, and MI.10 Twenty percent of ischemic events have been shown to occur in the preoperative period, 25% intraoperatively, and 41% postoperatively, with 97% of all events being clinically silent.11 Silent ischemic episodes in the preoperative and postoperative periods are associated with perioperative cardiac morbidity; both of these factors are more clearly associated with postoperative MI than intraoperative ischemia. The absence of preoperative ischemia may predict the absence of postoperative MI;12 however, this has not been a uniform finding.13

Radionuclide Scanning

Dipyridamole-thallium scanning (DTS) has been shown to be a sensitive predictor of perioperative cardiac complications for patients who cannot exercise.14 Dipyridamole dilates normal coronary vascular beds and diverts flow away from areas supplied by a stenotic vessel (usually indicative of >70% stenosis), resulting in a perfusion defect. Although DTS has been shown to be highly sensitive and specific in patients presenting for vascular surgery, it has proved less useful when consecutive patients have been studied and all researchers blinded.15 A similar result was shown when diabetic patients presenting for vascular surgery were evaluated with DTS in addition to clinical risk stratification.16 A large prospective study of DTS in 2000 vascular patients showed an OR of only 1.95.17 The authors added that the value of DTS depends on the patient's risk profile (i.e., low-risk patient with positive DTS is a more significant result than a high-risk patient with positive DTS).

Single photon emission tomography (SPECT) may improve accuracy of DTS.18 However, in patients presenting for abdominal aortic aneurysm surgery, SPECT has not accurately predicted adverse cardiac outcomes.19 While DTS reflects perfusion defects, a particular form of SPECT imaging (FDG SPECT) allows analysis of myocardial glucose utilization. In a study of 148 patients with ECG and echocardiographic evidence of Q-wave infarction, SPECT diagnosed viable myocardium in 61% of myocardial areas with a Q-wave pattern (i.e., previously deemed to be scar tissue), and 75% of dysfunctional non-Q-wave areas.20 Therefore, SPECT may have a further role in the evaluation of patients with evidence of myocardial damage.

Echocardiographic Imaging Techniques

Transthoracic echocardiography (TTE) provides the clinician with a detailed analysis of ventricular wall motion, ejection fraction, and if present, evidence of left ventricular hypertrophy. Resting TTE is noninvasive and readily applicable to the critically ill patient. Initial investigation of TTE as an assessment tool before noncardiac surgery did not support the routine use of TTE as a preoperative screening tool.21 However, a larger study showed that when TTE reveals systolic dysfunction in preoperative testing, TTE is associated with an increased OR of postoperative MI (OR = 2.8) and cardiogenic pulmonary edema (OR = 3.2). The authors concluded that in selected intermediate-risk patients, TTE before noncardiac surgery may provide independent information about perioperative cardiac risk.22

Dobutamine stress echocardiography (DSE) is echocardiography performed before, during, and after an intravenous infusion of dobutamine. A positive DSE is one that shows ventricular wall motion abnormalities or evidence of ischemia-induced diastolic dysfunction. Initial investigation of DSE to stratify patients at cardiac risk proved superior to DTS.23 The prognostic value of DSE in comparison to dipyridamole stress echocardiography (DiSE) and DTS was studied in over 2000 patients undergoing major vascular surgery. DSE and DiSE had significantly improved odds ratios for predicting cardiac complications (OR = 9.6 and 37.1, respectively) in comparison to DTS (OR = 1.95). The authors concluded that stress echocardiography effectively stratified patients into low- and high-risk groups regardless of clinical profile. In another study, when preoperative vascular surgery patients were selected to receive DSE according to protocol (one or more clinical risk factors), the prognostic value of DSE was again highlighted.24 DSE has little prognostic value in low-risk patients, but in the clinically intermediate and high-risk population, DSE may help identify patients in whom cardiac revascularization should be considered before noncardiac surgery.

Recent Myocardial Infarction

A history of recent MI remains one of the most significant predictors of perioperative myocardial infarction (PMI). Risk of reinfarction is increased substantially if the infarct occurred within 6 months of proposed surgery. Early studies reported reinfarction rates of 15% if the infarct occurred within 3 to 6 months prior to noncardiac surgery, increasing to 30% if it was within 3 months. Rao and coworkers subsequently studied patients who were aggressively managed in the ICU for up to 96 hours postoperatively. These retrospective data in 733 patients undergoing major noncardiac surgery identified reinfarction rates of 5.7% in patients within 3 months of MI and 2.3% in patients within 6 months of MI.25

Patients who have suffered an infarction are frequently assessed by exercise testing or angiography at approximately 6 weeks following infarction, in an effort to provide objective evidence of ongoing ischemia. On the basis of this testing, if patients have good exercise tolerance and are deemed to be at low risk for ongoing ischemia, noncardiac surgery does not necessarily result in increased cardiac morbidity and mortality.26

A history of angina was not included in Goldman's original CRI; however, symptoms or signs of ongoing ischemia, whether clinically silent or overt, have been repeatedly shown to be significant predictors of perioperative cardiac events. Based on current knowledge, patients within the first 6 months following MI should all be deemed to be at intermediate risk. These patients will benefit from noninvasive testing to quantify perioperative cardiac risk.

Congestive Heart Failure

Preoperative signs of congestive cardiac failure, a third heart sound, and signs of pulmonary edema are persistently strong predictors of perioperative cardiac morbidity. Left ventricular dysfunction serves as a marker of prior MI, ongoing ischemia, significant valvular disease, or cardiomyopathy. The highest risk for development of perioperative pulmonary edema is among patients >40 years of age who have evidence of cardiac failure preoperatively (approximately 15%, compared to 5% in patients with well-controlled cardiac failure). Patients with no pre-existing diagnosis of cardiac failure have approximately a 2% incidence of developing postoperative pulmonary edema.27

Age

Although multivariate analyses repeatedly show age to be a risk factor, other studies suggest age itself may not be a major risk factor independent of the general health of the patient.28 However, signs of ischemic heart disease may not be as clinically obvious in the elderly population. One study compared coronary angiography results from 260 elderly (>70 years) patients to 260 control patients (age <60 years). The results for patients with zero, one to two, and three or more clinical markers who exhibited severe stenoses were 17%, 55%, and 85% in the elderly group, in comparison to 14%, 36%, and 68%, respectively, in the control group. Thus the presence of clinical markers of ischemia in the elderly should prompt careful evaluation and a lower threshold for the initiation of noninvasive testing.

Diabetes

Diabetes, like age, may not be an independent risk factor per se. A retrospective multivariate analysis of 282 diabetics who underwent surgery found that the risk of a major cardiac event was related to pre-existing cardiac failure, valvular heart disease, and age >75 years.29 The age relation may reflect the duration of the diabetes. A large nationally based review of almost 15,000 patients who underwent vascular surgery revealed little influence of diabetes on subsequent cardiac mortality. After controlling for comorbid conditions, procedure type, and diabetic complications, only type 1 diabetics were at increased risk for cardiovascular complications. Neither type 1 nor type 2 diabetes was an independent risk factor for death.30

Implementation of Recommendations on Preoperative Testing

Recommendations from the American College of Cardiology/American Heart Association (ACC/AHA) have helped refine the clinician's approach to patients with known or suspected ischemic heart disease presenting for major surgery. Previously, 50% to 75% of patients presenting for major vascular surgery were investigated by noninvasive cardiac testing. By comparing historical data from the pre-guideline era (102 patients) and post-guideline implementation (94 patients), one center documented a dramatic reduction in preoperative stress testing (88% to 47%), cardiac catheterization (24% to 11%), and coronary revascularization (25% to 2%).32 Stress tests were equally distributed between low- and high-risk groups in the pre-guideline era, but were utilized significantly less frequently in the low-risk group in the post-guideline era (31% low-risk, 61% high-risk). As expected, associated costs fell dramatically to 10% of the pre-guideline average, but death (4% vs. 3%) and myocardial infarction rates (7% vs. 5%) were not statistically altered by the reduced number of investigations.31

Preoperative Percutaneous Transluminal Angioplasty

To date there are no randomized trials demonstrating the effectiveness of preoperative percutaneous transluminal angioplasty (PTCA) in patients presenting for major surgery with known ischemic heart disease. Elective PTCA carries inherent risks of acute MI (1% to 3%), emergency requirement for coronary artery bypass grafting (CABG) (0.2% to 3%), and in-hospital death (0.5% to 1.4%). The inevitable disruption of the atheromatous plaque in the stable patient may potentially increase the incidence of coronary thrombosis in the setting of major surgery. A retrospective study showed no reduction in perioperative MIs in patients treated with PTCA within 90 days of major surgery, compared to a known coronary artery disease control population.32 One uncontrolled retrospective study of patients receiving coronary stents before intermediate- or high-risk surgery showed a high rate of perioperative MI (28%), major bleeding (24%), or cardiac death (32%).33

Preoperative Coronary Artery Bypass Grafting

Recommendations for preoperative CABG34 were based on retrospective data, and either utilized historical controls or did not include the mortality associated with CABG itself. Preoperative prophylactic CABG still remains of unproven benefit. Since the first discernable survival benefit in CABG patients exists at 2 years post-CABG,35 the relative benefit of preoperative CABG over medical therapy remains unproven.

Perioperative Beta-Blockade

Identification of patients at risk for perioperative ischemia facilitates the introduction of perioperative β-blockade. In patients with known reversible ischemia, administration of perioperative β-blockers reduced the MI and death rate from 32% to 12%.36 This highly significant reduction in cardiac complications probably results from reduction in tachycardia and hypertension. The generalized acceptance of the safety and efficacy of β-blockers has led to calls for their routine use in patients who are at risk for ischemic heart disease. Mangano and coworkers studied 200 patients with known or suspected ischemic heart disease who underwent noncardiac surgery.37 Patients were randomized to receive atenolol or placebo pre-, intra-, and postoperatively until hospital discharge. In-hospital cardiac mortality was unchanged between groups, but overall mortality was significantly reduced in the three time spans studied: 0 to 6 months following discharge (0% vs. 8%), 0 to 12 months (3% vs. 14%), and 0 to 24 months (10% vs. 21%). In a large group of vascular surgery patients, in the subgroup of patients with fewer than three risk factors (83% of total), patients who received β-blockers had a lower risk of cardiac complications (0.8%) than those not receiving β-blockers (2.3%). The same group reported a 12% incidence of cardiac events in a high-risk vascular surgery population randomized to receive bisoprolol, compared to 32% in patients without β-blockade.38 They summarized that bisoprolol significantly reduced long-term cardiac death and MI in high-risk patients after major vascular surgery. Despite such widespread recommendations in favor of perioperative β-blockade, the optimal timing for stopping the β-blocker therapy has not been well elucidated. One small series suggested that abrupt withdrawal of β-blockers in the postoperative period may lead to increased postoperative MI (OR = 17.7).39

Modification of Surgery

Patients with concomitant unstable angina and recent MI rarely present for noncardiac surgery. However, such patients represent a significant risk for MI and cardiac death. Noncardiac operations in such circumstances should be deferred except in emergency situations. The substitution of major surgery for a minimally invasive approach is frequently impossible and may not lead to an improvement in cardiac risk. In a retrospective review of patients presenting for elective abdominal aortic aneurysm (AAA) repair, similar morbidity and mortality rates were found for an endovascular versus an open approach.40 While there was a significant increase in preoperative pulmonary compromise in the endovascular group, there were no differences in history and symptoms of ischemic heart disease, blood pressure, lipid profile, or diabetes between groups.

Intensive Perioperative Management

The rationale for using aggressive perioperative medical intervention to reduce cardiac risk is compelling. Many of the major cardiac risk factors such as congestive heart failure, myocardial ischemia, and dysrhythmias are detectable and amenable to therapy. The mainstay of intra- and postoperative monitoring for many years has been the pulmonary artery catheter (PAC; see below).

Preoperative ICU admission for investigation and “fine tuning” has been extolled by some centers. The use of such invasive monitoring with aggressive perioperative control of hemodynamics (i.e., admission to the ICU for 72 to 96 hours postoperatively) has been credited for a five- to sevenfold reduction in the reinfarction rate of patients undergoing surgery within 6 months of MI (24 Assessment Section). Precise factors responsible for this reduction are difficult to determine. Others have also reported low perioperative MI rates in high-risk patients when similar pre-, intra-, and postoperative strategies have been employed.41 However, the better outcomes could be due to factors such as better general medical management of the postinfarct patient prior to surgery, or the administration of β-blockade to the intensively managed patients (not commented upon in earlier papers). Postoperative cardiac morbidity following noncardiac surgery may be predicted. It may be that postoperative cardiac ischemia of long duration rather than the presence of ischemia itself is the factor most significantly associated with adverse cardiac outcome.

Perioperative optimization of oxygen delivery using intravenous fluid and inotropes has been reported to reduce overall mortality in high-risk surgical patients. Landmark papers in the 1980s suggested definitive end points of therapy (so-called “supranormalization”), but these were not specifically described for particular high-risk populations such as septic patients, the elderly, and those with ischemic heart disease. Some authors report higher mortality using an aggressive approach in such patients.42

Pulmonary Artery Catheter

The routine use of the pulmonary artery catheter to aid decision making for the high-risk surgical patient has been called into question. A large prospective cohort study of mixed medical and surgical patients in the ICU showed increased mortality, length of ICU stay, and costs associated with use of the PAC.43 A multicenter randomized trial of the use of the PAC in almost 2000 high-risk patients undergoing elective abdominal, thoracic, vascular, and major orthopedic surgery, showed no benefit to therapy directed by the PAC over standard care.44 In addition, eight patients in the PAC group showed evidence of pulmonary emboli in comparison to zero in the control patients. This interesting finding prompts even more speculation about the safety of the PAC in surgical patients.

Newer Cardiac Output Monitors

Such evidence suggests that the PAC may not be an appropriate monitoring tool in the elective surgical population. Newer less-invasive tools have come into clinical use. A central venous catheter capable of continuous measurement of central venous oxygen saturation was used by Rivers and associates to achieve early goal-directed therapy in medical patients, with dramatic improvement in mortality.45 Another tool that is less invasive than the PAC is the pulse contour cardiac output (PiCCO) monitor. This device requires brachial or femoral arterial cannulation with a specialized arterial catheter containing a thermistor. Using standard cold saline dilution technique, cardiac output analysis and measurement of intrathoracic blood volume (ITBV) and extravascular lung water are possible. Reports suggest that ITBV more reliably predicted cardiac preload than central venous pressure or pulmonary capillary wedge pressure.46

Transesophageal Echocardiography

Transesophageal echocardiography (TEE) is a sensitive marker of myocardial ischemia, often revealing segmental wall motion abnormalities before other signs of ischemia become obvious. TEE has been advocated to detect intraoperative ischemia, and has been shown to have superior sensitivity and specificity (sensitivity 75%, specificity 100%) in comparison to two-lead ECG (sensitivity 56%, specificity 98%) and pulmonary capillary wedge pressure (sensitivity 25%, specificity 93%).47 A larger study (224 patients) confirms this opinion, finding that TEE is frequently influential in guiding clinical decision making.48 In comparison to two-lead ECG and PAC, TEE was the most important intraoperative guiding factor in decision making for anti-ischemic therapy (56% of decisions based on TEE), fluid administration (28%), and vasopressor or inotrope administration (16%).

Resource Allocation

If aggressive hemodynamic monitoring in the ICU is responsible for the improved survival of patients with ischemic heart disease following noncardiac surgery, the financial implications are staggering. In Rao's study, more than 1300 ICU days of care were required to bring about a 2.4% reduction in the reinfarction rate. However, if admission criteria were restricted to congestive heart failure, angina plus congestive heart failure, or angina plus hypertension, this would account for almost 80% of the perioperative infarctions, but reduce ICU days to <300. A more recent study found that 77% of postoperative infarctions occurred within 26 hours of surgery, and 100% within 48 hours, suggesting a shorter period of monitoring may be sufficient.

Type of Surgery: Type of Anesthetic

Noncardiac operations may be divided into those that are likely to provoke perioperative ischemia and those that do not increase the risk of ischemia above normal. Major vascular procedures involving aortic cross-clamping and infrainguinal arterial bypass carry a high risk of postoperative ischemia,49 as do major abdominal and thoracic procedures. Orthopedic procedures such as total hip arthroplasty have a lower incidence of cardiac morbidity, and are deemed intermediate risk. Peripheral nonvascular procedures such as transurethral resection of the prostate, an operation frequently performed in patients with coexisting coronary artery disease, are associated with a low incidence of perioperative MI.50

Major surgery is associated with an intense sympathetic and procoagulant response, which may be implicated in the development of myocardial ischemia. These neurohumoral responses to surgery may be diminished with the use of epidural anesthesia and analgesia (EAA) extending into the postoperative period.51 Stress-mediated release of hormones such as cortisol, antidiuretic hormone, and catecholamines is blunted by epidural anesthesia.52 In addition, postoperative pain may contribute to tachycardia and resultant subendocardial ischemia. Early studies showed a dramatic reduction in cardiac complications in patients treated with EAA in comparison to general anesthesia (GA) alone.53 This opinion was further upheld by a large systematic review of relevant trials of epidural or spinal anesthesia vs. GA over the past 30 years, which showed a statistically and clinically significant reduction in mortality and morbidity after surgery.54 This retrospective work was prospectively tested by a large (915 patients) randomized controlled trial of high-risk surgical patients who either received EAA (intraoperatively and up to 72 hours postoperatively) with GA or GA alone with intraoperative and postoperative opioids as the mainstay of the analgesic regimen.55 The group observed no reduction in mortality or cardiac morbidity between groups. The only clinical benefit documented was a reduction in postoperative respiratory failure. However, the authors commented that 15 epidurals were required to prevent one episode of respiratory failure. The authors also commented that in no cases were there any serious complications with catheter placement or postoperative problems directly attributable to the placement of the epidural.

Recommendations

Evaluation and treatment of patients presenting for noncardiac surgery requires careful attention to history, functional status, and assessment of clinical evidence of reversible cardiac failure or dysrhythmias, in addition to consideration of the timing and indications for the proposed surgery. There is no doubt that clinical risk factors such as known ischemic heart disease, cardiac failure, age >70 years, and Q waves on a resting ECG are all independently documented to be associated with an increase in perioperative cardiac morbidity. Despite published and validated clinical risk indexes, the clinician will daily assess the patient on the basis of the above criteria, and ascribe the category of low, intermediate, or high risk according to the number of risk factors elucidated.

Following published guidelines and analysis of the literature, it can be recommended that noninvasive cardiac testing will not add to the clinicians knowledge or improve risk stratification in patients with none of the above clinical criteria. Similarly, patients who present a significant clinical risk or who have recently suffered an MI may require interventional cardiology and possible revascularization prior to proposed noncardiac surgery. This high-risk group should be intensively monitored in the perioperative period, including a stay in the ICU for approximately 48 to 72 hours. It is the intermediate-risk group of patients, presenting with one to two clinical risk factors, who will benefit most from noninvasive testing. Current evidence suggests DSE may be superior to DTS, TTE, or ExECG in preoperative cardiac risk stratification. However, DSE is not widely available and requires specialized centers and expertise. Current treatment protocols suggest a significant clinical benefit from the widespread administration of β-blockers to all intermediate- and high-risk patients. Some caution may need to be exercised regarding the duration of dose and discontinuation of the drug. Finally, despite recent negative reports regarding the use of the pulmonary artery catheter, evidence continues to support noninvasive or minimally-invasive perioperative monitoring of high-risk patients following noncardiac surgery.

Clinical Assessment of Pulmonary Risk

Postoperative alterations in pulmonary physiology predispose to the development of atelectasis (see Chap. 87). Marked decreases in forced expiratory volume in 1 second (FEV1) and forced vital capacity (FVC) have been documented with serial postoperative pulmonary function tests.56 A reduction in functional residual capacity (FRC) of approximately 70% of basal values may occur by about 18 hours after surgery, resulting in closure of small airways as FRC approximates closing volume.57 Progressive loss of functional lung tissue results and intrapulmonary shunting leads to worsening hypoxemia.

Many risk factors for the development of postoperative atelectasis have been highlighted. Each of the following has been shown to predict postoperative atelectasis: preoperative severe bronchitis, FEV1 more than two standard deviations less than predicted, obesity, malnutrition, abdominal surgery, and age. Analysis of risk factors in a group of 272 patients referred for preoperative assessment concluded that statistically significant predictors of postoperative pulmonary complications were: partial pressure of arterial carbon dioxide (PaCO2) ≥45 mm Hg (OR = 61.0), FVC ≤1.5 L/min (OR = 11.1), maximum laryngeal height ≤4 cm (OR = 6.9), forced expiratory time ≥9 seconds (OR = 5.7), smoking ≥40 pack-years (OR = 5.7), and body mass index (BMI) ≥30 kg/m2 (OR = 4.1).58

Bedside pulmonary function tests (PFTs) such as spirometry have been used to identify patients at risk of developing postoperative pulmonary complications, but lack of randomization, selection bias, and retrospective or unblended analysis of outcome invalidate conclusions.59 Spirometry as a screening procedure for high-risk patients remains unproved and its routine use has been discouraged.60 The American College of Physicians recommends preoperative pulmonary function testing in the following groups: patients with unexplained dyspnea, patients undergoing high-risk surgery (cardiac, thoracic, and upper abdominal), cigarette smokers, symptoms of dyspnea on exertion, and patients undergoing head and neck or orthopedic surgery with uncharacterized lung disease. All patients undergoing lung resection should have PFTs.61

The inability to improve pulmonary function despite adequate therapy may be a more sensitive predictor of postoperative respiratory failure.62 In a prospective study, those at risk of developing postoperative respiratory failure (defined as ventilator dependent >2 postoperative days) were best identified by the failure of 48 to 72 hours of intensive preoperative preparation to improve FVC, forced expiratory flow over 25% to 75% of the expiratory cycle (FEF25–75), and maximal voluntary ventilation measured over 1 minute (MVV). Five percent of the study group developed postoperative respiratory failure, and all of these patients had an FEF25–75 and MVV less than 50% of predicted values, which had not improved with preoperative therapy. The perioperative mortality in this subgroup was 60%.

Evaluation of Risk Prior to Pulmonary Resection Surgery

Approximately 80% of patients presenting for lung cancer surgery have concomitant chronic obstructive pulmonary disease (COPD) and 20% to 30% have severe pulmonary dysfunction.63 Pulmonary resection for lung cancer has been associated with morbidity of 12% to 50% and mortality of 2% to 12%.64 A more recent retrospective analysis confirms that these figures have not been markedly improved (morbidity 20%, mortality 3%).65 In addition to the general preoperative preparation of the surgical patient, those patients who will require pulmonary resection must have a preoperative estimation of postoperative pulmonary reserve.

A multifactorial risk index was proposed for patients undergoing thoracic surgery consisting of the Cardiac Risk Index (CRI) and a pulmonary risk index (PRI), known as the cardiopulmonary risk-factor index (CPRI). These pulmonary risk factors had previously been validated as independent risk factors in univariate analysis. The CPRI assesses obesity, cigarette smoking within 8 weeks of surgery, productive cough within 5 days of surgery, FEV1/FVC <70%, and PaCO2 >45 mm Hg. Each of these factors was assigned one point. By combining the CRI (0 to 4) and the PRI (0 to 6), patients classified as having a CPRI of 4 or greater were 17 times more likely to develop a postoperative pulmonary complication than patients with a CRPI less than 4.66

Guidelines for prediction of outcome following lung resection are generally based on preoperative whole lung function tests.67 MVV (% of predicted), FEV1 (liters), and FEV1 (% of predicted) have been most commonly used. Guideline values for proceeding with pneumonectomy, lobectomy, or wedge/segmental resection are:

• For pneumonectomy MVV >55%, FEV1 >2 L, FEV1% >55%;
• For lobectomy MVV >40%, FEV1 >1 L, FEV1% 40% to 50%; and
• For wedge/segmental resection MVV >35%, FEV1 >0.6 L, FEV1% >40%.

Predicted postoperative FEV1 has been suggested as a sensitive predictor of postoperative pulmonary complications. For this measurement, FEV1 and CT calculation of the number of preoperative functioning lung segments are required. Predicted postoperative FEV1 (ppoFEV1) may then be calculated using the formula:

Studies have suggested that if ppoFEV1 <40% of predicted, this may be a sensitive predictor of prohibitive operative risk and that resection should not be considered. More recent work suggests that a low ppoFEV1 may indeed be a sensitive predictor of postoperative pulmonary complications in lung cancer resection patients, but only in the group without pre-existing COPD. ppoFEV1 was not a significant predictor of postoperative pulmonary complications in patients with a preoperative diagnosis of COPD. This may be due to the fact that while these patients may have been losing lung tissue, a proportionally greater part of it was emphysematous and therefore less involved in gas exchange.68

In addition to these standard bedside tests, diffusing capacity (DLCO) may be helpful. If there is evidence of interstitial lung disease on chest x-ray or undue dyspnea on exertion, even if FEV1 and FVC are normal, a DLCO should be obtained. If preoperative FEV1 or DLCO is <40% of predicted, these are independent predictors of postoperative pulmonary complications. These patients may benefit from cardiopulmonary exercise testing. This is a sophisticated assessment of cardiopulmonary reserve and allows calculation of maximal oxygen consumption (VO2max). Risk of perioperative pulmonary complications can be stratified according the VO2max measured. Patients with a preoperative VO2max >20 mL/kg are not at increased risk of complications; however, patients with VO2max of <15 mL/kg and <10 mL/kg are at intermediate- and high-risk, respectively, for perioperative pulmonary complications.

Risk Modification

Preoperative Preparation

Standard preoperative preparation, including the use of intensive chest physiotherapy, bronchodilators, and appropriate use of antibiotics are considered routine practice in an effort to reduce the risks of postoperative pulmonary complications. High-risk patients who were assigned to receive a protocol of intensive pre- and postoperative therapy (including delay of surgery if indicated) had a 22% incidence of postoperative complications, compared to 60% in a group in whom the preoperative preparation was at the discretion of the admitting physician. Aggressive pulmonary therapy resulted in shorter hospital stay despite frequent delays of surgery in the treatment group to improve pulmonary function.69

Despite the clinical application of deep breathing exercises, intermittent positive-pressure breathing (IPPB), and incentive spirometry, these treatments have not been shown to be independently successful in the prevention of postoperative pulmonary complications. Incentive spirometry has been shown to be of benefit, but only in high-risk patients. Pre- and postoperative chest physiotherapy per se is of value only in the treatment of established pulmonary atelectasis.

However, high-risk morbidly obese (BMI >40 kg/m2) patients may benefit significantly from the application of biphasic positive end-expiratory pressure noninvasively and prior to signs of respiratory distress.70

Cessation of Cigarette Smoking

Continued smoking up to the time of surgery is associated with a significant increase in mortality. Cessation of smoking has been shown to result in a significant increase in FRC and reduced postoperative pulmonary complications. For maximum risk reduction, the patient should stop smoking at least 8 weeks prior to surgery.

Site of Surgical Incision

Extremity surgical intervention leads to little alteration in lung volume. However, lower abdominal incision results in a 25% to 30% decrease in FVC and a mild decrease in oxygenation. Upper abdominal and thoracic surgery produces significant impairment of pulmonary ventilation and defense systems independent of the effect of anesthesia.

Maintenance of FVC is essential for effective secretion clearance and is reduced by 50% to 60% immediately following upper abdominal surgery. Gradual restoration of FVC over the next 5 to 7 days is usual. During the first 24 hours after upper abdominal or thoracic surgery, reductions occur in tidal volume (20%) and FRC (70% to 80%) as a result of incision pain and reflex diaphragmatic splinting, resulting in rapid shallow breathing, absence of spontaneous sigh, and chest wall splinting.71

Laparoscopic surgery has revolutionized the postoperative care of patients undergoing common surgical procedures such as cholecystectomy and fundoplication. Comparing the open subcostal cholecystectomy to the laparoscopic approach, significant improvements in FVC (52% vs. 73% baseline), FEV1 (53% vs. 72% baseline), and FEF (53% vs. 81% baseline) were documented.72 A similar comparison of the laparoscopic versus open approach resulted in quicker return of postoperative respiratory mechanics towards but not quite to normal compared to open cholecystectomy.73 A review of over 300 patients following laparoscopic cholecystectomy showed no major postoperative pulmonary morbidity in the entire group, despite the fact that 45 were deemed to be obese, including 18 who were morbidly obese (BMI >45 kg/m2).74

Type of Anesthesia and Analgesia

Repeated attempts to demonstrate a consistent decrease in postoperative pulmonary complications with the use of regional anesthesia alone or in combination with general anesthesia have failed. Anesthetic technique per se is not a significant determinant of postoperative respiratory complications.75 All forms of regional anesthesia such as epidural local anesthetic, epidural narcotic, intercostal nerve blocks, and paravertebral nerve blocks have beneficial effects on postoperative FEV1, FVC, and partial pressure of arterial oxygen (PaO2).76 However, these beneficial effects have not proved to alter outcome in terms of pneumonia, respiratory failure, or death. Perhaps this is because the delayed reduction in FRC, resulting in atelectasis and hypoxemia, remains largely unaffected by the intra- and immediate postoperative analgesic regimens.

Repeated studies of intraoperative positive end-expiratory pressure (PEEP) have failed to show benefit in normal patients; however, the intraoperative application of PEEP 10 cm H2O to morbidly obese patients (BMI >40 kg/m2) resulted in an improvement in perioperative oxygenation (110 to 130 mm Hg).77 However, these patients were studied in the early postoperative period and this improvement may not have been maintained.

Recommendations

Although several questions concerning accurate preoperative assessment for prevention of clinically significant pulmonary complications remain unanswered, from available data a few guidelines may be generated. Likely to benefit the patient preoperatively are: cessation of smoking (at least 4 weeks, preferably >8 weeks abstinence), weight loss, and optimization of airway obstruction using bronchodilators. Intraoperative management options that may benefit the patient include: peripheral incision, limiting duration of surgery, endoscopic procedures, and the use of intraoperative PEEP. Postoperatively, good analgesia by the use of regional analgesia and patient-controlled analgesia delivery devices have been shown to be equally effective in reducing morbidity. Postoperative chest physiotherapy has proven value only for the treatment of atelectasis, but is useful as part of a program to encourage deep breathing exercises.

Hyperthyroidism

Hyperthyroidism in the surgical patient may significantly alter surgical risk by imposing additional burden on the cardiovascular system.78,79 Angina pectoris, atrial fibrillation, and cardiac failure may be precipitated, and other abnormalities include hypercalcemia, glucose intolerance, anorexia, myopathy, impaired drug absorption, and coagulation defects.

Preoperative preparation of a hyperthyroid patient includes deferral of elective surgery and optimization of thyroid function where possible. Most patients with hyperthyroidism have Graves disease and should be treated with a combination of antithyroid drugs (such as propylthiouracil and methimazole), iodide solutions, and possibly glucocorticoids to protect against adrenal insufficiency that may be associated with thyrotoxicosis.80

Surgical options for the treatment of a patient with Graves disease include subtotal or total thyroidectomy. A recent retrospective analysis of over 1200 patients over a 40-year period suggests that total thyroidectomy is the preferred surgical option for the treatment of Graves disease.81

If the surgery is emergent, the mainstay of treatment is perioperative β-blockade to reduce adrenergic thyroxine effects. Traditionally, propranolol has been recommended. However, esmolol may be preferred because of the rapid action and elimination, particularly when hemodynamic instability occurs.82

Thyroid storm is fortunately a rare event because of β-blockade prophylaxis, but when it occurs there is still a 40% mortality rate. Tachydysrhythmias, frequently atrial fibrillation, are seen. Cardioversion is usually unsuccessful. Hyperthermia, hypertension, leukocytosis, and evidence of adrenal insufficiency may occur. Treatment with β-blockade should begin once the diagnosis is suspected. Invasive hemodynamic monitoring should be instituted to guide fluid and electrolyte therapy in the intensive care environment where possible.

Hypothyroidism

The incidence of hypothyroidism in the general population is 0.5% to 0.8%, increasing to 1% to 3% in patients over 65 years of age. Treatment with lithium or amiodarone, drugs not uncommonly associated with drug-induced hypothyroidism, may alert the clinician to this diagnosis.

Although there are many physical signs of hypothyroidism, the only sign statistically correlated with biochemical abnormalities is the presence of delayed ankle reflexes. Retrospective analysis suggests that postoperative complications are not increased in patients with mild to moderate hypothyroidism, but patients with severe hypothyroidism undergoing surgery are more prone to periods of hypotension, neuropsychiatric disturbance, and intestinal hypomotility.83 Clinical assessment in addition to serum T3, T4, and thyrotropin levels should be used to assess thyroid status.84

Patients receiving regular thyroxine replacement, without signs and symptoms of hypothyroidism, can safely withhold oral thyroxine perioperatively for the majority of operations, as the half-life of levothyroxine is 1 week. If parenteral administration is required, thyroxine can safely be given intravenously at half the oral daily dose. Overtly hypothyroid patients undergoing emergency surgery should be treated with thyroxine replacement in addition to hydrocortisone 100 to 300 mg every 8 hours. Elderly patients with a clinical history or suspicion of coronary artery disease require cautious thyroxine replenishment and coadministration of nitrates.85

General anesthesia may be associated with profound cardiovascular depression in hypothyroid patients. Special consideration needs to be taken to titrate anaesthetic drugs to avoid this known effect. The use of newer anesthetic monitoring techniques such as bispectral index analysis may help titrate anaesthetic medications.86

Diabetes Mellitus

Diabetes mellitus is the most common endocrine disorder encountered in the perioperative period, since it occurs in almost 5% of the general population.87 Traditionally, diabetics presented for surgery for limb amputation and wound débridement, but owing to surgical advances in vitrectomy, cataract extraction, renal transplantation, and peripheral vascular repairs, diabetic patients are frequently presenting for preoperative assessment. Type I (insulin-dependent diabetes mellitus) comprises approximately 25% of the diabetic population, and affects a younger population (<35 years) which is prone to perioperative hyperglycemia and ketoacidosis because of inability to secrete insulin. Type II (non-insulin-dependent diabetes mellitus) patients are older and often obese and have a decrease in the number or responsiveness of insulin receptors, together with impaired insulin secretion, features which are accentuated in the perioperative period.88,89

The aim of therapy is to avoid excess morbidity and mortality which may be caused or exacerbated by extremes in blood glucose levels, undue protein catabolism, and fluid and electrolyte disturbances.

Risk Assessment

Although one study showed no difference in mortality and morbidity in vascular surgical patients, it is generally accepted that diabetics encounter increased perioperative morbidity and mortality.90–92

A perioperative myocardial infarction rate of 5.2% is reported in diabetics undergoing abdominal aortic reconstruction compared to 2.1% of nondiabetic patients. Inadequate control of blood glucose can lead to ketosis, acidemia, and dehydration. Decreased wound healing occurs at glucose levels greater than 200 mg/dL. Glucose concentrations greater than 250 mg/dL have been shown to impair leukocyte function and exacerbate ischemic brain damage. In addition to the effects of abnormal blood glucose levels, diabetics are at particular risk of atherosclerotic vascular disease, hypertension, and coronary artery disease.

Preoperative clinical markers for increased perioperative complications have been suggested. The presence of congestive cardiac failure, diabetic end-organ failure commonly associated with diabetes, valvular heart disease, and peripheral vascular disease associated with infection are statistically associated with increased mortality. Autonomic neuropathy is found in over 40% of patients presenting for surgery and may alter hemodynamic responses to intubation and surgery.93,94

Apart from anesthesia for cataract extraction, choice of anesthetic technique has not been associated with altered outcome.95 Thoracic epidural and spinal anesthesia techniques are associated with reduced intraoperative catecholamine release. However, to date, no study of exclusively diabetic patients has compared outcome between regional and general anesthesia techniques.

Risk Reduction

Preoperative evaluation should include thorough clinical assessment for cardiac, neurologic, and peripheral vascular abnormalities. A careful history for the presence of ischemic heart disease or prior myocardial infarction should be supported by cardiac investigations where appropriate. Glycosylated hemoglobin (HbA1c) levels less than 10% suggest satisfactory glycemic control. Renal failure may be present in up to 35% of diabetic patients upon presentation,110 and plasma creatinine, urinary sediment examination, and urine analysis for infection should be assessed. Recent evidence that intensive insulin therapy to achieve strict control of blood sugar (80 to 110 mg/dL) improves outcome in nondiabetic surgical ICU patients suggests that general practice (to maintain a blood sugar <200 mg/dL) may be inadequate in the postoperative diabetic.96

Nondiabetic patients recovering from surgery commonly show transient hyperglycemia and diminished insulin secretion and end-organ responsiveness in response to increased circulating catecholamines.97 However, these patients secrete sufficient insulin to suppress lipolysis and ketogenesis. Diabetics present with decreased or absent preoperative insulin secretion and a pre-existing insulin resistance, which serves to worsen the hyperglycemic response to surgery. Decreased peripheral use of glucose results in lipolysis, ketogenesis, possible acidemia, glycosuria, and dehydration.98

A variety of insulin regimens have been suggested for the routine perioperative management of diabetics undergoing surgery. No single regimen has proved markedly superior. Currently, perioperative intravenous glucose infusions are recommended, and insulin may be administered via a variety of dosages and routes. Subcutaneous administration of half the regular daily dose before surgery, using a variable-rate glucose infusion to maintain normoglycemia, has proved successful.99 Mixing insulin, potassium, and 5% or 10% glucose has also been suggested.100,101 Most authors suggest a variable-rate insulin infusion using an automated syringe device with a simultaneous glucose infusion through an alternative intravenous access.102,103

Insulin requirements vary widely in the perioperative patient. The normal state (approximately 0.25 unit of insulin per gram of glucose) is influenced by many factors such as obesity, concomitant glucocorticoid administration, and the septic state, which may increase insulin requirements to as high as 0.4 to 0.8 unit of insulin per gram of glucose. The highest insulin requirements have been observed in patients undergoing CABG (0.8 to 1.2 U/g of glucose).104

Recommendations for glucose infusion to prevent catabolism suggest between 5 and 10 g of glucose per hour,105 although the optimal dose of glucose necessary for prevention of fat and protein catabolism has not been clearly determined. However, clinical experience suggests that most surgical diabetics can be maintained within the normal blood glucose range with an insulin infusion set between 1 and 2 U/h.

Patients receiving total parenteral nutrition, which is generally up to 25% dextrose, usually require an additional 2 to 3 units of insulin per hour. Apart from careful monitoring of blood glucose levels, the patient's overall clinical status (particularly neurologic and hemodynamic) should be closely observed. Avoidance of hypoglycemic episodes while allowing mild hyperglycemia without ketosis is prudent in the diabetic whose blood sugar is extremely difficult to control.

What emerges from all the studies dealing with perioperative diabetes management is that the most important factor in optimal perioperative glycemic control is frequent measurement of blood sugar and appropriate therapeutic interventions by trained staff. Perioperative metabolic management should be planned and coordinated by surgeons, anesthetists, and diabetic care teams in conjunction with the patient when possible. With the exception of type II diabetic patients presenting for minor surgery, all diabetic patients should receive intravenous infusions of glucose with appropriate insulin to achieve normoglycemia until the preoperative regimen is resumed.

Glucocorticoid Supplementation in Chronic Glucocorticoid Users

Perioperative glucocorticoid supplementation for patients receiving steroid therapy is common. The rationale for its use is the avoidance of hypoadrenalism, resulting in a variety of clinical signs including fever, nausea, dehydration, abdominal pain, hypotension, and shock. Other evidence of hypoadrenalism includes low-voltage complexes on the ECG, hypoglycemia, and eosinophilia. Despite the common use of steroids in hospital patients, the incidence of perioperative adrenal insufficiency is low (0.01% to 0.1%).106

Retrospective and prospective data suggest that routine steroid supplementation for all glucocorticoid-treated patients may not be necessary.107–109. Well-known adverse effects of exogenous glucocorticoids include immunosuppression, exacerbation of osteoporosis, avascular necrosis of the femur, diabetes, peptic ulcer disease, diminished wound healing, and neuropsychiatric disorders.

Daily endogenous cortisol release in normal adults approximates 25 to 30 mg per day at rest. Stressors such as major surgery or critical illness increase endogenous production to 75 to 100 mg per day because of increased secretion of adrenocorticotropic hormone (ACTH) from the anterior pituitary gland. Increased cortisol secretion returns to normal within 24 hours of skin incision in uncomplicated minor surgery.110

The clinical rationale for steroid supplementation in the perioperative period is based on the known protracted recovery of the hypothalamus-pituitary-adrenal (HPA) axis following prolonged glucocorticoid administration.111 However, the stress response observed in perioperative patients results in ACTH levels far in excess of that required for maximal adrenocortical stimulation.112 A number of studies113–115 have suggested that patients on chronic steroid therapy undergoing elective major surgery may not require perioperative steroid supplementation in addition to their regular steroid regimen. In a study of 40 renal transplant recipients on chronic prednisone therapy, none of the patients received more than baseline glucocorticoid therapy during admission for surgery or critical illness. Despite biochemical evidence of decreased adrenal response to exogenous ACTH in 67% of the patients, none of the patients exhibited clinically overt hypoadrenalism, and 97% of all patients excreted normal or increased urinary cortisol concentrations during their hospital stay. This suggested that cortisol concentrations were sufficient to meet requirements during the time of stress.109

Identification of patients at risk for perioperative hypoadrenalism by history alone is difficult.113 Functional status of the adrenal gland may be assessed quickly and easily using the 30-minute ACTH test. This is the most effective and convenient tool for perioperative evaluation of HPA function.116

Baseline cortisol measurements are taken immediately prior to the intravenous administration of 250 μg of synthetic ACTH. Adequate adrenal function is diagnosed by baseline levels usually above 200 nmol/L, which rises above 500 nmol/L 30 minutes following ACTH administration (see Chap. 79).

Patients who exhibit a subnormal response to ACTH should receive supplemental steroids. Since endogenous cortisol secretion in normal individuals rarely exceeds 200 mg/d, exogenous steroid supplementation should range from 75 to 150 mg per day. To date, no data suggest that supplemental glucocorticoid therapy exceeding this amount is beneficial. Patients on chronic steroid therapy undergoing minor surgery should have their regular steroid dose on the morning of surgery and no additional doses if surgery is uncomplicated. Candidates for major surgery should receive 25 mg hydrocortisone intravenously on induction of anesthesia, and 100 to 150 mg per day over the following 24 to 72 hours. If a patient presenting for surgery is already receiving a maintenance steroid dose greater than the estimated requirement, additional steroid coverage is not necessary.115

Clinical and biochemical preoperative assessment of patients on chronic steroid therapy is invaluable in the identification of patients at risk for adrenal insufficiency in the perioperative period. Published recommendations for supplemental steroid coverage should be followed by dosing to physiologic levels. If doubt exists, a preoperative 30-minute ACTH test will clarify the need for perioperative supplemental glucocorticoid administration.