Colorectal Surgery in the High-Risk Patient

Introduction

Increasing life expectancy coupled with the growing incidence of both benign and malignant colorectal disease has resulted in a greater need for surgeons to operate on patients with a variety of comorbidities. It has been estimated that elderly persons require surgery four times more often than the remainder of the population. This elderly group has a high incidence of cardiac and pulmonary disease, and other comorbidities may be the result of other primary disease processes (e.g., malnutrition in the patient with inflammatory bowel disease). Patients with such comorbidities represent unique challenges and often require an individualized and multidisciplinary approach.

Identification of the High-Risk Patient

The identification and management of preoperative risks is an integral part of preparation for surgery. Such an evaluation is critically important because it may guide the anesthetic and surgical strategy and provide the patient with realistic postoperative expectations. When determining risk, both patient and operative factors should be taken into account. As the number or severity of comorbidities rises, and/or the operative complexity increases, the risk of postoperative complications rises. Many studies have established specific, validated criteria that relate morbidity after surgery to preoperative risk factors. Many of these studies have looked specifically at colorectal surgery.

One of the oldest risk-calculating tools is Ondrula’s multifactorial index, which assigns points to defined risk factors, which can then be correlated to postoperative morbidity and mortality ( Table 81-1 ). To create their index, Ondrula’s group reviewed 972 colorectal procedures using multivariate discriminant function analysis and determined 11 factors that predicted operative outcomes. Each of these factors was assigned a point value, with highest point values given to emergency (vs. elective) surgery, age of 75 years or older, and congestive heart failure. These values were then combined to determine a “risk score.” The mortality rate in four patient groups was found to be directly proportional to the risk score (0-4 points = 1% mortality, 5-8 points = 10% mortality, 9-13 points = 19% mortality, and >13 points = 33% mortality). However, cirrhosis and renal insufficiency were excluded because values could not be assigned to them accurately as a result of a very strong association with mortality and because a small number of patients with either comorbidity caused a disproportionate statistical effect. Notably, Ondrula’s model contains both nonmodifiable factors (e.g., age and previous treatment with radiation) and modifiable factors (e.g., albumin level and the presence of congestive heart failure) that can be optimized prior to surgery and presumably improve outcomes.

The American Society of Anesthesiologists score is commonly used by anesthetists and surgeons to broadly describe a patient’s health. Although it is not commonly used to predict postoperative outcomes, this score is used internationally and provides a standardized way to describe a patient’s health preoperatively. This system was developed in 1963 and has since been adapted to include six categories of physical status ranging from “normal/healthy” to “moribund” to “declared brain dead, for organ harvesting” ( Table 81-2 ).

Other scoring systems, such as the Physiological and Operative Severity Score for enumeration of Mortality and Morbidity (POSSUM), which originally was designed for auditing purposes, have been developed to allow comparison among different institutions using validated formulas of predicted and observed morbidity and mortality. Because of concerns about overprediction of complications, variations such as the Portsmouth modification have been applied, permitting the direct comparison of outcomes in specific patient groups such as those undergoing colorectal or laparoscopic surgery or the elderly. First published by Copeland et al. in 1991, the POSSUM score was developed by analyzing patient comorbidities, intraoperative findings (such as blood loss and peritoneal soiling), and operative outcomes of 1372 nontrauma, elective, and emergency surgeries. Initially 62 individual factors were included in the analysis. Through multiple multivariate analyses, this number was first reduced to 35 and then again to the presently used 18-factor score. Although 432 gastrointestinal surgeries were included in the analysis for the design of the original POSSUM score, Tekkis et al noted the need for a colorectal-specific scoring system. Similar to the original POSSUM design, physiologic and operative parameters were recorded and related to outcomes. Procedures ranging from minor (such as hemorrhoidal sclerotherapy) to complex major (such as abdominal perineal resection) were included, and a 10-parameter, colorectal surgery–specific scoring system was then devised ( Table 81-3 ).

Most recently, Web-based scoring tools have been developed. These online tools are gaining popularity because of their convenience and easy-to-use formats, as well as the increased availability of tablets and other smart devices in the hospital setting. The American College of Surgeons Risk Calculator is the most commonly used tool of this kind. Twenty-one preoperative factors including patient comorbidities and procedure type may be entered into this publicly available Web page. With these data, eight 30-day postoperative outcomes are calculated. These outcomes include morbidity, mortality, the more general “any complication” and “serious complication” categories, and specific complications such as pneumonia, cardiac events, surgical site infections, venous thromboembolism, renal failure, and urinary tract infection. To develop this calculator, standardized clinical data from more than 300 American College of Surgeons National Surgical Quality Improvement Program hospitals and more than 1.4 million operations from 2009 to 2012 were analyzed. A subjective “surgeon adjustment score” was incorporated into the initial computerized algorithm to account for varying outcomes due to experience between surgeons, and the model’s validity was tested by 80 surgeons using 10 case scenarios.

In addition to calculating patient risk based on comorbidities, the operative risk of the individual procedure also must be taken into account. In general, longer operative times, increased blood loss, emergency surgery, and surgery in the presence of fecal spillage all confer additional risk. When developing the colorectal-specific POSSUM model, Tekkis et al evaluated more than 6000 patients undergoing colorectal surgery in 15 United Kingdom hospitals between 1993 and 2001. The work of these investigators provides a general overview of which procedures carry high and low risk of mortality in both the emergency and elective setting, with right hemicolectomy and anterior resection having the highest number of mortalities. In this large cohort, it was also noted that increasing age was significantly associated with worse outcomes. The odds ratios of mortality, when compared with a control group of patients younger than 60 years, rose from 4.5 in the 61- to 70-year age group to 8.4 in persons aged 71 to 80 years and 16.1 in persons older than 80 years.

Minimizing Risk Associated with Emergency Surgery

Emergency surgery is associated with worse outcomes and higher mortality than elective surgery and is a key factor in many risk-scoring systems. Emergency surgery was given the highest index value in Ondrula’s study, with a mortality of 2.8% in the elective setting and 12% in the study by Tekkis et al. This negative impact on surgical outcomes is likely due to a combination of factors, including advanced presentation of disease, an unprepared colon, and the effects of acute sepsis, stress, and/or dehydration on many organ systems.

Clearly, the patient with an acute abdomen or signs and symptoms of an intra-abdominal catastrophe cannot avoid the operating room for long. Through the use of increasingly sophisticated radiologic and endolaparoscopic modalities, however, careful initial nonoperative management of the emergency colorectal patient may allow a subsequent definitive operation to take place under semi-elective or even fully elective circumstances, thus decreasing operative risk. The indications for emergency colorectal surgery generally fall into one of three categories:

• 1.

  Sepsis (e.g., acute diverticulitis and pelvic abscess)

• 2.

  Obstruction (e.g., large bowel cancer, volvulus, and stricturing Crohn disease [CD])

• 3.

  Bleeding (e.g., diverticular disease and colitis)

The initial management for any colorectal emergency is fluid resuscitation with use of blood products if required. Diagnostic maneuvers including radiologic imaging (e.g., an abdominal radiograph and computed tomography [CT] with or without use of contrast material), endoscopic techniques (e.g., sigmoidoscopy and full colonoscopy), or angiography should follow ( Fig. 81-1 ). Such diagnostic techniques may be used in conjunction with treatment modalities in an effort to further stabilize the patient and control sepsis so there is time to optimize nutrition and correct anemia, coagulopathy, and electrolyte imbalances. When such “bridging techniques” are used, a definitive surgical procedure in an appropriately prepared bowel on an elective basis with a more favorable outcome may then be possible. Procedures performed electively in such a manner often result in the resection of significantly less bowel when compared with the same procedures performed as an emergency. One example is CT-guided percutaneous drainage of an intra-abdominal abscess from diverticulitis or CD, which frequently can be followed by the subsequent performance of a single-stage resection. Similarly, a colonoscopy or a Gastrografin enema can be therapeutic when acute obstruction is caused by a volvulus and diagnostic when caused by a malignancy or severe stricturing CD. In many cases, the obstruction can be relieved by either a colonoscopically placed stent or a “trephine” stoma created proximal to the obstruction in less than an hour, without laparotomy or even laparoscopy, after induction of light general or spinal anesthesia. When the obstruction is due to a distal colorectal cancer, chemotherapy or chemoradiation may then be considered, reducing tumor size for later definitive resection. In cases of CD strictures without an abscess, high-dose intravenous steroids may relieve the obstruction, allowing time for nutritional optimization prior to a more definitive future procedure.

Minimizing Risk Associated with Cardiac Disease

Numerous studies have been performed to determine factors predictive of cardiac complications in patients undergoing major noncardiac general surgery. History or presence of congestive heart failure, previous myocardial infarction, old age, diabetes, and emergency surgery are recurring features in these studies. It should be noted that up to 60% of patients who have peripheral vascular disease (PVD) also have significant silent coronary vascular disease. Carotid stenosis may be asymptomatic, but it carries a risk of intra- or postoperative cerebrovascular accident. If carotid stenosis is suspected or indicated, including the presence of PVD, carotid Doppler scanning should be performed. Colorectal procedures can add cardiac stress through blood loss and prolonged reverse Trendelenburg positioning.

The categories of cardiac risk in noncardiac surgery can be divided into three groups, all of which predict negative outcomes: (1) left ventricular (LV) dysfunction; (2) myocardial ischemia; and (3) heart valve abnormalities.

Preoperative investigations are tailored to these categories. Examples of such investigations by category include serum brain natriuretic peptide, echocardiography and cardiac CT or magnetic resonance imaging (MRI) for LV dysfunction, an electrocardiogram, serum troponin and creatine kinase–myocardial bound for acute ischemia, exercise treadmill in the fit patient or nuclear perfusion imaging, a dipyridamole, adenosine, or dobutamine stress test in the less fit patient for chronic ischemia, and echocardiography, an electrocardiogram, and cardiac CT or MRI scanning to assess for valvular anomalies. Several guidelines are in place to determine the appropriate preoperative test for cardiac risk assessment. Many guidelines are the result of collaborations between expert groups such as the European Societies of Cardiology and Anaesthesiology’s guide and the guidelines from the American College of Cardiology and the American Heart Association, who have worked together since 1980 to produce such recommendations.

A common feature of cardiac evaluation guidelines is determination of functional capacity, usually in metabolic equivalents (METs). For example, one MET equals the basal metabolic rate at rest, climbing two flights of stairs is two METs, and walking two blocks is three METs. Inability to perform activities that require four METS (e.g., climbing four flights of stairs) predicts increased incidence of postoperative cardiac events. For patients with known cardiac comorbidities or poor functional capacity, stress echocardiography using exercise or pharmacologic agents such as dobutamine or dipyridamole is performed. Coronary angiography should be performed if ischemia is suspected. Stress echocardiography has a high negative predictive value (90% to 100%), but the positive predictive value is relatively low (25% to 45%). Thus a positive test demands further evaluation. MRI can assess cardiac perfusion and wall motion and can assist in detecting ischemia. CT may be used as a noninvasive angiographic modality for the detection of coronary calcium reflective of atherosclerosis.

The Goldman index and its modifications, the Lee index and the Modified Multifactorial Index by Detsky et al, are commonly used to predict cardiac risk in patients undergoing noncardiac surgery. The Lee index is the most concise of the three indices and was developed by prospectively following up on more than 2800 patients undergoing a variety of procedures and validated with an additional 1400 patients. The independent determinants of perioperative cardiac events are listed later ( Table 81-4 ). In the Lee index, each determinant is given one point. The risk of major cardiac complications according to the Lee index is as follows: 0 points = 0.4%, 1 point = 0.9%, 2 points = 7%, and 3 or more points =11%.

Perioperatively, cardiac medications should be continued, especially β-blockers, both to avoid rebound tachycardia and to prevent the myocardial ischemic effects of catecholamines released during surgery. Effective pain control is similarly very important. Pulmonary artery catheter monitoring is usually reserved for patients of at least moderate risk. If used, it should be maintained through the postoperative period of fluid mobilization—usually 48 to 72 hours after surgery.