Mack C. Mitchell Jr.
Gastrointestinal Bleeding
The presence of blood in the stool or in the upper gastrointestinal (GI) tract is a significant finding that requires thorough investigation. GI bleeding may manifest as occult blood, hematemesis, melena (black stool) or intermittent hematochezia (overtly bloody stool). Massive hemorrhage, whatever the source, requires immediate hospitalization and often emergency diagnostic procedures. In hemodynamically stable individuals, the evaluation of GI bleeding can often be performed in an ambulatory setting. Table 46.1 shows the common conditions associated with GI bleeding.
Tests for Detection of Blood in Stool
In normal subjects, the hemoglobin concentration of the stool is less than 2 mg hemoglobin per g of stool, as measured by tagged red cell assay. The most commonly used test for fecal occult blood is the modified guaiac slide test (Hemoccult). This test depends on the pseudoperoxidase activity of hemoglobin to detect blood in the stool. It is not influenced by supplemental iron in the diet, but false-positive tests may result from ingestion of rare red meat and peroxidase-rich foods (uncooked vegetables such as broccoli, turnips, and cauliflower). False negative results are more likely in patients taking large doses of vitamin C. The stool slides can be stored up to 6 days if necessary without a decrease in the sensitivity of the test. Rehydration of the fecal material on the slide also increases the false-positive rate and is not recommended.
In populations screened with Hemoccult testing to detect colon cancer or polyps, 1% to 5% of subjects have positive test results (1). However, the predictive value of a positive test is only 2% to 17% for colon cancer (1) and 9% to 36% for adenomatous polyps. The remainder of positive tests are because of other causes of bleeding (e.g., gastritis, peptic ulcer) or to bleeding without a detectable source. Salicylates and other nonsteroidal anti-inflammatory drugs (NSAIDs) can cause occult GI bleeding, either because of a direct irritant effect on the stomach or duodenum or because of unmasking of an underlying lesion. Because colonic neoplasms may bleed intermittently, sensitivity is improved when multiple stool specimens are evaluated. Although the optimal number and
P.706
timing of stool samples have not been determined, checking two samples from three separate daily stool specimens is recommended for colorectal cancer screening (2). Among asymptomatic patients with a positive slide test who have carcinoma, more than 80% have early lesions limited to the bowel. Therefore, a positive test for occult blood in the stool requires further investigation and may favorably influence prognosis.
|
TABLE 46.1 Common Causes of Gastrointestinal Bleeding |
|
|
Evaluation of Patients with Gastrointestinal Bleeding
Choosing Appropriate Tests
The history and physical examination direct the sequence of the various tests used to investigate GI bleeding. The patient's age and medical history; the nature of associated symptoms, and the severity of bleeding are all important factors. For example, patients younger than 50 years of age are less likely to have a colonic lesion than are older patients. Peptic disease and benign rectal lesions are more evenly distributed among adults of all ages.
In asymptomatic patients with occult fecal blood and in patients with hematochezia but no other symptoms, the lower bowel should be investigated first. In patients with hematemesis or melena but no other symptoms, the upper GI tract should be investigated first. A reasonable approach to the evaluation of lower and upper GI bleeding is described in the next two sections.
Lower Gastrointestinal Tract
For patients presenting with hematochezia, flexible sigmoidoscopy or colonoscopy is usually the first test done to evaluate the lower bowel. In patients older than 40 years of age, colonoscopy should be performed as the initial test. However, in patients younger than 40 years of age, if the pattern of bleeding is consistent with rectal disease (see Chapter 98) colonoscopy is not always necessary. Diverticulosis (see below), should not be considered the cause of intermittent mild hematochezia until colonoscopy has failed to provide an alternative explanation. For patients with occult fecal blood and no localizing symptoms, a colonoscopy or a flexible sigmoidoscopy plus barium enema is the minimum recommended workup (1).
Proctosigmoidoscopy
Anorectal lesions are poorly visualized by barium enema. Cryptitis, bleeding hemorrhoids, fissures, and proctitis can be seen only by endoscopy. Even rectal polyps and cancer are much better revealed by proctoscopy than by radiography.
Flexible sigmoidoscopy is used to evaluate the rectum and descending colon. It has replaced rigid proctosigmoidoscopy for evaluation of the rectum and sigmoid colon. It is useful in screening asymptomatic patients for colorectal adenoma or carcinoma (3), but is not recommended for evaluating patients older than the age of 40 years with GI bleeding because colonoscopy is required to exclude more proximal colonic lesions.
Colonoscopy
Colonoscopy is indicated in patients with GI bleeding of suspected colonic origin, either as the initial test or as a followup examination when flexible sigmoidoscopy and/or barium enema have not provided an unequivocal diagnosis. Most gastroenterologists perform colonoscopy instead of a barium enema plus flexible sigmoidoscopy because colonoscopy has a higher positive predictive value in the evaluation of rectal bleeding (4). In patients with polyps, colonoscopy provides a way to remove the polyps without major surgery. An experienced endoscopist can reach the cecum in more than 95% of cases (5). Complications from the procedure are mainly perforation and hemorrhage; the overall complication rate for diagnostic colonoscopy is 0.3% to 0.4%, with a mortality rate of 0.02% (6). If polypectomy is performed, the morbidity increases to 1% to 2%, but the mortality rate remains the same.
The sensitivity of colonoscopy in experienced hands is much higher than that of an air-contrast barium enema: Only 2% of polyps are not diagnosed by the former procedure. In a study of anemic patients with occult GI bleeding, colonoscopy revealed polyps (greater than 5 mm in diameter) or cancer in 15% of patients with negative barium enema examinations; of patients with rectal bleeding, 34% had a significant lesion (including 11% with cancer) when the barium enema was reported as negative or simply as showing diverticulosis (7).
Patient Experience.
Preparation for colonoscopy usually includes a liquid diet for 1 day and laxatives such as Fleets Phosphosoda. An alternative preparation is to drink 4 L of a nonabsorbed isosmolar solution of polyethylene glycol and electrolytes (Go-lytely, Colyte, Nu-lytely) the night before the procedure. Elderly patients may have difficulty ingesting the large amount of fluid. Just before the procedure, the patient is given moderate sedation (usually with meperidine or fentanyl plus midazolam). Propofol is becoming more popular for moderate sedation because of more rapid onset and shorter recovery time after sedation. Moderate sedation is used to avoid the discomfort associated with distention of the bowel with air for inspection and with stretching of the mesentery as the colonoscope is maneuvered through the bowel lumen. However, some patients may be able to tolerate the procedure with only reassurance and no sedatives. There is no additional discomfort when a biopsy or polypectomy is performed. The duration of the procedure is variable, depending on the tortuosity of the colon, the presence of disease, and the skill of the endoscopist, but the average is 20 to 45 minutes.
P.707
If moderate sedation is administered, someone must accompany the patient home after the procedure because of possible lingering sedation. Antibiotic prophylaxis must be given to patients who are at high risk of endocarditis (see Chapter 65) before and after the procedure (see Chapter 93).
Upper Gastrointestinal Tract
The sequence of tests performed in evaluating the upper GI tract depends on the severity of the bleeding and the suspected diagnosis.Upper endoscopy, or esophagogastroduodenoscopy (EGD), is the procedure of choice for most patients with upper GI bleeding. In patients with acute bleeding, actively bleeding lesions can often be treated at the time of the procedure, by endoscopic band ligation or sclerotherapy in the case of bleeding esophageal varices or by electrocautery or the placement of hemoclips in the case of bleeding ulcer or angiodysplasia. If an upper GI series has been performed and is negative or reveals a gastric ulcer (Chapter 43) or a tumor, upper endoscopy should be the next procedure. If the upper and lower GI tracts have been evaluated in a patient with GI bleeding and both the studies are negative, a small-bowel series should be considered to investigate the possibility of other disorders that affect mainly the small intestine, such as Crohn disease.
Radiologic Studies
The upper GI series may detect mass lesions in the esophagus and stomach and gastric and duodenal ulcerations, although it is insensitive for evaluating mucosal detail. It is well tolerated and inexpensive. Chapter 43 describes the patient's experience during the performance of an upper GI series.
The conventional small bowel series is very poor at detecting small lesions of the intestine (e.g., cancer, leiomyoma). Disorders such as Crohn disease or lymphoma are more likely to be revealed by radiography (although a definitive diagnosis can be made only by biopsy). These sources of bleeding are uncommon and should be suspected only when the more common conditions (peptic ulcer disease, colonic polyps) have been excluded. The patient should be warned that the small bowel series requires spending 1 to 5 hours in the radiology department, during which time films are taken every 30 minutes.
Endoscopy
Upper endoscopy is the most widely used means of investigating upper GI bleeding. This technique not only is more sensitive than radiography but also provides a direct means of obtaining specimens for histological examination. Because of its greater sensitivity, endoscopy is indicated when barium studies are negative in the evaluation of a suspected upper GI source of bleeding.
Patient Experience.
Upper endoscopy is an outpatient procedure that usually takes less than 15 minutes. The patient fasts overnight before the procedure. Just before the procedure, the patient is sedated with intravenous medication (meperidine or fentanyl plus midazolam) and the throat is sprayed with a topical anesthetic. Under direct vision, mucosal biopsies and cytologic brushings can be obtained from suspicious lesions for histologic diagnosis. Biopsies are completely painless. Complications include perforation and bleeding but are extremely rare. Newer, smaller-caliber endoscopes have greatly improved patient tolerance of the procedure. After the procedure, because of the sedation, someone must accompany the patient home. The patient typically has a sore throat for several hours.
Occult and Obscure Gastrointestinal Bleeding
Occult bleeding is defined as chronic slow bleeding occurring from the GI tract, presenting either as iron deficiency anemia or a positive fecal occult blood test. Evaluation of occult bleeding should always start with a lower GI examination, preferably with a colonoscopy, especially in individuals older than 40 years of age, unless the history is strongly suggestive of an upper GI lesion. If the colonoscopy is negative, proceeding to an upper endoscopy is reasonable.
Obscure bleeding is defined as persistent GI bleeding, occult or overt, for which no source is identified by upper and lower endoscopy. Arteriovenous malformations, peptic ulcer disease, erosions in large hiatus hernias, and NSAID use are some of the causes of obscure bleeding. Obscure bleeding does not necessarily correspond to small-volume bleeding: It can range from occult to massive bleeding. Evaluation of obscure bleeding should begin with repeat upper and lower endoscopy. A duodenal or small bowel biopsy should be obtained for celiac disease during upper endoscopy, especially in young white patients with a history of diarrhea and persistent iron deficiency anemia. If these are again unrevealing, evaluation of the entire small bowel is warranted. A small bowel series may be helpful in patients with a negative endoscopic examination. It can detect lesions larger than 1 to 2 cm in size but is not very sensitive for identifying many causes of obscure GI bleeding. If the small-bowel series is negative, then capsule endoscopy or push enteroscopy should be the next step. (Push enteroscopy is performed by passing a longer endoscope than is typically used for upper endoscopy as far as possible beyond the duodenum [8,9]). A radioisotope-labeled red blood cell scan (the so-called “bleeding scan”) and mesenteric angiography should be considered in cases of overt bleeding that remain undiagnosed. Bleeding scans and angiography are helpful only when the rate of bleeding exceeds 0.1 to 0.4 mL/minute and 0.5 mL/minute, respectively, at the time of the study (10,11). In the absence
P.708
of active bleeding, however, angiography may sometimes be helpful if the characteristic vascular pattern of a Dieulafoy lesion (a dilated submucosal blood vessel that has eroded through the overlying mucosa), tumor, or vascular malformation is seen. Exploratory laparotomy with intraoperative enteroscopy is sometimes necessary in patients who require repeated transfusions because of recurrent bleeding and in whom the source of bleeding remains unidentified despite extensive evaluation as described previously. All patients with persistent bleeding should be evaluated for a disorder of hemostasis (e.g., von Willebrand disease) if an anatomic cause of bleeding has not been identified (seeChapter 56).
Wireless Capsule Endoscopy
Wireless capsule endoscopy can be performed using a small digital camera with memory chip that is embedded in a capsule. After the capsule is swallowed, the chip transmits digital images (2 frames per second) to a recording device that is worn on the belt for up to 8 hours. The capsule is propelled through the GI tract by peristalsis and permits recording of images from the esophagus to the ileocecal valve. The images are then replayed at a faster speed allowing the observer to interpret the “video endoscopy.” The technique has been useful in detecting obscure causes of bleeding, particularly from small mucosal lesions in the small bowel and is also useful for the diagnosis of Crohn disease of the small bowel (12,13). The major disadvantage is the inability to perform biopsies at the time of the imaging.
Selected Lesions that Bleed
Chapter 43 discusses the most common cause of upper GI bleeding—peptic disease. Common causes of lower GI bleeding include benign anorectal disorders (Chapter 98), inflammatory bowel disease (discussed later in this chapter), angiodysplasia, colonic neoplasms, and diverticulosis (see below).
Colonic Polyps
Colonic polyps or colonic cancer should be suspected in any patient older than 40 years of age who has GI bleeding or a change in bowel habits. Bleeding from colonic polyps may be occult or may occur as intermittent hematochezia, although most patients are asymptomatic. It is believed that most cancers of the colon (except those associated with ulcerative colitis) arise from these benign adenomas, although only a small percentage of premalignant polyps grow into invasive cancers. The removal of polyps before they become malignant has the potential to reduce the occurrence of colonic cancer in predisposed individuals (14). Although polyps are most common in the rectosigmoid region, they may be found anywhere in the colon.
|
TABLE 46.2 Polyps: Relationship of Size, Histologic Type, and Risk of Carcinoma |
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The risk of a polyp becoming malignant is related to the histological type and size. Hyperplastic polyps make up 10% to 30% of all colorectal polyps. They tend to be small (less than 0.5 cm in diameter) and to be located in the distal colon or rectum, and probably have no malignant potential. Villous and tubular adenomas make up almost two thirds of all colorectal polyps, are found in 25% of people by age 50 years (and in 50% by age 80), and carry a definite risk of malignant transformation that increases as they increase in size (15). The risk that a villous adenoma larger than 2 cm in diameter is cancerous is greater than 50% (Table 46.2) (16). Fortunately, if the cancer remains confined to the mucosa of the polyp (carcinoma in situ), colonoscopic polypectomy is curative. Because the cancerous change in the polyp may be focal, single biopsies of a polyp are not sufficient to exclude the presence of a malignancy; instead, the entire polyp must be excised.
Once an adenomatous polyp has been detected, surveillance for additional polyps is indicated. In 30% of patients more than one polyp is present at the time of initial investigation, and the risk of recurrence increases with the number and size of polyps that are initially discovered (17). Subsequent development of new polyps occurs in at least 10% of patients. If one or two adenomatous polyps less than 1 cm are detected during colonoscopy, repeat colonoscopy is indicated in 5 years to continue screening for new polyps or colon cancer. If more than 2 adenomatous polyps are found or if a polyp is greater than 1 cm in size, then colonoscopy should be repeated in 3 years. If a followup examination reveals no further polyps, the screening interval may be increased to once every 5 years (18).
Multiple polyposis syndromes are rare inherited abnormalities that are important to recognize because of their malignant potential. Familial polyposis, Gardner syndrome, and Turcot syndrome all are associated with multiple adenomatous polyps of the colon and therefore carry a high risk for development of carcinoma. Gardner syndrome includes osteomas and soft tissue tumors, and Turcot
P.709
syndrome includes tumors of the central nervous system. Familial polyposis and Gardner syndrome are inherited as autosomal dominant defects, while Turcot syndrome is an autosomal recessive disorder. The diagnosis of a polyposis syndrome is usually made in patients in their 20s, with cancer developing in virtually all patients within the next 20 years. There is considerable controversy about the therapy for these conditions. Colonic resection is curative, but the timing and extent of resection are subjects of controversy. When rectal polyps are present, a proctocolectomy should be performed to eliminate the risk of cancer. However, because the patients are generally asymptomatic and young, the prospect of an ileostomy is often overwhelming. The results of ileorectal pull-through procedures are encouraging. Cyclooxygenase-2 (COX-2) inhibitors (e.g., celecoxib), have been demonstrated to reduce the number of polyps in patients with familial polyposis syndromes (19). Whether these agents are of use also in patients with sporadic adenomatous polyps and whether they reduce the risk of malignant transformation is currently being evaluated through long-term prospective trials.
Whereas colonic polyposis syndromes involve adenomatous polyps, other conditions are associated with juvenile polyps or hamartomas. ThePeutz–Jeghers syndrome consists of multiple hamartomas, predominantly of the small intestine, associated with buccal and cutaneous pigmentation. Although the malignant potential of the hamartomas is low, duodenal and ovarian carcinomas have been reported in these patients and the cumulative risk for colon cancer is 39% (20). Rarely, juvenile polyps may occur throughout the GI tract. In the absence of associated extracolonic manifestations, this syndrome is called generalized juvenile polyposis; when accompanied by alopecia, nail bed changes, hyperpigmentation, and malabsorption, it is called the Cronkhite–Canada syndrome.
Colorectal Cancer
Epidemiology and Etiology
Cancers of the colon and rectum account for 14% of all cancers and are the second leading cause of cancer death overall in the United States. Overall, there is approximately a 5% lifetime chance of developing a colorectal cancer.
Colorectal cancer occurs with increasing frequency in older age groups, with two-thirds occurring in people older than 65 years of age. Geographic differences in the mortality rate from this neoplasm suggest an etiologic role for dietary and environmental factors. In particular, a high-fat, low-fiber diet appears to be associated with an increased risk of colorectal cancer. For unclear reasons, the proportion of cancers in the right side of the colon has increased in recent years, with a commensurate drop in the proportion of rectosigmoid lesions (21). Currently, approximately half of colorectal cancers are within reach of the flexible sigmoidoscope. The remaining half is proximal to the splenic flexure and accessible only with a colonoscope. The three main predisposing conditions for colorectal cancer are colonic polyps, familial polyposis (see previous discussion), and ulcerative colitis (see Chapter 45). The presence of these conditions dictates the need for a strict colonoscopic surveillance program and, at times, even prophylactic surgery to prevent the development of cancer. (See Chapter 45 for the risk of cancer in ulcerative colitis.) Although it is accepted that a family history of colorectal cancer predisposes a person to polyps and colorectal cancer, there is some evidence that the risk does not rise significantly above that of the general population unless more than one first-degree relative has had colorectal cancer (22) or unless the family member has had cancer before age 60 (18). Based on these observations, individuals with one or more first-degree relatives with colon cancer should begin screening at age 40 or 10 years earlier than the onset of cancer in the youngest affected relative (18).
Screening Tests
Because of the high incidence of colon cancer and its precursor, the colon polyp, population screening has been recommended in all patients older than 50 years of age. Patients at high risk should be screened at an earlier age (Table 46.3). A number of large, randomized trials showed that annual or biennial fecal occult blood testing decreases the 8- to 15-year cumulative mortality from colorectal cancer by 16% to 33%, compared with no screening (23, 24, 25, 26). Hemoccult cards are convenient for the patient because they can be mailed to the practitioner's office without a significant loss in sensitivity. For asymptomatic patients, it was recommended that such screenings be combined with a yearly rectal examination and with flexible sigmoidoscopy every 5 years (27). A retrospective case-control study of patients 45 years of age and older who had been monitored for 17 years revealed that screening sigmoidoscopy reduced deaths from colon cancer by almost 60% (3).
However, because of the insensitivity of the test for fecal occult blood (see earlier discussion) and because flexible sigmoidoscopy is limited to the descending colon, a substantial number of cancers may be missed (34% of men in a VA cooperative study [28]). In women, 65% of advanced neoplasms (polyps >1 cm and cancers) would have been missed by flexible sigmoidoscopy alone (29). It is reasonable, therefore, to consider whether colonoscopy at age 50 years, repeated (if normal) every 10 years, might not be the best screening test for colon cancer (30).
Computerized tomographic scanning with oral contrast can visualize the bowel as well as the intra-abdominal organs. Computed tomography (CT) colonography has been advocated as a method of population screening for colorectal cancer. A meta-analysis of 33 studies including
P.710
6393 patients showed the sensitivity for detection of polyps >9 mm was 85% (95% confidence interval, 79% to 91%) and 70% (95% confidence intervals, 55% to 84%) for polyps 6 to 9 mm in size (31). While this technique shows promise for the future, there are still technical problems that limit its widespread use today.
|
TABLE 46.3 Colon Cancer Screening |
||||||||||||||||
|
||||||||||||||||
Diagnosis
History
The chromosomal changes that accompany the progression of normal colonic mucosa to adenoma and then to carcinoma have been elucidated (32) and may provide a means to detect people who are at increased risk for development of colorectal cancer so that surveillance can be focused on them. Currently, most patients with adenocarcinoma of the colon are diagnosed only after symptoms have developed. Fewer than one-third are asymptomatic at the time of diagnosis (33), yet it is in this group that the highest chance for cure exists. The major presenting symptoms (33,34) are abdominal pain (25% to 75%) and a change in bowel habits (20% to 50%), either constipation or diarrhea. Abdominal pain is least common among patients with cancer of the rectum, where even large lesions can be accommodated without producing symptoms. Gross blood in the stool is another common complaint, occurring in 75% of patients with rectal cancer and 30% to 40% of patients with colon cancer above the rectum. This hematochezia, however, is often mistakenly attributed to hemorrhoids. Presentation with anemia or with weight loss is also common.
Physical Examination
The findings on physical examination vary according to the location and extent of the lesion. The primary tumor may be palpable as an abdominal mass, particularly in cancer of the right colon, where lesions can remain asymptomatic for long periods. Metastatic disease may be suggested by the presence of a large, hard, nodular liver; ascites; or peripheral adenopathy or by the palpation of a mass in the cul-de-sac on rectal examination. Signs of anemia may be present, particularly in lesions of the cecum and ascending colon, which can bleed covertly for months or even years before the diagnosis is made. Most patients have a positive test result for occult blood sometime during their course of illness.
Radiographic Studies
The diagnosis of colon cancer is sometimes made by barium enema or by CT scanning. Findings may include a polypoid mass, stenosis (either as a stricture or with an “apple-core” appearance), distortion of the mucosa, and localized rigidity of the bowel wall. At times, distortion or fixation of adjacent structures may be seen. The accuracy of the barium enema for the diagnosis of colon cancer is excellent, except at opposite ends of the large bowel. The cecum is often difficult to evaluate because of the inability to cleanse the region completely or to distinguish a prominent ileocecal valve or sphincter from a mass. The rectum is also difficult to visualize optimally, because often it is obscured by the balloon through which the barium is administered. Proctoscopy or flexible sigmoidoscopy is recommended in addition to a barium enema for patients with suspected colorectal carcinoma (7).
Endoscopy
Endoscopy plays a major role in the diagnosis of colorectal cancer. Some polypoid lesions, even if large or sessile, can be removed via the colonoscope, avoiding surgery in many cases. Additionally, colonoscopy has an important role in the identification of other colonic lesions. The prevalence of coexistent polyps in patients with colon
P.711
cancer is high, ranging from 10% to 30%. These residual polyps may develop into carcinomas, accounting for the incidence (5% to 10%) of a second colon cancer in patients with cancer of the colon over 25 years (16). Additionally, synchronous colon carcinomas occur in 3% to 5% of patients. Therefore, colonoscopy is helpful in ensuring that the rest of the colon is free of neoplastic lesions. Unless emergent surgery is required for obstruction, colonoscopy should be performed preoperatively. In a very large survey, the sensitivity of colonoscopy for the detection of colorectal carcinoma was 95%, whereas that of barium enema was 83% (7). There is also a significant miss rate for colonoscopy: For precursor polyps, colonoscopy failed to detect 27% of lesions 0.5 cm or smaller and 6% of lesions 1 cm or larger (35).
Carcinoembryonic Antigen
Carcinoembryonic antigen (CEA) is a fetal antigen found in the blood of many patients with colorectal carcinoma (from 30% of patients with local disease to 83% of patients with metastatic disease). CEA levels may be elevated in the blood of patients with other malignancies, in cigarette smokers, and with a variety of benign conditions, including peptic ulcer, pancreatitis, diverticulitis, and inflammatory bowel disease. Therefore, the level of CEA is not a useful screening test for the presence of colorectal cancer.
Although its efficacy and cost effectiveness for this purpose have been questioned (36), the assay may be helpful in the postoperative treatment of patients who have increased blood CEA levels at the time of diagnosis. Persistently elevated CEA concentrations postoperatively suggest metastatic disease; falling levels that then rise on followup evaluation suggest reemergence of the malignancy, usually at a remote site.
Therapy
A detailed discussion of the treatment of colon cancer is beyond the scope of this chapter. Surgery is the most effective therapy for most cases of colon carcinoma. Inoperable tumors (including metastatic tumors) respond 20% to 40% of the time to 5-fluorouracil (5-FU) combined with leucovorin, preferably prescribed by or with the advice of an oncologist. However, there is no clear effect of such therapy on survival. (5-FU is usually well-tolerated but occasionally mucositis, with severe diarrhea, occurs.) Although 5-FU has been the mainstay of therapy for metastatic colon carcinoma for decades, other agents are now available including irinotecan, oxaliplatin, and two humanized monoclonal antibodies that target vascular endothelial growth factor (bevacizumab) and the epidermal growth factor receptor (cetuximab). These drugs are generally administered in combination, usually with 5-FU. Operable tumors at an advanced stage (stage III or IV; Table 46.4) should be treated with adjuvant systemic chemotherapy. Administration of 5-FU with leucovorin postoperatively decreases recurrence and improves survival (37). The addition of levamisole does not seem to add any benefit. Capecitabine is an oral fluoropyrimidine carbamate preferentially converted to 5-FU in tumor cells. The role of oral chemotherapy agents, in particular capecitabine, is still being defined. It is not yet clear whether patients with stage II colon cancer benefit from adjuvant therapy (38).
|
TABLE 46.4 Colon Carcinoma |
||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||
Patients with rectal cancer should be evaluated by an oncologist prior to surgery to determine the need for preoperative adjuvant radiation and/or chemotherapy. Preoperative chemoradiotherapy has been shown to reduce local recurrence rates and complications compared with postoperative chemoradiotherapy in patients with clinical stage T3 or T4 or node-positive (i.e., locally advanced) rectal cancer (39).
Preoperative Evaluation
Before surgery, most patients should undergo evaluation for metastases. Liver function tests and computed tomography (CT) or an ultrasound of the liver should be performed routinely. In patients with bowel obstruction or bleeding, surgery may still be needed as palliation, despite the presence of liver metastases. In patients who are asymptomatic from their bowel lesions, the presence of multiple hepatic metastases should deter surgical intervention, although single hepatic metastases are often resectable. Abnormal liver function tests alone should not be considered absolute evidence of metastatic disease, nor should normal test results be considered to exclude hepatic metastases. A histologic diagnosis should be made, if possible. Needle liver biopsy, guided by ultrasound or CT, is a simple way to obtain tissue. A preoperative CEA concentration should also be determined as a baseline.
For patients requiring an ostomy, preoperative evaluation by an enterostomal therapist is helpful, not only to
P.712
discuss with the patient problems and concerns about the ostomy, but also to mark the proper location of the ostomy preoperatively.
Prognosis
The prognosis of colorectal carcinoma is based on several variables. The major variable is the extent of the tumor, in terms of its invasion through the bowel wall and its lymph node involvement (Table 46.4). Vessel invasion and the degree of differentiation of the tumor histologically also affect survival.
Followup Care
Patients may develop diarrhea early in the course after partial colectomy, but it is usually transient and easily controlled with antidiarrheal medication. The patient with a colostomy needs continued followup care by the surgeon and the enterostomal therapist to ensure proper functioning and handling of the ostomy.
The goals of long-term followup are detection of recurrence or spread of the cancer and surveillance for new colonic lesions. Most commonly, metastases occur in adjacent nodes, with eventual spread to the liver. Physical examinations, liver enzymes, and CEA determinations usually are done at prescribed intervals for 3 to 5 years. Abdominal CT scans are done to assess a confirmed rise in CEA levels or liver enzymes, but whether long-term outcome is improved is unclear, despite earlier detection of recurrent tumor.
Colonoscopy should be performed within the first 6 to 12 months postoperatively, if it has not already been done preoperatively to determine the presence of synchronous lesions. If no lesions are found, colonoscopy should be repeated 3 years later. The interval may then be increased to once every 5 years if no recurrence or adenomatous polyps are detected. Yearly evaluation for occult fecal blood loss should also be performed, with three Hemoccult cards. If any of these are positive, colonoscopy should be repeated.
Arteriovenous Malformations of the Colon
Arteriovenous malformations of the colon are a common source of GI bleeding (40,41), most often in the elderly and in patients with chronic renal failure. A variety of terms have been used to describe these abnormalities, including angiodysplasia, hemangioma, and vascular ectasia. The etiology of the disorder is unknown. Although the lesions may occur throughout the GI tract, they appear most commonly in the mucosa of the cecum and ascending colon, where multiple lesions are often found, ranging in size from 1 mm to more than 1 cm. An association of angiodysplasia of the colon with aortic stenosis has been observed repeatedly (40).
The prevalence of angiodysplasias and the frequency with which they cause bleeding are uncertain. With increasing use of endoscopy and selective angiography, the disorder is being recognized more often. In one study of patients older than 60 years of age without a history of GI bleeding, submucosal vascular ectasia was detected in 53% and mucosal lesions in 27% (40). Angiodysplasias may be the most common cause of bleeding from the right colon. Together with diverticula, they are the most common causes of major lower intestinal bleeding in the elderly (41).
When these lesions bleed, they often produce hematochezia. The bleeding is sometimes brisk and may be massive, but occult blood loss may also occur and seems to be a common presentation of this disorder (42). Bleeding often stops spontaneously, but it commonly recurs. The lesions cannot be detected by barium enema, are not recognizable from the serosal surface by the surgeon, and are often overlooked by the pathologist. The diagnosis is best made by colonoscopy or selective arteriography (by which a malformation can be visualized even when the bleeding has stopped). However, as with diverticula, the mere presence of angiodysplasias does not incriminate them as the source of bleeding, and other potential sources should be sought. Endoscopic therapy with a heater probe, injection sclerotherapy, argon plasma coagulation, or laser ablation of discrete mucosal lesions can be performed by an experienced endoscopist. Lesions that are resistant to endoscopic therapy and continue to bleed should be treated by surgical resection of the involved segment of colon. However, angiodysplasias may be present diffusely throughout the intestinal tract and therefore not amenable to endoscopic or surgical therapy. Intra-arterial vasopressin or embolization at angiography may be helpful for actively bleeding lesions of the small and/or large bowel. Medical therapy with a combination of mestranol, 0.05 mg, and norethindrone, 1 mg, may be helpful in preventing rebleeding and decreasing transfusion requirements in patients with diffuse intestinal angiodysplasias (43). Combination therapy is associated with significant adverse effects in both men (gynecomastia, testicular atrophy) and women (vaginal bleeding), and bleeding may recur once the therapy is stopped. Nonetheless, medical therapy may be extremely helpful in selected patients.
Clinicians should be aware of this disorder, especially in elderly patients with GI bleeding in whom the initial evaluation is unrevealing, but the diagnosis is made only after consultation with a gastroenterologist or radiologist.
Diverticular Disease of the Colon
Definitions
The terminology for conditions subsumed under the phrase diverticular disease is widely misunderstood. The phrase refers to a variety of clinical states that may differ in etiology and prognosis. Table 46.5 lists the nomenclature of diverticular disease of the colon. As can be seen from this classification, diverticulosis is simply the presence
P.713
of colonic diverticula, without presuming that there are accompanying signs and symptoms. Symptomatic diverticular disease is diverticulosis associated with pain or altered bowel habits in the absence of evidence of diverticular inflammation. Diverticulitis is inflammation of one or more diverticula, generally implying perforation of a diverticulum, and is almost always symptomatic.
|
TABLE 46.5 Nomenclature of Diverticular Disease of the Colon |
|
|
Epidemiology and Pathogenesis
The prevalence of diverticular disease in Western countries increases with age. Approximately 20% of men and women older than 40 years of age, 50% of those older than 60 years, and as many as 66% over age 85 have diverticulosis of the colon (44). Vegetarians have a much lower prevalence of diverticular disease than nonvegetarians. In a prospective study of almost 48,000 men, a low-fiber diet increased the risk of diverticular disease by two- to threefold over a 4-year period (45). This evidence led to the hypothesis that a low-fiber diet increases intraluminal pressure and that the increased pressure leads to herniation of the mucosa through weakened or porous parts of the colonic muscle. In support of this hypothesis is the demonstration in some individuals with diverticula of higher resting pressures in the colon and of exaggerated contractile activity in response to meals and cholinergic stimulation (46). Therefore, both a low-fiber diet and disordered colonic motility have been implicated in the pathogenesis of diverticulosis.
Weakness in the colonic wall through which the mucosa herniates to form the diverticulum is hypothesized to play a role in the pathogenesis of diverticulosis. Herniation occurs at the site of least resistance, most often at points of penetration of intramural vessels through the circular muscle layer. The association of colonic diverticula with scleroderma and with Marfan and Ehlers–Danlos syndromes suggests that loss of muscle mass or defects in collagen may be important factors. Changes in collagen synthesis occur with aging and may explain the increased prevalence of diverticula in elderly people. Therefore, the formation of diverticula may also involve a degenerative process of the colonic muscle with a change in tensile strength of the wall of the colon.
Asymptomatic Diverticulosis
A substantial majority of patients with diverticulosis are entirely asymptomatic. The diverticula may be localized to the sigmoid colon or may involve the entire colon diffusely. The sigmoid colon is almost always involved (95% of the time), and sigmoid diverticula account for 75% of all colonic diverticula (44). It is believed that this predilection is explained by the narrow caliber of the sigmoid colon, which results in higher intraluminal pressures and hence a greater risk of herniation. By contrast, rectal diverticula rarely occur. Diverticula are more common in the ascending colon in Asians, but the reasons for this observation are unclear (47,48).
The natural history of diverticulosis is variable. A majority of patients remain asymptomatic or have symptoms that are not severe enough to cause them to seek medical attention. Symptomatic diverticular disease manifests either as painful diverticular disease (75%) or as diverticulitis or hemorrhage (25%) (49). In a minority of cases, however, typical symptoms precede anatomic disease (the “prediverticular” state).
Although a diet high in fiber may be beneficial in preventing the development of diverticula, there is little evidence that therapy for asymptomatic diverticulosis is of any value in preventing or even delaying the occurrence of symptomatic diverticular disease or of such complications as diverticulitis or hemorrhage. Maintenance of regular bowel habits without the use of laxatives is probably the best advice for patients who are asymptomatic. It is also prudent to alert patients to the manifestations of symptomatic diverticular disease and to urge them to seek medical care promptly should such symptoms develop.
Painful Diverticular Disease
Diagnosis
Diverticular disease may at times become symptomatic. When the predominant symptoms are abdominal pain and an alteration in bowel habits, the cause is usually painful diverticular disease. The hallmark of this disease is abdominal pain without evidence of an inflammatory process. The pain may be colicky or steady, is generally in the left lower quadrant, and is usually made worse by meals (presumably because of gastrocolic reflex) and at least partially relieved by having a bowel movement or by passing flatus. Bowel movements, usually during the painful episodes, often become irregular, with development of constipation, diarrhea, or both in an alternating fashion. Constipation is more common than the other alterations in bowel habits. These attacks are usually episodic rather than continuous. Symptoms may also include nausea, heartburn, and flatulence.
P.714
Physical examination may reveal tenderness, at times significant, in the left lower quadrant of the abdomen. A tender sigmoid loop, which feels like a sausage, may be palpable. The stool should be negative for occult blood, but rectal bleeding may be found because of coincidental rectal outlet disorders such as fissures or hemorrhoids. The presence of fever, leukocytosis, or peritoneal signs points toward the more serious diagnosis of diverticulitis.
Colonoscopy or barium enema (see Chapter 45) is important both for diagnosing diverticulosis and for excluding other reasons for the symptoms. Spasm may be a feature of diverticular disease, but fistulas or a mass suggests diverticulitis, carcinoma, or Crohn disease. Particularly in elderly patients, in whom the prevalence of diverticulosis is high, it is important not to assume that the patient's symptoms have been explained once diverticula are found; carcinoma, for example, may be the real cause.
Therapy
The therapy for symptomatic diverticular disease is based on the assumption that a low-fiber diet and increased colonic pressure are important pathogenetic factors. Diets high in fiber (Chapter 45, Table 45.2) are prescribed and have been shown to be effective in improving bowel transit and relieving symptoms (50). Commercial preparations of hydrophilic colloids made from vegetable fiber are available and convenient, but they are more expensive than dietary sources (Chapter 45, Table 45.3). Patients should be instructed about high-fiber diets and, if necessary, should be given fiber supplements at a dosage of 4 to 10 g (1 tablespoon one to three times a day in a glass of water or juice). Artificial fiber products in tablet form are also available.
In addition to dietary maneuvers, anticholinergic drugs or antispasmodic drugs may be helpful for the relief of abdominal pain. Although these agents are not of proven value for this condition, some patients do respond. Dicyclomine (Bentyl), at a dosage of 10 to 20 mg before meals and at bedtime, or hyoscyamine (e.g., Levsin, also available in sustained-release and sublingual formulations) 0.125 to 0.375 mg every 4 hours as needed may be helpful. Other, more potent anticholinergics may produce adverse side effects and may aggravate the constipation.
The patient should be told that the course of the disease is unpredictable and that attacks will probably be experienced at irregular intervals (months to years) for the rest of his or her life. There is no benefit in continuing to take medication for the condition between attacks, but maintaining a high-fiber diet is prudent.
Diverticulitis
Diverticulitis results from perforation of one or more diverticula, usually in the sigmoid colon. Perforation may result from persistently high colonic pressures, from obstruction by a fecalith, and/or from an inflammatory process that weakens the wall of the diverticulum. It is important to note that the colonic mucosa is not inflamed in diverticulitis, since the inflammation is within the diverticulum or outside the wall of the bowel. Hinchey and colleagues described four stages of diverticulitis (51). In stage I, the perforation is small and confined to the pericolonic tissue. In stage II, the collection is larger, but in stage III, there is generalized purulent peritonitis. Stage IV is reserved for generalized fecal soilage and peritonitis. Stages III and IV are sometimes referred to as “perforated diverticulitis” (52). Diverticulitis increases in incidence with age and with duration of the underlying diverticulosis, and it is more common in patients who have many diverticula. Fistulas may form to the bladder (colovesical fistula is most common), vagina (especially after hysterectomy), small bowel, or skin (53).
Diverticulitis is the most common complication of diverticulosis. The long-term risk for development of diverticulitis among patients with diverticulosis is 10% to 25% (44), but it is probably much lower in patients with asymptomatic diverticulosis.
Diagnosis
The cardinal symptoms of acute diverticulitis are abdominal pain and fever. In classic cases, the pain is severe, abrupt in onset, and persistent, worsening with time and localizing to the left lower quadrant. The pain is often accompanied by anorexia, nausea, and vomiting. Altered bowel habits, both diarrhea and constipation, are common. Urinary tract symptoms and purulent vaginal discharge may occur because of fistula formation or because of inflammation of contiguous structures.
Abdominal tenderness and fever are found on physical examination. Localized peritonitis may be indicated by marked direct and rebound tenderness over the involved area, usually most pronounced in the left lower quadrant. The abdomen may become distended and tympanitic to percussion and the bowel sounds diminished. A mass may be felt at the site of inflammation in the left lower quadrant or on pelvic or rectal examination. Occult rectal bleeding occurs in approximately 25% of patients.
Leukocytosis is almost always present. Pyuria and/or hematuria may be found when there is involvement of the bladder or ureter.
The presentation of acute diverticulitis may be muted in the elderly patient. A high degree of suspicion is required in this population, because there may be no fever, no leukocytosis, and minimal abdominal pain.
The differential diagnosis includes painful diverticular disease, carcinoma of the colon, and inflammatory or ischemic bowel disease. The presence of peritonitis, fever, and leukocytosis rules out simple symptomatic diverticular disease. The other conditions are distinguished from diverticulitis by their clinical course and by endoscopy.
The diagnosis of acute diverticulitis is made largely on clinical grounds, but some tests may be useful in
P.715
confirming the clinical impression. Plain abdominal films are useful initially to detect free air caused by perforation, a surgical emergency. Abdominal CT can be very helpful in visualizing diverticular abscesses and other signs of inflammation in the pericolonic tissues, and has largely supplanted other radiographic and endoscopic modalities in the clinical diagnosis of diverticulitis (54). The role of endoscopy is very limited in acute diverticulitis, since there is seldom enough mucosal inflammation to confirm the diagnosis and the risk of perforation is significantly increased. Weeks later, after successful medical therapy, it is safe to perform colonoscopy to exclude other conditions, such as carcinoma or Crohn disease.
Therapy
Most patients with diverticulitis should be hospitalized, placed on bowel and bed rest, and given analgesics, intravenous hydration, and antimicrobial drugs (such as fluoroquinolones and metronidazole, synthetic penicillins, or third-generation cephalosporins to treat both aerobic and anaerobic infection). Selected patients, particularly younger patients who are otherwise healthy and who have only mild tenderness and low-grade fever, may be treated on an ambulatory basis with oral broad-spectrum antibiotics (e.g., ciprofloxacin 500 mg twice daily plus metronidazole, 500 mg every 6 hours for 2 weeks).
Although more than 75% of patients respond to conservative medical management, surgical consultation should be obtained early in the hospital course to facilitate operative intervention if necessary. The patient's condition usually improves markedly in 3 to 10 days if medical therapy is successful. For patients who respond to conservative management, a recurrence rate of 20% to 25%, mostly in the first 5 years, can be expected (55).
Failure to resolve the acute inflammatory process, recurrent attacks of diverticulitis, and obstructive stricture formation are indications for surgical intervention. It seems reasonable that patients be placed on a high-fiber diet after recovery from an acute episode of diverticulitis.
Diverticular Bleeding
Diverticular disease is the most common cause of gross lower GI bleeding in adults, followed closely by bleeding from angiodysplasia (56). Both diverticulosis and angiodysplasia are common in the older population, and both are commonly found in the proximal colon. Diverticular bleeds, in contrast to diverticulitis, occur in the right colon in two thirds of cases (even though diverticula are much more common in the left colon) (57). The average age of patients with diverticular bleeding is approximately 70 years (58). Bleeding is the presenting manifestation of diverticular disease in approximately 16% of patients (59,60). The exact cause of diverticular bleeding is uncertain. Diverticulitis is rarely, if ever, associated with gross bleeding. There is no evidence that dietary therapy reduces the risk of hemorrhage. Most instances of bleeding occur in patients who are otherwise asymptomatic.
Massive hemorrhage is a common mode of presentation for diverticular bleeding, although in other cases the bleeding may be occult. Massive lower GI bleeding in a patient known to have diverticula is not necessarily diverticular in origin. In 30% of cases, colonoscopy detects a second lesion (e.g., cancer or angiodysplasia) (57). Occult bleeding also should be ascribed to diverticulosis only after other causes have been excluded by a thorough evaluation (see above).
Patients with diverticular hemorrhage require hospitalization for hemodynamic stabilization, diagnosis, and therapy. Approximately 70% of patients stop bleeding spontaneously. Colonoscopy can be performed after purging with polyethylene glycol (PEG) solutions. If the site of bleeding can be identified, intervention with bipolar electrocoagulation, injection of epinephrine or placement of hemoclips can be used to control the bleeding. In the subset of patients who continue to bleed, angiography is the initial procedure of choice, if the bleeding is very brisk. During angiography, transcatheter embolization (with autologous blood clot, Gelfoam, or another agent) can be attempted and is sometimes effective in patients who are poor surgical candidates. Angiography usually does not identify a site of bleeding when the bleeding is slow and intermittent, and in such patients a radionuclide bleeding scan, and a colonoscopy after a standard bowel preparation, should be considered. Emergency surgery is indicated when the bleeding is persistent. It involves a segmental colectomy when the bleeding site has been identified but colonoscopic interventions have failed, or subtotal colectomy with ileorectal anastomosis when the bleeding site in the colon has not been identified. The recurrence rate after the first episode of diverticular bleeding is 20% to 25%; it is approximately 50% after the second episode and increases with each subsequent episode of bleeding (61). Patients with two or more episodes of significant diverticular bleeding should be considered for elective surgical resection (62).
Specific References*
For annotated General References and resources related to this chapter, visit http://www.hopkinsbayview.org/PAMreferences.
P.716