Antibiotic-Associated Diarrhea and Colitis

Diarrhea and colonic inflammation that occur during or shortly after the administration of antibiotics or chemotherapy.

The vast majority of cases are caused by a cytotoxin produced by Clostridium difficile; this cytotoxin triggers epithelial necrosis and a characteristic inflammatory process. C. difficile, the most common agent of nosocomial diarrhea, is acquired most often by elderly patients in hospitals or nursing homes. Nosocomial transmission involving environmental contamination with C. difficile and carriage of the organism on the hands of hospital personnel has been documented. Acquisition of C. difficile is often asymptomatic, but it may have clinical consequences if elderly patients receive certain antibiotics or chemotherapeutic agents. Other possible risk factors include surgery, intensive care, nasogastric intubation, and length of hospital stay. Some patients have antibiotic-associated diarrhea without evidence of C. difficile infection.

Although almost all antibiotics have been implicated, cephalosporins, extended-spectrum penicillins (eg, ampicillin), and clindamycin are implicated most often. Other penicillins and erythromycin are involved less often.

Symptoms and Signs

The disease spectrum ranges from mild diarrhea (with little or no inflammation) to severe colitis often associated with pseudomembranes, which adhere to necrotic colonic epithelium. The typical clinical picture of antibiotic-associated colitis includes watery nonbloody diarrhea, lower abdominal cramps, fever, and leukocytosis. Fever is usually low grade but may be high occasionally. In severe cases, dehydration, hypotension, hypoproteinemia, toxic megacolon, or colonic perforation may occur. When diarrhea occurs without colitis, constitutional symptoms are usually absent. Weight loss and malnutrition are associated with C. difficile in the elderly.

Diagnosis

Diagnostic studies are used to define anatomic and histopathologic changes and to identify the causative organism. Certain tests can identify C. difficile or its toxin. The tissue culture assay for a cytopathic toxin neutralized by specific antitoxins is the standard; the toxin is present in 95 to 100% of cases of pseudomembranous colitis if three separate stool specimens are examined. However, many hospitals lack the facilities for these assays and must submit stool specimens to reference laboratories. The preferred alternative is an enzyme-linked immunoassay, which yields results comparable to the tissue culture assay. This test has a reported sensitivity of 85% and a specificity of 100%. Stool cultures for C. difficile require selective growth media, and inexperienced laboratories have reported difficulties in recovering the organism. Moreover, whether C. difficile can be implicated in antibiotic-associated diarrhea without identifying the cytotoxin is controversial.

In general, endoscopy should be performed in severely ill patients who present atypically and require a rapid diagnosis to expedite treatment. In severely ill patients, flexible sigmoidoscopy is usually satisfactory, because the distal colon is involved in most cases. However, changes may be confined to the right colon in up to one third of cases, making colonoscopy necessary when less extensive procedures do not confirm a diagnosis that is strongly suspected. The yellowish gray pseudomembranes are dense and adhere to the underlying colonic mucosa, but the mucosa between the pseudomembranes appears normal. When pseudomembranes are not grossly visible, mucosal biopsies may show characteristic findings. Barium x-rays and CT are less useful. Barium enemas should be performed gently to reduce the risk of colonic perforation.

Treatment

The implicated causative drug should be stopped, if possible. If symptoms persist or the disease is clinically severe, patients should receive metronidazole 250 mg po qid for 7 to 14 days or vancomycin 125 mg po qid for 7 to 14 days. If oral administration is not possible, metronidazole 500 mg IV q 6 h should be given until oral administration is possible. Metronidazole and vancomycin appear to be therapeutically comparable in mild cases, but metronidazole costs substantially less. However, if the patient is seriously ill, oral vancomycin is usually recommended. Fever usually resolves within 24 hours, and diarrhea decreases over 4 to 5 days. A worse prognosis has been reported in patients with a serum albumin < 25 g/L, a fall in albumin > 11 g/L, and persistent toxin in the stools >= 7 days after therapy. Prognosis does not seem to be affected by age or sex.

Relapse rates average 20 to 25% after successful treatment with either drug. Patients who have one relapse are more likely to have another; this phenomenon cannot be explained by antibiotic resistance but may involve sporulation, which leads to relapse within 4 weeks after successful treatment. Relapses invariably respond to another course of antibiotics. In the 5 to 10% of patients who have multiple relapses, metronidazole or vancomycin in conventional doses should be followed by a 3-week course of cholestyramine or colestipol 4 g tid and/or Lactobacillus acidophilus 500 mg po qid, or vancomycin 125 mg po every other day. As an adjunct to antibiotic therapy, Saccharomyces boulardii, a nonpathogenic yeast, appears to reduce relapses.