Antibiotics and Clostridium difficile: Cause and Cure

Clostridium difficile, the most common nosocomial infection of the gastrointestinal tract, causes a spectrum of disease from diarrhea to the severe complication pseudomembranous colitis (PMC). This can be a severe illness, with increased morbidity and mortality. The organism produces toxins A and B that cause colonic inflammation and colitis. Diagnosis is made by detection of one or both toxins in stool specimens. When severe, diagnostic pseudomembranes are seen at colonoscopy; these are attached to the colon mucosa and are composed of fibrin, polymorphonuclear cells, and debris. Clostridium difficile-associated diarrhea (CDAD) is usually treated with oral metronidazole or vancomycin. Most will respond but 10% to 20% will develop recurrent CDAD (RCDAD) thus requiring retreatment. C. difficile became widely recognized as a distinct entity in 1974 when Tedesco et al1 performed a prospective study of 200 patients receiving clindamycin. Forty-one (20%) had diarrhea, and 20 (10%) had PMC. In hamsters, clindamycin caused cecitis and death. Work using this animal model led to the isolation of toxins A and B, with subsequent identification of C. difficile as the causative bacterium.2,3 CDAD has been reported as a consequence of many antibiotics, but most commonly implicated have been those with broad spectrum, especially against anaerobes. Cephalosporins, especially third generation, ampicillin, and clindamycin have been the most common antibiotics with a narrow spectrum that are less likely to cause CDAD. Even a single dose of an antibiotic can lead to CDAD. Furthermore, CDAD has been reported as a consequence of any antibiotics, even those used to treat it. There are several reports of CDAD related to metronidazole, and one case report related to vancomycin therapy.4 Rifampicin is another antibiotic that has been used to treat CDAD and RCDAD, but can also cause CDAD. In this issue of the Journal, Jung et al5 report 6 cases of PMC associated with rifampicin diagnosed over a 4-year period. All patients were being treated for pulmonary tuberculosis with multidrug regimens. Patient age ranged from 54 to 82 years of age. The onset of diarrhea ranged from 6 to 40 days after the onset of antimicrobial therapy. The diarrhea was watery, though 1 patient had bloody stools. Only half of the patients had stools positive for C. difficile toxin-by-toxin assay. All patients had colonoscopy; left-sided colitis was most common. Colonic biopsy was diagnostic in all 6. Other laboratory results were interesting: all patients had leukocytosis. Although no data were given white blood cell count at diagnosis, this is a good reminder that a leukemoid reaction can be seen in CDAD. Two patients had hypoalbuminemia—this has also been reported as an early event in CDAD.6 Hypoalbuminemia can be due to exudative loss from colon ulcers. Of course, in these patients active tuberculosis may have contributed to their leukocytosis and hypoalbuminemia. These clinical findings serve as a good reminder that although C. difficile diagnosis rests on stool assays these diagnostic tests are imperfect. Cytotoxin B assay has been the gold standard for decades, but is time consuming and expensive. Most laboratories now use enzyme immunoassay tests for toxins A and/or toxin B. These tests, overall, have good sensitivity and specificity but do have false negatives. Thus, clinical suspicion should prompt early empiric treatment if patient are ill while awaiting confirmatory diagnosis with a different or repeat stool assay, or visualizing the colon mucosa as was done here. There has been concern over the risk of colonoscopy with severe colitis, and controversy about the need to evaluate the right colon by colonoscopy in suspected PMC. One review of confirmed C. difficile colitis showed similar detection rates using sigmoidoscopy and colonoscopy.7 In the current paper all 6 cases were localized to the left colon, up to the hepatic flexure. Perhaps in a sick patient where urgent diagnosis is needed and stool tests are inconclusive, a flexible sigmoidoscopy or limited colonoscopy may be a reasonable first step. Pseudomembranes are classic, when present, but mild cases can have normal colon mucosa, or various degrees of colitis. Most importantly, I want to stress giving empiric treatment to sick patients, even if initial tests are negative but clinical suspicion is high. All of these cases resolved with treatment; their physicians discontinued rifampicin and treated with metronidazole. Vancomycin was needed for 2 patients who did not get better with metronidazole. There have been recent reports of more cases of decreased responsiveness to metronidazole.8,9 If patients do improve within 3 to 4 days on metronidazole, they should be treated with oral vancomycin. One of 6 patients (16%) had recurrent C. difficile, which fits with published rates of 10% to 20% for RCDAD. How can the authors be sure rifampicin was the culprit in patients with multiple antibiotics? They point out that rifampicin has wide antibacterial activity, but other agents like isoniazid and ethambutol do not. After treating the CDAD, they substituted a quinolone for the rifamipicin with no further CDAD. CDAD has been receiving increased attention lately as a result of epidemics with increasing severity in the United States, Canada, and some European countries. A strain has been identified that produces more toxin and producing a third binary toxin.10–12 Quinolone use seems to select for this strain and has been implicated in several epidemics. Quniolones, initially when introduced, were an uncommon cause of CDAD. The increasing use of quinolones, and the fact that some quinolone resistant strains acquire characteristics that may account for the increased virulence are likely contributing to the current serious epidemics. In summary, this case series serves as a nice reminder of several clinical lessons: (1) any antibiotic can cause CDAD, even those used to treat it; (2) onset of diarrhea is usually within a week but can be much longer; and (3) Stool diagnostic tests are imperfect so diagnosis requires a high index of suspicion. Sigmoidoscopy or colonoscopy may be needed in some cases.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.