Effectiveness and appropriateness of empiric metronidazole for Clostridium difficile-associated diarrhea.

OBJECTIVE

Although Clostridium difficile is the most common infectious etiology of nosocomial diarrhea, noninfectious causes are far more common. Empiric initiation of therapy for all patients is of unknown value. The aim of this study was to determine benefits of empiric metronidazole for Clostridium difficile-associated diarrhea (CDAD).

METHODS

We conducted a 4-month prospective surveillance of all patients in two community teaching hospitals receiving metronidazole for empiric treatment of presumptive CDAD. A database including antibiotic usage, fever, white blood cell count, feeding formula usage, comorbidity, and response to therapy was maintained.

RESULTS

Seventy-one patients on the medical (50), surgical (18), obstetric (two), and trauma (one) service were identified. Sixty-two had nosocomial diarrhea; nine had diarrhea on admission. Seventy (97%) received antibiotics; one (3%) was on nelfinavir only. Eighteen (25%) were subsequently proven to have CDAD; two (3%) had laxative-induced diarrhea; two (3%) had diarrhea secondary to a medication (colchicine [one] and nelfinavir [one]); one (1%) had diarrhea caused by bowel preparation for colonoscopy. The remaining 49 (68%) did not have a clearly established diarrhea etiology. (Four did not undergo stool examination.) Statistical analysis (chi(2) test) demonstrated a significant decrease in symptoms for metronidazole-treated patients with CDAD versus those with a different diagnosis (p = 0.05). Not surprisingly, multivariate regression analysis identified a strong correlation of diagnosing CDAD with age >60 yr, antibiotics exposure, fever, elevated white blood cell count, and resolution of symptoms with specific metronidazole treatment. CDAD was definitively diagnosed in 25% of our hospitalized patients with diarrhea, consistent with published data. Although some cases might have been missed, most patients did not have CDAD and received no benefit (and were potentially harmed) by empiric metronidazole. There was no way a priori to distinguish CDAD from non-CDAD.

CONCLUSIONS

In the absence of clear guidelines, empiric metronidazole should be reserved for strongly presumptive CDAD patients (older patients with comorbid conditions receiving broad-spectrum antibiotics associated with CDAD) who cannot hemodynamically or otherwise tolerate diarrhea. Used judiciously, empiric therapy may more rapidly resolve symptoms, and could conceivably prevent/abate severe complications and nosocomial spread.