Inclusion Criteria:

• Age ≥ 18 years.
• In-patients with new onset of colitis evidenced by diarrhea (≥3 unformed stools within 24 hours), with one or more of the following: abdominal pain or cramps; peripheral leukocytosis, otherwise unexplained; or fever, otherwise unexplained.
• C. difficile toxin A or B detected in a stool specimen obtained within 7 days before enrollment by enzyme immunoassay.
• Patients able to take oral medications.
• Patients willing to avoid the following medications during the study: oral and intravenous metronidazole, oral vancomycin, anti-peristaltic drugs, opiates, Saccharomyces cerevisiae (baker’s yeast), Lactobacillus GC, cholestyramine or colestipol. [Patients on opiates may be included in the study as long as they were taking opiates prior to enrollment and the dose is not increased during the study].
• Patients willing to abstain from alcohol during the 10-day treatment duration and for two days following treatment.

Exclusion Criteria:

• Patients with other known causes of diarrhea or colitis (e.g., Shigella, Salmonella, Cryptosporidium parvum, Giardia lamblia, Entamoeba histolytica, inflammatory bowel disease, irritable bowel syndrome, advanced AIDS or chemotherapy for malignancy).
• Use within 1 week of enrollment of any drug or therapy with anti-C. difficile activity such as oral or intravenous metronidazole and oral vancomycin. [Patients that have taken up to 2 doses of metronidazole or vancomycin can be included in the study].
• Females of child bearing age who are either pregnant, breast-feeding or not using birth control. A double barrier method, oral birth control pills administered for at least 2 monthly cycles prior to study drug administration, an IUD, or medroxyprogesterone acetate administered intramuscularly for a minimum of one month prior to study drug administration are acceptable methods of birth control for inclusion into the study. In addition, female patients of child-bearing potential should have a baseline pregnancy test and should agree to continue an acceptable method of birth control for the duration of the study (including follow-up).
• Patients taking phenytoin, celecoxib, and/or losartan. [Patients taking Coumadin® (warfarin) may be included as long the prothrombin time is monitored at least twice weekly during the first 2 weeks of the study and at least weekly thereafter].
• Patients with severe renal or hepatic impairment.
• Patients who are clinically unstable (e.g., patients with signs of toxic megacolon or imminent perforation).
• Serious systemic disorders incompatible with the study.
• History of hypersensitivity to metronidazole.