A 10-day course of oral fidaxomicin (200 mg twice a day [b.i.d.]), a potent new macrocyclic drug, was compared to vancomycin (125 mg four times a day [q.i.d.]) in 1,164 adults (1,105 in the modified intent-to-treat [mITT] population) with Clostridium difficile infection in two phase III randomized, double-blind trials at sites in North America and 7 European countries. Of 1,105 mITT patients, 792 (71.7%), including 719/999 (72.0%) in the per-protocol (PP) population, provided a C. difficile strain at baseline, of whom 356 received fidaxomicin with 330 cures (92.7%) and 363 received vancomycin with 329 cures (90.6%). The susceptibilities (MIC(90)) of baseline isolates did not predict clinical cure, failure, or recurrence for fidaxomicin (MIC(90), 0.25 μg/ml for both; range, ≤ 0.007 to 1 μg/ml), but there was a one-dilution difference in the MIC(90) (but not the MIC(50)) for vancomycin (MIC(90), 2 μg/ml [range, 0.25 to 8 μg/ml] for cure and 4.0 μg/ml [range, 0.5 to 4 μg/ml] for failures). A total of 65 (7.9%) "rifaximin-resistant" (MIC > 256 μg/ml) strains were isolated in both treatment groups on enrollment, which increased to 25% for failures at the end of therapy. No resistance to either fidaxomicin or vancomycin developed during treatment in either of the phase III studies, although a single strain isolated from a cured patient had an elevated fidaxomicin MIC of 16 μg/ml at the time of recurrence. All isolates were susceptible to ≤ 4 μg/ml of metronidazole. When analyzed by restriction endonuclease analysis (REA) type, 247/719 (34.4%) isolates were BI group isolates, and the MICs were generally higher for all four drugs tested (MIC(90)s: fidaxomicin, 0.5; vancomycin, 2.0; metronidazole, 2.0; and rifaximin, >256 μg/ml) than for the other REA types. There was no correlation between the MIC of a baseline clinical isolate and clinical outcome. MIC(90)s were generally low for fidaxomicin and vancomycin, but BI isolates had higher MICs than other REA group isolates.