The efficacy, safety, and tolerability of voriconazole and fluconazole were compared in 391 immunocompromised patients with mycology- and biopsy-proven esophageal candidiasis. Primary efficacy analysis (256 patients) of esophageal treatment as assessed by esophagoscopy revealed success rates of 98.3% with voriconazole and 95.1% with fluconazole. The 95% confidence interval for the difference in success rates ranged from -1.0% to 7.5%. The overall safety and tolerability of both antifungals were acceptable. Fewer patients discontinued voriconazole treatment because of insufficient clinical response (4 patients [2.0%] vs. 5 patients [2.6%]). More patients discontinued voriconazole than fluconazole treatment because of laboratory test abnormalities (7 patients [3.5%] vs. 2 patients [1.0%]) or treatment-related adverse events (5 patients [2.5%] vs. 1 patient [0.5%]). The most frequent adverse events (23%) with voriconazole were mild, transient visual disturbances. Voriconazole (200 mg, b.i.d.) was shown to be at least as effective as fluconazole in the treatment of biopsy-proven esophageal candidiasis in immunocompromised patients.