BACKGROUND: Invasive candidosis is increasingly prevalent in seriously ill patients. Our aim was to compare micafungin with liposomal amphotericin B for the treatment of adult patients with candidaemia or invasive candidosis. METHODS: We did a double-blind, randomised, multinational non-inferiority study to compare micafungin (100 mg/day) with liposomal amphotericin B (3 mg/kg per day) as first-line treatment of candidaemia and invasive candidosis. The primary endpoint was treatment success, defined as both a clinical and a mycological response at the end of treatment. Primary analyses were done on a per-protocol basis. This trial is registered with ClinicalTrials.gov, number NCT00106288. FINDINGS: 264 individuals were randomly assigned to treatment with micafungin; 267 were randomly assigned to receive liposomal amphotericin B. 202 individuals in the micafungin group and 190 in the liposomal amphotericin B group were included in the per-protocol analyses. Treatment success was observed for 181 (89.6%) patients treated with micafungin and 170 (89.5%) patients treated with liposomal amphotericin B. The difference in proportions, after stratification by neutropenic status at baseline, was 0.7% (95% CI -5.3 to 6.7). Efficacy was independent of the Candida spp and primary site of infection, as well as neutropenic status, APACHE II score, and whether a catheter was removed or replaced during the study. There were fewer treatment-related adverse events--including those that were serious or led to treatment discontinuation--with micafungin than there were with liposomal amphotericin B. INTERPRETATION: Micafungin was as effective as--and caused fewer adverse events than--liposomal amphotericin B as first-line treatment of candidaemia and invasive candidosis.