The aim of this study was to determine the impact of a biomarker-based strategy on early discontinuation of empirical antifungal treatment. Prospective randomized controlled single-center unblinded study, performed in a mixed ICU. A total of 110 patients were randomly assigned to a strategy in which empirical antifungal treatment duration was determined by (1,3)-β-D-glucan, mannan, and anti-mannan serum assays, performed on day 0 and day 4; or to a routine care strategy, based on international guidelines, which recommend 14days of treatment. In the biomarker group, early stop recommendation was determined using an algorithm based on the results of biomarkers. The primary outcome was the percentage of survivors discontinuing empirical antifungal treatment early, defined as a discontinuation strictly before day 7. A total of 109 patients were analyzed (one patient withdraw consent). Empirical antifungal treatment was discontinued early in 29 out of 54 patients in the biomarker strategy group, compared with one patient out of 55 in the routine strategy group [54% vs 2%, p<0.001, OR (95% CI) 62.6 (8.1-486)]. Total duration of antifungal treatment was significantly shorter in the biomarker strategy compared with routine strategy [median (IQR) 6 (4-13) vs 13 (12-14)days, p<0.0001). No significant difference was found in the percentage of patients with subsequent proven invasive Candida infection, mechanical ventilation-free days, length of ICU stay, cost, and ICU mortality between the two study groups. The use of a biomarker-based strategy increased the percentage of early discontinuation of empirical antifungal treatment among critically ill patients with suspected invasive Candida infection. These results confirm previous findings suggesting that early discontinuation of empirical antifungal treatment had no negative impact on outcome. However, further studies are needed to confirm the safety of this strategy. This trial was registered at ClinicalTrials.gov, NCT02154178.