Résumé
Abstract Introduction Invasive mold disease is an important complication of patients with hematologic malignancies, and is associated with high mortality. A diagnostic-driven approach has been an alternative to the classical empiric antifungal therapy. In the present study we tested an algorithm that incorporated risk stratification using the D-index, serial serum galactomannan and computed tomographic-scan to guide the decision to start antifungal therapy in neutropenic patients. Patients and methods Between May 2010 and August 2012, patients with acute leukemia in induction remission were prospectively monitored from day 1 of chemotherapy until discharge or death with the D-index and galactomannan. Patients were stratified in low, intermediate and high risk according to the D-index and an extensive workup for invasive mold disease was performed in case of positive galactomannan (≥0.5), persistent fever, or the appearance of clinical manifestations suggestive of invasive mold disease. Results Among 29 patients, 6 (21%), 11 (38%), and 12 (41%) were classified as high, intermediate, and low risk, respectively. Workup for invasive mold disease was undertaken in 67%, 73% and 58% (p = 0.77) of patients in each risk category, respectively, and antifungal therapy was given to 67%, 54.5%, and 17% (p = 0.07). Proven or probable invasive mold disease was diagnosed in 67%, 45.5%, and in none (p = 0.007) of high, intermediate, and low risk patients, respectively. All patients survived. Conclusion A risk stratification using D-index was a useful instrument to be incorporated in invasive mold disease diagnostic approach, resulting in a more comprehensive antifungal treatment strategy, and to guide an earlier start of treatment in afebrile patients under very high risk.