Attributable mortality of candidemia – Results from the ECMM Candida III multinational European Observational Cohort Study

Abstract

Introduction

Despite antifungal advancements, candidaemia still has a high mortality rate of up to 40%. The ECMM Candida III study in Europe investigated the changing epidemiology and outcomes of candidaemia for better understanding and management of these infections.

Methods

In this observational cohort study, participating hospitals enrolled the first ten consecutive adults with blood culture-proven candidemia. Collected data included patient demographics, risk factors, hospital stay duration (follow-up of 90 days), diagnostic procedures, causative Candida spp., management details, and outcome. Controls were included in a 1:1 fashion from the same hospitals. The matching process ensured similarity in age (10-year range), primary underlying disease, hospitalization in intensive care versus non-ICU ward, and major surgery within 2 weeks before candidemia between cases and controls. Overall and attributable mortality were described, and a survival probability for cases and controls was performed.

Results

One hundred seventy-one pairs consisting of patients with candidemia and matched controls from 28 institutions were included. In those with candidemia, overall mortality was 40.4%. Attributable mortality was 18.1% overall but differed between causative Candida species (7.7% for Candida albicans, 23.7% for Candida glabrata/Nakaseomyces glabratus, 7.7% for Candida parapsilosis and 63.6% for Candida tropicalis). Regarding risk factors, the presence of a central venous catheter, total parenteral nutrition and acute or chronic renal disease were significantly more common in cases versus controls. Duration of hospitalization, and especially that of ICU stay, was significantly longer in candidemia cases (20 (IQR 10–33) vs 15 days (IQR 7–28); p = 0.004).

Conclusions

Although overall and attributable mortality in this subgroup analysis of matched case/control pairs remains high, the attributable mortality appears to have decreased in comparison to historical cohorts. This decrease may be driven by improved prognosis of Candida albicans and Candida parapsilosis candidemia; whereas candidemia due to other Candida spp. exhibits a much higher attributable mortality.