A multicentre study of antifungal strategies and outcome of Candida spp. peritonitis in intensive-care units.

Information on the species causing Candida peritonitis, their in vitro susceptibility, antifungal strategies in this setting and patient outcome is still scarce. AmarCand was a prospective, non-interventional study in 271 adult intensive-care unit (ICU) patients with proven invasive Candida infection who received systemic antifungal therapy (France, 2005-2006). Of these ICU patients, 93 (median age 65 years, simplified acute physiology score II 52) had Candida peritonitis, including 73 nosocomial peritonitis, 53 concomitant bacterial peritoneal infections and 26 candidaemias. Candida species were C. albicans (n = 63/108 isolates, 58%), C. glabrata (n = 22, 20%), C. krusei (n = 9), C. kefyr (n = 5), C. parapsilosis (n = 3), C. tropicalis (n = 3), C. ciferii (n = 2) and C. lusitaniae (n = 1). Of tested isolates, 28% were fluconazole-resistant or susceptible dose-dependent (C. albicans 3/32, C. glabrata 9/14, C. krusei 4/4). Empiric antifungal treatment was started 1 day (median) after peritonitis diagnosis, with fluconazole (n = 2 patients), caspofungin (n = 12), voriconazole (n = 3), amphotericin B (n = 2), or a combination (n = 4). Following susceptibility testing, empiric antifungal treatment was judged inadequate in 9/45 (20%) patients and modified in 30 patients (fluconazole was replaced by caspofungin (n = 14) or voriconazole (n = 4)). Mortality in ICU was 38% (35/93) and was not influenced by type of Candida species, fluconazole susceptibility, time to treatment, candidaemia, nosocomial acquisition, or concomitant bacterial infection. No specific factors for death were identified. In summary, a high proportion of fluconazole-resistant or susceptible dose-dependent strains was cultured. These results confirm the high mortality rates of Candida peritonitis and plead for additional investigation in this population. Antifungal treatment for severe cases of Candida peritonitis in ICU patients remains the standard care.