The relationship between death attributable to Candida spp and histopathological results is studied in an ICU population where Candida spp is isolated. Prospective study in non-neutropenic ICU patients in whom Candida spp. has been Candida spp has been detected. The following are defined: invasive candidiasis based on dissemination and multifocality, the therapeutic protocol, and the methodology of the postmortem study. Attributable mortality is distinguished according to statistical, histopathological and clinical criteria. The presence of Candida spp was observed in 145 cases of 3389 ICU discharges (4.3%). There are 120/145 cases (83%) classified as invasive candidiasis and 25/145 colonizations (17%). ICU mortality is 35% (51/145) and hospital mortality is 46% (67/145). The postmortem study has been carried out in 71% of those who died in the ICU (36/51). Candida albicans is the most frequently isolated species (87%), followed by Candida glabrata (18%). 24 candidemias and 3 endophthalmitis have been demonstrated. Isolation in bronchial secretions is the most frequent (86%), followed by pharynx (83%) and gastric secretions (78%); positive urine culture has been obtained in 36% of cases. The attributable mortality has been 25% by statistical methods, 28% by the postmortem study and 23% by clinical data. Abdominal surgery and non-Candida albicans are risk factors for attributable mortality, and correct antifungal treatment is a protective factor against said attributable mortality. In conclusion: histopathological attributable mortality is like that defined by both statistics and clinical data; the simplification of the definition of invasive candidiasis, in non-neutropenic critically ill patients, allows for faster and more targeted therapeutic action. The early antifungal treatment may improve therapeutic results in this population.