Direct treatment costs caused by candidemia in German intensive care unit (ICU) patients are currently unknown. We analyzed treatment costs and the impact of antifungal drug choice. Comprehensive data of patients who had at least one episode of candidemia while staying in the ICU between 01/2005 and 12/2010 were documented in a database using the technology of the Cologne Cohort of Neutropenic Patients (CoCoNut). A detailed analysis of all disease-associated treatment costs was performed. Patients treated with echinocandins (i.e., anidulafungin, caspofungin, micafungin) or fluconazole were analyzed separately and compared. Forty-one and 64 patients received echinocandins and fluconazole, respectively. The mean Acute Physiology and Chronic Health Evaluation (APACHE) IV score was 114 (95 % confidence interval [CI]: 106-122) vs. 95 (95 % CI: 90-101, p = <0.001). Twenty-three (56 %) and 33 (52 %, p = 0.448) patients survived hospitalization, while 17 (41 %) and 22 (34 %, p = 0.574) survived one year after diagnosis. In the echinocandin and fluconazole groups, the mean costs per patient of ICU treatment were 20,338 (95 % CI: 12,893-27,883) vs. 11,932 (95 % CI: 8,016-15,849, p = 0.110), and the total direct treatment costs per patient were 37,995 (95 % CI: 26,614- 49,376) vs. 22,305 (95 % CI: 16,817- 27,793, p = 0.012), resulting in daily costs per patient of 1,158 (95 % CI: 1,036- 1,280) vs. 927 (95 % CI: 828- 1,026, p = 0.001). Our health economic analysis shows the high treatment costs of patients with candidemia in the ICU. Sicker patients had a prolonged hospitalization and were more likely to receive echinocandins, leading to higher treatment costs. Outcomes were comparable to those achieved in less sick patients with fluconazole.