Retrospective Study of Outcomes Comparing Initial Treatment with Fluconazole or Micafungin in Immunosuppressed Patients with Candidemia

Background: Candidemia (CAND) is associated with poor outcomes including mortality rates as high as 47%, prolonged hospital stay, and increased healthcare costs. Echinocandins are recommended as front-line therapy for CAND; however, data among immunocompromised patients are lacking.

Methods: Retrospective cohort study of immunocompromised, adult patients with CAND at NewYork-Presbyterian Hospital from 2012-2014. Patients who received ≥3 consecutive days of initial treatment with fluconazole (FLUC) or micafungin (MICA) were included. Those with CAND in the prior 30 days, polymicrobial bloodstream infections, or C. krusei CAND were excluded. The primary outcome was complete response (CR) at day 14, defined as clinical improvement and sterilization of blood cultures. Secondary outcomes included microbiological and clinical success at day 14, survival at day 14 and 28, and recurrent CAND within 90 days.

Results: 71 patients were included: 26 received FLUC and 45 received MICA. Median age was 63 years, 78% had malignancy, 21% recent chemotherapy, 24% received solid organ transplant, and 4% were neutropenic. C. glabrata (39%) was most common species followed by C. albicans (35%). Baseline demographics were similar between FLUC and MICA groups. Patients who received MICA were more likely to require vasopressors (p=0.036), have acute kidney injury (AKI) (p=0.046), and have a higher Pitt Bacteremia score (PBS) (p=0.055) at time of CAND. There was no difference in CR at day 14 between the FLUC and MICA groups (54% vs. 49%; p=0.876). Time to antifungal therapy was not different between groups (42 vs. 47 hrs; p=0.185) or between those achieving a CR compared to those that did not (45 vs. 42 hrs; p=0.617) and outcomes did not differ by Candida spp. Microbiological and clinical success at day 14, survival at day 14 and 28, and recurrent CAND within 90 days were also similar between treatment groups. In a multivariable analysis including MICA therapy, PBS, AKI, and need for vasopressors, only PBS was an independent predictor of 14-day CR (OR 0.65 95% CI 0.48,0.90; p=0.008).

Conclusion: We have demonstrated that clinical responses rates are similar among immunosuppressed patients with CAND who are treated with FLUC or MICA. As with prior studies, these data indicate that severity of illness is a primary determinant of patient outcomes, and not necessarily the choice of antifungal agent.