Unlike bacterial empyema, which commonly arises from pneumonia, Candida empyema is mostly associated with underlying malignancy. We describe our experience with Candida empyema.
Retrospective review of cases from 01/2009-12/2015, in which Candida was recovered from pleural fluid. We excluded Candida from drains in place >24 hours.
49 patients (pts) were identified. Median age was 65 yrs. 63% were men. 34% had malignancy, 21% were transplant recipients, and 4% were neutropenic. 81% had surgery within 1 yr (cardiothoracic (CT; 38%), abdominal (13%), CT + abdominal (19%)). 26% developed sepsis. 22% were admitted with empyema. 83% were in an ICU within 3 d of diagnosis. Sources of empyema were 1) infection from another site (34%; abdomen, 27%; neck, 7%); 2) introduction during CT or liver surgery (23%; VATS , 11%; transplant, 9%; heart , repeated thoracentesis, 2% each); 3) iatrogenic esophageal leaks (17%); 4) spontaneous esophageal perforation (11%); 5) unknown/cryptic (15%). 38% of pts had concomitant bacterial infections. 11% developed empyema during antifungal prophylaxis. C. albicans (66%) and C. glabrata (28%) were most common. 11% were co-infected by C. albicans and C. glabrata. 6% had concomitant candidemia. Surgical drainage was performed in all pts. 19% required VATS and decortication; 15% underwent open thoracotomy. 83% received antifungal treatment (fluconazole, 64%; caspofungin, 17%; voriconazole, 2%). 30 days mortality was 19%. 19% had persistent/recurrent infection. Death was highest with esophageal perforation (60%). Death or recurrent infection occurred in 57% of pts with cryptic source. Sepsis and high SAPS II were associated with death (p≤0.01).
Candida empyema was usually not associated with malignancy, but rather most common in older pts with prior CT or abdominal surgery. Outcomes were best for cases stemming from another infection or medical intervention, in which the source was clear and correctable. Outcomes were poor in cryptic cases or cases due to acute esophageal rupture. Surgical drainage was a routine component of care. Numbers were too small to show benefit from antifungal therapy, but we advocate treatment for all pts since outcomes could not be reliably predicted from clinical findings.
P. Vergidis, None
R. K. Shields, Merck: Grant Investigator , Research support
Astellas: Grant Investigator , Research support
M. H. Nguyen, None
See more of: Poster Abstract Session