This study was performed to evaluate trends in species distribution in patients’ with hematogenous candidiasis at a comprehensive cancer center. The results of a retrospective analysis from January 1, 1993 to December 31, 1998 were compared with prior reports from Memorial Sloan-Kettering Cancer Center in the last forty years. In 570 total episodes since 1974, 43.9% were due to Candida albicans. During 1990’s, C. parapsilosis emerged as the most frequent yeast species in the non-C. albicans group (36.1% during 1993-1998 from 20.9% 1974-1982; p < 0.01). An increase in C. krusei from 5.9% (1974-1982) to 10.5% during the recent six years (1993-1998) was also noticed. The proportion of C. tropicalis among non-albicans fungemia during 1974-1982 was 42.8%, whereas in 1993 to 1998 a marked decline in C. tropicalis hematogenous infection was observed (27.8%; p < 0.01). During 1998, the incidence of candidemia declined from 7.1% (1972-1973) and 6.5% (1982) to 3.4% (p < 0.01), and improved survival among fungemic patients (33% mortality in 1998; 77.3% during 1974-1982; p < 0.001) was encouraging. The increase in C. parapsilosis bloodstream invasion during 1990’s was associated with a significant reduction in the endogenous non-albicans Candida tropicalis infection that probably resulted in part due to the common prophylaxis, and/or preemptive fluconazole given routinely in high-risk patients undergoing treatment for cancer. The widespread use of extraneous implantable and/or semi-implantable indwelling intra-vascular devices may also have played an important role in promoting (exogenous) C. parapsilosis infection. This study emphasizes the importance of periodic evaluation of candidemia, especially at centers caring for patients at risk.

Most hematogenous candidiasis originates from endogenous host flora. The impact of clinically prominent Candida colonization on short-term mortality (<or=14 weeks) was prospectively studied in 193 hospitalized patients from 1998 to 1999. Clinically prominent colonization included yeasts isolated from all sterile body sites and >50 colonies of Candida from non-sterile sites. Fourteen (7.1%) patients were granulocytopenic (ANC <or=100/microl). Nineteen (9.8%) had undergone marrow transplantation, 26 (13.5%) had a hematologic malignancy and 129 (66.5) had non-hematologic cancer. Candida isolates (216) were collected form 210 specimens. Fifty-three (27.5%) patients died; 25 (19.4%) with solid tumors, compared to 16 (61.5%) with hematological malignancy, and 11 (57.9%) BMT recipients (P < 0.001). No deaths were seen in patients with AIDS, and one (7.7%) in a patient with a benign condition (P < 0.001). Twenty-six (29.2%) patients with respiratory tract, 13 (23.2%) with gastrointestinal tract, and three (14.3%) with genitourinary tract colonization died. In patients with multiple-site colonization, mortality was significantly higher (45.5%) (P < 0.05). Mortality was higher in patients with C. glabrata (52.9%) and C. krusei (75%) colonization than with C. albicans (24.1%) (P < 0.025). This study shows that patients with hematologic cancer and recipients of marrow transplant with Candida colonization of multiple body sites, and colonization with C. glabata or C. krusei have poor survival.