The Premier Healthcare Database (PHD; https://offers.premierinc.com/rs/381-NBB-525/images/PremierHealthcareDatabaseWhitepaper.pdf) captures real-world information on healthcare utilization and pharmacy activity from >1400 US hospitals and health systems, accounting for ~25% of all national hospital admissions. Over 500 of these facilities also submit laboratory results to the database; from those, we identified all hospitalizations during 2016–2024 in which Candida was isolated from a blood culture. Concurrent conditions were identified using International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) diagnosis codes (Supplementary Table 1) and devices, antifungal treatments, and other healthcare utilization using billing records.We calculated candidemia rates per 100 000 hospitalizations among facilities contributing laboratory data and summarized clinical characteristics, outcomes, and AFST results, both overall and by species (C. albicans alone, C. glabrata alone, and other Candida species). We tabulated yearly numbers and percentage distributions of causative species. We did not perform statistical testing or present P-values because the large sample size can render nonmeaningful differences statistically significant, and the study sample is a large convenience sample of US hospitals.
RESULTS
During 2016–2024, 13 177 candidemia-associated hospitalizations were identified, involving at least 12 864 unique patients at 361 hospitals and representing a rate of 47.1 per 100 000 hospitalizations (Supplementary Table 2). The number and rate of candidemia-associated hospitalizations varied across the period, increasing from 1380–1620 hospitalizations (range: 35.9–50.8 per 100 000) during 2016–2019 to 1957–2020 (range: 66.6–67.1) during 2020–2021 and then decreasing to 837–1325 (range: 41.5–51.1) during 2022–2024. Overall, rates were highest for hospitalizations involving patients who were aged 45–64 years (68.4) and ≥65 years (55.3), male (58.3), non-Hispanic Black (57.3), and hospitalized in the Southern US Census region (51.6). Demographic distributions were generally similar across Candida species, except that median age was higher among hospitalizations involving C. albicans (64 years; interquartile range and (66; 55–75) compared with other species (59; 43–70).
C. glabrata
Candida
Candida albicans constituted 35%–37% of candidemia-associated hospitalizations during 2016–2022, decreasing to 32% in 2024 (Figure 1). Candida glabrata increased steadily, becoming the most common species in 2024 (34%). Candida parapsilosis fluctuated between 12% and 17% during 2016–2023, before declining to 8% in 2024. Candida tropicalis remained relatively stable (6%–9%) during 2016–2024. No candidemia-associated hospitalizations involved Candida auris during 2016–2018; its prevalence remained <2% during 2019–2021, and then increased to 2% in 2022 and 5% in 2023 and 2024, making it the fifth most common species by 2024. Hospitalizations in which more than 1 Candida species was isolated were categorized as mixed-species candidemia and accounted for 5%–7% of candidemia-associated hospitalizations during 2016–2024.
Overall, candidemia hospitalizations involving C. glabrata more often had diabetes mellitus documented (51%) than did those involving C. albicans (41%) or other Candida species (38%) (Supplementary Table 3). Candidemia hospitalizations involving other Candida species more often had documented opioid-related disorders (11%) than did those involving C. albicans (8%) or C. glabrata (5%). From 1 October 2020 (when the coronavirus disease 2019 [COVID-19] ICD-10-CM code was standardized and fully implemented) through 31 December 2024, candidemia hospitalizations involving C. albicans more often had diagnosis codes for COVID-19 (21%) than did those involving C. glabrata (6%) or other Candida species (7%).
Overall, echinocandin receipt was more common in hospitalizations involving non-albicans Candida species (76% for C. glabrata, 75% for other Candida species, 67% for C. albicans). Hospitalizations involving C. albicans also had a longer interval between admission and collection of the first Candida-positive blood culture (4 days; IQR, 0–13) than those involving C. glabrata (2; 0–8) or other Candida species (2; 0–12). Median length of stay was similar for hospitalizations across species (C. albicans [14; 7–27], C. glabrata [13; 7–24], and other Candida species [14; 7–29]). Invasive mechanical ventilation was more frequent among hospitalizations involving C. albicans (49%) compared with C. glabrata (45%) and other Candida species (41%). Crude mortality (inhospital death or discharge to hospice) was highest among hospitalizations involving C. albicans (42%) and C. glabrata (41%), compared with other Candida species (33%).
In total, 5431/13 177 (41%) candidemia-associated hospitalizations had any AFST data available. Among those, 135 of 2697 tested isolates (5%) were fluconazole resistant (30/1045 [3%] for C. albicans, 53/589 [9%] for C. glabrata, and 52/1063 [6%] for other Candida species), and 24/2242 (1%) were echinocandin resistant, primarily C. glabrata (20/589 [3%]). Cell sizes were too small to evaluate annual resistance trends.
DISCUSSION
In this analysis of >13 000 candidemia hospitalizations from 2016 to 2024, C. glabrata emerged as the most common causative species, surpassing C. albicans in 2024. Because this shift was most evident in the final year, continued surveillance is needed to determine whether it represents a sustained trend. This shift might reflect the broader use of fluconazole prophylaxis favoring selection for non-albicans Candida species [4]. Approximately 9% of tested C. glabrata isolates had fluconazole resistance, slightly higher than in EIP surveillance data (4.9%) [2], underscoring the continued importance of the Infectious Diseases Society of America’s (IDSA) recommendation of echinocandins as first-line, empiric therapy for most candidemia cases [5, 6]. Our findings emphasize the need for laboratory capacity for species-level Candida identification and antifungal susceptibility testing, which IDSA recommends for all bloodstream isolates [6]. This capacity supports appropriate deescalation from echinocandins to fluconazole and is especially critical given the rising national prevalence of C. auris [7], which accounted for 5% of candidemia hospitalizations in 2023–2024 and for which echinocandins remain first-line therapy (https://www.cdc.gov/candida-auris/hcp/clinical-care/index.html). Consistent with EIP data, C. parapsilosis was the third most common causative species; it remains an emerging public health concern owing to its transmissibility in healthcare settings and rising fluconazole resistance [2, 8]. Overall, echinocandin resistance remained infrequent (1%), similar to EIP surveillance (<1%) [2]. This analysis also identified higher candidemia hospitalization rates for patients who were male, non-Hispanic Black, and aged ≥45 years, consistent with what has been previously described [2].
Our analysis suggested species-specific differences in clinical characteristics, treatment patterns, and outcomes for candidemia by Candida species. COVID-19–associated candidemia cases more frequently involved C. albicans (21%) than those involving other Candida species (6%–7%), and crude mortality was similar for hospitalizations involving C. albicans (42%) and C. glabrata (41%) and was substantially higher for both than for other Candida species (33%). These mortality differences might be partly explained by the higher prevalence of COVID-19 among C. albicans hospitalizations, which could reflect the dominance of C. albicans as an endogenous colonizer of mucosal surfaces and the substantial mucosal barrier disruption associated with COVID-19 and invasive mechanical ventilation receipt [9-11]. However, C. albicans candidemia hospitalizations still demonstrated relatively high mortality rates even after stratifying by COVID-19 status (Supplementary Table 4). This might relate to the organism’s capacity for hyphal and pseudohyphal growth, enhanced tissue invasion, and the strong inflammatory host responses it elicits [12]. The longer median time to detection for C. albicans (4 days vs 2 days for other species) might reflect later-onset, healthcare-associated infection, which could partly contribute to the high mortality observed among hospitalizations involving C. albicans. In an April–August 2020 EIP candidemia cohort, C. albicans occurred in 43.8% (28/64) of COVID-19–associated cases versus 35.8% (67/187) of non-COVID-19–associated cases, but overall differences in species distribution were nonsignificant, potentially reflecting limited statistical power [10].
Premier Healthcare Database data represent a convenience sample and may not be generalizable to the entire United States. Participating hospitals can change year-to-year; hospitals in the Western region are particularly underrepresented in PHD laboratory data [8]. Another limitation is the potential for misclassification inherent in medical coding. Additionally, the dataset does not allow identification of unique patients hospitalized multiple times at different facilities or recognition of outbreak-associated cases—an important consideration for C. auris and C. parapsilosis, pathogens known to spread in healthcare settings [7, 13]. Lastly, a small proportion of cases were categorized as unspecified Candida species, which could modestly affect observed species trends.
Despite these limitations, this study is, to our knowledge, the largest cohort of candidemia hospitalizations to date. Our findings supplement traditional case-based surveillance efforts, highlight the continued emergence of non-albicans Candida species, and reinforce the importance of adherence to IDSA candidemia guidelines.
Supplementary Material
Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
JL
Dr. Jessica Little
Infectious Disease · Transplant/Immunocompromised StatesBrigham and Women's Hospitaljessical