Candida auris is a highly transmissible and frequently drug-resistant emerging fungal pathogen capable of causing severe infections. C. auris can colonize skin, leading to infection and transmission in healthcare settings. In the United States, reported clinical cases increased by 95% during 2020–2021 (1). US data on C. auris come primarily from case series and outbreak investigations and are geographically limited, and national surveillance data lack detail on patients’ underlying conditions, healthcare use, and outcomes. Therefore, we used a large healthcare services database to describe features of hospitalized patients with C. auris infection or colonization.

The PINC-A1 Healthcare Database (PHD) (https://offers.premierinc.com/rs/381-NBB-525/images/PINC_AI_Healthcare_Data_White_Paper.pdf) is a hospital-based all-payer database that contains healthcare use, financial, and pharmacy data from >1,000 US hospitals. Laboratory data are available from ≈25% of those hospitals. We identified all hospitalizations with a culture positive for C. auris from any specimen type during 2017–2022. We used diagnosis codes from the International Classification of Diseases 10th Revision, Clinical Modification, to identify underling conditions and complications (Appendix Table) and billing data to identify medical devices. We assessed features of C. auris hospitalizations and compared those with versus those without bloodstream infection (BSI) by using χ², Fisher exact, and Wilcoxon tests (α = 0.05).

A total of 192 C. auris hospitalizations (38 [20%] with BSI) occurred in 42 hospitals. C. auris hospitalizations primarily occurred among older adults (median age 68 years [range 21–89 years]), male patients (54%), and non-Hispanic White patients (60%). Non-Hispanic Black patients more frequently had BSI than did other races/ethnicities (39% vs. 29%; p = 0.022) (Table). The first positive C. auris specimen was collected within 2 days of admission for 63% of bloodstream and 48% of nonbloodstream C. auris hospitalizations. Among hospitalizations with bloodstream C. auris, 58% also had another positive specimen type. Among hospitalizations without bloodstream C. auris, the most common positive specimen types were axilla (38%) and urine (34%).

Underlying conditions and complications were similar for patients with bloodstream and nonbloodstream C. auris and most commonly were sepsis (64%), diabetes (55%), chronic kidney disease (44%), and pneumonia (43%). Compared with nonbloodstream C. auris, bloodstream C. auris hospitalizations more frequently involved central venous catheters (CVC) (76% vs. 53%; p = 0.010) and tracheostomies (29% vs. 12%; p = 0.008). Echinocandin use was more frequent for bloodstream (76%) versus nonbloodstream (25%) hospitalizations; median time from first positive culture to echinocandin use was 2 days (interquartile range 1–3 days).

Most (75.5%) hospitalizations involved an intensive care unit stay; mechanical ventilation was used in 43% of hospitalizations. Median hospitalization length was 13 days (range, 1–209 days). In-hospital mortality rate was 21%; discharge locations included hospice (13%), skilled nursing facility (28%), and long-term acute care (15%). Estimated crude mortality rates were 47% for bloodstream C. auris vs. 31% for nonbloodstream.

This analysis of a large convenience sample of C. auris‒associated hospitalizations provides information about clinical features that are currently unavailable through national public health surveillance. Our results support smaller previous investigations showing that infection and colonization with C. auris occurs most commonly in patients with complex medical conditions (2–5). The proportion of C. auris cases involving BSI (20%) was comparable to the 9%–28% BSI rate among clinical and screening cases found in previous state-specific studies (2,6). Including in-hospital deaths and discharges to hospice, the overall estimated crude mortality rate of 34% (47% for BSI) was similar to the 30-day mortality rate from a previous study in New York (27% overall and 39% for BSI) (2).

Consistent with candidemia treatment guidelines, most BSI hospitalizations involved echinocandin use, and treatment lag was typical (7). Hospitalizations involving C. auris BSI were associated with non-Hispanic Black race, similar to those for non–C. auris candidemia (8). The association between CVC use and C. auris BSI is not surprising, given that CVC use is a well-documented risk factor for candidemia and is common among patients with C. auris, because extensive healthcare exposure, intensive care unit stays, and use of medical devices are key factors in C. auris acquisition (2,9). Many BSI patients were probably admitted with C. auris, based on first positive blood specimens occurring soon after admission, similar to finding from a New York case series (3). However, we could not assess previous healthcare exposures and prehospitalization laboratory data, which are major considerations because patients with C. auris usually acquire it in healthcare settings and can remain colonized for months (9).

Other limitations of our study include lack of antifungal susceptibility testing data, possible underdetection of cases caused by potential incompleteness of 2022 data, and underrepresentation of the West and the Northeast regions in PHD laboratory data (10), which is particularly relevant because those regions report high case counts (https://www.cdc.gov/fungal/candida-auris/tracking-c-auris.html). PHD laboratory data also underrepresent rural, smaller hospitals (10), potentially further biasing this convenience sample of C. auris cases. In conclusion, this analysis of hospitalization data supports previous targeted reports and demonstrates a need for strengthened national surveillance and further studies to identify risk factors for C. auris infection and colonization.

Top