Accelerated Increase in Candida auris Bloodstream Infections during COVID-19 Pandemic, South Africa

Disclaimer: Early release articles are not considered as final versions. Any changes will be reflected in the online version in the month the article is officially released.

The COVID-19 pandemic coincided with rising secondary bloodstream infections (BSIs) from multidrug-resistant organisms, including Candida auris . To assess candidemia trends, we conducted a retrospective analysis of blood culture isolates from public and private laboratories in South Africa taken during January 2019–June 2022. We evaluated weekly aggregated Candida BSI counts and COVID-19 cases using segmented regression within an interrupted time-series framework. In total, 15,393 candidemia cases were identified, 70% from the private sector. C. parapsilosis accounted for 39% of cases, whereas C. auris represented 26%. The proportion of C. auris increased significantly from 17% in 2019 to 31% in 2021 (p Candida BSIs rose by 11 cases per week (p = 0.03), largely driven by C. auris (+5 cases/week; p C. auris in Candida BSIs and the urgent need for enhanced surveillance, diagnostics, and infection prevention.

The COVID-19 pandemic, caused by the novel respiratory virus SARS-CoV-2, created strains on the global healthcare system that were experienced mainly in the delivery of acute hospital services ( 1 , 2 ). Those challenges led to secondary healthcare-associated bloodstream infections (BSIs) caused by multidrug-resistant bacteria and fungi, partially related to increased background use of antimicrobial agents, steroids, and immunomodulatory agents ( 3 – 5 ). In addition, viral-induced immune-mediated damage by SARS-CoV-2 might have encouraged bacterial and fungal colonization ( 6 ).

Candida BSI is a growing concern worldwide. More than 90% of Candida BSIs are caused by 1 of 6 species: C. albicans , Nakaseomyces glabratus (previously C. glabrata ), C. tropicalis , C. parapsilosis , Pichia kudriavzevii (previously C. krusei ), and, increasingly, C. auris ( 7 ) . Those 6 species are also included in the World Health Organization Fungal Priority Pathogen List ( 8 ). In the EUROBACT-2 international cohort study, fungal pathogens accounted for ≈8% (230/2,927) of all healthcare-associated BSIs from 2,600 enrolled patients in 333 intensive care units in 52 countries during June 2019–January 2020 ( 8 ). Candida species were the most common fungi; 58% (133/230) were identified as non– C. albicans Candida species ( 9 ). Candida spp. was the third most common healthcare-associated BSI pathogen (≈12% [333/2,828]) in a prepandemic multicenter study from India ( 10 ).

The emergence of C. auris in South Africa was first recognized retrospectively from a BSI isolate collected in 2009, which was initially misidentified as C. haemulonii because of limitations in standard diagnostic methods ( 11 ). National surveillance during October 2012–November 2016 identified a total of 1,692 cases of confirmed or probable C. auris infections across South Africa; cases were concentrated in Gauteng Province, where 92% of cases were reported from private-sector hospitals ( 11 ). Nationally, the number of cases rose exponentially from 18 in the baseline period (October 2012–November 2013) to 861 in the corresponding final period (October 2015–November 2016) ( 11 ). By 2016–2017, C. auris was the third most common cause of BSI nationally, surpassing most other non– C. albicans species ( 12 ). Molecular epidemiology revealed that isolates in South Africa predominantly belonged to the African clade III, which is genetically distinct and almost universally resistant to fluconazole; ≈6% had reduced susceptibility to amphotericin B (MIC > 2 µg/mL) (13). Of note, 5% of isolates were resistant to both fluconazole and amphotericin B ( 13 ).

The COVID-19 pandemic coincided with a marked increase in C. auris outbreaks worldwide, particularly in intensive care settings ( 14 ). Several contributing factors were identified, including prolonged mechanical ventilation, widespread antimicrobial use, increased indwelling device exposure, and compromised infection prevention protocols ( 14 ). A systematic review ( 14 ) documented C. auris outbreaks in >10 countries, including India, Brazil, Mexico, Pakistan, the United States, and multiple countries in Europe; reported crude mortality rates among COVID-19 co-infected patients ranged from 30% to 72%. Evidence from Orange County, California, USA, further illustrated this trend ( 15 ); in long-term acute-care hospitals, the probability of new C. auris colonization during the first COVID-19 wave reached 22.5% (95% CI 18.5–26.6) within 30 days of admission. That increase coincided with widespread disruptions in infection prevention practices and shortages of protective equipment, and although incidence declined in later phases of the pandemic, the overall prevalence of colonization remained high ( 15 ). Schaefer et al. ( 16 ) reported that C. auris incidence at a major academic hospital in New York increased nearly 3-fold, from 2.6 cases/10,000 admissions in 2019 to 7.8 cases/10,000 admissions in 2022 ( 16 ). Similarly, national surveillance data from China revealed 1,846 laboratory-confirmed C. auris cases across 22 provinces; ≈80% of isolates were resistant to fluconazole and ≈7% were resistant to amphotericin B ( 17 ).

In this study, we aimed to evaluate the distribution of Candida species isolated from blood cultures from both private and public health sectors in South Africa and to describe changes during the COVID-19 pandemic. Permission to conduct this study was obtained from the South African Society of Clinical Microbiology and NHLS, and ethics approval was obtained from the Human Research Ethics Committee (Medical) of the University of the Witwatersrand (clearance no. M210752).

We conducted a secondary analysis of blood culture data archived in pathology laboratory information systems from January 1, 2019, through June 30, 2022, in South Africa. The study population consisted of all patients who had a blood culture submitted either to the National Health Laboratory Service (NHLS) or to 1 of 3 large amalgamated private pathology practices (Ampath [ https://www.ampath.co.za ], Lancet Laboratories [ https://www.lancet.co.za ], or PathCare/Vermaak and Partners [ https://www.vpath.co.za ]). The NHLS provides routine diagnostic pathology services to the public health sector, serving ≈83% of the population of South Africa, and has 60 laboratories that offer microbiology testing, including automated blood culture systems and culture identification ( 11 ). The 3 large private pathology groups provide a full range of microbiology testing for almost all inpatients in South Africa with private health insurance ( 11 ). The NHLS used several platforms to identify Candida to species level at the time of this study: Vitek 2 YST, API 20C Aux, or API ID 32C (bioMérieux, https://www.biomerieux.com ); Auxacolor (Bio-Rad Laboratories, https://www.bio-rad.com ); and Microscan (Beckman Coulter, https://www.beckmancoulter.com ) ( 11 ). The private pathology groups identified Candida primarily using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry methods (Bruker, https://www.bruker.com ; or bioMérieux). The total number of blood culture specimens collected during the study period was not available as a denominator. The private health sector represents a small proportion of the general population, and laboratory data from this sector might not be directly comparable with the public sector owing to factors such as specimen-taking practices ( 18 ). We included patients with blood cultures positive with any of the 6 most common Candida species ( C. albicans , C. auris , N. glabratus , P. kudriavzevii , C. parapsilosis , and C. tropicalis ) in this analysis. Blood culture data from NHLS were obtained from the surveillance data warehouse at the National Institute for Communicable Diseases (NICD) ( 19 ). Blood culture data from the private sector were obtained through the South African Society of Clinical Microbiology, which has formed a public–private partnership with the NICD ( 19 ).

We defined a case of Candida BSI was defined as illness in a person seen at a healthcare facility in South Africa who had a blood culture from which C. albicans , C. auris , N. glabratus , P. kudriavzevii , C. parapsilosis , or C. tropicalis was isolated. We regarded positive blood cultures with the same organism that had been collected within 30 days of a first positive blood culture as duplicates and excluded them.

We extracted COVID-19 hospital admission data from March 5, 2020, through June 30, 2022, from the inpatient national surveillance system, DATCOV. Established in April 2020 by the NICD, DATCOV served as an active surveillance program for COVID-19 hospital admissions for both public and private hospitals in South Africa ( 1 ). We extracted routine surveillance data for SARS-CoV-2 infections from the reengineered Notifiable Medical Conditions Surveillance System for the period March 1, 2020, through June 30, 2022. In July 2020, the NICD established the Notifiable Medical Conditions Surveillance System, which contains data for all SARS-CoV-2 tests (PCR or antigen detection) conducted in South Africa ( 20 ). January 2019–February 2020 was considered the prepandemic baseline, and March 2020–June 2022 was the COVID-19 pandemic period.

We reported categorical variables for blood culture data in tables as frequencies and percentages or presented them as bar charts and tables. We used the Pearson χ 2 test or Fisher exact test to compare Candida species distribution among years, sectors, and provinces. We performed all analyses using Stata version 15 (StataCorp LLC, https://www.stata.com ) or R Studio version 4.5.1 ( https://rstudio.com/products/rstudio ). We prepared and tabulated blood culture, DATCOV, and SARS-CoV-2 datasets in Microsoft Excel ( https://www.microsoft.com ). We used those tabulated datasets to plot graphs and apply 30-day moving averages for cases of C. auris and C. parapsilosis BSI. We conducted interrupted time series analysis to assess the effects of the COVID-19 pandemic on the incidence of Candida BSI, with a specific focus on C. auris . We defined the intervention point as the onset of the COVID-19 pandemic in South Africa (i.e., March 2020). We aggregated Candida BSI and COVID-19 case counts by week and scaled COVID-19 count data to enable visual comparison. We then plotted weekly counts to examine trends before and after the interruption. We used segmented regression to estimate changes in the level and slope of Candida BSI incidence associated with the pandemic. We annotated national lockdown levels 1–5 in increasing order of restrictiveness. We used interrupted time series models to evaluate both immediate and gradual changes in weekly Candida BSI rates.

Figure 1 . Total number of cases of Candida auris BSI by health sector in study of accelerated increase in C. auris BSIs during COVID-19 pandemic, by month and cumulative,...

Over the 42-month period, 15,393 cases of Candida BSI were reported: 3,415 cases in 2019, 4,229 cases in 2020, 5,315 cases in 2021 and 2,434 cases during January–June 2022. The private sector accounted for 70% (10,826/15,393) of all cases. Cases of Candida BSI increased across the country, from 4,048 during the 14-month baseline period (January 2019–February 2020) to 5,413 in the first 14 months of the pandemic (March 2020–April 2021), a 34% rise; cases then increased to 5,932 in the following 14-month period (May 2021–June 2022), a further 10% increase ( Figure 1 ). Cases in Gauteng Province accounted for 50% (7,706/15,393) of all Candida BSI cases, followed by KwaZulu-Natal Province (14% [2,163/15,393]) and the Western Cape Province (12% [1,809/15,393]). Similar distribution patterns were observed for those provinces in both private and public sectors ( Appendix Table).

C. parapsilosis accounted for 39% (6,080/15,393) of all Candida BSI cases in South Africa, followed by C. auris (26% [3,928/15,393]). Nationally, there were notable differences in the distribution of Candida species. The relative frequencies of cases of C. auris BSI increased from 17% (597/3,415) in 2019 to 31% (1,626/5,315) in 2021 (p C. parapsilosis , from 46% (1,555/3,415) in 2019 to 37% (1,955/5,315) in 2021 (p N. glabratus from 10% (332/3,415) in 2019 to 7% (359/5,315) in 2021 (p C. albicans, P. kudriavzevii , and C. tropicalis were relatively unchanged ( Table ). C. parapsilosis was the most common species in 8 of the 9 provinces in South Africa. C. auris was the dominant species in the North West Province and accounted for 42% (312/749) of cases ( Appendix Figure). In the private sector, C. parapsilosis was the dominant species, accounting for 44% (4,742/10,826) of cases, followed by C. auris, which accounted for 32% (3,507/10,826) of cases. We observed a more marked increase in the percentage of cases of C. auris BSI in the private sector, from 24% (580/2,393) in 2019 to 37% (1,439/3,880) in 2021 (p C. parapsilosis , from 51% (1,217/2,393) in 2019 to 40% (1,567/3,880) in 2021, with a further decrease to 37% (629/1,684) in 2022 (p C. albicans was the dominant species, accounting for 41% (1,883/4,567) of cases, followed by C. parapsilosis , which accounted for 29% (1,338/4,567) of cases. Similar to the private sector, the percentage of cases of C. auris BSI in the public sector increased substantially, from 2% (17/1,022) in 2019 to 13% (187/1,435) in 2021 (p C. parapsilosis , from 33% (338/1,022) in 2019 to 27% (388/1,435) in 2021 (p N. glabratus BSIs, from 16% (160/1,022) in 2019 to 12% (169/1,435) in 2021 (p = 0.01) ( Table ).

Figure 2 . Cases of BSI by type in study of accelerated increase in Candida auris BSIs during COVID-19 pandemic, by year and age group, South Africa, January 2019–June 2022. A, B)...

Of the 15,393 total cases, 54% (8,333) of patients were male and 46% (7,060) female. The highest number of cases of Candida BSI were seen in persons Figure 2 , panels A, B). With the exception of 2 age categories (21–30 years of age and 51–60 years of age) in the private sector, we observed distinct differences among the other age categories across all 4 years. We observed that 74% (201/273) of persons C. auris BSI were from the public sector. In contrast, almost all C. auris BSI cases in persons 51–60 years of age (98% [821/834]) and in persons 61–70 years of age (99% [769/775]) were reported from the private sector ( Figure 2 , panels C, D).

The first laboratory-confirmed case of SARS-CoV-2 infection was documented in South Africa on March 5, 2020. During the study period, 3,995,881 cases of COVID-19 were reported; 486,789 of those resulted in hospital admission. More than half of COVID-19 hospital admissions (52% [254,975/486,789]) occurred in the public sector.

Figure 3 . Total number of cases of Candida BSIs (n = 15,393), COVID-19 cases (n = 3,995,881), and COVID-19 hospital admissions (n = 486,789), by month and 30-day moving average, in...