Candidemia

We defined a case of candidemia as illness in any patient at a healthcare facility in South Africa who had Candida species isolated from a blood culture specimen processed by an NHLS or private-sector diagnostic laboratory. We defined a confirmed case of C. auris candidemia as illness in a patient with an isolate confirmed as C. auris at NICD, regardless of the referring laboratory’s initial identification. We also included probable cases for which the referring laboratory identified C. auris or Candida haemulonii but a viable isolate was not available for confirmation at NICD. Multiple Candida isolates cultured within 30 days of the first positive blood culture specimen were included in a single case. We classified cases of candidemia into 2 groups on the basis of NICD identification (or the referring laboratory’s identification if a viable isolate was not available): C. auris and non–C. auris candidemia.

We defined an episode of candidemia defined as the first positive blood culture drawn from a peripheral vein that yields a Candida species, where the clinical symptoms and signs were compatible with Candida bloodstream infections (BSIs) [9 (link),21 (link)]. Late recurrent candidemia was defined as the second episode of candidemia that occurred at least 30 days after the last positive blood culture positive for Candida [9 (link),16 (link),21 (link),22 (link)]. For two different Candida species recovered within one week in the same patient, it was considered to be polyfungal candidemia. In cases of candidemia with more than two negative blood cultures, resolution of all septic symptoms and completion of antifungal treatment, the next positive blood culture yielding the same Candida species was sent to molecular diagnostics to document whether it was relapse or re-infection. Relapses were episodes caused by the same genus and species; re-infection was considered to be the second episode growing different Candida species after one week.
The clinical information was from the review of medical charts, and included demographic characteristics, predisposing risk factors within the preceding 30 days from the onset of Candida bloodstream infection (BSI) (defined as the day of first positive blood culture for Candida spp.), underlying diseases, and the presence of an intravenous catheter or any other artificial device at the time of candidemia.

The targeted antimycotic prophylaxis included micafungin until the end of March 2019. Thereafter, anidulafungin was used based on the overarching recommendation of the local drug commission. In the case of pre-existing fungal colonization with echinocandin-resistant Candida spp. or Aspergillus spp., the prophylaxis was switched to fluconazole, voriconazole, or liposomal amphotericin B, and the patient was consequently excluded from the study.
Antimycotic prophylaxis was started as soon as the criteria for HR-LTRs were fulfilled. If TAP was started on the day of operation, it was rated as “immediate”, and “delayed” when started during the postoperative course. We used micafungin in a dosage of 100 mg/d or a loading dose of anidulafungin (200 mg), followed by 100 mg/d, both given intravenously.
In the case of a diagnosed infection, we adapted the antifungal regime as follows: fluconazole (800 mg loading dose, 1200–1600 mg with a body mass index >30 kg/m²), followed by a maintenance dose of at least 400 mg (600–800 mg with body mass index >30 kg/m²). The dosage was furtherly adjusted according to the renal function or in the case of renal replacement therapy. Voriconazole was initiated by two loading doses of 6 mg/kg every 12 h, followed by a maintenance dose according to a weekly performed therapeutic drug monitoring. Isavuconazole was started with 200 mg every 8 h for two days, followed by a daily dose of 200 mg. Liposomal amphotericin B (L-AmB) was dosed at 3 mg/kg per day.
The echinocandin prophylaxis was carried out over a period of minimum 7 to 14 days, and prolonged in the case of a diagnosed IFI or on clinical decision by the intensivist. Reasons for discontinuation of the prophylaxis were the completion of prophylaxis at discharge or missing clinical signs of infection, death, or switch to the therapeutic regime in the case of diagnosed infection.
Echinocandin therapy was continued in the case of invasive candidiasis and a positive response to therapy. In the case of confirmation of a fungal pathogen outside the echinocandin’s spectrum of activity or if a salvage therapy was indicated by the treating clinician, a switch to another agent was performed.
Antimycotic susceptibility was assessed according to the breakpoints determined by the European Committee on Antimicrobial Susceptibility Testing Subcommittee on Antifungal Susceptibility Testing [49 (link)]. Given the limited testing of echinocandins as first-line therapy of invasive aspergillosis, voriconazole was used as a first-line agent for aspergillosis, and isavuconazole as an alternative in the case of voriconazole-caused side effects or a suspected mucormycosis. [50 (link),51 (link)] Liposomal amphotericin B or combination therapy of different antifungal agents has been used as a last option in critically ill patients. Fluconazole was used as first-line therapy for invasive Candida parapsilosis infections [22 (link),23 (link),24 (link),33 (link),48 (link),52 (link),53 (link)].
The duration of treatment continued for at least 14 days after the time of the last negative blood culture in the case of candidemia or until all clinical signs and symptoms had resolved.

A “proven” IFI was defined according to the definitions of the European Organization for Research and Treatment of Cancer/Invasive Fungal Infections Cooperative Group and the National Institute of Allergy and Infectious Diseases Mycoses Study Group (EORTC/MSG) Consensus Group [43 (link)]. Infections during ongoing critical care treatment were rated as “probable” upon the recommendations of the EORTC/MSGERC ICU Working Group [44 (link)].
Besides clinical symptoms of an infectious disease process, fundoscopic findings or hepatosplenic lesions by computed tomography were accepted as clinical criteria. Mycological criteria were defined by positive serum 1,3-β-d-glucan in two consecutive samples or recovery of Candida in an intra-abdominal specimen obtained surgically or within 24 h from external drainage. Candidemia was defined as the isolation of Candida spp. from at least one blood culture. An isolated Candida peritonitis without candidemia was rated as invasive only in cases with histopathologic or direct microscopic examination of perioperatively sampled sterile fluid or tissue. Positive samples taken from drains more than 24 h after surgery, as well as Candida isolation from respiratory secretions, stool, skin, and wound sites, and an asymptomatic candiduria were interpreted as colonization or “possible” infection in the case of clinical signs of sepsis [33 (link),44 (link),45 (link),46 (link),47 (link)].
Detection of Aspergillus spp. was only rated as “probable” invasive aspergillosis (IA) after positive mycological evidence of Aspergillus spp. either via cytology, direct microscopy, and/or culture in a lower respiratory tract specimen, or via a galactomannan antigen index >0.5 in plasma/serum, and/or galactomannan antigen >0.8 in bronchoalveolar lavage fluid (BALF) in the case of at least one diagnostic sign in computed tomography or bronchoscopic proof of a tracheobronchitis.
In the case of an IFI, computed tomography, transesophageal echocardiography, and fundoscopy were routinely performed to detect organ involvement.
Breakthrough IFI was defined as any IFI occurring during ongoing prophylaxis, including fungi outside its spectrum of activity. The time of b-IFI occurrence was defined as the first attributable clinical sign or symptom, mycological finding, or radiological feature. A relapse IFI was defined as occurrence after treatment and being caused by the same pathogen at the same site, although dissemination can occur [48 (link)].

We performed a retrospective, matched, case-control study in two university-affiliated tertiary hospitals in South Korea from January 2019 to August 2020 (Samsung Changwon Hospital, 760-bed; Inje University Busan Paik Hospital, 818-bed). Patients (age ≥ 18 years) with candidemia were included if they had at least one positive blood culture for Candida species. For each case, one control matched for age (±5 years), duration of hospitalization, hospital ward (intensive care unit (ICU)/non-ICU), and type of surgery was identified within the same hospital [12 (link)]. Duration of hospitalization in the cases was calculated as the time from the day of admission to the day of collection of the first positive blood culture with Candida species. Matched controls remained hospitalized for the equivalent time and did not develop candidemia during hospitalization. The hospital ward was assessed based on the index date (day of occurrence of candidemia in the cases or matched day in the controls). Surgical procedures within the four weeks before the index date was identified and classified as no surgery, hepatobiliary/gastrointestinal surgery, genitourinary surgery, other abdominal/pelvic surgery, cardiothoracic surgery, or other major surgery [10 (link)]. Patients who had neutropenia (absolute neutrophil count of <500 cells/mm³) or for whom we could not identify previous antibiotic therapy were excluded. Moreover, we excluded cases for which matched control patients could not be identified. The primary objective was to identify antibiotic factors associated with candidemia. The secondary objective was to identify differences in antibiotic factors according to the origins of candidemia, and the origins of candidemia were classified into catheter-related bloodstream infection (CRBSI) and non-CRBSI. The study was approved by the Samsung Changwon Hospital and Busan Paik Institutional Review Boards (IRB numbers: SCMC 2021-08-005 and BPIRB 2021-08-051).