key: cord-0778380-bt9aktm6
authors: Arastehfar, Amir; Ünal, Nevzat; Hoşbul, Tuğrul; Özarslan, Muhammed Alper; Karakoyun, Ayşe Sultan; Polat, Furkan; Fuentes, Diego; Gümral, Ramazan; Turunç, Tuba; Daneshnia, Farnaz; Perlin, David S; Lass-Flörl, Cornelia; Gabaldón, Toni; Ilkit, Macit; Nguyen, M Hong
title: Candidemia among COVID-19 patients in Turkey admitted to ICUs: A retrospective multicenter study
date: 2022-02-13
journal: Open Forum Infect Dis
DOI: 10.1093/ofid/ofac078
sha: 9c44ece35e3ab243dac3f6d9c140f1fd63503c06
doc_id: 778380
cord_uid: bt9aktm6

OBJECTIVES: We evaluated the epidemiology of candidemia among COVID-19 patients admitted to intensive care units (ICUs). METHODS: We conducted a retrospective multicenter study in Turkey between April- December 2020. RESULTS: Twenty-eight of 148 enrolled patients developed candidemia, yielding an incidence of 19% and incidence rate of 14/1,000 patient-days. The probability of acquiring candidemia at 10, 20 and 30 days of ICU admission was 6%, 26% and 50%, respectively. Over 80% of patients received antibiotics, corticosteroid and mechanical ventilation. Receipt of a carbapenem (odds ratio and 95% confidence interval (OR, 95% CI) of 6.0 (1.6–22.3), P=0.008), central venous catheter (4.3 (1.3–14.2), P=0.02) and bacteremia preceding candidemia (6.6 (2.1–20.1), P=0.001) were independent risk factors for candidemia. Mortality rate did not differ between patients with and without candidemia. Age (1.05 (1.01–1.09), P=0.02) and mechanical ventilation (61 (15.8–234.9), P<0.0001) were independent risk factors for death. Candida albicans was the most prevalent species overall. In Izmir, C. parapsilosis accounted for 50% (2/4) of candidemia. Both C. parapsilosis isolates were fluconazole non-susceptible, harbored Erg11-Y132F mutation, and were clonal based on whole-genome sequencing. The two infected patients resided in ICUs with ongoing outbreaks due to fluconazole-resistant C. parapsilosis. CONCLUSIONS: Physicians should be aware of the elevated risk for candidemia among COVID-19 patients requiring ICU care. Prolonged ICU exposure and ICU practices rendered to COVID-19 patients are important contributing factors to candidemia. Emphasis should be placed on heightened infection control in the ICU, and developing antibiotic stewardship strategies to reduce irrational antimicrobial therapy.

Invasive fungal infections, especially those due to Candida spp., are associated with huge economical burdens and high mortality rate. 1 Furthermore, emergence of drug-resistant Candida species, such as Candida auris, Candida parapsilosis and Candida glabrata, imposes a growing threat due to limited options of antifungal therapy. 2 The classic risk factors associated with candidemia include leukopenia, chronic renal failure, abdominal surgery, intensive care unit (ICU) stay, central venous catheters (CVCs), mechanical ventilation, and long-term use of corticosteroids. 3 Most recently, COVID-19 pandemic has predisposed millions of patients to secondary infections, including fungal infections. [4] [5] [6] [7] Indeed, an increased incidence of candidemia associated with COVID-19 and its associated high mortality rate has been reported. [8] [9] [10] Recent outbreaks due to multidrug resistant C. auris have also been noted in some centers. 11, 12 How COVID-19 patients are at risk for Candida infections is not fully understood. The reasons may be multifactorial, including immune dysregulation and organ damage resulting from SARS-CoV-2, acquired immunodeficiency state stemming from immunomodulatory agents administered to treat severe cases of COVID-19, and breach in standard healthcare practices of infection prevention and antibiotic stewardship. 13 To provide a further insight into candidemia associated with COVID-19, we conducted a multicenter study in Turkey to assess the incidence rate and cumulative risk of ICUacquired candidemia, identify factors predisposing to candidemia, and evaluate the risk factors for mortality.

A c c e p t e d M a n u s c r i p t

This retrospective study was conducted at three hospitals in Turkey (Adana City Hospital, Adana, Ege University Medical Faculty Hospital, Izmir, and Gülhane Training and Research Hospital, Ankara). From April through December 2020, all patients with COVID-19 admitted to the participating ICUs were included in the study.

Candidemia was defined as positive blood culture for Candida species. Candida isolates were speciated using matrix assisted laser desorption ionization-time of flight mass spectrometry at individual centers, however antifungal susceptibility was not performed as per standard of care. For this study, susceptibility testing was performed in accordance with CLSI M60-A3, 14 and non-susceptibility to azole agents was defined according to CLSI breakpoints. 4, 7, 14, 15 Immunosuppressed was defined as presence of an active malignancy or immunocompromised state (primary or secondary due to Human Immunodeficiency virus, hematopoetic stem cell or solid organ transplant, receipt of corticosteroid or other immunosuppressed agents, including biologics). Since steroid was administered to almost all patients in this study, we only included patients receiving steroid (> 20 mg prednisone-equivalent daily for > 10 days) prior to COVID-19 diagnosis in the definition.

We reviewed patients' electronic medical records and collected demographic characteristics and underlying medical conditions. Established risk factors for candidemia from published studies were also collected, including the presence of CVCs (and their duration), mechanical ventilation, renal replacement therapy, and use of antibacterial agents. 3, 16 Data on bacterial blood stream infections were also included.

A c c e p t e d M a n u s c r i p t

Fluconazole-resistant C. parapsilosis isolates were subjected to ERG11 Sanger sequencing using primers and PCR conditions reported elsewhere. 17

The genomes of 2 C. parapsilosis isolates from Izmir Hospital (isolates 35 and 37) underwent Illumina WGS using NovaSeq 6000. 18 Eight previously sequenced C. parapsilosis isolates, including two clonal pairs recovered from the same patient were included in the analysis of genomic variants for comparison. 19,20 Single Nucleotide Polymorphisms (SNPs) were identified using Freebayes 21 as implemented within the PerSVade version 0.10 pipeline. 18 The genome CDR-317 was used as reference. A multiple correspondence analysis (MCA) was performed as previously described. 19

Data analyses were conducted using Stata/SE v16. 

Over the study period, 148 patients were admitted to the ICU (Adana City Hospital, 99 patients, Ege University Hospital, 24 patients, and Gülhane Hospital, 25 patients).

Demographics and underlying diseases are presented in Table 1 . Twenty-eight COVID-19 patients developed candidemia, yielding a candidemic rate of 19%. This rate was 13-fold higher than the rate of patients without COVID-19 admitted to the ICU in the same study period (1.5%) ( Table 2 ). The probability of acquiring candidemia among COVID-19 patients within 10 days of ICU admission was 6%, but this rate increased to 26% at 20 days and 50% at 30 days ( Figure 1 ). The median duration from hospital admission for COVID-19 to ICU admission was 20 days (interquartile range (IQR): 12-28 days), and from ICU admission to candidemia was 12 days (IQR: 8-18 days) ( Figure 1 ).

Candida albicans was the most common species recovered (43%, 12/28), followed by C. parapsilosis (25%, 7/28), C. tropicalis (21%, 6/28), and C. glabrata, C. krusei and C. lusitaniae (4%, 1 each) (Supplemental Figure 1 and Supplemental Table 1 ). Forty-thee percent (3/7) of C. parapsilosis isolates was non-susceptible to fluconazole (MIC ≥4µg/mL) (Supplemental Tables 1 and 2 ). All fluconazole-non-susceptible C. parapsilosis isolates harbored Y132F mutation in the 14α-demethylase Erg11p. Of note, both fluconazole non-A c c e p t e d M a n u s c r i p t susceptible C. parapsilosis isolates from Izmir (isolates 35 and 37) were recovered from azole-naïve patients who resided in units with an ongoing clonal outbreak of fluconazoleresistant C. parapsilosis that started in 2015. 22 To evaluate the genetic relatedness of these two isolates, we performed Illumina whole genome sequencing. For comparison, we included whole genome sequencing data from eight C. parapsilosis strains from two previous studies: five isolates from the same patient (including two pairs of clonal isolates) in one study, 19 and three unrelated strains from clinical and environmental sources in the second study. 20 Our genomic analyses demonstrated that the two isolates in this study were closely related. Table 1 ). The duration of indwelling CVC was significantly longer among patients with candidemia than those without (18.5 versus 8 days, respectively, P=0.0001). Using logistic regression analysis, receipt of a carbapenem and bacteremia (overall or due to coagulase-negative Staphylococcus or Acinetobacter baumannii complex) were independent risk factors for candidemia ( Table 1) .

Information of antifungal therapy was available for 25 patients. Two patients were not treated with antifungals (one of whom died on the day of candidemia detection). Among the three patients infected with fluconazole-resistant Candida isolates, two were initially treated with fluconazole which was then transitioned to amphotericin B and one to an anidulafungin, while the other one was azole-naïve. Among the 19 patients infected with fluconazolesusceptible isolates, 13 were treated with fluconazole, 2 with an echinocandin, and 4 were initially treated with fluconazole which was transitioned to amphotericin B (2 patients) or to an echinocandin (2 patients). The overall in-hospital mortality rate was 84% (125/148). There was no significant difference in mortality rate between patients with or without candidemia (86%, 24/28 versus 84%, 101/120, P=1.0). The median time from candidemia diagnosis to death was 14 days (IQR: 4-23 days). Older age (P=0.03) and requirement of mechanical ventilation (P<0.0001) were independent risk factors for death among critically ill COVID-19 patients in ICU (Table 3) .

This multicenter study of critically ill COVID-19 patients requiring ICU admission identified several important findings. First, 19% of patients developed candidemia, a rate that is higher than previously reported in the literature (range: 2.5 to 14%). 23 This rate was A c c e p t e d M a n u s c r i p t increased by 13-fold among non-COVID-19 patients admitted to the same hospitals in the same time period. Second, COVID-19 patients with candidemia had prolonged hospital stay, with a median time of 12 days from ICU admission to candidemia diagnosis. Our finding is in line with a recent study by the Center for Disease Control and Prevention which showed that most candidemia was acquired more than one week after COVID-19. 7 The estimated cumulative risk for candidemia among our patients increased with longer ICU stay, from 6% at day 10 to 26% at day 20 and 50% at day 30 ( Figure 1) . Third, the majority of patients with candidemia received antibiotics (99%) and corticosteroid therapy (84%), and required mechanical ventilation (84%), all well-established risk factors for candidemia. Lastly, we identified bacterial bloodstream infection (especially due to coagulase-negative Staphylococcus and A. baumannii complex), receipt of a carbapenem and presence of CVC as independent risk factors for candidemia. Altogether, our data support the notion that prolonged ICU exposure and healthcare therapeutic interventions of COVID-19 patients were responsible for a higher rate of candidemia observed in our study.

During the COVID-19 pandemic, in order to limit patient contact, hospitals utilized mobile and out-of-room monitoring and device controls and extended dwell intravenous catheters. 24 Such approaches could impact infection control practices. Furthermore, the presence of airborne and contact isolation and cumbersome personal protective equipment might have rendered CVC placement more technically difficult. Along the same line, the increase in ICU patient census might have interfered with routine infection control practice such as surveillance and maintenance of CVC devices and favored utilization of these devices for regular blood draws. All of these factors likely adversely increase the risk of catheter contamination and catheter-associated infection. Indeed, 26% (39/148) of our patients had at least one episode of bacteremia, which were most commonly caused by Gram positive bacteria, especially coagulase-negative Staphylococcus, a finding that echoes published A c c e p t e d M a n u s c r i p t studies. 25, 26 In our study, 71% (20/28) of candidemic patients had preceding bacteremia, and bacteremia was an independent risk factor for subsequent candidemia. All the candidemic cases in our study were primary blood stream infection, as there was no apparent infection at another site, and 86% (24/28) of these cases was central line associated bloodstream infections (CLABSI). Increase in Candida-associated CLABSIs among hospitalized COVID-19 patients has been previously reported. 27

Over 80% of our patients received broad spectrum antibiotics, corticosteroid and mechanical ventilation, which are all classical risk factors for candidemia. 3, 16 Utilization of corticosteroid has significantly improved outcome in severely ill hospitalized patients with COVID-19, 28 and mechanical ventilation is required for patients with COVID-19 associated respiratory failure, thus their use is unlikely to be modifiable. However, the fact that 99% of patients received antibiotics is very concerning. Literature to date showed that bacterial coinfection with COVID-19 occurred in only ~14% of patients in ICU, 29 which is less prevalent than in patients with influenza. 30 Nevertheless, empiric antibiotic prescription in COVID-19 patients was widespread, 31 and similar to our finding, >90% of hospitalized COVID-19 patients were noted to receive empirical antibiotics. 32 As the manifestations of COVID-19 patients with cytokine release syndrome mimic bacterial sepsis, it could be difficult for physicians to withhold antibiotics in this setting.

The association with prolonged ICU acquisition, preceding bacteremia (especially with coagulase-negative Staphylococcus), and Candida CLABSI, suggest that candidemia is linked to infection control issues. Indeed, breach in infection prevention practices has been linked to outbreaks of C. auris throughout the globe. 11, 25, 33 Furthermore, the high rate of administration of antibacterial agents especially carbapenem (80%) might impact bacterial flora and promoted Candida growth, 34 and along with corticosteroid utilization, lead to the A c c e p t e d M a n u s c r i p t selection of Candida superinfection in these critically ill and medically complexed COVID-19 patients.

Overall, C. albicans was the most common species recovered, followed by C. parapsilosis and C. tropicalis. Two patients from Izmir were infected with genetically-related fluconazole non-susceptible C. parapsilosis isolates carrying a Y132F mutation in Erg11p.

The clonal outbreaks due to this particular strain of C. parapsilosis have been reported in numerous countries, [35] [36] [37] [38] and present a particular challenge as the strains persist and cause outbreaks despite application of disinfectants. 39 Centers dealing with clonal outbreak due to C. parapsilosis like ours should closely monitor the emergence of fluconazole-resistance, given its association with poorer outcomes. 39 Recently, C. parapsilosis isolates with Erg11-Y132F mutation has also been linked to echinocandin resistance, which further complicates treatment strategy. 40 Since these strains retain susceptibility to amphotericin B, 40 and the efficacy of this agent was shown in in vivo study, 41 we recommend amphotericin B as empiric antifungal therapy until susceptibility data are available among centers experiencing problem with fluconazole-resistant C. parapsilosis.

The mortality rates of COVID-19 in the ICU ranged from 50-65%, and higher among patients requiring mechanical ventilation. Age and need for mechanical ventilation were predictors for mortality among our patients. Unlike previous studies, 5,7,42 we did not detect a worse outcome among patients with candidemia compared with those without (mortality of 86% vs 84%, respectively). It is possible that, with an overall mortality rate of 84%, it is difficult to distinguish attributable mortality from death caused by candidemia or underlying diseases. Previous studies have suggested that the mortality attributed to candidemia is not significant in a population of patients with high expected mortality. 43 Moreover, in this setting, candidemia might merely a marker for severity of illness. Along this line, our data showed that the mortality rate for patients with candidemia was ~2-fold higher among A c c e p t e d M a n u s c r i p t patients with COVID-19 than those without (84% versus 50%). This finding suggests that COVID-19 may amplify the risk of death due to candidemia.

It is important to acknowledge that our study is limited by its retrospective design, and results may have been influenced by practices and patient populations at our hospitals.

Experiences at other centers may differ. Furthermore, the number of candidemia among COVID-19 patients was small, and we do not have detailed clinical data associated with non- M a n u s c r i p t M a n u s c r i p t [15] . The remaining five isolates correspond to serial isolates from the same patients where three distinct clonal complexes were identified [14] . Strains considered from the same clonal complex are circled.

A c c e p t e d M a n u s c r i p t 

Hidden killers: human fungal infections

The Quiet and Underappreciated Rise of Drug-Resistant Invasive Fungal Pathogens

Invasive candidiasis

Coronavirus Disease 2019-Associated Invasive Fungal Infection

COVID-19-Associated Candidiasis (CAC): An Underestimated Complication in the Absence of Immunological Predispositions?

COVID-19 Associated Pulmonary Aspergillosis (CAPA)-From Immunology to Treatment

The landscape of candidemia during the COVID-19 pandemic

Candidemia in Coronavirus Disease 2019 (COVID-19) Patients: Incidence and Characteristics in a Prospective Cohort Compared With Historical Non-COVID-19 Controls

Characteristics of candidemia in COVID-19 patients; increased incidence, earlier occurrence and higher mortality rates compared to non-COVID-19 patients

Candidemia among Iranian Patients with Severe COVID-19 Admitted to ICUs

Specialty Care Unit -Florida

Sopirala MM. Predisposition of COVID-19 patients to secondary infections: set in stone or subject to change?

Institute CaLS. Reference method for broth dilution antifungal susceptibility testing of yeasts. Fourth edition. CLSI Standard M60 CLSI

Progress in antifungal susceptibility testing of Candida spp. by use of Clinical and Laboratory Standards Institute broth microdilution methods

Clinical prediction rules for invasive candidiasis in the ICU: ready for prime time?

Evaluation of Molecular Epidemiology, Clinical Characteristics, Antifungal Susceptibility Profiles, and Molecular Mechanisms of Antifungal Resistance of Iranian Candida parapsilosis Species Complex Blood Isolates

Narrow mutational signatures drive acquisition of multidrug resistance in the fungal pathogen Candida glabrata

Unexpected genomic variability in clinical and environmental strains of the pathogenic yeast Candida parapsilosis

Freebayes: haplotype-based variant detection from short-read sequencing

First Report of Candidemia Clonal Outbreak Caused by Emerging Fluconazole-Resistant Candida parapsilosis Isolates Harboring Y132F and/or Y132F+K143R in Turkey

Epidemiology and Mycology of Candidaemia in nononcological medical intensive care unit patients in a tertiary center in the United States: Overall analysis and comparison between non-COVID-19 and COVID-19 cases

Conserving Supply of Personal Protective Equipment-A Call for Ideas

Spread of Carbapenem-Resistant Gram-Negatives and Candida auris during the COVID-19 Pandemic in Critically Ill Patients: One Step Back in Antimicrobial Stewardship? Microorganisms

Bacteremia and Blood Culture Utilization during COVID-19 Surge in New York City

pandemic, central-line-associated bloodstream infection (CLABSI), and catheter-associated urinary tract infection (CAUTI): The urgent need to refocus on hardwiring prevention efforts

Dexamethasone in Hospitalized Patients with Covid-19

Co-infections in people with COVID-19: a systematic review and meta-analysis

Improving antibiotic stewardship in COVID-19: Bacterial co-infection is less common than with influenza

Antibiotic use in patients with COVID-19: a 'snapshot' Infectious Diseases International Research Initiative (ID-IRI) survey

Bacterial and Fungal Coinfection in Individuals With Coronavirus: A Rapid Review To Support COVID-19 Antimicrobial Prescribing

Multidrug-Resistant Candida auris Infections in Critically Ill Coronavirus Disease Patients

Effects of carbapenems and their combination with amikacin on murine gut colonisation by Candida albicans

Fluconazole-Resistant Candida parapsilosis Bloodstream Isolates with Y132F Mutation in ERG11 Gene

Hospital outbreak of fluconazole-resistant Candida parapsilosis: arguments for clonal transmission and long-term persistence

Fluconazole-resistant Candida parapsilosis strains with a Y132F substitution in the ERG11 gene causing invasive infections in a neonatal unit

Prevalence and Clonal Distribution of Azole-Resistant Candida parapsilosis Isolates Causing Bloodstream Infections in a Large Italian Hospital

Environmental Clonal Spread of Azole-Resistant Candida parapsilosis with Erg11-Y132F Mutation Causing a Large Candidemia Outbreak in a Brazilian Cancer Referral Center

Genetically related micafunginresistant Candida parapsilosis blood isolates harbouring novel mutation R658G in hotspot 1 of Fks1p: a new challenge?

Efficacy of LAMB against Emerging Azoleand Multidrug-Resistant Candida parapsilosis Isolates in the Galleria mellonella Model

Is the Frequency of Candidemia Increasing in COVID-19 Patients Receiving Corticosteroids?

Effects of nosocomial candidemia on outcomes of critically ill patients

A c c e p t e d M a n u s c r i p t A c c e p t e d M a n u s c r i p t A c c e p t e d M a n u s c r i p t M a n u s c r i p t Preceding bacteremia due to Acinetobacter baumannii complex 9% (13) 39% (11) 2% (2) <0.0001 <0.0001 48.1 (5.9-391.0) 1 Among patients receiving carbapenem, all received meropenem; 2 patients received both meropenem and ertapenem but at different times. 2 Please refer to Supplemental Figure 2 for enumeration of specific bacteria responsible for bloodstream infections among patients with and without bacteremia A c c e p t e d M a n u s c r i p t A c c e p t e d M a n u s c r i p t