Candidemia ranks the fourth most prevalent infection in healthcare settings. Notably, non-Candida albicansCandida species, particularly Candida parapsilosis, have experienced a worldwide increase as causative agents of candidemia in recent years.
AimsThis study aimed to assess shifts in prevalence and antifungal susceptibility of C. parapsilosis among candidemia-causing species over time, alongside investigating clonal relationships among isolates.
MethodsWe analyzed Candida species in candidemia episodes from January 2016 to December 2022. MALDI-TOF MS identified Candida isolates at the species level, and the identification of the species within C. parapsilosis complex was achieved via the restriction fragment length polymorphism method. Fluconazole susceptibility testing followed EUCAST guidelines. Clonal relationships among C. parapsilosis isolates were assessed using three methods: AP-PCR, REA-PFGE, and electrophoretic karyotyping.
ResultsAcross a seven-year period, Candida prevalence among all blood cultures was 2.1% (679 cases). The most prevalent Candida species were C. parapsilosis (33.7%), C.albicans (32.7%), Nakaseomyces glabratus (formerly Candida glabrata) (14.4%), and Candida tropicalis (8%). Notably, C. parapsilosis isolation rates increased annually from 2016 to 2022. Fluconazole susceptibility of C. parapsilosis isolates declined over time, with rates decreasing from 100% in 2016 to 72.6% in 2022. Electrophoretic karyotyping exhibited superior discrimination in clonal relationship analysis (D=0.9875).
ConclusionsOur findings highlight a concerning trend of increasing fluconazole resistance in C. parapsilosis sensu stricto over the study period. Electrophoretic karyotyping has emerged as a robust method for assessing clonal relationships. While no dominant outbreak isolate was identified, the high clustering rate suggests the potential of C. parapsilosis as a significant nosocomial infection agent in the future.
La candidemia es la cuarta infección más prevalente en los centros sanitarios. En particular, las especies de Candida no-Candida albicans, especialmente Candida parapsilosis, han experimentado en los últimos años un aumento mundial como agentes causantes de candidemia.
ObjetivosEl objetivo de este estudio fue evaluar los cambios en la prevalencia de especies responsables de candidemia, así como la sensibilidad antifúngica de C.parapsilosis a lo largo del tiempo, además de investigar las relaciones clonales entre los aislamientos.
MétodosSe identificaron las especies de Candida en episodios de candidemia desde enero de 2016 hasta diciembre de 2022. Mediante la técnica MALDI-TOF MS se identificaron las especies de Candida; para la identificación de las especies dentro del complejo C.parapsilosis se utilizó el método de polimorfismo de longitud de fragmentos de restricción. Las pruebas de sensibilidad al fluconazol siguieron las directrices marcadas por EUCAST. Las relaciones clonales entre los aislamientos de C.parapsilosis se evaluaron mediante tres métodos: AP-PCR, REA-PFGE y cariotipado electroforético.
ResultadosA lo largo de un periodo de siete años, el aislamiento de Candida en todos los hemocultivos fue del 2,1% (679 casos). Las especies de Candida más prevalentes fueron C.parapsilosis (33,7%), C.albicans (32,7%), Nakaseomyces glabratus (antes Candida glabrata) (14,4%) y Candida tropicalis (8%). Cabe destacar que la tasa de aislamiento de C.parapsilosis aumentó anualmente desde 2016 a 2022. La sensibilidad al fluconazol de los aislamientos de C.parapsilosis disminuyó con el tiempo, siendo esta sensibilidad del 100% en 2016 y del 72,6% en 2022. El cariotipado electroforético mostró una discriminación superior en el análisis de la relación clonal (D=0,9875).
ConclusionesNuestros hallazgos destacan una tendencia preocupante en el aumento de la resistencia al fluconazol de C.parapsilosis sensu stricto durante el período de estudio. El cariotipado electroforético se ha revelado como un método sólido para evaluar las relaciones clonales. Aunque no se identificó ningún aislamiento dominante responsable de algún brote, la elevada tasa de agrupación sugiere que C.parapsilosis puede ser un importante agente de infección nosocomial en el futuro.
Candidemia, the fourth most prevalent healthcare-associated bloodstream infection,21 has seen a global surge in non-Candida albicansCandida species as causative agents. This shift in trend could potentially modify empirical treatment protocols and the prognosis of candidemia. Therefore, it is crucial to identify this rise in non-C.albicansCandida species and devise appropriate strategies. The introduction of advanced diagnostic tools, such as mass spectrometry (MALDI-TOF MS) and T2 magnetic resonance, is expected to reshape the future candidemia epidemiology.25 The routine application of MALDI-TOF MS for yeast identification has enhanced our understanding of previously misidentified or unidentified Candida species.28
Candida parapsilosis ranks the third- or fourth most common species causing candidemia worldwide.15,28 In our country, C.parapsilosis has been reported as the second most common Candida species after C.albicans.3,30 In countries like Brazil, Argentina, Peru, Spain, Russia, and China, C.parapsilosis accounts for more than 20% of the Candida species recovered from blood cultures,50 whereas in Türkiye, Greece, Croatia, Romania, South Africa, Nigeria, and Paraguay, it constitutes more than 30%.15,51
The C.parapsilosis complex is significant due to its ability to form biofilms on central venous catheters and medically implanted devices, posing a threat to patients undergoing invasive medical procedures.17 Infections caused by species within the C.parapsilosis complex can be life-threatening in immunocompromised individuals. Therefore, it is essential to swiftly identify the species implied and perform antifungal susceptibility tests in light of the increasing global antifungal resistance.51
Significant advances in molecular methodologies have enabled the investigation of the genetic diversity of C.parapsilosis isolates. Techniques such as restriction fragment length polymorphism (RFLP) have made it possible dividing the C.parapsilosis complex into three different groups: Group I, II, and III.38 A 2005 study revealed that these groups are so distinct one another that they can be considered independent species. Consequently, C.parapsilosis Group I was renamed as C.parapsilosis sensu stricto, Group II as Candida orthopsilosis, and Group III as Candida metapsilosis.42 There are differences in antifungal resistance and virulence factors among the species within C.parapsilosis complex. C.parapsilosis sensu stricto, the most common member of this species complex, is the most frequent cause of candidemia due to its colonizing ability.5
In recent years, there has been a notable trend of diminishing susceptibility to azole compounds among C.parapsilosis isolates. Although azole resistance among these strains vary geographically, a global increase in resistance has been reported. In light of the rising frequency of C.parapsilosis isolates and fluconazole resistance, the World Health Organization has classified this species complex as a high-risk fungal pathogen.49C.parapsilosis isolates are recognized for causing healthcare-associated outbreaks in hospitals.20 These outbreaks typically result from contaminated intravenous solutions or horizontal transmission from the environment via healthcare workers’ hands to patients.36 Various methods, including arbitrarily primed polymerase chain reaction (AP-PCR), restriction enzyme analysis-pulsed field gel electrophoresis (REA-PFGE), electrophoretic karyotyping (EK), random amplified polymorphic DNA (RAPD), multilocus sequence typing (MLST), and analysis of polymorphic microsatellite regions, can be employed to investigate outbreaks caused by C.parapsilosis isolates.4,13,46 However, an ideal method that is standardized, highly reproducible, stable, inexpensive, and rapidly applicable, is yet to be established.46
This study aims to examine the frequency and species distribution of C.parapsilosis complex isolates obtained from blood cultures in a university hospital, which includes a liver transplant hospital, an oncology hospital, and a bone marrow transplantation unit. Furthermore, we aim to determine the clonal relationship between the isolates to uncover the dynamics of infection.
1Material and methods1.1Candida isolatesThis study is a single-center retrospective study conducted in a tertiary care university hospital in Türkiye. Isolates obtained from blood cultures between January 2016 and December 2022 were included, sourced from the Microbiology laboratory at İnönü University Turgut Özal Medical Center. Our center, located in the Eastern Anatolia region of Türkiye, was severely impacted by a major earthquake on February 6, 2023. Consequently, data from 2023 were excluded from our study in order to prevent the changing patient population following the earthquake from affecting the epidemiologic data. Our hospital is a tertiary complex with a total of 1585 beds, including the main building, liver transplant institute, oncology hospital, and 317 intensive care beds. Blood cultures were incubated in the BACT/ALERT 3D (Becton Dickinson, Maryland, USA) automated system. Yeast isolates were identified using MALDI-TOF MS (bioMérieux, France), and all isolates were preserved in a skim milk medium at −80°C for long-term storage. Before further analysis, isolates were revived by subculturing them on Sabouraud dextrose agar, and were incubated at 35°C for 24–48h. Protein extraction was performed by applying a portion of a fresh colony directly as a thin film on a spot of a MALDI disposable target plate, followed by the addition of 1μl of matrix and 1μl of 28.9% formic acid solution (bioMérieux, France), as per the manufacturer's instructions. After drying, the target plate was loaded into the Vitek MS instrument for identification, using the database outlined in the manufacturer's instructions. Epidemiologic data of the patients, including the date of yeast isolation, hospitalization service, date of hospitalization, and invasive procedures, were recorded.
1.2Antifungal susceptibility testingThe in vitro susceptibility of the isolates to fluconazole was tested in accordance with EUCAST guidelines (EUCAST EDef 7.3 document). The minimum inhibitory concentration (MIC) values (mg/L) obtained were interpreted based on the clinical cutoff values in the EUCAST guidelines.19C.parapsilosis American Type Culture Collection (ATCC) 22019 and Candida krusei ATCC 6258 strains were utilized for quality control in each study.
1.3Subtype-level identificationThe restriction fragment length polymorphism (RFLP) method was employed for the subspecies determination of C.parapsilosis complex isolates. DNA was extracted from fresh yeast colonies subcultured on Sabouraud dextrose agar using the EZ1 DNA Tissue Kit (Qiagen, Germany) in the BioRobot EZ1 (Qiagen, Germany) automated isolation device, with minor modifications following the manufacturer's instructions. The target region was amplified using primers specific for the secondary alcohol dehydrogenase (SADH) gene region [(SADH-F GTTGATGCTGTTGGATTGT) and (SADH-R CAATGCCAAATCTCCCAA)].42 The SADH-PCR product (716bp) obtained was incubated with BanI (20U) restriction enzyme (Biolabs, UK) at 37°C for 2h, followed by electrophoresis in a 2% agarose gel. Amplicons were photographed post-electrophoresis with the Gel Logic 2200 imaging system (Kodak Co., Rochester, NY, USA). C.parapsilosis sensu stricto, C.metapsilosis and C.orthopsilosis PCR sequences contain, respectively, one specific restriction site, which produces two bands (521 and 196bp); three restriction sites, which produce four bands (370, 188, 93, and 60bp); and zero restriction sites (716bp).
1.4Sequence analysis for non-typable isolatesIsolates finally identified as C.orthopsilosis, but not identified at the beginning despite repeated RFLP, were typed at the species level using sequence analysis. Specifically, the large subunit rRNA (D1/D2 region) gene of these isolates was amplified as previously described, and the base sequences of the obtained PCR product were determined using the Sanger sequencing method.24
1.5Sequencing of ERG11The ERG11 gene, one of the most common molecular mechanisms of fluconazole resistance, was sequenced using primers and conditions previously described by Grossman et al., and then compared to C.parapsilosis isolate ATCC 22019 wild-type ERG11 sequence.22 The obtained sequences were compared using the Basic Local Alignment Search Tool (BLAST) available at the National Center for Biotechnology Information Database (http://www.ncbi.nlm.nih.gov).
1.6Investigation of clonal relationships among isolatesTo investigate the clonal relationships between C.parapsilosis isolates, AP-PCR, EK, and REA-PFGE methods were employed.
Arbitrarily primed polymerase chain reaction: For the genotyping of C.parapsilosis isolates by AP-PCR, a protocol using the M13 primer (5′-GAG GGT GGC GGT TCT-3′) was utilized, as previously standardized by Kuzucu et al.27
Electrophoretic karyotyping: EK was performed as previously described by Shin et al. with minor modifications.40 For in situ lysis of the prepared cell suspensions, 25μl of 2500U lyticase enzyme were used per sample. Subsequently, plugs were meticulously prepared with a cell-agarose-lithiase mixture prepared with 2% low-melting agarose. Electrophoresis was conducted with a CHEF DR II (pulsed-field certified agarose, Bio-Rad Laboratories) at a pulse time of 90–325s and 6V/cm2 in a 0.8% agarose gel with 0.5× TBE (50mM Tris, 45mM boric acid, and 0.5mM EDTA) for 48h.
Restriction enzyme analysis-PFGE: Plugs, prepared as outlined for EK, were transferred to a buffer containing 20U of BssHII and incubated at 37°C overnight. Electrophoresis was then conducted using a CHEF DR II (Bio-Rad laboratories, Nazareth, Belgium) at a pulse time of 90–360s, and 3.5V/cm2 in a 0.7% agarose gel with 0.5× TBE for 36h.
All three methods utilized the Gel Logic 2200 system (Kodak Co., Rochester, NY, USA) for imaging. Band profiles were analyzed using the GelCompar II software system (version 6.5; Applied Maths, Sint-Martens-Latem, Belgium). The Dice correlation coefficient was used to calculate the similarity in band analysis, and the UPGMA method was used for cluster analysis.12
1.7Statistical analysisThe results obtained were subjected to multivariate logistic regression analysis, as well as to a trend odds model (p-trend) to evaluate the temporal trends in fluconazole-susceptibility rates and the overall increase in prevalence of C.parapsilosis sensu stricto across all Candida species recovered from BSI. A non-parametric Mann-Kendal test was conducted to determine if the prevalence of fluconazole susceptibility decreased overall or if the incidence of C.parapsilosis increased over time. All statistical tests were performed using IBM SPSS Statistics 26.0, with p<0.05 considered statistically significant.
2ResultsThis study analyzed a total of 32,576 blood cultures sent to our laboratory from various intensive care units and wards over a seven-year period. In our center, the blood culture positivity rate was 7.7% (n=2514), and the frequency of candidemia was 2.1% (n=680). Candidemia accounted for 27.1% (680/2514) of all positive blood cultures. When the distribution of the species of Candida was analyzed, 229 (33.7%) were C.parapsilosis, 222 (32.7%) C.albicans, 98 (14.4%) Nakaseomyces glabratus (formerly Candida glabrata), 54 (8%) Candida tropicalis, 31 (4.6%) Kluyveromyces marxianus (formerly Candida kefyr), 14 (2.1%) Pichia kudriavzevii (formerly C.krusei), 10 (1.5%) Clavispora lusitaniae (formerly Candida lusitaniae), 9 (1.3%) Meyerozyma guilliermondii (formerly Candida guilliermondii), 7 (1.%) Candida dubliniensis, 4 (0.6%) Candida inconspicua, and 2 (0.3%) Wickerhamomyces anomalus (formerly Candida pelliculosa).
The study included a total of 229 C.parapsilosis isolates obtained from blood cultures, each corresponding to a unique patient. For the analysis of clonal relationships, 121 C.parapsilosis isolates from the culture collection of our laboratory that had been isolated from 115 patients were included. Due to the development of resistance during the antifungal treatment in six patients, a second isolate of C. parapsilosis from each patient was recovered from blood culture.
In our study, the mean age of the 229 patients with C.parapsilosis was 52 years (min–max 0–101). Fourteen (6.1%) patients were aged 18 years and younger, 51 (22.3%) were aged 55–64 years, and 49 (21.4%) were aged 65–74 years.
In our study, 183 (79.9%) C.parapsilosis isolates were susceptible to fluconazole. The first fluconazole-resistant C.parapsilosis was isolated in 2017. The fluconazole susceptibility of C.parapsilosis isolates each year, from 2016 to 2022, was 100%, 95.2%, 91.3%, 85%, 80.7%, 73.2%, and 72.6%. A statistically significant decrease over time in fluconazole susceptibility was observed (mean estimate: −5.42, [95% CI −6.56; −4.29], slope of linear trend p<0.001). The fluconazole susceptibility of the Candida isolates recovered each year are presented in Table 1 according to species.
Yearly distribution of fluconazole-susceptible Candida isolates. Susceptibility was determined according to EUCAST breakpoints.
| Year | C. albicans | C. parapsilosis | N. glabratus | C. tropicalis | Other | Total |
|---|---|---|---|---|---|---|
| Fluconazole-susceptible strains (n)/All strains (n) | ||||||
| 2016 | 2/3 (66.6%) | 1/1 (100%) | – | – | 1/1 (100%) | 4/5 (80%) |
| 2017 | 28/31 (90.3%) | 20/21 (95.2%) | 4/5 (80%) | 1/1 (100%) | 9/10 (90%) | 62/68 (91.2%) |
| 2018 | 25/28 (89.2%) | 21/23 (91.3%) | 6/6 (100%) | 3/3 (100%) | 10/11 (90.9%) | 65/71 (91.5%) |
| 2019 | 39/44 (88.6%) | 34/40 (85%) | 11/12 (91.7%) | 7/8 (87.5%) | 13/15 (86.7%) | 104/119 (87.4%) |
| 2020 | 20/24 (83.3%) | 21/26 (80.7%) | 12/12 (100%) | 10/10 (100%) | 8/9 (88.8%) | 71/81 (87.6%) |
| 2021 | 37/40 (92.4%) | 41/56 (73.2%) | 27/32 (84.4%) | 15/15 (100%) | 17/20 (85%) | 137/163 (84%) |
| 2022 | 47/52 (90.4%) | 45/62 (72.6%) | 30/31 (96.8%) | 17/17 (100%) | 9/11 (81.8%) | 148/173 (85.5%) |
| Total | 198/222 (89.1%) | 183/229 (79.9%) | 90/98 (91.8%) | 53/54 (98.1%) | 67/77 (87%) | 591/680 (87%) |
| p-Trend | 0.206 | <0.001 | 0.15 | 0.163 | 0.008 | 0.956 |
Other species: K. marxianus, P. kudriavzevii, C. lusitaniae, M. guilliermondii, C. dubliniensis, C. inconspicua, and W. anomalus.
Among the 46 fluconazole-resistant C.parapsilosis isolates collected, 25 strains that were successfully revived were subjected to ERG11 gene analysis. Several mutations were identified, with A395T (Y132F) being the most common (21 isolates), indicating a potential mechanism of resistance. Despite being classified as a silent mutation, T591C (Silent, I197I) was detected in 23 isolates, suggesting possible implications in resistance development. G1372A (G458S) was also identified in one isolate.
Among all the isolates collected during the study period, all were identified as C.parapsilosis, except for one isolate identified as C.orthopsilosis by RFLP and one isolate identified as C.tropicalis by sequence analysis. C. parapsilosis was isolated from one patient in 2016, from 21 patients in 2017, 23 patients in 2018, 40 patients in 2019, 26 patients in 2020, 56 patients in 2021, and 62 patients in 2022. There was a statistically significant increase in C.parapsilosis isolates among Candida species over the years (mean estimate: 1.92, [95% CI 0.21; 3.64], slope of linear trend p=0.035). The distribution of Candida species over the years is presented in Table 2. Upon analyzing the inpatient units, it was observed that C.parapsilosis isolates were most frequently isolated from patients in the medical intensive care units, accounting for 52.8% of cases (Table 3).
Distribution of Candida species over time.
| Year | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| 2016 | 2017 | 2018 | 2019 | 2020 | 2021 | 2022 | Total | p-trend | |
| C. albicans | 3 (60%) | 31 (45.5%) | 28 (39.4%) | 44 (37.3%) | 24 (29.6%) | 40 (24.5%) | 52 (30.1%) | 222 (32.7%) | 0.004 |
| C. parapsilosis | 1 (20%) | 21 (31%) | 23 (32.4%) | 40 (33.9%) | 26 (32.1%) | 56 (34.4%) | 62 (35.8%) | 229 (33.7%) | 0.035 |
| N. glabratus | 0 (0%) | 5 (7.3%) | 6 (8.5%) | 12 (10.2%) | 12 (14.8%) | 32 (19.6%) | 31 (17.9%) | 98 (14.4%) | 0.001 |
| C. tropicalis | 0 (0%) | 1 (1.5%) | 3 (4.2%) | 8 (6.8%) | 10 (12.4%) | 15 (9.2%) | 17 (9.8%) | 54 (8%) | 0.006 |
| Other | 1 (20%) | 10 (14.7%) | 11 (15.5%) | 14 (11.8%) | 9 (11.1%) | 20 (12.3%) | 11 (6.4%) | 76 (11.2%) | 0.004 |
| Total | 5 | 68 | 71 | 118 | 81 | 163 | 173 | 679 | |
Other species: K. marxianus, P. kudriavzevii, C. lusitaniae, M. guilliermondii, C. dubliniensis, C. inconspicua, and W. anomalus.
In our study, AP-PCR, REA-PFGE, and EK methods were employed for studying the clonal relationship of C.parapsilosis isolates, and 34, 39, and 69 different genotypes, respectively, were found among a total of 115 isolates. The isolates were grouped into 15, 20, and 25 different clusters, respectively (optimization 3.0, tolerance 1.0, cutoff 95%). The clustering rate of the isolates was 83.4% in the AP-PCR and REA-PFGE methods, and 60.8% in the EK method. The clustering rate of C.parapsilosis isolates was 67.6% in the inpatient units and 52.6% in intensive care units with the EK method. Representative gel images of all three methods are depicted in Fig. 1.
Representative band patterns of Candida parapsilosis species complex isolates analyzed using AP-PCR, REA-PFGE, and EK methods. Isolates numbered 38, 39, 40, 41, 42, and 43 were selected for this figure based on the sharp of their bands. No relationship between these isolates was expected after comparing the epidemiologic data. (a) AP-PCR analysis showed indistinguishable band patterns among all isolates. (b) REA-PFGE analysis revealed that five isolates were closely related, while one isolate exhibited a distinct banding pattern. (c) EK analysis demonstrated that all isolates were unrelated, with Candida orthopsilosis (isolate 42) being clearly differentiated from C.parapsilosis sensu stricto isolates only through this method.
Candida species rank among the most common fungal pathogens causing fungemia in humans.48 Despite advancements in antifungals and treatment protocols, healthcare-associated Candida infections are escalating globally, driven by increasing resistance. Consequently, novel strategies are imperative to combat candidemia,5 with molecular epidemiologic research leading the charge. However, conventional identification methods present challenges that hinder the establishment of infection epidemiology and make determining the epidemiology of the infection difficult.
This field encompasses a broad spectrum of investigations, from promptly identifying the causative agent at the species level to delineating antimicrobial resistance and clonal relationships among isolates. Given the presence of naturally resistant species and varied transmission dynamics, species-level typing of Candida yeasts is pivotal. The advent of MALDI-TOF MS systems has revolutionized species-level identification of non-C.albicansCandida species in routine clinical microbiology laboratories.25,39 At our center, MALDI-TOF MS has enabled species-level identification of Candida yeasts since 2016, spanning a seven-year period.
Candida species rank as the fourth most common cause of nosocomial bloodstream infections.9,32 Incidence rates of candidemia vary globally, ranging from 0.68 to 4.1% in intensive care units, to 0.02–0.14% in the community.26 In a cross-sectional study in Egypt, Candida species accounted for 1.6% of bloodstream infections in adults.37 Similarly, a study in our country identified Candida yeasts as the causative agent in 1.12% of patients based on blood cultures.23 Our own study over a seven-year period revealed a candidemia incidence of 2.1%. This higher rate in our center may be attributed to the substantial number of immunocompromised patients, particularly in the liver transplant and oncology units. However, given the low sensitivity of blood culture systems, estimating the burden of candidemia accurately remains challenging.14
Although C.albicans is the most frequently isolated species in candidemia worldwide, an increase in non-C.albicansCandida species in recent years has been reported. Specifically, candidemia caused by C.parapsilosis and N.glabratus has had a notable increase in the last decade.3,28 In the SENTRY study, which involved over 200 medical centers in 45 countries, N.glabratus was the most frequently isolated non-C.albicansCandida species in regions outside Latin America, while C.parapsilosis and C.tropicalis were reported in Latin America.33 Two studies conducted in Türkiye reported that C.parapsilosis was the most frequently isolated agent of candidemia, being recovered in 43% and 60% of the patients, followed by C.albicans (33%, 20%) and C.tropicalis (12%, 8%).10,23 In our study, C.parapsilosis (33.7%) was the most isolated species, followed by C.albicans (32.7%), N.glabratus (14.4%), C.tropicalis (8%), K.marxianus (4.6%), P.kudriavzevii (2.1%), C.lusitaniae (1.5%), M.guilliermondii (1.3%), C.dubliniensis (1%), C.inconspicua (0.6%), and W.anomalus (0.3%).
We hypothesized that C.parapsilosis would be a predominant species among pediatric patients with invasive candidiasis, particularly due to its known association with catheter-related infections and hospital outbreaks. Our hypothesis is supported by the fact that C.parapsilosis isolates are known for their transmission through hyperalimentation-intravenous devices, leading to outbreaks in pediatric intensive care unit patients.36 A study has shown that C.parapsilosis accounts for 30% of candidemia outbreaks in neonatal intensive care units.1 In a study investigating the distribution of Candida species in pediatric patients with invasive candidiasis, C.albicans (50%) was the most frequently isolated, followed by C.parapsilosis (24%).1,41 In our pediatric patient population, the distribution of Candida species was as follows: C.parapsilosis (37.9%), C.albicans (27.6%), N.glabratus (14.3%), C.tropicalis (14.3%), P.kudriavzevii (5.9%). Compared to previous studies, our findings indicate a higher incidence of C.parapsilosis in pediatric patients. C. parapsilosis, which is known for its ability to form biofilms on medical devices, and its association with nosocomial outbreaks in neonates, has been extensively found in neonatal and pediatric wards.
Our statistical p-trend analysis revealed a significant upward trend in the occurrence of C.parapsilosis at our center compared to other species, with a slope of linear trend p-value of 0.035. This is in line with the findings of Papas et al., who reported that C.parapsilosis is more frequently associated with invasive human infections in certain regions, including the Mediterranean basin, India, and Latin America.31 These regions are also known for widespread inappropriate use of antimicrobials and high levels of antimicrobial resistance (AMR). Almirante et al. discovered that nearly 90% of the patients with C.parapsilosis sepsis had undergone antibacterial treatment. They also found that previous antifungal treatment significantly increased the risk of C.parapsilosis bloodstream infection by more than 5-fold.2 Given the high AMR burden in our country, the significant increase in the frequency of C.parapsilosis observed in our study over a relatively short seven-year period could be attributed to selective pressure resulting from high antimicrobial consumption and increased AMR. Unlike C.albicans and N.glabratus, which are frequently associated with endogenous infections caused by the host microbiota, C.parapsilosis is primarily considered a nosocomial species due to its strong association with healthcare settings, particularly parenteral nutrition solutions and intravascular devices frequently used in critically ill patients.8C.parapsilosis has been reported as a colonizer in medical devices, especially plastic surfaces, for up to four weeks due to its biofilm-forming ability, making it a significant concern for hospital-acquired infections.44
Therefore, we hypothesize that the increased prevalence of C.parapsilosis in healthcare-associated candidemia may be due to a decrease in infection control measures. Additionally, in recent years, the increasing use of MALDI-TOF MS in clinical laboratories has significantly enhanced the accuracy of yeast species identification compared to traditional culture and biochemical tests. This advancement has improved the identification of Candida species and may have influenced the prevalence reported. The widespread adoption of this technology has facilitated the identification of previously misidentified non-C. albicansCandida species, potentially contributing to the observed increase in their incidence. Lastly, while MALDI-TOF MS plays a crucial role in this shift, other contributing factors may also exist, and further studies are needed to draw definitive conclusions.
A significant concern about C. parapsilosis is the recent reports on its decreasing susceptibility to azole antifungal agents.1,47 The fluconazole resistance rates generally reported in C.parapsilosis strains have been 2–5%.45 In a multicenter study in our country, 7.7% fluconazole resistance was reported in 575 C.parapsilosis isolates, with variability between centers (0–47.1%).6 In our study, the first fluconazole-resistant C. parapsilosis strain was isolated in 2017, and the incidence of resistant strains has gradually increased. In our study, we found that the cumulative fluconazole resistance rate in C. parapsilosis strains was 20.1%.
Molecular biological studies have revealed the genetic diversity of C.parapsilosis isolates, leading to the term C.parapsilosis complex. Since 2005, it has been shown that the species within this species complex are so different one another that they can be considered independent species. This fact has added depth to the molecular epidemiology of C.parapsilosis. However, the number of studies in the literature in which species belonging to the C.parapsilosis complex were identified is limited. In these studies, C.parapsilosis sensu stricto was reported as the most common species causing candidemia, with C.orthopsilosis and C.metapsilosis found more rarely.16,17 In the SENTRY study, out of 2548 invasive C.parapsilosis species complex isolates sent from 135 centers worldwide between 1997–2016, 95.5% were C.parapsilosis sensu stricto, 3.2% were C.orthopsilosis, and 1.3% were C.metapsilosis.33 In the ARTEMIS antifungal surveillance program, C.parapsilosis sensu stricto (92.1%), C.orthopsilosis (6.1%), and C.metapsilosis (1.8%) were identified among 1929C.parapsilosis species complex isolates between 2001–2006.29 The study observed differences in the frequency of species depending on regions and countries, reporting that C.orthopsilosis was the most common in Asia (11%), while this frequency was 4.9% in the USA.29 In a study by Demirci-Duarte et al., among 181 C.parapsilosis isolates, 99.4% were C.parapsilosis sensu stricto and 0.4% were C.orthopsilosis.17 In our research, we had 228 C.parapsilosis sensu stricto and one C.orthopsilosis among 229 C.parapsilosis isolates, with no C.metapsilosis isolates found.
Studies on the different antifungal susceptibility pattern of the species within the C.parapsilosis complex have been emerging. The ARTEMIS study found that fluconazole MIC values for C.orthopsilosis and C.metapsilosis isolates were lower than those for C.parapsilosis sensu stricto isolates, although all the isolates were susceptible to fluconazole.29 In contrast, a study by Tavanti et al. with 27 C.orthopsilosis isolates from Italy, reported higher mean MIC values for fluconazole (5.89μg/ml) than the ARTEMIS study (2.1μg/ml).43 The fluconazole MIC value of the only C.orthopsilosis isolate included in our study was 0.75μg/ml. We think that the differences in antifungal susceptibility described among the species of the complex are not categorical differences, suggesting that it is not essential for clinical laboratories to identify the species within the complex in routine laboratory practice. By contrast, our study did not identify any polifungal infections. However, this does not rule out the possibility entirely, given that we used subcultures from single colonies. Recent studies have reported that different species in the C.parapsilosis species complex are being identified concurrently as causative agents of candidemia. In a study by Barbedo et al., polifungal infections, including C.parapsilosissensu stricto and C.orthopsilosis, were described in 9.5% of fungemia episodes caused by the C.parapsilosis complex.7 In our study, given that we used subcultures from single colonies, we do not know if any patient would suffer a polifungal infection with species of the C.parapsilosis complex. Further research could provide valuable insights into the prevalence and role of mixed infections caused by different members of this species complex.
The potential of the C.parapsilosis complex to be transmitted horizontally via medical instruments and the hands of healthcare workers is a significant concern, particularly in the context of hospital outbreaks involving fluconazole-resistant isolates. Analysing clonal associations is vital for preventing healthcare-associated infections and outbreaks. Such studies are a crucial part, as well, of outbreak investigations.5,46 These analyses can also serve as a measure of the effectiveness of infection control procedures. Furthermore, conducting these studies in a prospective or real-time manner can yield data that could be used to prevent outbreaks from escalating.
Several methods, including AP-PCR, REA-PFGE, EK, and microsatellite genotyping, can be employed to both determine the clonal relationships among isolates within C.parapsilosis complex, and identify outbreaks.11,36 However, the perfect method for determining these clonal relationships has yet to be found. In our research, we utilized AP-PCR, REA-PFGE, and EK methods and obtained, out of 115 C.parapsilosis isolates, 34 and 39 different genotypes using the AP-PCR and REA-PFGE methods, respectively, and 69 different genotypes by using the EK methodology. The clustering rates of these methods were 83.4%, 83.4%, and 60.8%, respectively. Using the mathematical formula developed by Hunter and Gaston, the value of the discrimination power was 0.9393, 0.9492, and 0.9875, respectively.
When we evaluated the fingerprinting results along with the epidemiological data of the patients from whom the yeasts were isolated, we found that the EK method provided the best discrimination. Even a single band difference in the chromosomal band patterns obtained by this method can categorize isolates as related or unrelated.18 According to the literature, the EK method has been shown to be effective in investigating cases of persistent or recurrent fungemia caused by C.parapsilosis.34,35,40 Shin et al. studied C.parapsilosis strains from 17 patients who experienced recurrent episodes of fungemia. Using the EK method, the authors discovered that these recurrent infections were usually caused by reinfection with a second strain.40 In our study, based on the EK results, we determined that recurrent candidemia episodes in three patients were caused by a second C.parapsilosis strain. Although no dominant epidemic clone emerged among the isolates using the EK method, the 60.8% clustering rate suggests that infection control measures should be enhanced.
Our study has certain limitations. Although it was a single-center analysis, which may have facilitated more efficient access to records, the retrospective nature of the study resulted in incomplete data on patient outcomes. Nevertheless, our study has demonstrated an increase in the prevalence of C.parapsilosissensu stricto, one of the causative agents of candidemia, as well as a significant increase in its fluconazole resistance rate, over the years. This suggests a significant epidemiological shift. In our study, the most frequently identified ERG11 mutation in fluconazole-resistant C.parapsilosis strains was Y312F. The EK method proved to be superior to the other two methods when investigating clonality in C.parapsilosis isolates as it had a better discriminatory power. Fluconazole-resistant C.parapsilosissensu stricto has the potential to colonize in-hospital ecological niches where other Candida species, particularly C.albicans, are present. In light of the inevitable global increase in antibiotic resistance, which will force even countries that use antimicrobials more judiciously to use them more frequently, it is likely that many countries will experience a change in the frequency of pathogens in the coming years, similar to that observed in our study.
Ethics statementThis study was carried out with the approval of the Inonu University Scientific Research and Publication Ethics Committee, dated March 29, 2022, and numbered 2022/3281.
Conflict of interestThe authors declare that they have no conflict of interest.
Data availability statementThis study is part of the corresponding author's thesis, which was completed in September 2022.
Yusuf Yakupoğulları: 0000-0002-5545-3467; Yücel Duman: 0000-0002-9090-2096; Mehmet Sait Tekerekoğlu: 0000-0001-7284-3427; Baris Otlu: 0000-0002-6220-0521.
This study was presented at the 33rd European Congress of Clinical Microbiology and Infectious Diseases, Copenhagen, Denmark, April 15–18, 2023.
|