Introduction

Candidiasis is the most common fungal infection and usually affects children with chronic illness, prematurity, immunodeficiency, and critical diseases (1). In the United States, Candida spp. is the widespread cause of invasive fungal disease and is the second most common problem of central line-associated bloodstream infections (CLABSI) in pediatric patients (2).

Candidiasis leads to increased mortality, morbidity, prolonged hospitalization time, and health care costs in children (3). Regional distribution of Candida species, antifungal susceptibility and risk factors for invasive candidiasis (IC) are important for effective empirical treatment and prevention strategies.

We aimed to establish the distribution of Candida species, antifungal susceptibility, risk factors and mortality in hospitalized children with IC.

Materials and Methods

Patient Data and Definitions

This cross-sectional study was performed at a tertiary university hospital. Pediatric patients with culture-proven invasive Candida spp. infection during hospitalization during a four year (from September 2014 to September 2018). The demographic data, clinic features, and microbiological results of the pediatric patients with IC were analyzed retrospectively.

IC was characterized as the isolation of Candida spp. from sterile body fluids. Candidemia was described as the isolation of Candida spp. in blood culture from a peripheral vessel or a central venous catheter (CVC). The isolation of Candida spp. from any blood culture in a patient with a CVC or Candida spp. growth from a catheter-tip culture was defined as catheter-associated candidemia.

The previous hospitalization (history hospitalization) was defined as hospitalization history within three months before the IC (4). The presence of mechanical ventilation was recorded within two days before IC. Arterial catheter, CVC, urinary catheter, and use of total parenteral nutrition (TPN) were recorded within a week before infection. In the two weeks before infection, surgical intervention (abdominal and non-abdominal), broad-spectrum antibiotic use, immunosuppressive drug use, presence of neutropenia, and previous use of antifungal drugs were recorded. Initial therapy was deemed to be delayed when the time elapsed between taking the culture sample and the start of antifungal therapy was more than 72 hours. The time of first culture positivity was always considered as a benchmark to evaluate time to infection and mortality.

Identification of Organism and Susceptibility Testing

The isolates were cultured from several clinical samples (blood, CVC, and peritoneal fluid). Blood and sterile body fluid cultures were processed by the BACTEC FX 200 system (Becton Dickinson, United States of America). Growth was determined from the culture of specimens on blood and “Eosine Methylene Blue” agar plates. The colonies identified as yeast growth were transferred to the mycology laboratory for identification and antifungal susceptibility tests. The colonies were subcultured to “sabouraud dextrose agar” (Oxoid, England), and identification was performed after the strains were determined to be pure. Yeasts were identified by germ tube test, microscopic morphology on Cornmeal tween 80 agar (Oxoid, England) and CHROMagar Candida (CHROMagar, France), and API 20C AUX (BioMérieux, France).

The susceptibility of Candida isolates to fluconazole, amphotericin B (AmB), and caspofungin was determined by microdilution method performed based on The Clinical and Laboratory Standards Institute (CLSI) M27-A3 standards (5). Minimal inhibitory concentrations (MICs) for fluconazole, AmB, and caspofungin were evaluated according to the CLSI species specific clinical breakpoints. These were as follows: for fluconazole C. albicans, C. tropicalis and C. parapsilosis susceptible (MIC ≤2 μg/mL), susceptible-dose dependent (MIC 4 μg/mL), resistant (MIC ≥8 μg/mL); for fluconazole C. glabrata susceptible-dose dependent (MIC ≤ 32 μg/mL), resistant (MIC ≥64 μg/mL); for AmB susceptible (MIC ≤1 μg/mL), resistant (MIC >1 μg/mL); for caspofungin C. albicans, C. krusei, C. tropicalis, and C. glabrata resistant (MIC >0,5 μg/mL), C. parapsilosis resistant (MIC >4 μg/mL). C. krusei was considered as naturally resistant to fluconazole regardless of MIC (6).

Statistical Analysis

Statistical analyses were performed using SPSS version 20 (SPSS, Inc, Chicago, IL) software. Categorical factors were shown as numbers and percentages. Shapiro-Wilk normality test was used for test normality. Non-normal distributed continuous variables were presented as medians and interquartile ranges unless stated otherwise. Qualitative data were compared using the chi-square test, while quantitative variables were compared between groups using the Mann-Whitney U test or Fisher’s exact test. The response variable of this study was mortality. Patients were right-censored at the 7^th and 30^th days after the initial positive culture. Kaplan-Meier survival curves were prepared, and the log-rank test was used to assess the differences between survival curves. Statistical analyses were applied in R (7) version 3.4.3, using the packages “survival” and “survminer”. Significance was defined at the double-sided p-value of <0.05.

This study was approved by the Dokuz Eylül University Non-Invasive Research Ethics Committee (decision no: 2016/13-23, date: 12.5.2016).

Results

Demographics, Characteristics and Risk Factors

There were 45 IC patients during the four years. Four patients had one more recurrent episode, but only the first episodes were included. Two patients had more than one Candida spp. Two candida strains (C. albicans, C. tropicalis) were isolated in one patient, and three Candida strains (C. parapsilosis, C. albicans, C. krusei) in another patient.

Demographics and characteristics of the patients with IC are shown in Table 1. The most prominent risk factors were having underlying diseases (100%), a previous antibiotic use (95.6%), CVC (73.3%), and previous hospitalization (55.6%). Candida endocarditis and peritonitis developed in 2 (4.4%) and 1 (2.2%) patients, respectively.

Distribution and Antifungal Susceptibility of the Candida Species

Forty-eight Candida strains and seven different Candida species were isolated. The most common species were C. albicans and C. parapsilosis, both 20 (41.7%). Antifungal susceptibility testing was performed for 34 Candida isolates. The susceptibility of the isolates to the antifungal agents is shown in Table 2.

Therapy

Antifungal therapy was given in 42 episodes (93.3%), monotherapy in 17 (37.8%), and sequential treatment in 25 (55.6%) episodes. Twelve (26%) patients were receiving antifungal agents before IC. Initial therapy was delayed in seven patients (15.5%).

Outcome

Six of the patients died within 30 days following a positive culture, and the overall mortality rate was 13.3%. The demographic and clinical risk factors for 7-day and 30-day mortality were presented in Table 3. Patients with neutropenia and dialysis had a higher rate of mortality in 7-day (both p=0.02) and 30-day (p=0.04 and p=0.003, respectively). Mean survival times for 7-day mortality were lower for the patients with neutropenia (Mean ± SE: 5.86±0.97 vs 6.97±0.04), dialysis (Mean ± SE: 6.14±0.97 vs 6.92±0.06), and abdominal surgery (5.67±0.90 vs 6.97±0.03) when compared to patients who had not. Also, mean survival times for 30-day mortality were lower for the patients with neutropenia (Mean ± SE: 19.00±4.84 vs 28.29±0.97), dialysis (Mean ± SE: 15.86±4.36 vs 28.66±0.92) when compared to patients who had not (Figure 1).

Of the total of 32 patients with CLABSI, 29 had a catheter removed following positive culture, and 16 of them were removed within 72 hours. There was no difference for mortality (both p=0.33) between catheter withdrawals within 72 hours and those not withdrawn within 72 hours.

Discussion

In our study, 58.3% and 41.7% of IC episodes were related to non-albicans Candida species and C. albicans, respectively. C. albicans and C. parapsilosis were the most common isolated species (both 41.7%). Fluconazole, caspofungin, and AmB resistance rates were 38.2%, 3.1%, and 2.9%, respectively. 73.3% of C. parapsilosis isolates were resistant to fluconazole. Risk factors associated with mortality were neutropenia, dialysis, and abdominal surgery. Also, most of the IC cases were catheter-related candidemia and nosocomial candidiasis.

According to recent studies, non-albicans Candida species cause more than half of IC cases in children (8). Neu et al. (9), 74% of 203 episodes of pediatric candidiasis were found to be related to non-albicans species (43% C. parapsilosis). Our study confirmed that non-albicans Candida species have been increasing in recent years in IC as in previous studies, and it was remarkable that the frequency of C. parapsilosis was at least as frequent as C. albicans.

Various risk factors have been identified in candidiasis, and these are the presence of immunodeficiency, underlying disease, neutropenia, prematurity, transplantation, parenteral nutrition, surgery (especially gastrointestinal surgery), bacterial infections, malignancy, colonization with Candida spp., broad-spectrum antibiotics, corticosteroids and chemotherapeutic agents, CVCs, dialysis, endotracheal intubation, and stay in the ICU (10-12). Risk factors for IC in this study were similar to those reported in previous studies, and all patients had an underlying disease. Prematurity, congenital heart disease, and solid organ tumor were the most common underlying diseases. Most of our patients had CVCs and received broad-spectrum antibiotic treatment. Particularly, candidemia was catheter-related in 97% of patients with CVC. Also, 41 of 45 IC episodes (91.1%) were considered nosocomial. The high incidence of catheter-related infections and nosocomial candidiasis emphasized the importance of infection control measures and careful care.

Candida species are reported to have differences in clinical features and outcomes. Celebi et al. (12) reported that neutropenia and pre-infection hospitalization were more frequent in non-albicans candidemia. Mortality and disseminated candidiasis were higher in patients with C. albicans. Similarly, in a study with pediatric 248 candidiasis cases, the mortality rate for C. albicans was 34.1% contrast to a rate of 22.4% with non-albicans species (13). Dotis et al. (14) showed that using a mechanical ventilator in the two days before Candida infection is a significant risk factor for C. parapsilosis infection. In a retrospective case-control study at Texas Children’s Hospital in 276 episodes of candidemia, there was no difference between C. albicans and non-albicans candidemia in terms of demographics, underlying diagnosis, risk factors, clinical features, dissemination, or 30-day mortality (15). There was no difference between the invasive infections due to C. albicans versus non-albicans Candida species and C. parapsilosis versus non-parapsilosis, in our study. Our sample size might be not enough to determine a small difference.

With the increase of infections caused by non-albicans Candida spp., there are problems in the choice of empirical treatment of patients with IC. Studies show that antifungal resistance to non-albicans Candida species is higher (9). The sensitivity of Candida spp. to antifungal agents can usually be predicted if the species of infectious isolation are known.

However, some Candida spp. are not in parallel with the general sensitivity pattern similar to our study. We found that 61.8% of all Candida isolates were fluconazole sensitive, and 38.2% were fluconazole-resistant. The most prominent feature of our study was high fluconazole resistance in C. parapsilosis strains, which was 73.3%. However, fluconazole is the first recommended antifungal agent for C. parapsilosis infections (16). A prospective study which includes 1,218 episodes of Candida BSI conducted by using species-specific CLSI reference broth microdilution method for the sensitivity of antifungal agents established that 2.9% of C. parapsilosis was non-susceptible to fluconazole (17). Devrim et al. (18) reported that fluconazole resistance was 58.4% in 12 pediatric hematology and oncology patients with catheter-associated C. parapsilosis blood supply infection. In our country, revised susceptibility of 453 Candida strains isolated from adult and pediatric patients, according to CLSI M27-A3 criteria was evaluated. Fluconazole resistance was 1.4% in C. albicans, 18.2% in C. parapsilosis, 2.6% in C. tropicalis, and %14.3 in C. glabrata. There were no AmB-resistant isolates in this study. The highest resistance to fluconazole was in C. parapsilosis species (19). Therefore, antifungal susceptibility testing is beneficial for effective antifungal therapy. Considering the current dominance of non-albicans strains and fluconazole resistance rates in our hospital, the treatment of echinocandin or AmB as initial treatment is seen as a good option.

It is difficult to assess the death attributed to Candida, and the published attributable mortality rates vary depending on the type of study. The overall mortality rate in children with candidiasis ranges from 10% to 26% (8,20,21) In most studies, the presence of an arterial catheter, neutropenia, steroid treatment, insufficiency of antifungal therapy, prolonged antibiotic therapy, immunosuppressive conditions, disseminated candidiasis, use of TPN and mechanical ventilation, intensive care, and C. albicans isolation was established to be associated with mortality due to candidemia (12,22). In our study, the overall mortality rate was 13.3%, in C. albicans isolated patients 10%, in non-albicans Candida spp. isolated 14.2% and in C. parapsilosis isolated 15%. This mortality rate was within the ranges reported in the literature. Although there are studies analyzing the risk factors affecting mortality in the first 30 days, there are not enough studies investigating the risk factors affecting mortality in the first 7 days. In our study, dialysis, neutropenia were risk factors that increased mortality on the 7^th and 30^th days. Additionally, abdominal surgery is also a risk factor in the first 7 days of mortality. So in the first 7 days of follow-up of patients undergoing abdominal surgery, dialysis, and neutropenia with IC, caution should be exercised for IC and may be evaluated for prophylactic antifungal therapy.

The principal limitations of this study are relatively small sample size and it may not be possible to evaluate multivariate analyzes. Another limitation is that our studies are single-centered and cannot be generalized to other centers.

Conclusions

In summary, we found that C. parapsilosis and C. albicans are seen more frequently. When considering that the fluconazole resistance of C. parapsilosis is 73.3%, the most effective treatment seems to be AmB and caspofungin. Dialysis, neutropenia, and abdominal surgery were risk factors that increased mortality. These data will help us to identify patients who are at risk for IC and will guide us in the selection of empirical treatment.

Ethics

Ethics Committee Approval: This study was approved by the Dokuz Eylül University Non-Invasive Research Ethics Committee (decision no: 2016/13-23, date: 12.5.2016).

Informed Consent: Retrospective study.

Peer-review: Externally and internally peer-reviewed.

Authorship Contributions

Surgical and Medical Practices: Z.G.K., N.B., Concept: Z.G.K., N.B., M.C.E., M.D.D., Design: Z.G.K., N.B., M.C.E., A.N.E., M.D.D., Data Collection or Processing: Z.G.K., N.B., M.C.E., M.D.D., Analysis or Interpretation: Z.G.K., N.B., A.N.E., Literature Search: Z.G.K., N.B., A.N.E., Writing: Z.G.K., N.B., M.C.E., A.N.E., M.D.D.

Conflict of Interest: No conflict of interest was declared by the authors.

Financial Disclosure: The authors declared that this study received no financial support.