Antifungal Susceptibility and Risk Factors in Patients with Candidemia

Materials and Methods: A case-control study was designed, and data collected between December 2013 and December 2014 were retrospectively evaluated. The case group consisted of patients with candidemia. The control group was selected from the inpatients that did not develop candidemia but were admitted in the same clinic and during the same period as the candidemia group. The diagnosis of candidemia was based on a compatible clinical picture and positive blood culture of Candida spp. The demographic characteristics, sequential organ failure assessment (SOFA) scores, comorbidities, use of invasive devices, antibiotics administered, and duration of antibiotic uses were compared between both the groups.


Introduction
Candidemia and invasive candidiasis are major causes of nosocomial infections linked to a number of risk factors such as prior antimicrobial therapy, venous and urinary catheters, intensive care unit admission, parenteral nutrition, major surgery, and immunosuppressive therapies [1,2].In recent years, a serious increase in the number of invasive Candida infections has been reported.According to recent data, Candida species are the fourth leading cause of circulatory infections [3].Candida species have the potential to cause a wide spectrum of diseases, ranging from simple superficial infections in the skin and mucosa to more serious infections in deeper tissues, which may prove fatal.Serious Candida infections include deep, invasive, systemic, disseminated, and hematogenous infections [4].In the past, 70% -90% of the isolated agents in Candida infections have been reported to be C. albicans, while 5% comprised C. glabrata and C. tropicalis.The other Candida species were rarely isolated.Epidemiological studies have indicated that the infection pattern has shifted from C. albicans in favor of the non-albicans species, including C. glabrata, C. tropicalis, C. parapsilosis, and C. krusei [5].Besides the shift in the epidemiological pattern, a serious change in the resistance of the Candida species to conventional therapies has also been reported [6].It is rather difficult to diagnose invasive Candida infections other than candidemia and candiduria.Therefore, clinicians should meticulously evaluate the risk factors for Candida infections and act based on these data [7].The increase in the incidence of fungal infections and the related morbidity and mortality rates lead to a parallel increase in the empirical antifungal use, to the emergence of resistant fungal species, and consequently to increased resistance rates.Thus, the need for in vitro antifungal susceptibility tests in order to select an appropriate and efficient antifungal therapy is ever increasing [8].The aim of this study is to observe the risk factors, isolated Candida species, and antifungal susceptibility in patients with candidemia.

Materials and Method
The study was conducted using a retrospective approach on patients who developed candidemia at the Dicle University Hospital between December 2013 and December 2014.The approval of the Ethics Committee of the Dicle University School of Medicine was obtained for this study.The study was conducted on the patients admitted at the Dicle University School of Medicine Hospital, clinics, and intensive care units.Dicle University Hospital is a tertiary referral hospital providing healthcare to approximately 3.5 million people living in Diyarbakir and the peripheral provinces.The study was designed as a case-control research.Patients in whose blood culture Candida spp. was observed were included in the case group.We randomly selected at least 2 control patients for each case who were matched according to the age groups, hospitalized in the same department in the same period, with no signs and symptoms of candidemia, and with negative blood cultures for Candida spp.A corresponding control subject for each patient was selected in the control group.A form was designed in order to record the data of the study and control subjects.This form included a number of risk factors that we suspected, which were also observed in previous similar studies.Variables including the hemogram, AST, ALT, urea, creatinine, CRP, SOFA (score, length of the hospital stay, and time until mortality in the case and control groups were statistically compared.The cultures from the case group were tested at the microbiology laboratory of the Dicle University School of Medicine Hospital.Informed consent was obtained from the patients.The blood samples were collected from both arms of the subjects using aseptic techniques and incubated in blood culture bottles.The volume of each sample was 10 ml in adults and 1-3 mL in pediatric patients.The blood sample bottles were placed in BACTEC 9240 and BACTEC 9120 (Becton Dickinson, USA) devices to be stored.The blood cultures in which the growth was detected by the BACTEC device were inoculated in blood agar, eosin methylene blue agar, and Sabouraud dextrose agar (SDA).The samples incubated in the media were stored in an incubator at 37°C for 18-24 h (WTB Binder, Germany).For identification, the isolates that grew in the SDA were both directly examined with fresh preparation and Gram staining was applied.Further, the germ tube growth of the isolates in a human serum was tested.The fully automated identification was performed using the VITEK 2 AST and YST commercial test kits.The antifungal sensitivity of the isolates was also tested using a fully automated method through the VITEK 2 AST and YST commercial test kits.The VITEK AST and VITEK YST commercial kits were procured within the framework of the Scientific Research Project.These tests were performed at another center.

Statistical analysis
The data were analyzed using the SPSS 22.0 (IBM statistics for Windows version 22, IBM Corp., Armonk, New York, United States) software package.The normality of the distribution of the data was assessed based on the Shapiro-Wilk test, and the variables with a normal distri-bution were tested using parametric methods, while the variables outside the normal distribution were tested using non-parametric methods.The comparison of the two independent groups was performed with the help of the independent samples T-test, while the Mann-Whitney U-test was used with the Monte Carlo simulation technique.The correlation of the variables was examined through Spearman' s rho test.For the comparison of the categorical data, Pearson' s chi-square and Fisher' s exact tests were compared with the Monte Carlo simulation and exact test results.The logistical regression test was used to observe the causal relationship of the categorical response variable with the explanatory variables in the binomial and multinomial categories.The quantitative data are expressed as mean±SD.(standard deviation), median±IQR (interquartile range), and median range (maximum-minimum).The categorical variables are expressed as number (n) and percentage (%).Data were evaluated with a 95% confidence interval, and statistical significance was based on a p value below 0.05.

Results
For the purposes of this study, 42 candidemia patients and 42 controls were enrolled.The mean age of the case group was 47.3±32 years, and the mean age of the control group was 56.7±27.2years.In the case group, 28 patients (67%) were male, while 25 patients (60%) were male in the control group.No statistically significant difference was observed in terms of the mean age and gender between the groups (Table 1).The case and control groups were most frequently admitted to the internal medicine intensive care, pediatrics, and thoracic diseases intensive care units.The distribution of the clinics where the patients were admitted is presented in Table 2.When the clinical findings of the study and control groups were compared, the prevalence of nosocomial infections (infections that have developed within 48-72 h after admittance or within 10 days after discharge, which were not in the incubation period at the time of admittance), sepsis (demonstration of the systemic inflammatory syndrome to be associated with the infection and 2 or more of these symptoms), candiduria, and fever was statistically significantly higher in the case group (p=0.012,p=0.004, p≤0.001, and p=0.027, respectively).Further, abdominal surgery, CVP catheters, total parenteral nutrition, endotracheal intubation, frequency of blood transfusions, and SOFA scores were found to be significantly higher in the case group in comparison to the control group (p=0.002,p=0.028, p<0.001, p=0.045, p<0.001, and p=0.024, respectively) (Table 1).The logistic regression test demon-strated that TPN and blood transfusion are the most important predictors for candidemia (OR=8.14 and OR=5.96, respectively) (Table 3).Among the 42 strains isolated in the study, 22 (52.4%) were identified as C. albicans, while the others were non-albicans Candida strains.The non-albicans Candida strains were subgrouped as C. parapsilosis (16.7%), C. glabrata (9.5%), C. tropicalis (7.1%), C. krusei (4.8%), and non-typed non-albicans Candida species (9.5%) (Table 4).Only one C. albicans strain (4.5%) was resistant to fluconazole, while 7 among the non-albicans Candida strains (35%) were resistant to fluconazole.No resistance was observed among the C. albicans strains to the other antifungal agents such as amphotericin B, flucytosine, caspofungin, and voriconazole.The non-albicans Candida strains were found to be resistant to flucytosine in 4 patients, voriconazole in 1 patient, and amphotericin B in 1 patient.No resistance was observed to caspofungin.

Discussion
In recent years, besides the increase in the number of Candida infections, a shift is also observed in the variety or the species causing these infections.Although C. albicans takes the lead in the nosocomial Candida infections, the frequency of non-albicans Candida species such as C. tropicalis, C. lusitaniae, C. krusei, C. parapsilosis, and C. glabrata, which are known to be more resistant to antifungal treatment, is increasing [9,10].According to the studies, the incidence of invasive Candida infections is similar in males and females [9,10].Also, in our study, although 67% of the patients with candidemia were males, gender has not been observed to be a factor in the development of candidemia.In 52.4% of candidemia patients in our study, C. albicans was found to be the agent, and this result was in compliance with the previous studies [11][12][13][14][15]. C. parapsilosis is observed worldwide, except for North America [16,17].In our study, C. parapsilosis was the second most commonly isolated agent (16.7%).In our study, only 4 candidemia episodes (9.5%) were caused by C. glabrata, and this low prevalence was partly associated with the low levels of prophylactic fluconazole use.In a multicenter study, the frequency of C. albicans was reported as 68%, while this ratio was reduced to 63% in 2005.The frequency of C. glabrata was around 10%.Interestingly, the ratio of C. tropicalis was increased from 5.2% to 7.3% [2].In a study from our country, C. tropicalis was reported as the second most frequent agent [18] [20].In a study conducted in Italy, the length of hospitalization, CVP catheterization, history of bacteremia, total parenteral nutrition, and chronic kidney failure were also observed to be independent risk factors [21].In a study from our country, history of antibiotherapy, blood transfusion, total parenteral nutrition, and urethral catheterization were reported to be independent risk factors for candidemia [18].In another study from our country focusing on the risk factors in patients in which candidemia has developed, CVP placement was found to be an independent risk factor [22].Geographical differences, patient characteristics, differences in the hospital environment, numeric difference of candidemia episodes, and differences in the sensitivity of the isolation method may partially explain the difference in the study results.Among the risk factors of candidemia investigated in our study, only CVP catheterization was found as an independent risk factor in the multivariate logistic regression analysis.In case-control studies, the mortality rate varies between 49% and 55% [16,18,20].In our study, the mortality rate was determined as 46.4%.The duration of our patients' hospital stay and the longer follow-up (30 days) to evaluate mortality in our study (30 days) may explain the higher mortality rate in our series.In the study by Diekema et al. [23], fluconazole resistance was found as 3% for C. albicans, 10% for C. glabrata, and 7% for C. tropicalis.In another study where antifungal susceptibility was investigated in 35 Candida isolates, all the isolates were observed to be susceptible to amphotericin B, while fluconazole and itraconazole susceptibilities were 89% and flucytosine susceptibility was 97% [24].In the study by Yuksekkaya et al. [25].conducted on 56 urinary candidiasis isolates, no resistance to amphotericin B was observed, although resistance to fluconazole was observed in 1 C. glabrata (5.2%) and 1 (6.7%) C. tropicalis isolate.
In a study conducted in our region, the distribution and antifungal susceptibility of the Candida species isolated from the blood cultures were investigated.According to the results of this study, a majority of the isolated species were non-albicans Candida species with a 52.3% ratio.Our study has been observed to be in compliance with previous studies.The fact that no routine Candida typing and susceptibility testing are performed at the microbiology laboratory of our hospital constitutes a limitation of our study.Consequently, since there was no data about the previous Candida strains and their susceptibility at hand, no comparison could be made.Further, the number of patients in the case-control study was limited.The control group did not correspond to the case group, and it would have been a better alternative if 2 controls were enrolled for each patient.This is a further limitation of our study.Although the number of patients we had within a year was higher, since some of the isolated strains were not stored and subjected to typing and susceptibility testing, the number of our patients remained at 42.
In conclusion, the risk of candidemia should not be overlooked in patients who do not respond to antibiotic treatment, received TPN, with inserted CVP catheters, received blood transfusion, or have undergone abdominal surgery.In the study by Eksi et al. [26] conducted in our region, Candida species were typed and tested for antifungal susceptibility.However, since the risk factors were not evaluated in this study, our study becomes prominent as the first one to take these factors into consideration.Furthermore, our study is the first to focus upon the types and antifungal susceptibility of Candida strains isolated at our hospital.Knowing the types and susceptibility of Candida strains will be an important guiding factor in our choice of antifungal therapy.
In particular, since approximately half (47.6%) of Candida strains isolated at our hospital are non-albicans Candida strains and a 35% fluconazole resistance was observed, this resistance to fluconazole at our hospital should be considered during the treatment of severe Candida infections.

Table 1 .
Distribution of demographic characteristics of case and control groups . Between 2000 and 2003, C. albicans (57.7%), C. tropicalis CVP: central venous pressure; SOFA: sequential organ failure assessment

Table 2 .
Departments developing candidemia ICU: intensive care unit

Table 3 .
Risk factors for candidemia case and control groups: "logistic regression" test CVP: central venous pressure; SOFA: sequential organ failure assessment

Table 4 .
Distribution of candida strains [25]ur study, the most frequently isolated fungi were C. albicans species.While 15.5% of the non-albicans Candida species in this study were fluconazole-resistant, only 5.66% of the C. albicans species were resistant to fluconazole[25].