Characteristics and clinical outcomes of patients with Candida bloodstream infections in a tertiary care hospital in Jordan

Introduction: Candida species are important causes of bloodstream infections, accounting for significant morbidity and mortality in hospitalized patients. Methodology: A retrospective observational study was conducted in an academic tertiary hospital of Jordan. The medical records of patients hospitalized over a ten-year period. were reviewed and patients with candidemia were identified. Data analysis included the infecting Candida species, resistance to antifungals, and risk factors associated with mortality. Results: A total of 158 cases of candidemia were identified, with an overall incidence rate of 0.48 episodes/1,000 admissions. The proportion of candidemia caused by Candida albicans (44.3%) was higher than that of candidemia caused by non-albicans Candida species (42.4%). Exposure to antibiotic therapy before hospitalization was the only independent factor associated with non-albicans Candida infection (OR 2.454; p = 0.033). The overall crude 30-day mortality was 38.7%. Central venous catheterization (OR 0.255; p = 0.026), mechanical ventilation (OR 0.162; p = 0.003), severe sepsis and septic shock (OR = 0.073; p = 0.008), admission to intensive care unit (OR 0.78; p = 0.001), C. albicans (OR 0.235; p = 0.018), length of stay (OR 1.057; p = 0.001), number of comorbidities (OR 0.580; p = 0.008) were independent risk factors for 30-day mortality. Conclusion: This study identified several risk factors associated with blood stream infections caused by Candida over 10-years period. Continuous surveillance programs to monitor such types of infection are of great value to antimicrobial stewardship programs.


Introduction
Nosocomial bloodstream infections caused by Candida species have increased considerably in the latest years. They are associated with prolonged hospitalization time and increased mortality [1][2][3][4]. Candida species are estimated to be the fourth pathogen isolated in blood cultures in North America and the fifth to tenth in Europe [5,6]. In the past, the most common causative pathogen was Candida albicans but recent studies reported a global change in Candida spp. distribution toward non-albicans Candida species [7,8]. This change in the epidemiology of candidemia has challenged the empirical treatment strategies in infected patients. This is complicated by two factors; resistance or variable susceptibility to azole antifungals which are commonly prescribed in empirical therapy and delay in the identification of candidemia due to suboptimal sensitivity of traditional blood cultures [9][10][11]. Accordingly, it is of great importance the identification of risk factors and predictors of mortality in patients with candidemia to guide empirical therapy and develop strategies to improve clinical outcomes. The aims of this study were to describe the annual incidence of candidemia over a 10-years period in a Jordanian tertiary care hospital and to identify risk factors associated with Candida bloodstream infection (BSI) and mortality in infected patients.

Study Design
This was a retrospective analysis of patients with blood cultures positive for Candida species, admitted at King Abdullah University Hospital (KAUH), a tertiary care hospital in Jordan, between January 1, 2005 and December 31, 2014. The approval to conduct this study was granted by the ethics committee in KAUH.

Data collection
Candidemia was defined as the identification of Candida species in a blood culture specimen with concomitant clinical signs and symptoms of infection. The day of the first positive blood culture was considered as the onset of candidemia. If multiple episodes of candidemia occurred in the same patient during the study period, the patient was considered as one case, using only the first episode.
The following demographic and clinical parameters were obtained from electronic patients' medical records and charts: Patient demographic characteristics, drug allergy, co-morbidities, hospitalization in the previous 90 days, previous antibiotic use (within 90 days before admission), and hospital transfer. Events that occurred during hospitalization before and after Candida development were also examined, including primary critical care unit admission, prior antibiotic or antifungal use, prior corticosteroids use, length of hospital stay and length of stay in critical care unit, complete blood count analysis, blood product transfusion during hospital stay, type of Candida isolated from blood, antifungal used for the empirical and the definitive treatment and Candida spp isolated from other body sites.
Neutropenia was defined as absolute neutrophil count ≤ 500 cells/mm during the previous 14 days of the onset of candidemia. Acute renal failure was defined as an increase in serum creatinine by 1.5 times or more from baseline [4]. Finally, 30-day in-hospital mortality was defined as death within 30 days of the first Candida positive blood culture.

Statistical Analysis
Descriptive data were presented as mean±SD or median and interquartile range for continuous data, whereas frequencies and percentages were used to summarize categorical data. To identify independent risk factors associated with 30-day morality in patients with candidemia, multivariable stepwise logistic regression analysis was conducted. Variables with a pvalue of ≤ 0.10 on univariate analysis were included in the multivariable model. Odds ratio (OR) with 95% confidence intervals (CI) were calculated. The same method was used to identify risk factors of C. albicans vs. non-albicans Candida species infections. All tests performed were 2-tailed tests of significance and pvalue less than 0.05 was considered significant. Statistical analysis was performed using SPSS (Statistical Package for the Social Sciences) version 20.

Trends in Candida Blood Stream Infections Over a 10-Years Period
A total of 158 cases of candidemia were identified during the study period. The most commonly isolated Candida species were C. albicans (44%), followed by C. parapsilosis (15%) and C. tropicalis (11%). Other Candida species accounted for 17% of the isolates (6% C. glabrata, 2% C. krusei, 2% C. famata, 1% C. guilliermondii and 6% other Candida species). The remaining isolates (13%) were mixed or unknown Candida species. The average time of candidemia incidence after admission was 20 ± 12 days.
The overall incidence of candidemia was 0.48 cases/1000 admissions. The incidence rate fluctuated over the study period, with the lowest incidence in 2008 (0.32 cases/1000 admissions) and the highest incidence in 2009 (0.72 cases/1000 admissions). The annual incidence rates of candidemia over the 10-year period are shown in Table 1. Overall, the resistance to antifungal agents was low over the study period. The percentage of susceptibility to fluconazole was 100.0% for C. albicans, C. parapsilosis and C. tropicalis. C.glabrata was 87.5% susceptible to fluconazole, while C. famata and C. krusei were resistant. All C. parapsilosis and C. tropicalis isolates were susceptible to amphotericin, voriconazole, fluconazole, and flucytosine.

Clinical Characteristics and Risk Factors of Patients with C. albicans vs. non-albicans Candida BSIs
A total of 119 patients out of 158 had complete medical records and 104 patients of them had one species of Candida species. When patients' baseline characteristics and clinical details stratified by two categories of Candida species were compared, it was shown that patients with C. albicans BSIs had a higher overall death rate (48%) compared to patients with nonalbicans Candida BSIs (32%) (p = 0.008). Table 2 displays the results of the univariate and multivariable logistic regression analyses. In univariate analysis, patients with non-albicans Candida were more likely to have received antibiotics in the previous 90 days (p = 0.034) and empirical or prophylactic fluconazole treatment (p = 0.091) compared to patients with C. albicans. However, in multivariable regression, only previous exposure to antibiotics in the previous 90-days remained independently associated with non-albicans Candida BSI (OR 2.454, 95% CI 1.072-5.615; p = 0.033).

Risk factors associated with the 30-day mortality of patients with candidemia
Previous antifungal therapy (prophylaxis or empirical) was administered in 22.7% of patients, while 9.24% of patients did not receive any antifungal treatment. Fluconazole was the most frequently used antifungal (59.7%), followed by amphotericin B  Length of hospital stay (days) (23.5%) and caspofungin (10.9%). Only three patients (2.5%) received a combination therapy (fluconazole + voriconazole or fluconazole + amphotericin B). Mortality was lower if antifungal therapy was initiated within 24 hours of culture result compared to therapy after 24 hours (37.0% vs 41.3%, p = 0.031).

Discussion
This study examined the annual incidence of Candida BSIs and species distribution over a 10-year period in a tertiary hospital in Jordan. The overall incidence of candidemia was found to be comparable to that of other studies around the world [8,[12][13][14][15][16]. However, the reported incidence rate is higher in hospitals in Latin America and Italy (1.2 to 1.7 episode/1000 admission) and lower in hospitals in Northern Europe (0.01 to 0.08 episode/1000 admission) [17][18][19]. On regional level, the incidence rate was higher (3 episodes/1000 inpatient days) in a pediatric tertiary hospital in Egypt [20]. While in a university hospital in Saudi Arabia 4.7% of blood cultures were positive for Candida [21]. The incidence of candidemia was fluctuating over the study period. This fluctuation could be explained by an inconsistency in hospital's staff adherence to preventive measures and infection control practices, including hand washing, skin disinfections and removal of unnecessary catheters [22]. While C. albicans is still considered the most frequent species causing candidemia, trends for growing rates of non-albicans Candida have been reported [2,7,13,23]. The current study observed a fluctuation in non-albicans Candida species over the study period with the highest incidence in 2014.
In this study, previous exposure to antibiotics was found to be the only independent risk factor associated with non-albicans Candida BSIs. This would call attention to the importance of appropriate antibiotic use in hospitals and of antimicrobial stewardship programs in controlling and preventing unnecessary use of antibiotics [24,25]. Previous studies have identified other risk factors associated with non-albicans Candida, including glucocorticoids use, central venous catheter placement, candiduria, male gender, total parental nutrition and recent history of solid tumors [24][25][26][27][28][29][30].
The overall 30-day crude mortality was 38.7%, a finding comparable to that of other studies. In a tertiary center in the United Kingdom, the 30-day crude mortality in patients with candidemia was 37% [13]. Similarly, the reported 30-day crude mortality in two Italian tertiary hospitals was 38.8% [3]. However, in a university hospital in Saudi Arabia, the overall mortality rate was higher (71%) [21]. In addition, 30day crude mortality was higher in patients with C. albicans compared to patients with non-albicans Candida (48% vs.31.5%) BSI. In the US, Chow et al have reported no variation in hospital mortality among the two groups (57% versus 58%) [24]. Playford et al reported higher mortality in patients with non-albicans Candida compared to C. albicans infections (53% versus 41%) [26]. In contrast, Wang et al reported higher crude mortality with C. albicans BSIs (44.3% versus 29.8%) [25]. In this study, C. albicans was also a significant predictor of mortality. While C. albicans is more virulent than non-albicans Candida which could explain the higher mortality rate, mortality is not solely related to the virulence of the Candida pathogens but it is also related to the severity of the underlying illness, the comorbidities, and the integrity of the defense mechanisms of the host as has been shown previously [27]. In our study, ICU stay, central venous catheterization, mechanical ventilation, length of hospitalization, number of comorbidities, severe sepsis and septic shock were identified as predictors of mortality in patients with candidemia. These results are in agreement with previously published studies that reported risk factors for mortality in patients with candidemia [4,16,20,28,[31][32][33].
This study has some limitations. First, it is a retrospective study performed at a single center and the results may not be applicable to other settings. Second, the severity of illness score (such as APACHE II score) was not included because of unavailable data; however, number of comorbidities was used. Third, mortality measured in this study is crude mortality which may be affected by patient's medical condition. Finally, there were limitations in data accessibility and availability.

Conclusion
This study is the first longitudinal epidemiological study of Candida BSIs in a Jordanian tertiary care hospital. The annual incidence of candidemia was fluctuating over the 10-year surveillance period. Previous antibiotic use was the only risk factor associated with non-albicans Candida BSIs and multiple underlying medical conditions were identified as predictors of 30-day crude mortality in candidemia patients. Based on these findings, continuous local surveillance programs, strict adherence to infection control measures and implementation of antimicrobial stewardship programs should be considered to improve clinical outcomes in patients with candidemia.