Summary
Introduction: Mixed candidemia (MC) and late recurrent candidemia (LRC) are rare conditions. Studies on these issues are limited. Our aim was to investigate the characteristics, risk factors, and outcomes of these rare conditions.
Materials and Methods: The study was carried out between May 2019-May 2021 in Ankara City Hospital as a cross-sectional descriptive study. Mixed candidemia was defined as the isolation of at least two Candida species in two blood culture bottles taken within 72 hours or one blood culture bottle taken simultaneously (from a patient with clinical symptoms). Late recurrent candidemia was defined as the recurrence of candidemia at least 30 days after the treatment of the first candidemia episode was completed and the symptoms resolved. In the study, MC patients were compared with monomicrobial candidemia patients. The LRC group was compared with the single episode candidemia group.
Results: During the study period, 549 candidemia episodes were detected in 533 patients. Mixed candidemia was detected in 38 (7.1%) of these patients. Late recurrent candidemia was seen in 16 (10.7%) of 149 patients who recovered after treatment and survived above 30 days. History of abdominal surgery and instrumentation, chemotherapy and transplantation were significantly higher in the MC group compared to the monomicrobial candidemia group. Compared to the single episode candidemia, the LRC group had significantly higher rate of abdominal surgery and instrumentation history, concomitant bacteremia, Candida colonization index of >0.5 and 1-year follow-up mortality.
Conclusion: Both MC and LRC were significantly more common in patients who had undergone abdominal surgery and instrumentation. With the development of diagnostic methods, we may encounter these rare cases more often. Since candidemia has a high mortality rate, recurrent candidemia may be overlooked. According to the results of the study, there is a 10% risk of recurrence in recovered candidemia cases.
Introduction
Candida bloodstream infections (BSIs) are an increasingly important healthcare-associated infection. In addition to being a cause of serious morbidity and mortality, Candida BSIs also cause increased length of hospitalization and higher costs[1, 2]. Numerous studies have identified risk factors for candidemia[3-5]. However, mixed candidemia (MC) and late recurrent candidemia (LRC) are rare conditions[6, 7]. Therefore, there are limited data on the characteristics, risk factors, and outcomes of the patients with MC and LRC. In this study, we aimed to answer these questions and fill this gap since early diagnosis and treatment are crucial in these infections, which have crucially higher mortality rates.
Methods
Study Design, Settings and Participants
The study was carried out between May 2019 and May 2021 in Ankara City Hospital (3,810 bed capacity including 690 intensive care unit beds) as a cross-sectional descriptive study. Ethics committee approval was obtained from Ankara City Hospital Ethics Committee (no: E1-21-1911, date: 23.06.2021). Patients over 18 years of age with Candida isolated in clinically significant blood cultures were included in the study. The MC patients were compared with monomicrobial candidemia. The LRC group was compared with the single episode candidemia (Figure 1). For LRC, data from the first episode were used in the study. Patients with recurrent candidemia within 30 days of recurrence were not included in the LRC group. According to the case definitions, they were included in one of the other two groups. Demographic characteristics, risk factors, type of Candida spp. and its susceptibility pattern, and outcomes of the patients were collected from the hospital electronic record system.
Definitions
Candidemia was defined as the isolation of one or more Candida isolates in at least one blood culture bottle with findings consistent with infection. MC was defined as the isolation of at least two Candida species in two blood culture bottles taken within 72 hours or one blood culture bottle taken simultaneously (from a patient with clinical symptoms). Late recurrent candidemia was defined as the recurrence of candidemia (any of the Candida species) at least 30 days after the treatment of the candidemia episode was completed and the symptoms resolved. The single episode candidemia group included patients who recovered from candidemia, lived at least 30 days, and did not have a recurrence of candidemia.
Nosocomial candidemia was defined as the patient who had candidemia at least two days after hospitalization without recent hospitalization history. In order for candidemia to be accepted as central venous catheter (CVC) related, the same Candida species should be isolated in the simultaneous catheter culture with the peripheral blood culture. Previous use of broad-spectrum antibiotics was defined as the use of one or more antipseudomonal cephalosporins, piperacillin-tazobactam, carbapenems, fosfomycin, colistin, and tigecycline antibiotics within one month. High-dose steroid was defined as using more than 15 mg of prednisolone (or its’ equivalent) for more than three weeks. Receiving chemotherapy within the last month was classified as a possible risk factor. Neutropenia was defined as the presence of absolute neutrophil count less than 0.5x109/L. Abdominal intervention history was defined as having abdominal surgery or gastric instrumentation within three months. In the study, in order for total parenteral nutrition (TPN), mechanical ventilation and central venous catheter use to be considered as possible risk factors, it should have started at least two days before candida isolation. Candida colonization index (CCI) (CCI=Number of sites colonized/number of sites cultured, threshold >0.5)[8] and Candida score [multifocal colonization (1 point), sepsis (2 points), surgery (1 point), TPN (1 point), threshold >2.5][9] were also evaluated among the possible risk factors in the study.
Early mortality was defined as deaths occurring within seven days of Candida isolation. Related mortality was defined as death occurring within five days of candidemia without any other concurrent signs of infection and no other apparent cause[6]. In the one-year follow-up of mortality in LRC patients, the one-year period after the second attack of candidemia was considered. One-year mortality follow-up was followed from hospital electronic records and national death notification system. The taxonomy of the species previously included in the genus Candida has changed. Since this update would not make a difference to clinical practice, the old taxonomy was followed.
Microbiological Identification
Microbiological tests of the study were performed in Ankara City Hospital central microbiology laboratory. The BacT/Alert (bioMérieux) automated blood culture system was used for monitoring blood culture bottles. Candida spp. isolates were identified using the VitekMS (bioMérieux) device and the MALDI-TOF MS method. Susceptibility tests were evaluated with the VITEK® 2 Compact automated system (bioMérieux, France) according to the European Committee on Antimicrobial Susceptibility Testing.
Statistical Analysis
For those without normal distribution in descriptive statistics on continuous variables were calculated as median, interquartile range (IQR) [Quartile 1 - Quartile 3 (Q1 - Q3)]. For categorical variables, count and percentages were calculated. The conformity of the variables to the normal distribution was examined using visual (histogram and probability graphs) and analytical methods (Kolmogorov-Smirnov test). The Pearson chi-square test and the Fisher’s exact probability test were used to compare categorical variables. Student’s t-test and the Mann-Whitney U test were used for continuous variables. Statistical significance level was accepted as p<0.05. IBM Statistical Package for the Social Sciences statistics for Windows (IBM Corp. Released 2013. version 22.0. Armonk, NY: IBM Corp.) program was used for statistical analyses.
Results
During the study period, 549 candidemia episodes were detected in 533 patients. Mixed candidemia was observed in 38 (7.1%) patients. Late recurrent candidemia was seen in 16 (10.7%) of 149 patients who recovered after treatment and lived longer than 30 days. In 549 candidemia episodes, 587 Candida spp. were isolated, such as C. albicans (n=278, 47.4%), C. parapsilosis (n=124, 21.1%), C. glabrata (n=79, 13.4%), and C. tropicalis (n=68, 11.6%), in order of frequency. In addition, a total of 10 different Candida species were isolated from the patients in the study (Table 1). When all groups were examined separately, it was determined that C. albicans was the most frequently isolated one. Isolated Candida species in MC and LRC were detailed in Tables 2 and 3.
Mixed Candidemia
Of 38 MC patients, 15 were male and the median age of the patients was 68 years (IQR: 56-79). The most common comorbidity in MC patients was hypertension (39.5%), and the most common risk factors for candidemia in these patients were CVC (100%), previous use of broad-spectrum antibiotics (97.4%), and mechanical ventilation (89.5%), respectively. Concomitant bacteremia was detected in 15 (39.5%) patients. Characteristics, treatments, and outcomes of MC patients are shown in Table 2. During the follow up, 33 (86.8%) of the MC patients died. It was recorded as a candidemia-related death in 12 (31.5%) patients.
In the comparison of MC patients with monomicrobial candidemia patients, MC was found to be significantly higher in female patients [p=0.044; odds ratio (OR)=1.981, 95% confidence interval (CI)=1.009-3.887], and those with a history of abdominal surgery and instrumentation (p=0.018; OR=2.276, 95% CI=1.136-4.557), chemotherapy (p=0.042; OR=2.338, 95% CI=1.054-5.185), and transplantation (p=0.047; OR=4.629, 95% CI=1.199-17.871). In addition, ocular involvement (p=0.04; OR=9.219, 95% CI=1.57-54.142) was significantly higher in the MC group compared to the control group. Detailed comparison of MC and monomicrobial candidemia is shown in Table 4.
Late Recurrent Candidemia
The median time between patients’ candidemia episodes was 70 (IQR: 40-138) days. Seven of the 16 LRC patients were male and the median age of the patients was 61.5 (IQR: 57-75). Hypertension (n=7, 43.8%) was the most common comorbidity in LRC patients. In addition, the most common risk factors in patients were CVC (n=16, 100%) and previous use of broad-spectrum antibiotics (n=15, 93.8%), in order of frequency. The Candida species isolated in the patients, antifungal resistance status of the isolates, treatments, and outcomes are summarized in Table 3.
Compared to the single episode candidemia group, the LRC group had a significantly higher rate of abdominal surgery and instrumentation history (p=0.042; OR=3.03, 95% CI=1.054-8.712), concomitant bacteremia (p=0.024; OR=3.214, 95% CI=1.117-9.249), CCI of >0.5 (p=0.011; OR=4.936, 95% CI=1.078-22.596), and 1-year follow-up mortality (p<0.001; OR=6.5, 95% CI=1.979-21.349). The detailed comparison of the LRC and single episode candidemia groups for risk factors and outcomes is presented in Table 5.
Discussion
In our study, the proportion of MC was 7.1%. Different rates (1.9-5.3%) of MC have been reported in the literature[10, 11]. One reason for this is differences in the definition of MC in the studies. In some studies, isolation of different Candida spp. in a single culture is accepted, while in others, isolating different Candida within three days is accepted as MC[10, 11]. Another reason for the variation in incidence values between studies may be the difference in the microbiological methodology. In our study, the MALDI-TOF MS method was used for Candida identification. The use of MALDI-TOF has a high accuracy in the identification of clinical pathological yeasts[12]. This can be considered as one of the strengths of our study. Increasing detection of MC with developing methods and the detection of resistant strains may lead to appropriate treatments. Mixed candidemia has been associated with high mortality in the literature[10, 11] More studies are needed as it is a rare condition.
Overall, in our study C. albicans has been the most frequently isolated Candida species, and C. parapsilosis has been the second common. In many other studies, C. albicans has been seen most frequently[10, 13]. In the study, the combination of C. glabrata and C. parapsilosis was the most frequently isolated combination in patients with MC. In other studies, C. albicans and C. glabrata were the most common combination[6, 10]. According to the literature, the incidence of C. parapsilosis in nosocomial candidemia is gradually increasing[13-15].
Compared with the monomicrobial candidemia group, the history of abdominal surgery and chemotherapy was found to be higher in the MC group. In the study of Boktour et al.[16] in cancer patients, it was found that receiving chemotherapy within one month was significant for MC, as in our study. Disruption of the integrity of the intestinal mucosa by abdominal surgery and chemotherapy may cause microbial translocation.
Late recurrent candidemia was seen in 16 (10.7%) of the 149 patients who recovered after treatment and lived longer than 30 days. In other studies, in the literature, LRC has been reported between 2-9%[7, 17]. Candidemia is a clinical condition with high mortality[2]. Most of the patients die in the early period[18]. Approximately 70% of the 495 patients in the monomicrobial candidemia group died within 30 days. For this reason, there is a group of patients for whom we did not know whether candidemia will recur. This situation complicates the investigation of LRC cases. According to recovered candidemia patients, there is a risk of the recurrence of candidemia in 1 out of every 10 patients with a history of candidemia. Therefore, LRC may be an overlooked condition.
The median time between the first and second episodes was 70 (2.3 months) days. The longest recurrence period was 520 days. The longest time to recurrence in the literature was reported as 14 years[19]. In the literature, the median duration of recurrence varies between 1 and 6 months[7, 17, 19, 20]. In our study, abdominal surgery and instrumentation were associated with LRC. In other studies, gastrointestinal diseases have been found to be associated with LRC[7, 19, 20]. Gastrointestinal Candida colonization is likely to cause hematogenous spread in the LRC patients. In our study, 1-year mortality was observed as 75% in LRC patients (p=<0.001). In previous studies, mortality rates were found to be higher (45-50%) in the LRC group compared to the control group, but it was not statistically significant[7, 20]. C. parapsilosis was isolated in the second episode in 10 (62.5%) of the LRC patients. In seven of these patients, the source of candidemia was found to be the catheter. Munoz et al.[7] also found similar results in their studies. C. parapsilosis typically forms a biofilm on catheters and similar implanted devices[2].
The number of cases (549) is not relatively low for a two-year follow-up. Guerra-Romero et al.[22] reported 645 cases in 13 years of follow-up (1972-1985). Jensen et al.[6] reported 747 cases in their 21-year follow-up (1985-2006). Ramos et al.[10] reported 779 candidemia episodes in a multicenter study with the participation of 29 hospitals between 2010 and 2011. According to the literature, the frequency of candidemia is increasing[14]. In particular, nosocomial candidemia has increased in the last few decades[13, 15]. In addition, its frequency may vary according to the region where candidemia is reported[23].
Study Limitations
The limitation of the study is that the follow-up period of the study was two years. Since rare cases were investigated in the study, the small number of patients and the retrospective nature of the study are the limitations of the study. Multicenter studies may be performed for these rare conditions. In addition to the current limitations, the fact that many cases may have been missed when the growth performance of Candida species in blood culture is considered, no matter how well the identification is performed. Another limitation of the study is that confounding control could not be made through multivariate analyses. The strengths of the study are the clear definition of LRC and the exclusion of patients who died within one month from the single episode candidemia group. This situation prevents the bias that may occur in the study. Another strength of our study is the use of MALDI-TOF for Candida isolation. The use of MALDI-TOF has high accuracy in the identification of clinical pathological yeasts.
Conclusion
Both MC and LRC were significantly more common in patients who had undergone abdominal surgery and instrumentation. With the development of diagnostic methods, we may encounter these rare cases more often. Since candidemia has a high mortality rate, recurrent candidemia may be overlooked. According to the results of the study, it should be taken into account that late recurrence may occur in one out of every ten surviving candidemia cases.
Ethics
Ethics Committee Approval: Ethics committee approval was obtained from Ankara City Hospital Ethics Committee (no: E1-21-1911, date: 23.06.2021).
Informed Consent: Retrospective study.
Authorship Contributions
Surgical and Medical Practices: B.O.Ö., A.B., H.B., Concept: B.O.Ö., A.B., D.G., B.D., H.B., Design: B.O.Ö., A.B., N.K., D.G., B.D., H.B., Data Collection or Processing: B.O.Ö., A.B., Ö.A., N.K., H.B., Analysis or Interpretation: B.O.Ö., A.B., Ö.A., N.K., Literature Search: B.O.Ö., Ö.A., N.K., B.D., Writing: B.O.Ö., H.B.
Conflict of Interest: No conflict of interest was declared by the authors.
Financial Disclosure: The authors declared that this study received no financial support.
