Introduction

Candida infections tend to emerge in immune-compromised hosts with dosing such as steroids, immune suppressants or chemotherapy.1, 2, 3 However, although immune-compromised status is a risk factor for Candida infection, some patients with normal immune states may have a Candida infection.4, 5 Bloodstream infections (BSIs) with Candida species become easily fatal not only in immune-compromised hosts but also in those with normal immune status. The number of such infections is increasing, especially in intensive care units (ICUs) patients.6, 7, 8 The clinical risk factors for BSIs with Candida need to be understood for the prevention of mortality associated with Candida BSIs.

Candida spp. are basically classified into Candida albicans and Candida non-albicans. Krcmery and Barnes9 reported that Candida non-albicans may cause more severe disease and higher mortality especially in BSI cases. The significant risk factors for severe Candida BSIs from Candida albicans and Candida non-albicans need to be differentiated.

In this study, we investigated the characteristics of BSIs with Candida spp. in our consecutive cases over a 5-year period at Kobe University Hospital, Kobe, Japan, and discriminated between the risk factors for Candida albicans spp. and Candida non-albicans BSIs especially focusing on investigation of Candida non-albicans spp., which in general tend to show more resistance to antifungal agents. We also examined the resistance to antifungal agents such as amphotericin B (AMPH), miconazole (MCZ), micafungin (MCFG), flucytosin (5-FC), fluconazole (FLCZ), itraconazole (ITCZ) and voriconazole (VRCZ).

Materials and methods

Patients

Candida species (spp.) were isolated from patients with BSIs from January 2008 to June 2013 in Kobe University Hospital (920 beds). We investigated the comparative clinical risk factors for Candida BSIs between Candida albicans and Candida non-albicans spp. and the susceptibilities of isolated fungi to representative antifungal agents. BSI was defined as symptomatic fungus isolation of one or more in 20 ml blood. We recorded the patient’s gender, age and clinical risk factors for Candida albicans or Candida non-albicans spp. causing BSI. Candida albicans BSI and Candida non-albicans BSI were defined as isolation of any species of Candida albicans or Candida non-albicans from at least one blood culture. In patients with more than one positive blood culture, only the first BSI was included. The detection of Candida spp. was performed using CHROMagar Candida (Japan BD Co. Ltd, Tokyo, Japan) and if necessary RapID Yeast plus kit (AMCO Incorporated, Tokyo, Japan). Risk factors included underlying cardiovascular diseases, liver–biliary–pancreatic disease, cancer-bearing, diabetes mellitus, steroid dosing and chronic renal failure. Antifungal agents tested were AMPH, MCZ, MCFG, 5-FC, FLCZ, ITCZ and VRCZ. The study was retrospectively performed with methods based on the Helsinki Declaration.

Susceptibility testing

Antifungal agent susceptibilities were tested and the results were interpreted and reported using the reference broth microdilution method as described by the Clinical and Laboratory Standards Institute (CLSI) M27-S3 (2010 CLSI Document M100-S20). The minimal inhibitory concentration was defined as the lowest antimicrobial concentration that totally inhibited bacterial growth. Susceptibilities were evaluated by CLSI category. We tested Candida strains against AMPH, MCZ, MCFG, 5-FC, FLCZ, ITCZ and VRCZ antifungal agents using drug susceptibility test kits for yeast-like fungus (ASTY) (Kyokuto Pharmaceutical Industrial Co. Ltd, Takahagi, Japan) after 48 h of culture. For quality control, Candida parapsilosis ATCC22019 or Candida Krusei ATCC6258 were used.

Statistical analyses

Statistical analysis was conducted using Fisher’s exact test, Student’s t-test, and univariate and multivariate logistic regression analysis with PASW Statistics 17.0 software packages (for Windows; SPSS Inc., Chicago, IL, USA). The latter two analyses were performed in order to seek for the odds ratio for comparison of extent of relationship between significant risk factors and Candida non-albicans BSI. Statistical significance was established at the 0.05 level.

Results

Patient characteristics

There were 61 patients with Candida albicans BSIs and 49 patients with Candida non- albicans BSIs. The Candida non-albicans spp. isolated included 24 (49%) Candida parapsilosis, 14 (29%) Candida glabrata, 6 (12%) Candida tropicalis, 2 (4%) Candida guilliermondii, 1 (2%) Candida stellatoidea, 1 (2%) Candida krusei and 1 (2%) other Candida spp. Patients included 74 males and 36 females, and 62 patients were age >70. Cardiovascular diseases, liver–biliary–pancreatic disease, cancer, diabetes mellitus, steroid dosing, chronic renal failure, postoperative status, death, chemotherapy and radiation therapy were noted in 23, 24, 40, 20, 8, 16, 49, 18, 12 and 2 cases, respectively. The susceptibility ratios to antifungal agents were 100, 100, 100, 97.3, 82.7, 40.0 and 94.5% for AMPH, MCZ, MCFG, 5-FC, FLCZ, ITCZ and VRCZ, respectively (Table 1).

Table 1 Characteristics of Candida albicans and Candida non-albicans in bloodstream infection cases
Full size table

Comparison between Candida albicans BSIs and Candida non-albicans BSIs

The comparison data between Candida albicans and Candida non-albicans spp. showed that BSI patients with Candida albicans had older age and a higher ratio of underlying cardiovascular diseases (P=0.018) and postoperative status (P=0.009) and those with Candida non-albicans spp. had a higher ratio of cancer (P=0.040) and chemotherapy (P=0.035). Resistance to antifungal agents was higher in FLCZ (P=0.003) and ITCZ (P<0.001) in Candida non-albicans spp. than Candida albicans (Table 1).

Risk Factors for Candida non-albicans BSIs compared with Candida albicans BSIs

Univariate analyses showed that Candida non-albicans species had significantly higher rate in patients with cancer (odds ratio (OR) (95% confidence interval (CI))=2.29 (1.04–5.06) and P=0.040), chemotherapy (OR=4.35 (1.11–17.1) and P=0.035), FLCZ resistance (OR=77.3 (4.51–1324) and P=0.003) and ITCZ resistance (OR=15.6 (5.39–45.1) and P<0.001) and a lower rate of underlying cardiovascular diseases (OR=0.27 (0.09–0.80) and P=0.018) and postoperative status (OR=0.35 (0.16–0.77) and P=0.035) than Candida albicans. Multivariate analyses demonstrated that Candida non-albicans spp. had significantly higher rates of chemotherapy (OR=4.44 (1.04–19.0) and P=0.045), FLCZ resistance (OR=5.87 (2.01–17.1) and P=0.001) and ITCZ resistance (OR=18.7 (5.77–60.4) and P<0.001) and a lower rate of underlying cardiovascular diseases (OR=0.25 (0.08–0.82) and P=0.022) than Candida albicans (Table 2).

Table 2 Results of univariate and multivariate logistic regression analyses of risk factors associated with Candida non-albicans BSIs
Full size table

Discussion

One of the main findings of this study was a significantly higher rate of Candida non-albicans species in BSI patients with cancer, chemotherapy, FLCZ resistance and ITCZ resistance, and lower rate of underlying cardiovascular diseases and postoperative status than Candida albicans. Multivariate analyses demonstrated that Candida non-albicans spp. had significantly higher rate of chemotherapy, FLCZ resistance and ITCZ resistance, and lower rate of underlying cardiovascular diseases than Candida albicans. Regarding the risk factors for Candida BSIs, Ma et al.10 showed that complicated abdominal surgery, presence of central venous catheter, neutropenia and Candida tropicalis were seen at significantly higher rates and treatment with FLCZ were fewer in the mortality group than the survival group in their comparison analyses, and central venous catheter was an independent risk factor for a 30-day Candida BSI mortality by multivariate analyses. On the other hand, increases in common risk factors such as ICU admission,11, 12 or immunosuppressive therapies13 may have resulted in an increase in the overall pool of patients at high risk for candidemia. Our suggested significant risk factors for Candida non-albicans BSIs demonstrated a significantly higher rate of chemotherapy than Candida albicans in multivariate analyses, which agrees with these previous works.

Regarding antifungal therapy, Ha et al.14 reported that treatment failure of first-line antifungal drugs was associated with 30-day mortality and salvage therapy using VRCZ for the treatment of candidemia in patients intolerant of or refractory to other antifungal agents was reported to be successful.15 VRCZ may be a suitable agent for salvage therapy for invasive candidiasis, even in the setting of previous azole exposure and Candida krusei infection.16, 17 Our current data showed that FLCZ and ITCZ were independent risk factors for Candida non-albicans BSIs. Taken together, these studies suggest it may be necessary to predict the causative Candida (Candida albicans or Candida non-albicans) based on the suggested risk factors in order to avoid treatment failure.

Chow et al.18 showed that ICU patients had a major pre-ICU operation, total parenteral nutrition duration, hemodialysis duration, mean number of red blood cell (RBC) transfusions per day, gastrointestinal procedure, and enteric bacteremia were risk factors for Candida non-albicans spp. BSIs, with total parenteral nutrition duration per days at risk, hemodialysis duration per days at risk, major operation during ICU stay and enteric bacteremia were risk factors for Candida non-albicans spp. BSIs in their multivariate analyses. While the differences between our study comparing Candida albicans spp. and Candida non-albicans BSIs and the previous study make direct comparison difficult, our data on risk factors for Candida BSIs are useful for the understanding and predicting the outcome of these infections.

From a mycological aspect, clinically important Candida spp. are subgrouped into albicans, grabrata, parapsilosis, tropicalis, guilliermondii or krusei and can be generally categorized as Candida albicans and Candida non-albicans spp. for investigative purposes.19 The epidemiology of Candida BSIs appears to be changing and especially the proportion of Candida non-albicans BSIs has been increasing.20, 21, 22, 23, 24, 25 We designed this study to discriminate between Candida albicans spp. and Candida non-albicans, and demonstrated that underlying cardiovascular diseases and postoperative status were significant risk factors for Candida albicans BSI and cancer and chemotherapy for Candida non-albicans spp. BSI. The information of the difference of the clinical risk factors between Candida albicans and Candida non-albicans BSIs may be useful to the physicians who take care of Candida patients because the susceptibilities to antifungal agents were different between them. That is, Candida non-albicans spp. BSIs showed a higher resistance ratio to FLCZ and ITCZ than Candida albicans BSIs, taken together with the risk factor data mentioned above, has implications for first-line treatments of Candidal BSI.

We would like to emphasize the limitations of this study. First, the number of cases was not enough for definitive conclusions although they do represent all the consecutive cases in our institution, Kobe University Hospital, over the last 5 years. Second, we lacked a control group for comparison, unlike previous studies.18, 19 Third, the potential risk factors we investigated may not be comprehensive and did not include some risk factors assessed in other studies. These limitations will be addressed in our future work.

In conclusion, this study differentiated several risk factors for BSI with Candida albicans (underlying cardiovascular diseases and postoperative status) and Candida non-albicans spp. (cancer and chemotherapy), and demonstrated that Candida non-albicans spp. were more resistant to FLCZ and ITCZ than Candida albicans. These findings might be affected by Candida epidemiology in our University Hospital.