ABSTRACT
Candiduria is common in hospitalized patients, and asymptomatic candiduria contributes to antifungal overuse. The guidelines for management of asymptomatic candiduria do not recommend antifungal use, but rather the elimination of predisposing factors. It is unknown whether these recommendations are being followed. The primary objective of this study was to characterize candiduria management among hospitalized patients. This was a retrospective cohort study of a random sample of 305 hospitalized patients with candiduria at four U.S. medical centers from January 2010 to December 2013. Patients were classified as asymptomatic or symptomatic based on established criteria, and data were collected by chart review. Infectious Diseases Society of America (IDSA) treatment guideline adherence and its association with clinical outcomes, including candiduria recurrence (short- and long-term) and 30-day readmission, were assessed. Eighty percent of patients were classified as having asymptomatic candiduria. Overall, 143 (47%) patients were not managed according to recommended guidelines, including 105/243 (43%) in the asymptomatic candiduria group and 38/62 (61%) in the symptomatic group (P = 0.01). Discordance among asymptomatic patients was driven by overtreatment with an antifungal (98/105 [93%]). Thirty-three percent of patients with asymptomatic candiduria not managed according to the guidelines were treated for over 7 days, and 5% received over 14 days of therapy. Fluconazole was the most commonly used empirical antifungal among asymptomatic candiduria patients (96%), followed by micafungin (4%). Asymptomatic candiduria patients not managed according to the guidelines had a trend toward higher 30-day readmission (35% versus 26%, P = 0.27). Inappropriate management of candiduria among hospitalized patients was high, leading to overtreatment with antifungal therapy.
INTRODUCTION
Candida is a common finding in the urine of hospitalized patients and represents a therapeutic challenge (1, 2). It can be difficult for clinicians to determine when treatment is warranted given the range of possible clinical scenarios. Most patients with candiduria are symptomless and have no features to suggest urinary tract infection (UTI) (2). Urine cultures may be ordered due to unexplained fever, leukocytosis, or foul-smelling urine, but recent data show that urine is frequently cultured without a clear indication (3). An increase in ordering urine cultures can lead to the detection of asymptomatic candiduria and unnecessary therapy.
A prospective study of the management of Candida UTIs showed that asymptomatic candiduria almost always represents colonization and that eliminating underlying risk factors is sufficient management (4). Treatment with antifungal therapy in asymptomatic candiduria is recommended only for high-risk adult patients, such as those with neutropenia or who are undergoing urologic manipulation (5). Despite clinical practice guidelines discouraging the treatment of asymptomatic candiduria, many patients still receive antifungal therapy even when not indicated. Inappropriate treatment increases the risk of selecting for antifungal-resistant organisms, drug-drug interactions, and adverse drug events (6–9). Thus, eliminating unnecessary antifungal use is critical.
Candiduria must be accurately diagnosed and managed to minimize overutilization of antifungals. The first step is to understand whether patients are being managed according to the Infectious Diseases Society of America (IDSA) guidelines for the management of candidiasis (5). The primary objective of this study was to characterize the management of candiduria among hospitalized patients and determine whether patients are managed in accordance with IDSA guidelines. The secondary objectives were to evaluate clinical and microbiological outcomes, including (i) candiduria recurrence rates (short- and long-term) and (ii) 30-day hospital readmission rates. We hypothesized that hospitalized asymptomatic adult patients frequently receive antifungal therapy for candiduria irrespective of current IDSA guideline recommendations.
RESULTS
Patient characteristics.Three hundred five patients met the inclusion criteria during the study period. The average age of patients with candiduria was 69 ± 15 years, and the majority were female (67%) (Table 1). A majority of patients were admitted from home (56%), and the primary patient category was general medicine (88%). The most common comorbidities were hypertension (66%), diabetes (43%), and renal disease (32%), and the average Charlson comorbidity index was 4.9 ± 2.8. Most positive urine cultures represented asymptomatic candiduria (80%). A urinary catheter was present in 197 (65%) patients and was replaced after index culture in 67 of 197 (34%) patients.
Demographics and other characteristics of the candiduria cohort (n = 305)
Microbiology.Candida albicans was isolated in 47% of asymptomatic candiduria patients (115/243) and 24% of symptomatic candiduria patients (24/62, P < 0.01) (Table 2). Other organisms isolated from asymptomatic candiduria patients included: C. glabrata (19%) and other non-albicans Candida (12%). Symptomatic candiduria patients showed a similar distribution, including C. glabrata (24%) and other non-albicans Candida (10%). A concomitant bacterial infection was common among asymptomatic (71%) and symptomatic (74%) patients.
Yeast isolates and adherence to IDSA guidelines stratified by presence of candiduria symptomsa
Management according to IDSA guidelines and antifungal treatment characteristics.Overall, 143 of 305 (47%) patients were not managed according to IDSA guidelines, including 105 of 243 (43%) in the asymptomatic group and 38 of 62 (61%) in the symptomatic group (P = 0.01) (Table 2). The reasons for guideline nonadherence differed between asymptomatic and symptomatic candiduria patients. The major reason for nonadherence in asymptomatic candiduria was no indication for antifungal therapy (98/105 [93%]). Among symptomatic candiduria patients, reasons for nonadherence to the guidelines were inappropriate choice of antifungal agent (5/35 [14%]), inappropriate antifungal dose (14/35 [40%]), and therapy being indicated but without administration of an antifungal agent (16/35 [46%]).
One hundred forty-five (48%) patients received empirical antifungal treatment for candiduria, including 107 of 243 (44%) asymptomatic patients and 38 of 62 (61%) symptomatic patients (Table 3). Fluconazole was the most frequently used antifungal in both the asymptomatic and symptomatic groups (139/145 [96%]), with micafungin used in 6 patients (4%). No patients received either amphotericin B (AmB) deoxycholate or flucytosine. Among asymptomatic candiduria patients, 100 (93%) were not managed in accordance with the guidelines, and 97% received fluconazole therapy when it was not indicated. Among patients who received antifungal therapy, the median durations of therapy for guideline nonadherent asymptomatic and symptomatic candiduria patients were 4 (interquartile range [IQR], 1 to 7) and 2.5 (IQR, 1 to 5) days, respectively. Additionally, 33% of asymptomatic patients received greater than 7 days of antifungal therapy when treatment was not indicated. For symptomatic patients, the median durations of therapy among those treated according to the guidelines were 5 (IQR, 3 to 7) days, and 2 of 22 (9%) patients were treated for >14 days.
Antifungal treatment and patient outcomes between guidelines adherent and nonadherent patients by presence of candiduria symptomsa
Patient and clinical factors associated with nonadherence to IDSA guidelines.Nonadherence to IDSA guidelines for the management of candiduria was more common in older patients (median, 74; [IQR, 61 to 83]; P < 0.01), patients admitted from a nursing home (29%, P < 0.01), and those with a higher Charlson comorbidity index (6 versus 4, P < 0.01) (see Table S1 in the supplemental material). Other factors associated with guideline nonadherence included the presence of symptomatic candiduria (27% versus 15%, P < 0.01), receipt of antibiotics in previous 30 days (80% versus 68%, P = 0.02), and a concomitant bacterial infection at time of index culture (80% versus 64%, P < 0.01). The presence of a urinary catheter at the time of index culture and the urinary catheter being replaced after index culture were also more common in the nonadherent group (P ≤ 0.02 for both). In the final multivariable model, a higher Charlson comorbidity index (adjusted odds ratio [aOR], 1.25; 95% confidence interval [CI], 1.14 to 1.38), presence of symptomatic candiduria (aOR, 3.11; 95% CI, 1.67 to 5.81), a urinary catheter at the time of index culture (aOR, 1.73; 95% CI, 1.03 to 2.90), and a concomitant bacterial infection (aOR, 2.09; 95% CI, 1.20 to 3.63) were independently associated with guideline nonadherence for the management of candiduria (Table S2).
Patient outcomes.Among asymptomatic patients not managed according to the guidelines, there was a nonsignificant trend toward higher overall candiduria recurrence (60% versus 45%, P = 0.34) and 30-day readmission (35% versus 26%, P = 0.27) rates (Table 3). No significant differences in outcomes were observed among patients with symptomatic candiduria based upon guideline adherence. However, the 30-day readmission rate was lower, though not statistically significant, among guideline-adherent than in guideline nonadherent patients (13% versus 52%, respectively; P = 0.09).
DISCUSSION
Nearly 80% of candiduria patients were asymptomatic in this evaluation of 305 cases of candiduria within four medical centers, and 43% of patients with asymptomatic candiduria were not managed according to current IDSA guidelines. Furthermore, symptomatic Candida UTIs were treated primarily with fluconazole at inappropriate doses and for shorter durations than recommended by the IDSA. No patients were treated with AmB deoxycholate or flucytosine in the presence of non-albicans Candida spp.; and in some cases, micafungin was used for treatment.
Currently, the recommended management of asymptomatic candiduria is removal of any predisposing risk factors, and antifungal agents are not warranted unless the patient is at high risk for dissemination (5). We found that nearly half of asymptomatic patients were treated with an antifungal agent, typically fluconazole. Radosevich and colleagues performed a single-center retrospective review of 132 candiduria patients and found that 65% were treated with an antifungal agent, with the majority started on fluconazole (10). Chen et al. had similar observations, with 33% of patients receiving treatment for candiduria and 78% initiated on fluconazole (11). Although these studies did not evaluate whether management and treatment were in accordance with recommended guidelines, they both indicated a small percentage of patients with documented signs and symptoms of a Candida UTI. Our findings are similar in that 48% of patients were treated with an antifungal agent, but we found that only 20% of antifungal-treated patients were managed according to the guidelines.
Micafungin was used as an alternative to fluconazole in a small subset of patients rather than guideline-recommended alternatives of AmB deoxycholate or flucytosine, which were not used at all. Treatment with micafungin for a Candida UTI is surprising, given its extensive protein binding and negligible urinary excretion of unchanged drug (12). There are numerous published reports of both the success and failure of echinocandin use for the treatment of Candida UTIs (13–16). It is not currently clear whether echinocandins are a viable treatment option for candiduria, and further research is necessary to elucidate their role.
We found higher Charlson comorbidity index scores, symptomatic candiduria, urinary catheterization at the time of index culture, and concomitant bacterial infection to be independently associated with guideline-discordant decision-making. An increasing Charlson score is not unexpected as a predictor of nonadherence, as clinicians may have less apprehension and therefore a greater likelihood to add on therapy to sicker patients. This may also be the reason that patients with concomitant bacterial infections were treated for candiduria when it was not indicated. Although it has been documented that the addition of an antifungal for the treatment of candiduria may lead to superior short-term clearance, treatment is not associated with long-term clinical benefit (4, 17, 18). Additionally, unnecessary antifungal exposure can lead to antifungal resistance. The majority of patients in our study received fluconazole, many at subtherapeutic doses. While the emergence of triazole-resistant C. albicans is rare, triazole exposure can select for triazole-resistant Candida spp., such as C. glabrata and C. krusei (19, 20). For patients with symptomatic candiduria, the recommended management is antifungal therapy based on susceptibility testing (5). We found that the mismanagement of symptomatic candiduria was primarily driven by a lack of treatment initiation. Furthermore, a urinary catheter as a predictor of mismanagement may simply be due to a disproportionately larger number of asymptomatic candiduria patients with a urinary catheter at time of index culture.
These findings have important implications for growing antifungal stewardship (AFS) efforts (21, 22). Many of the reasons for guidelines discordance, such as overtreatment and inappropriate choice of drug or dose, are common antimicrobial stewardship program (ASP) interventions. Several reports have shown the effectiveness of ASP interventions in the management of candidemia (23, 24). Given the high mortality and poor outcomes associated with invasive fungal disease, this seems to be a logical first step. Inappropriate management of UTIs is not specific to yeast, as similar trends exist for bacterial UTIs (25). Although investigators have reported stewardship interventions focusing on the management of asymptomatic bacteriuria, the same is not true for candiduria (26, 27). Our data show that there is a need for stewardship programs that focus on Candida UTIs given the high percentage of inappropriate management decisions and increasing risk of resistance.
This study has several limitations. First, due to its retrospective design, it is possible that clinicians did not document the symptoms of Candida UTIs, leading to possible misclassification. A previous prospective surveillance study of candiduria showed that only 2% to 4% of all patients have signs and symptoms associated with candiduria (4). Other studies of candiduria have reported similar rates, so the extent of this bias may be attenuated (10, 11). Also, our study definition of symptomatic candiduria included fever, which may have led to a higher percentage of patients being identified as symptomatic. Second, the management of candiduria may vary between institutions and be dependent on hospital-specific policies or antimicrobial stewardship interventions. In order to enhance the generalizability of our findings, we included four different hospitals with similar numbers of beds in different regions of the United States. All hospitals had an established antimicrobial stewardship program, and fluconazole use was not restricted at any hospital. Also, the policies regarding the identification of yeast in the urine may differ between laboratories. Given our retrospective design, no further work-up was possible for these cultures. Third, over half of the fungal cultures were identified without the Candida species, thus underestimating the true prevalence of C. albicans and non-albicans Candida. Also, it is possible that antifungal therapy was administered to patients on broad-spectrum antibiotics to prevent the emergence of invasive fungal infection (selective gut decontamination). However, this was not documented in any chart, and typically, enteral nonabsorbable antifungal preparations are employed for this purpose. Finally, we did not systematically collect the ordering prescriber's fluconazole use, although we did exclude patients with any evidence of extraurinary fungal infection.
In conclusion, approximately half of the candiduria patients in our study were not managed in accordance with guidelines and were commonly overtreated with fluconazole. Recommended therapeutic alternatives, such as AmB deoxycholate and flucytosine, were not utilized, with patients instead being treated with an echinocandin. Multiple variables were independently associated with nonadherence to IDSA guidelines, and it is possible that these can be rectified within an antifungal stewardship program.
MATERIALS AND METHODS
Study design and patient population.This was a multicenter retrospective cohort study of adults with candiduria who were hospitalized at four general hospitals in the United States: Eastern Maine Medical Center (Bangor, ME), Wheaton Franciscan Healthcare-St. Francis Hospital (Milwaukee, WI), Erie County Medical Center (Buffalo, NY), and Baylor-St. Luke's Medical Center Hospital (Houston, TX). The study period was January 2010 to December 2013, and study data were collected and managed using REDCap electronic data capture tools hosted at the University at Buffalo (28). The institutional review board at the University at Buffalo and at each participating institution approved the study.
Patients were identified using each institution's microbiology database, and each site included at least 60 patients. Patients were included in the study if they were ≥18 years of age with at least one urine culture yielding ≥103 yeast colonies/ml, an inoculum cutoff used in two previous studies of candiduria (4, 17). The first urine culture yielding Candida spp. was considered the index culture, and patients were eligible to enter the study only once. Patients were excluded if they had any evidence of extraurinary fungal infection, a diagnosis of vaginitis as per the medical chart, documentation of a spinal cord injury, and/or had received systemic antifungal treatment within the preceding 7 days.
Data collection.Data collected on standardized case report forms included demographic information (age, sex, race, and past medical history); patient location at the time of candiduria (intensive care unit [ICU] or non-ICU); patient category (general medicine, surgical hematologic malignancy, solid organ transplant, or solid tumor); corticosteroid use (in past 30 days); Charlson comorbidity index (at time of index culture); presence of risk factors, including antibacterial agent treatment within the preceding 30 days; and recent surgery (≤30 days) and hospitalization (≤30 days) (29). Microbiological data, including Candida species, if available, and concomitant bacterial infection during the candiduria episode were also collected. Candiduria treatments, including antifungal agent, dose, and duration, as well as the management (removal or not) of an indwelling urinary catheter, if present, were also documented.
Definitions and outcomes.All definitions were selected prospectively during the initial study design. Patients were categorized as either asymptomatic or symptomatic candiduria, as this directs UTI management due to Candida species. Asymptomatic candiduria was defined as a positive urine culture with ≥103 yeast colonies/ml in the absence of dysuria, polyuria, flank pain, and/or fever (temperature of 37.8°C [100°F]) (30). Asymptomatic candiduria was inclusive of patients that may have been minimally symptomatic at time of the index urine culture and for whom symptoms may not have been detected due to the retrospective nature of this study. Symptomatic candiduria was defined as a positive urine culture with ≥103 yeast colonies/ml and at least one documented symptom mentioned above without an alternative etiology (17). While fever is objectively measured in the medical record, other UTI signs and symptoms are less objectively documented. Medical notes were extensively queried for mentions of dysuria, polyuria, and/or flank pain to determine patients with symptomatic candiduria.
The IDSA guidelines on the management of candidiasis from 2009 were followed to evaluate whether patients with Candida UTIs were managed according to the guidelines (31). These guidelines were updated in 2016 but did not contain appreciable changes with regard to the management of candiduria (5). For asymptomatic candiduria, the IDSA guidelines recommend eliminating predisposing factors, such as indwelling catheters, whenever feasible. Treatment with antifungal agents is not recommended unless the patient belongs to a group at high risk for dissemination (e.g., neutropenia). For symptomatic candiduria, the recommended treatment is fluconazole 3 mg/kg of body weight daily or 200 mg for 2 weeks in patients with fluconazole-susceptible Candida; for those with triazole-resistant Candida spp., the recommended treatment is either amphotericin B (AmB) deoxycholate or flucytosine for 7 to 10 days.
For each patient, the empirical management of candiduria was noted and categorized as guideline adherent (if the initial management aligned with IDSA-recommended management) or guideline nonadherent (if the initial management did not align with IDSA-recommended management). For patients categorized as guideline nonadherent, the reason for nonadherence was recorded as: (i) therapy not indicated, (ii) therapy indicated but antifungal agent was not administered, (iii) inappropriate choice of drug, and/or (iv) inappropriate drug dose. The creatinine clearance at the time of treatment was taken into account when evaluating the appropriateness of treatment dose. Outcomes recorded were short- and long-term candiduria recurrence and 30-day hospital readmission. Short-term candiduria recurrence was defined as a positive urine culture with ≥103 yeast colonies/ml in less than 1 month following discontinuation of therapy for the index candiduria. Candiduria recurrence not meeting the definition of short-term was classified as long-term recurrence.
Statistical analysis.Descriptive characteristics were summarized using means and standard deviations (SD) or median and interquartile ranges (IQR) for continuous variables and percentages for categorical variables. Clinical and demographic characteristics were compared between the two groups (guideline adherent and guideline nonadherent) using the Wilcoxon rank-sum, Pearson chi-square, and Fisher's exact tests, as appropriate. A logistic regression model was used to explore patient and clinical characteristics associated with guideline nonadherence for candiduria management. Factors with a P value of <0.2 in the bivariate analysis were included in the multivariable logistic model. A backward elimination strategy was employed, with variables at a P value of <0.05 kept in the model. All tests were 2-tailed, and a P value of <0.05 was considered to be statistically significant. SAS version 9.4 (SAS Institute, Cary, NC) was used for all analyses.
ACKNOWLEDGMENTS
The research reported in this publication was supported by the National Center for Advancing Translational Sciences of the National Institutes of Health under award UL1TR001412. This study was supported by an investigator-initiated research grant from Astellas, Inc.
The content of this paper is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
We thank Terry Mashtare for his assistance with the preliminary analysis.
D.M.J. has received research support from Astellas, Inc.; N.D.B. has received research support and is an advisory board member for Astellas, Inc.; and T.J.D., A.M.C., and D.R.B. disclose no conflicts of interest.
FOOTNOTES
- Received 18 July 2017.
- Returned for modification 28 August 2017.
- Accepted 29 October 2017.
- Accepted manuscript posted online 6 November 2017.
Supplemental material for this article may be found at https://doi.org/10.1128/AAC.01464-17.
- Copyright © 2017 American Society for Microbiology.
