The effect of current management on morbidityand mortality in hospitalised adults with funguria

Christine Simpsona, Sandra Blitzb, Stephen D. Shafranc,*

aDepartment of Medicine, University of Alberta, 2F1.13 Walter C. Mackenzie Health Sciences Centre, 8440-112 Street, Edmonton, Alta., Canada T6G 2R7bData Management, Breast Cancer International Research Group, Suite 1100, 9925 - 109t Street,Edmonton, Alta., Canada T5K 2J8cDivision of Infectious Diseases, University of Alberta, 2E4.16 Walter C. Mackenzie Health Sciences Centre,8440-112 Street, Edmonton, Alta., Canada T6G 2B7

Accepted 15 August 2003

Available online 22 September 2003

KEYWORDSFunguria; Asymptomatic

funguria; Antifungal

treatment; Risk

reduction; Invasive

fungal infection

Summary Background. To compare morbidity and mortality in inpatients withasymptomatic funguria between those treated and those observed for funguria.Methods. Retrospective analyses were performed in 149 consecutive adult tertiary

care inpatients with asymptomatic funguria. The primary endpoints were death,length of hospitalisation and progression to invasive fungal infection (IFI).Results. Of the 149 subjects, 70% were female, 55% were .65 years, recent

antibiotic and urinary catheter use occurred in .70%, diabetes in 32%, recent ICUadmission in 29%, and concomitant bacteriuria in 28%. Forty-seven percent did notreceive active intervention. Of the remainder, 46% were managed by controlling oreliminating risk factors for funguria or progression to IFI; fluconazole or amphotericin Bwere used to treat the other 54%. Fourteen percent died and 2.7 % progressed to IFI,with no significant difference between the treated versus observed groups for eitherendpoint ðp . 0:2Þ: Median length of hospitalisation was significantly greater in thetreated group ðp , 0:01Þ;multivariate analysis demonstrated an exclusive relationshipto the greater number of risk factors present in the treated group.Conclusion. Asymptomatic funguric patients who were managed with risk reduction

and/or antifungal therapy were older, had more risk factors for funguria andsubsequent progression to IFI, and had a longer hospital admission than those managedwith observation. Treatment of asymptomatic funguria with risk reduction and/orantifungal therapy did not impact adult inpatient morbidity or mortality in this review;rather, the presence of multiple risk factors for funguria or IFI appeared to serve as a‘sickness indicator’.Summary. In this study, we found that treatment for asymptomatic funguria in

hospitalised adults did not impact morbidity or mortality. Rather, the presence ofmultiple risk factors for funguria or IFI correlated with a longer duration ofhospitalisation, suggesting that funguria may be a ‘sickness indicator’, similar tobacteriuria in the elderly.Q 2003 The British Infection Society. Published by Elsevier Ltd. All rights reserved.

Journal of Infection (2004) 49, 248–252

www.elsevierhealth.com/journals/jinf

0163-4453/$30.00 Q 2003 The British Infection Society. Published by Elsevier Ltd. All rights reserved.doi:10.1016/j.jinf.2003.08.008

*Corresponding author. Tel.: þ1-780-407-8077; fax: þ1-780-407-7137.E-mail address: sshafran@uah.ualberta.ca

Introduction

The yeast-growing urine specimen remains aclinical dilemma for the clinician. Five to 10% ofall urinary isolates among hospital inpatients areidentified as ‘yeast’.1 Several studies document themost efficient and least deleterious means forsterilizing the urine of funguric patients, but thereis currently insufficient evidence to indicatewhether or not asymptomatic funguria (absence ofdysuria, frequency, urgency, incontinence andsuprapubic pain) should be treated.1,2 Funguriacan be an indication of contamination duringsample collection or of colonization of the urinarytract and/or intrabladder devices. Yet, 17–20% ofpatients with funguria die in hospital.3 Invasivefungal infection (IFI) develops in 1.3 –10% ofpatients who initially present with funguria, and0.4% of these die.4 Outcomes such as these have ledphysicians to actively treat patients with riskreduction and/or antifungal therapy, although nostudies have demonstrated a clinical benefit orsurvival advantage.

Several algorithms have been proposed to aid inthe decision to use risk reduction and/or antifungaltherapy for asymptomatic funguria. Fisher et al.5

suggest that patients with funguria be stratified intothree main groups based on clinical characteristics:the previously healthy and asymptomatic; high-riskpatients unlikely to develop IFI; and those highlylikely to develop IFI. They propose that antifungaltherapy should be used in the latter two groups,based on ‘clinical discretion’.5 Others havesuggested that only symptomatic patients betreated.6 Duma7 states that before any clinicaltreatment is considered, a repeat clean-catch urinespecimen should be submitted to the laboratory,and the same strain of fungus isolated beforeconsidering the funguria to be ‘significant’. IFIshould then be excluded, and all predisposing riskfactors for funguria eliminated. If the funguriapersists for greater than one week, or it isimpossible to eliminate the primary predisposingrisk factors, antifungal therapy should then beinstituted.7 In favour of observation, spontaneousresolution of funguria occurs in 28–35% of patientswhen intrabladder drainage devices are removed.3,8

Other studies suggest that 40–62% of repeat urinespecimens eventually resolve without interven-tion.9,10 Clearly, a consensus regarding whichfunguric patients should be treated is lacking.

Clinical evidence indicates that risk factors forfunguria and/or IFI include any immunodeficient/suppressed state, neutropaenia, recent use ofantibiotics or corticosteroids, presence of intra-

venous or intrabladder catheters, recent aggressivesurgery, hyperalimentation, transplantation, dia-betes mellitus (DM), pregnancy, and urinary tractdysfunction or anatomical distortion.7 In a study offactors predisposing to funguria, DM was shown tobe present in 39%, urinary tract abnormalities in38%, concomitant urinary tract bacterial infectionsin 85%, intrabladder drainage devices in 83%, and90% had recently been treated with antimicrobialmedications.3 DM was also associated with non-clearance of the urine unless antifungal treatmentwas administered.3

The purpose of this study is to investigate theincidence and characteristics of funguria in adulthospital inpatients in a tertiary care facility, andto evaluate the hypothesis that treatment ofasymptomatic funguria has a positive morbidityand/or mortality benefit for such patients.

Methods

The University of Alberta Hospital (UAH) isEdmonton’s largest academic tertiary care centre.All adult hospital inpatients whose urine sampleswere identified by the microbiology laboratory asgrowing yeast in concentrations exceeding 106 cfu/lby the Bac-Tube method during a 7-month period(mid-August 1999–mid-March 2000) were assessedprospectively for study inclusion. Subjects wereexcluded if there was evidence of IFI at the time oftheir first yeast-positive urine, or if they had grownyeast in their urine at the UAH within the previoustwelve months. A thorough chart review andmedical record computer search was performed atthe time of discharge, so as not to influencemanagement. Information regarding patient demo-graphics, risk factors for funguria and IFI, symptomssuggestive of urinary tract infection, follow-upinvestigations, and means of controlling and/ortreating the infection was obtained and recorded.Primary endpoints included length of hospitalis-ation, death and progression to IFI.

All data were entered into a database, and wereanalyzed using SASe. Subjects were split into twocomparative groups: those whose funguria wasuntreated, and those whose funguria was treatedwith local or systemic antifungal therapy and/orrisk factor reduction/elimination. Differences inpatient demographics and specific risk factorsbetween the two groups are outlined in Tables 1and 2.

Outcomes between the treated and untreatedgroups were assessed using a chi-square test for thecategorical variables and a Wilcoxon sign-rank test

Effect of funguria management 249

for the continuous variables. The multivariateanalysis of the length of hospitalisation wasassessed by least squares regression (the dependentvariable was transformed by taking the square-rootin order to adjust for its skewness). These analyses

were replicated using Cox proportional hazardsmodels to evaluate the consistency of the model.No data points were censored.

The study was approved by the Research EthicsBoard of the University of Alberta Hospital.

Table 1 Demographics

Variable Treated (79)a nb (range) Observed (70)c n (range) All (149) n (range) p-value

Age 69 (19, 94) 62 (18, 90) 67 (18, 94) 0.04GenderMale 22 (28%) 22 (31%) 44 (30%) NSFemale 57 (72%) 48 (69%) 105 (70%)No. hospital daysd 29 (3, 250) 17 (1, 249) 22 (1, 250) ,0.01No. enrolled dayse 16 (0, 207) 8.5 (0, 235) 12 (0, 235) ,0.01No. ICU daysf 16 (1, 91) 7 (1, 100) 13 (1, 100) 0.05No. risk factorsg 4 (1, 8) 3 (1, 8) 3 (1, 8) ,0.01No. assessmentsh 2 (0, 4) 0 (0, 5) 1 (0, 5) ,0.01

a Sub-group of patients treated by controlling and/or removing identified risk factors for funguria or progression to invasive fungalinfection, and/or by local systemic antifungal therapy.

b All values are median averages.c Sub-group of patients observed only after funguria.d Total number of days from admission to discharge.e Total number of days from funguria to discharge.f Total number of days in ICU (general systems, coronary care, cardiovascular, neurosurgical and burns).g Risk factors for funguria or invasive fungal infection, as per Table 2.h Assessments following funguria, as per Table 3.

Table 2 Risk factors for funguria or invasive fungal infection

Variable Treated (79)a n (%) Observed (70)b n (%) All (149) n (%) p-valuec

Age .65d 50 (63) 32 (46) 82 (55) N/AAntibioticse 71 (90) 60 (86) 131 (88) NSUrinary catheterf 67 (85) 41 (59) 108 (72) ,0.01Urological abnormalitiesg 11 (14) 6 (8.6) 17 (11) NSGI/GU surgeryh 11 (14) 8 (11) 19 (13) NSICUi 26 (33) 17 (24) 43 (29) NSImmuno-suppressionj 18 (23) 16 (23) 34 (23) NSNeutropaeniak 2 (2.5) 0 (0.0) 2 (1.3) NSDiabetes mellitus 22 (28) 25 (36) 47 (32) NSHaemodialysisk 9 (11) 7 (10) 16 (11) NSCentral catheterk 34 (43) 18 (26) 52 (35) 0.03Hyper-alimentationk 14 (18) 1 (1.4) 15 (10) ,0.01Transplant 6 (7.6) 3 (4.3) 9 (6.0) NSBurnsk 2 (2.5) 0 (0.0) 2 (1.3) NSSevere mucositisk 1 (1.3) 1 (1.4) 2 (1.3) NS

a Sub-group of patients treated by controlling and/or removing identified risk factors for funguria or progression to invasive fungalinfection, and/or by local systemic antifungal therapy.

b Sub-group of patients observed only after funguria.c p . 0:2 is non-significant (NS).d Category chosen a priori; NOTE: age and gender were not included as risk factors.e Within previous two weeks.f At time of funguria or within previous two weeks; intermittent or indwelling.g Suprapubic catheter, nephrostomy drainage tube, ileal conduit, ureteral stent, etc.h Gastrointestinal or genitourinary surgery during admission prior to funguria.i Intensive care treatment within previous one week.j Steroids, anti-rejection medications, HIV infection, etc. within previous two weeks.k Present at time of funguria.

C. Simpson et al.250

Results

Of the 172 adult hospital inpatients identified bythe microbiology laboratory as growing yeast intheir urine, seven were excluded due to thepresence of IFI at the time of identification and 15because of a previously positive urine within thepast twelve months. One was lost-to-follow-up (noendpoint achieved after one year of enrollment).The remaining 149 eligible subjects did not reportsymptoms of urinary tract infection (as documentedin the patient record) and were enrolled into thestudy with demographics and characteristics aslisted in Tables 1 and 2.

The incidence of funguria in all hospital urinespecimens during the study period was found to be3%. Speciation was performed in 73% of cases and70% were identified as Candida albicans. Subjectswere assessed for follow-up investigations, as wellas types of treatment/risk factor control (Table 3).There was no significant difference between thetreated and untreated groups with respect tomortality rate (13% vs. 16%) or progression to IFI(2.5% vs. 2.9%). The treated group experienced asignificantly longer length of hospitalisation com-pared to the untreated group ðp , 0:01Þ; as well as alonger length of study enrollment ðp , 0:01Þ and anincreased number of days in the ICU ðp ¼ 0:05Þ:However, they also had significantly more riskfactors for funguria and progression to IFI ðp ,

0:01Þ: After adjusting for the number of risk

factors using multivariate analyses, there were nolonger any significant differences between thetreated and observed groups with respect tolength of hospitalisation, study enrollment anddays in the ICU.

Discussion

The management of asymptomatic funguria inhospital inpatients is currently an issue of debateand uncertainty, largely because the natural historyof funguria is unclear and it is also unknown whethera morbidity or mortality benefit exists for thetreated patient. In this study, funguria was found inapproximately 3% of all submitted urine specimensin a tertiary care referral centre where it tended tooccur in elderly, female patients. Forty-nine per-cent of our subjects had more than four risk factorsfor the development of funguria and/or IFI, 88% hadbeen treated with antibiotics, 72% had an intra-bladder drainage device inserted during the pre-vious two weeks, 32% had DM, 29% had beenadmitted to the ICU within the past week, and28% had concomitant bacteriuria. Seventy percentof all speciated specimens were identified asC. albicans. These results are consistent withthose previously reported.1,3

In subjects treated with risk factor controland/or antifungal therapy, a significantly greaternumber of repeat urine cultures, urinalyses or bloodcultures were submitted within one week followingthe identification of funguria. However, there wasnot a significantly greater number of follow-upabdominal computed tomography scans, abdominalsonograms or endourological procedures performedin the treated compared to the observed group. It ispossible that those treated were thought to be athigher risk for developing IFI and thus werefollowed up more closely with further investi-gations once the funguria was identified.

In this study, 47% of all subjects were simplyobserved, while 53% were treated for funguria. Ofthe treated group, 46% were treated with risk factorcontrol alone, 28% with systemic antifungal therapyalone, and 1.3% with local antifungal therapy alone.In the same group, 70% received some form of riskfactor control, with or without antifungal treat-ment, 52% received antifungal therapy with orwithout risk factor control, and 24% received acombination of risk factor control and antifungaltherapy. There was no significant differencebetween the two groups with respect to the overallin-hospital mortality rate or of the rate of pro-gression to IFI, even taking into consideration the

Table 3 Antifungal treatment/risk factor eliminationcombinations

Treatment protocol n/79 (%)

Risk factor control/elimination onlya 36 (46)Amphotericin B bladder irrigation(ABBI) only

1 (1.3)

Systemic treatment onlyb 22 (28)

Risk factor control/elimination plusAll antifungal therapyc 19 (24)Just systemic therapy 18 (23)Just ABBI 1 (1.3)

ABBI plusSystemic therapy and risk factor

control/elimination2 (2.5)

Just systemic therapy 1 (1.3)Just risk factor control/elimination 1 (1.3)

a Including removal or replacement of urinary catheter,removal of urological stents, etc. change to intermittentbladder catheterizations, and discontinuation of antibioticswithin one week of identification of funguria.

b IV amphotericin B or PO/IV fluconazole, in variousregimens.

c Local or systemic therapy.

Effect of funguria management 251

larger number of risk factors within the treatedgroup. However, the treated group had a signifi-cantly longer length of hospitalisation, enrollmentperiod and ICU stay. This was found to be solelyrelated to the higher number of presenting riskfactors within the treated group.

Asymptomatic funguriaappears tohave similaritieswith asymptomatic bacteriuria in the elderly, whichis also associated with a higher rate of functionalimpairment than in age-matched controls withoutbacteriuria.11,12 Treatment of asymptomatic bac-teriuria in the elderly does not reduce mortality inmen13 or women,14 just as treatment of asympto-matic bacteriuria did not affect mortality in ourstudy.

In this study, we found that treatment forasymptomatic funguria did not impact adult hospi-tal inpatient morbidity or mortality. Rather, thepresence of multiple risk factors for funguria or IFIcorrelated with the length of hospitalisation intertiary care inpatients. The high incidence of co-morbidities in these patients suggests that thepresence of funguria may be a ‘sickness indicator’,similar to asymptomatic bacteriuria in the elderly.

Currently available data do not support treat-ment of asymptomatic funguria.

Acknowledgements

A special acknowledgment to the research nurses inthe Division of Infectious Diseases who tirelesslyhelped to collect the data, as well as to those inMedical Records who enabled us to find theinpatient charts on those subjects enrolled. Wewould also like to acknowledge the microbiologylaboratory personnel who identified the subjects forthis study.

References

1. Wise GJ. Fungal infections of the urinary tract. In: Walsh CP,Retik BA, Stamey AT, Vaughen ED, editors. Campbell’sUrology, 6th ed. Philadelphia: Saunders; 1992. p. 928—50.

2. Weber DJ, Rutala WA, Samsa GP, Wilson MB, Hoffmann KK.Relative frequency of nosocomial pathogens at a universityhospital during the decade 1980 to 1989. Am J Infect Control1992;20:192—7.

3. Kauffmann CA, Vazquez JA, Sobel JD, et al. Prospectivemulticenter surveillance study of funguria in hospitalizedpatients. Clin Infect Dis 2000;30:14—8.

4. Storfer SP, Medoff G, Fraser VJ, Powderly WG, Dunagan WC.Candiduria: retrospective review in hospitalized patients.Infect Dis Clin Pract 1994;3:23—9.

5. Fisher JF, Newman CL, Sobel JD. Yeast in the urine: solutionsfor a budding problem. Clin Infect Dis 1995;20:183—9.

6. Munoz P, Gijon P, Akala L, Bouza E. How to managecandiduria. Rev Med Microbiol 1998;9(4):225—30.

7. Duma RJ. Commentary: Candiduria—recognition and treat-ment. Infect Dis Clin Pract 3;1:29—31

8. Fan-Havard P, O’Donovan C, Smith SM, Oh J, Bamberger M,Eng RHK. Oral fluconazole versus amphotericin B bladderirrigation for treatment of candidal funguria. Clin Infect Dis1995;21:960—5.

9. Leu H-S, Huang C-T. Clearance of funguria with short-courseantifungal regimens: a prospective, randomized, controlledstudy. Clin Infect Dis 1995;20:1152—7.

10. Schoenbeck J. Studies on Candida infection of the urinarytract and on the antimycotic drug 5-fluorocytosine. Scand JUrol Nephrol 1972;(Suppl. 11):7—48.

11. Boscia JA, Kobasa WD, Knight RA, Abrutyn E, Levison ME,Kaye D. Epidemiology of bacteriuria in an elderly ambulatorypopulation. Am J Med 1986;80:208—14.

12. Abrutyn E, Mossey J, Berlin JA, Boscia J, Levison M, PitsakisP, Kaye D. Does asymptomatic bacteriuria predict mortalityand does antimicrobial treatment reducemortality in elderlyambulatory women? Ann Intern Med 1994;120:827—33.

13. Nicolle LE, Bjornson J, Harding GKM, MacDonell JA.Bacteriuria in elderly institutionalized men. N Engl J Med1983;309:1420—5.

14. Nicolle LE, Mahew WJ, Bryan L. Prospective randomizedcomparison of therapy and no therapy for asymptomaticbacteriuria in institutionalized elderly women. Am J Med1987;83:27—33.

C. Simpson et al.252