Bladder Volume at O xIndependent Risk Factor for Breakthrough Febrile UrinaryTract Infection

Siobhan E. Alexander, AngChristopher S. Cooper*

From the Departments of Urology and Pediat

affecting children, with an estimatedprevalence of 0.4% to 1.8% in thegeneral pediatric population and 30%in those with a history of febrile

fections and renal injury, and mini-mizing morbidity of treatment andfollowup.3 Factors influencing thedecision to surgically correct reflux

Abbreviationsand Acronyms:

BBD bladder-bowel dysfunctionfUTI febrile urinary tractinfection

PBC predicted bladder capacityUTI urinary tract infectionVCUG voiding cystourethrogramVUR vesicoureteral reflux

Accepted for publication October 2, 2014.Study received institutional review board

approval.* Correspondence: Department of Urology,

University of Iowa, 200 Hawkins Dr., 3RCP, IowaCity, Iowa 52242-1089 (telephone: 319-353-8673;FAX: 319-356-3900; e-mail: christopher-cooper@uiowa.edu).HE JOURNAL OF UROLOGY

1342 j www.jurology.com

0022-5347/15/1934-1342/0T

2015 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION AND RESEARCH, INC.

http://dx.doi.org/10.1016/j.juro.2014.10.002Vol. 193, 1342-1346, April 2015

Printed in U.S.A.provides additional prognostic information about the risk of pyelonephritis andresolution, and may assist with counseling and clinical decision making.

Key Words: pyelonephritis, urinary tract infections,urination, vesico-ureteral reflux

VESICOURETERAL reflux is one of themost common urological diagnoses

urinary tract infection.1,2 Treatmentgoals include preventing febrile in-ela M. Arlen, Douglas W. Storm, Kathleen Kieran and

rics, University of Iowa Hospitals and Clinics, Iowa City, Iowa

Purpose: Improved identification of children with vesicoureteral reflux at risk forrecurrent febrile urinary tract infection may impact management decisions. Wehypothesized that reflux occurring earlier during bladder filling increases theduration of exposure of the kidneys to bacteria, and, therefore, increases the riskof pyelonephritis.

Materials and Methods: Children with vesicoureteral reflux and detailed voidingcystourethrogram data were identified. Bladder volume at onset of reflux wasnormalized for age. Demographics, reflux grade, laterality, presence/absence ofbladder-bowel dysfunction and breakthrough febrile urinary tract infectionswere assessed. Median followup was 24 months (IQR 12 to 52).

Results: A total of 208 girls and 47 boys were analyzed with a mean SD age atdiagnosis of 3.1 2.6 years. On univariate analysis history of febrile urinarytract infection (HR 2.17, 95% CI 1.33e2.85, p 0.01), dilating vesicoureteralreflux (HR 1.6, 95% CI 1.05e2.42, p 0.03) and bladder-bowel dysfunction (HR1.66, 95% CI 0.99e2.75, p 0.05) were associated with an increased risk ofbreakthrough febrile urinary tract infection. Median bladder volume at onset ofreflux in children with breakthrough febrile urinary tract infection was signifi-cantly less (33.1%) than in those without infection (49.5%, p 0.003). Refluxonset at 35% predicted bladder capacity or less was associated with a signifi-cantly increased risk of breakthrough febrile urinary tract infection on multi-variate analysis (HR 1.58, 95% CI 1.05e2.38, p 0.03).Conclusions: Children with early filling vesicoureteral reflux are at increasedrisk for breakthrough febrile urinary tract infection independent of reflux grade.Bladder volume at onset of reflux should be recorded during cystograms since itnset of Vesicoureteral Reflu is an

formula, age 2 30 ml.All children were placed on prophylactic antibiotics

neous VUR resolution at a mean SD of 32.4 26.9 months (range 6 to 170) after initial diagnosis.

REAand followed with yearly cystograms until resolution ofreflux or operative repair, and updated PBC according toage was used with each subsequent VCUG. VUR resolu-tion was defined as a single negative VCUG. Childrenwith secondary VUR or not prescribed continuous anti-biotic prophylaxis were excluded from the study.

Patient demographics, history of UTI with or withoutfever, presence or absence of bladder and/or boweldysfunction, and subsequent imaging were assessed.Febrile UTI was defined as a positive urine culture ofmore than 100,000 cfu of a single organism associatedwith a body temperature of 101.5F or greater. Urinespecimens were obtained via bag specimen, catheteriza-tion or clean catch, depending on patient age. Cathe-terized specimens were obtained if bag specimens werepositive. BBD was defined as constipation, incontinencebeyond expected for age/developmental stage, and/or uri-nary urgency and frequency necessitating anticholinergicmedicine. Median duration of followup was 24 monthsinclude further risk of febrile urinary tract infection,risk of new renal parenchymal scarring anddecreased likelihood of spontaneous resolutionof reflux.4e7

Spontaneous resolution of VUR is dependent oninitial grade of reflux, age, bladder volume at onsetof VUR, gender, voiding dysfunction and presenceof renal scarring.4 High grade VUR has also beenassociated with increased risk of pyelonephritis andrenal scarring.8,9 We compared bladder volumeat VUR onset to risk of breakthrough fUTI andhypothesized that reflux occurring in early bladderfilling would increase the risk of pyelonephritisindependent of reflux grade due to increased dura-tion of exposure of the upper tracts to bacteria fromthe bladder, compared to cases of reflux occurringlater during bladder filling or only during voiding.

MATERIALS AND METHODSInstitutional review board approval was obtained. A totalof 255 children with primary VUR in whom bladder vol-ume at onset of reflux on initial VCUG was documentedwere identified between 1984 and 2010. VCUG protocol atour institution routinely includes 2 filling and voidingcycles in a nonsedated child. The contrast material isroom temperature and infused from a height of approxi-mately 70 cm above the pubic symphysis.

Cystogram findings including grade, laterality, volumeof contrast material instilled and bladder volume at VURonset were abstracted. VUR grade was determined usingthe International Reflux Study in Children classificationsystem.5 Dilating VUR was defined as grade III to V. Incases of bilateral reflux the highest VUR grade was usedfor analysis. Bladder volume at VUR onset was normal-ized for PBC according to age for all patients using the

REFLUX IN EARLY BLADDER FILLING INC(IQR 12 to 52).The association of bladder volume at VUR onset with

the outcome of breakthrough fUTI while on prophylacticA total of 102 children (40%) underwent surgicalcorrection of VUR during clinical followup at amean SD of 39.2 38.8 months (range 1 to 174)after initial diagnosis. Indications for surgeryincluded breakthrough fUTI in 40 patients (39.2%),persistent VUR in 31 (30.4%), differential loss ofrenal function in 27 (26.5%) and parental prefer-ence alone in 4 (3.9%). A total of 68 children (26.7%)experienced 90 breakthrough fUTI events, of whom51 (20%) had 1, 12 (4.7%) had 2 and 5 (2%) had3 fUTIs while on antibiotic prophylaxis. Mean SDinterval from VUR diagnosis to initial break-through fUTI was 23.2 32 months (range 1 to172). Of the patients 35 (13.7%) were lost to fol-lowup before either VUR resolution or surgicalcorrection, although all had undergone at least 1followup VCUG and followup data were analyzedup to the date of the last VCUG.

Univariate proportional means model analysisantibiotics was analyzed using data obtained from403 VCUGs performed in the 255 children included inthe study. Recurrent event data were analyzed usingthe proportional means model, with inference to test forthe effect of the model independent variables based onthe robust sandwich covariance estimate.10 Models werefitted with a single independent variable. In addition tobladder volume at VUR onset as percentage of PBC, othervariables tested included age, gender, VUR grade, later-ality, BBD and history of fUTI before diagnosis of VUR.From these univariate models those that had p

Variable All Pts Pts with Breakthrough fUTI Pts without fUTI

.5) 3.22 2.39 (0e10.8) 3.07 2.69 (0e12.5)

.6) 64/68 (94.1) 144/187 (77.0)

.2) 31/68 (45.6) 64/187 (34.2)

.8) 5/68 (7.4) 15/187 (8.0)

.9) 32/68 (47.1) 108/187 (57.8)

.6) 24/68 (35.3) 49/187 (26.2)

.8) 6/68 (8.8) 14/187 (7.5)

.8) 1/68 (1.5) 1/187 (0.5)

.0) 44/68 (67.7) 81/187 (43.3)

.7) 23/68 (33.8) 40/187 (21.4)

.2) 4/68 (5.9) 40/187 (21.4)

.5) 35/68 (51.5) 104/187 (55.6)

.2) 29/68 (42.6) 43/187 (23.0)

.6) 52/68 (76.5) 100/187 (53.5)49)

G INCREASES RISK OF URINARY INFECTIONfUTI (HR 1.74, 95% CI 1.15e2.63, p 0.01). Thisassociation was also observed comparing medianbladder volume percentage PBC at onset of VUR,with those with breakthrough fUTI having signifi-cantly lower median bladder volumes (33.1%,IQR 18.3% to 69.6%) compared to those withoutbreakthrough infection (49.5%, IQR 33.0% to 76.4%,p 0.003). Univariate assessment of the associationof patient demographics and clinical variables withincreased risk of breakthrough fUTI was also per-formed (table 2). As anticipated, a history of fUTI(HR 2.17, 95% CI 1.33e2.85, p 0.01) or dilatingVUR (HR 1.6, 95% CI 1.05e2.42, p 0.03) wasassociated with a significantly increased risk ofbreakthrough fUTI. Other variables that suggestedan association with breakthrough UTI were BBD(HR 1.66, p 0.05), 2 UTIs or more before diagnosis

Mean SD yrs age at initial VCUG (range) 3.11 2.61 (0e12No. female/total No. (%) 208/255 (81No. dilating VUR/total No. (%) 95/255 (37No. VUR grade/total No. (%):

I 20/255 (7II 140/255 (54III 73/255 (28IV 20/255 (7V 2/255 (0

No. bilat VUR/total No. (%) 125/255 (49No. BBD/total No. (%) 63/255 (24No. prior UTIs/total No. (%)

0 44/255 (171 139/255 (542 Or more 72/255 (28

No. history of fUTI/total No. (%) 152/255 (59Median mos followup (IQR) 23 (12eTable 1. Patient demographics and imaging characteristics

1344 REFLUX IN EARLY BLADDER FILLINof VUR (HR 2.98, p 0.05) and female gender(HR 2.74, p 0.09).

Using the demographic and clinical variables inthe univariate analysis that had an association withbreakthrough fUTI (ie p

which VUR was observed may have been impactedby not measuring the volume of contrast within the

REFLUX IN EARLY BLADDER FILLING INCREASES RISK OF URINARY INFECTION 1345normalized for PBC demonstrating higher resolu-tion rates.15e17 Clinicians and parents often opt forintervention based on the likelihood of spontaneousresolution as well as clinical course, thereby placingan emphasis on the ability to predict the likelihoodof recurrent febrile infections.

Antibiotic prophylaxis in children with VUR orrecurrent UTI has been widely used, and thereduction in UTIs has been reaffirmed by 2 recentlypublished large multi-institutional studies.9,18 Morethan a fourth of our patients experienced a febrilebreakthrough infection. Similar to previous reports,dilating reflux and a history of fUTI were indepen-dently predictive of breakthrough infections.

We also found that VUR occurring at low bladdervolumes was associated with an increased likelihoodof fUTI through time independent of other vari-ables, including VUR grade. Children with reflux at35% PBC or less were more than 1.5 times likelier toexperience a breakthrough fUTI than those withVUR occurring at greater bladder volumes. Onepossible explanation for this association is thatchildren with VUR occurring earlier may haveincreased exposure of the kidneys to the occasionalbacteria from within the bladder compared to thosewith reflux that occurs late in filling or only duringvoiding, which may amplify the risk of pyelone-phritis. Early VUR has been associated withanatomically incompetent ureteral orifices,19 whichmay also facilitate entry of bacteria from thebladder into the upper urinary tract. Converselychildren with reflux occurring later in bladderfilling or voiding are at decreased risk for recurrentpyelonephritis. The impact of bladder volume at theonset of VUR constitutes an additional clinicalvariable that should be considered along with mul-tiple other clinical variables in determining clinicalmanagement to maximize benefit and minimizerisk.

Bladder and/or bowel dysfunction is known toaffect UTI rate as well as spontaneous resolution ofreflux in children with VUR.3 Our results also sug-gest that a history of BBD increases the likelihood ofbreakthrough fUTI, although this association didnot quite reach significance. This result may reflecta limitation of our study by categorizing patients ashaving a history of BBD and not further assessingongoing BBD treatment and outcomes. Diagnosiswas based on overt BBD symptoms, and it is likelythat the use of a validated questionnaire or ancillarytest such as uroflowmetry would have increased thepercentage of children categorized as having BBD.In addition, the recently published RIVUR (Ran-domized Intervention for Children with Vesicoure-

teral Reflux) trial reveals the significant benefitof antibiotic prophylaxis in decreasing UTIs intubing and catheter and subtracting it from thetotal volume. The diagnosis of BBD was based onclinical symptoms and treatment and not urody-namics or validated symptom scores, and it seemslikely that use of the latter methods would haveresulted in more children being diagnosed withBBD. While some urine cultures were obtained viasterile catheterization, others were voided speci-mens, which may have yielded false-positive results,although there is currently no consensus amongmultiple international guidelines for either themethod of obtaining urine in toilet trained childrenor the number of colony forming units that consti-tute a positive urine culture.

Renal scarring is a known risk factor for fUTI.However, 99mtechnetium dimercaptosuccinic acidscans were not routinely obtained. Therefore, theincidence of renal scarring/dysplasia in our patientcohort is unknown. While all children were placedon continuous antibiotic prophylaxis, compliancewas not formally assessed.

Despite these limitations, when controlling forother variables, our study shows a significant in-dependent correlation between VUR occurring atlow bladder volumes and increased risk of break-through fUTI. This finding provides an additionalobjective measure to help assess projected clinicalcourse in individual patients.

CONCLUSIONSChildren with vesicoureteral reflux occurring at lowbladder volumes are at increased risk for break-through febrile urinary tract infection independentof reflux grade. This easily reportable variable onVCUG has previously been demonstrated as anindependent predictive factor for spontaneous res-olution of VUR. Our findings emphasize the impor-tance of routinely noting bladder volume at theonset of VUR during VCUG to provide additionalprognostic information about risk of pyelonephritis,children with BBD compared to those with BBDreceiving placebo,18 and all of the patients in ourstudy were prescribed antibiotic prophylaxis.

Our study has several limitations aside from itsretrospective nature. Since our institution is a ter-tiary referral center, there are uncontrolled vari-ables that may cause selection bias. In addition,only cystograms obtained within our hospital wereevaluated, so many children who were initially seenelsewhere for diagnosis and treatment of fUTI andinitial VCUG were excluded. The bladder volume atand to assist with clinical decision making andcounseling.

REFERENCES

1. Sargent MA: What is the normal prevalenceof vesicoureteral reflux? Pediatr Radiol 2000; 30:587.

2. Skoog SJ, Peters CA, Arant BS Jr et al: PediatricVesicoureteral Reflux Guidelines Panel summaryreport: clinical practice guidelines for screeningsiblings of children with vesicoureteral refluxand neonates/infants with prenatal hydro-nephrosis. J Urol 2010; 184: 1145.

3. Peters CA, Skoog SJ, Arant BS Jr et al: Summaryof the AUA guideline on management of primaryvesicoureteral reflux in children. J Urol 2010;184: 1134.

7. Sjostrom S, Sillen U, Bachelard M et al:Spontaneous resolution of high grade infantilevesicoureteral reflux. J Urol 2004; 172: 694.

8. Lin KY, Chiu NT, Chen MJ et al: Acutepyelonephritis and sequelae of renal scar inpediatric first febrile urinary tract infection.Pediatr Nephrol 2003; 18: 363.

9. Brandstrom P, Esbjorner E, Herthelius M et al:The Swedish Reflux Trial in Children: III. Urinarytract infection pattern. J Urol 2010; 184: 286.

10. Lin DY, Wei LJ, Yang I et al: Semiparametricregression for the mean and rate functions of

R Stat Soc Series B Stat: 711.

attooiatr N

HJ: Pupdate. Curr Opin Pediatr

ualizedephro

of primary vesicoureteral reflux: results from2,462 children. J Urol 2009; 182: 1535.

15. McMillan ZM, Austin JC, Knudson MJ et al:Bladder volume at onset of reflux on initialcystogram predicts spontaneous resolution.J Urol 2006; 176: 1838.

16. Kirsch AJ, Arlen AM, Leong T et al: Vesicoure-teral reflux index (VURx): a novel tool to predictprimary reflux improvement and resolution inchildren less than 2 years of age. J Pediatr Urol2014; 10: 1249.

17. Knudson MJ, Austin JC, Wald M et al: Compu-tational model for predicting the chance of early

laxis for children with vesicoureteral reflux.

eafeornothdeonuffid

eireod t

1346 REFLUX IN EARLY BLADDER FILLING INCREASES RISK OF URINARY INFECTIONThis study provides data that revincreased risk of breakthroughpatients with onset of VUR at 35%bladder capacity. As the authorseasily reportable variable at VCUGadditional information for clinicalThe future of VUR treatment restsdetermine which patients will squences of VUR. This study provl a significantbrile UTI inless estimatedte, this is anat can providecision making.our ability to

er the conse-es us with an

important pieccontinue to ddetermining whfebrile UTIs anto the VUR puzzle. We mustct our research efforts towardwith VUR is at increased risk forheir long-term effects.

Hillary L. CoppDepartment of Urology

UCSF Medical Center

San Francisco, CaliforniaEDITORIAL COMMENT6. Schwab CW Jr, Wu H, Selman H et al:Spontaneous resolution of vesicoureteral reflux:a 15-year perspective. J Urol 2002; 168: 2594.

14. Estrada CR Jr, Passerotti CC, Graham DA et al:Nomograms for predicting annual resolution rate

voiding cystourethrogram correlates withdynamic endoscopic hydrodistention grade ofureteral orifice. J Urol 2014; 192: 1503.4. Knudson MJ, Austin JC, McMillan ZM et al:Predictive factors of early spontaneous resolu-tion in children with primary vesicoureteralreflux. J Urol 2007; 178: 1684.

5. Lebowitz RL, Olbing H, Parkkulainen KV et al:International system of radiographic grading ofvesicoureteric reflux. International Reflux Studyin Children. Pediatr Radiol 1985; 15: 105.

Methodol 2000; 62

11. Saadeh SA and Mtract infections. Ped

12. Mak RH and Kuotract infection: an2006; 18: 148.

13. Cooper CS: Individicoureteral reflux. Nmanagement of ves-urol Mon 2012; 4: 530.

N Engl J Med 2014; 370: 2367.

19. Arlen AM, Broderick KM, Huen KH et al:Temporal pattern of vesicoureteral reflux onTK: Managing urinaryephrol 2011; 26: 1967.

athogenesis of urinary

resolution in children with vesicoureteral reflux.J Urol 2007; 178: 1824.

18. Hoberman A, Greenfield SP, Mattoo TK et al;RIVUR Trial Investigators: Antimicrobial prophy-recurrent events. J

Bladder Volume at Onset of Vesicoureteral Reflux is an Independent Risk Factor for Breakthrough Febrile Urinary Tract InfectionMaterials and MethodsResultsDiscussionConclusionsReferences