The Role of Bladder Urine Transforming Growth Factor-�1Concentrations in Diagnosis and Management of UnilateralPrenatal Hydronephrosis

Fayez Almodhen, Oleg Loutochin, John Paul Capolicchio, Roman Jednakand Mohamed El-Sherbiny*From Montreal Children’s Hospital, McGill University Health Center, Montreal, Quebec, Canada

Abbreviations

and Acronyms

BU � bladder urine

T1/2 � half-time

TGF � transforming growth factor

UPJO � ureteropelvic junctionobstruction

US � ultrasound

Submitted for publication November 21, 2008.Study received approval of ethical and re-

search committees of Montreal Children’s Hospi-tal Research Institute.

* Correspondence: Montreal Children’sHospital, 2300 Rue Tupper, Montreal, Quebec,Canada H3H1P3 (telephone: 514-412-4366;e-mail: mohamed.el-sherbiny@muhc.mcgill.ca).

Purpose: We evaluated the relationship between bladder urine transforminggrowth factor-�1 concentration and severity of hydronephrosis in newborns withunilateral prenatal hydronephrosis.Materials and Methods: We prospectively studied all newborns presenting with uni-lateral prenatal hydronephrosis between January 2005 and 2007. Patients with associ-ated anomalies, vesicoureteral reflux, contralateral pathology or ipsilateral ureteraldilatation were excluded from study. Postnatal evaluation included voiding cystoure-thrography, renal ultrasonography and determination of bladder urine transforminggrowth factor-�1 concentration. Diuretic renal scans were performed in patients withinitial grade 3 or 4 hydronephrosis or increasing hydronephrosis during followup. Py-eloplasty was performed when a well tempered renogram showed an obstructive drain-age curve with a half-time greater than 20 minutes and/or an obstructive washout curvepattern during the diuretic phase. Patients were analyzed in observational and surgicalgroups. We studied the longitudinal changes in bladder urine transforming growthfactor-�1 in each group and compared concentration levels in the first 3 months of life inboth groups.Results: A total of 42 newborns were included. The observational group consisted of31 patients followed for a mean of 14 � 6 months. During the first 3 months, from 3to 12 months and in the second year of life mean ultrasound grade and bladder urinetransforming growth factor-�1 decreased from 2.3 to 1.7 to 1.2 (p �0.05) and from11.5 to 8.6 to 6.1 pg/mmol creatinine (p �0.05), respectively. Pyeloplasty was per-formed in 11 patients at a mean age of 6 � 5 months. Mean followup was 7 � 5months. In the first 3 months, preoperatively and at 3 to 12 months postoperativelymean ultrasound grade and bladder urine transforming growth factor-�1 were 3.5, 4and 3 (p �0.05), and 23, 29 (p �0.05) and 8 pg/mmol creatinine (p �0.003), respec-tively. Mean bladder urine transforming growth factor-�1 levels in the first 3 monthsof life were 23 � 14 and 11.5 � 8 pg/mmol creatinine in the surgical and observa-tional groups, respectively (p �0.001). Limiting comparison to the 23 patients withinitial grades 3 and 4 hydronephrosis revealed levels of 23 � 14 and 13 � 9 pg/mmolcreatinine in the surgical and observational groups, respectively (p �0.02). At acutoff of 17 pg/mmol creatinine bladder urine transforming growth factor-�1 in thefirst 3 months of life was 82% sensitive and 86% specific in predicting surgery.Conclusions: Bladder urine transforming growth factor-�1 changes throughtime are associated with similar changes in hydronephrosis grade. Bladder urinetransforming growth factor-�1 in the first 3 months of life can predict the need forsurgery in newborns with prenatal hydronephrosis.

Key Words: hydronephrosis, kidney, transforming growth factor beta,

ureteral obstruction, urinary bladder

292 www.jurology.com0022-5347/09/1821-0292/0 Vol. 182, 292-298, July 2009THE JOURNAL OF UROLOGY® Printed in U.S.A.Copyright © 2009 by AMERICAN UROLOGICAL ASSOCIATION DOI:10.1016/j.juro.2009.02.132

BLADDER URINE TRANSFORMING GROWTH FACTOR-�1 AND NEONATAL HYDRONEPHROSIS 293

HYDRONEPHROSIS is the most common urological find-ing on prenatal ultrasound. Ureteropelvic junctionobstruction is the most common cause of prenatalhydronephrosis (60%).1,2 Most prenatally detectedhydronephrosis will resolve spontaneously duringfollowup. Postnatal diagnosis of obstruction vsnonobstructive dilatation remains difficult. Ultra-sonography has typically assumed a more impor-tant role in the surveillance of children with hy-dronephrosis but is not used alone to establish thepresence of obstruction. Presently diuretic renog-raphy is the most widely used investigation toevaluate renal function and drainage.3 However,the test is invasive and the use of ionizing radia-tion remains a concern, especially in infants andyoung children.4 Moreover, repeat tests are oftenrequired, especially when the drainage curve isequivocal.

Biochemical markers are another potentiallyuseful diagnostic tool that has recently become afocus of clinical research. Preliminary investiga-tions looking at urinary concentrations of TGF-�1have suggested that they may prove useful in thiscapacity.

Furness et al demonstrated that urinaryTGF-�1 levels were significantly increased in chil-dren with upper urinary tract obstruction com-pared to normal controls.5 They were the first toidentify a urinary marker that correlates withupper urinary tract obstruction with greater than90% sensitivity. However, their study was criti-cized for lack of correlation to patients with di-lated nonobstructed uropathy treated nonopera-tively. Subsequently El-Sherbiny et al comparedurinary TGF-�1 levels in patients with symptom-atic upper urinary tract obstruction to those inpatients with dilated nonobstructed upper urinarytracts.6 TGF-�1 was 80% sensitive and 82% spe-cific in diagnosing obstruction. More recentlyTaha et al showed that BU TGF-�1 is useful notonly in diagnosing upper urinary tract obstruc-tion, but also during postoperative followup.7

They found that BU TGF-�1 significantly de-creased following successful corrective surgery.We sought to evaluate its role in the diagnosis andlongitudinal followup of a homogeneous group ofnewborns with prenatal unilateral hydronephro-sis.

MATERIALS AND METHODS

We prospectively studied all newborns presenting withunilateral prenatal hydronephrosis between January2005 and 2007. Patients were excluded from study ifthey had associated congenital anomalies, vesi-coureteral reflux, contralateral renal pathology or ipsi-lateral ureteral dilatation. Patient assessment was per-

formed according to the current standards for the

evaluation of hydronephrosis and no additional invasivetesting was performed. The protocol was approved bythe ethical and research committees of the MontrealChildren’s Hospital Research Institute. A signed con-sent was obtained from the parent(s) or the legal guard-ian of all patients.

Postnatal evaluation was performed within 3 monthsafter birth and included clinical examination, renal ultra-sonography and voiding cystourethrography. Hydrone-phrosis was graded by the treating pediatric urologistaccording to the Society for Fetal Urology classification.8

Bladder urine specimen was collected from all patientsfor determination of BU TGF-�1 concentration. Urinesamples were obtained by catheterization during voidingcystourethrography or by applying a urine collection bag.TGF-�1 was measured in duplicate using a commerciallyavailable, quantitative sandwich enzyme linked immuno-assay kit (R&D Systems Inc, Minneapolis, Minnesota).The assay has been previously detailed by Furness et al.5

Briefly standards were prepared from a stock solution of2,000 pg/ml TGF-�1 by serial dilutions. TGF-�1 concen-trations of the samples were determined by extrapolatingfrom the standard curve. To account for sample differ-ences in urinary concentration and osmolality, the rawurinary TGF-�1 concentration was corrected for urinarycreatinine concentration (TGF-�1 pg/mmol creatinine),which is reported as the corrected value. Urinary creati-nine was measured in duplicate by a modified rate Jaffetechnique, and the mean value was used for calculation.9

A well tempered renogram was performed in patientswith initial grade 3 or 4 hydronephrosis.10 Patients werehydrated with 15 cc/kg dextrose 5% and 0.3% normalsaline intravenously for 30 minutes beginning 15 minutesbefore the study. Continuous bladder catheterization wasmaintained. 99mTechnetium mercaptoacetyltriglycine wasused with a minimum dose of 1 mCi and 50 �Ci/kg. Onemg/kg furosemide was administered intravenously afterthe renography phase (20 to 30 minutes) but not before theentire collecting system was believed to have been com-pletely filled, as determined by comparison with the com-panion ultrasound images.

All patients were followed periodically every 3 to 6months, and then annually with US. Bladder urine sam-ples collected at each followup were obtained for determi-nation of TGF-�1 concentration. Repeat renography wasperformed in patients with worsening hydronephrosis, ini-tial equivocal results or symptoms consistent with renalobstruction. Indications for pyeloplasty included initialipsilateral split function less than 35%, obstructive drain-age curve with T1/2 greater than 20 minutes, obstructivewashout curve pattern during diuretic phase, worseningdifferential renal function more than 5% or worseningdrainage curve.

Patients were analyzed in observational and surgicalgroups. We studied the longitudinal changes in US gradeand BU TGF-�1 in the observational group at 3 timepoints, namely the first 3 months, from 3 to 12 months andin the second year of life. These time points were decidedbased on available samples at the time of data analysisbecause the urine collection was obtained only duringroutine followup visits, as ordered by the treating pediat-

ric urologist. In the surgical group US grade and BU

BLADDER URINE TRANSFORMING GROWTH FACTOR-�1 AND NEONATAL HYDRONEPHROSIS294

TGF-�1 were compared at 3 time points, namely in thefirst 3 months, preoperatively and at 3 to 12 monthspostoperatively. Finally we compared BU TGF-�1 levelsin the first 3 months of life in both groups.

Statistical analysis was performed using the Mann-Whitney U and Wilcoxon rank sum tests, with p �0.05considered significant. Pearson correlation coefficientswere computed to evaluate correlations between vari-ables. The receiver operating characteristic curve was con-structed to determine the threshold value of the BUTGF-�1 concentration giving the best sensitivity and spec-ificity.

RESULTS

The study included 42 newborns (33 males and 9females) who were followed for 15 � 6 months(range 4 to 30). Hydronephrosis was on the left sidein 34 patients (81%) and on the right side in 8 (19%).Initial grade was 1 in 3 patients, 2 in 16, 3 in 15 and4 in 8.

A total of 31 patients (74%) were observed nonop-eratively for a mean of 14 � 6 months. In thisobservational group mean US grade and BU TGF-�1decreased from 2.3 to 1.7 to 1.2 and from 11.5 to 8.6to 6.1 pg/mmol creatinine, respectively, during thefirst 3 months, from 3 to 12 months and in thesecond year of life (figs. 1 and 2). All changes werestatistically significant.

Pyeloplasty was performed in 11 patients (26%)at mean age of 6 � 5 months (range 1 to 14). Thesepatients constituted the surgical group that was fol-lowed for 7 � 5 months postoperatively. Indicationsfor and timing of surgery are outlined in the table. Inthe surgical group the mean US grade and BUTGF-�1 were 3.5, 4 and 3 (p �0.05), and 23, 29

Figure 1. Changes in mean ultrasound grade through time in

observational group.

(p �0.05) and 8 pg/mmol creatinine (p �0.003), re-spectively, during the first 3 months, preoperativelyand at 3 to 12 months postoperatively (figs. 3 and 4).

Mean BU TGF-�1 concentrations in the first 3months of life were 23 � 14 and 11.5 � 8 pg/mmolcreatinine in the surgical and observational groups,respectively (p �0.001, fig. 5). Comparison limited tothe 23 patients with initial grades 3 and 4 hydrone-phrosis was associated with levels of 23 � 14 and 13� 9 pg/mmol creatinine for the surgical and obser-vational groups, respectively (p �0.02, fig. 6). At acutoff of 17 pg/mmol creatinine BU TGF-�1 in thefirst 3 months of life was 82% sensitive and 86%specific in predicting surgery (fig. 7).

We studied the possible influence of split renalfunction on the BU TGF-�1 concentration at theinitial 3-month evaluation. There was no significantlinear correlation of BU TGF-�1 with the split func-tion. Correlation coefficient was �0.42 (p �0.07). Asignificant linear correlation was observed be-tween initial hydronephrosis grade and initial BUTGF-�1 concentration. Correlation coefficient was0.40 (p �0.01, fig. 8).

DISCUSSION

TGF-�1 is a multifunctional cytokine that partici-pates in wound healing and regulates the depositionof collagen through modulation of collagen produc-tion and degradation.11 Increasing evidence sug-gests that in a variety of pathological states includ-ing ureteral obstruction excessive renal TGF-�1production is responsible for the dysregulation ofextracellular matrix production and the develop-

Figure 2. Changes in BU TGF-�1 concentration through time inobservational group.

ment of progressive renal fibrosis.12–15 Moreover,

BLADDER URINE TRANSFORMING GROWTH FACTOR-�1 AND NEONATAL HYDRONEPHROSIS 295

Isaka et al reported that interstitial fibrosis in uni-lateral ureteral obstruction could be blocked withTGF-�1 antisense oligodeoxynucleotides.16

Seremetis and Maizels observed that increasedrenal tissue levels of TGF-�1 in response to urinarytract obstruction were associated with an increase inurinary TGF-�1.12 Further studies have revealedsignificant differences in renal pelvic TGF-�1 con-centrations in children with UPJO compared tobladder urine.5,6 Moreover, BU TGF-�1 concentra-tions were significantly increased in children withUPJO compared to normal controls and those withdilated nonobstructed kidneys. Interestingly BUTGF-�1 concentrations decrease following correctivesurgery.6,7 Studies have demonstrated that meanBU TGF-�1 concentrations were lower in older com-

Characteristics of patients undergoing surgery

PtNo.

Age at First UrineSample (mos)

Initial Evaluation/Imaging Follow

USGrade

SplitFunction (%) T1/2 (mins) US Grade F

1* 0.6 4 30 More than 602 1.7 4 45 253 2 4 43 More than 604 0.2 4 30 275 0.2 4 50 More than 606 2.9 4 49 8 Worse/47 0.5 3 45 15 Worse/48† 3 3 45 9 Worse/49 0.3 3 55 17 Worse/4

10 0.2 3 45 20 Worse/411† 1.5 4 54 15 Worse/4

Patients 1 to 5 underwent early surgery following initial evaluation and patients* Patient had palpable renal mass at initial evaluation.† Patient had urinary tract infection at followup evaluation.

Figure 3. Changes in mean ultrasound grade through time in

surgical group.

pared to younger children.5–7 This finding in olderchildren might reflect a lower and more steady-stateproduction of this growth factor with longstandingobstruction.6 Taha et al found that the higher base-line level of BU TGF-�1 concentrations in youngerchildren increased its diagnostic accuracy to 100%.7

This evidence led us to propose that BU TGF-�1concentrations might be helpful in detecting signif-icant obstruction among newborns with prenatalUPJO.

We prospectively evaluated the role of BU TGF-�1concentrations in newborns with isolated prenatalunilateral hydronephrosis. This design allowed us tostudy a homogeneous group, and thereby minimizethe effects of confounding factors. Patients were fol-lowed longitudinally and finally analyzed as part of

ation/Imaging

Age at Preop UrineSample (mos)

Postop Imaging

(%) T1/2 (mins) US GradeSplit

Function (%)T1/2

(mins)

Improved/3 53 11Improved/4 49 6Improved/4 — —Improved/4 40 10Improved/4 48 12

More than 60 14 Improved/3 — —65 9 Improved/3 45 9

More than 60 4 Improved/1 45 6More than 60 14 Improved/1 49 8

25 4 Improved/3 45 12More than 60 5.4 Improved/4 52 9

underwent delayed surgery due to worsening obstruction.

Figure 4. Changes in BU TGF-�1 concentration through time in

up Evalu

Splitunction

574332492753

6 to 11

surgical group.

BLADDER URINE TRANSFORMING GROWTH FACTOR-�1 AND NEONATAL HYDRONEPHROSIS296

a surgical or observational group. The data showedthat in the observational group the decrease in hy-dronephrosis grade through time was associatedwith a similar decrease in BU TGF-�1 concentra-tions. This correlation between changes in BU TGF-�1concentration and ultrasound grade, which is the prin-cipal test for monitoring hydronephrosis, may have animportant implication in the followup of such patients.Should our results be confirmed in a larger series,periodic measurement of BU TGF-�1 concentrationscould be useful in reducing the number of ultrasoundsand renograms during followup, and in reassuring thephysician and parents. To our knowledge this practicalnoninvasive method of monitoring has not been sug-gested before.

Figure 5. Mean BU TGF-�1 concentrations in first 3 months oflife in observational and surgical groups.

Figure 6. Mean BU TGF-�1 concentrations in first 3 months oflife in patients with grades 3 and 4 hydronephrosis in observa-

tional and surgical groups.

In the surgical group there was a marginal in-crease in the BU TGF-�1 concentration just beforepyeloplasty compared to the initial level. Althoughthis increase was not statistically significant, oneshould interpret these data with caution since only 6patients had the pyeloplasty performed during fol-lowup. Another explanation could be related to the

Figure 7. Receiver operating characteristic curve depicting thatat cutoff of 17 pg/mmol creatinine BU TGF-�1 in first 3 monthsof life was 82% sensitive and 86% specific in predicting surgery.

Figure 8. Relationship of initial hydronephrosis grade to initial

BU TGF-�1 concentrations.

BLADDER URINE TRANSFORMING GROWTH FACTOR-�1 AND NEONATAL HYDRONEPHROSIS 297

initial higher BU TGF-�1 concentration in thisgroup. BU TGF-�1 concentration significantly de-creased after pyeloplasty during a mean followup of7 months. The significant decrease in BU TGF-�1concentration postoperatively is consistent with pre-vious studies.6,7 These data provide further supportfor the role of BU TGF-�1 concentration as a usefulnoninvasive tool for followup after pyeloplasty.

One of the most common dilemmas presented byprenatal hydronephrosis is determining which pa-tients will ultimately need pyeloplasty. All existingtests are radiological and either invasive or operatordependent. Consequently we focused our effort onaddressing this aspect in our study.

A point of interest in our observations is the signif-icant difference between the initial (first 3 months oflife) levels of BU TGF-�1 concentration determined forthe observational and surgical groups. We conducted asubanalysis limited to high risk cases (grades 3 and 4hydronephrosis), and the significant difference be-tween the observational and surgical groups wasmaintained. It is noteworthy that obstruction was di-agnosed in the surgical group based on a well tem-pered renal scan either initially or during followup.

To our knowledge our study is the first to evaluatethe initial (first 3 months of life) value of the BUTGF-�1 concentration as a prognostic factor in new-borns with hydronephrosis. The sensitivity of thetest in our study (82%) was lower than that reportedby Furness et al (92%),5 which could be attributed tothe difference in the control groups in both studies.Furness et al included normal children without up-per tract dilatation in the control group, which is incontrast to our control group of children with a di-lated nonobstructed upper tract who did not requiresurgery for a mean followup of 14 months. Moreover,in a previous study El-Sherbiny el al obtained asimilar sensitivity when they used a control group ofchildren with dilated but nonobstructed uppertracts.6 We believe that the 82% sensitivity deter-

mined in our study is more reflective of the natural

REFERENCES

as a function of patient age, presentation and neim MA: Role of urinary tr

clinical situation since the question of obstructionwill be raised only in the presence of dilatation.

The level of TGF-�1 expression has previouslybeen correlated with the outcome following surgicalcorrection of obstruction. Seremetis and Maizels re-ported that cases with a higher mean level of TGF-�1expression had more significant improvement on thepostoperative renogram compared to those withlower expression.12 However, the relationship be-tween renal function and BU TGF-�1 concentrationhas not been confirmed in other series.6,7,17 In ourstudy the lack of correlation between BU TGF-�1concentration and renal function could be explainedat least partly by our small number of patients. Astudy of the correlation between TGF-�1 and renalfunction in a larger group of patients might helppredict recoverability of renal function following re-lief of obstruction. Our study highlights the role ofBU TGF-�1 concentration as an additional usefulnoninvasive test that significantly correlates withhydronephrosis grade and can predict the futureneed for pyeloplasty.

One of the limitations of our study is that theincrease in the level of TGF-�1 is not specific toobstructive uropathy. Moreover, although the crite-ria we used for determination of significant obstruc-tion have been reported previously,5–7 they remainarbitrary. Another limitation is the small samplesize and the small number of patients requiringsurgery. Certainly a larger study with high gradehydronephrosis and longer followup is warranted tocharacterize further the role of this noninvasive test.

CONCLUSIONS

BU TGF-�1 changes through time are associatedwith similar changes in hydronephrosis grade. BUTGF-�1 in the first 3 months of life can predict theneed for surgery in newborns with prenatal hydro-nephrosis. A study with a larger number of patientsis required to evaluate further the role of this non-

invasive tool.

1. Tam JC, Hodson EM, Choong KK, Cass DT, CohenRC, Gruenewald SM et al: Postnatal diagnosisand outcome of urinary tract abnormalities de-tected by antenatal ultrasound. Med J Aust 1994;160: 633.

2. Aksu N, Yavascan O, Kangin M, Kara OD, AydinY, Erdogan H et al: Postnatal management ofinfants with antenatally detected hydronephrosis.Pediatr Nephrol 2005; 20: 1253.

3. Salem YH, Majd M, Rushton HG and BelmanAB: Outcome analysis of pediatric pyeloplasty

differential renal function. J Urol 1995; 154:1889.

4. Kass EJ, Majd M and Belman AB: Comparison ofthe diuretic renogram and the pressure perfusionstudy in children. J Urol 1985; 134: 92.

5. Furness PD III, Maizels M, Han SW, Cohn RA andCheng EY: Elevated bladder urine concentrationof transforming growth factor-beta1 correlateswith upper urinary tract obstruction in children.J Urol 1999; 162: 1033.

6. El-Sherbiny MT, Mousa OM, Shokeir AA and Gho-

ansforming growth fac-

tor-beta1 concentration in the diagnosis of upperurinary tract obstruction in children. J Urol 2002;168: 1798.

7. Taha MA, Shokeir AA, Osman HG, Abd El-AzizAel-A and Farahat SE: Pelvi-ureteric junction ob-struction in children: the role of urinary trans-forming growth factor-beta and epidermal growthfactor. BJU Int 2007; 99: 899.

8. Fernbach SK, Maizels M and Conway JJ: Ultra-sound grading of hydronephrosis: introduction tothe system used by the Society for Fetal Urology.

Pediatr Radiol 1993; 23: 478.

BLADDER URINE TRANSFORMING GROWTH FACTOR-�1 AND NEONATAL HYDRONEPHROSIS298

9. Heinegard D and Tiderstrom G: Determination ofserum creatinine by a direct colorimetric method.Clin Chim Acta 1973; 43: 305.

10. Conway JJ and Maizels M: The “well tempered”diuretic renogram: a standard method to examine theasymptomatic neonate with hydronephrosis or hy-droureteronephrosis. A report from combined meetingsof the Society for Fetal Urology and members of thePediatric Nuclear Medicine Council–the Society of Nu-clear Medicine. J Nucl Med 1992; 33: 2047.

11. Basile DP: The transforming growth factor beta

EDITORIAL COMMENT

REFERENCES

REPLY BY AUTHORS

assessing obstruction and function. Th

progress and future directions. Curr OpinNephrol Hypertens 1999; 8: 21.

12. Seremetis GM and Maizels M: TGF-beta mRNAexpression in the renal pelvis after experimental andclinical ureteropelvic junction obstruction. J Urol1996; 156: 261.

13. Yamamoto T, Noble NA, Miller DE and BorderWA: Sustained expression of TGF-beta 1 under-lies development of progressive kidney fibrosis.Kidney Int 1994; 45: 916.

14. Chung KH and Chevalier RL: Arrested develop-ment of the neonatal kidney following chronic

e detailed in-need for pyeloplas

15. Seseke F, Thelen P, Heuser M, Zoller G andRingert RH: Impaired nephrogenesis in ratswith congenital obstructive uropathy. J Urol2001; 165: 2289.

16. Isaka Y, Tsujie M, Ando Y, Nakamura H,Kaneda Y, Imai E et al: Transforming growthfactor-beta 1 antisense oligodeoxynucleotidesblock interstitial fibrosis in unilateral ureteralobstruction. Kidney Int 2000; 58: 1885.

17. Chiou YY, Chiu NT, Wang ST, Cheng HL and TangMJ: Factors associated with the outcomes ofchildren with unilateral ureteropelvic junction ob-

system in kidney disease and repair: recent ureteral obstruction. J Urol 1996; 155: 1139. struction. J Urol 2004; 171: 397.

The search for biomarkers of congenital urinary ob-struction is on. One is always a bit concerned aboutputting all of the eggs in the basket of TGF-�. Howwell will this approach select those patients we aremost concerned with? Are the sensitivity and specific-ity appropriately balanced? Those are not simple deci-sions, but we need to be thinking along those lines.1,2

Similarly we must consider what end points havebeen used to pick a threshold value for any param-eter. In this study the authors have used loss ofrelative uptake on renal scan, as well as increasedwashout times. They recognize that these are indeedarbitrary choices with no real basis in clinical out-comes. They are “surrogate outcomes” that we must

arguing about the value of these outcomes for sev-eral decades. The authors do not provide us with thedetailed information as to which patients were op-erated on due to loss of function or due to increasedwashout times.

It is also important to realize, based on our studyof renal pathology in UPJO, that normal relativefunction on a scan may not always correlate withnormal renal pathology.3 Careful selection of pa-tients, end points and panels of indicators, and suf-ficient followup will be critical to this ongoing effort.

Craig A. Peters

Division of Pediatric UrologyUniversity of Virginia

use to permit clinical decision making. We have been Charlottesville, Virginia

1. Chevalier RL: Biomarkers of congenital obstructivenephropathy: past, present and future. J Urol 2004;172: 852.

2. Decramer S, Wittke S, Mischak H, Zurbig P, Wal-

outcome of congenital unilateral ureteropelvicjunction obstruction in newborn by urinary pro-teome analysis. Nat Med 2006; 12: 398.

3. Huang WY, Peters CA, Zurakowski D, Borer JG,

congenital ureteropelvic junction obstruction: evi-dence for parenchymal maldevelopment. KidneyInt 2006; 69: 137.

den M, Bouissou F et al: Predicting the clinical Diamond DA, Bauer SB et al: Renal biopsy in

We agree that there is no gold standard test for thediagnosis of obstruction, which creates more diffi-culty with any study designed to assess the role ofbiomarkers. We limited our patient population tothose with unilateral UPJO and excluded patientswith VUR, megaureter or other genitourinary pa-thology to increase the accuracy of renography in

formation as to which patients were operated on dueto loss of function or elevated washout times is pro-vided in the table. Our study highlights the role ofBUTGF-�1 concentration as an additional usefulnoninvasive test that significantly correlates withhydronephrosis grade and can predict the future

ty.