Urinary and Serum Carbohydrate Antigen 19-9 as a Biomarker in

Ureteropelvic Junction Obstruction in Children

Abdol-Mohammad Kajbafzadeh,* Azadeh Elmi, Saman Shafaat Talab,Hamed Emami, Shadi Abdar Esfahani and Parisa SaeediFrom the Pediatric Urology Research Center, Department of Urology, Pediatric Center of Excellence, Children’s Hospital Medical Center,Tehran University of Medical Sciences, Tehran, Iran

Purpose: We evaluated the predictive role of serum and urinary carbohydrateantigen 19-9 in the diagnosis and followup of pediatric ureteropelvic junctionobstruction.Materials and Methods: The study included 27 children with ureteropelvic junc-tion obstruction who underwent pyeloplasty (group 1), and 41 controls consistingof 27 healthy children (group 2) and 14 children with hydrocele/renal cyst (group3). Serum and voided urine were evaluated for carbohydrate antigen 19-9 in eachgroup. Additionally urine from the affected pelvis and fluid in hydrocele/renalcyst were collected at surgery in groups 1 and 3. Serum and voided urine sampleswere obtained at 3, 6 and 9 months after pyeloplasty for carbohydrate antigen19-9 assessment, and were correlated with clinical factors.Results: Preoperative carbohydrate antigen 19-9 level was significantly greaterin group 1 than in controls. The best cutoff values for serum and urinarycarbohydrate antigen 19-9 were 13.21 U/ml and 30.6 U/ml, respectively, withsignificantly higher sensitivity and specificity for urinary values. Obstructionrelease was followed by improvement of renal function together with significantreduction in urinary and serum carbohydrate antigen 19-9 at 3 months. Initialpelvis diameter and renographic function significantly correlated with urinarycarbohydrate antigen 19-9. No significant correlation was found regarding serumcarbohydrate antigen 19-9.Conclusions: Voided urine carbohydrate antigen 19-9 is a noninvasive, clinicallyapplicable marker in congenital obstructive nephropathy. The practical implica-tions of these data for diagnosis and long-term followup in ureteropelvic junctionobstruction are significant. Our findings suggest that proper decrease in urinarycarbohydrate antigen 19-9 after pyeloplasty is predictive of excellent surgicaloutcomes and resolution of renal damage.

Key Words: CA-19-9 antigen; congenital abnormalities; tumor markers,

Abbreviations

and Acronyms

AP � anteroposterior renal pelvic

CA-19-9 � carbohydrate antigen19-9

DTPA � diethylenetriaminepentaacetic acid

SFU � Society for Fetal Urology

UPJ � ureteropelvic junction

Submitted for publication September 15,2009.

Funded by Iranian National Science Founda-tion.

Study received institutional review board ap-proval.

* Correspondence: No. 34, 2nd Floor, 7th St.,Saadat-Abad, Ave. Tehran 1998714616, Iran (tele-phone: 98-21-2208-9946; FAX: 98-21-2206-9451;e-mail: kajbafzd@sina.tums.ac.ir).

biological; ureteral obstruction

URETEROPELVIC junction obstruction isthe most common congenital abnor-mality of the ureter, and an identifi-able cause of hydronephrosis and re-nal failure in children.1 Prolongedobstruction in the pelvicaliceal systemleads to tissue damage and tubulo-in-

terstitial fibrosis, and consequently de-

0022-5347/10/1836-2353/0THE JOURNAL OF UROLOGY®

© 2010 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION AND RES

terioration of renal growth and loss ofnephrons.

The key point in treatment of pa-tients with severe UPJ obstruction isrelief of the obstruction and restora-tion of normal urine passage throughthe junction, which results in kidney

preservation. However, the timing of

Vol. 183, 2353-2360, June 2010Printed in U.S.A.

EARCH, INC. DOI:10.1016/j.juro.2010.02.031www.jurology.com 2353

CARBOHYDRATE ANTIGEN 19-9 LEVELS IN URETEROPELVIC JUNCTION OBSTRUCTION2354

surgical intervention remains controversial and isoften based on interpretation of differential renalfunction, degree of obstruction and washout mea-surements from diuretic renography.2 Although nogold standard is available to identify pathologicalobstruction, conventional imaging modalities usedfor diagnosis of UPJ obstruction occasionally areinsufficient for decision making in managing thecondition. Additionally these methods mandate ex-posure to radiation. Thus, there is a need to discovernoninvasive, simple and cost-effective alternativemethods for diagnosing congenital obstructive ne-phropathy with high accuracy.

Identification of sensitive biomarkers seems to bepotentially applicable in diagnosis and followup ofpediatric UPJ obstruction. The markers may eitherhave a role in the embryogenesis pathway of thecondition or be the consequence of cellular responseof the developing kidney to obstruction. Several fac-tors have been investigated for this purpose, withvarying degrees of success (transforming growth fac-tor-�1, epidermal growth factor, monocyte chemo-tactic peptide, endothelin-1).3–6 These factors dem-onstrate significant changes in marker levels afterobstruction release associated surgery.

Serum CA-19-9 has been clinically applied as avaluable tumor marker for pancreatic and gastroin-testinal carcinoma.7 It has been shown that CA-19-9is expressed in normal tissues with excretory epithe-lium. Moreover, increased CA-19-9 has been ob-served in uroepithelial tumors.8 Although its increasein nonmalignant disease is unusual, some rare hydro-nephrosis cases with high serum levels due to un-known etiology have also been documented. These con-ditions include hydronephrosis secondary to ureteralstone, transitional cell carcinoma of the renal pelvisand ureteral cancer.9–11 We investigated whether se-rum and urine CA-19-9 levels are increased in congen-ital obstructive nephropathy caused by UPJ obstruc-tion. Additionally we examined possible clinicalapplication of urinary CA-19-9 as a noninvasive diag-nostic and predictive biomarker in UPJ obstruction.

PATIENTS AND METHODS

The study included 17 boys and 10 girls with unilateralUPJ obstruction who underwent pyeloplasty between May2005 and October 2007 (group 1). Mean patient age was27.62 months (range 0.5 to 98). Criteria for pyeloplastyincluded clinical, ultrasound and DTPA scan evaluations,and/or failure of conservative management.

This study was approved by the institutional reviewboard, and written informed consent was obtained from allparents after the purpose of the study was explained.Study inclusion criteria were absence of malignant tumor,normal liver function and absence of benign disease thatcould possibly cause increased CA-19-9. Diagnosis was

made by renal ultrasound and confirmed by 99mtechne-

tium DTPA renal scan. Voiding cystourethrogram wasperformed to assess for vesicoureteral reflux. All childrenhad sterile urine and normal blood urea nitrogen andcreatinine levels.

Serum and voided urine samples were obtained fromthe 2 control groups. The first group consisted of genderand age matched children (17 boys and 10 girls, mean 30.1months) who were healthy without urinary dilatation onultrasound, and free of malignant or benign disease caus-ing increased CA-19-9 (group 2). The second group con-tained 8 boys with hydrocele and 6 patients with renalcyst (mean age 32.4 months, group 3). In this latter groupCA-19-9 levels were also evaluated in fluids obtained dur-ing surgery.

In group 1 first morning voided urine and serum sam-ples were obtained preoperatively and after induction ofanesthesia, with urine samples collected directly from theobstructed pelvis with a 22 gauge needle. All patientsunderwent dismembered pyeloplasty. Serum and voidedurine samples were also obtained from all patients at 3, 6and 9 months postoperatively.

Interpreted clinical features were age, gender, lateral-ity, grade of hydronephrosis, AP diameter and renal func-tion. Hydronephrosis grade was determined according toSFU guidelines.12 Carbohydrate antigen 19-9 was mea-sured using an electrochemiluminescence enzyme immu-nometric kit (Roche Elecsys® Kits). Patients were followedwith clinical examinations, urinalysis, ultrasound and di-uretic renography at 3 and 9 months postoperatively.

Nonpaired Student’s t test was used to compare meanswhen the experimental design consisted of 2 samples.Data were expressed as mean � SEM. Correlation coeffi-cients were Pearson’s r values. ROC curve was used todetermine the cutoff values of urinary and serum CA-19-9that gave the best sensitivity and specificity.

RESULTS

Preoperative CA-19-9 levels from serum and voidedurine were significantly greater in group 1 (19.10 �2.17 U/ml and 319.20 � 36.71 U/ml, respectively) thanin controls (7.01 � 1.53 U/ml and 16.19 � 2.34 U/ml,see table). No significant association was found re-garding age and urinary CA-19-9 levels (fig. 1).

Serum and voided urine CA-19-9 in patients with UPJobstruction and controls

Mean � SEMSerum CA-19-9

(U/ml)

Mean � SEMVoided Urine

CA-19-9 (U/ml)

Mean � SEMSurgical Fluid

CA-19-9 (U/ml)

UPJ obstruction:Before pyeloplasty 19.10 � 2.17* 319.20 � 36.70* 1,765.67 � 447.703 Mos postop 10.25 � 1.67† 53.82 � 6.12*,‡ —9 Mos postop 6.75 � 0.98† 18.40 � 3.05‡ —

Healthy controls 7.10 � 1.53 16.91 � 2.34 —Renal cyst, hydrocele 8.25 � 1.61 14.80 � 3.95 3.91 � 0.69

CA-19-9 levels in samples obtained from dilated renal pelvis and hydrocele or renalcyst during surgery were also evaluated.* p �0.001 compared to controls.† p �0.05.

‡ p �0.001 compared to preoperative values.

CARBOHYDRATE ANTIGEN 19-9 LEVELS IN URETEROPELVIC JUNCTION OBSTRUCTION 2355

CA-19-9 level in urine sample obtained from thedilated renal pelvis during surgery was 1,765.67 �447.70 U/ml, which was significantly greater thanin voided urine (p �0.001). Compared to preoper-ative levels, significant decreases in serum andurinary values were noted at 3 months postoper-atively (p �0.05 and p �0.001, respectively, seetable). The fluid obtained from renal cyst or hy-drocele exhibited no difference compared to nor-mal urine.

ROC curve for serum CA-19-9 revealed 74.1%sensitivity and 87% specificity at a reference valueof 13.21 U/ml. For voided urine CA-19-9 the sen-sitivity was 100% and specificity was 82.6% at acutoff of 30.6 U/ml (fig. 2).

Two patients with poor renal function and con-tinuously increasing washout curve on diuretic

Figure 1. Correlation between age and voided urine CA-19-9 incontrols (r � 0.034, p � 0.866) and pyeloplasty group (UPJO, r �0.029, p � 0.888). UPJO, UPJ obstruction.

Figure 2. Receiver operating characteristic curve constructed todetermine CA-19-9 reference value for optimal sensitivity and

specificity for voided urine and serum.

renography demonstrated slightly higher thancutoff values in voided urine, although this findingdid not correlate with severity of obstruction. Ac-cordingly pyeloplasty was performed and mea-sured intrapelvic CA-19-9 exhibited levels of morethan 5,000 U/ml.

Intriguingly the initial AP diameter and voidedurine CA-19-9 level showed significant correlation(r � 0.403, p �0.05, fig. 3). Moreover, comparison ofrenal parenchymal thickness of the obstructed kid-ney preoperatively and at 3 months postoperativelyrevealed statistical significance (3.93 � 2.37 mm vs9.26 � 3.41 mm, p �0.001), although no significantcorrelation was found regarding renal parenchymalthickness and CA-19-9 levels. While the differencein SFU grade was significant using chi-square test(p �0.001), hydronephrosis grade demonstrated in-significant correlation with CA-19-9.

Initial renographic function showed significantinverse correlation with intrapelvic urine CA-19-9 inall cases (r � 0.613, p �0.001, fig. 4, A). Additionallysignificant correlation was found between voidedurine CA-19-9 and initial renal function after ex-cluding 2 cases with a discrepancy between urinaryCA-19-9 and clinical findings (r � 0.474, p �0.05, fig.4, B).

Preoperatively mean � SEM differential functionwas 31.51% � 2.26%. This value reached 41.40% �1.46% at 3 months and 46.20% � 1.94% at 9 monthsafter obstruction release, which was a significantimprovement (p �0.001).

No significant correlation was found regardingserum CA-19-9 levels and aforementioned clinicalfactors. Figure 5 represents the initial diuretic renalscan and magnetic resonance urography in 1 case

Figure 3. Correlation between voided urine CA-19-9 and initialAP diameter (r � 0.403, p �0.05).

with associated CA-19-9 levels.

CARBOHYDRATE ANTIGEN 19-9 LEVELS IN URETEROPELVIC JUNCTION OBSTRUCTION2356

DISCUSSION

To our knowledge this study represents the firstpublished attempt to elucidate increased serumand urinary CA-19-9 levels with significant corre-lation in patients with congenital UPJ obstruc-

Figure 4. A, correlation between intrapelvic urine CA-19-9 and inurine CA-19-9 and initial renal function (r � 0.474, p �0.05).

Figure 5. Prenatally diagnosed left hydronephrosis in 7.5-year-enhanced T1-weighted magnetic resonance urography sequencesevere hydronephrosis (length 185 mm, cortex 6 mm) on left sidvoided urine and 1,046.9 U/ml for intrapelvic urine. Postoperativ

urine. GFR, glomerular filtration rate.

tion. Based on our results, significant decrease inCA-19-9 occurred at 3 months after obstructionrelease and reached normal levels at 9 months.Thus, we propose urinary CA-19-9 as a monitoringbiomarker in congenital obstructive nephropathy,

nal function (r � 0.613, p �0.001). B, correlation between voided

l. A, diuretic renal scan with 99mtechnetium DTPA. B, contrastls morphology of kidney/ureter systems representing persistentperative CA-19-9 levels were 19.8 U/ml for serum, 633 U/ml for

ls (3 months) were 4.4 U/ml for serum and 59.5 U/ml for voided

itial re

old girrevea

e. Preoe leve

CARBOHYDRATE ANTIGEN 19-9 LEVELS IN URETEROPELVIC JUNCTION OBSTRUCTION 2357

with significant correlation with initial reno-graphic function and AP diameter.

Patients with UPJ obstruction pose evaluationand treatment challenges. Although diureticrenography and ultrasound measurements areavailable for establishing the presence and sever-ity of UPJ obstruction, correlation of histologicalfindings and extent of deterioration of renal func-tion has been unsatisfactory. While there are noclear indications for timing of surgical interven-tion to prevent kidney injury, discovery of poten-tial biomarkers of disease progression resultingfrom urinary tract obstruction could be useful indetermining targets for therapeutic manipulationand response monitoring. Grandaliano et al re-ported reduced urinary epidermal growth factorconcentration in obstructive nephropathy, whilemonocyte chemotactic peptide 1 urine concentra-tion was markedly increased.4 After surgical cor-rection they noted a marked change in the factorlevels. Taha et al evaluated endothelin 1 in voidedurine for diagnosis of UPJ obstruction, with 100%sensitivity and specificity in children youngerthan 1 year.5

CA-19-9 is widely used as a diagnostic and prog-nostic tumor marker. However, it is also known toincrease in patients with benign disorders. A fewrecent studies have addressed this issue in benignurological disease associated with significant hy-dronephrosis.10,12,13 These series revealed a dra-matic decrease after obstruction release surgery,and indicate that serum and urinary CA-19-9 lev-els correlate with hydronephrosis. They also con-firmed in vivo expression of CA-19-9 in hydrone-phrotic pelvic mucosa, as well as in tubules inresponse to obstructive injury. The proposedmechanisms include inflammation or proliferationof noncancerous tissues resulting in enhanced ex-pression of marker, as well as impairment in theCA-19-9 discharge pathway.14

The standard cutoff of serum CA-19-9 for ma-lignancy (less than 37 U/ml) has also been re-ported in patients with hydronephrosis, with anincidence of 25%.10 This finding was inconsistentwith that of the present series. We observed thatalthough increased serum CA-19-9 was detectedin most cases, serum levels greater than 37 U/mlwere found in only 11.1%, demonstrating a mini-mal role for serum CA-19-9 in the diagnosis andmanagement of UPJ obstruction. On the otherhand, all cases exhibited urinary CA-19-9 levelsgreater than 98 U/ml. Moreover, the striking in-crease in urinary levels, when compared to serumvalues, suggests a renal rather than systemic or-igin for either increased production of or decline indegradation of this tumor marker in this subset of

patients. It is noteworthy that a cutoff of 44.45

U/ml represents 100% sensitivity and specificityfor predictive values of urinary CA-19-9.

Although bladder urine CA-19-9 was signifi-cantly lower than intrapelvic levels in all cases, asignificant correlation was established betweenthe 2 values, suggesting a valuable role for preop-erative voided urine CA-19-9 in predicting in-trapelvic levels. One exception to this finding wasin patients with severely obstructed UPJ and ex-tensive renal damage. In this setting drainage ofurine and accordingly CA-19-9 were impaired, anddespite extremely high levels of CA-19-9 in therenal pelvis, the level in voided urine was inap-propriately low. However, sensitivity of themarker for predicting UPJ obstruction is main-tained in cases with low voided urine CA-19-9,although with evidence of severe obstruction andongoing renal damage the decision for surgery isin line with clinical judgment rather than basedmerely on CA-19-9 levels.

An important limitation of previous studies in-troducing urinary biomarkers for obstructive ne-phropathy could be the lack of different types ofcontrol to determine the exact specificity of themarker. In this regard we evaluated CA-19-9 inserum, urine and intrarenal cyst fluid to confirmthe exclusive increase of this biomarker in ob-structive pathology rather than cystic lesions ofthe kidney. In view of the fact that serum CA-19-9is a sensitive marker for malignancies and someother benign conditions involving epithelial struc-tures, along with our findings, we suggest a low sen-sitivity of serum CA-19-9 for predicting urinary ob-struction. We propose that CA-19-9 measured invoided urine might be an alternative noninvasive testwith high predictive value and practical clinical impli-cations compared to serum CA-19-9 in patients withUPJ obstruction.

Grignon et al reported the efficiency of ultrasoundin early identification and serial assessment of UPJstenosis.15 Subsequently the value of SFU grading ingrades 1 and 2 hydronephrosis was confirmed, whiledifferential renal scan seemed helpful in higherSFU grades.16 In this series we measured renalparenchymal thickness to monitor kidney damagestatus and changes were correlated to biomarkerlevel. Our data established a significant correla-tion of initial AP diameter as a manifestation ofconsiderable obstruction, rather than hydrone-phrosis grading (including renal parenchymalthickness) with urinary CA-19-9. It is also note-worthy that CA-19-9 in voided and intrapelvicurine was directly correlated to the extent of ini-tial renal function loss, and in this regard it maybe an indicator of split renographic function statuspreoperatively. However, in our cases the correla-

tion of intrapelvic urine was stronger than voided

CARBOHYDRATE ANTIGEN 19-9 LEVELS IN URETEROPELVIC JUNCTION OBSTRUCTION2358

urine, and the significant level of the correlationbetween voided urine and renal function was onlyobtained after excluding the 2 cases with discrep-ancy between urinary CA-19-9 and clinical fea-tures. One explanation for these findings could bethat voided specimens represent the contributionof both kidneys rather than the affected kidney.

Elder et al reported that kidneys with renalfunction less than 35% had less postoperativefunctional enhancement due to significant tissuechanges.17 Interestingly our patients with differ-ential renal function less than 35% showed highurinary CA-19-9 and less improvement. We hy-pothesize that pediatric UPJ obstruction and lowCA-19-9 levels in voided and intrapelvic urine re-cover more rapidly or completely after pyelo-plasty.

Decreasing serum and urinary CA-19-9 levels af-ter pyeloplasty were the most clinically significantparameters in the implication of this marker. At 3months the voided urine level, although significantlylower than the preoperative level, was greater than incontrols. Additionally diuretic renography at this pointrevealed significant improvement but further followupis essential to allow prediction of long-term renal func-tional status. We may conclude that renal steady statephase is established several months after surgical cor-rection of the obstruction. This finding suggests uri-nary CA-19-9 as a noninvasive marker for followup of

pediatric UPJ obstruction after surgical release.

REFERENCES

epidermal growth factor and transforming growth 191.

It should be remembered that these results wereobtained in a subset of patients who were selectedfor pyeloplasty according to clinical, ultrasound andDTPA scan evaluations, and/or failure of conserva-tive treatment. Thus, further studies including alarger population and a subset of patients with a lowrisk, dilated pelvicaliceal system should be per-formed to confirm the accuracy of this biomarker inUPJ obstruction diagnosis, treatment and fol-lowup. Moreover, such studies would facilitatedemonstration of the role of urinary CA-19-9 as asurrogate for determining the appropriate manage-ment approach for congenital hydronephrosis, ob-struction that causes renal functional deterioration,from conditions that will not affect renal functionaldevelopment.

CONCLUSIONS

Carbohydrate antigen 19-9 in voided urine is anoninvasive, clinically applicable marker in con-genital obstructive nephropathy. In children withUPJ obstruction the sensitivity of voided urineCA-19-9 was 100%. The practical clinical implica-tions of these data for diagnosis and long-termfollowup of pediatric UPJ obstruction are signifi-cant. Our findings suggest that the proper de-crease in urinary CA-19-9 level during followup ispredictive of excellent surgical outcomes and res-

olution of renal damage.

1. Zhang PL, Peters CA and Rosen S: Ureteropelvicjunction obstruction: morphological and clinicalstudies. Pediatr Nephrol 2000; 14: 820.

2. Stock JA, Krous HF, Heffernan J et al: Correlationof renal biopsy and radionuclide renal scan dif-ferential function in patients with unilateral ure-teropelvic junction obstruction. J Urol 1995; 154:716.

3. Furness PD, Maizels M, Han SW et al: Elevatedbladder urine concentration of transforminggrowth factor-beta1 correlates with upper urinarytract obstruction in children. J Urol 1999; 162:1033.

4. Grandaliano G, Gesualdo L, Bartoli F et al: MCP-1and EGF renal expression and urine excretion inhuman congenital obstructive nephropathy. Kid-ney Int 2000; 58: 182.

5. Taha MA, Shokeir AA, Osman HG et al: Diagnosisof ureteropelvic junction obstruction in children:role of endothelin-1 in voided urine. Urology2007; 69: 560.

6. Yang Y, Zhou X, Gao H et al: The expression of

factor-beta1 in the stenotic tissue of congenitalpelvi-ureteric junction obstruction in children.J Pediatr Surg 2003; 38: 1656.

7. Koprowski H, Herlyn M, Steplewski Z et al: Spe-cific antigen in serum of patients with coloncarcinoma. Science 1981; 212: 53.

8. Vestergaard EM, Wolf H and Orntoft TF: In-creased concentrations of genotype-interpretedCa 19-9 in urine of bladder cancer patients markdiffuse atypia of the urothelium. Clin Chem 1998;44: 197.

9. Aoki D, Nomata K, Kanda S et al: A case ofpyonephrosis caused by ureteral stones with el-evated serum levels of CA19-9. Hinyokika Kiyo1999; 45: 629.

10. Suzuki K, Muraishi O and Tokue A: The correla-tion of serum carbohydrate antigen 19-9 withbenign hydronephrosis. J Urol 2002; 167: 16.

11. Taki T, Honda N, Yamada Y et al: CA19-9-pro-ducing transitional cell carcinoma of the renalpelvis: a case report. Hinyokika Kiyo 2001; 47:

12. 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.

13. Aybek H, Aybek Z, Sinik Z et al: Elevation ofserum and urinary carbohydrate antigen 19-9 inbenign hydronephrosis. Int J Urol 2006; 13: 1380.

14. Meyer A, Kausch I, Kruger S et al: Elevation of CA19-9 in giant hydronephrosis induced by a renalcalculus. Urology 2004; 63: 381.

15. Grignon A, Filiatrault D, Homsy Y et al: Uretero-pelvic junction stenosis: antenatal ultrasono-graphic diagnosis, postnatal investigation, andfollow-up. Radiology 1986; 160: 649.

16. Moon DH, Park YS, Jun NL et al: Value ofsupranormal function and renogram patterns on99mTc-mercaptoacetyltriglycine scintigraphy inrelation to the extent of hydronephrosis for pre-dicting ureteropelvic junction obstruction in thenewborn. J Nucl Med 2003; 44: 725.

17. Elder JS, Stansbrey R, Dahms BB et al: Renalhistological changes secondary to ureteropelvic

junction obstruction. J Urol 1995; 154: 719.

CARBOHYDRATE ANTIGEN 19-9 LEVELS IN URETEROPELVIC JUNCTION OBSTRUCTION 2359

EDITORIAL COMMENTS

The answer to the question of when we need tooperate on a dilated kidney remains one of the mostdifficult issues in modern pediatric urology. The cur-rent management is essentially based on imagingand relative renal function evaluation. These meth-ods are presently incapable of determining the rel-ative risk of renal damage in borderline cases.

The authors have chosen the promising biologicalmarker method, believing it will clarify the indicationsfor surgery. I congratulate them for choosing an al-ready well studied biological marker, CA-19-9, as apredictor for UPJ obstruction in children. The currentstudy shows a significant correlation between urinaryCA-19-9 and differential renal function. With respectto the noninvasive advantage of a biological marker,the relevance of such a study should be accepted basedon results obtained from noninvasive methods such as

REPLY BY AUTHORS

obstructive nephropathy are prequisites for intro-

both kidneys. Intrapelvic urine assessment is used inthis study as the optimal method, although this ap-proach requires that either the operative decision al-ready be made, while for nonoperated cases percuta-neous urine sampling is needed to measure CA-19-9.Either way, the marker would lose its great advantageof being noninvasive.

Nevertheless, the contribution of this series to clin-ical research is significant, and it will open the way forfurther clinical studies. I am sure that in the nearfuture the authors will provide us with the next step oftheir study, focusing on the main clinical issues—anoninvasive method and reproducible predictive valuefor assessing the nonoperated dilated system.

Alaa El-Ghoneimi

Department of Pediatric SurgeryUniversity Diderot-Paris VII

voided urine, although voided urine is coming from Paris, France

The authors have presented a compelling argumentfor the potential clinical usefulness of a biomarkerfor unilateral UPJ obstruction by applying a well-known biomarker for unrelated conditions. CA-19-9has been used in several types of malignancy, andwhile the reasons for its being increased in the ob-structed kidney are unclear, the association is clear.In the search for clinical means by which we canidentify children at risk for renal functional deteri-oration in UPJ obstruction several biomarkers havebeen explored, often with great initial enthusiasm,only to fade in apparent usefulness. Much of thisproblem is due to the challenge of correlating bi-omarker changes with clinically useful outcomes.Even agreement on what outcomes are relevant con-tinues to be lacking. Without agreement on how toassess outcomes in UPJ obstruction validating auseful biomarker remains challenging.

The definition of “obstructed” itself remains con-troversial, which represents one of the major limita-tions of this study, in that the fact of undergoingpyeloplasty is not proof of obstruction. The criteriafor surgery vary widely, with no universal gold stan-

there are those with reduced relative uptake andseverely dilated kidneys, just as there are those withnormal relative uptake and limited dilatation. How-ever, this variation can be useful since we see acorrelation between these parameters of severityand the levels of CA-19-9. It is also important torecognize that the definition of obstruction may reston change with time, and in this pilot study theimpact of time is not assessed preoperatively. Thereis the suggestion of an important relationship be-tween CA-19-9 levels and obstructive effect in thatthere is a clear and progressive decline in CA-19-9postoperatively. As we continue to look for robustindicators of renal risk, CA-19-9 may be a useful toolfor clinical application, as well as perhaps providingnovel insight into pathophysiological mechanisms.The authors should be congratulated on this initialeffort and encouraged to pursue this approach fur-ther.

Craig A. Peters

Division of Pediatric UrologyUniversity of Virginia

dard. Clearly in this population of surgical patients Charlottesville, Virginia

We used CA-19-9 in voided urine as a noninvasive,clinically applicable marker for congenital ob-structive nephropathy with significant correlationwith initial renographic function and AP diame-ter. The noninvasive nature of the biomarkeralong with the high sensitivity and specificity for

ducing this method as a surrogate for determiningthe management of congenital hydronephrosis. Al-though voided urine CA-19-9 demonstratedweaker correlation due to the dilutional effect ofthe contralateral kidney, its role is confirmed bythe significant correlation between the intrapelvic

and voided urine levels.

CARBOHYDRATE ANTIGEN 19-9 LEVELS IN URETEROPELVIC JUNCTION OBSTRUCTION2360

The treatment approach for UPJ obstruction iscontroversial. Decisions regarding surgery are mul-tifactorial and a renal scan is not consistent fordetermining the extent of obstruction.1 However,the current strategies are still inadequate for deter-mining the timing of surgical intervention in border-line cases. The generally accepted approach is pres-ervation of renal function. In this study patientswere selected for pyeloplasty according to the clini-cal, ultrasound and DTPA scan evaluations, and/orfailure of conservative management rather thanmere obstruction status.

It is noteworthy that CA-19-9 expression hasbeen demonstrated in different normal tissues in-volving epithelial structure, such as pancreas,gallbladder, colon, renal pelvis and prostate (ref-erences 8 and 11 in article). Urinary hydronephro-

REFERENCES

has been described as a cause of increased CA-19-9 (references 10 and 14 in article). Proposedmechanisms include inflammation and prolifera-tion of noncancerous tissues resulting in enhancedexpression of the marker as a response to theobstructed pelvis. In this setting impairment ofthe CA-19-9 discharge pathway may augment thelevel of the biomarker as well.2

Considering that obstruction status may changewith time, we are currently performing a comple-mentary study on the value of CA-19-9 to distin-guish the high risk from low risk dilated pyelocal-iceal system. The CA-19-9 level in association withserial renography and ultrasonography is used toassess the evolution of renal function in correlationwith CA-19-9 changes. Accordingly the biomarkerwill be a practical valuable adjunctive in decision

sis due to renal stones or urogenital malignancies making for UPJ obstruction management.

1. Sheu JC, Koh CC, Chang PY et al: Ureteropelvic junction obstruction in children: 10 years’ experience in one institution. Pediatr Surg Int 2006; 22: 519.

2. Ito S and Gejyo F: Elevation of serum CA19-9 levels in benign diseases. Intern Med 1999; 38: 840.