- Email: [email protected]
Dynamic Hydrodistention of the Ureteral Orifice: A Novel Grading System With High Interobserver Concordance and Correlation With Vesicoureteral Reflux Grade
Dynamic Hydrodistention of the Ureteral Orifice: A Novel Grading System With High Interobserver Concordance and Correlation With Vesicoureteral Reflux Grade Andrew J. Kirsch,*,† Jonathan D. Kaye, Wolfgang H. Cerwinka, Justin M. Watson, James M. Elmore, Robert H. Lyles, Joseph A. Molitierno and Hal C. Scherz‡ From the Department of Pediatric Urology, Children’s Healthcare of Atlanta, Emory University School of Medicine and Department of Biostatistics, Rollins School of Public Health of Emory University (RHL), Atlanta, Georgia
Abbreviations and Acronyms FS ⫽ failed reflux surgery HD ⫽ hydrodistention HIT ⫽ hydrodistention implantation technique NC ⫽ normal control SS ⫽ successful reflux surgery VCUG ⫽ voiding cystourethrography VUR ⫽ vesicoureteral reflux Study received institutional review board approval. * Correspondence: Children’s Healthcare of Atlanta, Emory University School of Medicine, 5445 Meridian Mark Rd., Suite 420, Atlanta, Georgia 30342 (telephone: 404-256-1502; FAX: 404-256-1268; e-mail: [email protected]). † Financial interest and/or other relationship with Q-Med and Cook. ‡ Financial interest and/or other relationship with Q-Med.
Purpose: We evaluated the usefulness and interobserver concordance of a novel grading system for dynamic ureteral hydrodistention. Materials and Methods: Between May 1, 2002 and July 1, 2008 the hydrodistention grade in 697 ureters was prospectively assigned and recorded, including H0 —no hydrodistention, H1— ureteral orifice open but tunnel not evident, H2— tunnel seen only and H3— extravesical ureter visualized. Specifically 489 refluxing ureters (vesicoureteral reflux group) were compared to 100 normal control ureters (normal control group). Additionally, the posttreatment hydrodistention grade in 56 ureters in which surgery for reflux failed was compared to that in 52 ureters with successful surgery. Hydrodistention grades assigned to an additional 77 ureters by 3 blinded observers were compared to assess the interobserver concordance of this system. Results: Vesicoureteral reflux and hydrodistention grades correlated significantly (p ⬍0.001). Ureters with a higher reflux grade also showed a higher hydrodistention grade. The normal control group (mean ⫾ SEM hydrodistention grade 0.62 ⫾ 0.07) showed a statistically lower hydrodistention grade than the reflux groups (overall mean hydrodistention grade 2.26 ⫾ 0.01). Mean posttreatment hydrodistention grade in the failed reflux surgery group was statistically higher than that in the mean successful reflux surgery group (2.03 ⫾ 0.09 vs 1.33 ⫾ 0.08). By defining the degree of hydrodistention as normal (H0 –H1) and abnormal (H2–H3) the concordance between observers was 95% and 96%. Conclusions: The dynamic hydrodistention classification is a reliable method of evaluating the presence or absence of vesicoureteral reflux as it correlates significantly with radiographic reflux grade. It has high interobserver concordance. Key Words: bladder, ureter, vesico-ureteral reflux, urinary tract infections, polytetrafluoroethylene
1688
www.jurology.com
IN 2002 HIT was introduced for endoscopic treatment for VUR.1 This technique uses HD of the distal ureter to visualize the intraureteral injection site and determines appropriate bulking agent volume. Larger, more capacious ureters show a higher HD grade and require more injection volume.2 A
loss of dynamic HD after implantation implies that a sufficient volume of bulking material has been injected and it provides an objective means of determining the end point of injection. Monitoring HD following injection(s), and using HIT and/or double HIT methods with adequate volumes of bulking ma-
0022-5347/09/1824-1688/0 THE JOURNAL OF UROLOGY® Copyright © 2009 by AMERICAN UROLOGICAL ASSOCIATION
Vol. 182, 1688-1693, October 2009 Printed in U.S.A. DOI:10.1016/j.juro.2009.02.061
DYNAMIC HYDRODISTENTION CLASSIFICATION OF URETERAL ORIFICE
terial are among the key factors for achieving success rates as high as 90% after a single treatment.1,3,4 The technique subsequently evolved into a method that classifies the ureteral orifice based on its degree of HD, which allows a more meaningful description in static and dynamic terms.5 We compared HD grades in ureters with known VUR to HD grades in NC ureters to determine whether HD correlates with VUR grade. Control ureters were in patients with no history of UTI, VUR, ureteropelvic junction obstruction or any other indication for VCUG. We also compared posttreatment HD grades in ureters with SS to grades in those with FS. Furthermore, to confirm the reproducibility of this novel grading system we assessed the interobserver concordance of HD grade among blinded reviewers.
MATERIALS AND METHODS Institutional review board approval was obtained for this study. Between May 1, 2002 and July 1, 2008 a total of 697
Hydrodistention Grade
Description
H0
No orifice distention evident
H1
Orifice opens Intramural tunnel not evident
H2
Orifice opens Intramural tunnel evident Extramural ureter not evident
H3
Orifice opens Extramural ureter evident or ureter can accept cystoscope
1689
ureters in 478 patients with a mean age of 3 years (range 6 months to 12 years) who were undergoing cystoscopy were prospectively assigned an HD grade (fig. 1). A 9.5Fr Wolfe cystoscope with a 5-degree offset lens was used in all cases. Normal saline irrigation was hung 80 cm above the patient. This method has been shown to maintain a bladder pressure of less than 5 cm H2O at low bladder volumes.6 The degree of ureteral orifice and tunnel expansion with HD was evaluated with the bladder less than half full based on expected bladder volume calculated by age. The cystoscope tip was placed at the ureteral orifice and saline was irrigated at full flow. The actual site of saline efflux is 1 cm proximal to the lens tip. All data were collected prospectively. We compared HD grade in 489 ureters in patients with VCUG demonstrated grades I to V primary VUR to HD grades in the NC group of 100 ureters. All patients had had at least 1 culture proven UTI and had undergone at least 2 VCUGs before cystoscopy. Patients with secondary reflux, ie posterior urethral valves, prior surgery, exstrophy etc, were excluded from analysis, as were patients with aberrant anatomy, ie a duplex system, periureteral diverticulum etc. In
Cystoscopic appearance
Figure 1. Ureteral orifice HD classification
Example
1690
DYNAMIC HYDRODISTENTION CLASSIFICATION OF URETERAL ORIFICE
patients with known VUR the location and configuration of the ureteral orifices were also recorded. The NC group had no history of febrile UTI and no VUR. Indications for cystoscopy in the NC group were hematuria, urolithiasis, retrograde pyelogram and/or stent placement, among other diagnoses (see table). In urolithiasis cases only the ureteral orifice contralateral to the calculus was included to eliminate the possibility that edema from the calculus could affect hydrodistention. All patients with of posterior urethral valves, urethral stricture due to trauma and hypospadias, voiding dysfunction and dysuria underwent VCUG, which demonstrated absent VUR. Three attending pediatric urologists (AJK, JME and HCS) contributed patients to this series and assigned a HD grade to their own patients. We also assigned HD grades to an additional 108 ureters that had previously undergone surgery, which was endoscopic in 105 and open in 3. At the time of repeat cystoscopy for re-treatment, which was at least 4 months postoperatively, we compared HD grades in the 56 ureters with FS to the grades in 52 with SS to further validate this grading system. The FS group included ureters that showed VUR on VCUG at cystoscopy for re-treatment, whereas the SS group included those in which every posttreatment VCUG was negative for VUR. The mean ⫾ SEM HD grade in each group was determined. The chi-square test of correlation between 2 ordinal categorical variables was used to determine whether HD grade correlated with VUR grade in each group. Fisher’s exact test was performed to determine whether statistically significant differences in HD grade existed between the VUR groups and the NC group with each grade considered individually and statistical significance considered at p ⬍0.05. Fisher’s exact test was also used to determine whether a statistically significant difference in HD grade existed between the FS group and the SS groups. The NC group had no history of febrile UTI. To establish the interobserver consistency of HD grading 2 novices, that is a year 4 medical student (JMW) and a year 1 pediatric urology fellow (JAM), were instructed in the HD grading system by an experienced pediatric urol-
Indications for cystoscopy in normal control group Indication
No. Pts
No. Ureters
Ureteral calculus* Hematuria evaluation Traumatic or congenital urethral stricture Urethral stricture after hypospadias repair Dysuria/persistent pelvic pain Voiding dysfunction Urinary retention Congenital adrenal hyperplasia/ambiguous genitalia Posterior urethral valves Genital anomalies Other
12 5 5 4 4 4 4 3
12 10 10 8 8 8 8 6
4 6 5
8 12 10
56
100
Totals * Only ureters contralateral to calculi were evaluated.
ogist (AJK) using the cystoscopic video recordings of patients not included in this study. The medical student and fellow were blinded to patient history, including VCUG results, and each graded 77 ureters. HD grades were compared to determine concordance among the 3 observers. Interobserver agreement on the absolute grade assignment (H0 to H3) was compared. Interobserver agreement regarding low grade (H0 –H1) and high grade (H2–H3) designations were also compared. Fisher’s exact test was used to determine the level of statistical significance of this concordance.
RESULTS VUR Group of 489 Ureters The mean HD grade in all 489 ureters was 2.26 ⫾ 0.01. VUR grade was I to V in 49, 160, 189, 68 and 23 patients, respectively. Ureters with grade I VUR were found in patients with higher grade contralateral VUR. The mean HD grade was 2.01 ⫾ 0.10, 2.13 ⫾ 0.06, 2.29 ⫾ 0.06, 2.49 ⫾ 0.09 and 2.77 ⫾ 0.16 for VUR grades I to V, respectively (fig. 2). All cases of VUR demonstrated H1 or greater and the grade corresponded with the VUR grade on VCUG (p ⬍0.05, fig. 3). Of the patients with grades I to III VUR 29%, 15% and 13% had low grade HD (H0 –H1), while 71%, 85% and 87%, respectively, had high grade HD (H2–H3). On the other hand, 5% and 0% of the patients with grades IV and V VUR had low grade HD, while 95% and 100%, respectively, had high grade HD (fig. 3). We also noted that 95% of ureteral orifices with any degree of reflux were laterally displaced in the bladder. The ureteral orifice before HD had a horseshoe configuration in 95% of cases, while it was a golf hole in 4% and normal in 1%. NC Group of 100 Ureters The mean HD grade in this group was 0.62 ⫾ 0.07. The group consisted of 56 patients (100 ureters). The table lists the indications for cystoscopy. HD grade was H0 to H3 in 49%, 40%, 11% and 0% of cases, respectively. Notably 89% of the cases did not demonstrate significant HD, including H0 in 49% and H1 in 40% (fig. 3). The NC group showed a significantly lower HD grade than any of the other groups mean 0.62 ⫾ 0.07, p ⬍0.05). Posttreatment Groups of 108 Ureters HD grade was determined in patients who had previously undergone antireflux therapy. Of the 108 ureters in this group 105 underwent endoscopic treatment, while 3 underwent open surgery. Of the latter ureters 2 were extravesical and 1 was intravesical, and surgery failed in all 3. Posttreatment HD grade in the FS group of 56 ureters was also statistically higher than that in the SS group of 52 (mean 2.03 ⫾ 0.09 vs 1.33 ⫾ 0.08). It was similar to the mean HD grade in the grade I VUR group (2.01 ⫾
DYNAMIC HYDRODISTENTION CLASSIFICATION OF URETERAL ORIFICE
3 2.5
2.29 +0.06
2.13 +0.06
2.01 +0.11
2.49 +0.10
1691
2.77 +0.16
2 1.5 1
0.62 +0.07
0.5 0
Controls*
VUR I
VUR II
VUR III
VUR IV
VUR V
Figure 2. Mean HD grade (y-axis) in VUR groups. Values above bars represent mean ⫾ SEM. Control vs VUR I to V p ⬍0.05. Mean HD grade correlated with VUR grade for VUR grades I to V (p ⬍0.05).
0.10). There were no differences in patient gender or age between the VUR group and NC groups, or between the FS and SS groups. Interobserver Agreement The absolute numerical (H0 –H3) concordance between the grades assigned by the attending physician and the fellow was 78% and clinical (H0 –H1 vs H2– H3) concordance was 91%. Similarly the absolute concordance between the medical student and the attending physician was 79% with 96% clinical concordance. Finally, the absolute concordance between the fellow and the medical student was 82% with clinical concordance on 95% of the ureters (fig. 4). Agreement in absolute and clinical concordance between the attending physician, the fellow and the medical student was statistically significant (p ⬍0.05).
DISCUSSION Beginning in the late 1960s the ureteral orifice has been classified to determine the potential for spontaneous reflux resolution on surveillance cystoscopy
HD Grade (%)
H0
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%
11%
26%
H1
28%
H2
H3
42%
54%
77%
40% 43%
57% 45%
49%
41% 30%
Controls N=100
15%
I
N=49
II
N=160
13%
III
N=189
23% 5%
IV
N=68
V
N=23
VUR Grades
Figure 3. Percent of HD grade in each NC and VUR group. All ureters in VUR group had some degree of hydrodistention.
Figure 4. Concordance between blinded observers of HD grading. 1, attending pediatric urologist. 2, pediatric urology fellow. 3, year 3 medical student. Agree Precisely, absolute agreement, ie H0, H1, H2 or H3. Agree Clinically, clinical concordance, ie H0 –H1 vs H2–H3. Absolute and clinical concordance was significant for all 3 pairings (p ⬍0.05).
or voiding cystography and the need to correct the nonrefluxing contralateral ureter at open surgery. Ureters were classified based on their configuration, location and tunnel length to predict renal dysplasia and VUR resolution.7–9 These classification schemes were often cumbersome and not reproducible. Their usefulness has since been refuted and they currently have little role in VUR evaluation and management.10 It is not our practice to rely on ureteral orifice location or configuration to determine the need for contralateral treatment during unilateral antireflux surgery. We have noted a change in the configuration of the orifice upon ureteral hydrodistention. For example, a horseshoe ureter may appear as a golf hole with dynamic HD. Indeed, the ubiquity of the laterally displaced horseshoe orifice in our series (95%) supports the findings by Bellinger and Duckett of the futility of a grading system based on orifice shape and location.10 These findings also suggest that HD is a more accurate and objective cystoscopic measure of ureterovesical junction competency. The data in the current study, which demonstrate a significant correlation between HD and VUR, further support this contention. The statistically significant difference in HD grade in SS and FS cases in our series underscores the validity of this grading system. It is noteworthy that the difference applied to all antireflux therapies. That is, intravesically, extravesically and endoscopically treated ureters in which reflux persisted or recurred after treatment had higher posttreatment HD grades as a group than their successfully treated counterparts as a group. With respect to interobserver agreement the statistically significant absolute and clinical concordance among all 3 graders indicates that the HD grading system is learned and adopted quite readily. Its interobserver concordance proved extremely reliable when
1692
DYNAMIC HYDRODISTENTION CLASSIFICATION OF URETERAL ORIFICE
determining high vs low grade HD. It is likely that these differences, rather than H0 –H3, will ultimately prove clinically important, just as the differences between specific grades of VUR are less clinically significant than the differences between low and high grade VUR. A few limitations of the current study warrant mention. The occasional difficulty in visualizing the ureteral orifice in a head-on manner for lateral ureters may represent a technical limitation of our proposed grading system. Moreover, the degree of bladder filling and the height of the irrigation column may affect saline flow and the degree of HD. However, Edmondson et al reported that these factors have no bearing on VUR when positioning the instillation of contrast medium.11 Another potential limitation relates to our NC group. Because some of these patients, ie those with calculus, had no indication for VCUG, we do not know with certainty the VUR status of all patients in this group. Therefore, we assumed that the prevalence of reflux in the NC group reflected that in the general population, which is generally accepted to be approximately 1%. Finally, surgeons were not blinded to VUR grade when performing cystoscopy and assigning a HD grade. The high concordance of the independent interobserver variability part of this study and the fact
that HD grades in this series were assigned to such a large number of ureters by multiple surgeons makes potential bias unlikely. We are currently investigating novel clinical applications of HD. Specifically we are measuring the HD grade in nonrefluxing contralateral ureters, hypothesizing that those in which de novo contralateral VUR develops have a higher grade of HD than those that do not. We are also investigating the HD grade in ureters that are VCUG negative and positive for positioning the instillation of contrast medium as well as those in patients who experience recurrent febrile UTIs despite repeatedly negative VCUGs. The advisability of empirically treating all of these groups based on HD grade will be the subject of future study.
CONCLUSIONS Dynamic HD classification of the ureter is a highly reliable grading system that provides a simplified and meaningful way for urologists to communicate cystoscopic findings as they correlate with the degree of VUR. Ureters with hydrodistention grades H0 –H1 should be considered normal and at low risk for VUR, while H2–H3 ureters should be considered abnormal because they are more likely to reflux.
REFERENCES 1. Kirsch AJ, Perez-Brayfield M, Smith EA et al: The modified sting procedure to correct vesicoureteral reflux: improved results with submucosal implantation within the intramural ureter. J Urol 2004; 171: 2413. 2. McMann LP, Scherz HC and Kirsch AJ: Long-term preservation of dextranomer/hyaluronic acid copolymer implants after endoscopic treatment of vesicoureteral reflux in children: a sonographic volumetric analysis. J Urol 2007; 177: 316. 3. Yu RN and Roth DR: Treatment of vesicoureteral reflux using endoscopic injection of nonanimal stabilized hyaluronic acid/dextranomer gel: initial experience in pediatric patients by a single surgeon. Pediatrics 2006; 118: 698.
4. Molitierno JA, Scherz HC and Kirsch AJ: Endoscopic treatment of vesicoureteral reflux using dextranomer hyaluronic acid copolymer. J Pediatr Urol 2008; 4: 221.
8. King LR, Kazmi SO and Belman AB: Natural history of vesicoureteral reflux. Outcome of a trial of nonoperative therapy. Urol Clin North Am 1974; 1: 441.
5. Cerwinka WH, Scherz HC and Kirsch AJ: Dynamic hydrodistention classification of the ureter and the double HIT method to correct vesicoureteral reflux. Arch Esp Urol 2008; p 61.
9. Mackie GG and Stephens FD: Duplex kidneys: a correlation of renal dysplasia with position of the ureteral orifice. J Urol 1975; 114: 274.
6. Hagerty JA, Maizels M and Cheng EY: The PIC cystogram: its place in the treatment algorithm of recurrent febrile UTIs. Adv Urol 2008; 2008: 763620. 7. Lyon RP, Marshall S and Tanagho EA: The ureteral orifice: its configuration and competency. J Urol 1969; 102: 504.
10. Bellinger MF and Duckett JW: Vesicoureteral reflux: a comparison of non-surgical and surgical management. Contrib Nephrol 1984; 39: 81. 11. Edmondson JD, Maizels M, Alpert SA et al: Multi-institutional experience with PIC cystography- incidence of occult vesicoureteral reflux in children with febrile urinary tract infections. Urology 2006; 67: 608.
EDITORIAL COMMENT These authors aim to validate their system of ureteral orifice HD from 2 perspectives, that is to 1) demonstrate the association of ureteral HD with VUR and 2) show the interobserver concordance of ureteral HD grading. They are continuing to thoroughly study their widely used system. Ultimately they conclude that dynamic HD classification is a reliable method of evaluating VUR and it has high interobserver concordance.
The interrogation of interobserver agreement was done in multiple comparison groups based on disparate experience levels. When grouping low vs high grade HD, which they believe is the clinically relevant grouping scheme, the concordance rate was 91% to 96%. In reference to their first aim, HD indeed correlated strongly with VUR. However, despite their claim that HD grade statistically corresponded with
DYNAMIC HYDRODISTENTION CLASSIFICATION OF URETERAL ORIFICE
VUR grade (fig. 2), I was more impressed with the apparent promiscuity of high grade (H2–H3) HD. Specifically in children with no VUR the rate of high grade HD was 11%, 69% of those with grade I VUR had high grade HD and in patients with grade II VUR that rate increased to 85%. This point is relevant to an ever increasingly common question, that is whether we should treat a contralateral nonrefluxing ureter at the time of endoscopic treatment for unilateral VUR if we evaluate that orifice with HD and find an abnormality? Ultimately the fundamental question in the context of this article is which HD grade requires treatment? The authors conclude that grades H0-H1 should be considered normal and yet 13% of the children with grade III VUR and 5% with grade IV VUR had H1 HD. Given the data presented, I would still have a hard time with that decision and would find it difficult to base
1693
my decision conclusively on HD grade. The authors state that this is the subject of an ongoing study and I suggest that we must wait for those data to be better informed in this particular clinical situation. Finally, the authors state that the statistically significant difference between HD grades in SS and FS cases underscores the validity of the grading system and the difference applied to all antirefluxing therapies (open and endoscopic treatments). Because only 3 of the 108 patients in this cohort (posttreatment group) underwent open surgery, this conclusion cannot be made. It would seem more appropriate to consider the validity of HD grading only in the context of endoscopic treatment for VUR. Carlos Estrada Children’s Hospital Boston Boston, Massachusetts
REPLY BY AUTHORS Not only do we agree that high grade HD is somewhat ubiquitous to all VUR grades, but we believe that this promiscuity may have important clinical significance. We know from previous studies that predictions about the behavior of ureters is unreliable. Why, for instance does grade IV reflux resolve spontaneously and without incident in some cases, while in others it does not? Why does grade I reflux persist radiographically and continue to be associated with clinical problems in some patients while in most this degree of reflux is insignificant? Why do some children with repeatedly negative VCUGs experience recurrent bouts of pyelonephritis? We believe that HD may help explain these situations. The proportion of each VUR grade that did not resolve spontaneously in the International Reflux Study resembles the proportion of each VUR group that demonstrated HD grade 3 in our series. Furthermore, if one were to use abnormal hydrodistention as a criterion to treat contralateral ureters that did not show reflux on VCUG, more than 60% would be abnormal and, thus, candidates for treatment.
Our data suggest that VUR is likely a bilateral process. Our finding that no ureters in the control group, as well as most nonrefluxing contralateral ureters (reference 5 in article), had abnormal HD supports this hypothesis. VCUG is not always diagnostic for VUR, and in many patients being observed VUR tends to change sides, resolve and recur. VCUG is inexact at diagnosing reflux and yet it is the gold standard for making this diagnosis. With this in mind, the strong correlation between HD and VUR grade justifies nontreatment of the contralateral ureters with low grade HD. Conversely, treating all ureters with a higher HD grade may result in over treatment of some ureters. However, because of the negligible risk associated with endoscopic injection we are comfortable using HD to determine which contralateral ureters to inject. In fact, our recent data show that new contralateral VUR has been eliminated using this approach. The cost and clinical significance of treating ureters proactively in this manner are under investigation.
Abbreviations and Acronyms FS ⫽ failed reflux surgery HD ⫽ hydrodistention HIT ⫽ hydrodistention implantation technique NC ⫽ normal control SS ⫽ successful reflux surgery VCUG ⫽ voiding cystourethrography VUR ⫽ vesicoureteral reflux Study received institutional review board approval. * Correspondence: Children’s Healthcare of Atlanta, Emory University School of Medicine, 5445 Meridian Mark Rd., Suite 420, Atlanta, Georgia 30342 (telephone: 404-256-1502; FAX: 404-256-1268; e-mail: [email protected]). † Financial interest and/or other relationship with Q-Med and Cook. ‡ Financial interest and/or other relationship with Q-Med.
Purpose: We evaluated the usefulness and interobserver concordance of a novel grading system for dynamic ureteral hydrodistention. Materials and Methods: Between May 1, 2002 and July 1, 2008 the hydrodistention grade in 697 ureters was prospectively assigned and recorded, including H0 —no hydrodistention, H1— ureteral orifice open but tunnel not evident, H2— tunnel seen only and H3— extravesical ureter visualized. Specifically 489 refluxing ureters (vesicoureteral reflux group) were compared to 100 normal control ureters (normal control group). Additionally, the posttreatment hydrodistention grade in 56 ureters in which surgery for reflux failed was compared to that in 52 ureters with successful surgery. Hydrodistention grades assigned to an additional 77 ureters by 3 blinded observers were compared to assess the interobserver concordance of this system. Results: Vesicoureteral reflux and hydrodistention grades correlated significantly (p ⬍0.001). Ureters with a higher reflux grade also showed a higher hydrodistention grade. The normal control group (mean ⫾ SEM hydrodistention grade 0.62 ⫾ 0.07) showed a statistically lower hydrodistention grade than the reflux groups (overall mean hydrodistention grade 2.26 ⫾ 0.01). Mean posttreatment hydrodistention grade in the failed reflux surgery group was statistically higher than that in the mean successful reflux surgery group (2.03 ⫾ 0.09 vs 1.33 ⫾ 0.08). By defining the degree of hydrodistention as normal (H0 –H1) and abnormal (H2–H3) the concordance between observers was 95% and 96%. Conclusions: The dynamic hydrodistention classification is a reliable method of evaluating the presence or absence of vesicoureteral reflux as it correlates significantly with radiographic reflux grade. It has high interobserver concordance. Key Words: bladder, ureter, vesico-ureteral reflux, urinary tract infections, polytetrafluoroethylene
1688
www.jurology.com
IN 2002 HIT was introduced for endoscopic treatment for VUR.1 This technique uses HD of the distal ureter to visualize the intraureteral injection site and determines appropriate bulking agent volume. Larger, more capacious ureters show a higher HD grade and require more injection volume.2 A
loss of dynamic HD after implantation implies that a sufficient volume of bulking material has been injected and it provides an objective means of determining the end point of injection. Monitoring HD following injection(s), and using HIT and/or double HIT methods with adequate volumes of bulking ma-
0022-5347/09/1824-1688/0 THE JOURNAL OF UROLOGY® Copyright © 2009 by AMERICAN UROLOGICAL ASSOCIATION
Vol. 182, 1688-1693, October 2009 Printed in U.S.A. DOI:10.1016/j.juro.2009.02.061
DYNAMIC HYDRODISTENTION CLASSIFICATION OF URETERAL ORIFICE
terial are among the key factors for achieving success rates as high as 90% after a single treatment.1,3,4 The technique subsequently evolved into a method that classifies the ureteral orifice based on its degree of HD, which allows a more meaningful description in static and dynamic terms.5 We compared HD grades in ureters with known VUR to HD grades in NC ureters to determine whether HD correlates with VUR grade. Control ureters were in patients with no history of UTI, VUR, ureteropelvic junction obstruction or any other indication for VCUG. We also compared posttreatment HD grades in ureters with SS to grades in those with FS. Furthermore, to confirm the reproducibility of this novel grading system we assessed the interobserver concordance of HD grade among blinded reviewers.
MATERIALS AND METHODS Institutional review board approval was obtained for this study. Between May 1, 2002 and July 1, 2008 a total of 697
Hydrodistention Grade
Description
H0
No orifice distention evident
H1
Orifice opens Intramural tunnel not evident
H2
Orifice opens Intramural tunnel evident Extramural ureter not evident
H3
Orifice opens Extramural ureter evident or ureter can accept cystoscope
1689
ureters in 478 patients with a mean age of 3 years (range 6 months to 12 years) who were undergoing cystoscopy were prospectively assigned an HD grade (fig. 1). A 9.5Fr Wolfe cystoscope with a 5-degree offset lens was used in all cases. Normal saline irrigation was hung 80 cm above the patient. This method has been shown to maintain a bladder pressure of less than 5 cm H2O at low bladder volumes.6 The degree of ureteral orifice and tunnel expansion with HD was evaluated with the bladder less than half full based on expected bladder volume calculated by age. The cystoscope tip was placed at the ureteral orifice and saline was irrigated at full flow. The actual site of saline efflux is 1 cm proximal to the lens tip. All data were collected prospectively. We compared HD grade in 489 ureters in patients with VCUG demonstrated grades I to V primary VUR to HD grades in the NC group of 100 ureters. All patients had had at least 1 culture proven UTI and had undergone at least 2 VCUGs before cystoscopy. Patients with secondary reflux, ie posterior urethral valves, prior surgery, exstrophy etc, were excluded from analysis, as were patients with aberrant anatomy, ie a duplex system, periureteral diverticulum etc. In
Cystoscopic appearance
Figure 1. Ureteral orifice HD classification
Example
1690
DYNAMIC HYDRODISTENTION CLASSIFICATION OF URETERAL ORIFICE
patients with known VUR the location and configuration of the ureteral orifices were also recorded. The NC group had no history of febrile UTI and no VUR. Indications for cystoscopy in the NC group were hematuria, urolithiasis, retrograde pyelogram and/or stent placement, among other diagnoses (see table). In urolithiasis cases only the ureteral orifice contralateral to the calculus was included to eliminate the possibility that edema from the calculus could affect hydrodistention. All patients with of posterior urethral valves, urethral stricture due to trauma and hypospadias, voiding dysfunction and dysuria underwent VCUG, which demonstrated absent VUR. Three attending pediatric urologists (AJK, JME and HCS) contributed patients to this series and assigned a HD grade to their own patients. We also assigned HD grades to an additional 108 ureters that had previously undergone surgery, which was endoscopic in 105 and open in 3. At the time of repeat cystoscopy for re-treatment, which was at least 4 months postoperatively, we compared HD grades in the 56 ureters with FS to the grades in 52 with SS to further validate this grading system. The FS group included ureters that showed VUR on VCUG at cystoscopy for re-treatment, whereas the SS group included those in which every posttreatment VCUG was negative for VUR. The mean ⫾ SEM HD grade in each group was determined. The chi-square test of correlation between 2 ordinal categorical variables was used to determine whether HD grade correlated with VUR grade in each group. Fisher’s exact test was performed to determine whether statistically significant differences in HD grade existed between the VUR groups and the NC group with each grade considered individually and statistical significance considered at p ⬍0.05. Fisher’s exact test was also used to determine whether a statistically significant difference in HD grade existed between the FS group and the SS groups. The NC group had no history of febrile UTI. To establish the interobserver consistency of HD grading 2 novices, that is a year 4 medical student (JMW) and a year 1 pediatric urology fellow (JAM), were instructed in the HD grading system by an experienced pediatric urol-
Indications for cystoscopy in normal control group Indication
No. Pts
No. Ureters
Ureteral calculus* Hematuria evaluation Traumatic or congenital urethral stricture Urethral stricture after hypospadias repair Dysuria/persistent pelvic pain Voiding dysfunction Urinary retention Congenital adrenal hyperplasia/ambiguous genitalia Posterior urethral valves Genital anomalies Other
12 5 5 4 4 4 4 3
12 10 10 8 8 8 8 6
4 6 5
8 12 10
56
100
Totals * Only ureters contralateral to calculi were evaluated.
ogist (AJK) using the cystoscopic video recordings of patients not included in this study. The medical student and fellow were blinded to patient history, including VCUG results, and each graded 77 ureters. HD grades were compared to determine concordance among the 3 observers. Interobserver agreement on the absolute grade assignment (H0 to H3) was compared. Interobserver agreement regarding low grade (H0 –H1) and high grade (H2–H3) designations were also compared. Fisher’s exact test was used to determine the level of statistical significance of this concordance.
RESULTS VUR Group of 489 Ureters The mean HD grade in all 489 ureters was 2.26 ⫾ 0.01. VUR grade was I to V in 49, 160, 189, 68 and 23 patients, respectively. Ureters with grade I VUR were found in patients with higher grade contralateral VUR. The mean HD grade was 2.01 ⫾ 0.10, 2.13 ⫾ 0.06, 2.29 ⫾ 0.06, 2.49 ⫾ 0.09 and 2.77 ⫾ 0.16 for VUR grades I to V, respectively (fig. 2). All cases of VUR demonstrated H1 or greater and the grade corresponded with the VUR grade on VCUG (p ⬍0.05, fig. 3). Of the patients with grades I to III VUR 29%, 15% and 13% had low grade HD (H0 –H1), while 71%, 85% and 87%, respectively, had high grade HD (H2–H3). On the other hand, 5% and 0% of the patients with grades IV and V VUR had low grade HD, while 95% and 100%, respectively, had high grade HD (fig. 3). We also noted that 95% of ureteral orifices with any degree of reflux were laterally displaced in the bladder. The ureteral orifice before HD had a horseshoe configuration in 95% of cases, while it was a golf hole in 4% and normal in 1%. NC Group of 100 Ureters The mean HD grade in this group was 0.62 ⫾ 0.07. The group consisted of 56 patients (100 ureters). The table lists the indications for cystoscopy. HD grade was H0 to H3 in 49%, 40%, 11% and 0% of cases, respectively. Notably 89% of the cases did not demonstrate significant HD, including H0 in 49% and H1 in 40% (fig. 3). The NC group showed a significantly lower HD grade than any of the other groups mean 0.62 ⫾ 0.07, p ⬍0.05). Posttreatment Groups of 108 Ureters HD grade was determined in patients who had previously undergone antireflux therapy. Of the 108 ureters in this group 105 underwent endoscopic treatment, while 3 underwent open surgery. Of the latter ureters 2 were extravesical and 1 was intravesical, and surgery failed in all 3. Posttreatment HD grade in the FS group of 56 ureters was also statistically higher than that in the SS group of 52 (mean 2.03 ⫾ 0.09 vs 1.33 ⫾ 0.08). It was similar to the mean HD grade in the grade I VUR group (2.01 ⫾
DYNAMIC HYDRODISTENTION CLASSIFICATION OF URETERAL ORIFICE
3 2.5
2.29 +0.06
2.13 +0.06
2.01 +0.11
2.49 +0.10
1691
2.77 +0.16
2 1.5 1
0.62 +0.07
0.5 0
Controls*
VUR I
VUR II
VUR III
VUR IV
VUR V
Figure 2. Mean HD grade (y-axis) in VUR groups. Values above bars represent mean ⫾ SEM. Control vs VUR I to V p ⬍0.05. Mean HD grade correlated with VUR grade for VUR grades I to V (p ⬍0.05).
0.10). There were no differences in patient gender or age between the VUR group and NC groups, or between the FS and SS groups. Interobserver Agreement The absolute numerical (H0 –H3) concordance between the grades assigned by the attending physician and the fellow was 78% and clinical (H0 –H1 vs H2– H3) concordance was 91%. Similarly the absolute concordance between the medical student and the attending physician was 79% with 96% clinical concordance. Finally, the absolute concordance between the fellow and the medical student was 82% with clinical concordance on 95% of the ureters (fig. 4). Agreement in absolute and clinical concordance between the attending physician, the fellow and the medical student was statistically significant (p ⬍0.05).
DISCUSSION Beginning in the late 1960s the ureteral orifice has been classified to determine the potential for spontaneous reflux resolution on surveillance cystoscopy
HD Grade (%)
H0
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%
11%
26%
H1
28%
H2
H3
42%
54%
77%
40% 43%
57% 45%
49%
41% 30%
Controls N=100
15%
I
N=49
II
N=160
13%
III
N=189
23% 5%
IV
N=68
V
N=23
VUR Grades
Figure 3. Percent of HD grade in each NC and VUR group. All ureters in VUR group had some degree of hydrodistention.
Figure 4. Concordance between blinded observers of HD grading. 1, attending pediatric urologist. 2, pediatric urology fellow. 3, year 3 medical student. Agree Precisely, absolute agreement, ie H0, H1, H2 or H3. Agree Clinically, clinical concordance, ie H0 –H1 vs H2–H3. Absolute and clinical concordance was significant for all 3 pairings (p ⬍0.05).
or voiding cystography and the need to correct the nonrefluxing contralateral ureter at open surgery. Ureters were classified based on their configuration, location and tunnel length to predict renal dysplasia and VUR resolution.7–9 These classification schemes were often cumbersome and not reproducible. Their usefulness has since been refuted and they currently have little role in VUR evaluation and management.10 It is not our practice to rely on ureteral orifice location or configuration to determine the need for contralateral treatment during unilateral antireflux surgery. We have noted a change in the configuration of the orifice upon ureteral hydrodistention. For example, a horseshoe ureter may appear as a golf hole with dynamic HD. Indeed, the ubiquity of the laterally displaced horseshoe orifice in our series (95%) supports the findings by Bellinger and Duckett of the futility of a grading system based on orifice shape and location.10 These findings also suggest that HD is a more accurate and objective cystoscopic measure of ureterovesical junction competency. The data in the current study, which demonstrate a significant correlation between HD and VUR, further support this contention. The statistically significant difference in HD grade in SS and FS cases in our series underscores the validity of this grading system. It is noteworthy that the difference applied to all antireflux therapies. That is, intravesically, extravesically and endoscopically treated ureters in which reflux persisted or recurred after treatment had higher posttreatment HD grades as a group than their successfully treated counterparts as a group. With respect to interobserver agreement the statistically significant absolute and clinical concordance among all 3 graders indicates that the HD grading system is learned and adopted quite readily. Its interobserver concordance proved extremely reliable when
1692
DYNAMIC HYDRODISTENTION CLASSIFICATION OF URETERAL ORIFICE
determining high vs low grade HD. It is likely that these differences, rather than H0 –H3, will ultimately prove clinically important, just as the differences between specific grades of VUR are less clinically significant than the differences between low and high grade VUR. A few limitations of the current study warrant mention. The occasional difficulty in visualizing the ureteral orifice in a head-on manner for lateral ureters may represent a technical limitation of our proposed grading system. Moreover, the degree of bladder filling and the height of the irrigation column may affect saline flow and the degree of HD. However, Edmondson et al reported that these factors have no bearing on VUR when positioning the instillation of contrast medium.11 Another potential limitation relates to our NC group. Because some of these patients, ie those with calculus, had no indication for VCUG, we do not know with certainty the VUR status of all patients in this group. Therefore, we assumed that the prevalence of reflux in the NC group reflected that in the general population, which is generally accepted to be approximately 1%. Finally, surgeons were not blinded to VUR grade when performing cystoscopy and assigning a HD grade. The high concordance of the independent interobserver variability part of this study and the fact
that HD grades in this series were assigned to such a large number of ureters by multiple surgeons makes potential bias unlikely. We are currently investigating novel clinical applications of HD. Specifically we are measuring the HD grade in nonrefluxing contralateral ureters, hypothesizing that those in which de novo contralateral VUR develops have a higher grade of HD than those that do not. We are also investigating the HD grade in ureters that are VCUG negative and positive for positioning the instillation of contrast medium as well as those in patients who experience recurrent febrile UTIs despite repeatedly negative VCUGs. The advisability of empirically treating all of these groups based on HD grade will be the subject of future study.
CONCLUSIONS Dynamic HD classification of the ureter is a highly reliable grading system that provides a simplified and meaningful way for urologists to communicate cystoscopic findings as they correlate with the degree of VUR. Ureters with hydrodistention grades H0 –H1 should be considered normal and at low risk for VUR, while H2–H3 ureters should be considered abnormal because they are more likely to reflux.
REFERENCES 1. Kirsch AJ, Perez-Brayfield M, Smith EA et al: The modified sting procedure to correct vesicoureteral reflux: improved results with submucosal implantation within the intramural ureter. J Urol 2004; 171: 2413. 2. McMann LP, Scherz HC and Kirsch AJ: Long-term preservation of dextranomer/hyaluronic acid copolymer implants after endoscopic treatment of vesicoureteral reflux in children: a sonographic volumetric analysis. J Urol 2007; 177: 316. 3. Yu RN and Roth DR: Treatment of vesicoureteral reflux using endoscopic injection of nonanimal stabilized hyaluronic acid/dextranomer gel: initial experience in pediatric patients by a single surgeon. Pediatrics 2006; 118: 698.
4. Molitierno JA, Scherz HC and Kirsch AJ: Endoscopic treatment of vesicoureteral reflux using dextranomer hyaluronic acid copolymer. J Pediatr Urol 2008; 4: 221.
8. King LR, Kazmi SO and Belman AB: Natural history of vesicoureteral reflux. Outcome of a trial of nonoperative therapy. Urol Clin North Am 1974; 1: 441.
5. Cerwinka WH, Scherz HC and Kirsch AJ: Dynamic hydrodistention classification of the ureter and the double HIT method to correct vesicoureteral reflux. Arch Esp Urol 2008; p 61.
9. Mackie GG and Stephens FD: Duplex kidneys: a correlation of renal dysplasia with position of the ureteral orifice. J Urol 1975; 114: 274.
6. Hagerty JA, Maizels M and Cheng EY: The PIC cystogram: its place in the treatment algorithm of recurrent febrile UTIs. Adv Urol 2008; 2008: 763620. 7. Lyon RP, Marshall S and Tanagho EA: The ureteral orifice: its configuration and competency. J Urol 1969; 102: 504.
10. Bellinger MF and Duckett JW: Vesicoureteral reflux: a comparison of non-surgical and surgical management. Contrib Nephrol 1984; 39: 81. 11. Edmondson JD, Maizels M, Alpert SA et al: Multi-institutional experience with PIC cystography- incidence of occult vesicoureteral reflux in children with febrile urinary tract infections. Urology 2006; 67: 608.
EDITORIAL COMMENT These authors aim to validate their system of ureteral orifice HD from 2 perspectives, that is to 1) demonstrate the association of ureteral HD with VUR and 2) show the interobserver concordance of ureteral HD grading. They are continuing to thoroughly study their widely used system. Ultimately they conclude that dynamic HD classification is a reliable method of evaluating VUR and it has high interobserver concordance.
The interrogation of interobserver agreement was done in multiple comparison groups based on disparate experience levels. When grouping low vs high grade HD, which they believe is the clinically relevant grouping scheme, the concordance rate was 91% to 96%. In reference to their first aim, HD indeed correlated strongly with VUR. However, despite their claim that HD grade statistically corresponded with
DYNAMIC HYDRODISTENTION CLASSIFICATION OF URETERAL ORIFICE
VUR grade (fig. 2), I was more impressed with the apparent promiscuity of high grade (H2–H3) HD. Specifically in children with no VUR the rate of high grade HD was 11%, 69% of those with grade I VUR had high grade HD and in patients with grade II VUR that rate increased to 85%. This point is relevant to an ever increasingly common question, that is whether we should treat a contralateral nonrefluxing ureter at the time of endoscopic treatment for unilateral VUR if we evaluate that orifice with HD and find an abnormality? Ultimately the fundamental question in the context of this article is which HD grade requires treatment? The authors conclude that grades H0-H1 should be considered normal and yet 13% of the children with grade III VUR and 5% with grade IV VUR had H1 HD. Given the data presented, I would still have a hard time with that decision and would find it difficult to base
1693
my decision conclusively on HD grade. The authors state that this is the subject of an ongoing study and I suggest that we must wait for those data to be better informed in this particular clinical situation. Finally, the authors state that the statistically significant difference between HD grades in SS and FS cases underscores the validity of the grading system and the difference applied to all antirefluxing therapies (open and endoscopic treatments). Because only 3 of the 108 patients in this cohort (posttreatment group) underwent open surgery, this conclusion cannot be made. It would seem more appropriate to consider the validity of HD grading only in the context of endoscopic treatment for VUR. Carlos Estrada Children’s Hospital Boston Boston, Massachusetts
REPLY BY AUTHORS Not only do we agree that high grade HD is somewhat ubiquitous to all VUR grades, but we believe that this promiscuity may have important clinical significance. We know from previous studies that predictions about the behavior of ureters is unreliable. Why, for instance does grade IV reflux resolve spontaneously and without incident in some cases, while in others it does not? Why does grade I reflux persist radiographically and continue to be associated with clinical problems in some patients while in most this degree of reflux is insignificant? Why do some children with repeatedly negative VCUGs experience recurrent bouts of pyelonephritis? We believe that HD may help explain these situations. The proportion of each VUR grade that did not resolve spontaneously in the International Reflux Study resembles the proportion of each VUR group that demonstrated HD grade 3 in our series. Furthermore, if one were to use abnormal hydrodistention as a criterion to treat contralateral ureters that did not show reflux on VCUG, more than 60% would be abnormal and, thus, candidates for treatment.
Our data suggest that VUR is likely a bilateral process. Our finding that no ureters in the control group, as well as most nonrefluxing contralateral ureters (reference 5 in article), had abnormal HD supports this hypothesis. VCUG is not always diagnostic for VUR, and in many patients being observed VUR tends to change sides, resolve and recur. VCUG is inexact at diagnosing reflux and yet it is the gold standard for making this diagnosis. With this in mind, the strong correlation between HD and VUR grade justifies nontreatment of the contralateral ureters with low grade HD. Conversely, treating all ureters with a higher HD grade may result in over treatment of some ureters. However, because of the negligible risk associated with endoscopic injection we are comfortable using HD to determine which contralateral ureters to inject. In fact, our recent data show that new contralateral VUR has been eliminated using this approach. The cost and clinical significance of treating ureters proactively in this manner are under investigation.