Abstract
Background and objectives
Standard clinical assessments do not predict surgical intervention in patients with a moderate degree of upper tract hydronephrosis. This study investigated whether combined measures of renal calyceal dilation and anteroposterior diameter (APD) of the renal pelvis at the first postnatal ultrasound better predict surgical intervention beyond standard assessments of the APD or Society of Fetal Urology (SFU) grading system.
Design, setting, participants, & measurements
A retrospective cohort of 348 children with antenatal hydronephrosis followed from 2003 to 2013 were studied. Using Cox regression, the risk for surgery by APD, SFU, and combined grading on the basis of the first postnatal ultrasound was calculated. The predictive capability of each grading system for surgery was determined by calculating the positive likelihood ratio (LR+).
Results
The combination of APD≥6–9 mm and diffuse caliectasis had a hazard ratio (HR) of 19.5 (95% confidence interval [95% CI], 3.94 to 96.9) versus 0.59 (95% CI, 0.05 to 6.53) for APD≥6–9 mm alone and a similar risk of 8.9 for SFU grade 3 (95% CI, 3.84 to 20.9). The combination of APD≥9–15 mm and diffuse caliectasis had an HR of 18.7 (95% CI, 4.36 to 80.4) versus 1.75 (95% CI, 0.29 to 10.5) for APD≥9–15 mm alone. The LR+ for surgery for diffuse caliectasis and APD≥6–9 mm was higher than for APD≥6–9 mm alone (HR=2.62; 95% CI, 0.87 to 7.94 versus HR=0.04; 95% CI, 0.01 to 0.32) and was higher for APD≥9–15 mm and diffuse caliectasis than APD≥9–15 mm alone (HR=2.0; 95% CI, 1.15 to 3.45 versus HR=0.14; 95% CI, 0.04 to 0.43). Both combined groups of moderate hydronephrosis (APD≥6–9 mm or ≥9–15 mm with diffuse caliectasis) had only slightly higher LR+ than SFU grade 3 (HR=1.89; 95% CI, 1.17 to 3.05).
Conclusions
These results suggest a grading system combining APD and diffuse caliectasis distinguishes those children with moderate degrees of upper tract hydronephrosis that are at higher risk of surgery.
Keywords: antenatal hydronephrosis, congenital hydronephrosis, grading system, pelviectasis, ureteropelvic junction obstruction
Introduction
Antenatal hydronephrosis (ANH), defined as dilation of the renal pelvis and/or calyces, occurs in 1%–5% of pregnancies and is the most common abnormality diagnosed in utero (1–4). The severity of ANH in utero and during postnatal life is variable, with a subset of patients requiring surgical intervention. A critical decision in clinical management of these patients is the need for surgical intervention postnatally.
Renal ultrasound is widely accepted as a noninvasive instrument for monitoring hydronephrosis in the postnatal period. The Society for Fetal Urology (SFU) grading system (5,6) and the measurement of the anteroposterior diameter of the renal pelvis (APD) represent the two most common standardized systems for evaluating hydronephrosis (1,7). The SFU grading system provides a qualitative assessment of hydronephrosis on the basis of the degree of pelvicaliectasis and the presence of cortical thinning (Figure 1A). Subjective evaluation of nonquantitative features is limited by high interobserver variability (8). APD is a quantitative assessment of the degree of pelviectasis on the basis of the measurement of the greatest diameter of the renal pelvis on ultrasound images acquired in a transverse plane (9) (Figure 1B).
Figure 1.
Ultrasound images demonstrating the Society for Fetal Urology and anteroposterior diameter of the renal pelvis grading systems. (A) Society for Fetal Urology grading system. Part I shows grade 1 (urine in pelvis barely splits sinus), part II shows grade 2 (urine fills intrarenal pelvis±major calyces dilated), part III shows grade 3 (virtually all calyces are visualized), and part IV shows grade 4 (similar to grade 3 plus parenchymal thinning). (B) anteroposterior diameter of the renal pelvis. 1, ultrasound image showing APD measured in the transverse section; 2, extrarenal pelvis.
A requirement for surgery and the ability of ultrasound imaging to predict a need for surgical intervention in ANH has been reviewed (1). Clinical outcomes for patients at the extremes of the spectrum of severity (i.e., mild and severe) of hydronephrosis are quite easily predicted. Mild hydronephrosis, defined as APD<10mm and SFU grade 1, stabilized or improved in 98% of patients reviewed (1). Up to 100% of children with SFU grade 4 worsen and require surgical intervention. In contrast, great variability in the natural history of ANH has been reported in patients with moderate grades of hydronephrosis, by either the APD or SFU grading systems. Studies demonstrate that 52%–65% of children with APD>12 mm stabilize or improve. Similarly, 42%–100% of patients with SFU grade 2 and 27%–89% of patients with SFU grade 3 stabilize or improve (1). Given the wide variation in outcomes in children with moderate degrees of hydronephrosis, there exists no clear consensus on criteria for surgical intervention on the basis of ultrasound measurements (3,4,10–12). As a result, current practice is far from uniform, often on the basis of physician or institutional individual practices. Previous studies have suggested that caliectasis may be associated with worse postnatal outcomes than pelviectasis without calyceal dilation (13–15). These studies proposed that the combination of APD grading and assessment of the renal parenchyma, including caliectasis and parenchymal thinning, could be a more predictive diagnostic tool for postnatal hydronephrosis. However, the results of these studies were inconclusive possibly because of the small study populations.
Here, we investigated the predictive capacity of an ultrasound-based grading system combining APD measurement and caliectasis to predict surgical intervention in ANH. We studied the largest patient cohort in which a combined grading system has been proposed, including almost four times more children than the aforementioned studies, and for the first time, to our knowledge, we compared a combined grading system of caliectasis with APD and SFU grading alone.
Material and Methods
This is a retrospective cohort study of 348 patients diagnosed with ANH and referred to either nephrology and/or urology clinics at The Hospital for Sick Children between April 1, 2003, and December 31, 2010, and followed longitudinally until June 30, 2013. Ethics approval was obtained at the Hospital for Sick Children. Ultrasound reports and clinical data were extracted and validated from electronic medical records.
Inclusion criteria are as follows: (1) children diagnosed with unilateral or bilateral ANH, (2) renal ultrasounds performed within the first 90 days of life, (3) first follow-up visit during the first 6 months of life, and (4) minimum of 1-year follow-up.
Exclusion criteria are as follows: (1) patients diagnosed with neurologic abnormalities involving the lower urinary tract (e.g., neurogenic bladder disorder, myelomeningocele); (2) renal malformations affecting at least one kidney, such as renal dysplasia, multicystic dysplastic kidney, renal agenesis, cross-fused ectopia, pelvic kidney, and autosomal dominant and recessive polycystic kidney disease; (3) lower urinary tract abnormalities, including posterior urethral valves, obstructive megaureter, and ureterocele; (4) and vesicoureteral reflux, diagnosed by voiding cystourethrogram, in the same kidney with hydronephrosis.
Ultrasound images were reanalyzed by two independent, trained, physician reviewers (nephrologist and urologist). The degree of hydronephrosis was determined on two consecutive coronal images of the kidney obtained at the hilum. SFU grades were generated according to the SFU definitions for hydronephrosis grades 1–4 (6). The SFU score corresponding to the image displaying the highest grade of hydronephrosis was recorded. For patients with bilateral hydronephrosis, SFU scores were generated for each renal unit, and the highest grade of hydronephrosis was considered. APD measurements were determined by the greatest diameter of the renal pelvis acquired in a transverse plane on ultrasound images. A total of 47 children with only ultrasound reports and no images were excluded, resulting in a final study population of 348 children.
We divided APD into quartiles generating four groups according to the measurement of the renal pelvis: (APD<6 mm, APD≥6–9 mm, APD≥9–15 mm, and APD>15 mm), similar to cut points reported in previous clinical studies (1,7,13,14). A combined grading system including each of the four APD groups plus the presence or absence of diffuse caliectasis was created. Diffuse caliectasis was defined by uniform dilation of major (central) and minor (peripheral) calyces, consistent with the SFU consensus definition (6).
The primary outcome was the requirement for surgical intervention to treat upper urinary tract hydronephrosis.
Additional data included age, sex, side of affected kidney, history of urinary tract infection confirmed by urine culture, and antibiotic prophylaxis.
Descriptive statistics were used to summarize data. Frequencies and proportions were calculated. ANOVA and Kruskal–Wallis tests were used to determine statistically significant differences (P<0.05) for categorical and continuous variables, respectively.
Using Cox proportional hazards analyses, we determined the association of ANH grading with surgical outcomes. Hazard ratios were calculated for APD, SFU, and also the combined grading systems to estimate the risk for surgery on the basis of worsening grades of ANH disease. Analyses were also adjusted for initial clinic visit with either a nephrologist or urologist.
Additionally, we calculated the positive likelihood ratios with 95% confidence intervals to estimate the probability of surgical intervention using the baseline ultrasound by the APD, SFU, and the combined grading systems. Similarly, to illustrate the performance of each grading system, we conducted a receiver operating characteristic analysis to determine the area under the curve across the range of sensitivity and 1-specificity.
Results
A total of 1193 children consecutively seen in either the nephrology or urology clinic were initially screened for ANH. Most were seen in the urology clinic (72.9%). There were 348 patients who met the inclusion criteria and constitute the final study population.
Demographic and clinical characteristics of the cohort of children with ANH are shown in Table 1. Overall, hydronephrosis occurred more frequently in boys (83%). Of the patients, 68% were treated with antibiotic prophylaxis. Only 4.3% of all children had one or more episode of confirmed urinary tract infection.
Table 1.
Baseline characteristics of 348 children with antenatal hydronephrosis by anteroposterior diameter of the renal pelvis and Society for Fetal Urology grading systems
| Characteristic | All Patients (N=348) | Anteroposterior Diameter of the Renal Pelvis | Society for Fetal Urology Grading System | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| <6 mm (n=75) | ≥6–9 mm (n=72) | >9–15 mm (n=104) | >15 mm (n=97) | P Value | I (n=68) | II (n=131) | III (n=55) | IV (n=94) | P Value | ||
| Male sexa | 289 (83) | 61 (81.3) | 61 (84.7) | 87 (83.6) | 80 (82.5) | 0.95 | 55 (80.9) | 112(85.5) | 48 (87.3) | 74 (78.7) | 0.44 |
| Age at first ultrasound (d)b | 18 (7–42) | 18 (7–43) | 16.5 (7–42) | 16 (8.5–35) | 20 (7–46) | 0.97 | 21(7.5–37.5) | 18 (9–42) | 14 (7–33) | 17.5(6–46) | 0.57 |
| Duration of follow-up (d)b | 1554±739 | 1387±755 | 1361±704 | 1551±697 | 1831±722 | 0.001 | 1295±707 | 1450±721 | 1796±657 | 1746±752 | 0.001 |
| Laterality,a left-sided, right-sided, and bilateral | 174 (50) | 41 (54.6) | 40 (55.5) | 54 (51.9) | 39 (40.2) | 0.29 | 37 (54.4) | 76 (58) | 22 (40) | 39 (41.5) | 0.50 |
| 44 (12.6) | 10 (13.4) | 6 (8.4) | 11 (10.6) | 17 (17.5) | 6 (8.8) | 5 (3.8) | 12 (21.8) | 21 (22.3) | |||
| 130 (37.4) | 24 (32) | 26 (36.1) | 39 (37.5) | 41 (42.3) | 25 (36.8) | 50 (38.2) | 21 (38.2) | 34 (36.2) | |||
| Prophylaxisa | 237 (68.1) | 48 (64) | 43 (59.7) | 70 (67.3) | 76 (78.3) | 0.11 | 37 (54.4) | 86 (65.6) | 39 (70.9) | 75 (79.8) | 0.07 |
| UTIa | 15 (4.3) | 6 (8) | 2 (2.7) | 6 (5.7) | 1 (1) | 0.11 | 4 (5.8) | 8 (6.1) | 1 (1.8) | 2 (2.1) | 0.34 |
| VCUGa | 250 (71.8) | 39 (52) | 50 (69.4) | 78 (75) | 83 (85.5) | 0.001 | 38 (55.8) | 83(63.3) | 50 (90.9) | 79 (84) | 0.001 |
| VURa | 35 (14) | 4 (10.2) | 6 (12) | 10 (12.8) | 15 (18) | 0.60 | 5 (13.1) | 8 (9.6) | 7 (14) | 15 (18.9) | 0.40 |
Values are n (%), median (interquartile range), mean±SD, or as otherwise indicated. UTI, history of urinary tract infection during follow-up; VCUG, number of patients that had voiding cystourethrograms done; VUR, vesicoureteral reflux diagnosed.
Analyzed using ANOVA test.
Analyzed using Kruskal–Wallis test; statistically significant at P<0.05.
Table 2 describes the outcomes of ANH for children with different grades of hydronephrosis in both the APD and SFU grading systems. A total of 96 patients (27.5%) underwent surgery. Surgical procedures included the following: unilateral pyeloplasty (88.5%; n=85), bilateral pyeloplasty (1%; n=1), laparoscopic pyeloplasty (3.1%; n=3), heminephrectomy (5.2%; n=5), and nephrectomy (2%; n=2). Children with higher grades of hydronephrosis underwent surgery more frequently than children with less severe grades of hydronephrosis. Children with high-grade hydronephrosis had surgery at younger ages compared with children with lower grades.
Table 2.
Surgical outcomes of 348 children with antenatal hydronephrosis
| Grading Systems | Group | Surgerya | Age at Surgeryb | Definitive Evidence of Obstruction in Operative or Pathology Report |
|---|---|---|---|---|
| APD | All children (N=348) | 96 (27.5) | 145 (103–370.5) | 95 (98.9) |
| APD<6 mm (n=75) | 2 (2.1) | 405–455c | 2 (2.1) | |
| APD≥6–9 mm (n=72) | 7 (7.3) | 113–1974c | 7 (7.3) | |
| APD>9–15 mm (n=104) | 22 (22.9) | 234.5 (130–328) | 21 (22.1) | |
| APD>15 mm (n=97) | 65 (67.7) | 123 (90–280) | 65 (68.5) | |
| P value | 0.001 | 0.01 | ||
| SFU | SFU grade 1 (n=68) | — | — | — |
| SFU grade 2 (n=131) | 7 (7.3) | 215–1528c | 7 (7.3) | |
| SFU grade 3 (n=55) | 23 (23.9) | 290 (130–636) | 23 (24.2) | |
| SFU grade 4 (n=94) | 66 (68.8) | 122.5 (90–214) | 65 (68.5) | |
| P value | 0.001 | 0.001 |
Values are n (%) or median (interquartile range). APD, anteroposterior diameter of the renal pelvis; SFU, Society for Fetal Urology.
Analyzed using ANOVA test.
Analyzed using Kruskal–Wallis test; statistically significant at P<0.05.
Minimum-maximum because of small numbers.
Surgical pathology and/or operative reports confirmed complete or partial urinary tract obstruction in all 96 patients who underwent surgery. Of those, 95 (98.9%) were scored as having complete obstruction, as characterized by a pinhole lumen/ureter, ureteral stricture, atretic or narrowed ureter, hypoplastic or dysplastic ureter, or presence of a crossing vessel. Only one child did not have ureteropelvic obstruction confirmed in the operative or pathology report and was reported as mild chronic inflammation. Also, another single child had a confirmed crossing vessel as the cause of ureteropelvic obstruction, and all other patients had evidence of intrinsic obstruction of the ureter at the ureteropelvic junction.
We calculated the hazard for surgery postnatally for APD, SFU, and the combined grading systems with the baseline ultrasound (Table 3). Using Cox regression, we calculated the hazard ratio for each category per grading system with APD<6mm, SFU grade 2 (none of the patients in the SFU grade 1 group had surgery), and APD<6 mm with no/few dilated calyces as our reference groups for each of the grading systems. Our data show that diffuse caliectasis had higher hazard ratios in the APD≥6–9 mm and APD ≥9–15 mm groups. The risk of surgery was 33-fold higher in the combined grading system APD≥6–9 mm with diffuse caliectasis and 10.7-fold higher in the combined grading system APD≥9–15 mm with diffuse caliectasis compared with APD alone in either the ≥6–9 mm or ≥9–15 mm groups, respectively. Both combined groups of moderate hydronephrosis had two times higher hazard than a similar SFU category defined by the presence of diffuse caliectasis.
Table 3.
Risk for surgery in children with antenatal hydronephrosis by baseline anterior posterior diameter of the renal pelvis, Society for Fetal Urology grading system and a combined grading system
| Grading Systems (N=348) | Children Who Underwent Surgery (n=96) | Adjusted Hazard Ratioa | 95% CI |
|---|---|---|---|
| APD | |||
| <6 mm | 2 (2.1) | ref | — |
| ≥6–9 mm | 7 (7.3) | 3.7 | 0.76 to 17.8 |
| ≥9–15 mm | 22 (22.9) | 8.5 | 1.99 to 36.1 |
| >15 mm | 65 (67.7) | 41.3 | 10.1 to 168.8 |
| SFU grade | |||
| 1 | — | — | — |
| 2 | 7 (7.3) | ref | — |
| 3 | 23 (23.9) | 8.9 | 3.84 to 20.9 |
| 4 | 66 (68.8) | 23.9 | 10.9 to 52.38 |
| Combined grading system | |||
| APD<6 mm, no/few dilated calyces (n=71) | 2 (2.1) | ref | |
| APD<6 mm, with diffuse caliectasis (n=4) | — | — | — |
| APD≥6–9 mm, no/few dilated calyces (n=60) | 1 (1.1) | 0.59 | 0.05 to 6.53 |
| APD≥6–9 mm, with diffuse caliectasis (n=12) | 6 (6.2) | 19.5 | 3.94 to 96.9 |
| APD≥9–15 mm, no/few dilated calyces (n=60) | 3 (3.1) | 1.75 | 0.29 to 10.5 |
| APD>9–15 mm, with diffuse caliectasis (n=44) | 19 (19.8) | 18.7 | 4.36 to 80.4 |
| APD>15 mm, no/few dilated calyces (n=7) | 1 (1.1) | 4.73 | 0.42 to 52.2 |
| APD>15 mm, with diffuse caliectasis (n=90) | 64 (66.6) | 44.7 | 10.9 to 183.1 |
Values are n (%) or as otherwise indicated. SFU 1 excluded from analysis because no events occurred in this group. APD, anteroposterior diameter of the renal pelvis; SFU, Society for Fetal Urology; ref, reference; no/few dilated calyces, absence of dilation of minor calyces only; diffuse caliectasis, presence of dilation of major and minor calyces; 95% CI, 95% confidence interval.
Hazard ratio adjusted for clinic where patient had been first seen: urology or nephrology.
We determined the predictive capability of the combined score by calculating the positive likelihood ratio for each grading system using the baseline ultrasound (Table 4). The likelihood ratio provides test characteristics of one group compared with all other groups by a specific grading system. A positive likelihood ratio of <1 indicates a negative probability for surgery, and >1 indicates that the test is likely to predict surgery. Likelihood ratio testing demonstrated that the presence of diffuse caliectasis in comparison with the same APD range with no/few dilated calyces was better than APD alone at predicting a need for surgical intervention. The likelihood ratio for APD≥6–9 mm with diffuse caliectasis was 65 times higher than APD≥6–9 mm alone. Predictive capability for surgery was 14 times higher for APD≥9–15 mm with diffuse caliectasis than APD≥9–15 mm alone. The likelihood ratio for severe hydronephrosis in the combined grading system was only slightly higher than APD and SFU alone. Both specificity and positive predictive values were higher in the combined grading system for moderate hydronephrosis than in similar APD groups (Table 4). Sensitivity was not high in any of the groups. The area under the curve varied by grading system with 0.83 for APD, 0.88 for SFU, and 0.86 for the combined score (Figure 2). There was a statistical difference between the APD and combined group (P<0.01), and there was no difference between the SFU and combined group (P=0.21). Together, these data demonstrate that the combined grading system is superior to APD alone for moderate degrees of hydronephrosis but is similar to SFU grade 3.
Table 4.
Likelihood of each grading system predicting need for surgery
| Grading Systems (N=348) | Children Who Had Surgery (n=96) | LR+ | Sensitivity | Specificity | PPV |
|---|---|---|---|---|---|
| APD | |||||
| <6 mm | 2 (2.1) | 0.07 (0.01 to 0.28) | 2 (0.01 to 0.07) | 71.0 (65.1 to 76.2) | 2.67 (0.40 to 9.32) |
| ≥6–9 mm | 7 (7.3) | 0.28 (0.13 to 0.59) | 7.29 (2.99 to 14.4) | 74.2 (68.3 to 79.4) | 9.72 (4.02 to 19) |
| ≥9–15 mm | 22 (22.9) | 0.70 (0.46 to 1.05) | 22.9 (14.9 to 32.6) | 67.4 (61.3 to 73.2) | 21.1 (13.7 to 30.6) |
| >15 mm | 65 (67.7) | 5.33 (3.75 to 7.58) | 67.7 (57.3 to 76.8) | 87.3 (82.5 to 91.1) | 67.1 (56.7 to 76.2) |
| SFU grade | |||||
| 1 | — | — | — | — | — |
| 2 | 7 (7.3) | 0.15 (0.07 to 0.31) | 7.29 (2.99 to 14.4) | 50.7 (44.4 to 57.1) | 5.34 (2.18 to 10.7) |
| 3 | 23 (23.9) | 1.89 (1.17 to 3.05) | 23.9 (15.8 to 33.7) | 87.3 (82.5 to 91.1) | 41.8 (28.6 to 55.8) |
| 4 | 66 (68.8) | 6.19 (4.26 to 9) | 68.7 (58.4 to 77.8) | 88.8 (84.3 to 92.4) | 70.2 (59.9 to 79.2) |
| Combined grading system | |||||
| APD<6 mm, no/few dilated calyces | 2 (2.1) | 0.08 (0.02 to 0.30) | 2.08 (0.31 to 7.34) | 72.6 (66.6 to 78) | 2.82 (0.42 to 9.83) |
| APD<6 mm, with diffuse caliectiasis | — | — | — | — | — |
| APD≥6–9 mm, no/few dilated calyces | 1 (1.1) | 0.04 (0.01 to 0.32) | 1.04 (0.17 to 5.69) | 76.5 (70.8 to 81.6) | 1.67 (0.28 to 8.97) |
| APD≥6–9 mm, with diffuse caliectasis | 6 (6.2) | 2.62 (0.87 to 7.94) | 6.25 (2.34 to 13.1) | 97.6 (94.8 to 99.1) | 50 (21.2 to 78.7) |
| APD≥9–15 mm, no/few dilated calyces | 3 (3.1) | 0.14 (0.04 to 0.43) | 3.12 (0.69 to 8.87) | 77.3 (71.7 to 82.3) | 5.0 (1.1 to 13.9) |
| APD≥9–15 mm, with diffuse caliectiasis | 19 (19.8) | 2.0 (1.15 to 3.45) | 19.7 (12.3 to 29.1) | 90.0 (85.7 to 93.4) | 43.1 (28.3 to 58.9) |
| APD>15 mm, no/few dilated calyces | 1 (1.1) | 0.44 (0.05 to 3.59) | 1.04 (0.17 to 5.69) | 97.6 (94.8 to 99.1) | 14.2 (2.37 to 57.7) |
| APD>15 mm, with diffuse caliectasis | 64 (66.6) | 6.97 (4.71 to 10.3) | 66.7 (56.3 to 75.9) | 90.4 (86.3 to 93.6) | 71.1 (60.6 to 80.1) |
Values are n (%) or % (95% confidence interval). APD, anteroposterior diameter of the renal pelvis; SFU, Society for Fetal Urology; no/few dilated calyces, absence of dilation of minor calyces only; diffuse caliectasis, presence of dilation of major and minor calyces; 95% CI, 95% confidence interval; LR+, positive likelihood ratio; PPV, positive predictive value.
Figure 2.
Receiver operating characteristic analysis comparing anteroposterior diameter of the renal pelvis, Society for Fetal Urology, and the combined grading systems. The area under the curve was 0.83 for APD, 0.88 for SFU, and 0.86 for the combined grading system. There was a statistical difference between the APD and combined group (P<0.01) and no difference between the SFU and combined group (P=0.21). APD, anteroposterior diameter of the renal pelvis; ROC, receiver operating characteristic; SFU, Society for Fetal Urology.
Discussion
We report on the independent association of the APD, SFU, or a combined grading systems with the need for surgical intervention. Our results support the use of diffuse caliectasis in the assessment of moderate hydronephrosis because it suggests higher predictive capability for surgical intervention compared with APD alone. Because the diagnosis of hydronephrosis in utero is defined, in clinical practice, using the APD system (5,6,8,13,16–18) and because the degree of hydronephrosis is assessed postnatally by pediatric radiologists using APD, we propose using this combined grading system to enhance the currently used APD system. This combined method will also provide continuity of assessment tools already used in the clinical evaluation of ANH.
Moderate hydronephrosis is associated with variable outcomes postnatally, and the inability to predict outcomes is equally frustrating to parents and clinicians. Using the first renal ultrasound after birth and adding the presence of diffuse caliectasis to standard APD grading significantly improves the predictive probability of surgery and provides potentially more information for clinical management. Recently, a multidisciplinary consensus on the classification of antenatal and postnatal urinary tract dilation has been proposed by a joint committee involving the Society for Pediatric Urology, American Society of Pediatric Nephrology, and American Academy of Radiology (19). This classification is on the basis of several ultrasound parameters, including APD and caliectasis (central and peripheral). Currently, the urinary tract dilation classification is associated with the risk of postnatal uropathies, but clinical outcomes, such as risk for surgery or urinary tract infection, are not yet defined. Our proposed grading system demonstrates for the first time, to our knowledge, that moderate APD values (≥6–9 mm and ≥9–15mm) in combination with diffuse caliectasis are associated with subsequent surgical intervention.
Consistent with our results, two previous studies have proposed the association of different APD ranges with aspects of the renal parenchyma, including caliectasis and parenchymal thinning, as a potentially more accurate diagnostic tool for postnatal assessment of ANH (13,14). In the first study, Grignon et al. (13) applied such a grading system to the analysis of 92 hydronephrotic kidneys in utero. In this study, all patients (16 kidneys) with moderate or severe caliectasis underwent surgery postnatally in contrast with the 62% assessed as mild (10 kidneys). Blachar et al. also applied a quantitative measure of pelviectasis and caliectasis to the assessment of 100 patients diagnosed with ANH. All 11 patients who underwent surgery presented with caliectasis, with nine of these patients presenting with severe calyceal dilation (14). In another study including 56 fetuses, the presence of concomitant caliectasis with APD measurement between 7–20 mm on the third trimester ultrasound increased the risk of postnatal surgical intervention 3-fold (15). Together, although these studies are limited by the inclusion of small numbers of patients, they do suggest that caliectasis is associated with surgical intervention. Our cohort has a substantially larger number of children followed for at least 1 year and demonstrates similar findings.
At this time, there is no gold standard for guiding clinical decisions regarding surgical intervention in ANH. Surgical repair is undertaken at the clinical discretion of the treating surgeon. Given this limitation, we used pathologic reports and observation at time of surgery as independent evidence that true urinary tract obstruction existed. Indeed, in all patients the operative and/or pathologic reports either demonstrated definite ureteropelvic junction obstruction or objective signs consistent with obstruction. These findings support the concept that APD grading in combination with diffuse caliectasis can predict a need for surgical intervention, especially for moderate hydronephrosis.
Repeated imaging after detecting ANH is commonly performed to determine the need for therapeutic intervention, including surgery. A limitation of our study resides in the fact that we did not analyze the temporal trend of hydronephrosis on ultrasound images. Retrospectively, time intervals for repeat ultrasounds varied significantly among patients (median interval between the first and second postnatal ultrasound was 150 days; interquartile range, 91–230 days). Additionally, many children had repeated measures at outside hospitals, and only 33% had available studies at 6-month follow-up. Future studies, however, need to consider change over time in the scoring system. Furthermore, we recognize that diuretic renography is commonly used in practice to evaluate upper tract obstruction. However, in our institution, it was not performed routinely for all children with hydronephrosis or for all children who underwent surgery. Another limitation is the potential for selection bias because 47 out of 395 eligible children did not have ultrasound images available and were therefore excluded from the study. This constitutes approximately 13% of the cohort and therefore is less likely to bias the results.
Strengths of our study were scoring by two independent physician-trained reviewers and consistent scoring of a higher grade in patients in which a difference of one grade was found between reviewers. A higher score in these patients may have underestimated the risk for surgery for lower grades of hydronephrosis. The standardization likely improved the precision of SFU, which may have overestimated prediction. Discordance between trained reviewers for SFU grading is a result of the lack of standardization in the manner in which calyces are imaged, resulting in variation in the number of calyces imaged for any particular patient and the number of sections of kidney tissue that are imaged. Scoring of calyces across a defined and consistent number of kidney cross-sectional images is suggested to minimize bias in scoring severity of hydronephrosis.
We suggest that the proposed scoring system would be a helpful tool used by nephrologists to counsel parents about the likelihood for surgery. The combined grading system needs to be tested further in the clinical setting to determine utility and validity in predicting surgical intervention.
A grading system using APD combined with diffuse caliectasis distinguishes those children at higher risk of surgery. This effect is particularly significant for moderate grades of hydronephrosis.
Disclosures
None.
Acknowledgments
We would like to thank the patients and families seen at the Hospital for Sick Children for their participation in this study. We would also like to thank Yi Zhang and Alex Salei for their contribution for data acquisition.
This work has been made possible through an International Society of Nephrology funded fellowship to Joana Dos Santos.
An abstract with preliminary data was presented at the International Pediatric Nephrology Association meeting in Shanghai, China, in September 2013.
Footnotes
Published online ahead of print. Publication date available at www.cjasn.org.
References
- 1.Sidhu G, Beyene J, Rosenblum ND: Outcome of isolated antenatal hydronephrosis: A systematic review and meta-analysis. Pediatr Nephrol 21: 218–224, 2006 [DOI] [PubMed] [Google Scholar]
- 2.Mallik M, Watson AR: Antenatally detected urinary tract abnormalities: More detection but less action. Pediatr Nephrol 23: 897–904, 2008 [DOI] [PubMed] [Google Scholar]
- 3.Timberlake MD, Herndon CD: Mild to moderate postnatal hydronephrosis--grading systems and management. Nat Rev Urol 10: 649–656, 2013 [DOI] [PubMed] [Google Scholar]
- 4.Davenport MT, Merguerian PA, Koyle M: Antenatally diagnosed hydronephrosis: Current postnatal management. Pediatr Surg Int 29: 207–214, 2013 [DOI] [PubMed] [Google Scholar]
- 5.Fernbach SK, Maizels M, Conway JJ: Ultrasound grading of hydronephrosis: Introduction to the system used by the Society for Fetal Urology. Pediatr Radiol 23: 478–480, 1993 [DOI] [PubMed] [Google Scholar]
- 6.Nguyen HT, Herndon CD, Cooper C, Gatti J, Kirsch A, Kokorowski P, Lee R, Perez-Brayfield M, Metcalfe P, Yerkes E, Cendron M, Campbell JB: The Society for Fetal Urology consensus statement on the evaluation and management of antenatal hydronephrosis. J Pediatr Urol 6: 212–231, 2010 [DOI] [PubMed] [Google Scholar]
- 7.Cheng AM, Phan V, Geary DF, Rosenblum ND: Outcome of isolated antenatal hydronephrosis. Arch Pediatr Adolesc Med 158: 38–40, 2004 [DOI] [PubMed] [Google Scholar]
- 8.Arger PH, Coleman BG, Mintz MC, Snyder HP, Camardese T, Arenson RL, Gabbe SG, Aquino L: Routine fetal genitourinary tract screening. Radiology 156: 485–489, 1985 [DOI] [PubMed] [Google Scholar]
- 9.Wong DC, Anderson PA, Macken M, Jackson JR: Congenital hydronephrosis who requires intervention? Can J Urol 6: 812–818, 1999 [PubMed] [Google Scholar]
- 10.Chertin B, Pollack A, Koulikov D, Rabinowitz R, Hain D, Hadas-Halpren I, Farkas A: Conservative treatment of ureteropelvic junction obstruction in children with antenatal diagnosis of hydronephrosis: Lessons learned after 16 years of follow-up. Eur Urol 49: 734–738, 2006 [DOI] [PubMed] [Google Scholar]
- 11.Onen A, Jayanthi VR, Koff SA: Long-term followup of prenatally detected severe bilateral newborn hydronephrosis initially managed nonoperatively. J Urol 168: 1118–1120, 2002 [DOI] [PubMed] [Google Scholar]
- 12.Bouzada MC, Oliveira EA, Pereira AK, Leite HV, Rodrigues AM, Fagundes LA, Gonçalves RP, Parreiras R: Diagnostic accuracy of postnatal renal pelvic diameter as a predictor of uropathy: A prospective study. Pediatr Radiol 34: 798–804, 2004 [DOI] [PubMed] [Google Scholar]
- 13.Grignon A, Filion R, Filiatrault D, Robitaille P, Homsy Y, Boutin H, Leblond R: Urinary tract dilatation in utero: Classification and clinical applications. Radiology 160: 645–647, 1986 [DOI] [PubMed] [Google Scholar]
- 14.Blachar A, Blachar Y, Livne PM, Zurkowski L, Pelet D, Mogilner B: Clinical outcome and follow-up of prenatal hydronephrosis. Pediatr Nephrol 8: 30–35, 1994 [DOI] [PubMed] [Google Scholar]
- 15.Oktar T, Acar O, Atar A, Salabas E, Ander H, Ziylan O, Kalelioglu I, Has R, Yuksel A: How does the presence of antenatally detected caliectasis predict the risk of postnatal surgical intervention? Urology 80: 203–206, 2012 [DOI] [PubMed] [Google Scholar]
- 16.Kitchens DM, Herndon CD: Antenatal hydronephrosis. Curr Urol Rep 10: 126–133, 2009 [DOI] [PubMed] [Google Scholar]
- 17.Pates JA, Dashe JS: Prenatal diagnosis and management of hydronephrosis. Early Hum Dev 82: 3–8, 2006 [DOI] [PubMed] [Google Scholar]
- 18.Herndon CD: Antenatal hydronephrosis: Differential diagnosis, evaluation, and treatment options. ScientificWorldJournal 6: 2345–2365, 2006 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Nguyen HT, Benson CB, Bromley B, Campbell JB, Chow J, Coleman B, Cooper C, Crino J, Darge K, Herndon CD, Odibo AO, Somers MJ, Stein DR: Multidisciplinary consensus on the classification of prenatal and postnatal urinary tract dilation (UTD classification system). J Pediatr Urol 10: 982–998, 2014 [DOI] [PubMed] [Google Scholar]