OBJECTIVES:

Posterior urethral valves predispose children to renal replacement therapy (RRT) and bladder dysfunction. Researchers of single-institutional series were unable to refine risk stratification because of rarity of the disease. We aimed to identify clinical variables associated with the risk of RRT and clean intermittent catheterization (CIC) in a large multicenter cohort study.

METHODS:

Children with posterior urethral valves born between 1995 and 2005 who were treated before 90 days of life at 5 children’s hospitals were retrospectively reviewed. Outcomes included RRT and recommendation for CIC. Predictors and outcomes were assessed by using survival analysis.

RESULTS:

A total of 274 patients were managed for a median of 6.3 years, and 42 progressed to RRT. On survival analysis, 16% progressed to RRT by 10 years of age. RRT varied by the serum nadir creatinine level in the first year of life (SNC1) (log-rank P < .001). After stratifying by the SNC1, the estimated risk of progressing to RRT by 10 years of age was 0%, 2%, 27%, and 100% for an SNC1 of <0.4, an SNC1 of 0.4 to 0.69, an SNC1 of 0.7 to 0.99, and an SNC1 of ≥1.0 mg/dL, respectively. CIC was recommended in 60 patients, which translated on survival analysis to a risk of 26% by 10 years of age.

CONCLUSIONS:

Risk of RRT and CIC recommendation increased with age. The SNC1 strongly predicted need for RRT. These results allow for both improved family counseling and the potential for more appropriate screening and intervention strategies for those identified in higher-risk groups.

What’s Known on This Subject:

A serum nadir creatinine level of <1 mg/dL is considered a prognostic indicator for need for renal replacement therapy in posterior urethral valves. However, the ability to narrow risk stratification cutoffs to multiple creatinine level ranges has been lacking.

What This Study Adds:

In this multicenter study of children with posterior urethral valves, we identified multiple nadir creatinine level ranges for prediction of renal replacement therapy. These results allow for improved family counseling and more risk-based screening and intervention strategies for higher-risk groups.

Obstructive uropathy, which includes posterior urethral valves (PUVs), remains a frequent cause of pediatric end-stage renal disease (ESRD) requiring renal replacement therapy (RRT).1 Progression to ESRD in PUVs is likely multifactorial, caused by diminished renal reserve due to a combination of renal dysplasia, obstructive pressure–induced renal injury, vesicoureteral reflux (VUR), and acquired scars from urinary tract infection.2,4 Authors of previous studies have described several prognostic factors for progression to ESRD, such as the serum nadir creatinine level in the first year of life (SNC1), presence of VUR, age at presentation, severity of upper tract dilation, and severe bladder dysfunction.5,13 Bladder dysfunction occurs to variable degrees in patients with PUVs, thus resulting in behavioral intervention, pharmacologic intervention, clean intermittent catheterization (CIC), and/or surgical intervention. The use of CIC in the PUV population can be challenging in these patients because they have a sensate urethra; however, evidence exists that CIC is beneficial in patients with PUV who have refractory bladder dysfunction.14,16 

Existing literature on PUV outcomes is based on single-center experience rather than multicenter observations. Our aim for this study was to measure 2 relevant clinical outcomes in the PUV population, RRT and CIC recommendations, among 5 children’s hospitals. These outcomes would serve as a baseline measure toward future initiatives to risk stratify patients, implement interventions, and report outcomes within our collaborative effort. We present age-based RRT and CIC risks as well as RRT risk stratified by SNC1. We hypothesized that increasing SNC1 would be associated with an increased risk of RRT.

Pediatric Urology Midwest Alliance (PUMA), a multi-institutional clinical and research collaborative, was initiated with the goal of improving the management and investigation of urologic diseases. We conducted a retrospective cohort study using medical records from patients diagnosed with PUVs and treated at 1 of the 5 PUMA hospitals from 1995 to 2005. The eligible study population was composed of male infants with the diagnosis of PUVs confirmed on cystography or cystoscopy. Patients were excluded if they underwent a first intervention for PUVs after 90 days of life to ensure a cohort of early diagnosis and treatment of PUVs. The institutional review boards at each PUMA center approved this study and data-sharing agreements obtained to allow for multicenter collaboration.

The primary outcomes of interest were (1) RRT and (2) recommendation for CIC. RRT was defined as either onset of dialysis or kidney transplantation. We determined recommendation for CIC by chart review of office and/or inpatient provider notes and/or durable medical equipment orders for CIC supplies. Sociodemographic and clinical characteristics were abstracted to identify potential risk factors for both CIC and RRT. The analysis included birth year, race and ethnicity, presence of solitary kidney, VUR, the SNC1, differential kidney function of ≤10% of 1 kidney, and anticholinergic use. Birth year was used to classify patients into 2 birth cohorts, 1995–2004 and 2005–2015, to evaluate potential changes in practice patterns and outcomes over time. The SNC1 was initially classified into 4 categories (<0.4, 0.4 to 0.69, 0.7 to 0.99, and ≥1 mg/dL) on the basis of clustering of data points after analysis. The SNC1 measurement was not adjusted for BMI.

Patient sociodemographic and clinical characteristics were summarized by using frequencies and percentages for categorical variables and medians and interquartile ranges (IQRs) for continuous variables. For each of the 2 study outcomes, risk of RRT and CIC recommendation, Kaplan-Meier curves and log-rank tests were used to examine the unadjusted association between patient characteristics and time to outcome occurrence by using patient age in years as the time scale.

The difference in follow-up times is the primary reason for using time-to-event analyses to examine the risk of the primary outcome over time (Kaplan-Meier curves and proportional hazards regression). These methods allow for including all patients in the risk estimation up until they experience the outcome (RRT and/or CIC recommendation) or their last follow-up visit for those who do not experience either outcome during the study period. By using this approach, we are able to obtain estimates of the risk on the basis of the entire study population rather than a restricted set of patients with a specific length of follow-up. For each analysis, age at last follow-up appointment was used as the censoring age for patients who did not experience the outcome of interest. We used time-to-event modeling (Kaplan-Meier curves and proportional hazards regression) to estimate the risk of RRT or CIC recommendation. By using the terminology for time-to-event analyses, patients are said to be censored if they do not experience the outcome of interest (RRT or CIC recommendation) by the end of the study.

Characteristics associated with time to outcome occurrence in bivariable analyses at P < .20 were added to a proportional hazards regression model to identify significant independent risk factors for the outcome. From a statistical perspective, we initially used a P < .2 to screen for statistically significant factors (factors that we identified a priori that may be clinically significant). When characteristics were unknown, this was recorded; however, there was no interest in including the unknown category in the screening step because this group may include a combination of different patients with different characteristics. We did not want to prematurely exclude potential clinically relevant confounders, and if we identified that these unknown factors were statistically significant in the screening step, the goal was to include all of these factors in the modeling on the basis of a priori consideration of their clinical significance and statistical significance at P < .2. When a substantial proportion of patients had the unknown category, we then included this category in the modeling because otherwise, the alternative would have been to exclude the patients from the analysis or to impute, which we determined was less appropriate when there was a high proportion of subjects with missing data.

The final model retained significant risk factors at P < .05. Study site (1 of 5 PUMA hospitals) was included for adjustment in the regression model for RRT and CIC on the basis of results from both the univariable and multivariable analyses.

A total of 274 subjects met inclusion criteria, with a median follow-up time (age at last follow-up appointment) of 6.3 years (IQR: 3.5–9.7). Therefore, 25% of patients were managed up to 3.5 years, 75% of patients were managed up to 9.7 years, and 25% of patients were managed for >9.7 years. Overall, 184 (67.1%) patients were lost to follow-up by 10 years of age. Ninety patients were managed for ≥10 years from birth to date of last follow-up. Two hundred and twelve and 169 patients were managed for ≥3 and 5 years from birth to date of last follow-up, respectively.

One hundred and seventy-four patients (63.5%) had antenatal findings leading to clinical suspicion of diagnosis of PUVs. Patients were more commonly non-Hispanic white (69.3%) and born in the younger cohort (2005–2015; 76.6%). Only 1.5% were noted to have a solitary kidney; however, ≤10% differential function of 1 kidney on a nuclear medicine scan occurred in 18 (6.5%) patients. Half of the patients had VUR present at the time of diagnosis. A majority of patients had an SNC1 of <0.4 (47.1%) or 0.4 to 0.69 mg/dL (21.5%), with less patients in the more severe ranges of 0.7 to 0.99 (5.1%) and ≥1.00 mg/dL (8.1%; Table 1). A total of 53.9% of patients were found to have anticholinergic use at any age at follow-up.

TABLE 1

Patient Characteristics in the PUV Cohort

Results
Overall, n (%) 274 (100.0) 
Birth cohort, n (%)  
 1995–2004 64 (23.4) 
 2005–2015 210 (76.6) 
Race and ethnicity, n (%)  
 NH white 190 (69.3) 
 NH African American 48 (17.5) 
 NH other 15 (5.5) 
 Hispanic 11 (4.0) 
 Unknown 10 (3.6) 
Solitary kidney, n (%)  
 Yes 4 (1.5) 
 No 256 (93.4) 
 Unknown 14 (5.1) 
VUR present at diagnosis, n (%)  
 Yes 137 (50.0) 
 No 105 (38.3) 
 Unknown 32 (11.7) 
≤10% kidney, n (%)  
 Yes 50 (18.2) 
 No 101 (36.9) 
 Unknown 123 (44.9) 
Anticholinergic, n (%)  
 Yes 146 (53.3) 
 No 124 (45.3) 
 Unknown 4 (1.5) 
SNC1, n (%), mg/dL  
 <0.40 129 (47.1) 
 0.40–0.69 59 (21.5) 
 0.7–0.99 14 (5.1) 
 ≥1.00 24 (8.8) 
 Unknown 48 (17.5) 
Age at last follow-up, median (IQR), y 6.3 (3.5–9.7) 
Results
Overall, n (%) 274 (100.0) 
Birth cohort, n (%)  
 1995–2004 64 (23.4) 
 2005–2015 210 (76.6) 
Race and ethnicity, n (%)  
 NH white 190 (69.3) 
 NH African American 48 (17.5) 
 NH other 15 (5.5) 
 Hispanic 11 (4.0) 
 Unknown 10 (3.6) 
Solitary kidney, n (%)  
 Yes 4 (1.5) 
 No 256 (93.4) 
 Unknown 14 (5.1) 
VUR present at diagnosis, n (%)  
 Yes 137 (50.0) 
 No 105 (38.3) 
 Unknown 32 (11.7) 
≤10% kidney, n (%)  
 Yes 50 (18.2) 
 No 101 (36.9) 
 Unknown 123 (44.9) 
Anticholinergic, n (%)  
 Yes 146 (53.3) 
 No 124 (45.3) 
 Unknown 4 (1.5) 
SNC1, n (%), mg/dL  
 <0.40 129 (47.1) 
 0.40–0.69 59 (21.5) 
 0.7–0.99 14 (5.1) 
 ≥1.00 24 (8.8) 
 Unknown 48 (17.5) 
Age at last follow-up, median (IQR), y 6.3 (3.5–9.7) 

NH, non-Hispanic.

A total of 42 (15.3%) patients progressed to RRT during follow-up. Taking into account the differing lengths of follow-up, 9%, 16%, and 25% of patients had experienced RRT by 1, 10, and 13 years respectively (Fig 1). On univariable analyses, VUR (P = .01), a higher SNC1 (P < .0001), and anticholinergic use (P = .01) were significantly associated with progression to RRT (Table 2). The remaining patient characteristics (birth cohort, race and ethnicity, a solitary kidney, and ≤10% kidney function) were not associated with age at RRT. In a multivariable model, the SNC1was the only significant independent factor (P < .0001).

FIGURE 1

Kaplan-Meier estimates of age at RRT (PUV cohort; 1995–2015).

FIGURE 1

Kaplan-Meier estimates of age at RRT (PUV cohort; 1995–2015).

TABLE 2

Summary of Events (RRT) and Association Between Patient Characteristics and Age at RRT

Patients, nOutcomeAnalysis of Age at RRT
RRT, n (%)Censoreda, nUnadjustedbAdjustedcP
Log-rank PHR (95% CI)
Overall 274 42 (15.3) 232 — — — 
Birth cohort       
 1995–2004 64 12 (18.8) 52 .29 Not included Not included 
 2005–2015 210 30 (14.3) 180 — Not included Not included 
Race and ethnicity       
 NH white 190 34 (17.9) 156 .50 Not included Not included 
 NH African American 48 4 (8.3) 44 — Not included Not included 
 NH other 15 1 (6.7) 14 — Not included Not included 
 Hispanic 11 2 (18.2) — Not included Not included 
 Unknown 10 1 (10.0) — Not included Not included 
Solitary kidney       
 Yes .44 Not included Not included 
 No 256 36 (14.1) 220 — Not included Not included 
 Unknown 14 6 (42.9) — Not included Not included 
VUR present at diagnosis       
 Yes 137 24 (17.5) 113 .01 2.31 (0.91–5.85) .08 
 No 105 6 (5.7) 99 — Reference 
 Unknown 32 12 (37.5) 20 — 3.57 (1.25–10.20) .02 
≤10% kidney       
 Yes 50 3 (6.0) 47 .76 Not included Not included 
 No 101 7 (6.9) 94 — Not included Not included 
 Unknown 123 32 (26.0) 91 — Not included Not included 
Anticholinergic       
 Yes 146 30 (20.5) 116 .01 0.97 (0.45–2.10) .93 
 No 124 10 (8.1) 114 — Reference 
 Unknown 2 (50.0) — 0.41 (0.08–2.16) .29 
SNC1, mg/dL       
 <0.40 129 129 <.0001 Reference 
 0.40–0.69 59 3 (5.1) 56 —   
 0.7–0.99 14 5 (35.7) — 22.14 (5.27–93.12) <.0001 
 ≥1.00 24 24 (100.0) — 173.25 (47.45–632.56) <.0001 
 Unknown 48 10 (20.8) 38 — 10.64 (2.81–40.25) .01 
Patients, nOutcomeAnalysis of Age at RRT
RRT, n (%)Censoreda, nUnadjustedbAdjustedcP
Log-rank PHR (95% CI)
Overall 274 42 (15.3) 232 — — — 
Birth cohort       
 1995–2004 64 12 (18.8) 52 .29 Not included Not included 
 2005–2015 210 30 (14.3) 180 — Not included Not included 
Race and ethnicity       
 NH white 190 34 (17.9) 156 .50 Not included Not included 
 NH African American 48 4 (8.3) 44 — Not included Not included 
 NH other 15 1 (6.7) 14 — Not included Not included 
 Hispanic 11 2 (18.2) — Not included Not included 
 Unknown 10 1 (10.0) — Not included Not included 
Solitary kidney       
 Yes .44 Not included Not included 
 No 256 36 (14.1) 220 — Not included Not included 
 Unknown 14 6 (42.9) — Not included Not included 
VUR present at diagnosis       
 Yes 137 24 (17.5) 113 .01 2.31 (0.91–5.85) .08 
 No 105 6 (5.7) 99 — Reference 
 Unknown 32 12 (37.5) 20 — 3.57 (1.25–10.20) .02 
≤10% kidney       
 Yes 50 3 (6.0) 47 .76 Not included Not included 
 No 101 7 (6.9) 94 — Not included Not included 
 Unknown 123 32 (26.0) 91 — Not included Not included 
Anticholinergic       
 Yes 146 30 (20.5) 116 .01 0.97 (0.45–2.10) .93 
 No 124 10 (8.1) 114 — Reference 
 Unknown 2 (50.0) — 0.41 (0.08–2.16) .29 
SNC1, mg/dL       
 <0.40 129 129 <.0001 Reference 
 0.40–0.69 59 3 (5.1) 56 —   
 0.7–0.99 14 5 (35.7) — 22.14 (5.27–93.12) <.0001 
 ≥1.00 24 24 (100.0) — 173.25 (47.45–632.56) <.0001 
 Unknown 48 10 (20.8) 38 — 10.64 (2.81–40.25) .01 

PUV cohort (1995–2015). NH, non-Hispanic; —, not applicable.

a

Patients who were censored were either lost to follow-up or did not undergo RRT by the end of their follow-up.

b

Log-rank test for the differences in age at RRT (univariable).

c

Multivariable proportional hazards regression model for age at RRT. The model included only factors with a P < .20 in the univariable log-rank test.

Compared with patients with an SNC1 <0.7 mg/dL (reference group for the regression model), those with an SNC1 of 0.7 to 0.99 mg/dL were 22.1 times more likely to undergo RRT (95% confidence interval [CI]: 5.3–93.1). All 24 patients (100.0%) with an SNC1 ≥1 mg/dL had RRT, which was 173 times higher than that of the reference group (95% CI: 47.4–632.6; Table 2). After stratifying by the SNC1, the estimated risk of progressing to RRT by 10 years of age was 0%, 2%, 27%, and 100% for an SNC1 <0.4, an SNC1 of 0.4 to 0.69, an SNC1 of 0.7 to 0.99, and an SNC1 ≥1.0 mg/dL, respectively (Fig 2).

FIGURE 2

Kaplan-Meier estimates of age at RRT, stratified by SNC1 (PUV cohort; 1995–2015).

FIGURE 2

Kaplan-Meier estimates of age at RRT, stratified by SNC1 (PUV cohort; 1995–2015).

Sixty (21.9%) patients had CIC recommended during the follow-up period. Accounting for the differing lengths of follow-up, by 10 years of age, 26% of patients had CIC recommended, whereas 50% of the patients were recommended CIC by 14 years of age (Fig 3). In univariable analyses, younger birth cohort (2005–2015; P < .0001), VUR (P = .02), and the SNC1 (<.0001) were significantly associated with earlier recommendation of CIC (Table 3). In a multivariable model, younger birth cohort (hazard ratio [HR]: 15.3; 95% CI: 4.6–51.1), an SNC1 ≥1 mg/dL (HR: 8.3; 95% CI: 3.6–19.2), and anticholinergic use (HR: 6.8; 95% CI: 2.7–17.2) were associated with an increased risk of CIC.

FIGURE 3

Kaplan-Meier estimates of age at recommendation for CIC (PUV cohort; 1995–2015).

FIGURE 3

Kaplan-Meier estimates of age at recommendation for CIC (PUV cohort; 1995–2015).

TABLE 3

Summary of Events (CIC) and Association Between Patient Characteristics and Age at CIC

Patients, nOutcomeAnalysis of Age at CIC
CIC, n (%)Censoreda, nUnadjustedbAdjustedcP
Log-rank PHR (95% CI)
Overall 274 — — — — — 
Birth cohort       
 1995–2004 64 13 (20.3) 51 <.0001 Reference 
 2005–2015 210 47 (22.4) 163 — 15.27 (4.57–51.09) <.0001 
Race and ethnicity       
 NH white 190 44 (23.2) 146 .65 Not included Not included 
 NH African American 48 12 (25.0) 36 — Not included Not included 
 NH, other 15 2 (13.3) 13 — Not included Not included 
 Hispanic 11 2 (18.2) — Not included Not included 
 Unknown 10 10 — Not included Not included 
Solitary kidney       
 Yes 1 (25.0) .91 Not included Not included 
 No 256 56 (21.9) 200 — Not included Not included 
 Unknown 14 3 (21.4) 11 — Not included Not included 
VUR present at diagnosis       
 Yes 137 40 (29.2) 97 .02 1.36 (0.68–2.72) .38 
 No 105 12 (11.4) 93 — Reference 
 Unknown 32 8 (25.0) 24 — 3.34 (1.24–8.99) .02 
≤10% kidney       
 Yes 50 11 (22.0) 39 .68 Not included Not included 
 No 101 23 (22.8) 78 — Not included Not included 
 Unknown 123 26 (21.1) 97 — Not included Not included 
Anticholinergic       
 Yes 146 53 (36.3) 93 <.0001 6.78 (2.67–17.20) <.0001 
 No 124 6 (4.8) 118 — Reference 
 Unknown 1 (25.0) — 8.07 (0.77–84.42) .08 
SNC1, mg/dL       
 <0.40 129 20 (15.5) 109 <.0001 Reference 
 0.40–0.69 59 12 (20.3) 47 — 1.35 (0.55–3.31) .51 
 0.7–0.99 14 7 (50.0) — 2.43 (0.98–6.00) .05 
 ≥1.00 24 13 (54.2) 11 — 8.34 (3.62–19.19) <.0001 
 Unknown 48 8 (16.7) 40 — 0.43 (0.15–1.24) .12 
Patients, nOutcomeAnalysis of Age at CIC
CIC, n (%)Censoreda, nUnadjustedbAdjustedcP
Log-rank PHR (95% CI)
Overall 274 — — — — — 
Birth cohort       
 1995–2004 64 13 (20.3) 51 <.0001 Reference 
 2005–2015 210 47 (22.4) 163 — 15.27 (4.57–51.09) <.0001 
Race and ethnicity       
 NH white 190 44 (23.2) 146 .65 Not included Not included 
 NH African American 48 12 (25.0) 36 — Not included Not included 
 NH, other 15 2 (13.3) 13 — Not included Not included 
 Hispanic 11 2 (18.2) — Not included Not included 
 Unknown 10 10 — Not included Not included 
Solitary kidney       
 Yes 1 (25.0) .91 Not included Not included 
 No 256 56 (21.9) 200 — Not included Not included 
 Unknown 14 3 (21.4) 11 — Not included Not included 
VUR present at diagnosis       
 Yes 137 40 (29.2) 97 .02 1.36 (0.68–2.72) .38 
 No 105 12 (11.4) 93 — Reference 
 Unknown 32 8 (25.0) 24 — 3.34 (1.24–8.99) .02 
≤10% kidney       
 Yes 50 11 (22.0) 39 .68 Not included Not included 
 No 101 23 (22.8) 78 — Not included Not included 
 Unknown 123 26 (21.1) 97 — Not included Not included 
Anticholinergic       
 Yes 146 53 (36.3) 93 <.0001 6.78 (2.67–17.20) <.0001 
 No 124 6 (4.8) 118 — Reference 
 Unknown 1 (25.0) — 8.07 (0.77–84.42) .08 
SNC1, mg/dL       
 <0.40 129 20 (15.5) 109 <.0001 Reference 
 0.40–0.69 59 12 (20.3) 47 — 1.35 (0.55–3.31) .51 
 0.7–0.99 14 7 (50.0) — 2.43 (0.98–6.00) .05 
 ≥1.00 24 13 (54.2) 11 — 8.34 (3.62–19.19) <.0001 
 Unknown 48 8 (16.7) 40 — 0.43 (0.15–1.24) .12 

PUV cohort (1995–2015). NH, non-Hispanic; —, not applicable.

a

Patients who were censored were either lost to follow-up or did not undergo CIC recommendation by the end of their follow-up.

b

Log-rank test for differences in age at CIC (univariable).

c

Multivariable proportional hazards regression model forage at CIC. The model included only factors with a P < .20 in the univariable log-rank test in addition to study site (1 of 5 participating hospitals).

Multicenter studies have the advantage over single-center studies because of the inclusion of a larger number of subjects and may help facilitate the generalizability of findings. To our knowledge, this current series represents the largest infant PUV population described in the literature. Bolstering its large numbers is that this series represents a multi-institutional cohort rather than a single-center cohort, allowing for determination of clinically meaningful cutoffs for SNC1. On the survival analysis, 16% of our PUV cohort progressed to RRT by 10 years of age, which is similar to previous single-center publications.17 We anticipate the number of patients with PUVs requiring RRT to increase as our cohort continues to be managed over time because a significant number of patients with PUVs progress to RRT during adolescence and into adulthood.17,18 

We observed 2 separate inflection points in the Kaplan-Meier survival curve for risk of RRT. Rapid progression to RRT within the first year of life occurred in 9% of patients, presumably secondary to severe renal dysplasia accounting for swift need for RRT in these patients. An additional 17% of the remaining cohort progressed to RRT between ages 10 and 15 years. We are unable to comment on the reasons for delayed progression to RRT in all patients because this study was not designed to answer that question. We did not include CIC recommendation as a variable predicting RRT because CIC is meant to treat underlying bladder dysfunction, therefore suggesting CIC recommendation as a confounding factor. Accurate assessment of bladder dysfunction as a predictor of RRT would necessitate a review of all urodynamics; however, this analysis would not be reliable given the limitations of the retrospective urodynamic data from multiple sites without a standardized protocol.

Our inclusion criterion affords the benefit of data that are relevant to patients with PUVs who have been diagnosed within 3 months of life with early intervention. Authors of the majority of previous studies of patients with PUVs did not limit the age of their study groups, and therefore, direct comparisons between our data set are limited because of variability in the study design and data collection. A homogenous sample allows for a more accurate risk assessment for prognosis of certain outcomes at defined end points. We were able to estimate the risk of progressing to RRT by 10 years of age being 0%, 2%, 27%, and 100% for an SNC1 <0.4, an SNC1 of 0.4 to 0.69, an SNC1 of 0.7 to 0.99, and an SNC1 of ≥1.0 mg/dL, respectively.

Assessment and stratification of RRT risk at variable time points have clinical utility in the care of the child with PUVs. Counseling parents of a child with PUVs may be facilitated with improved knowledge from risk stratification, especially in families for whom rapid progression to RRT is likely. Additionally, providers may adjust the frequency of urologic or renal testing on the basis of the risk level to identify changes in hopes of mitigating or delaying progression to RRT. Future prospective research within our collaboration (PUMA) is now directed to better understand how standardization of newborn management may alter the risk of RRT in patients with PUVs. Results from the current study reveal that patients with an SNC1 of 0.4 to 0.99 mg/dL may represent the group with the greatest potential for disease course modification.

Approximately one-quarter (26%) of the patients in our cohort were recommended CIC by 10 years of age. We intentionally chose the outcome of CIC recommendation and cannot comment on compliance of CIC. Assessment of compliance with any recommendation can be difficult to accurately measure. We did not record the indications for recommending CIC in our study design, and indications for recommending CIC were not uniform among the participating institutions. The risk of recommendation of CIC differed by institution, VUR at time of diagnosis, and SNC1. We can only infer that severe bladder dysfunction with renal impairment were factors leading to recommending CIC. Ansari et al4 and DeFoor et al9 examined risk factors for ESRD in children with PUVs, with the authors of both studies defining severe bladder dysfunction as patients requiring CIC. Both studies revealed on a multivariable logistic regression analysis that severe bladder dysfunction was an independent risk factor for RRT in addition to increased serum nadir creatinine. Holmdahl et al19 examined 19 boys with PUVs who were started on CIC at 8 months of life, with results revealing that early treatment of bladder dysfunction counteracted renal deterioration seen during childhood in patients with PUVs. Lopez Pereira et al20 also evaluated urodynamic variables in 59 patients with PUVs who also had ESRD and those with normal renal function. Of the 22 patients with ESRD, 15 had abnormal urodynamics (68%) and only 7 had bladders with normal urodynamic findings (32%). Of the 37 boys with normal renal function, 19 had dysfunctional bladders (51%) and 18 had normal bladders (49%). Although the authors did not specifically comment on rates of CIC usage in their series, they noted that 8 of 9 patients with poor bladder compliance and 2 of 3 patients with myogenic failure developed ESRD. The authors concluded that bladder dysfunction should be considered a prognostic factor for the development of ESRD.

We observed a steady increase in the proportion of patients who did have CIC recommended during the follow-up period. This observation was likely related to the clinical phenomenon of valve bladder syndrome, which was described in 1980 by Mitchell.21 The syndrome describes an entity in which polyuria from renal-concentrating defect, poor bladder compliance, and residual urine volume promotes bladder dysfunction even after successful surgical elimination of the anatomic bladder outlet obstruction from PUVs.16 Because of the sensate nature of the male urethra in patients with PUVs, compliance with performing CIC is a concern when studying CIC effect on clinical outcomes. This has led to recommendation to perform abdominal wall catheterizable channel for bladder catheterization as originally described by Mitrofanoff.22 King et al23 described the use of the catheterizable channel in the PUV population, with improvement in hydronephrosis and bladder dysfunction, but this did not prevent renal deterioration. In prospective studies in which patient and family compliance was accounted for, use of a catheterizable channel and controlling for the level of bladder dysfunction with standardized urodynamic assessment are necessary to better understand the impact CIC has on progression to RRT in PUV.

Our purpose for this article was to describe 2 readily available binary outcomes for a PUV population within the framework of a multicenter collaborative research network. All retrospective studies experience limitations that need to be acknowledged because these may affect data interpretation. Data collection was standardized, and every attempt was made to ensure complete data while accounting for loss to follow-up. Given the geographic proximity of PUMA institutions, we were able to capture some migratory patients who left 1 PUMA institution but continued their care in another PUMA institution. Although a large number of patients are included in this series, the length of follow-up is relatively short for a disease condition that requires follow-up through adolescence into adulthood for meaningful long-term renal outcome measurement. The large amount of patients lost to follow-up is a significant limitation, and unfortunately, we were unable to ascertain all the reasons for all patients who dropped out (ie, moved out of the area or were healthy and felt that further urologic follow-up was unnecessary). The SNC1 was not adjusted for BMI, and it is possible that interpretation of the SNC1 without this adjustment could affect results. Although body weight (more importantly, muscle mass) does become a more clinically significant factor for interpreting serum creatinine as the child grows, we felt that during infancy, it was unnecessary to adjust for this as a variable. Finally, another limitation was that clinical care for PUVs was not standardized in the cohort, and the outcomes may have been affected without similar standards of care for the various treatment modalities of RRT, CIC, and/or anticholinergic medication.

In our multicenter collaborative, we identified 274 patients with early treatment of PUVs, with 16% of the patients who were managed at risk for RRT at age 10 years. In a similar analysis, we identified 26% of patients at risk for CIC recommendation at age 10 years. The SNC1 strongly predicted the risk of this outcome. In fact, in patients with an SNC1 <0.4 mg/dL, none progressed to RRT, whereas all of those with an SNC1 ≥1.0 mg/dL progressed to RRT. Patients with an SNC1 between 0.4 and 0.99 mg/dL showed a more variable course, suggesting a patient subset that may benefit the most from future study of factors that affect disease modification. Future prospective research is underway to standardize early treatment of patients with PUVs to better understand how management differences may alter the disease course.

Drs McLeod, Szymanski, Gong, Granberg, and Whittam conceptualized and designed the study, designed the data collection instruments, collected data, revised the manuscript, and critically reviewed the manuscript for important intellectual content; Dr VanderBrink conceptualized and designed the study, designed the data collection instruments, collected data, coordinated and supervised data collection, drafted the initial manuscript, revised the manuscript, and critically reviewed the manuscript for important intellectual content; Mr Sebastião conduted the initial data analyses; Drs Fuchs, Reddy, and Gargollo critically reviewed the manuscript for important intellectual content; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

FUNDING: No external funding.

Mark P. Cain, MD; Martin Kaefer, MD; Rosalia Misseri, MD; Richard C. Rink, MD; Konrad M. Szymanski, MD; Benjamin Whittam, MD (Riley Hospital for Children at IU Health, Indianapolis, IN); Earl Y. Cheng, MD; Edward Gong, MD; Emilie K. Johnson, MD, MPH; Ilina Rosoklija; Elizabeth Yerkes, MD (Ann and Robert H. Lurie Children’s Hospital of Chicago, Chicago, IL); Charles Concodora, MD; William Robert DeFoor, MD, MPH; Eugene Minevich, MD; Paul Noh, MD; Pramod P. Reddy, MD; Abbey Riazzi, PA; Andrew Strine, MD; Brian A. VanderBrink, MD (Cincinnati Children’s Hospital Medical Center, Cincinnati, OH); Daniel DaJusta MD; Molly Fuchs, MD; Venkata Rama Jayanthi, MD; Daryl J. McLeod, MD, MPH; Yuri Sebastiao (Nationwide Children’s Hospital, Columbus, OH); Patricio Gargollo, MD; and Candace Granberg, MD (Mayo Clinic, Rochester, MN).

     
  • CI

    confidence interval

  •  
  • CIC

    clean intermittent catheterization

  •  
  • ESRD

    end-stage renal disease

  •  
  • HR

    hazard ratio

  •  
  • IQR

    interquartile range

  •  
  • PUMA

    Pediatric Urology Midwest Alliance

  •  
  • PUV

    posterior urethral valve

  •  
  • RRT

    renal replacement therapy

  •  
  • SNC1

    serum nadir creatinine level in the first year of life

  •  
  • VUR

    vesicoureteral reflux

1
North American Pediatric Renal Trials and Collaborative Studies
. NAPRTCS annual reports. 2014. Available at: https://web.emmes.com/study/ped/annlrept/annlrept.htmlwww.NAPRTCS.org. Accessed January 7,
2019
2
Lopez Pereira
P
,
Martinez Urrutia
MJ
,
Espinosa
L
,
Jaureguizar
E
.
Long-term consequences of posterior urethral valves.
J Pediatr Urol
.
2013
;
9
(
5
):
590
596
[PubMed]
3
Drozdz
D
,
Drozdz
M
,
Gretz
N
,
Möhring
K
,
Mehls
O
,
Schärer
K
.
Progression to end-stage renal disease in children with posterior urethral valves.
Pediatr Nephrol
.
1998
;
12
(
8
):
630
636
[PubMed]
4
Ansari
MS
,
Surdas
R
,
Barai
S
,
Srivastava
A
,
Kapoor
R
.
Renal function reserve in children with posterior urethral valve: a novel test to predict long-term outcome.
J Urol
.
2011
;
185
(
6
):
2329
2333
[PubMed]
5
Bilgutay
AN
,
Roth
DR
,
Gonzales
ET
 Jr
, et al
.
Posterior urethral valves: risk factors for progression to renal failure.
J Pediatr Urol
.
2016
;
12
(
3
):
179.e1
179.e7
[PubMed]
6
Sarhan
OM
,
El-Ghoneimi
AA
,
Helmy
TE
,
Dawaba
MS
,
Ghali
AM
,
Ibrahiem
el-HI
.
Posterior urethral valves: multivariate analysis of factors affecting the final renal outcome.
J Urol
.
2011
;
185
(
suppl 6
):
2491
2495
[PubMed]
7
Engel
DL
,
Pope
JC
 IV
,
Adams
MC
,
Brock
JW
 III
,
Thomas
JC
,
Tanaka
ST
.
Risk factors associated with chronic kidney disease in patients with posterior urethral valves without prenatal hydronephrosis.
J Urol
.
2011
;
185
(
suppl 6
):
2502
2506
[PubMed]
8
Ansari
MS
,
Gulia
A
,
Srivastava
A
,
Kapoor
R
.
Risk factors for progression to end-stage renal disease in children with posterior urethral valves.
J Pediatr Urol
.
2010
;
6
(
3
):
261
264
[PubMed]
9
DeFoor
W
,
Clark
C
,
Jackson
E
,
Reddy
P
,
Minevich
E
,
Sheldon
C
.
Risk factors for end stage renal disease in children with posterior urethral valves.
J Urol
.
2008
;
180
(
suppl 4
):
1705
1708; discussion 1708
[PubMed]
10
Heikkilä
J
,
Holmberg
C
,
Kyllönen
L
,
Rintala
R
,
Taskinen
S
.
Long-term risk of end stage renal disease in patients with posterior urethral valves.
J Urol
.
2011
;
186
(
6
):
2392
2396
[PubMed]
11
Coleman
R
,
King
T
,
Nicoara
CD
, et al
.
Combined creatinine velocity and nadir creatinine: a reliable predictor of renal outcome in neonatally diagnosed posterior urethral valves.
J Pediatr Urol
.
2015
;
11
(
4
):
214.e1
214.e3
12
Merguerian
PA
,
McLorie
GA
,
Churchill
BM
,
McKenna
PH
,
Khoury
AE
.
Radiographic and serologic correlates of azotemia in patients with posterior urethral valves.
J Urol
.
1992
;
148
(
5
):
1499
1503
[PubMed]
13
Denes
ED
,
Barthold
JS
,
González
R
.
Early prognostic value of serum creatinine levels in children with posterior urethral valves.
J Urol
.
1997
;
157
(
4
):
1441
1443
[PubMed]
14
Hale
JM
,
Wood
DN
,
Hoh
IM
, et al
.
Stabilization of renal deterioration caused by bladder volume dependent obstruction.
J Urol
.
2009
;
182
(
suppl 4
):
1973
1977
[PubMed]
15
Montané
B
,
Abitbol
C
,
Seeherunvong
W
, et al
.
Beneficial effects of continuous overnight catheter drainage in children with polyuric renal failure.
BJU Int
.
2003
;
92
(
4
):
447
451
[PubMed]
16
Koff
SA
,
Mutabagani
KH
,
Jayanthi
VR
.
The valve bladder syndrome: pathophysiology and treatment with nocturnal bladder emptying.
J Urol
.
2002
;
167
(
1
):
291
297
[PubMed]
17
Smith
GH
,
Canning
DA
,
Schulman
SL
,
Snyder
HM
 III
,
Duckett
JW
.
The long-term outcome of posterior urethral valves treated with primary valve ablation and observation.
J Urol
.
1996
;
155
(
5
):
1730
1734
[PubMed]
18
Holmdahl
G
,
Sillén
U
.
Boys with posterior urethral valves: outcome concerning renal function, bladder function and paternity at ages 31 to 44 years.
J Urol
.
2005
;
174
(
3
):
1031
1034; discussion 1034
[PubMed]
19
Holmdahl
G
,
Sillen
U
,
Hellström
AL
,
Sixt
R
,
Sölsnes
E
.
Does treatment with clean intermittent catheterization in boys with posterior urethral valves affect bladder and renal function?
J Urol
.
2003
;
170
(
4, pt 2
):
1681
1685; discussion 1685
[PubMed]
20
Lopez Pereira
P
,
Martinez Urrutia
MJ
,
Espinosa
L
,
Lobato
R
,
Navarro
M
,
Jaureguizar
E
.
Bladder dysfunction as a prognostic factor in patients with posterior urethral valves.
BJU Int
.
2002
;
90
(
3
):
308
311
[PubMed]
21
Mitchell
ME
.
Valve bladder syndrome.
In:
American Urological Association Annual Meeting of North Central Section
;
November 9–16, 1980
;
Hamilton, Bermuda
22
Mitrofanoff
P
.
Trans-appendicular continent cystostomy in the management of the neurogenic bladder [in French].
Chir Pediatr
.
1980
;
21
(
4
):
297
305
[PubMed]
23
King
T
,
Coleman
R
,
Parashar
K
.
Mitrofanoff for valve bladder syndrome: effect on urinary tract and renal function.
J Urol
.
2014
;
191
(
suppl 5
):
1517
1522
[PubMed]

Competing Interests

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.