Late Kidney Effects of Nephron-Sparing vs Radical Nephrectomy for Wilms Tumor: A Systematic Review and Meta-Analysis
Abstract
Purpose:
We compare differences in long-term kidney function between patients undergoing either radical nephrectomy (RN) or nephron-sparing surgery (NSS) in unilateral and bilateral Wilms tumor (WT), respectively.
Materials and Methods:
A systematic search was performed in September 2020. Comparative studies were evaluated according to Cochrane collaboration recommendations. Assessed long-term (>1-year postoperative) outcomes included chronic kidney disease, hypertension and glomerular filtration rate, among others. Odds ratio and mean difference with 95% confidence interval were extrapolated from available data for quantitative synthesis. Random-effects meta-analysis and meta-regression were performed according to study design and techniques. The systematic review was prospectively registered on PROSPERO (CRD42020205378).
Results:
A total of 23 studies describing 293 cases of unilateral WT and 386 cases of bilateral WT were included in the qualitative synthesis. Overall effect estimates demonstrate that patients undergoing RN have significantly increased odds of developing abnormal kidney function (OR 4.29, 95% CI 1.02, 18.00) and lower estimated glomerular filtration rate at long-term followup (mean difference −8.99, 95% CI −16.40, −1.58) compared to those undergoing NSS. In bilateral WT, patients undergoing RN with contralateral NSS have higher odds of developing abnormal kidney function (OR 3.82, 95% CI 1.76, 8.33) and hypertension (OR 5.81, 95% CI 1.31, 25.68) compared to bilateral NSS.
Conclusions:
Current evidence is low quality but suggests that NSS for unilateral and bilateral WT may be associated with better kidney function or blood pressure at late followup. Further research to investigate sources of heterogeneity is recommended.
| BP |
blood pressure |
| CKD |
chronic kidney disease |
| eGFR |
estimated glomerular filtration rate |
| ESRD |
end-stage renal disease |
| GFR |
glomerular filtration rate |
| HTN |
hypertension |
| NSS |
nephron-sparing surgery |
| PRISMA |
Preferred Reporting Items for Systematic Reviews and Meta-Analyses |
| RN |
radical nephrectomy |
| SCr |
serum creatinine |
| WT |
Wilms tumor |
Wilms tumor (WT) is the most common type of pediatric kidney cancer. Advancements in surgical and medical therapy have contributed to an overall survival rate greater than 90%.1 As with other cancers, increased survival for patients with WT may be coupled with late adverse effects from therapy. WT survivors are more likely to develop late kidney abnormalities, including chronic kidney disease (CKD) or hypertension (HTN).2,3 In unilateral and bilateral WT, the incidence of kidney failure (stage 5 CKD, chronic dialysis, transplant) is estimated to be less than 0.5% and 12%, respectively.4,5 Patients are variably subjected to many types of kidney injury mechanisms, including nephrotoxic chemotherapy, abdominal radiation, and nephrectomy. It is thus challenging to determine causation between individual kidney injury causes and late kidney outcomes.3
Radical nephrectomy (RN) or nephron-sparing surgery (NSS) is performed for local control in conjunction with treatment strategies developed by large oncology groups. To date, studies report similar relapse and survival rates.1,5,6 Traditionally, NSS was reserved for patients with bilateral WT, high risk of metachronous lesions (ie predisposition syndromes) or high risk of kidney failure. In recent years, increasing familiarity with and refinement of the procedure, reduced morbidity and favorable outcome data on NSS in treating renal cell carcinoma7–9 have led to interest in wider adoption of NSS in children with kidney neoplasms. The most common current practice indicates that NSS for unilateral WT should be favored for patients at high risk for kidney failure or later bilateral disease.10 These lesions tend to be smaller (often detected during screening of patients with predisposition syndromes) and/or managed with neoadjuvant chemotherapy to reduce volume and facilitate dissection from the surrounding normal parenchyma. Although this introduces a selection bias, the opportunity to compare outcomes with patients requiring radical resection provides valuable information.
Several observational studies have investigated late kidney outcomes in children treated surgically for unilateral and bilateral WT; however, a knowledge gap remains on the association of surgical approach with kidney outcomes.5,11–14 The objective of this systematic review is to assess the association of RN vs NSS with late kidney outcomes in patients with unilateral and bilateral WT.
Methods
The review protocol was registered on PROSPERO (CRD42020205378) and performed according to Cochrane Collaboration recommendations.15 Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)16 were applied to ensure high-quality reporting.
Search Strategy
An experienced information specialist was consulted in the design of our search strategy. We searched MEDLINE®, Embase®, CINAHL®, Scopus® and Web of Science™ for studies in any language with no date restrictions. The MEDLINE search strategy is provided in supplementary table 1 (https://www.jurology.com). We searched the references of published studies and hand-searched corresponding full-text articles from gray literature to identify additional studies. The final search was conducted on September 3, 2020.
Evidence Acquisition
Articles were eligible if they assessed late kidney or blood pressure (BP) outcomes in patients undergoing curative unilateral or bilateral WT surgery. Late kidney outcomes were defined as reduced glomerular filtration rate (GFR), CKD, kidney failure, proteinuria, HTN, electrolyte abnormalities, dialysis or kidney transplant occurring at least 1 year following surgery. We included all studies that compared RN vs NSS in unilateral WT, and RN + NSS vs NSS + NSS in bilateral WT. Studies with fewer than 5 patients, not reporting late kidney outcomes (>1 year after surgery) or not reporting outcomes compared between groups were excluded. Previously published systematic reviews, narrative reviews and commentaries were searched for references.
Two reviewers (AK, AJ) independently screened titles/abstracts and full texts based on inclusion and exclusion criteria. All stages of the process were piloted between the 2 reviewers. Conflicts at all stages were resolved by consensus between 2 reviewers. After full-text screening, all studies reporting desired outcome statistics, or data to calculate them, were included in quantitative synthesis. Two reviewers (AK, AJ) independently abstracted data into a standardized, pilot-tested data form for study characteristics and relevant covariates for both study arms, when possible. Study-level covariates were determined a priori based on literature review and collected, including: percent with low-risk cancer, followup time after surgery, patient age, chemotherapy, radiation and predisposing syndromes. If an outcome was not directly reported, but determinable from presented data, the reviewers performed the required calculations and noted for which outcomes this occurred. If primary outcomes were collected and not reported, study authors were contacted; this occurred in only 1 case.17
Measures of Effect
A composite outcome was used to express late kidney effects as abnormal kidney function: presence of estimated GFR (eGFR) <90 ml/minute/1.73 m2, serum creatinine (SCr) >1.5 mg/dl or kidney failure. The composite outcome components were selected based on initial a priori discussion among the investigators and then modified slightly after brief review of the selected articles, based on how kidney outcomes were expressed in most studies and collected at the time of followup or the latest followup in studies with multiple time points. HTN was defined as systolic or diastolic BP ≥95th percentile for age, height, sex or treatment with anti-HTN medications.18
Statistical Analysis
Odds ratios were estimated to evaluate the relation of these dichotomous outcomes with nephrectomy group. Means and standard deviations for continuous measures were estimated to allow for pooling of continuous outcomes19 or estimated from error bars if information was solely provided in figures, when required. Continuous outcomes (eg eGFR) were compared using mean difference. All estimates were expressed with their 95% CIs.
A random-effects model was chosen for all meta-analyses. The random-effects meta-analysis was conducted using an inverse-variance method. A sensitivity analysis was conducted post hoc to explore the impact of zero-event studies.20 Heterogeneity was quantified using I2 values,21 and between-study heterogeneity was calculated using the DerSimonian and Laird method.22 We defined significant statistical heterogeneity as I2 ≥50%. Weighted mean followup time was calculated from the mean values provided for each study. Planned meta-regression was used to investigate potential sources of clinical and methodological heterogeneity. Residual maximum likelihood estimates for between-study heterogeneity were used in all instances of meta-regression. Covariates of interest defined a priori were followup time, presence of predisposing syndromes, nephrotoxic chemotherapy or radiation. Covariates assessed in meta-regression were based on the availability of covariates in either unilateral or bilateral WT. Meta-analyses were conducted using RevMan 5.4 (Cochrane, London, UK) and meta-regressions were conducted using the metareg command in Stata® 16.1.
Risk of Bias Assessment
In accordance with Cochrane recommendations,15 risk of bias for individual nonrandomized studies was appraised using all domains of the ROBINS-I tool23 and randomized controlled trials were appraised using the RoB 2.0 tool.24 Two reviewers (AK, MLG) independently assessed each study for risk of bias; discrepancies were resolved by consensus between 2 reviewers. Studies with critical risk of bias were removed from quantitative synthesis. Funnel plots were generated to examine the risk of publication bias for all meta-analyses.
Results
Search Results and Study Characteristics
A total of 3,923 unique records were retrieved from literature after 3,047 duplicates were removed; 3,736 were excluded based on title and abstract screening, 187 moved on to full-text screening and 166 were excluded (PRISMA flow diagram, fig. 1). Two additional studies were included from review of references during full-text review (from a conference abstract and a narrative review).17,25 Ultimately, 23 articles were included for this systematic review, including 9 studies investigating unilateral WT11,13,14,17,26–30 and 14 studies investigating bilateral WT.5,25,31–42 There was only 1 randomized intervention study comparing RN vs NSS to include in this review. The inter-reviewer kappa was >0.75 at each stage. Study characteristics, overall risk of bias assessment and outcome measures are provided for unilateral WT in table 1 and bilateral WT in table 2.
Figure 1. PRISMA flow diagram of search strategy for systematic review and meta-analysis.
| Study | Study Design | Surgery | No. | Yrs to Followup (range or ±SD) | Mean ml/min/1.73 m2 eGFR at Followup (range or ±SD) | % Pts with Renal Insufficiency (eGFR <90 or SCr >1.5) | Other Renal Outcomes* | % Pts Given Chemotherapy | Overall Bias | |
| At Diagnosis | At Followup | |||||||||
| Cost 201411 | Retrospective cohort | RN | 15 | 8.4† (0.5–31.8) | 131.0† (98.6–161.2) | 0.0 | 0.0 | – | 13.3 preop,
|
Moderate risk |
| NSS | 15 | 2.1† (0.6–10.5) | 135.3† (57.5–158.8) | 46.7 | 6.7 | – | 40.0 preop,
|
|||
| Cozzi 200526 | Retrospective cohort | RN | 16 | 6.0±3.4 | – | – | 18.8 | HTN: 12.5%; abnormal albumin-to-creatine ratio: 12.5%;
|
100.0 postop | Severe risk |
| NSS | 11 | 5.4±3.2 | – | – | 0.0 | HTN: 0.0%; abnormal albumin-to-creatine ratio: 0.0%;
|
100.0 postop | |||
| Cozzi 201213 | Retrospective cohort | RN | 15 | 12.4±4.0 | 88.6±13.5 | 86.7 | 53.3 | – | 100.0 | Severe risk |
| NSS | 10 | 12.3±4.0 | 106.2±24.1 | 50.0 | 0.0 | – | 100.0 | |||
| Cozzi 201329 | Retrospective cohort | RN | 42 | 11.7±6.5 | 95.1±18.5 | – | 58.1 | – | 75.3 preop, | Severe risk |
| NSS | 12 | 11.4±7.8 | 109.8±18.4 | – | 8.3 | – | 100.0 preop | |||
| Cozzi 201612 | Retrospective cohort | RN | 25 | 11.3±7.5‡ | 99.8±20.0 | 52.0 | 32.0 | – | 94.4 preop | Moderate risk |
| NSS | 11 | 10.8±4.6‡ | 36.3 | 0.0 | – | |||||
| Janeczko 201527 | Retrospective cohort | RN | 41 | 2.0 | 94.0±22.0‡ | – | – | HTN: 8.0%
|
92.0 preop, (radiation: 12.0%) | Severe risk |
| NSS | 9 | 95.6±22.0‡ | – | – | ||||||
| Nerli 202017 | Randomized controlled trial | RN | 13 | 4.5±2.0 | 98 | – | 38.4 | HTN: 23.1% abnormal albumin-to-creatine ratio: 15.4%;
|
100.0 preop | Low risk |
| NSS | 15 | 110 | – | 6.6 | HTN: 0.0% abnormal albumin-to-creatine ratio: 0.0%;
|
100.0 preop | ||||
| Neu 201728 | Retrospective cohort | RN | 33 | 24.8±11.2 | – | 29.7 | 54.5 | – | 100.0 | Severe risk |
| NSS | 2 | – | 50.0 | – | 100.0 | |||||
| Romao 201214 | Retrospective cohort | RN | 2 | 3.7±3.2 | 121.6±33.4 | 50.0 | 0.0 | HTN: 50.0% | 50.0 preop | Severe risk |
| NSS | 6 | 2.8±1.6 | 125.0±7.5 | 0.0 | 0.0 | HTN: 16.7% | 66.7 preop | |||
| Study | Surgery | No. Pts | Yrs to Followup (range or ±SD) | Mean ml/min/1.73 m2 eGFR at Followup (range) | % Pts with Renal Insufficiency (eGFR <90, SCr >1.5 or ESRD) | Other Measures* | % Pts Given Adjuvant Therapy | Overall Bias | |
| Chemotherapy | Radiation | ||||||||
| Aldrink 201825 | RN + NSS | 5 | 1.8 (0.6–11.3) | 97.6 (68.6–105.8) | 40.0 (eGFR) | – | 100.0 | – | Moderate risk |
| NSS + NSS | 10 | 176.1 (113.0–279.8) | 0.0 (eGFR) | – | |||||
| Aronson 201136 | RN + NSS | 14 | 10.5† (5.5–34.0) | – | 28.5 (SCr) | HTN: 7.1% | 100.0 | 20.0‡ | Severe risk |
| NSS + NSS | 8 | – | 0.0 (SCr) | HTN: 0.0% | |||||
| Davidoff 201537 | RN + NSS | 3 | 3.7 (0.03–13.4) | – | 33.3 (eGFR) | Proteinuria: 13.9%‡;
|
95.2 | 42.9 | Moderate risk |
| NSS + NSS | 39 | – | 36.4 (eGFR) | ||||||
| Fuchs 199938 | RN + NSS | 4 | 3.3 (1.0–4.0) | 103.8 (66.0–122.0) | 25.0 (creatinine clearance) | – | 92.8‡ | – | Severe risk |
| NSS + NSS | 9 | 8.3 (2.0–18.0) | 119.4 (91.0–169.0) | 0.0 (creatinine clearance) | – | – | |||
| Giel 200739 | RN + NSS | 9 | 6.0 (1.3–17.3) | – | 33.3 (1 SCr, 2 ESRD) | HTN: 52.9% | 100.0 | – | Moderate risk |
| NSS + NSS | 8 | – | 12.5 (ESRD) | – | |||||
| Hamilton 20115 | RN + NSS | 53 | 13.9 (0.01–19.8)‡ | – | 5.7 (ESRD)‡ | – | – | – | Severe risk |
| NSS + NSS | 35 | – | 0.0 (ESRD)‡ | – | |||||
| Hubertus 201540 | RN + NSS | 12 | 4.7 (0.6–13.8) | 111.0 (30.0–87.0) | 25.0 (SCr) | HTN: 66.7%; albuminuria: 33.3%; proteinuria: 16.7%;
|
100.0 | 16.7 | Moderate risk |
| NSS + NSS | 10 | 4.9 (1.2–8.4) | 130.0 (82.0–178.0) | 0.0 (SCr) | HTN: 20%; albuminuria: 0%;
|
90.0 | 20.0 | ||
| Kumar 199834 | RN + NSS | 27‡ | 6.0 (1.0–15.0) | – | – | – | 100.0 | 30.0 | Critical risk |
| NSS + NSS | 9‡ | – | – | – | |||||
| Oue 201435 | RN + NSS | 15 | 8.0† (1.3–13.1) | – | 53.3 (SCr) | – | 93.3 | 16.0 | Severe risk |
| NSS + NSS | 10 | – | 20.0 (SCr) | – | 100.0 | ||||
| Sarhan 201042 | RN + NSS | 16 | 3.0 (1.0–11.0) | – | 0.0 (SCr) | – | 75.0 | 0 | Severe risk |
| NSS + NSS | 3 | – | 0.0 (SCr) | – | 100.0 | 0 | |||
| Shaul 199243 | RN + NSS | 11 | 7.1 (1.3–16.9) | – | 36.4 (ESRD) | – | 100.0 | 36.4 | Moderate risk |
| NSS + NSS | 2 | 6.4 (5.3–7.5) | – | 0.0 (ESRD) | – | 100.0 | 0 | ||
| Sudour 201232 | RN + NSS | 29 | 8.0† (7.0–9.1) | – | 24.1 (SCr) | ESRD: 14.6%;
|
100.0 | 22.9 | Moderate risk |
| NSS + NSS | 19 | – | 10.5 (SCr) | 100.0 | |||||
| Sulkowski 201233 | RN + NSS | 2 | 3.0 (0.3–6.6)† | 81.0 (71.7–90.3) | 50.0 (eGFR) | – | 83.3‡ | 25‡ | Moderate risk |
| NSS + NSS | 6 | 99.1 (62.8–129.4) | 33.3 (eGFR) | – | |||||
| Tan 201841 | RN + NSS | 2 | 3.5 (0.8–7.4)† | – | 5.5 (eGFR) | HTN: 50% | 100.0 | 27.8 | Severe risk |
| NSS + NSS | 16 | – | HTN: 12.5% | ||||||
Risk of Bias
Based on the risk of bias tools used in this review, almost all studies were determined to have concerns for at least moderate risk of bias, with most meeting criteria for serious risk of bias. In almost all studies, the primary cause of severe risk of bias was issues of confounding or lack of reporting on important covariates such as chemotherapy, radiation or predisposing syndromes. An outline of important covariates, which were collected for each study, is provided in supplementary tables 2 and 3 (https://www.jurology.com).
Unilateral WT
Patients treated with RN were compared to patients treated with NSS in unilateral WT with a weighted mean followup of 9.4 years (range 2.0–24.8). From 5 studies with available data, patients who underwent RN had increased odds of developing abnormal kidney function (OR 4.29; 95% CI 1.20, 18.00) compared to NSS (fig. 2, A). Notably, 2 studies with sufficient information were not pooled26,29 due to a high likelihood of overlapping cohorts with a pooled study that included more comprehensive kidney outcomes reporting.13 Pooled mean eGFR across 5 studies showed a significantly decreased eGFR by 8.99 ml/minute/1.73 m2 (95% CI −16.40, −1.58) in patients with RN vs NSS (supplementary fig. 1, https://www.jurology.com). Two studies compared the difference in eGFR at diagnosis and followup in patients who underwent RN or NSS.11,13 In both studies, eGFR improved for patients undergoing NSS by 17.5 and 43.6 ml/minute/1.73 m2, respectively. For patients undergoing RN, eGFR improved in 1 study by 10.4 ml/minute/1.73 m2 and dropped in the other by 18.9 ml/minute/1.73 m2. From 3 studies with available data, there was no statistically significant difference in the odds of HTN by end of followup (OR 5.83, 95% CI 0.91, 37.43) between patients undergoing RN vs NSS (fig. 2, B).
Figure 2. A, forest plot pooled effect estimates for the outcome of abnormal kidney function in unilateral WT; comparison: RN vs NSS; subgroup: nonrandomized and randomized studies. Statistical method: inverse variance with random-effects model (OR with 95% CI). B, forest plot pooled effect estimates for the outcome of HTN in unilateral WT; comparison: RN vs NSS; subgroup: nonrandomized and randomized studies. Statistical method: inverse variance with random-effects model (OR with 95% CI). IV, instrumental variable.
For studies contributing to meta-regression on effect estimates for unilateral WT,11–14,17,28 there were no significant differences in log ORs of abnormal kidney function at followup with respect to a unit increase in followup time (β=0.13, p=0.54) or presence of stage I/II WT (β=6.90, p=0.40; supplementary table 4, https://www.jurology.com). Other covariates such as presence of` predisposing syndromes, chemotherapy and radiation were not included due to inadequate availability of information. Subgroup analysis between patients treated under Children’s Oncology Group protocol vs the International Society of Pediatric Oncology protocol showed no significant difference in the incidence of abnormal kidney outcomes (chi2=0.41, p=0.52; supplementary fig. 2, https://www.jurology.com).
Based on funnel plots generated from studies of unilateral WT patients (supplementary fig. 3, https://www.jurology.com), there was a likelihood of publication bias favoring the RN approach. The risk of bias assessment for studies on unilateral WT is provided in supplementary tables 5 and 6 (https://www.jurology.com). Of the 9 studies, 1 had low risk, 2 had moderate risk and 6 had severe risk of bias.
Bilateral WT
Patients treated with RN + NSS were compared to patients treated with bilateral NSS in bilateral WT with a weighted mean followup of 7.7 years (range 1.8–13.9). Abnormal kidney function was assessed in 11 studies. There were increased odds of developing abnormal kidney function (OR 3.82; 95% CI 1.76, 8.33) in patients after RN + NSS compared to bilateral NSS. There was a nonsignificant association between groups in prevalence of kidney failure (OR 3.73; 95% CI 0.71, 19.56; fig. 3, A). Pooled mean eGFR in 4 studies showed a significant decrease in GFR of 27.91 ml/minute/1.73 m2 (95% CI −38.84, −16.98) after RN + NSS vs bilateral NSS; however, this outcome had high heterogeneity (I2=89%; supplementary fig. 4, https://www.jurology.com). There was increased odds of developing HTN (OR 5.81; 95% CI 1.31, 25.68) in those receiving RN + NSS vs bilateral NSS across 3 studies (fig. 3, B).
Figure 3. A, forest plot pooled effect estimates for the outcome of abnormal kidney function in bilateral WT; comparison: RN + NSS vs NSS + NSS; subgroup: abnormal kidney function and kidney failure. Statistical method: inverse variance with random-effects model (OR with 95% CI). B, forest plot pooled effect estimates for the outcome of HTN in bilateral WT; comparison: RN + NSS vs NSS + NSS. Statistical method: inverse variance with random-effects model (OR with 95% CI). IV, instrumental variable.
For studies contributing to meta-regression on effect estimates for bilateral WT,5,25,31,32,34–39,42 there were no significant differences in log ORs of abnormal kidney function at followup with respect to a unit increase in followup time (β=–0.196, p=0.58), presence of predisposing conditions to poor kidney outcomes (β –1.54, p=0.71) or receipt of chemotherapy (β=7.76, p=0.49; supplementary table 7, https://www.jurology.com). Other covariates such as treatment with radiation and stage I/II WT were not included due to inadequate availability of information. Subgroup analysis between patients treated under the Children’s Oncology Group protocol vs the International Society of Pediatric Oncology protocol showed no significant difference in the incidence of abnormal kidney outcomes (chi2=0.13, p=0.72; supplementary fig. 2, https://www.jurology.com).
In studies of patients with bilateral WT (supplementary fig. 5, https://www.jurology.com), funnel plots suggest a low risk of publication bias for assessed outcomes. The risk of bias assessment for studies on bilateral WT is provided in supplementary table 8 (https://www.jurology.com). With regard to the 14 studies included in the systematic review on bilateral WT, 7 studies had moderate risk, 6 had serious risks and 1 had critical risk of bias.
Discussion
This meta-analysis suggests that, overall, NSS is associated with reduced likelihood of abnormal kidney function compared to RN. These findings support the notion that NSS should be considered in selected patients with WT to reduce the incidence of kidney conditions. Our findings should be considered in the context of relevant covariates, potential sources of heterogeneity and predominantly nonrandomized study designs. Recent systematic reviews have suggested that RN and NSS appear to have similar oncologic outcomes with regard to relapse rate and overall survival,6 and a prospective trial that encouraged NSS in children with bilateral tumors or with predisposition syndromes had excellent oncologic outcomes.43 However, NSS has not been tested in a prospective manner for nonsyndromic patients with unilateral WT. NSS has traditionally been reserved for bilateral WT, small unilateral WT or in patients at higher risk for kidney failure.7,44 Ultimately, our findings provide potentially useful data summaries for families and health care practitioners to aid with treatment decisions, as well as provide direction for future studies on the long-term consequences of surgical strategies for WT.
A recent Cochrane Review on late kidney outcomes in childhood cancer survivors determined that inconsistent reporting of kidney outcomes limited determining accurate effect sizes.2,3 Here, only 1 included article reported all the outcomes described in our initial protocol.39 Future studies should similarly report kidney outcomes more comprehensively and in a standardized fashion. Long-term kidney outcome studies should include measures of GFR and proteinuria (the components of the CKD definition45), as CKD is a well-known potent cardiovascular risk factor.46 The simplest and most feasible way to evaluate GFR is to measure SCr or cystatin C to estimate GFR; however, both methods may have limitations in certain pediatric populations.47 Reference standard GFR measurement methods (eg analyte clearance studies) are most accurate but are cumbersome and expensive. Studies evaluating kidney outcome at a single time point may be best suited for reference standard GFR measurement, whereas longitudinal studies with multiple time points, evaluating kidney function trajectory, are likely best suited for eGFR measurement methods. Only 5 studies reported prevalence of HTN and none used 24-hour ambulatory BP monitoring; this should be done more frequently and reported in the same manner as done recently in a noncomparative cohort study.48 Findings from the recent AREN 0534 clinical trial have shown excellent oncologic outcomes out to 7 years of followup with NSS, but kidney outcomes are still pending for bilateral or bilateral-predisposed WT.1,43
There are important clinical considerations for interpreting kidney function post WT surgery, especially as some patients with WT are expected to be ideal candidates for NSS.8,44 This suggests there are inherent differences between patients receiving RN vs NSS by tumor and patient factors. Patients receiving NSS are expected to have smaller tumors and lower disease burden, may be more prone to reduced kidney function requiring preservation of renal parenchyma and are more likely to be treated at a referral center.26 We did not find a significant difference in tumor stage between RN and NSS groups, but this may be confounded by downstaging from up-front chemotherapy or unreported markers of initial cancer status. Moreover, WT pathology and positive margin status determine the need for more intensive therapy and affect kidney function for patients at followup.8 Conditions like Denys-Drash syndrome predispose patients to nephropathy, and they should be screened for kidney dysfunction regularly.7 In this review, 11 studies reported on patients with predisposing syndromes or other genitourinary abnormalities, constituting 25% of the study population. Our review highlights that although kidney function in WT survivors is associated with surgical approach, other treatments impact long-term kidney function in WT survivors.3
Study findings should be interpreted in light of limitations. Inconsistent outcome reporting limited the ability to pool results for primary outcomes, and despite our attempts to summarize data in a standardized fashion this summary was limited by the published data available. We defined a composite outcome for abnormal kidney function and performed subgroup analysis for those with kidney failure in bilateral WT. Reporting in the literature was an important limitation; only 4 studies reported the number of patients given adjuvant nephrotoxic chemotherapy, information for adjuvant radiation therapy was not stratified by type of surgery in most studies and there was significant heterogeneity in reporting of followup time. We performed meta-regression with covariates that were most available but found no significant association between covariates assessed and abnormal kidney outcomes. Most included studies had selection bias, as patients with lower risk tumors or worse preoperative eGFR are likely offered NSS more readily than RN.11,13,26,36 This issue cannot be addressed without randomized studies. Moreover, all but 1 study was nonrandomized, with associated limitations.
Overall, the use of NSS in unilateral tumors is a more recent development to mitigate excessive kidney parenchymal loss when metachronous contralateral tumors are a possibility, and are recommended primarily for polar or noninfiltrating tumors.49 Yet, important challenges remain, limiting the implementation of NSS. These include limited surgeon experience with NSS in the pediatric population, the relative rarity of this condition, the lack of clear guidelines on indications for the procedure and the lack of prospective data on oncologic outcomes.1 Importantly, NSS should not be performed at the cost of adequate cancer control, which can result in tumor spillage, recurrence and poorer survival, along with a need for further nephrotoxic therapy. NSS in unilateral disease should be reserved for cases where good oncologic outcomes can be assured and where there is a risk of renal dysfunction. These procedures should ideally be done at high-volume centers with specific expertise in pediatric renal cancer. Future studies on WT should attempt to discern the role of predisposition syndromes, chemotherapy, radiation and cancer pathology in surgical outcomes. This review suggests that, when indicated, the use of NSS over RN may result in improved kidney function in the long term for WT survivors. However, further research with robust clinical trials and more consistent reporting of potential confounders is required to conclusively determine whether NSS is superior to RN, and which populations are most likely to be affected.
Conclusions
Based on this meta-analysis of low-quality evidence comparing RN vs NSS, utilizing nephron-sparing procedures in patients with unilateral and bilateral WT is associated with a lower incidence of abnormal kidney function. Larger comparative studies, with comprehensive reporting of confounders, are needed to conclusively determine the role of nephrectomy type to prevent negative kidney consequences many years after the treatment of WT.
Acknowledgment
We thank our information expert, Kaitlin Fuller, from the University of Toronto Libraries, for consultation on the search strategy.
References
- 1. : Update on Wilms tumor. J Pediatr Surg 2019; 54: 390. Google Scholar
- 2. : Early and late renal adverse effects after potentially nephrotoxic treatment for childhood cancer. Cochrane Database Syst Rev 2013; 10: CD008944. Google Scholar
- 3. : Early and late adverse renal effects after potentially nephrotoxic treatment for childhood cancer. Cochrane Database Syst Rev 2019; 3: CD008944. Google Scholar
- 4. : Renal failure in Wilms’ tumor patients: a report from the National Wilms’ Tumor Study Group. Med Pediatr Oncol 1996; 26: 75. Google Scholar
- 5. : The management of synchronous bilateral Wilms tumor: a report from the National Wilms Tumor Study Group. Ann Surg 2011; 253: 1004. Google Scholar
- 6. : Oncologic outcomes of partial versus radical nephrectomy for unilateral Wilms tumor. Pediatr Blood Cancer 2012; 58: 898. Google Scholar
- 7. : Renal function in patients with Wilms tumor. Urol Oncol 2016; 34: 33. Google Scholar
- 8. : Pathological review of Wilms tumor nephrectomy specimens and potential implications for nephron sparing surgery in Wilms tumor. J Urol 2012; 188: 1506. Link, Google Scholar
- 9. : Oncology: the role of partial nephrectomy in Wilms tumor. Curr Urol Rep 2013; 14: 350. Google Scholar
- 10. National Cancer Institute: Wilms Tumor and Other Childhood Kidney Tumors Treatment (PDQ®)–Health Professional Version. Available at https://www.cancer.gov/types/kidney/hp/wilms-treatment-pdq. Accessed September 25, 2020. Google Scholar
- 11. : A comparison of renal function outcomes after nephron-sparing surgery and radical nephrectomy for nonsyndromic unilateral Wilms tumor. Urology 2014; 83: 1388. Google Scholar
- 12. : Adverse effects of nephrectomy. Pediatr Nephrol 2016; 31: 1195. Google Scholar
- 13. : Chronic kidney disease in children with unilateral renal tumor. J Urol 2012; 187: 1800. Link, Google Scholar
- 14. : Nephron sparing surgery for unilateral Wilms tumor in children with predisposing syndromes: single center experience over 10 years. J Urol 2012; 188: 1493. Link, Google Scholar
- 15. : Cochrane Handbook for Systematic Reviews of Interventions Version 6.1 (updated September 2020). Available at Training.cochrane.org/handb. Google Scholar
- 16. : The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. MetaArXiv. Available at https://osf.io/preprints/metaarxiv/v7gm2/. Google Scholar
- 17. : Oncological and renal function outcome in children with unilateral Wilms’ tumors treated with nephron sparing surgery or ablative nephrectomy. J Cancer Res Pract 2020; 7: 116. Google Scholar
- 18. : Clinical practice guideline for screening and management of high blood pressure in children and adolescents. Pediatrics 2017; 140: e20171904. Google Scholar
- 19. : Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range. BMC Med Res Methodol 2014; 14: 35. Google Scholar
- 20. : Inclusion of zero total event trials in meta-analyses maintains analytic consistency and incorporates all available data. BMC Med Res Methodol 2007; 7: 5. Google Scholar
- 21. : Measuring inconsistency in meta-analyses. BMJ 2003; 327: 557. Google Scholar
- 22. : Meta-analysis in clinical trials. Control Clin Trials 1986; 7: 177. Google Scholar
- 23. : ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. BMJ 2016; 355: i4919. Crossref, Medline, Google Scholar
- 24. : Cochrane Handbook for Systematic Reviews of Interventions. New York: John Wiley & Sons 2019. Google Scholar
- 25. : Technical considerations for nephron-sparing surgery in children: what is needed to preserve renal units?. J Surg Res 2018; 232: 614. Google Scholar
- 26. : Renal function adaptation in children with unilateral renal tumors treated with nephron sparing surgery or nephrectomy. J Urol 2005; 174: 1404. Link, Google Scholar
- 27. : Evaluation of renal function in pediatric patients after treatment for Wilms’ tumor. Adv Clin Exp Med 2015; 24: 497. Google Scholar
- 28. : Prospective analysis of long-term renal function in survivors of childhood Wilms tumor. Pediatr Nephrol 2017; 32: 1915. Google Scholar
- 29. : Renal function adaptation up to the fifth decade after treatment of children with unilateral renal tumor: a cross‐sectional and longitudinal study. Pediatr Blood Cancer 2013; 60: 1534. Google Scholar
- 30. : Renal function of patients with synchronous bilateral Wilms tumor. Ann Surg 2017; 266: e74. Google Scholar
- 31. : Bilateral Wilms tumors (WT) treated with the SIOP 93 protocol in France: epidemiological survey and patient outcome. Pediatr Blood Cancer 2012; 59: 57. Google Scholar
- 32. : Nephron-sparing partial nephrectomy for bilateral Wilms’ tumor. J Pediatr Surg 2012; 47: 1234. Google Scholar
- 33. : Conservative surgical management of bilateral Wilms tumor: results of the United Kingdom Children’s Cancer Study Group. J Urol 1998; 160: 1450. Link, Google Scholar
- 34. : Bilateral Wilms tumors treated according to the Japan Wilms Tumor Study Group protocol. Pediatr Blood Cancer 2014; 61: 1184. Google Scholar
- 35. : Long-term outcome of bilateral Wilms tumors (BWT). Pediatr Blood Cancer 2011; 56: 1110. Google Scholar
- 36. : Overall survival and renal function of patients with synchronous bilateral Wilms tumor undergoing surgery at a single institution. Ann Surg 2015; 262: 570. Google Scholar
- 37. : Nephron-sparing surgery in synchronous bilateral Wilms’ tumors. J Pediatr Surg 1999; 34: 1505. Google Scholar
- 38. : Renal function outcomes in patients treated with nephron sparing surgery for bilateral Wilms tumor. J Urol 2007; 178: 1786. Link, Google Scholar
- 39. : Development of hypertension is less frequent after bilateral nephron sparing surgery for bilateral Wilms tumor in a long-term survey. J Urol 2015; 193: 262. Link, Google Scholar
- 40. : Retrospective analysis to determine outcomes of patients with bilateral Wilms tumor undergoing nephron sparing surgery: 15-year tertiary single-institution experience. Pediatr Surg Int 2018; 34: 427. Google Scholar
- 41. : Bilateral Wilms’ tumors: single-center experience with 22 cases and literature review. Urology 2010; 76: 946. Google Scholar
- 42. : Treatment of bilateral Wilms’ tumor: comparison of initial biopsy and chemotherapy to initial surgical resection in the preservation of renal mass and function. J Pediatr Surg 1992; 27: 1009. Google Scholar
- 43. : Results of treatment for patients with multicentric or bilaterally predisposed unilateral Wilms tumor (AREN0534): a report from the Children's Oncology Group. Cancer 2020; 126: 3516. Google Scholar
- 44. : Nephron-sparing surgery for Wilms tumor: a systematic review. Urol Oncol 2016; 34: 24. Google Scholar
- 45. : KDIGO 2012 clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney Int 2013; 3: 5. Google Scholar
- 46. : Cardiovascular disease in children with chronic kidney disease. J Am Soc Nephrol 2012; 23: 578. Google Scholar
- 47. : Measurement and estimation of glomerular filtration rate in children. Adv Chronic Kidney Dis 2017; 24: 348. Google Scholar
- 48. : Kidney outcomes and hypertension in survivors of Wilms tumor: a prospective cohort study. J Pediatr 2021; 230: 215. Google Scholar
- 49. : Nephron sparing surgery (NSS) for unilateral Wilms tumor (UWT): the SIOP 2001 experience. Pediatr Blood Cancer 2014; 61: 2175. Google Scholar
Funding: None applicable.
Ethical Approval: Ethical approval was not required for a systematic review.
Conflicts of Interest: The authors declare no relevant conflicts of interest.
