Systematic review of the effectiveness of antenatal intervention for the treatment of congenital lower urinary tract obstruction

Abstract

Please cite this paper as: Morris R, Malin G, Khan K, Kilby M. Systematic review of the effectiveness of antenatal intervention for the treatment of congenital lower urinary tract obstruction. BJOG 2010;117:382–390.

Background  Congenital lower urinary tract obstruction is associated with high mortality and morbidity. Antenatal detection has improved with advances in ultrasound technology, and has allowed the option of antenatal intervention.

Objectives  To systematically review the literature to evaluate the effectiveness of antenatal interventions to improve perinatal survival and postnatal renal function in congenital lower urinary tract obstruction.

Search strategy  Extensive electronic searches (database inception 2009) using Medical Subject Headings (MeSH) and keywords, without restrictions. Reference lists of included studies were checked, and all authors were contacted.

Selection criteria  Studies were selected according to a predefined protocol. The included studies were observational or randomised trials, where an intervention was performed in utero to treat congenital lower urinary tract obstruction, compared with another intervention or no treatment.

Data collection and analysis  Data were extracted on study design, quality and results to construct 2 × 2 tables. Meta-analysis was performed where possible. Peto ORs with 95% CIs were computed.

Main results  Prenatal bladder drainage improved perinatal survival compared with no treatment (OR 3.86, 95% CI 2.00–7.45). This effect was amplified in a subgroup with poor predicted prognosis (OR 12.85, 95% CI 1.25–153.03). However, although treatment increases survival, it appears that the residual risk of poor long-term postnatal renal function is uncertain (OR 0.50, 95% CI 0.13–1.90).

Author’s conclusions  Antenatal bladder drainage appears to improve perinatal survival in cases of congenital lower urinary tract obstruction, but may confer a high residual risk of poor postnatal renal function, based on observational studies. Randomised research with long-term follow up is necessary to determine the role of antenatal treatment in clinical practice.

Background

Congenital lower urinary tract (bladder neck) obstruction (LUTO) comprises a heterogeneous group of conditions: commonly, congenital posterior urethral valves and urethral atresia. The incidence is approximately 2.2 in 10 000 births.¹ The natural history of the condition has a high mortality, up to 45%, largely as a result of pulmonary hypoplasia or severe early onset renal failure.² Those that survive the neonatal period also have a poor long-term prognosis, with obstructive uropathy accounting for large numbers of childhood renal failure.³ The accuracy of antenatal ultrasound for detection of the condition has improved markedly in recent years, with more than 85% of cases detected in utero.⁴

Several prenatal techniques have been employed in attempts to improve the prognosis of the condition through alleviating the obstruction in utero, including serial ultrasound-directed vesicocentesis, vesico-amniotic shunting, fetal cystoscopy and valve ablation. There is a small volume of literature describing open surgical correction of the fetal bladder neck obstruction. The rationale behind these treatments is that restoring the amniotic fluid volume in the second trimester will minimise the risk of pulmonary hypoplasia, and that relieving the urinary tract pressure may attenuate cystic renal parenchymal disease.⁵ However, observational studies have reported mixed results.^6,7 A systematic review performed in 2003 examined the effect of bladder drainage on perinatal survival.⁸ This review, performed in our centre, found evidence of improved survival in fetuses with a poor predicted prognosis who underwent bladder drainage. However, a limited search strategy was employed, and the conclusions were based on a small number of primary studies (four studies in the meta-analysis, including 137 fetuses). We performed an updated and more extensive search, in line with contemporary recommendations. This has included data relating to any antenatal intervention, including cystoscopy and open surgery. In addition, we have extended the scope of the outcome examined to include postnatal renal function, in order to establish the effectiveness of antenatal intervention.

Methods

A protocol-driven systematic review was performed in accordance with recommended methods.^9,10

Identification of studies

We searched MEDLINE (1966–2009), EMBASE (1980–2009), Cochrane Library (2009:1), CINAHL (inception–2009), MEDION, SIGLE, Index of scientific and technical proceedings, DARE, and the British Nursing Index for relevant citations. In MEDLINE the search (Appendix S1) consisted of a combination of Medical Subject Headings (MeSH; e.g. urethral obstruction, prune belly syndrome) ‘or’ keywords (e.g. enlarged bladder, congenital urinary tract obstruction, posterior urethral valves) for disease. These were combined using ‘and’ with MeSH (e.g. fetal therapies) ‘or’ keywords (e.g. vesicoamniotic shunt, fetal cystoscopy) for intervention. This search strategy was adapted for use in the other electronic databases. A comprehensive database was constructed using Reference Manager 11.0. The reference lists of all known primary and review articles were examined to identify cited articles not captured by electronic searches. No language restrictions were applied.

Study selection and data extraction

Initially, the database was scrutinised by two reviewers (RKM or GM, partly in duplicate), and full articles of all citations that were likely to meet the predefined selection criteria were obtained. Translations of articles in languages other than English were obtained. Final inclusion or exclusion decisions were made through examination by two reviewers (GM and RKM), in accordance with the most recent guidance,^9,10 strictly adhering to the following criteria:

• 1
  Population: fetuses with ultrasonographic evidence of lower urinary tract obstruction (enlarged bladder, bilateral hydronephrosis, keyhole sign).
• 2
  Intervention: bladder drainage via vesicocentesis, vesico-amniotic shunt, fetoscopic surgery, e.g. cystoscopy and ablation of valves, or open fetal bladder surgery.
• 3
  Outcome: perinatal mortality, measurements of renal function in survivors (e.g. serum creatinine, need for dialysis or transplantation), or other morbidity indicators (e.g. need for ventilation).
• 4
  Study design: randomised controlled trials, and controlled and uncontrolled observational studies. Case reports and case series of fewer than five were excluded.

Information was extracted on study design, quality and data to construct 2 × 2 tables by two reviewers (GM and RKM, partly in duplicate). The 2 × 2 tables compared intervention with no treatment according to the outcomes examined, or one treatment with another. In order to optimise the completeness of the data and to identify duplications in the population, the authors of the included studies were contacted and asked to provide extra information regarding the individuals included and the techniques used in the interventions. Where duplications in the population existed, the largest data set was included in the meta-analysis. Any disputes regarding data extraction were resolved by input from a third reviewer (KSK).

Quality assessment

All studies included were assessed for quality by the use of the full Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement.¹¹ A randomised controlled trial was considered to be the ideal study design: efforts to minimise bias, adequate reporting of the characteristics of the study population, including potential confounding factors, and an adequate definition of all variables were considered to be important elements for overall study quality.

Data synthesis

Using the 2 × 2 tables we computed Peto ORs for survival and their 95% CIs. The Peto OR is based on the observed number of events and number of events expected if there were no differences between the experimental and control groups. It is particularly useful where the numbers are small, and the event rate is low, as is the case in this review.¹⁰ An Excel spreadsheet was used to summarise all of the results. We performed meta-analysis to explore the effect of antenatal interventions on outcome according to pre-defined subgroups. The effect of any intervention compared with no treatment was computed for each outcome: overall survival, survival excluding voluntary termination of pregnancy (VTP), perinatal survival [excluding VTP and intrauterine death (IUD)], and survival with normal postnatal renal function. Subgroup analyses were performed according to the predicted fetal prognosis (according to antenatal ultrasound features or fetal urinalysis), and comparing treatments with one another. Forest plots were constructed for each group. We inspected for heterogeneity visually and statistically, calculating the Cochrane Q score and the inconsistency square (I²).¹² An I² of >50% was felt to demonstrate significant heterogeneity between studies.¹³ A fixed-effects model was used throughout. Where tables contained cells with the value 0, we excluded the 2 × 2 table from the meta-analysis if the Peto OR could not be calculated. We did not use the recommended method of adding 0.5 to allow for the continuity correction of sparse data¹⁴ because it is usually not required for the Peto OR, and studies where there are no events or all participants experience the event do not add information regarding the relative probability.¹⁰

We constructed a funnel plot to examine for the presence of publication bias, and performed the Egger test to explore this statistically. All statistical analyses were performed using StatsDirect 2.7.2 (http://www.statsdirect.com).

Results

The process of literature searching and study selection is illustrated in Figure 1. There were 23 primary articles that met the selection criteria (Table S1); however, after contact with the authors it was apparent that there was significant overlap in the population in four of these studies.^15–18 We therefore included the largest and most complete data set in meta-analyses,¹⁷ which gave a total of 20 articles with a total of 369 fetuses eligible for inclusion in the meta-analysis.^{1,6,7,19–34} This gave 16 additional studies and 232 additional fetuses from the previous meta-analysis.⁸ A funnel plot was constructed to explore for publication bias. There was no visual funnel plot asymmetry and the Egger test was not statistically significant, suggesting that all relevant studies had been included.

Figure 1.  Process from initial search to final inclusion for the systematic review of effectiveness of antenatal intervention for congenital lower urinary tract obstruction (up to January 2009).

All fetuses included had ultrasound features suggestive of lower urinary tract obstruction (i.e. enlarged fetal bladder with dilated proximal urethra, with or without associated hydronephrosis). The reported gestational age at diagnosis ranged from 13 to 38 weeks. A total of 261 fetuses received an antenatal intervention intended to relieve the obstruction. The majority of these were percutaneous vesico-amniotic shunts (87%). Nine fetuses underwent an open procedure (by maternal laparotomy and hysterotomy), including open shunt insertion, bladder marsupialisation or cutaneous ureterostomy. Twenty-six underwent fetal cystoscopy, and 14 of these had ablation of posterior urethral valves using laser fulguration, urethral stent or hydroablation. Vesicocentesis was considered a diagnostic or therapeutic procedure, depending on the technique used. Where the bladder was completely drained on one or more occasions, this was considered to be a treatment intended to relieve obstruction, and therefore for the purposes of meta-analysis these were included in the intervention group (one fetus underwent serial vesicocentesis and no other procedure). Where a small bladder aspirate was performed for diagnostic purposes, or complete drainage was not described by the authors, this was not considered to be an intervention and the fetuses were included in the ‘no treatment’ group. Fetuses were classified as having a good or poor predicted prognosis by the study authors based on fetal urinalysis results. Criteria for a good predicted prognosis included a urine sodium (Na) concentration of ≤100 mEq/l, chloride (Cl) of ≤90 mEq/l, and osmolarity of ≤200–210 mOsm/l. The results for all studies included in the review are given in Table S2. Eight studies were excluded from the meta-analysis because the number of cells equal to zero in the 2 × 2 tables did not allow for the calculation of the Peto OR.

Overall, the study quality was variable (Figure S1). There were no eligible randomised controlled trials, and the included studies represented a combination of prospective and retrospective cohort studies. Over 80% of studies complied with the STROBE statement elements describing study design, explanation of study size, participant eligibility criteria, patient characteristics and follow up, and number of outcome events.¹¹ However, few studies made efforts to address bias, and most were poor in their reporting of the technique used for the intervention, the overall results and precision of their findings.

There were 12 studies included in the meta-analysis of the effect of any antenatal intervention on perinatal survival compared with no treatment (Figure 2). Antenatal intervention improved perinatal survival (Peto OR 3.82, 95% CI 2.14–6.82, I² 0%). When analysed by subgroups according to predicted prognosis based on fetal urinalysis, within the good prognosis group the direction of effect remained the same, but did not achieve significance (OR 2.58, 95% CI 0.79–8.45, I² = 0). However, where there was a poor prognosis, the size of the effect was increased (OR 26.19, 95% CI 4.39–156.25, I² = 0). The effect on the overall population remained significant when VTPs were excluded from the analysis (OR 2.43, 95% CI 1.18–5.02, I² 5.3%). The size of the effect remained large in the poor prognosis group (OR 8.05, 95% CI 1.23–52.87, I² 0%). When we focused on the effect of antenatal intervention on postnatal survival, excluding VTP and intrauterine death, antenatal intervention only showed a significant benefit in the poor prognosis group (OR 9.36, 95% CI 1.41–62.05).

Figure 2.  Effect of antenatal intervention on perinatal survival compared with no treatment (including voluntary termination of pregnancies) stratified by predicted prognosis.

We examined the effect of antenatal intervention on survival with normal postnatal renal function. Seven studies were included in the analysis (Figure 3). Five studies gave an OR favouring no treatment, and the point estimate summary OR for the overall population suggested that intervention may be associated with survival with impaired renal function, although this was not statistically significant (OR 0.67, 95% CI 0.22–2.00, I² = 0%). When limited to a subgroup with good predicted prognosis based on fetal urinalysis, the direction of the effect favoured intervention (OR 2.98, 95% CI 0.45–19.62, I² 0%), but this did not achieve statistical significance. No fetuses in the predicted poor prognosis group survived with normal renal function; therefore, it was not possible to calculate ORs for this outcome.

Figure 3.  Effect of antenatal intervention compared with no treatment on postnatal survival with normal renal function (excluding intrauterine deaths and voluntary termination of pregnancy) stratified according to predicted prognosis.

We performed subgroup analysis to compare fetal cystoscopy alone with no treatment, or fetal cystoscopy with vesico-amniotic shunting (Table 1). None of the results achieved statistical significance, and are difficult to interpret in view of the relatively small numbers of fetuses who underwent fetal cystoscopy. When these cases and open procedures were excluded from the analysis, and bladder drainage procedures (vesico-amniotic shunt or vesicocentesis) were compared with no treatment, the results were similar (Table 1). The effect for the overall population on perinatal survival favoured treatment (OR 3.86, 95% CI 2.00–7.45), with a large effect in the poor prognosis group (OR 13.85, 95% CI 1.25–153.03). None of the other results changed significantly in effect size or direction.

Table 1.   Summary of results for the systematic review of effectiveness of antenatal intervention for congenital lower urinary tract obstruction

Treatment comparison	Peto OR (95% CI) n = number of included studies
	Perinatal survival			Perinatal survival (excluding VTP)			Postnatal survival (excluding VTP and IUD)			Postnatal survival with normal renal function
	All patients	Good prognosis	Poor prognosis	All patients	Good prognosis	Poor prognosis	All patients	Good prognosis	Poor prognosis	All patients	Good prognosis	Poor prognosis
IUD, intrauterine death; VAS, vesicoamniotic shunt; VC, vesicocentesis; VTP, voluntary termination of pregnancy.
Cystoscopy or open treatment compared with VC only	7.38 (0.15, 372.38) n = 1	–	0.02 (0.00, 1.69) n = 1	–	–	–	–	–	–	–	–	–
Cystoscopy or open treatment compared with VAS	0.65 (0.12, 3.55) n = 2	0.08 (0.00, 1.82) n = 1	–	0.70 (0.11, 4.36) n = 2	0.08 (0.00, 4.49) n = 1	–	0.70 (0.11, 4.36) n = 2	0.08 (0.00, 4.49) n = 1	–	0.78 (0.05, 12.39) n = 1	–	–
VAS and/ or VC compared with no treatment	3.86 (2.00, 7.45) n = 11	2.06 (0.49, 8.70) n = 1	13.85 (1.25, 153.03) n = 1	2.38 (1.06, 5.34) n = 10	1.59 (0.32, 7.81) n = 1	9.89 (0.78, 125.53) n = 1	1.62 (0.65, 4.02) n = 8	0.52 (0.07, 3.92) n = 1	13.08 (1.01, 170.32) n = 1	0.50 (0.13, 1.90) n = 5	0.92 (0.08, 10.55) n = 1	–
Any treatment compared with no treatment	3.82 (2.14, 6.84) n = 12	2.58 (0.79, 8.45) n = 3	26.19 (4.39, 156.25) n = 2	2.43 (1.18, 5.02) n = 11	2.45 (0.61, 9.86) n = 2	8.05 (1.23, 52.87) n = 2	1.82 (0.82, 4.03) n = 9	1.38 (0.26, 7.20) n = 2	9.36 (1.41, 62.05) n = 2	0.67 (0.22, 2.00) n = 7	2.98 (0.45, 19.62) n = 2	–

All studies except three reported the occurrence of procedure-related complications. The most common was shunt dislodgement or occlusion, occurring in 77 (34%) of cases where vesico-amniotic shunt was performed. Two cases of preterm prelabour rupture of membranes were reported, one related to an open fetal procedure. Chorioamnionitis complicated the procedure in four cases. There were three deaths felt to be related to the antenatal intervention, including one intrauterine death, one neonatal death secondary to preterm birth and one neonatal death related to chorioamnionitis. Four traumatic injuries in association with vesicoamniotic shunt were reported, including three hernias at the shunt insertion site and one bladder rupture.

Discussion

The current evidence is limited, but suggests that prenatal intervention in cases of congenital bladder neck obstruction improves perinatal survival, particularly in cases where the predicted prognosis is poor. When considering the effect on long-term survival with normal postnatal renal function, the results suggest a trend towards intervention having adverse sequalae (in terms of morbidity), although this finding was not statistically significant. The evidence presented in this review largely relates to vesico-amniotic shunting, and it is difficult to comment on the potential effects of other treatments, such as fetal cystoscopy (as the sample sizes in these cohort studies were extremely small).

Our group published a review in 2003 looking at the effect of bladder drainage on outcome.⁸ This review included 16 studies and 342 fetuses; however, only four studies and 137 fetuses were included in the meta-analysis because of the difficulty the authors had in correctly identifying the effectiveness data. Since the publication of this review there have been significant methodological advances in search strategies, quality assessment and statistical analysis in this area. We thus felt it appropriate to repeat this review using these techniques, and in addition contacted all the authors of the original papers to ensure the correct and most complete data was obtained. Contact with the authors allowed us to obtain complete data for all 20 studies (369 fetuses) included in this review, thus they could all be included in the meta-analysis, and to accurately identify studies that were duplications. We were also able to clarify where vesicocentesis had been employed (as well as) or instead of shunting, allowing us to perform a much more detailed analysis according to the antenatal intervention performed as shown in Table 1. This review thus represents the best and most contemporaneous evidence available at the time of publication regarding the efficacy of antenatal intervention in cases of congenital lower urinary tract obstruction. The strength of our review lies in the robustness of our methodology. We have complied with current guidelines on the performance and reporting of systematic reviews.^10,12,35 An extensive literature search was performed in multiple databases without language restrictions. Funnel plot analysis reassured us that there was no evidence of missing studies. Although there was clinical heterogeneity between the included studies regarding the interventions and size and direction of the effect, there was no statistical heterogeneity within our meta-analysis.

There are several limitations to our review. The overall quality of included studies was variable, and all included studies represented small cohorts or case series (Figure S1). The reporting of interventions was generally inadequate, which was a particular limitation when considering vesicocentesis, which may be a diagnostic or therapeutic procedure. Although we requested further information and clarification from all authors regarding this point, none were able to provide the original data set, and those that responded were only able to provide general information about the techniques used. We only considered vesicocentesis to be a therapeutic procedure if the authors stated that the bladder was fully drained on one or more occasions. We acknowledge the possibility that this technique may have been used in some of the fetuses included in the ‘no intervention’ group, where this was inadequately described in the manuscript. However, if this is the case, the observed effect size in favour of treatment would be less than the actual effect; therefore, we do not believe this affects the validity of our finding that antenatal intervention increases the odds of perinatal survival. Unfortunately, because of inadequate reporting and paucity of information, it was not possible to perform a separate comparison between fetuses where no invasive procedure was performed and those who had a diagnostic vesicocentesis. Although this issue may be addressed by individual patient data meta-analysis,³⁶ this is impossible in this case, as most authors no longer have the primary data sets.

The ability of prenatal intervention to relieve obstruction and improve perinatal survival in cases of congenital bladder neck obstruction is biologically plausible. A large number of neonatal deaths were secondary to pulmonary hypoplasia. The crucial gestation period for fetal human lung development is during the canalicular phase of development, between 16 and 25 weeks of gestation. Severe oligohydramnios occurring for more than 14 days before 25 weeks of gestation is associated with neonatal mortality of >90%.³⁷ We were unable to assess the influence of gestational age at the time of the procedure within our review, because of the lack of reporting in primary studies.

Animal models have suggested a benefit of antenatal intervention for renal development. Fetal urethral obstruction produces hydronephrosis in the kidneys of sheep fetuses, similar to that seen in human fetuses with LUTO, and relief of the obstruction by bladder marsupialisation lessens the degree of renal tract dilatation seen.^5,38 Experimental obstruction of a single ureter in fetal lambs produced unilateral renal dysplastic changes similar to those seen in human fetuses with LUTO, with decompression preventing or ameliorating the development of dysplasia.^39,40 However, they were unable to replicate these observations with bladder outlet obstruction, and human studies have failed to show a consistent benefit of antenatal intervention for renal function in human fetuses with LUTO.^7,18

Our review could not show any benefit of antenatal intervention for preserving normal renal function: the results suggest that odds of survival with normal renal function may be reduced. It has been postulated that the dysplastic changes seen in fetal kidneys may be part of a different pathological process, rather than solely caused by obstruction.⁴¹ Our findings would support this: the apparent reduction in survival with normal renal function following antenatal intervention may be caused by increased numbers surviving with severe LUTO and renal complications, who would otherwise have died from pulmonary hypoplasia. It must be noted, however, that although the point summary estimate suggested a risk of poor long-term renal function in fetuses undergoing antenatal intervention, the evidence was uncertain and based on observational data.

Our results highlight the need for further primary randomised research to clarify this issue, and the role of antenatal intervention for the relief of LUTO in clinical practice. The percutaneous shunting in lower urinary tract obstruction (PLUTO) is a randomised controlled trial comparing in utero vesico-amniotic shunting with conservative management and long-term follow-up. Recruitment is ongoing and expected to continue until 2013. Current National Institute for Clinical Excellence (NICE) interventional procedure guidance recommends that fetal cystoscopy and vesico-amniotic shunting should not be performed outside a research context, and recommends randomisation to the PLUTO trial.^42,43

The results of our review suggest that antenatal intervention to relieve obstruction in cases of congenital LUTO may improve perinatal survival, particularly in fetuses with poor predicted prognosis on fetal urinalysis. However, because of the lack of evidence of benefit for survival with normal renal function, we believe that current evidence is insufficient to dictate clinical practice.

Disclosure of interests

We declare no conflicts of interest.

Contribution to authorship

All authors were responsible for the design of the study. GM and RKM were responsible for the data extraction and analysis. All authors checked the analysis and were involved in the drafting and critical revision of the manuscript.

Details of ethics approval

Ethical approval was not required.

Funding

This work is funded by Wellbeing of Women, the Hannah Eliza Guy Charity (Birmingham Children’s Hospital Charity) and the NIHR HTA programme. RKM is funded by an MRC/RCOG Clinical Research Training Fellowship. GM is funded by the Mary Crosse Fellowship (Birmingham Women’s Foundation NHS Trust Research and Development).

Acknowledgements

We acknowledge the help of Mr Justin Clarke and authors of the review published in 2003 for providing further information regarding their review.⁸