The effectiveness of non-surgical treatments for women with stress urinary incontinence was investigated. The authors concluded that intensive forms of pelvic floor muscle training were effective, but the optimal form of intervention for the NHS needs to be established and definitive research on effectiveness undertaken. The authors' conclusions reflected the evidence, including the uncertainties, and are likely to be reliable.

To assess the clinical effectiveness of non-surgical treatments for women with stress urinary incontinence.

Cochrane Incontinence Group Specialised Register, CINHAL, EMBASE, BIOSIS Previews, Current Controlled Trials, Science Citation Index, Social Science Citation Index, ClinicalTrials.gov and UKCRN Portfolio Database were searched up to early to mid 2008. There were no language restrictions. Websites of relevant organisations were searched. Search terms were reported.

Randomised controlled trials (RCTs) and quasi-RCTs that evaluated nonsurgical treatment for stress urinary incontinence in women were eligible for inclusion. Treatments of interest (alone or in combination) were lifestyle interventions, physical or behavioural therapy and pharmacotherapy. Hormonal treatment was included only if given as an adjunct to another treatment and complementary therapies were excluded. Eligible comparators were no treatment, one of the included interventions and surgical treatment. Studies were included of women with stress urinary incontinence alone (type 1 population), where at least 50% of the women had stress urinary incontinence alone and the remaining had urgency urinary incontinence or mixed urinary incontinence (type 2 population), and where under 50% had stress urinary incontinence alone but at least 50% had mixed urinary incontinence with stress symptoms (type 3 population). Women with incontinence during pregnancy or in the early postpartum period were included.

The primary outcomes of interest were number of women cured, number improved (including cured or improved), adverse events and quality of life.

The included studies evaluated 37 distinct treatments (individual or combined interventions). Most treatments included pelvic floor muscle training as part of the intervention. Pelvic floor muscle training programmes varied considerably between studies. The supervisory intensity provided with the training was classified in the review as basic (up to two clinic visits per month) or with extra sessions (more than two clinic visits per month). The other interventions evaluated were biofeedback, electrical stimulation, vaginal cones (passive and active), motor relearning, modified Pilates, bladder training and serotonin-noradrenaline reuptake inhibitor (SNRI) drug therapy (duloxetine of various doses). A large proportion of included participants came from 11 trials that compared a SNRI with placebo. In some included studies the outcomes cure and improvement were based on patient assessment and in others were based on clinician assessment.

Two reviewers independently assessed studies for inclusion. Disagreements were resolved through consensus or arbitration by a third reviewer.

Two reviewers independently assessed the risk of bias of included studies using an adapted version of the checklist developed by the Cochrane Incontinence Group.

Data were extracted to allow calculation of odds ratios for dichotomous outcomes and standardised mean differences (SMD) for continuous outcomes, each with 95% confidence intervals. The denominator used was the number of participants with the available outcome data.

Data were extracted by one reviewer and checked by a second. Disagreements were resolved by discussion and where necessary by referral to a third reviewer.

A quantitative synthesis was undertaken for primary outcomes only, using data at the end of treatment or at first follow-up (if later). Studies were pooled to derive a summary estimate using a random-effects model. Heterogeneity was assessed using X² and I². Possible reasons for heterogeneity were explored, such as differences in the study populations (type 1, 2 or 3) and the way in which outcomes were assessed. A narrative synthesis was undertaken where quantitative synthesis was not possible. A mixed treatment comparison was undertaken to produce pooled estimates and 95% credible intervals (Cr I) for cure and improvement based on direct and indirect comparisons. Publication bias was not formally assessed.

Eighty-eight studies (n=9,721 participants) were included. Sample size ranged from 11 to 683. Fourteen studies (16%) reported an adequate random allocation sequence generation and allocation concealment.

Based on both direct comparisons and mixed treatment comparison analysis, the treatments were on average more effective than no treatment. Based on the mixed treatment comparison, the most effective treatment was pelvic floor muscle training with intensive supervision (more than two clinic visits per month) or with biofeedback. For pelvic floor muscle training with extra sessions compared to no treatment the odds ratio for cure was 10.7 (95% CrI 5.03 to 26.2) and for pelvic floor muscle training plus biofeedback compared to no treatment the cure odds ratio was 12.3 (95% CrI 5.35 to 32.7). Basic pelvic floor muscle training was better than no treatment only when success was measured in terms of improvement (OR 4.47, 95% CrI 2.03 to 11.9). The models for the mixed treatment comparisons were based on the assumption that patient and clinician assessed success were similar and that that the different time intervals over which an outcome was assessed in the trials was unimportant.

With the exception of SNRI medication, adverse effects were not common.

For cure, the strategy of lifestyle changes and pelvic floor muscle training with extra sessions followed by tension-free vaginal tape procedure had a greater than 70% probability of being cost-effective for all threshold values, for a willingness to pay for a quality adjusted life year (QALY) up to £50,000. For improvement, using the same strategy there was a greater than 50% probability of cost-effectiveness at £10,000 for an additional QALY.

More intensive forms of pelvic floor muscle training were effective and appeared worthwhile. Further research was needed the establish the optimal form of more intensive therapy that was feasible and efficient for the National Health Service (NHS) to provide, as well as further definitive evidence from large well-designed studies.

The review had explicitly defined inclusion criteria. Several relevant sources were searched for studies. There were no language restrictions. Appropriate methods were used to reduce error and bias in the review processes. A detailed account was provided for methods of analysis. Use of a mixed treatment comparison was appropriate given the range of nonsurgical interventions identified in the review.

Inevitably, given the complexity of the analysis, a number of assumptions had to be made about the equivalence of various trial characteristics and the authors discussed these in the report. In particular they highlighted the mixed population, variation between studies in outcome definition and variation between studies in the nature and duration of the pelvic floor training. They also highlighted the need for caution in interpreting the results due to the sparseness of data available for interventions and the wide credible intervals for the treatment effects which suggested uncertainty about the relative effectiveness of the interventions.

The authors' conclusions reflected the evidence presented, including uncertainty in the evidence, and are likely to be reliable.

Practice: The authors stated that there was clear evidence that pelvic floor muscle training plus biofeedback or with extra sessions was effective. Some other treatments seemed promising, but there was insufficient evidence to recommend routine use. Basic pelvic floor muscle training improved symptoms compared to no treatment, but there is no evidence that it was any better in terms of cure.

Research: The authors stated that further research was required to define an optimal form of intensive pelvic floor muscle training that was feasible and efficient for the NHS to provide followed by large trials of the most promising regimens, with long-term follow-up. Any new long-term data should be incorporated into an updated economic model.

National Institute for Health Research (NIHR) Health Technology Assessment programme (project number 06/41/02).

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.