Urinary incontinence (involuntary or abnormal urine loss) is a common condition affecting an estimated 423 million individuals worldwide (Irwin et al, 2011). Its prevalence varies, ranging from 5–15.2% in Asian populations, 1.8–30.5% in European populations and 1.7–36.4% in US populations (Milsom et al, 2014). This condition can have a significant clinical (Yang et al, 2018), psychological (Sims et al, 2011) and financial impact (Thom et al, 2010) on an individual's life. The development of urinary incontinence is associated with several non-modifiable risk factors, such as age (Milsom and Gyhagen, 2019), gender (Nitti, 2001) and family history (von Gontard et al, 2011), as well as modifiable risk factors of weight (Aune et al, 2019), smoking (Kawahara et al, 2020), diet (Maserejian et al, 2010) and caffeine intake (Gleason et al, 2013). Despite the link between weight, smoking, diet and caffeine intake and urinary incontinence, there is limited evidence supporting interventions that aim to affect these modifiable risk factors (Imamura et al, 2015). The Cochrane systematic review discussed in this commentary (Imamura et al, 2015) aimed to evaluate the effectiveness of a range of lifestyle-focused interventions that are commonly used in the management of urinary incontinence.
This commentary critically appraises the methods used in the Cochrane systematic review and considers the importance of the findings for clinical practice.
Methods
This Cochrane systematic review undertook a comprehensive search of the Cochrane Central Register of Controlled Trials and MedLine using a previously validated search procedure from date of inception to 3 July 2013. Additional handsearching, citation list and clinical trial registry searches were performed. There were no language restrictions placed on the search procedure. Only randomised control trials (RCTs) or quasi-randomised trials that compared lifestyle interventions to no active control treatment for adults with urinary incontinence were included. A post-hoc decision was made to include studies where not all participants had urinary incontinence on entering the trial.
A robust study selection, data extraction and assessment of risk of bias (Cochrane assessment of bias criteria) was undertaken independently by two reviewers, with disagreements being resolved by discussion (Higgins et al, 2020). A summary statement of quality for each outcome was produced using the GRADE criteria (Guyatt et al, 2008). A meta-analysis was undertaken using a fixed-effect model, estimating relative risks for dichotomous outcomes and Mantel-Haenszel weighted mean differences for continuous outcomes. A range of subgroup analyses were planned but were not undertaken due to a lack of data. A narrative synthesis was undertaken for adverse events due to a similar lack of numerical data.
Findings
Eleven studies were included of mixed design, of which six were RCTs, four were crossover trials and one was a quasi-randomised trial. Only two studies had a low risk of bias on four out of the seven criteria for assessment of bias, namely, randomisation, allocation concealment and blinding of outcome assessment and incomplete outcome data. The main areas of concern were a lack of blinding of participants, appropriate allocation concealment and selective reporting bias due to limited protocols being identified for the included studies. Finally, of the 11 studies included, seven had participants who did not have urinary incontinence at baseline. With the percentage of the population included having urinary incontinence ranging from 27% to 83% at baseline and one study where it was not reported but assumed that patients had some level of urinary incontinence at baseline.
The review found that, with weight loss interventions, adults with urinary incontinence were statistically and clinically significantly more likely to report improvements in symptoms, incontinence impact questionnaire, cure rates by symptom quantification (low-quality evidence: one study) and improvement rates by symptom quantification compared with non-active controls (low-quality evidence: two studies). On visual inspection of the forest plot for risk ratio of quantification of symptoms, there was some evidence that weight loss interventions become less effective over time (18 months follow-up), with the main effect moving towards the line of no effect. There was also a statistically and clinically significant reduction in the prevalence of weekly urinary incontinence (very low-quality evidence: one study). One study reported that the intervention had few side effects.
One study demonstrated a statistical and clinically significant reduction in the prevalence of weekly urinary incontinence for weight loss compared with the pharmacological intervention of metformin. For the intervention of restricting fluid intake, there were mixed findings. One study of three demonstrated a statistically significant reduction in urinary incontinence within a week (very low quality of evidence), and one study demonstrated improvement in quality of life compared with baseline measurements (very low quality of evidence). Two studies reported adverse events that included concentrated urine, constipation, headache and thirst. Evidence on the effects of caffeine reduction and a soy-rich diet was limited and of low quality, preventing any conclusions.
Commentary
Although the systematic review appeared robust when assessed using the Amstar2 critical appraisal tool (only considered inadequate on the basis of study type selection), it had an important limitation that rendered it of low confidence (Shea et al, 2017). A post-hoc decision to include data from trials where not all participants were incontinent when they entered the trial reduced the relevance of its findings to adults with urinary incontinence. It is unclear as to when this decision was made, which may have led to important studies being missed during study selection. This limitation is further compounded by the limited and low-quality evidence identified by the systematic review. Consequently, these findings should be viewed with caution.
Based on this limited evidence, there are signs that weight loss interventions may be considered as an option for a lifestyle treatment for urinary incontinence. Previous systematic reviews have shown a positive association with increased body mass index (BMI) and risk of urinary incontinence (Lamerton et al, 2018; Aune et al, 2019). Furthermore, previous reviews have shown that weight management interventions can be cost-effective (Forster et al, 2011; Loveman et al, 2011; Finkelstein and Verghese, 2019). However, the Cochrane systematic review found that there was some evidence to suggest that the intervention of weight loss maybe less effective over time. This should be considered when recommending this intervention as regular assessments may be required to ensure continuation of effectiveness.
There was limited evidence to suggest a benefit from reducing fluid intake, despite previous reviews finding the association between fluid intake and urinary frequency and urgency in men and women (Bradley et al, 2017). A range of adverse effects were also identified. As such, reduced fluid intake interventions should be considered with caution.
There was limited or inconsistent evidence on the effectiveness of caffeine reduction or a soy-rich diet, and no evidence found for physical activity, sweetened fizzy drinks or diet drinks, smoking cessation, constipation and alcohol. Even though there are inconsistencies in the evidence base identified in this Cochrane review, the current National Institute for Health and Care Excellence (NICE) guidelines recommend that caffeine reduction, fluid modification and weight loss management should be considered (NICE, 2015; 2019). Where these uncertainties still exist, there are other important aspects that can be taken into consideration, such as sensitivity to the intervention, type of urinary incontinence, patient preference and any caregiver needs (Demaagd and Davenport, 2012; Lobchuk and Rosenberg, 2014; Lukacz et al, 2017).
Further high-quality RCTs are required to assess the effectiveness of lifestyle interventions for adults with urinary incontinence. Priority should be given to weight loss management due to this intervention type currently having the strongest evidence of effect. Due to the sparsity of the evidence, the areas of physical activity, sweetened fizzy drinks or diet drinks, smoking cessation, constipation, and alcohol should be explored. All future RCTs in this area should ensure that there is consistency in outcomes used by adopting relevant urinary incontinence outcome sets for the type of urinary incontinence (Comet Initiative, 2020). Due to the scoping issues of this review and the out-of-date search procedure, an update of this Cochrane review is required.
KEY POINTS
- It has been indicated that weight loss management may be effective in improving symptoms of urinary incontinence, quality of life and prevalence of weekly urinary incontinence
- There is limited evidence on the effectiveness of caffeine reduction and a soy-rich diet, and no evidence of the efficacy of physical activity, sweetened fizzy drinks or diet drinks, smoking cessation, constipation and alcohol on urinary incontinence in adults
- Further high-quality random controlled trials are required to assess the effectiveness of lifestyle interventions for adults with urinary incontinence
- Future research on the effectiveness of lifestyle interventions for adults with urinary incontinence should use specific urinary incontinence outcome sets
CPD REFLECTIVE QUESTIONS
- What was the main limitation of this Cochrane systematic review?
- What factors should be considered when prescribing lifestyle interventions for adults with urinary incontinence?
- What additional strategies can be put in place if uncertainties lie within the effectiveness of an intervention?