Conservative interventions for treating functional daytime urinary incontinence in children

Abstract

Background

In children, functional daytime urinary incontinence is the term used to describe any leakage of urine while awake that is not the result of a known underlying neurological or congenital anatomic cause (such as conditions or injuries that affect the nerves that control the bladder or problems with the way the urinary system is formed). It can result in practical difficulties for both the child and their family and can have detrimental effects on a child’s well‐being, education and social engagement.

Objectives

To assess the effects of conservative interventions for treating functional daytime urinary incontinence in children.

Search methods

We searched the Cochrane Incontinence Specialised Register, which contains studies identified from CENTRAL, MEDLINE, MEDLINE In‐Process, MEDLINE Epub Ahead of Print, CINAHL, ClinicalTrials.gov, WHO ICTRP and handsearching of journals and conference proceedings (searched 11 September 2018). We also searched Chinese language bibliographic databases: Chinese Biomedical Literature Database (CBM), China National Knowledge Infrastructure (CNKI), and Wanfang. No language restrictions were imposed.

Selection criteria

We included randomised controlled trials (RCTs), quasi‐randomised, multi‐arm studies, cross‐over studies and cluster‐randomised studies that included children aged between 5 and 18 years with functional daytime urinary incontinence.

Data collection and analysis

Two review authors independently screened records and determined the eligibility of studies for inclusion according to predefined criteria. Where data from the study were not provided, we contacted the study authors to request further information. Two review authors assessed risk of bias and processed included study data as described in the Cochrane Handbook for Systematic Reviews of Interventions. Where meta‐analysis was possible, we applied random‐effects meta‐analysis using the Mantel‐Haenszel method for dichotomous outcomes.

Main results

The review included 27 RCTs involving 1803 children. Of these, six were multi‐arm and one was also a cross‐over study. Most studies were small, with numbers randomised ranging from 16 to 202. A total of 19 studies were at high risk of bias for at least one domain. Few studies reported data suitable for pooling due to heterogeneity in interventions, outcomes and measurements.

Individual conservative interventions (lifestyle, behavioural or physical) versus no treatment

Transcutaneous electrical nerve stimulation (TENS) versus sham (placebo) TENS. More children receiving active TENS may achieve continence (risk ratio (RR) 4.89, 95% confidence interval (CI) 1.68 to 14.21; 3 studies; n = 93; low‐certainty evidence).

One individual conservative intervention versus another individual or combined conservative intervention

Pelvic floor muscle training (PFMT) with urotherapy versus urotherapy alone. We are uncertain whether more children receiving PFMT with urotherapy achieve continence (RR 2.36, 95% CI 0.65 to 8.53, 95% CI 25 to 100; 3 studies; n = 91; very low‐certainty evidence).

Voiding education with uroflowmetry feedback and urotherapy versus urotherapy alone. Slightly more children receiving voiding education with uroflow feedback and urotherapy may achieve continence (RR 1.13, 95% CI 0.87 to 1.45; 3 studies; n = 151; low‐certainty evidence).

Urotherapy with timer watch versus urotherapy alone. We are uncertain whether urotherapy plus timer watch increases the number of children achieving continence compared to urotherapy alone (RR 1.42, 95% CI 1.12 to 1.80; 1 study; n = 58; very low‐certainty evidence).

Combined conservative interventions versus other combined conservative interventions

TENS and standard urotherapy versus PFMT with electromyographic biofeedback and standard urotherapy. We are uncertain whether there is any evidence of a difference between treatment groups in the proportions of children achieving continence (RR 1.11, 95% CI 0.73 to 1.68; 1 study; n = 78; very low‐certainty evidence).

PFMT with electromyography biofeedback and standard urotherapy versus PFMT without feedback but with standard urotherapy. We are uncertain whether there is any evidence of a difference between treatment groups in the proportions of children achieving continence (RR 1.05, 95% CI 0.72 to 1.52; 1 study; n = 41; very low‐certainty evidence).

Individual conservative interventions versus non‐conservative interventions (pharmacological or invasive, combined or not with any conservative interventions)

PFMT versus anticholinergics. We are uncertain whether more children receiving PFMT than anticholinergics achieve continence (RR 1.92, 95% CI 1.17 to 3.15; equivalent to an increase from 33 to 64 per 100 children; 2 studies; n = 86; very low‐certainty evidence).

TENS versus anticholinergics. We are uncertain whether there was any evidence of a difference between treatment groups in the proportions of children achieving continence (RR 0.81, 95% CI 0.05 to 12.50; 2 studies; n = 72; very low‐certainty evidence).

Combined conservative interventions versus non‐conservative interventions (pharmacological or invasive, combined or not with any conservative interventions)

Voiding education with uroflowmetry feedback versus anticholinergics. We are uncertain whether there was any evidence of a difference between treatment groups in the proportion of children achieving continence (RR 1.02, 95% CI 0.58 to 1.78; 1 study; n = 64; very low‐certainty evidence).

Authors' conclusions

The review found little reliable evidence that can help affected children, their carers and the clinicians working with them to make evidence‐based treatment decisions. In this scenario, the clinical experience of individual clinicians and the support of carers may be the most valuable resources. More well‐designed research, with well‐defined interventions and consistent outcome measurement, is needed.

Plain language summary

Conservative (non‐pharmaceutical and non‐surgical) treatments for children who have daytime urinary incontinence

Background

Many children continue to have daytime wetting accidents (urinary incontinence) long after most become dry, some into teenage years. When there is no known cause, it is referred to as functional daytime urinary incontinence. Daytime wetting can be a practical problem for the child and their carers, leading to additional laundry, expenses on absorbent pads or underwear, wet or stained furnishings and difficulties attending school or travelling. It can also lead to emotional stress, poor school attendance and performance, difficulties with social activities and making friends, and even depression and behaviour problems. Finding an effective treatment can improve the quality of life of children and their carers.

Review question

Children, their carers, doctors and nurses may want to avoid treatments that involve drugs or surgery, at least until they have tried treatments that do not involve these. In this review researchers brought together the results of studies conducted around the world to combine the evidence and assess which treatment(s) are most effective for managing daytime urinary incontinence in children.

How up‐to‐date is this review?

The evidence is current to 11 September 2018.

Study characteristics and funding sources

The review identified 27 studies involving 1803 children.

Twelve studies declared funding from outside sources. One of these was funded by Astellas Pharma, while another study was provided with special batches of placebo and medication free of charge, as well as materials for pad tests free of charge from another company.

Certainty of evidence

Cochrane Reviews assess the 'certainty' or reliability of the evidence using standardised methods that consider the way studies were designed, conducted and reported, differences between studies or populations, and the combined results of studies. Most of the studies identified for this review were small and many were poorly designed and not reported clearly. Most of the evidence was considered to be of very low certainty, meaning that little can be said with any certainty about the effectiveness of treatments.

Key results

Transcutaneous electrical nerve stimulation (TENS) may be more effective than no treatment for ending or reducing daytime wetting.

We are uncertain whether urotherapy (behavioural programmes in which children ‐ and sometimes carers ‐ are taught about how the bladder works, proper toileting postures and methods, scheduled toileting, and planning what and how much to drink) is more effective when supplemented with PFMT, voiding education with feedback, or watches with alarms set to remind children when to go to the toilet.

We are uncertain whether feedback that shows children how their muscles are working or how their bladder is emptying improves the effectiveness of TENS with urotherapy compared to PFMT plus feedback and urotherapy. We are also uncertain whether PFMT and urotherapy plus feedback improves the effectiveness of PFMT and urotherapy alone.

We are uncertain whether pelvic floor muscle training (PFMT) or TENS are more effective than anticholinergics (drugs that can reduce signals from the brain that cause the bladder to contract and empty).

We are uncertain whether voiding education plus uroflowmetry (a test to measure the volume of urine) and feedback increases the number of continent children compared to anticholinergics.

No serious adverse events were reported that were considered to be related to study treatments. Most non‐serious adverse events and side effects were mild or moderate in severity and were in children receiving pharmaceutical interventions. These included common pharmaceutical side effects such as nausea, abdominal pain, dry mouth, drowsiness and headache.

Authors' conclusions

There is a lack of good‐quality research evidence that can help children, their carers and doctors and nurses to make decisions about treatments. More well‐designed research may provide much needed evidence about the effectiveness of promising interventions in children with daytime urinary incontinence, such as TENS, PFMT and timers on watches (or mobile phones) to remind children about toileting schedules. However, it is hoped that this review will draw attention to the need for research into effective treatments for daytime wetting in children.

Summary of findings

Summary of findings for the main comparison. Transcutaneous electrical nerve stimulation (TENS) versus placebo (sham TENS) for functional daytime urinary incontinence in children.

Transcutaneous electrical nerve stimulation (TENS) versus placebo (sham TENS) for functional daytime urinary incontinence in children
Patient or population: children aged 3 to 18 years with daytime urinary incontinence Settings: four studies in Brazil, India and Denmark Intervention: transcutaneous electrical nerve stimulation (TENS), either parasacral or tibial Comparison: sham TENS
Outcomes	Study event rates*		Relative effect (95% CI)	No of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Placebo	TENS
Resolution of daytime urinary incontinence after 7 to 20 weeks' treatment	4/42 complete resolution of daytime urinary incontinence	33/51 complete resolution of daytime urinary incontinence	RR 4.89 (1.68 to 14.21)	93 (3)	⊕⊕⊝⊝ Low1,2	No assumed risk or corresponding risk were estimated due to severe imprecision, inconsistency (masked by imprecision) and high risk of bias
Frequency of incontinence episodes	See comment	See comment	See comment	78 (3)	⊕⊝⊝⊝ Very low2,3,4	Three studies reported improvements in terms of reduced frequency of incontinence episodes in the TENS groups (de Paula 2017; Hagstroem 2009; Lordelo 2010). No assumed risk or corresponding risk were estimated due to unsuitable data (different measurements).
Objective measurement of volume of incontinent episodes	No data	No data	No data	No data	N/A
Resource implications (e.g. cost of incontinence management methods)	No data	No data	No data	No data	N/A
Quality of life measurements	No data	No data	No data	No data	N/A
Adverse effects	No data	No data	No data	No data	N/A
*No assumed risk or corresponding risk were estimated for reasons outlined in comments. CI: confidence interval; RR: risk ratio; TENS: transcutaneous electrical nerve stimulation
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

¹Downgraded one level for imprecision: small sample size, wide confidence intervals.
 ²Downgraded one level for high or unclear risk of bias in several domains across studies.
 ³Downgraded one level for serious imprecision: small sample size.
 ⁴Downgraded one level for inconsistency: differing outcome measurements.

Summary of findings 2. Pelvic floor muscle training (PFMT) with urotherapy versus urotherapy alone for functional daytime urinary incontinence in children.

Pelvic floor muscle training (PFMT) with urotherapy versus urotherapy alone for functional daytime urinary incontinence in children
Patient or population: children aged 3 to 17 years with daytime urinary incontinence Settings: two studies from Iran (paediatric urology clinic) and Serbia (physical rehabilitation clinic) Intervention: pelvic floor muscle training (PMFT) with urotherapy Comparison: urotherapy
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Urotherapy	PFMT with urotherapy
Resolution of daytime urinary incontinence after 12 months of treatment	21 per 100	80 per 100 (25 to 100)	RR 2.36 (0.65 to 8.53)	91 (3)	⊕⊝⊝⊝ Very low1,2,3	The studies included children with dysfunctional voiding symptoms, some without daytime urinary incontinence. Data were available for the subgroups with daytime urinary incontinence and only these are considered in this review.
Frequency of incontinence episodes at 12 months	See comment	See comment	See comment	See comment	⊕⊝⊝⊝ Very low4,5	One study, not included in the outcome above, reported this outcome (Nemett 2008). The study compared osteopathy‐based physical therapy alongside standard urological care with standard urological care alone. 6/6 children in the osteopathy group and 6/8 in the standard‐care‐alone group reported improvement in daytime incontinence. Nemett 2008 did not report additional treatments provided as indicated during standard care by urologists so we cannot determine any effect of the osteopathy‐based physical therapy.
Objective measurement of volume of incontinent episodes	No data	No data	No data	No data	N/A
Resource implications (e.g. cost of incontinence management methods)	No data	No data	No data	No data	N/A
Quality of life measurements	No data	No data	No data	No data	N/A
Adverse effects	No data	No data	No data	No data	N/A
*The assumed risk is based on the median control group risk across studies. CI: confidence interval; PFMT: pelvic floor muscle training; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

¹Downgraded one level for imprecision: small sample size, very wide confidence intervals.
 ²Downgraded one level for inconsistency: substantial heterogeneity.
 ³Downgraded one level for high risk of bias in two domains in all studies and unclear risk of bias in other domains.
 ⁴Downgraded one level for imprecision: small sample size.
 ⁵Downgraded two levels for high risk of bias in four domains and unclear risk in three.

Summary of findings 3. Voiding education with uroflow feedback and urotherapy versus urotherapy alone for functional daytime urinary incontinence in children.

Voiding education with uroflow feedback and urotherapy versus urotherapy alone for functional daytime urinary incontinence in children
Patient or population: children aged 5 to 16 years with daytime urinary incontinence Settings: three studies in the Netherlands and Iran (paediatric urology units) and Belgium, Germany, Netherlands, Sweden (multicentre study in outpatient clinics) Intervention: voiding education with uroflow feedback and urotherapy Comparison: urotherapy alone
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Urotherapy	Bladder re‐education with feedback
Resolution of daytime urinary incontinence at 6 to 12 months after 6 months treatment	54 per 100	61 per 100 (47 to 78)	RR 1.13 (0.87 to 1.45)	151 (3)	⊕⊕⊝⊝ Low1,2	One study additionally provided PFMT to the children in the re‐education with feedback group (Kajbafzadeh 2011). The relative effect was little altered when we removed this study from the analysis (RR 1.18, 95% CI 0.88 to 1.57). Two of the studies included children without daytime urinary incontinence. Data were available for the subgroups with daytime urinary incontinence and only these are considered in this review.
Frequency of incontinence episodes	No data	No data	No data	No data	N/A
Objective measurement of volume of incontinent episodes	No data	No data	No data	No data	N/A
Resource implications (e.g. cost of incontinence management methods)	No data	No data	No data	No data	N/A
Quality of life measurements	No data	No data	No data	No data	N/A
Adverse effects	N/A	N/A	N/A	133 (2)	N/A	Van Gool 2014 reported 1 child with nervous tics and 1 with constipation in the placebo arm and none in the voiding education with feedback arm. Klijn 2006 reported that there were no adverse events in any study arm.
*The assumed risk is the median control group risk across studies. CI: confidence interval; PFMT: pelvic floor muscle training; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

¹Downgraded one level for imprecision: confidence intervals that include both benefit and detriment.
 ²Downgraded one level for high or unclear risk of bias in several domains across studies.

Summary of findings 4. Timer watch plus standard urotherapy versus standard urotherapy alone for functional daytime urinary incontinence in children.

Timer watch plus standard urotherapy versus standard urotherapy alone for functional daytime urinary incontinence in children
Patient or population: children aged 5 to 14 years with overactive bladder and at least one episode of daytime urinary incontinence per week Settings: one study at a university hospital children's incontinence clinic in Denmark Intervention: standard urotherapy plus timer watches with alarms set to remind children of the voiding schedule Comparison: standard urotherapy
Outcomes	Study event rates*		Relative effect (95% CI)	No of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Standard urotherapy	Timer watches
Resolution of daytime urinary incontinence after 3 months of treatment	0/28 complete resolution of daytime urinary incontinence	9/30 complete resolution of daytime urinary incontinence	RR 17.77 (1.08 to 291.82)	58 (1)	⊕⊝⊝⊝ Very low1,2	No assumed risk or corresponding risk were estimated due to single study with risk of bias and small sample size.
Frequency of incontinence episodes at 3 months	Mean (SD) increase in dry days per week 0.6 (1.8)	Mean (SD) increase in dry days per week 3.5 (2.5)	MD 2.90 (1.78 to 4.02)	58 (1)	⊕⊝⊝⊝ Very low1,2	No assumed risk or corresponding risk were estimated due to single study with risk of bias and small sample size.
Objective measurement of volume of incontinent episodes	No data	No data	No data	No data	N/A
Resource implications (e.g. cost of incontinence management methods)	No data	No data	No data	No data	N/A
Quality of life measurements	No data	No data	No data	No data	N/A
Adverse effects	No data	No data	No data	No data	N/A
*No assumed risk or corresponding risk were estimated for reasons outlined in comments. CI: confidence interval; MD: mean difference; RR: risk ratio; SD: standard deviation
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

¹Downgraded one level for imprecision: small sample size.
 ²Downgraded two levels for high risk of bias in one domain and unclear risk of bias in four domains.

Summary of findings 5. Transcutaneous electrical nerve stimulation (TENS) and standard urotherapy versus pelvic floor muscle training (PFMT) with electromyographic biofeedback and standard urotherapy for functional daytime urinary incontinence in children.

Transcutaneous electrical nerve stimulation (TENS) and standard urotherapy versus pelvic floor muscle training (PFMT) with electromyographic biofeedback and standard urotherapy for functional daytime urinary incontinence in children
Patient or population: children aged 5 to 16 years with daytime urinary incontinence Settings: one study in Brazil Intervention: parasacral transcutaneous electrical nerve stimulation (TENS) with urotherapy Comparison: pelvic floor muscle training with feedback and urotherapy
Outcomes	Study event rates*		Relative effect (95% CI)	No of participants (studies)	Certainty of the evidence (GRADE)	Comments
	PFMT	TENS
Resolution of daytime urinary incontinence	17/31 complete resolution of daytime urinary incontinence	20/33 complete resolution of daytime urinary incontinence	RR 1.11 (0.73 to 1.68)	64 (1)	⊕⊝⊝⊝ Very low1,2	No assumed risk or corresponding risk were estimated due to single study with high risk of bias and small sample size.
Frequency of incontinence episodes	See comments	See comments	See comments	64 (1)	⊕⊝⊝⊝ Very low1,2	There were significant changes in mean number of urinary incontinence episodes per day in both the TENS group (from 1.85 (SD 1.48) to 0.12 (SD 0.33); P = 0.001) and in the PFMT with biofeedback group (from 1.94 (SD 1.50) to 0.14 (SD 0.39); P = 0.001), but no difference between groups
Objective measurement of volume of incontinent episodes	No data	No data	No data	N/A	N/A
Resource implications (e.g. cost of incontinence management methods)	No data	No data	No data	N/A	N/A
Quality of life measurements	See comments	See comments	See comments	64 (1)	⊕⊝⊝⊝ Very low1,2	The study reported no difference in the quality of life of children in either group before and after treatment, measured using the AUQUEI questionnaire (Assumpcao 2000).
Adverse effects	No data	No data	No data	N/A	N/A
*No assumed risk or corresponding risk were estimated for reasons outlined in comments. AUQUEI: Autoquestionnaire Qualité de Vie Enfant Imagé; CI: confidence interval; PFMT: pelvic floor muscle training; RR: risk ratio; SD: standard deviation; TENS: transcutaneous electrical nerve stimulation
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

¹Downgraded one level for imprecision: small sample size and confidence intervals that include both benefit and detriment.
 ²Downgraded two levels for high risk of bias in three domains and unclear risk of bias in two domains.

Summary of findings 6. Pelvic floor muscle training (PFMT) with electromyography biofeedback and standard urotherapy versus pelvic floor muscle training without biofeedback but with standard urotherapy for functional daytime urinary incontinence in children.

Pelvic floor muscle training (PFMT) with electromyography biofeedback and standard urotherapy versus pelvic floor muscle training without biofeedback but with standard urotherapy for functional daytime urinary incontinence in children
Patient or population: children aged 5 to 15 years with dysfunctional elimination syndrome, 41/59 with daytime urinary incontinence Settings: one study in Brazil (nephro‐urology unit) Intervention: pelvic floor muscle training (PFMT) with electromyography biofeedback and urotherapy Comparison: pelvic floor muscle training (PFMT) without biofeedback and urotherapy
Outcomes	Study event rates*		Relative effect (95% CI)	No of participants (studies)	Certainty of the evidence (GRADE)	Comments
	PFMT	PFMT with feedback
Resolution of daytime urinary incontinence at 12 months (9 months after 3 months of treatment)	15/21 complete resolution of daytime urinary incontinence	15/20 complete resolution of daytime urinary incontinence	RR 1.05 (0.72 to 1.52)	41 (1)	⊕⊝⊝⊝ Very low1,2	The study included children with dysfunctional voiding symptoms, some without daytime urinary incontinence. Data were available for the subgroup with daytime urinary incontinence and only these are considered in this review. It should be noted that children in the feedback arm received fewer sessions (16) than the no‐feedback arm (24). No assumed risk or corresponding risk were estimated due to single study with risk of bias and small sample size.
Frequency of incontinence episodes at 12 months	See comment	See comment	See comment	41 (1)	⊕⊝⊝⊝ Very low1,2	Similar proportions of children in each treatment group who did not achieve continence reported ≥ 50% reduction in incontinence episodes.
Objective measurement of volume of incontinent episodes	No data	No data	No data	No data	N/A
Resource implications (e.g. cost of incontinence management methods)	No data	No data	No data	No data	N/A
Quality of life measurements	No data	No data	No data	No data	N/A
Adverse effects	No data	No data	No data	No data	N/A
*No assumed risk or corresponding risk were estimated for reasons outlined in comments. CI: confidence interval; PFMT: pelvic floor muscle training; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

¹Downgraded two levels for serious imprecision: small sample size, wide confidence intervals that include both benefit and detriment.
 ²Downgraded one level for high risk of bias in one domain and unclear risk of bias in two.

Summary of findings 7. Pelvic floor muscle training versus anticholinergics for functional daytime urinary incontinence in children.

Pelvic floor muscle training versus anticholinergics for functional daytime urinary incontinence in children
Patient or population: children aged 5‐14 years with daytime urinary incontinence or post‐micturition dribbling Settings: two studies from Brazil and China (urology departments) Intervention: pelvic floor muscle training (with urotherapy in one study) Comparison: anticholinergics (with urotherapy in one study)
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Anticholinergics	PFMT
Resolution of daytime urinary incontinence at end of treatment or 1 month after 3 months of treatment	33 per 100	64 per 100 (39 to 100)	RR 1.92 (1.17 to 3.15)	86 (2)	⊕⊝⊝⊝ Very low1,2,3
Frequency of incontinence episodes	Not estimated ‐ see comments	Not estimated ‐ see comments	Not estimated ‐ see comments	86 (2)	⊕⊝⊝⊝ Very low1,2,3	Outcomes were not suitable for pooling. Campos 2013 reported higher increase in median dry days/month in the PFMT group (9 days) than in the anticholinergics group (5 days). Zeng 2012 reported significant improvement of postmicturition dribbling (from ≥ 5 to 1 episode in the month after treatment) in 2/21 in the PFMT plus biofeedback group (in addition to 16/21 with complete resolution) and in 6/18 in the tolterodine group (in addition to 6/18 with complete resolution).
Objective measurement of volume of incontinent episodes	No data	No data	No data	No data	N/A	N/A
Resource implications (e.g. cost of incontinence management methods)	No data	No data	No data	No data	N/A	N/A
Quality of life measurements	No data	No data	No data	No data	N/A	N/A
Adverse effects				47 (1)	⊕⊝⊝⊝ Very low1,2,3	Campos 2013 reported that there were no adverse events
*The assumed risk is the median control group risk across studies. CI: confidence interval; PFMT: pelvic floor muscle training; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

¹Downgraded one level for imprecision: small sample size.
 ²Downgraded one level for heterogeneity in study participants.
 ³Downgraded one level for high and unclear risk of bias in several domains in both studies.

Summary of findings 8. Transcutaneous electrical nerve stimulation (TENS) versus anticholinergics for functional daytime urinary incontinence in children.

Transcutaneous electrical nerve stimulation (TENS) versus anticholinergics for functional daytime urinary incontinence in children
Patient or population: children aged 4 to 17 years with daytime urinary incontinence Settings: two studies in Brazil and Denmark Intervention: parasacral transcutaneous electrical nerve stimulation (TENS) with placebo drug Comparison: anticholinergics with sham TENS
Outcomes	Study event rates*		Relative effect (95% CI)	No of participants (studies)	Certainty of the evidence (GRADE)	Comments
	TENS	Anticholinergics
Resolution of daytime urinary incontinence at 3 months	6/34 complete resolution of daytime urinary incontinence	6/38 complete resolution of daytime urinary incontinence	RR 0.81 (0.05 to 12.50)	72 (2)	⊕⊝⊝⊝ Very low1,2	No assumed risk or corresponding risk were estimated due to small sample size severe imprecision, inconsistency (masked by imprecision) and high risk of bias
Frequency of incontinence episodes	4/21 ≥ 50% resolution of daytime urinary incontinence	8/23 ≥ 50% resolution of daytime urinary incontinence	See comments	44 (1)	⊕⊝⊝⊝ Very low1,2	No assumed risk, corresponding risk or relative effect were estimated due to small sample size, severe imprecision, inconsistency (masked by imprecision) and high risk of bias.
Objective measurement of volume of incontinent episodes	No data	No data	No data	N/A	N/A
Resource implications (e.g. cost of incontinence management methods)	No data	No data	No data	N/A	N/A
Quality of life measurements	No data	No data	No data	N/A	N/A
Adverse effects	See comments	See comments	See comments	61 (2)	⊕⊝⊝⊝ Very low1,2	One study reported side effects or adverse event in the anticholinergics plus sham TENS group: dry mouth in 58%, hyperthermia in 25% and hyperemia in 50%. One study reported the development of significant postvoid residual in 7 children and UTI in one child in the anticholinergics plus sham TENS group and one serious adverse event (appendicitis) in the active TENS plus placebo group, not considered to be related to the study treatment.
*No assumed risk or corresponding risk were estimated for reasons outlined in comments. CI: confidence interval; RR: risk ratio; TENS: transcutaneous electrical nerve stimulation; UTI: urinary tract infection
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

¹Downgraded two levels for serious imprecision: small sample size, very confidence intervals that mask inconsistency.
 ²Downgraded one level for high or unclear risk of bias in several domains across studies.

Summary of findings 9. Voiding education with uroflowmetry and feedback versus anticholinergics for functional daytime urinary incontinence in children.

Voiding education with uroflowmetry and feedback versus anticholinergics for functional daytime urinary incontinence in children
Patient or population: children aged 6 to 12 years with overactive bladder with urge incontinence Settings: one multicentre study in outpatient clinics in Belgium, Germany, Netherlands, Sweden Intervention: voiding education with uroflowmetry and feedback and urotherapy Comparison: anticholinergics and urotherapy
Outcomes	Study event rates*		Relative effect (95% CI)	No of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Anticholinergics	Training and feedback
Resolution of daytime urinary incontinence at 12 months after 6 months of treatment	13/30 complete resolution of daytime urinary incontinence	15/34 complete resolution of daytime urinary incontinence	RR 1.02 (0.58 to 1.78)	64 (1)	⊕⊝⊝⊝ Very low1,2	No assumed risk or corresponding risk were estimated due to single study with risk of bias and small sample size.
Frequency of incontinence episodes	No data	No data	No data	No data	N/A
Objective measurement of volume of incontinent episodes	No data	No data	No data	No data	N/A
Resource implications (e.g. cost of incontinence management methods)	No data	No data	No data	No data	N/A
Quality of life measurements	No data	No data	No data	No data	N/A
Adverse effects	N/A	N/A	N/A	64 (1)	N/A	The study reported minor behavioural problems in two children in the anticholinergics arm and none in the voiding education with feedback arm.
*No assumed risk or corresponding risk were estimated for reasons outlined in comments. CI: confidence interval; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

¹Downgraded two levels for serious imprecision: small sample size, wide confidence intervals that include both benefit and detriment.
 ²Downgraded one level for high risk of bias in one domain and unclear risk of bias in three.

Background

This is one of two reviews of interventions for functional daytime urinary incontinence in children. This review focuses on non‐surgical and non‐pharmacological interventions, while the other focuses on alpha blockers, one class of medication for treating this condition (Eliezer 2019).

Plain language descriptions of treatments and methods considered by this review are provided in Table 10.

1. Glossary of terms.

Abdominal muscle training	Exercises intended to improve the tone of muscles that, if not well controlled, can affect continence
Anticholinergics	Drugs that can reduce signals from the brain that cause the bladder to contract and empty
Bladder diary	A daily record of time and amount of urination or incontinence
Bladder training	Bladder training is a treatment approach that employs strategies such as delayed voiding with the intention of gradually increasing bladder capacity.
Core stability exercises	Exercises intended to improve the tone of muscles that, if not well controlled, can affect continence
Cystometry/cystometric evaluation	Measurement of bladder pressure
Electromyography	Measurement of nerve/muscle activity
Feedback or biofeedback	Terms used to describe ways in which children are shown how their muscles are working or how their bladder is emptying, sometimes on computer screens or tablets, and these are sometimes used to try to improve the effectiveness of other treatments such as pelvic floor muscle training or voiding education.
Micturition cycle	Cycle of bladder emptying, followed by gradual refilling with urine produced by the kidneys, followed by voluntary emptying
Nocturnal enuresis	Involuntary leakage of urine when asleep, commonly referred to as bedwetting
Pelvic floor muscle training (PFMT)	Exercises that are normally taught by a nurse specialist or physiotherapist and can then be performed at home, in which children learn to identify and exercise the muscles supporting the bladder and its outlet
Scheduled voiding	Emptying the bladder according to a schedule or time interval defined by a clinician
Storage symptoms or storage problems	To achieve continence the bladder should be able to reliably store urine until it is voluntarily emptied. Storage symptoms or problems are terms used when this is not the case: for example, incontinence (when the bladder involuntarily leaks), urgency (sudden sensation of needing to urinate), frequency (frequent need to urinate).
Transcutaneous electrical nerve stimulation (TENS)	A treatment in which adhesive electrodes are used to stimulate muscles or nerves in order to improve their function. Normally children may use the stimulator for 20‐30 minutes several times a week either at home or at a clinic.
Urotherapy	A name used to describe behavioural programmes in which children (and sometimes carers) are taught about how the bladder works, proper postures and methods for going to the toilet, planning what and how much to drink and scheduling when to go to the toilet
Vesico‐ureteric reflux	Flow of urine back from the bladder towards the kidneys via the ureters
Voiding	Emptying the bladder by passing urine
Voiding education	A treatment in which children practice emptying their bladders efficiently and completely, usually in a specially adapted toilet in which sensors measure the amount of urine passed and the rate at which it is passed
Voiding problems or voiding symptoms	Voiding problems or symptoms are terms used to describe difficulties emptying the bladder when attempting to do so, such as urinary retention (incomplete emptying of the bladder), hesitancy (difficulty starting or maintaining urination), and poor or weak flow of urine

Description of the condition

Functional daytime urinary incontinence in children is the term used to describe any leakage of urine while awake that is not due to some known cause. In general, by the age of three most children have achieved an adequate degree of continence and can use the toilet for voiding. By five years of age, most can void at will and postpone voiding. Involuntary wetting while asleep or awake becomes a problem that may require some therapeutic intervention (Milsom 2013).

Although no cause is determined in a large proportion of children with functional daytime urinary incontinence, a number of independent risk factors have been identified including stress, neglect or trauma, a paternal history of daytime wetting and a history of daytime wetting in male siblings (Lai 2015; Milsom 2013). Some physical factors are also known to contribute to daytime urinary incontinence such as constipation, vesico‐ureteric reflux (urine flowing back from the bladder towards the kidneys) and a history of urinary tract infection. Functional daytime urinary incontinence has been found in a significant number of children who had functional constipation and encopresis (bowel leakage) and urinary symptoms improved when bowel symptoms were treated (Loening‐Baucke 1997).     

Determining the prevalence of functional daytime urinary incontinence in children from existing research is problematic for several reasons. Definitions of urinary incontinence have differed between studies (frequency and extent of leakage). In some settings, increased reported prevalence may be attributed to greater public understanding of incontinence as a health issue and of treatments available as a result of awareness campaigns and information disseminated through health care, education and the media. Variations in reported prevalence may also result from societal norms and cultural differences. Limited access to or experience of health care may affect what degree of daytime wetting is perceived to constitute a health problem.

At seven years of age, studies have reported prevalence of daily daytime wetting, ranging from 0.5% to 0.8% (Hansen 1997; Hellstrom 1990), and of wetting more than once per week from 0.9% to 2.5% (Hansen 1997; Hellstrom 1990; Joinson 2006; Swithinbank 2010). Estimates for the prevalence of daytime wetting more than once per month range from 3.9% to 9.0% and combined daytime and night‐time wetting range from 6.3% to 9.0% (Joinson 2006; Kajiwara 2004; Lee 2000; Soderstrom 2004). At 11 to 13 years, estimates for daytime and night‐time wetting more than once per month range from 1.1% to 4.2% (Kajiwara 2004; Lee 2000; Soderstrom 2004). At 15 to 17 years of age, estimates of daytime wetting more than once every three months range from 1.8% to 3.0% (Hellstrom 1995; Swithinbank 1998).

At seven years of age, estimates of daytime and night‐time wetting more than once per month range from 5.8% to 8.9% in girls and from 6.8% to 9.2% in boys and of more than once per week from 1.2% to 3.1% in girls and 0.7% to 3.8% in boys (Hellstrom 1995; Joinson 2006; Swithinbank 2010). However, at ages 11 to 13 girls are more likely to be affected and this difference becomes more marked as they grow older. At ages 11 to 13, estimates of daytime and night‐time wetting more than once per month range from 3.9% to 4.3% in girls and from 1.0% to 4.1% in boys (Kajiwara 2004; Soderstrom 2004). Regular daytime wetting has been reported by 16.6% of girls and 7.2% of boys aged 11 to 12 (Swithinbank 1998). By the ages of 15 to 17, estimates for daytime wetting range from 3.6% to 4.7% of girls and from 0.3% to 0.9% of boys (Hellstrom 1995; Swithinbank 1998).

In many children daytime urinary incontinence is treated successfully or resolves spontaneously. With understanding, support and management using appropriate products, the impact on a child’s life can be minimised. However, daytime urinary incontinence can create practical and social difficulties for both the child and their family and has the potential to impact on a child’s well‐being in more serious ways. Daytime urinary incontinence can impact on children’s access and engagement with their education in both practical and performance aspects (Whale 2016). Research has indicated that children experiencing incontinence will have increased absenteeism, poorer academic performance and potential social difficulties during their school years (Filce 2015). Psychological and behavioural problems have been reported to be more common in children experiencing daytime wetting. (Joinson 2006; Lettgen 2002; von Gontard 2012). Psychological and behavioural problems in childhood have been found to be predictive of a range of personal, psychological and social problems later in life (McCulloch 2000).

Description of the intervention

Conservative treatments for functional daytime urinary incontinence in children encompass a broad range of non‐surgical and non‐pharmacological interventions, which may be employed individually or in combination.

Lifestyle and behavioural interventions can be implemented by the child and parents with little input from healthcare professionals. Keeping bladder diaries can be useful in establishing voiding patterns. Toilet plans and scheduled voiding can be used as initial, simple interventions. Incentives and rewards such as star charts may also be used to motivate children as well as provide positive reinforcement for changes to voiding and drinking schedules. Educational materials can be provided relating to diet, fluid intake and good posture when sitting on the toilet. Alarm devices, although primarily used in children with nocturnal enuresis, can also be useful for children with daytime urinary incontinence by providing auditory signals on the occurrence of incontinence (Glazener 2005; NICE 2010).

More complex lifestyle and behavioural interventions may require greater involvement from healthcare professionals. Methods involving bladder training, double voiding and voiding postponement are commonly used treatment options. Managing constipation is an important consideration, as it can be a causative factor in children with urinary incontinence (Loening‐Baucke 1997).

Physical therapies such as pelvic floor muscle training (PFMT), abdominal muscle training and core stability exercises can also be used to treat daytime functional urinary incontinence in children. Physical therapies can be used with or without biofeedback, which assists the children in learning good exercising or bladder emptying technique by means of visual or auditory feedback. Transcutaneous electrical stimulation (TENS) of the sacral root or tibial nerve can also be considered for treatment of daytime urinary incontinence (Bower 2003).

How the intervention might work

Conservative treatment options work in varying ways, targeting different pathological processes implicated in urinary incontinence.

Behavioural therapies, often collectively referred to as bladder rehabilitation or urotherapy, use a range of non‐pharmacological approaches. Bladder diaries aim to assess voiding patterns of the individual child and identify subsequent schedules for voiding. Scheduled voiding and toilet plans aim to retrain the bladder and the brain into a new pattern of micturition to improve the child's ability to suppress urges to urinate or respond to signals to void. Star charts and other reward strategies are intended to use positive reinforcement to provide incentives to the child to reach achievable goals, such as improved voiding patterns or continence. Good toileting posture is intended to develop and heighten awareness of correct voiding processes, to facilitate relaxed pelvic floor activity during voiding and to encourage children to take time to void. Modification of diet may reduce intake of agents that may aggravate urinary storage or voiding problems. Although reduction of fluid intake may in itself aggravate low functional bladder capacity, modification of fluid intake timing may reduce the need to void at times when incontinence typically occurs. Alarms, more commonly used in nocturnal enuresis, are intended to provide auditory signals on occurrence of incontinence and increase the child's awareness of the wetting. Bladder training employs strategies such as delayed voiding with the intention of increasing functional bladder capacity, thereby reducing the need to void frequently. Management of constipation and reduction of impaction can reduce pressure on the bladder, which can subsequently improve urgency and wetting in some children. Behavioural therapies may also be combined with psychotherapy, family therapy or education, or a combination of these, which aim to identify and address potential psychological causative factors and to modify the environment or emotional conditions that have led to or have compounded incontinence (Hellstrom 1987; Glazener 2005; NICE 2010; Nijman 2017).

Physical therapy focuses on strengthening and relaxing the supportive muscles of the urinary system, primarily the pelvic floor muscles, and consequently improving voiding patterns and emptying which may improve continence. Abdominal muscle training and core stability exercises are likewise intended to improve the tone of muscles that, if not well controlled, can affect continence (Nijman 2017; Vasconcelos 2007; Zivkovic 2012).

Biofeedback can also be a useful teaching and performance enhancing aid to improve the patient’s awareness of the physiological processes of micturition, in particular in identifying and using the pelvic floor muscles correctly and learning to understand good bladder emptying techniques. This can be achieved by auditory or visual signals using electromyography to identify contraction and relaxation of the pelvic floor muscles, thereby improving the effectiveness of the exercises (Nijman 2017; Vasconcelos 2007).

Electrical stimulation involves either direct stimulation or transcutaneous stimulation of the sacral root or tibial nerve. Although the direct mechanism of action of electrical stimulation is unclear, it is thought to centre on alterations in the afferent and efferent nerve fibres to the lower urinary tract in children (Bower 2003).

Why it is important to do this review

Daytime urinary incontinence can result in practical, social and educational difficulties for children and their families and can impact significantly on various psychosocial aspects of children’s lives (Joinson 2006). Therefore, identifying effective treatments is particularly important both to alleviate symptoms and to improve the mental well‐being of children affected.

Existing Cochrane Reviews relating to urinary incontinence in children focus on the treatment of nocturnal enuresis (Caldwell 2013; Deshpande 2012; Glazener 2002; Glazener 2004; Glazener 2005; Huang 2011). Consequently, there is a need for further systematic reviews of treatments for daytime urinary incontinence in the same population. This review focuses only on conservative interventions and a separate review will cover the use of alpha blockers for treating daytime urinary incontinence in children (Eliezer 2019).

Objectives

To assess the effects of conservative interventions for treating functional daytime urinary incontinence in children.

Methods

Criteria for considering studies for this review

Types of studies

We included randomised controlled trials (RCTs) and quasi‐randomised (e.g. using allocation by alternation) studies of conservative interventions for treating functional daytime urinary incontinence in children. We also included cross‐over and cluster‐randomised studies.

Types of participants

We included studies of children aged between 5 and 18 years with daytime urinary incontinence (where no comorbidities or conditions were considered to be the cause). We selected the lower cut‐off age of five years because the majority of children will have achieved continence by this age. Beyond five years of age, daytime urinary incontinence merits consideration for treatment.

Types of interventions

We included studies of conservative interventions for treating daytime urinary incontinence. We considered non‐pharmacological and non‐surgical interventions to be conservative. These included the following.

• Lifestyle, behavioural, psychosocial interventions: bladder diaries; toilet plans; timed voiding; rewards and incentives; constipation management; posture education; bladder training; double voiding and voiding postponement; dietary education and changes to drinking schedules and type of fluid consumed. With the child central to all these types of interventions, parents or carers are often also involved, with varying degrees of training and supervision from healthcare professionals. Education and training interventions may be aimed at both children and at parents or carers.

• Physical interventions: PFMT; abdominal muscle training and core stability exercises, with or without biofeedback. These may be performed in clinic settings under professional supervision or at home, or both.

• Non‐invasive feedback interventions: alarm devices or other techniques to improve awareness of voiding

• Non‐invasive stimulation interventions: stimulation of nerves or muscles transcutaneously

We did not include ingestion of foods, drinks, herbal agents or supplements specifically aimed at treating urinary incontinence. Similarly, we did not include devices or techniques that involved any invasive element, such as implants or needles.

We included the following comparisons:

• one individual conservative intervention (lifestyle, behavioural or physical) versus no treatment;

• combined conservative interventions versus no treatment;

• one individual conservative intervention versus another individual or combined conservative intervention;

• combined conservative interventions versus other combined conservative interventions;

• individual conservative interventions versus non‐conservative interventions (pharmacological or invasive, combined or not with any conservative interventions); and

• combined conservative interventions versus non‐conservative interventions (pharmacological or invasive, combined or not with any conservative interventions).

Types of outcome measures

Primary outcomes

• Number of children no longer experiencing daytime urinary incontinence: child‐ or carer‐reported

Secondary outcomes

• Frequency of incontinence episodes, of pad or clothes changes (whichever measure is reported by studies)

• Objective measurement of volume of incontinent episodes (e.g. pad tests)

• Psychometric or behavioural measurements (whichever measure is reported by studies)

• Quality of life measurements. The measurements considered will be dependent upon those reported in studies, but may include general paediatric quality of life tools such as PedQol or incontinence‐specific tools developed for use by children (Bower 2006; Varni 2001).

• Cost of interventions, cost‐effectiveness of interventions, resource implications (e.g. cost of incontinence management methods), missed schooldays

• Adverse events or child‐ or carer‐perceived harms associated with interventions reported by studies.

Main outcomes for 'Summary of findings' tables

The GRADE working group recommends including no more than seven critical outcomes in a systematic review (Guyatt 2011a; Guyatt 2011b). In this review, the six outcomes for which the certainty of evidence will be assessed are:

• Number of children no longer experiencing daytime urinary incontinence – child or carer reported

• Frequency of incontinence episodes, of pad or clothes changes (whichever measure is reported by studies)

• Objective measurement of volume of incontinent episodes (e.g. pad tests)

• Resource implications (cost of incontinence management methods)

• Quality of life measurements

• Adverse events or child‐ or carer‐perceived harms associated with interventions reported by studies

Search methods for identification of studies

We did not impose any restrictions, for example language or publication status, on the searches described below.

Electronic searches

This review drew on the search strategy developed for Cochrane Incontinence. We identified relevant studies from the Cochrane Incontinence Specialised Register to 11 September 2018. For more details of the search methods used to build the Specialised Register, please see the Group's webpages where details of the Register's development (from inception) and the most recent searches performed to populate the Register can be found. The register contains studies identified from the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, MEDLINE In‐Process, MEDLINE Epub Ahead of Print, CINAHL, ClinicalTrials.gov, the World Health Organization’s International Clinical Trials Registry Platform (WHO ICTRP), UK Clinical Research Network Portfolio (now known as Be Part of Research) and handsearching of journals and conference proceedings. Many of the studies in the Cochrane Incontinence Specialised Register are also contained in CENTRAL. The terms that were used to search the Cochrane Incontinence Specialised Register are given in Appendix 1. The search strategy was adapted for searching of the following Chinese language bibliographic databases: Chinese Biomedical Literature Database (CBM), China National Knowledge Infrastructure (CNKI), and Wanfang (Xia 2008).

The Specialised Register also contains studies identified in international clinical trials registries via the Cochrane Register of Studies search portal. However, the search terms used for the clinical trials registry searches are not as comprehensive as used in searches of the bibliographic databases. As it has been observed that academic papers on childhood incontinence may use different terms than those used in adult incontinence (such as 'day wetting', 'daytime wetting' and 'toileting problems'), we conducted supplementary electronic searches of international clinical trials registries via WHO ICTRP to 15 August 2018, using a review‐specific set of search terms. The WHO platform is a regularly updated database of studies registered in 16 international registries (China, European Union, Netherlands, ClinicalTrials.gov (USA), Japan, Australian & New Zealand, ISRCTN, Brazil, India, Korea, Cuba, Germany, Pan African, Sri Lanka, Thailand). Registries include much of the information needed to judge whether a study be may eligible for inclusion in a review: population, study interventions and comparisons, methods, outcome measures and follow‐up, and sample size.

The search terms encompassed conditions but no intervention terms to ensure all potential interventions were included. We did not impose any language or other limitations.

Searching other resources

In addition to electronic searches, we searched the reference lists of the identified study reports and reviews to identify other papers for inclusion. Furthermore, we made attempts to contact study authors for additional results or clarification of methods within the study where required.

Data collection and analysis

As per our published protocol (Buckley 2016), we conducted the following data collection and analysis in accordance with the guidance in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011a).

Selection of studies

Two review authors independently screened studies identified from the electronic searches for potential eligibility for inclusion. We screened titles and abstracts of studies identified in electronic databases of published studies and titles and metadata of studies identified in trials registers. We sought full‐text reports for all potentially eligible studies. For studies identified in trials registries that were not duplicated in searches of bibliographic databases (unpublished studies), we contacted the authors or institutions recorded in the registry to request study reports.

Two review authors independently determined the eligibility of studies for inclusion in the review from the full reports according to predefined criteria. The two review authors resolved any differences by discussion and, if necessary, referred to a third review author for arbitration. We employed a similar process on screening studies identified from the reference lists of included studies. Native speakers assessed papers in languages other than English for eligibility and subsequently for data extraction; we conducted a similar selection process for Chinese‐language studies.

Data extraction and management

Two review authors independently carried out data extraction using forms that we developed and piloted for the review. The two review authors resolved any differences by discussion and, if necessary, referred to a third review author for arbitration. Where data from the study were not provided, we contacted the study author(s) to request further information. Included study data were processed as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011b).

Assessment of risk of bias in included studies

Two review authors independently assessed the risk of bias in eligible studies using Cochrane's 'Risk of bias' tool (Higgins 2017). Factors considered included random allocation and concealment (where appropriate), description of missing individual data, blinding during intervention and at outcome assessment (where appropriate), description of and protection against possible contamination (where appropriate) and other biases. We judged the risk of bias for each domain as 'low risk', 'unclear risk' or 'high risk'. The two review authors resolved any differences by discussion and, if necessary, referred to a third review author for arbitration. Other sources of bias considered relevant for the review included substantial involvement in study conduct, interpretation or reporting by funding sources, or other confounders reported by the study authors.

Measures of treatment effect

Where data allowed, we calculated risk ratios (RR) for dichotomous data and mean differences (MD) with 95% confidence intervals (CI) for continuous data. If the same continuous outcome was measured using a different scale, we used the standardised mean difference (SMD).

Unit of analysis issues

Included studies with more than two treatment arms were handled according to the available comparisons. Where there were more than two arms, each with distinct interventions of interest to the review, we selected pairs of interventions and included them in the relevant comparisons in the review. Where there was more than one arm, with two pre‐crossover arms receiving an identical intervention of interest and the third arm a different intervention of interest, we merged the arms receiving the identical intervention into a single arm for comparison. We used outcomes from the first treatment period in our primary analysis due to concerns about a carry over of treatment effect from one period to the next.

Where data allowed, we planned to include cluster‐randomised studies in the analyses along with individually‐randomised studies. We would have adjusted the sample sizes of cluster‐randomised studies using the methods described in the Cochrane Handbook of Systematic Reviews of Interventions using an estimate of the intra‐cluster correlation co‐efficient (ICC) derived from the study (if possible), from a similar study or from a study of a similar population (Higgins 2011c). If we used ICCs from other sources, we would have reported it, and conducted sensitivity analyses to investigate the effect of variation in the ICC. If we identified both cluster‐randomised studies and individually‐randomised studies, we would pool the data from both using the sample size adjustment methods described above if there was little heterogeneity between the study designs and we considered the interaction between the effect of intervention and the choice of randomisation unit to be unlikely. We would have considered heterogeneity in the randomisation unit and performed sensitivity analyses to investigate the effects of the randomisation unit.

Dealing with missing data

We noted levels of attrition for included studies and planned to use sensitivity analyses to assess the impact of including studies with high levels (above 20%) of missing data (Sackett 1997). However, data were only missing for a small proportion of participants in two studies. We conducted analyses for all outcomes, insofar as was possible, on an intention‐to‐treat basis by using a conservative approach and imputing missing participants as not having experienced the outcome in all studied intervention arms. Then, the denominator for each outcome in each study was the number of eligible participants randomised. Whenever possible, we contacted investigators to request missing data.

Assessment of heterogeneity

Where data allowed, we intended to assess statistical heterogeneity using the I² statistic (Higgins 2003) and Chi² test, with heterogeneity regarded as substantial if the I² statistic value was greater than 50% or the Chi² P value less than 0.10 (Deeks 2017).

Assessment of reporting biases

In order to consider whether there may be publication bias, we intended to generate funnel plots for any meta‐analyses containing 10 or more studies, assessing asymmetry visually and formal tests for funnel plot asymmetry conducted as appropriate. For dichotomous outcomes, we intended to use the test proposed by Harbord 2006, and for continuous outcomes that proposed by Egger 1997. We also used electronic searching of trials registries to identify unpublished studies. This approach can identify only unpublished studies conducted relatively recently, since registration became a widespread research requirement, but there remains the possibility that older studies remain unidentified.

Had data been available, we planned to assess the risk of bias of evidence on “costs of interventions, cost‐effectiveness, resource implications, missed school days” using the methods set out in Chapter 15 of the Cochrane Handbook for Systematic Reviews of Interventions (Shemilt 2011).

Data synthesis

Where possible, we pooled data from studies and produced RRs or MDs for the relative effectiveness of interventions compared. We applied a random‐effects model because the types of interventions considered by the review are known to involve differing treatment protocols, which can introduce some clinical and methodological heterogeneity, which may manifest as statistical heterogeneity. With a random‐effects model, the observed variability in results of the studies (beyond sampling error) is fully incorporated in the meta‐analysis results. We used the Mantel‐Haenszel method for dichotomous outcomes and planned to use the inverse‐variance method to pool results in MD or SMD. Where pooling of data was not possible, or where single studies reported on comparisons, we reported results of individual studies narratively.

Had data been available, we planned to synthesise evidence on “costs of interventions, cost‐effectiveness, resource implications, missed school days” using the methods set out in Chapter 15 of the Cochrane Handbook (Shemilt 2011).

Subgroup analysis and investigation of heterogeneity

Where data were available, we planned subgroup analyses to be undertaken by age (children between 5 to 12 years and adolescents between 13 to 18 years), sex and pathological cause of daytime urinary incontinence where possible.

Sensitivity analysis

Where data were available, we planned to carry out sensitivity analyses for factors that may affect the results of meta‐analyses. We included in our main meta‐analyses, studies that we considered to have potentially important differences in populations or interventions or that we judged to be at high or unclear risk of bias for the investigated bias domains, and we subsequently excluded these in sensitivity analyses to assess their effect on the overall results.

Where studies included children with known causes of their incontinence or children with only voiding symptoms, we sought data for the eligible subgroup of children with daytime urinary incontinence with no known cause. Where data for the eligible subgroup were not published, we made efforts to retrieve data from authors. Where this was not possible, we included studies with small subgroups of children with known causes of incontinence or voiding symptoms alone in analyses, and, as appropriate, conducted sensitivity analyses to determine whether their inclusion affected results.

GRADE and 'Summary of findings' tables

We assessed the certainty of evidence in the review using the GRADE approach, considering the following factors:

• limitations in the study design (risk of bias);

• inconsistency of results (statistical heterogeneity);

• indirectness of evidence (applicability);

• imprecision (number of events and confidence intervals);

• publication bias.

We generated 'Summary of findings' tables for comparisons that we considered most likely to be relevant or helpful to clinicians, parents or children choosing between conservative treatments. In line with recommendations on summarising evidence for rare diseases or where there was a paucity of studies, we did not calculate assumed and corresponding risks for 'Summary of findings' tables with single studies or where pooled estimates were affected by severe imprecision or inconsistency or considerable risk of bias (Pai 2015).

Results

Description of studies

Results of the search

The combined Cochrane Incontinence Specialised Register, WHO ICTRP and Chinese databases searches identified 1516 records of potentially eligible studies (Figure 1). After removal of duplicates, we screened the titles and abstracts of 1221 records and excluded 1104. We assessed the remaining 117 full‐text published reports or registration records relating to 86 individual studies. Of these, we included 27 studies (documented in 50 published reports or registration records) in the review (Ayan 2007; Borch 2017; Brownrigg 2015; Campos 2013; de Paula 2017; Fielding 1980; Golli 2013; Hagstroem 2009; Hagstroem 2010; Halliday 1987; Hellstrom 1989; Kajbafzadeh 2011; Klijn 2006; Kroll 2006; Ladi Seyedian 2014; Lordelo 2010; Nemett 2008; Newgreen 2017; Patidar 2015; Quintiliano 2015; Reis 2014; Rhodes 2008; Sillen 2014; Van Gool 2014; Vasconcelos 2007; Zeng 2012; Zivkovic 2010); and we excluded 51 studies (documented in 55 reports). We have reported reasons for exclusion in the Characteristics of excluded studies.

1.

Study flow diagram

We identified the registration records of three ongoing studies that might become relevant to include in future updates of this review (ACTRN12611000828921; NCT02336906; NCT03478813; see the Characteristics of ongoing studies). We identified five studies (nine published reports or registration records) that we could not classify due to insufficient information on methods, populations or outcomes (Braga 2017; Kueter 2000; Mourani 2012; Stauber 2007; Weber 2011; see the Characteristics of studies awaiting classification).

Included studies

Design

Of the 27 included studies, 21 were individually‐randomised controlled trials and five were multi‐arm studies (Ayan 2007; Borch 2017; Klijn 2006; Newgreen 2017; Van Gool 2014). One study was both a multi‐arm study and a cross‐over study (Hellstrom 1989).

Sample sizes

In total, the included studies involved 1803 children; most studies were small, with numbers randomised ranging from 16 to 202 (median sample size 60). Further information is reported in the Characteristics of included studies.

Setting

Most studies were conducted in Europe, with three in Denmark (Borch 2017; Hagstroem 2009; Hagstroem 2010), three in the UK (Fielding 1980; Halliday 1987; Rhodes 2008), two in Sweden (Hellstrom 1989; Sillen 2014), and one each in the Netherlands (Klijn 2006), Poland (Kroll 2006), Serbia (Zivkovic 2010) and Slovenia (Golli 2013). A further multicentre European study was conducted in Germany, Belgium, Netherlands and Sweden (Van Gool 2014). In South America, six studies were conducted in Brazil (Campos 2013; de Paula 2017; Lordelo 2010; Quintiliano 2015; Reis 2014; Vasconcelos 2007). In North America, one study was conducted in Canada (Brownrigg 2015); and one in USA (Nemett 2008). In Asia, two studies were conducted in Iran (Kajbafzadeh 2011; Ladi Seyedian 2014), and one study each in China (Zeng 2012), India (Patidar 2015) and Turkey (Ayan 2007). One study was reported as having been conducted in 16 unspecified countries worldwide (Newgreen 2017).

Studies were conducted in outpatient settings linked to urology departments (Brownrigg 2015; Campos 2013; Kajbafzadeh 2011; Quintiliano 2015; Zeng 2012), paediatric urology or nephrology departments (Golli 2013; Klijn 2006; Ladi Seyedian 2014; Lordelo 2010; Vasconcelos 2007), paediatric departments (Halliday 1987; Hellstrom 1989; Rhodes 2008), specialist continence clinics (Fielding 1980; Hagstroem 2009; Hagstroem 2010), or physical therapy departments (Nemett 2008; Zivkovic 2010). Some studies were hospital‐based but did not report more precise settings (Borch 2017; Patidar 2015; Reis 2014; Sillen 2014; Van Gool 2014). No information on setting was reported in the remaining studies (Ayan 2007; de Paula 2017; Kroll 2006; Newgreen 2017).

Participants

The included studies involved children with daytime urinary incontinence. In some studies, the participants also had overactive bladder or urinary urgency symptoms (Ayan 2007; Borch 2017; de Paula 2017; Hagstroem 2009; Hagstroem 2010; Hellstrom 1989; Lordelo 2010; Newgreen 2017; Patidar 2015; Quintiliano 2015; Van Gool 2014); or broader voiding dysfunction, sometimes alongside other symptoms such as post‐void residual urine, detrusor sphincter discoordination, constipation and faecal incontinence (Ayan 2007; Brownrigg 2015; de Paula 2017; Kajbafzadeh 2011; Klijn 2006; Kroll 2006; Ladi Seyedian 2014; Nemett 2008; Vasconcelos 2007; Zivkovic 2010). One study included children with a range of defined lower urinary tract symptoms, including overactive bladder, stress urinary incontinence, urgency urinary incontinence, voiding postponement and underactive bladder (Golli 2013).

The age of children in the included studies ranged from 3 years to 18 years. Two studies did not include an age range but reported review‐appropriate mean ages (Patidar 2015; Vasconcelos 2007).

Interventions

Physical interventions

Six included studies evaluated physical interventions (Campos 2013; Kajbafzadeh 2011; Ladi Seyedian 2014; Nemett 2008; Zivkovic 2010; Zeng 2012).

Four of these six studies compared physical therapies combined with behavioural interventions with behavioural interventions alone (Kajbafzadeh 2011; Ladi Seyedian 2014; Nemett 2008; Zivkovic 2010).

Kajbafzadeh 2011 compared PFMT, voiding education with animated biofeedback and urotherapy with urotherapy alone in 80 children aged 5 to 16 years with voiding dysfunction with either constipation or faecal incontinence, some of whom also had daytime urinary incontinence (21/80). Data were available for the subgroup with daytime urinary incontinence and only these are considered in this review. All therapies were delivered over six months. Children in the PFMT/biofeedback arm had two sessions per week; sessions stopped once electromyography was normal. The mean number of sessions was 9.6 (range 6 to 12). PFMT was physiotherapist‐supervised. Children practiced voiding with feedback presented graphically to children via animated computer game displays, which reflected their muscle activity. PFMT was also practiced at home for at least 15 minutes twice daily. It was recommended that children in the urotherapy only arm attend monthly visits for six months to reinforce training and enhance compliance.

Ladi Seyedian 2014 compared PFMT and urotherapy with urotherapy alone in 60 children aged 5 to 14 years with urinary tract symptoms or dysfunctional voiding, some of whom had daytime urinary incontinence (28/60). Data were available for the subgroup with daytime urinary incontinence and only these are considered in this review. PFMT was physiotherapist‐supervised in 12 sessions (2 x 30‐minute sessions per week for six weeks) with and without Swiss ball. Urotherapy involved education for parents and children in both groups about the urinary and gastrointestinal tract function, normal drinking habit, scheduled voiding, the normal mechanism of toilet training, and the effects of a high‐fibre diet at bimonthly visits for 12 months.

Nemett 2008 compared osteopathy‐based physical therapy plus standard urological care with standard urological care alone (including urotherapy) in 32 children aged 4 to 11 years with dysfunctional voiding, more than half of whom had daytime urinary incontinence. Data were available for the subgroup with daytime urinary incontinence and only these are considered in this review. Osteopathy‐based physical therapy during four, one‐hour sessions over 1 to 12 weeks included mobilisation of bodily tissues to relieve movement restrictions, achieve balanced alignment and mobility and postural symmetry. Visits were scheduled to coincide with fortnightly urology visits. Standard urological care comprised four, one‐hour urology appointments at two‐week intervals, providing (as indicated): medication; behavioural modification (advice on timed voiding and evacuation schedules; dietary modification); PFMT; PFMT with feedback; and treatment of constipation.

Zivkovic 2010 compared PFMT combined with diaphragmatic exercises and standard urotherapy with standard urotherapy alone in 86 children aged 3 to 13 years (mean 7 years) with dysfunctional voiding, half of whom had daytime urinary incontinence (42/86). Data were available for the subgroup with daytime urinary incontinence and only these are considered in this review. PFMT was physiotherapist‐supervised in monthly therapeutic sessions with total number determined by clinical results (mean 10 sessions over one year), with request to perform at home daily (three‐second contractions, 30‐second relaxation, 30 repetitions daily). In addition, children were taught diaphragmatic breathing exercises with the aim of achieving abdominal muscle relaxation. Standard urotherapy included education about urinary tract function, fluid intake, regular voiding, genital hygiene, voiding posture and treatment of constipation at monthly sessions over one year. Pharmacotherapy was not stopped during the study.

Two studies compared PFMT with anticholinergics (Campos 2013; Zeng 2012).

Campos 2013 compared PFMT with oxybutynin, both combined with behavioural urotherapy, in 47 children aged 5 to 10 years with a median of three episodes of daytime urinary incontinence per week. Oxybutynin was administered at a dose of 0.2 mg/kg, split into twice a day, over three months. PFMT was supervised once a week by a physiotherapist and practised twice a week at home for three months. Both groups received similar urotherapy: scheduled voiding; advice on fluid intake and appropriate timing and type of drinks; use of 500 mL bottle to facilitate monitoring for fluid intake; voiding posture instruction.

Zeng 2012 compared PFMT plus biofeedback with anticholinergics in 39 children aged 8 to 14 years, with at least five episodes of post‐micturition dribbling in the previous month due to detrusor after‐contraction. PFMT with electromyography biofeedback was conducted in two to three, 20– to 30‐minute training sessions a week for 12 weeks. Patients were taught to perceive relaxation and contraction of the anal sphincter, and instructed to contract while simultaneously maintaining stable abdominal pressure. Anticholinergic treatment was tolterodine 1 mg, orally, twice a day for 12 weeks.

Stimulation interventions

We included eight studies that evaluated stimulation interventions (Borch 2017; de Paula 2017; Hagstroem 2009; Lordelo 2010; Patidar 2015; Quintiliano 2015; Reis 2014; Sillen 2014).

Four of the eight studies compared TENS with sham (placebo) TENS (de Paula 2017; Hagstroem 2009; Lordelo 2010; Patidar 2015).

de Paula 2017 compared parasacral TENS with sham TENS in 16 children and adolescents aged 3 to 18 years with overactive bladder and daytime incontinence. For parasacral TENS, electrodes were placed bilaterally to the lumbar spine at S2 to S3, with stimulation at a frequency of 10 Hz, pulse width of 700 ms, and variable intensity determined by the tolerance threshold of the child. In the sham TENS group electrodes were placed in the scapular region. Both groups had weekly 20‐minute, clinic‐based sessions for 20 weeks. Both groups received similar urotherapy: advice on voiding habit; voiding before bed and upon waking; voiding schedule every three to four hours; no delay in micturition; toilet position and posture; increase drink volumes; no caffeine; avoid drinking two hours before bed; increased fibre.

Hagstroem 2009 compared sacral TENS with sham TENS in 25 children aged 5 to 14 years with daytime urinary incontinence at least twice weekly. Participants were provided with TENS units and instructed to use them for two hours daily for four weeks. Half of the stimulators were modified by the manufacturer not to deliver any current for sham treatment. Sacral TENS was neurostimulation with digital TENS stimulator with self‐adhesive surface electrodes at the level of the S2 to S3 roots, with stimulation at a frequency of 10 Hz with a 200 sec pulse duration and biphasic waveform. Childrens' incontinence symptoms were recorded for a week before study entry and during the last week of treatment.

Lordelo 2010 compared parasacral TENS with sham TENS in 37 children aged over four years (mean seven years) with overactive bladder, the majority of whom had daytime incontinence (30/37). Parasacral TENS involved two 3.5 cm electrodes placed on each side of S3 and S2 with stimulation at a frequency of 10 Hz with a generated pulse of 700 ms, and intensity increased to maximum level tolerated. Electrodes were also placed in the scapular area 3 cm above the inferior scapular edge. In the sham group the stimulation was at the scapular electrodes. Both groups had 20, 20‐minute, clinic‐based sessions, three times a week (over seven weeks). Both groups also underwent urotherapy reinforced by a booklet: voiding before sleeping; increasing volume of liquid ingested daily; eating foods rich in fibre; refraining from postponement of voiding when experiencing symptoms of urgency; postural advice on micturition; toilet seat adapters and foot supports as needed.

Patidar 2015 compared tibial TENS with sham tibial TENS in 40 children with a mean age of eight years with overactive bladder and daytime urinary incontinence (of 37 who completed, 12 had mild incontinence, 15 moderate, 10 severe). TENS involved two self‐adhesive surface electrodes placed in the medial region of the foot, approximately 3 cm to 4 cm cephalad to the medial malleolus between the posterior margin of the tibia and the soleus muscle. Stimulation was an adjustable voltage pulse intensity of 0 to 10 mA, a fixed pulse width of 200 ms, and a frequency of 20 Hz. TENS was administered in weekly, 30‐minute clinic sessions for 12 weeks. Sham TENS followed the same schedule, with patch electrodes applied above the medial malleolus to simulate the test group but TENS settings were not applied and no stimulation was given.

One study compared TENS plus behavioural urotherapy with behavioural urotherapy alone (Sillen 2014).

Sillen 2014 compared sacral TENS plus behavioural urotherapy (bladder education, timed voiding, voiding posture, fluid intake, prevention of constipation) with behavioural urotherapy alone in 62 children aged 5 to 12 years with daytime urinary incontinence at least once weekly. Sacral TENS involved self‐adhesive surface electrodes placed bilaterally at sacral region (S2 to S3), with stimulation at frequency of 10 Hz and current intensity increased to maximum tolerated (to a maximum of 40 mA). Treatment was administered for 20 minutes, twice daily over 12 weeks. The child and the parents were instructed in handling a stimulator at home. Urotherapy included explanation of normal body and lower urinary tract function, voiding position, timed voiding (before urgency attacks), drinking habits, quality of beverages and prevention of constipation, and behavioural therapy and detailed lifestyle advice in order to normalise bladder function.

Two studies compared sacral TENS with anticholinergics (Borch 2017; Quintiliano 2015).

Borch 2017 compared anticholinergics plus sham TENS with active TENS plus placebo in 66 children aged 5 to 14 years with urge urinary incontinence at least twice weekly. The study also included a third arm in which children received both active TENS and anticholinergics. The anticholinergics plus sham TENS received oxybutynin (5 mg twice daily) with sham TENS for 10 weeks. The active TENS plus placebo group received home‐based TENS with placebo (twice daily) for 10 weeks. Sacral TENS involved electrodes placed at S2 to S3, with stimulation at a frequency of 10 Hz, 40 mA, with a biphasic wave, pulse width 200. Parents and children were trained in using the TENS device and instructed to apply stimulation for two hours daily. Sham TENS followed the same schedule.

Quintiliano 2015 compared anticholinergics plus sham TENS with active TENS plus placebo in 28 children aged 4 to 17 years with urinary urgency at least three times per week, the majority of whom had daytime incontinence (25/28). The anticholinergics plus sham TENS received oxybutynin (0.3 mg/kg twice daily) with sham TENS (three times/week). The active TENS plus placebo group received TENS (three times/week) with placebo (twice daily). TENS involved two surface electrodes placed symmetrically on parasacral region (and two unused placed on one scapula) and stimulation at a frequency of 10 Hz with a biphasic square current pulse, pulse width 700 milliseconds, and intensity increased to threshold. Sham TENS followed the same schedule, but only scapular electrodes were active. Both had 20 sessions, 20 minutes per session, three times a week for seven weeks. Both groups received standard urotherapy, with instruction on scheduled voiding every three hours, not retaining urine when there was urinary urgency, urinating before bedtime, avoiding coffee, tea, soda, chocolate and citrus fruits during treatment, ingesting a greater amount of liquid during the day, eating high‐fibre foods, and using a toilet seat reducer or footrest as necessary.

One study compared PFMT and feedback plus urotherapy with TENS plus urotherapy (Reis 2014).

Reis 2014 compared PFMT with electromyographic biofeedback with sacral TENS, both with behavioural urotherapy, in 78 children aged 5 to 16 years with daytime urinary incontinence and nocturnal enuresis. PFMT with the child able to see graphic electromyographic biofeedback trace on a screen was delivered in a physiotherapy clinic in two, 20‐minute sessions per week, with a maximum of 20 sessions per child (mean number of sessions 10.9). Sacral TENS was delivered in a physiotherapy clinic in two, 20‐minute sessions per week, with a maximum of 20 sessions per child (mean number of sessions 10.9); silicone electrodes placed at S2 to S4, current with frequency of 10 Hz, pulse width 700 μs and intensity at the sensitive threshold. Urotherapy in both groups included advice on timing of voiding, avoiding delay or restraint manoeuvres, voiding posture, fluid intake and dietary advice.

Feedback interventions

Five studies were included that evaluated feedback or alarm interventions (Halliday 1987; Kajbafzadeh 2011; Klijn 2006; Van Gool 2014; Vasconcelos 2007).

One study compared a wetting alarm with a sham alarm (Halliday 1987).

Halliday 1987 compared a wetting alarm with a sham wetting alarm in 44 children aged 5 to 15 years with daytime urinary incontinence. Daytime urinary incontinence was sufficiently problematic to be the principal complaint of parents at consultation (for most underwear sufficiently wet to soak outer clothing) and in 34 had been present continuously since birth. Children were randomised to up to three months' use of a buzzer alarm that activated when a pad worn inside underwear became wet, compared with a sham alarm, which sounded randomly every few hours, which may remind the child to use the toilet and establish regular voiding.

Three studies compared voiding education with uroflowmetry feedback alongside standard urotherapy, with standard urotherapy alone (Kajbafzadeh 2011; Klijn 2006; Van Gool 2014).

Kajbafzadeh 2011 compared PFMT, voiding education with animated biofeedback and urotherapy with urotherapy alone. Further details are presented in the physical interventions section above.

Klijn 2006 compared 24 weeks of urotherapy alone with 24 weeks of urotherapy supplemented during weeks 1 to 8 with home uroflow biofeedback in 143 children and adolescents aged 6 to 16 years with bladder instability, the majority of whom had daytime incontinence (95/143). Data were available for the subgroup with daytime urinary incontinence and only these are considered in this review. A third arm involved standard urotherapy plus a home video. Urotherapy included bladder education, scheduled voiding, voiding charts, voiding posture, instructions on proper defecation patterns with or without laxatives as needed, and prophylactic antibiotics. For biofeedback, uroflow units with palmtop displays were developed that could be used in the home and children instructed on their use and for the shape of curve to aim.

Van Gool 2014 compared three interventions (urotherapy plus anticholinergics, urotherapy plus placebo, and urotherapy plus voiding education with uroflowmetry and feedback) in 202 children aged 6 to 12 years, 97 of whom had overactive bladder with urge incontinence, uninhibitable urge to void and numerous small voidings per day. All received verbal and written standard urotherapy including instructions on fluid intake, voiding habits, hygiene, bladder diaries, and education on urgency, voided volumes and voiding frequency during three urotherapy visits over six months. Additionally, children were randomised to either a six‐month treatment period of 6 to 12 sessions of voiding education with uroflowmetry and feedback, or to oxybutynin (sugar‐free syrup, in a dosage of 0.3 mg/kg bodyweight per day, in three doses a day) or to placebo.

One study compared PFMT with feedback with PFMT alone (Vasconcelos 2007).

Vasconcelos 2007 compared PFMT with or without electromyography biofeedback in 59 children aged 5 to 15 years with dysfunctional elimination syndrome (abnormal pattern of elimination for child’s age characterised by urinary and bowel incontinence and withholding), the majority of whom had daytime urinary incontinence (41/59). Data were available for the subgroup with daytime urinary incontinence and only these are considered in this review. Children in the feedback group received PFMT in 16, 60‐minute sessions over a two‐month period, with biofeedback using a commercially available system that generated electromyography tracing during the contraction and relaxation phases. Children in the no‐feedback group received PFMT in 24, 60‐minute sessions over a three‐month period. Sessions consisted of a series of seven pelvic floor musculature contraction and relaxation exercises. Both groups also received behavioural urotherapy, which included scheduled voiding and drinking, training in toilet posture, and reinforcement using voiding diaries. Daytime urinary incontinence during four‐week periods was assessed at baseline and at 6 and 12 months.

Behavioural interventions

Eleven studies evaluated behavioural interventions (Ayan 2007; Brownrigg 2015; Fielding 1980; Golli 2013; Hagstroem 2010; Hellstrom 1989; Klijn 2006; Kroll 2006; Newgreen 2017; Rhodes 2008; Van Gool 2014).

One study compared a voiding re‐education programme with no treatment (a delayed treatment group) in 64 children aged 4 to 15 years with urinary tract dysfunction, a small majority of whom (37/64; 57.8%) had daytime urinary incontinence (Golli 2013). The programme featured a five‐day inpatient programme led by a nephrologist and delivered by trained nurses and a daily psychologist followed by monthly outpatient sessions. The inpatient programme included instruction on fluid intake, regular voiding habits, preventing and treating constipation, and PFMT and one‐to‐one sessions with a psychologist positively reinforcing the programme and supporting self‐image. The monthly outpatient programme supported the transfer of inpatient techniques to home setting, including voiding and bowel postponement, pelvic floor strengthening and relaxing. A bladder diary was used to record fluid intake, voiding, bowel habits, urinary incontinence, urgency, exercises and was reviewed at each session.

Five studies compared one behavioural modification intervention with another (Brownrigg 2015; Fielding 1980; Hagstroem 2010; Klijn 2006; Rhodes 2008).

Brownrigg 2015 compared individual one‐to‐one urotherapy with group urotherapy in children aged 6 to 10 years with urinary tract dysfunction, 28 of whom out of 60 had daytime incontinence. Both individual and group urotherapy included education on the urinary system, voiding education, progressive relaxation, work on psychosocial aspects of bladder and bowel dysfunction. Checklists and patient handouts were used to standardise urotherapy content.

Fielding 1980 compared bladder training combined with a night‐time enuresis alarm with an enuresis alarm alone in children aged 5 to 15 years with nocturnal enuresis with or without daytime incontinence. Daily bladder training for four weeks comprised of drinking 500 mL of fluid, timing of first urge to void, and encouragement to postpone voiding by a few minutes per session. In both groups a nocturnal enuresis alarm was used for 14 weeks. We consider only a subgroup of children with daytime incontinence who completed the study in this review.

Hagstroem 2010 compared 12 weeks of standard urotherapy alone with urotherapy supplemented with a timer watch in 58 children aged 5 to 14 years with overactive bladder and at least one episode of daytime urinary incontinence per week. Standard urotherapy included advice on fluid intake, scheduled voiding at two‐hour intervals until bedtime and bladder diaries. Timer watches had seven alarms set to remind children of the voiding schedule.

Klijn 2006 compared 24 weeks of urotherapy alone with urotherapy supplemented during weeks 1 to 8 with a home video in 143 children and adolescents aged 6 to 16 years with bladder instability, the majority (95/143) of whom had daytime incontinence. A third arm involved standard urotherapy plus home uroflow biofeedback. Urotherapy included bladder education, scheduled voiding, voiding charts, voiding posture, instructions on proper defecation patterns with or without laxatives as needed, and prophylactic antibiotics. The home video featured urotherapy instructions presented by a popular children’s television host, with an additional personalised section recorded during the first clinic visit, and was watched at least once daily during the study period.

Rhodes 2008 compared standard urotherapy with urotherapy supplemented with a bladder education work book in 20 children aged 6 to 10 years with daytime incontinence with or without nocturnal enuresis. The bladder workbook provided a fortnightly lesson plan for bladder rehabilitation, including delayed voiding and advice on fluid intake. Standard urotherapy included verbal urotherapy advice only, analogous to the contents of the workbook, provided by a urology nurse. At weeks 2 and 4 all parents received phone calls to offer further advice on bladder control, encourage adherence with advice and collect outcome data. At weeks 2 and 4 the mean number of days per week on which daytime urinary incontinence occurred were 3.9 and 2.8 respectively in the 10 children in the workbook group, and 5.2 and 3.5 days respectively in the standard urotherapy group.

Five studies compared behavioural modification interventions alone with behavioural modification interventions combined with pharmaceutical therapies (Ayan 2007; Hellstrom 1989; Kroll 2006; Newgreen 2017; Van Gool 2014).

Ayan 2007 compared urotherapy alone with urotherapy combined with placebo or with anticholinergics in 72 children aged 4 to 12 years with dysfunctional voiding, half of whom had daytime urinary incontinence. Urotherapy included instruction on timed voiding, double voiding and relaxation of the pelvic floor muscles during voiding. Anticholinergic treatment was with tolterodine 1 mg twice daily.

Hellstrom 1989 included 61 children aged 6 to 14 years with urinary urgency and instability, the majority of whom had daytime urinary incontinence (46/58 at least daily, 8/58 once or several times per week, one child once or twice per month; 3/58 had urgency or frequency without incontinence). All received information and training on how to gain bladder control, how to initiate voiding, voiding four to six times per day, recording of time of each void and incontinence episodes and conduct of 12‐hour pad test (for hospital measurement of volume voided). In addition, they were randomised to receive either placebo with cross‐over to 12.5 mg of terodiline, 12.5 mg of terodiline followed by placebo or 12.5 mg of terodiline followed by terodiline.

Kroll 2006 compared timed voiding, the use of a voiding diary and instruction on toilet posture and pelvic floor relaxation with the same instruction on posture and relaxation combined with an α₁‐selective alpha blocker (doxasosin, 0.5 to 2.0 mg/day depending on body weight) in 60 children and adolescents aged 5 to 17 years with detrusor‐sphincter disco‐ordination. Of the children with daytime urinary incontinence at entry, incontinence had resolved at six weeks in 11 out of 19 in the alpha blocker group and 13 out of 20 in the urotherapy only group (data provided by the study author).

Newgreen 2017 included 189 children and adolescents aged 5 to 17 years with daytime urinary incontinence four or mor‐ times daily. All had hour‐long urotherapy sessions at four and two weeks before baseline that included education about normal urinary tract function, regular voiding habits, voiding posture, avoiding holding manoeuvres, fluid intake and preventing constipation, use of bladder diaries, use of a timer watch to promote seven voids in 24 hours. At baseline they were randomised to one of four groups based on their age (Group I and Group II consisted of children, while Group III and Group IV consisted of adolescents). The participants were randomised to receive anticholinergic (solifenacin oral suspension once daily) with urotherapy or placebo alongside continuing urotherapy for 12 weeks.

Van Gool 2014 compared three interventions: urotherapy plus anticholinergics; urotherapy plus placebo; and urotherapy plus voiding education with uroflowmetry and feedback. Further details are presented in the feedback interventions section above.

Outcomes

Seventeen studies reported the number of children no longer experiencing daytime urinary incontinence after treatment, with follow‐up ranging from six weeks to 12 months (Borch 2017; Campos 2013; de Paula 2017; Golli 2013; Halliday 1987; Kajbafzadeh 2011; Klijn 2006; Kroll 2006; Ladi Seyedian 2014; Lordelo 2010; Patidar 2015; Quintiliano 2015; Sillen 2014; Van Gool 2014; Vasconcelos 2007; Zeng 2012; Zivkovic 2010).

Twelve studies reported frequency of daytime urinary incontinence episodes following treatment, with follow‐up ranging from four weeks to 12 months (Borch 2017; Campos 2013; de Paula 2017; Fielding 1980; Hagstroem 2009; Hagstroem 2010; Hellstrom 1989; Newgreen 2017; Reis 2014; Rhodes 2008; Vasconcelos 2007; Zeng 2012). One study reported parental report of "improvement" in daytime urinary incontinence at three months (Nemett 2008). Several studies reported composite symptom scales that included frequency and severity of symptoms: one study reported a study‐specific incontinence score (Hagstroem 2009), one a Vancouver Bladder and Bowel Dysfunction score (Brownrigg 2015), and four reported the Dysfunctional Voiding Scoring System (DVSS) (Ayan 2007; Golli 2013; Quintiliano 2015; Reis 2014).

Two studies reported a quality‐of‐life measure: the Pediatric Incontinence Questionnaire (PinQ) (Brownrigg 2015); the AUQUEI (Autoquestionnaire Qualité de Vie Enfant Imagé) questionnaire (Reis 2014).

Eight studies reported adverse events (Ayan 2007; Campos 2013; Hellstrom 1989; Klijn 2006; Kroll 2006; Quintiliano 2015; Sillen 2014; Van Gool 2014).

None of the studies reported data on objective measurements of volume of incontinent episodes, psychometric or behavioural measurements, cost or cost‐effectiveness of outcomes or resource implications.

Excluded studies

Reasons for exclusion of 55 reports or records (from 51 studies) during full‐text screening are included in the Characteristics of excluded studies.

Common reasons for exclusion included study populations (small proportions with daytime incontinence, ages below or above our inclusion criteria, populations with known causes of incontinence, studies on nocturnal enuresis only), interventions compared (only reporting on pharmacological, surgical or other invasive interventions), study design (observational studies, studies with no comparator, reviews) and lack of usable or relevant outcome data.

Ongoing studies

We identified three potentially relevant ongoing studies that may be eligible for inclusion in future updates of this review:

• One study compares a watch with preset alarms as a voiding reminder with a standard watch and standard advice on timed voiding in children aged 3 to 15 years with daytime urinary incontinence at least twice weekly (ACTRN12611000828921).

• A second study compares urotherapy with constipation treatment with urotherapy alone in children aged 5 to 17 years with lower urinary tract dysfunction and daytime (or daytime and night‐time) urinary incontinence (NCT02336906).

• A third study compares a "voiding school", in which small groups of children aged 5 to 6 with daytime urinary incontinence or nocturnal enuresis are educated using child‐orientated methods to try to achieve better bladder control, with usual care (individual advice concerning voiding habits and general lifestyle advice) (NCT03478813).

Further details can be found in the Characteristics of ongoing studies.

Risk of bias in included studies

The risk of bias for each included study is detailed in the Characteristics of included studies and overviews are presented in Figure 2 and Figure 3. Only one study was at low risk of bias for all domains (Hagstroem 2009), seven studies were at unclear risk of bias for at least one domain (Ayan 2007; Halliday 1987; Hellstrom 1989; Lordelo 2010; Newgreen 2017; Patidar 2015; Quintiliano 2015), and 19 were at high risk of bias for at least one domain (Borch 2017; Brownrigg 2015; Campos 2013; de Paula 2017; Fielding 1980; Golli 2013; Hagstroem 2010; Kajbafzadeh 2011; Klijn 2006; Kroll 2006; Ladi Seyedian 2014; Nemett 2008; Reis 2014; Rhodes 2008; Sillen 2014; Van Gool 2014; Vasconcelos 2007; Zeng 2012; Zivkovic 2010). The most common sources of bias related to sequence generation, concealment of allocation and lack of blinding of study participants or personnel during the interventions or at outcome assessment.

2.

'Risk of bias' summary: review authors' judgements about each 'Risk of bias' item for each included study

3.

'Risk of bias' graph: review authors' judgements about each 'Risk of bias' item presented as percentages across all included studies

Allocation

Random sequence generation

Fourteen studies reported adequate random sequence generation and we therefore judged these to be at low risk of bias (Borch 2017; Brownrigg 2015; Campos 2013; de Paula 2017; Hagstroem 2009; Hellstrom 1989; Kajbafzadeh 2011; Klijn 2006; Lordelo 2010; Quintiliano 2015; Sillen 2014; Vasconcelos 2007; Zeng 2012; Zivkovic 2010). The remaining thirteen studies did not report adequate methods of sequence generation and so we judged these studies to be at an unclear risk of bias (Ayan 2007; Fielding 1980; Golli 2013; Hagstroem 2010; Halliday 1987; Kroll 2006; Ladi Seyedian 2014; Nemett 2008; Newgreen 2017; Patidar 2015; Reis 2014; Rhodes 2008; Van Gool 2014).

Allocation concealment

Seven studies reported adequate allocation concealment procedures and we assessed them as being at low risk of selection bias (Brownrigg 2015; Hagstroem 2009; Halliday 1987; Hellstrom 1989; Klijn 2006; Newgreen 2017; Van Gool 2014). Twenty studies did not adequately report methods to conceal allocation (Ayan 2007; Borch 2017; Campos 2013; de Paula 2017; Fielding 1980; Golli 2013; Hagstroem 2010; Kajbafzadeh 2011; Kroll 2006; Ladi Seyedian 2014; Lordelo 2010; Nemett 2008; Patidar 2015; Quintiliano 2015; Reis 2014; Rhodes 2008; Sillen 2014; Vasconcelos 2007; Zeng 2012; Zivkovic 2010). We assessed these studies as being at unclear risk of bias.

Blinding

Only five studies reported blinding of participants and personnel and we assessed them at low risk of performance bias (Borch 2017; Hagstroem 2009; Halliday 1987; Hellstrom 1989; Newgreen 2017). We assessed six studies as unclear risk of performance bias due to not clearly reporting blinding status of participants and personnel (Ayan 2007; de Paula 2017; Lordelo 2010; Patidar 2015; Quintiliano 2015; Van Gool 2014), while we assessed 16 studies as high risk of performance bias due to no or partial blinding of participants, personnel, or both (Brownrigg 2015; Campos 2013; Fielding 1980; Golli 2013; Hagstroem 2010; Kajbafzadeh 2011; Klijn 2006; Kroll 2006; Ladi Seyedian 2014; Nemett 2008; Reis 2014; Rhodes 2008; Sillen 2014; Vasconcelos 2007; Zeng 2012; Zivkovic 2010).

Seven studies reported adequate blinding of outcome assessors (Borch 2017; Brownrigg 2015; Hagstroem 2009; Hellstrom 1989; Lordelo 2010; Patidar 2015; Quintiliano 2015). We assessed 11 studies to be at unclear risk of detection bias due to insufficient information on blinding of outcome assessment (Ayan 2007; Campos 2013; de Paula 2017; Hagstroem 2010; Halliday 1987; Klijn 2006; Newgreen 2017; Rhodes 2008; Van Gool 2014; Vasconcelos 2007; Zeng 2012), while nine were at high risk of bias because they reported outcomes by unblinded participants or parents and guardians (Fielding 1980; Golli 2013; Kajbafzadeh 2011; Kroll 2006; Ladi Seyedian 2014; Nemett 2008; Reis 2014; Sillen 2014; Zivkovic 2010).

Incomplete outcome data

We assessed 20 studies to be at low risk of bias because of low attrition rates and accounting for losses to follow‐up or missing data (Ayan 2007; Brownrigg 2015; Golli 2013; Hagstroem 2009; Hagstroem 2010; Halliday 1987; Hellstrom 1989; Kajbafzadeh 2011; Klijn 2006; Kroll 2006; Ladi Seyedian 2014; Lordelo 2010; Newgreen 2017; Patidar 2015; Rhodes 2008; Sillen 2014; Van Gool 2014; Vasconcelos 2007; Zeng 2012; Zivkovic 2010). We assessed two studies as being at unclear risk of bias because of insufficient or unclear information about attrition (Campos 2013; Quintiliano 2015). We assessed five studies to be at high risk of bias due to high or unexplained levels of attrition, or both (Borch 2017; de Paula 2017; Fielding 1980; Nemett 2008; Reis 2014).

Selective reporting

We assessed 19 included studies to be at low risk of reporting bias because they reported all expected outcomes (Ayan 2007; Brownrigg 2015; de Paula 2017; Hagstroem 2009; Hagstroem 2010; Halliday 1987; Kajbafzadeh 2011; Klijn 2006; Kroll 2006; Ladi Seyedian 2014; Lordelo 2010; Newgreen 2017; Patidar 2015; Quintiliano 2015; Reis 2014; Sillen 2014; Vasconcelos 2007; Zeng 2012; Zivkovic 2010). We assessed four studies to be at unclear risk of bias. Of these, two (Fielding 1980; Rhodes 2008), did not fully describe their methods and outcome measurements by current standards (although there was no reason to suspect that outcomes were reported selectively). Although a third study did not report fully some outcomes, they provided reasons for the omission (Hellstrom 1989); and the authors of a fourth study provided explanations for the unclear reporting of outcomes and further data when contacted (Campos 2013).

We assessed four studies to be at high risk of bias because they did not report outcomes mentioned in the methods, or because they exclusively reported outcomes favourable to an intervention or gave them precedence (Borch 2017; Golli 2013; Nemett 2008; Van Gool 2014).

Other potential sources of bias

We assessed 23 included studies to be at low risk of bias from other sources (Ayan 2007; Borch 2017; Brownrigg 2015; Campos 2013; de Paula 2017; Fielding 1980; Golli 2013; Hagstroem 2009; Halliday 1987; Hellstrom 1989; Kajbafzadeh 2011; Klijn 2006; Kroll 2006; Ladi Seyedian 2014; Lordelo 2010; Patidar 2015; Quintiliano 2015; Reis 2014; Sillen 2014; Van Gool 2014; Vasconcelos 2007; Zeng 2012; Zivkovic 2010).

We assessed four studies to be at unclear risk of bias from other sources: one because most participants had previously had unsuccessful treatment by one of the interventions being compared, which have confounded the results; one because the treatment site was identified as being associated with outcome, which may indicate lack of uniform intervention delivery, and because the industry funder was involved in study design and interpretation; one because all study participants were also attending unregulated standard care in which they received a wide variety of treatments, including medications, on the advice of paediatric urologists as indicated; and one because participants were a small convenience sample, with no information on those who declined to enrol (Hagstroem 2010; Nemett 2008; Newgreen 2017; Rhodes 2008).

Effects of interventions

See: Table 1; Table 2; Table 3; Table 4; Table 5; Table 6; Table 7; Table 8; Table 9

One individual conservative intervention (lifestyle, behavioural or physical) versus no treatment

Seven studies were eligible for this comparison (de Paula 2017; Fielding 1980; Golli 2013; Hagstroem 2009; Halliday 1987; Lordelo 2010; Patidar 2015).

Four studies compared active TENS versus sham (placebo) TENS (de Paula 2017; Hagstroem 2009; Lordelo 2010; Patidar 2015). One study compared an active wetting alarm versus a sham (placebo) alarm (Halliday 1987). Two studies compared behavioural interventions versus no treatment for daytime urinary incontinence (Fielding 1980; Golli 2013).

Primary outcomes

Number of children no longer experiencing daytime urinary incontinence: child‐ or carer‐reported

TENS versus sham TENS

Of the 16 participants in de Paula 2017, two did not complete the study. Both at end of treatment and at 60 days post‐treatment, five out of seven in the TENS group and two out of seven in the sham TENS group reported resolution of daytime urinary incontinence. In Lordelo 2010, parents reported complete resolution of symptoms (including daytime urinary incontinence) at the end of treatment in 13 of 21 in the TENS group versus none of 16 in the sham TENS group. In Patidar 2015, 15 of 21 (71.42%) in the TENS group versus 2 of 16 (12.5%) in the sham TENS group reported complete resolution of daytime urinary incontinence after 12 weeks of treatment. Hagstroem 2009 did not report on this outcome.

In pooled analysis of data from three studies (de Paula 2017; Lordelo 2010; Patidar 2015), more children in the active TENS group experienced resolution of daytime urinary incontinence versus the sham TENS group (RR 4.89, 95% CI 1.68 to 14.21; I² 25%, tau² 0.23; low‐certainty evidence; Analysis 1.1; Table 1). There was little difference when we removed the study with a subgroup of children without daytime incontinence (Lordelo 2010), from the analysis (RR 3.89, 95% CI 1.52 to 9.93; I² 0%, tau² 0.0). There was little difference when we removed the small number of children with missing outcome data from the denominators for two studies (de Paula 2017; Patidar 2015), from the analysis (RR 4.69, 95% CI 1.62 to 13.58; I² 27%, tau² 0.24).

1.1. Analysis.

Comparison 1 Transcutaneous electrical nerve stimulation (TENS) versus sham (placebo) TENS, Outcome 1 Resolution of daytime urinary incontinence at treatment completion.

Incontinence alarm versus sham alarm

Three months after starting treatment in Halliday 1987, 16 of 22 children in the genuine wetting alarm group had been dry for six weeks versus 13 of 22 children in the sham alarm group (RR 1.23, 95% CI 0.80 to 1.89); 16 of 20 and 13 of 19 if we include only those who completed the study (RR 1.17, 95% CI 0.80 to 1.70).

Behavioural intervention versus no treatment

At the end of the three‐month programme in Golli 2013, 15 of 29 children (51.7%) who completed the programme reported being completely dry, while no children in the delayed treatment comparison group reported resolution of symptoms. However, although Golli 2013 did not report proportions by group, overall at baseline 27 of 64 children recruited (42.2%) did not have daytime urinary incontinence and they reported that there were no differences between groups. Thus, any effect of the intervention in terms of fully resolving daytime urinary incontinence was modest.

Secondary outcomes

Frequency of incontinence episodes, of pad or clothes changes

TENS versus sham TENS

In Hagstroem 2009, median (interquartile range (IQR)) wet days per week changed from seven (6 to 7) to four (3 to 7) (before and after comparison P < 0.01; Mann‐Whitney) in the TENS group and from five (4 to 7) to 5.5 (4.5 to 7) in the sham TENS group. While there was a reduction in the number of wet days per week in the TENS group (median −3; IQR −3 to 0), the same was not true in the sham TENS group (median 0; IQR −1.5 to 1.5; between‐arms comparison at follow‐up P < 0.05; Mann‐Whitney). Median (IQR) daytime incontinence episodes per day changed from 1.5 (1 to 2) to 0.5 (0 to 1.5; before and after comparison P < 0.05; Mann‐Whitney) in the TENS group and from 0.75 (0 to 2.5) to 1 (1 to 2.5) in the sham TENS group. The median (IQR) change (median of differences from baseline to follow‐up time) in daytime incontinence episodes per day was −1 (−1 to −0.5) in the TENS group and 0 (−0.5 to 1) in the sham TENS group (between‐arms comparison at follow‐up P < 0.01; Mann‐Whitney). A study‐specific one‐week incontinence score was calculated by adding scores for each day (0 = a day with no incontinence, 1 = a day with incontinence limited to wet underwear, 2 = a day with wet outer clothes, 3 = a day with wet socks). In the TENS group the median (IQR) incontinence score fell from 13 (11 to 14) before the study to 5 (3 to 14; before and after comparison P < 0.01; Mann‐Whitney). In the sham TENS group, the median score changed from 10 (5 to 13) to 7.5 (5.5 to 10.5).

In de Paula 2017, parents reported their perception of improvement of their child's overall symptoms (overactive bladder and incontinence) by a visual analogue scale (VAS) from 0 to 10, with zero meaning 'no improvement' and 10 meaning 'the complete resolution of symptoms'. Both at end of treatment and at 60 days post‐treatment, the median (range) response was 8 (7 to 8) in the TENS group versus 6 (5 to 7) in the sham TENS group.

In Lordelo 2010, at the end of treatment amongst children who completed the study, parents reported improvement of symptoms in 8 of 21 in the TENS group versus 5 of 16 in the sham TENS group.

Behavioural intervention versus no treatment

In Fielding 1980, in the subgroup of children with daytime incontinence who completed the study (n = 8 in each arm) the mean number of days with urinary incontinence per four weeks reduced by 6.7 in the bladder training group (from 17.6 days per four weeks at baseline) and by 7.5 in the alarm only group (from 14.4).

Objective measurement of volume of incontinent episodes (e.g. pad tests)

Not reported.

Psychometric or behavioural measurements

Not reported.

Resource implications (cost of incontinence management methods)

Not reported.

Quality of life measurements

Not reported.

Adverse events or child‐ or carer‐perceived harms associated with interventions reported by studies

Not reported.

Combined conservative interventions versus no treatment

We did not identify any studies for this comparison.

One individual conservative intervention versus another individual or combined conservative intervention

Ten studies were eligible for this comparison (Brownrigg 2015; Hagstroem 2010; Kajbafzadeh 2011; Klijn 2006; Ladi Seyedian 2014; Nemett 2008; Rhodes 2008; Sillen 2014; Van Gool 2014; Zivkovic 2010). Three of these were multi‐arm studies (Klijn 2006; Nemett 2008; Van Gool 2014).

Three studies, including one multi‐arm study, compared PFMT combined with behavioural interventions versus behavioural interventions alone (Ladi Seyedian 2014; Nemett 2008; Zivkovic 2010). Two multi‐arm studies compared voiding education with feedback alongside standard urotherapy versus standard urotherapy alone (Klijn 2006; Van Gool 2014). One study compared PFMT, biofeedback and behavioural modification versus behavioural modification alone (Kajbafzadeh 2011).

One study compared TENS plus behavioural urotherapy versus behavioural urotherapy alone (Sillen 2014). One study compared standard urotherapy alone versus urotherapy plus timer watch (Hagstroem 2010). Two studies, including one multi‐arm study, compared standard urotherapy versus enhanced urotherapy plus supportive media (Klijn 2006; Rhodes 2008). One study compared one‐to‐one urotherapy versus group urotherapy (Brownrigg 2015).

Primary outcomes

Number of children no longer experiencing daytime urinary incontinence: child‐ or carer‐reported

PFMT combined with behavioural interventions versus behavioural interventions alone

Of the children with daytime urinary incontinence at baseline in Ladi Seyedian 2014, 15 of 17 children in the PFMT and urotherapy group reported that it was resolved compared to 4 of 11 children in the urotherapy‐alone group. Zivkovic 2010 reported that baseline daytime urinary incontinence was resolved in 20 of 24 children in the PFMT, diaphragmatic exercises and urotherapy group compared to 2 of 18 in the urotherapy‐alone group. Kajbafzadeh 2011 reported that baseline daytime urinary incontinence was resolved in 8 of 10 children at six months and 7 of 10 children at 12 months in the PFMT, voiding education with animated biofeedback and urotherapy group. In the urotherapy‐alone group, daytime urinary incontinence was resolved in 8 of 11 children at both 6 and 12 months.

Pooled analysis of data from the three studies suggests that there was little evidence to support any difference between PFMT with urotherapy versus urotherapy alone in terms of resolution of daytime urinary incontinence at 12 months (RR 2.36, 95% CI 0.65 to 8.53; Kajbafzadeh 2011; Ladi Seyedian 2014;Zivkovic 2010; I² 86%, tau² 1.08; very low‐certainty evidence; Analysis 2.1; Table 2). When we removed the study with additional voiding education with feedback from the analysis (Kajbafzadeh 2011), more children in the PFMT with urotherapy group experienced resolution of daytime incontinence but the estimate of effect had greater imprecision (RR 3.83, 95% CI 1.18 to 12.41; I² 59%, tau² 0.44).

2.1. Analysis.

Comparison 2 Pelvic floor muscle training and urotherapy versus urotherapy alone, Outcome 1 Resolution of daytime urinary incontinence at 12 months.

Voiding education with feedback alongside standard urotherapy versus standard urotherapy alone

At 12 months (28 weeks after completion) in Klijn 2006, 19 of 30 children in the standard urotherapy group reported resolution of daytime urinary incontinence, compared with 25 of 33 in the home uroflow biofeedback group (17/27 and 22/28 if only those completing the study are included). Of children with overactive bladder with urge incontinence at baseline in Van Gool 2014, the proportions reporting resolution of daytime urinary incontinence were similar between groups: 15 of 34 in the voiding education with uroflowmetry and feedback alongside urotherapy arm versus 13 of 33 in the urotherapy with placebo arm (15/31 and 13/24 respectively, if only those completing the study are included). Kajbafzadeh 2011 reported that daytime urinary incontinence at baseline was resolved in 8 of 10 children at six months and 7 of 10 children at 12 months in the PFMT, voiding education with animated biofeedback and urotherapy group. In the urotherapy‐alone group, daytime urinary incontinence was resolved in 8 of 11 children at both 6 and 12 months.

Pooled analysis of data from the three studies (Kajbafzadeh 2011; Klijn 2006; Van Gool 2014), showed negligible difference in resolution of incontinence between urotherapy alone or with bladder education with feedback in terms of resolution of daytime urinary incontinence (RR 1.13, 95% CI 0.87 to 1.45; I² 0%, tau² 0.00; low‐certainty evidence; Analysis 3.1; Table 3). The same was true when we removed Kajbafzadeh 2011 (the study with additional PFMT) from the analysis (RR 1.18, 95% CI 0.88 to 1.57; I² 0%, tau² 0.00). It should be noted that the time interval between end of treatment and outcome assessment was different in the studies: Van Gool 2014 did not include the treatment period in the 12‐month follow‐up period but Kajbafzadeh 2011 and Klijn 2006 did.

3.1. Analysis.

Comparison 3 Voiding education with feedback alongside standard urotherapy versus standard urotherapy alone, Outcome 1 Resolution of daytime urinary incontinence at 12 months.

Standard urotherapy alone versus urotherapy plus timer watch

In Hagstroem 2010, none of the 28 children in the standard urotherapy‐alone group achieved complete resolution of daytime urinary incontinence after 12 weeks compared to 9 of 30 in the timer watch group (RR 17.77, 95% CI 1.08 to 291.82; very low‐certainty evidence; Table 4).

TENS plus behavioural urotherapy versus behavioural urotherapy alone

At the end of 12 weeks of treatment in one study (Sillen 2014), 16 of 24 children in the TENS plus urotherapy group reported resolution of daytime incontinence versus 13 of 28 in the urotherapy‐alone group. Subgroup analysis found that of the children who had not previously had any treatment for their symptoms (urotherapy or anticholinergics), 12 of 18 in the TENS plus urotherapy group experienced resolution of daytime incontinence compared to 10 of 22 in the urotherapy‐alone group.

Standard urotherapy versus enhanced urotherapy plus supportive media

At 12 months in one study (Klijn 2006; 28 weeks after completion), 19 of 30 children with daytime urinary incontinence in the standard urotherapy group reported resolution of incontinence, compared with 22 of 32 in the home‐video group (17 of 27, and 22 of 32 if only those completing the study are included).

Secondary outcomes

Frequency of incontinence episodes, of pad or clothes changes

PFMT combined with behavioural interventions versus behavioural interventions alone

Of the 21 children who completed one study (Nemett 2008), 6 of 10 in the osteopathy group had daytime urinary incontinence at baseline compared to 8 of 11 in the standard urological care‐alone group. Of these, 6 of 6 and 6 of 8 reported improvement in daytime urinary incontinence at 10 to 12 weeks, but treatments provided during standard urological care were not reported, so that any effect of the osteopathy‐based physical therapy is difficult to determine.

Standard urotherapy alone versus urotherapy plus timer watch

In Hagstroem 2010 after 12 weeks, 0 of 28 children in the standard urotherapy group achieved a 90% or greater reduction in daytime urinary incontinence episodes versus 10 of 30 in the timer watch group, 5 of 28 versus 8 of 30 achieved a 50% to 89% reduction and 23 of 28 versus 12 of 30 less than a 50% reduction. The mean (SD) increase in dry days per week was +0.6 (+1.8) in the standard urotherapy group compared to +3.5 (+2.5) in the timer watch group (mean difference 2.90, 95% CI 1.78 to 4.02; Table 4). At baseline, children receiving standard urotherapy alone reported median (IQR 25% to 75%) number of wet days per week of 6 (3 to 7) while those in the additional timer watch group reported 7 (6 to 7). These figures were changed at 12 weeks to 5 (3 to 7) and 2 (0 to 5) respectively (P < 0.01; Mann‐Whitney).

TENS plus behavioural urotherapy versus behavioural urotherapy alone

Sillen 2014 reported a decreased number of daytime urinary incontinence episodes in 19 of 24 children in the TENS plus urotherapy group versus 24 of 28 in the urotherapy‐alone group. The mean (SD) change in the number of daytime urinary incontinence episodes from baseline to 12 weeks was reduced within both groups (−1.5 (1.3) in the TENS plus urotherapy group and −1.5 (1.4) in the urotherapy‐alone group, both P < 0.0001; Mann‐Whitney) but there was no difference between groups.

Standard urotherapy versus enhanced urotherapy plus supportive media

At weeks 2 and 4 in one study (Rhodes 2008), the mean number of days per week on which daytime urinary incontinence occurred were 3.9 and 2.8 respectively in the 10 children in the workbook group, compared to 5.2 and 3.5 days respectively in the standard urotherapy group. At baseline, participants were asked how many times they wet per day rather than how many days wet per week (not reported), so it was not possible to determine any difference between baseline and end of study.

One‐to‐one urotherapy versus group urotherapy

Brownrigg 2015 measured bladder and bowel dysfunction symptoms such as incontinence, urgency, frequency, dysuria, hesitancy, straining and constipation using the Vancouver Questionnaire, a 14‐item, 5‐point Likert scale instrument at baseline and at a median of 14 weeks' follow‐up (Afshar 2009). There were reductions in mean (SD) symptom scores in both groups: from 19.5 (7.2) to 13.4 (6.3; P < 0.01; students t‐test) in the individual urotherapy arm and from 18.3 (7.6) to 14.7 (7.9; P = 0.03; students t‐test) in the group urotherapy arm.

Objective measurement of volume of incontinent episodes (e.g. pad tests)

Not reported.

Psychometric or behavioural measurements

Not reported.

Resource implications (cost of incontinence management methods)

Not reported.

Quality of life measurements

One‐to‐one urotherapy versus group urotherapy

Brownrigg 2015 used PinQ to measure six domains of quality of life: social, self‐esteem, family, body image, independence and mental health (Bower 2006). The child completed questionnaires at baseline and at follow‐up visit, under the guidance of a research assistant, with limited input from parents/guardians. There was an improvement in mean (SD) scores in the individual urotherapy arm from 31.0 (14.3) to 20.1 (15.3) but little change in the group urotherapy arm, from 21.1 (10.8) to 21.0 (14.2). However, the difference in quality of life measures between the two arms at baseline makes the result difficult to interpret.

Adverse events or child‐ or carer‐perceived harms associated with interventions reported by studies

Voiding education with feedback alongside standard urotherapy versus standard urotherapy alone

Van Gool 2014 reported one child with nervous tics and one with constipation in the placebo arm, compared to none in the voiding education with feedback arm. Klijn 2006 reported that there were no adverse events in any study arm.

TENS plus behavioural urotherapy versus behavioural urotherapy alone

Sillen 2014 reported two adverse events in the TENS plus urotherapy group. One child had an unspecified adverse reaction to the electrodes and one child had increased urinary incontinence during the treatment period.

Standard urotherapy versus enhanced urotherapy plus supportive media

Klijn 2006 reported that there were no adverse events in any study arm.

Combined conservative interventions versus other combined conservative interventions

Two studies were eligible for this comparison (Reis 2014; Vasconcelos 2007).

One study compared PFMT plus behavioural modification versus PFMT combined with electromyography biofeedback plus behavioural modification (Vasconcelos 2007). One study compared PFMT plus electromyographic biofeedback and standard urotherapy versus TENS plus standard urotherapy (Reis 2014).

Primary outcomes

Number of children no longer experiencing daytime urinary incontinence: child‐ or carer‐reported

TENS plus urotherapy versus PFMT plus electromyographic feedback and urotherapy

At the end of treatment in one study (Reis 2014), there was little difference in proportions with complete response in daytime symptoms according to the ICCS criteria: 60.6% (20 of 33 completing the study) in the TENS group compared to 54.9% (17 of 31 completing the study) in the PFMT with biofeedback group (RR 1.11, 95% CI 0.73 to 1.68; P = 0.483; very low‐certainty evidence; Table 5).

PFMT plus behavioural modification versus PFMT combined with electromyography biofeedback and behavioural modification

In Vasconcelos 2007, when only those who had daytime urinary incontinence at baseline were included, 12 of 20 in the feedback group reported no incontinence episodes in four weeks at six months compared to 13 of 21 in the no‐feedback group. At 12 months, 15 of 20 in the feedback group reported no incontinence episodes versus 15 of 21 in the no‐feedback group (RR 1.05, 95% CI 0.72 to 1.52; very low‐certainty evidence; Table 6).

Secondary outcomes

Frequency of incontinence episodes, of pad or clothes changes

TENS plus urotherapy versus PFMT plus electromyographic feedback and urotherapy

In Reis 2014, there were marked changes in mean number of urinary incontinence episodes per day in both the TENS group (from 1.85 (SD 1.48) to 0.12 (SD 0.33); P = 0.001) and in the PFMT with biofeedback group (from 1.94 (SD 1.50) to 0.14 (SD 0.39); P = 0.001), but no difference between groups. Changes in mean DVSS scores were also similar between groups: from 10.18 (SD 3.31) to 2.03 (SD 2.53) in the TENS group and from 10.20 (SD 3.16) to 2.22 (SD 2.60) in the PFMT with biofeedback group.

PFMT plus behavioural modification versus PFMT combined with electromyography biofeedback and behavioural modification

Vasconcelos 2007 reported on improvement of daytime urinary incontinence, defined as 50% or more reduction in wetting episodes during four‐week periods at 6 and 12 months. When we included only those who had daytime urinary incontinence at baseline and excluded those reporting no incontinence episodes. In the feedback group, 4 of 20 reported improvement at six months compared to 3 of 21 in the no‐feedback group. At 12 months, 3 of 20 in the feedback group reported improvement versus 2 of 21 in the no‐feedback group.

Objective measurement of volume of incontinent episodes (e.g. pad tests)

Not reported.

Psychometric or behavioural measurements

Not reported.

Resource implications (cost of incontinence management methods)

Not reported.

Quality of life measurements

TENS plus urotherapy versus PFMT plus electromyographic feedback and urotherapy

Reis 2014 reported no difference in the quality of life of children in either group before and after treatment, measured using the AUQUEI questionnaire (Assumpcao 2000).

Adverse events or child‐ or carer‐perceived harms associated with interventions reported by studies

Not reported.

Individual conservative interventions versus non‐conservative interventions (pharmacological or invasive, combined or not with any conservative interventions)

Nine studies were eligible for this comparison (Ayan 2007; Borch 2017; Campos 2013; Hellstrom 1989; Kroll 2006; Newgreen 2017; Quintiliano 2015; Van Gool 2014; Zeng 2012). Five of these studies were multi‐arm studies (Ayan 2007; Borch 2017; Hellstrom 1989; Newgreen 2017; Van Gool 2014). One of these multi‐arm studies was also a cross‐over study (Hellstrom 1989).

Two studies compared PFMT with anticholinergics (Campos 2013; Zeng 2012). Two studies, including one multi‐arm study, compared sacral TENS versus anticholinergics (Borch 2017; Quintiliano 2015). One three‐arm study compared one TENS plus anticholinergics versus TENS alone (Borch 2017).

One multi‐arm study compared urotherapy and oxybutynin versus urotherapy and placebo (Van Gool 2014). One multi‐arm study compared solifenacin plus urotherapy versus placebo plus urotherapy, split into separate groups for children and adolescents (Newgreen 2017). One study compared doxazosin versus behavioural modification (Kroll 2006). One multi‐arm, cross‐over study compared placebo versus cross‐over to terodiline, terodiline with cross‐over to placebo and terodiline followed by terodiline, and all groups received behavioural information and training (Hellstrom 1989). One multi‐arm study compared behavioural modification versus behavioural modification and tolterodine and behavioural modification and placebo (Ayan 2007). These five studies were not suitable for meta‐analysis.

Primary outcomes

Number of children no longer experiencing daytime urinary incontinence: child‐ or carer‐reported

PFMT versus anticholinergics

At four months (one month after completion) in Campos 2013, 7 of 21 in the oxybutynin arm were completely dry day and night by parental report, compared to 14 of 26 in the PFMT arm (P = 0.07). After 12 weeks of treatment, Zeng 2012 observed complete resolution of symptoms in children with post‐micturition dribbling in 16 of 21 in the PFMT group compared to 6 of 18 in the tolterodine group (P = 0.01; 16/20 and 6/16 respectively if only children receiving full treatment as planned are included).

In pooled analysis of data from the two studies (Campos 2013; Zeng 2012), more children receiving PFMT than anticholinergics experienced resolution of daytime urinary incontinence (RR 1.92, 95% CI 1.17 to 3.15; tau² 0.00, I² 0%; very low‐certainty evidence; Analysis 4.1; Table 7).

4.1. Analysis.

Comparison 4 Pelvic floor muscle training versus anticholinergics, Outcome 1 Resolution of daytime urinary incontinence.

Sacral TENS versus anticholinergics

In Borch 2017, 3 of 23 (3/14 of those completing) in the anticholinergics plus sham TENS group reported resolution of daytime urinary incontinence at 11 weeks compared to 0 of 21 (0/19 of those completing) in the active TENS plus placebo group. Quintiliano 2015, at three months, reported resolution of daytime urinary incontinence in 3 of 15 in the anticholinergics plus sham TENS group versus 6 of 13 in the active TENS plus placebo group.

In pooled analysis of data from the two studies (Borch 2017; Quintiliano 2015), there was insufficient evidence to support any difference between anticholinergics plus sham TENS compared to active TENS plus placebo (RR 0.81, 95% CI 0.05 to 12.50; I² 69%, tau² 2.82; very low‐certainty evidence; Analysis 5.1;Table 8).

5.1. Analysis.

Comparison 5 Transcutaneous electrical nerve stimulation (TENS) versus anticholinergics, Outcome 1 Resolution of daytime urinary incontinence.

TENS versus TENS plus anticholinergics

At 11 weeks, Borch 2017 reported that 8 of 22 children (8/18 of those completing) in the anticholinergics plus TENS group had resolution of daytime urinary incontinence compared to in 0 of 21 children (0/19 of those completing) in the active TENS plus placebo group.

Behavioural modification versus behavioural modification plus pharmaceutical therapies

At 12 months after the end of the six‐month treatment in one study (Van Gool 2014), proportions reporting resolution of daytime urinary incontinence were similar between groups: 13 of 33 in the urotherapy with placebo arm versus 13 of 30 in the urotherapy with oxybutynin arm (RR 1.02, 95%CI 0.58 to 1.78; very low‐certainty evidence; Table 9), 13 of 24 versus 13 of 25, respectively, if only those completing the study are included. Of the children with daytime urinary incontinence at entry into Kroll 2006, incontinence had resolved at six weeks in 11 of 19 in the alpha blocker group compared to 13 of 20 in the urotherapy‐only group (data provided by the study author).

Secondary outcomes

Frequency of incontinence episodes, of pad or clothes changes

PFMT versus anticholinergics

At four months in Campos 2013, the median (range) number of dry days per month was 11 (1 to 24) during the first treatment month and 16 (0 to 27) in the third month in the oxybutynin arm. These figures were 15.5 (0 to 27) and 24.5 (6 to 30) respectively in the PFMT arm (data provided by the study authors). After 12 weeks of treatment, Zeng 2012 observed significant improvement of post‐micturition dribbling (only one episode in the month after treatment) in 2 of 21 in the PFMT plus biofeedback group compared to 6 of 18 in the tolterodine group (18/21 and 12/18 respectively if those with complete resolution are included).

Sacral TENS versus anticholinergics

Borch 2017 reported response to treatment (defined according to ICCS guidelines as 50% to 99% reduction and not including complete resolution) in 5 of 23 (5/14 of those completing) in the anticholinergics plus sham TENS group versus 4 of 21 (4/19 of those completing) in the active TENS plus placebo group. When we combined complete resolution and response to treatment, these proportions were 8 of 23 (8/14) and 4 of 21 (4/19) respectively.

In Quintiliano 2015, mean (SD) DVSS scores improved in both groups from baseline to three months, from 12.2 (2.9) to 3.6 (1.8) in the anticholinergics plus sham TENS group compared to 11 (2.2) to 3.5 (2.7) in the active TENS plus placebo group, but there was no difference between groups.

TENS versus TENS plus anticholinergics

Borch 2017 reported response to treatment (defined according to ICCS guidelines as 50% to 99% reduction and not including complete resolution) in 6 of 22 (6/18 of those completing) in the anticholinergics plus TENS group compared to 4 of 21 (4/19 of those completing) in the active TENS plus placebo group. When we combined complete resolution and response to treatment, these proportions were 14 of 22 (14/18) and 4 of 21 (4/19) respectively. Active TENS plus anticholinergics was reportedly more effective compared to active TENS with placebo regarding improvement in mean number of wet days per week (P < 0.01; student t‐test) and frequency (P < 0.001; student t‐test).

Behavioural modification versus behavioural modification plus pharmaceutical therapies

While Hellstrom 1989 reported a marked difference between urotherapy with anticholinergics versus urotherapy with placebo in mean change in urinary incontinence episodes per day from baseline to four weeks (−1.1 and −0.3 respectively, P = 0.001; Wilcoxon sign rank test), Newgreen 2017 reported little difference between urotherapy with anticholinergics versus urotherapy with placebo at 12 weeks (−1.1 versus −1.2 in children and −0.6 versus −0.7 in adolescents). In Hellstrom 1989, the reduction in the mean number of urinary incontinence episodes per day from baseline to four weeks was marked in the urotherapy with anticholinergics group (from 2.6 to 1.7, P = 0.05; Wilcoxon sign rank test), but not in the urotherapy with placebo group (from 1.5 to 1.6). After cross‐over, children who had been taking placebo showed similar improvement to those taking anticholinergics. No standard deviations were provided for the majority of measurements in the studies and no suitable standard deviations or correlation coefficients from other studies were available for imputing, so we did not undertake any meta‐analysis (Higgins 2011c).

Similarly, while Hellstrom 1989 reported a marked difference between urotherapy with anticholinergics versus urotherapy with placebo in mean number of dry days per week (+3.0 versus +0.7, P = 0.003; Wilcoxon sign rank test), Newgreen 2017 reported no such difference between groups. Hellstrom 1989 also reported subjective improvement in overall bladder symptoms in 27 of 32 in the urotherapy with anticholinergics group compared with 18 of 26 in the urotherapy with placebo group.

At one and three months in Ayan 2007, the mean symptom score in the tolterodine group was markedly lower compared to baseline, while scores remained comparatively steady in the urotherapy‐alone and urotherapy plus placebo groups. Mean scores for a subgroup of DVSS questions relating to detrusor overactivity were markedly different at one month of treatment in the tolterodine group (8.90 at baseline changing to 3.50 at one month, P 0.001; Mann‐Whitney), while there was a smaller difference in the urotherapy plus placebo group (9.40 at baseline, changing to 5.25 at one month, P < 0.05) and little difference in the urotherapy‐alone group (7.60 versus 6.90, P > 0.05).

Objective measurement of volume of incontinent episodes (e.g. pad tests)

Not reported.

Psychometric or behavioural measurements

Not reported.

Resource implications (cost of incontinence management methods)

Not reported.

Quality of life measurements

Not reported.

Adverse events or child‐ or carer‐perceived harms associated with interventions reported by studies

PFMT versus anticholinergics

Campos 2013 reported that there were no adverse events.

Sacral TENS versus anticholinergics

Quintiliano 2015 reported adverse events in the anticholinergics plus sham TENS group: 58% had dry mouth, 25% had hyperthermia and 50% had hyperemia. Borch 2017 reported the development of significant post‐void residual in seven children and urinary tract infection in one child in the anticholinergics plus sham TENS group. One serious adverse event (appendicitis) occurred in the active TENS plus placebo group, not considered to be related to the study treatment (Borch 2017).

TENS versus TENS plus anticholinergics

Borch 2017 reported the development of significant post‐void residual in one child in the anticholinergics plus TENS group versus one serious adverse event (appendicitis) in the active TENS plus placebo group, not considered to be related to the study treatment.

Behavioural modification versus behavioural modification plus pharmaceutical therapies

Kroll 2006 reported side effects in 6 of 30 children receiving an alpha blocker (headache, hypotonia, vertigo, epistaxis) compared with none in children receiving urotherapy only. Ayan 2007 discontinued tolterodine in one child in whom hand and foot oedema developed, although a direct correlation with the drug was not established. Two children receiving tolterodine reported headache and five reported dry mouth. No adverse events were reported in the urotherapy‐alone and urotherapy plus placebo groups. Hellstrom 1989 reported one case of mild nausea in the terodiline group compared with one case each of mild nausea, mild headache and moderate trembling in the placebo group. Van Gool 2014 reported minor behavioural problems in two children in the oxybutynin arm compared with one child with nervous tics and one with constipation in the placebo arm.

Newgreen 2017 reported both serious adverse events and non‐serious treatment‐emergent adverse events. Serious adverse events in children aged 5 to 12 years included frontal lobe epilepsy (1/73) and pyelonephritis (1/73) in the solifenacin group, compared with lymphadenitis (1/73), hypertension (1/73) and tachycardia (1/73) in the placebo group. In adolescents, there was one case of appendicitis in the solifenacin group one of abdominal pain in the placebo group. No serious adverse events were considered to be related to solifenacin treatment. In children in the solifenacin group, non‐serious treatment‐emergent adverse events included: constipation (4/73); electrocardiogram (ECG) QT prolonged (4/73); dry mouth (2/73); faecal incontinence (1/73); rectal fissure (1/73); flu‐like illness (1/73); nasopharyngitis (1/73); attention disturbance (1/73); altered mood (1/73); tic (1/73); haematuria (1/73); erythema (1/73); and urticaria (1/73). In children in the placebo group, non‐serious treatment‐emergent adverse events included: constipation (2/73); ECG QT prolonged (2/73); dry mouth (1/73); diarrhoea (1/73); nausea (1/73); and fatigue (1/73). All adverse events were considered mild or moderate in severity, except for severe faecal incontinence in one solifenacin‐treated child. In adolescents in the solifenacin group, non‐serious treatment‐emergent adverse events included: ECG QT prolonged (1/22); dry throat (1/22); and cystitis (1/22). In adolescents in the placebo group non‐serious treatment‐emergent adverse events included: ECG QT prolonged (1/19); dry mouth (1/19); abdominal pain (1/19); dizziness (1/19); headache (1/19); somnolence (1/19); and blurred vision (1/19).

Combined conservative interventions versus non‐conservative interventions (pharmacological or invasive, combined or not with any conservative interventions)

Two arms of a three‐armed study compared voiding education with uroflowmetry and feedback with anticholinergics and urotherapy (Van Gool 2014).

Primary outcomes

Number of children no longer experiencing daytime urinary incontinence: child‐ or carer‐reported

Voiding education plus uroflowmetry and feedback versus anticholinergics

In one study (Van Gool 2014), at 12 months after the end of the six‐month treatment, proportions reporting resolution of daytime urinary incontinence were similar between groups: 13 of 30 in the oxybutynin arm compared with 15 of 34 in the with‐feedback arm (RR 1.02, 95% CI 0.58 to 1.78; very low‐certainty evidence; Table 9), 13 of 25 versus 15 of 31 respectively, if we include only those completing the study.

Secondary outcomes

Frequency of incontinence episodes, of pad or clothes changes

Not reported.

Objective measurement of volume of incontinent episodes (e.g. pad tests)

Not reported.

Psychometric or behavioural measurements

Not reported.

Resource implications (cost of incontinence management methods)

Not reported.

Quality of life measurements

Not reported.

Adverse events or child‐ or carer‐perceived harms associated with interventions reported by studies

Voiding education plus uroflowmetry and feedback versus anticholinergics

Van Gool 2014 reported minor behavioural problems in two children in the oxybutynin arm versus none in the voiding education with feedback arm.

Sensitivity analyses

We noted levels of attrition for included studies and planned to use sensitivity analyses to assess the impact of including studies with high levels (above 20%) of missing data (Sackett 1997). However, data were only missing for a small proportion of participants in two studies. We conducted analyses for all outcomes, insofar as was possible, on an intention‐to‐treat basis by using a conservative approach and imputing missing participants as not having experienced the outcome in all studied intervention arms.

We preferred to impute rather than exclude missing outcome data in order to maintain the randomised sample and by extent, the benefits of randomisation. For that purpose, we used a conservative approach and we imputed missing participants as not having experienced the outcome in all studied intervention arms.

Due to the very limited number of studies in the forest plots, we decided not to exclude studies with high or unclear risk of attrition bias in a sensitivity analysis, but to repeat the meta‐analysis for these outcomes by excluding missing participants from all studied intervention arms – this results in reduction of the study sample size. This sensitivity analysis strategy is in line with the recommendations in Section 16.1.2 of the Cochrane Handbook (Higgins 2011c).

This only applies to one analysis (Analysis 1.1), and we have added the resulting relatively unchanged effect estimate to the Results.

Discussion

Summary of main results

We aimed to assess the effectiveness of conservative (non‐invasive and non‐pharmaceutical) interventions for treating functional daytime urinary incontinence in children. The review therefore included studies that compared many different conservative treatments or combinations with other conservative treatments or combinations, as well as studies that compared conservative treatments or combinations with pharmaceutical treatment.

We included 27 studies involving 1803 children. Most studies were small, with numbers randomised ranging from 16 to 202. Few studies reported data suitable for pooling due to heterogeneity in interventions, outcomes and measurements. We assessed the certainty of evidence for all comparisons and outcomes as low or very low, meaning that we can say little about any intervention's true effect.

Number of children no longer experiencing daytime urinary incontinence

There is low‐certainty evidence that TENS may be more effective than sham stimulation (de Paula 2017; Lordelo 2010; Patidar 2015; Table 1).

Very low‐certainty evidence means that we are uncertain whether there is any difference between PFMT with urotherapy and urotherapy alone in resolving daytime urinary incontinence (Kajbafzadeh 2011; Ladi Seyedian 2014; Zivkovic 2010; Table 2). There was low‐certainty evidence that there may be no difference between voiding education with uroflow feedback and urotherapy and urotherapy alone (Kajbafzadeh 2011; Klijn 2006; Van Gool 2014; Table 3). We are uncertain whether standard urotherapy supplemented with timer watches with alarms set to remind children of the voiding schedule is more effective than urotherapy alone due to very low‐certainty evidence (Hagstroem 2010; Table 4).

We are uncertain whether there is any difference between TENS and PFMT (Reis 2014; Table 5). We are uncertain whether there is any difference between PFMT with or without electromyography biofeedback (Vasconcelos 2007; Table 6).

We are uncertain whether PFMT is more effective than anticholinergics in children with daytime urinary incontinence and postmicturition dribble due to very low‐certainty evidence (Campos 2013; Zeng 2012; Table 7). We are uncertain whether there is any difference between transcutaneous electrical nerve stimulation and anticholinergics (Borch 2017; Quintiliano 2015; Table 8), due to very low‐certainty evidence.

We are uncertain whether there is any difference between voiding education with uroflowmetry feedback and anticholinergics (Van Gool 2014; Table 9).

Frequency of incontinence episodes

We are uncertain whether active sacral TENS may be more effective than sham TENS in improving a composite incontinence score in children with daytime urinary incontinence (de Paula 2017; Hagstroem 2009; Lordelo 2010; Table 1).

We are uncertain whether there is any difference between standard urological care alongside osteopathy‐based physical therapy and standard urological care alone (Nemett 2008; Table 2). For urotherapy and other behavioural interventions, we are uncertain whether standard urotherapy supplemented with timer watches with alarms set to remind children of the voiding schedule is more effective than urotherapy alone in reducing incontinence episodes due to very low‐certainty evidence (Hagstroem 2010; Table 4).

We are uncertain whether there is any difference between TENS and PFMT (Reis 2014; Table 5).

We are uncertain whether there is any difference between PFMT with or without electromyography biofeedback (Vasconcelos 2007; Table 6).

The very low‐certainty evidence means that we are uncertain whether PFMT is better than anticholinergics in children with daytime urinary incontinence and postmicturition dribble (Campos 2013; Zeng 2012; Table 7), or whether TENS plus anticholinergics is more effective than stimulation alone (Borch 2017; Table 8).

No studies comparing voiding education with uroflow feedback and urotherapy versus urotherapy alone, or voiding education with uroflowmetry and feedback compared with anticholinergics, assessed this outcome (Table 3; Table 9).

Objective measures of volume of incontinence episodes

Not reported.

Psychometric or behavioural measurements

Not reported.

Quality of life measurements

We are uncertain whether there is any difference in quality of life outcomes between PFMT with electromyographic biofeedback and sacral TENS in children with daytime urinary incontinence and nocturnal enuresis due to very low‐certainty evidence (Reis 2014; Table 5). This outcome was not reported for any other of our main comparisons (Table 1; Table 2; Table 3; Table 4; Table 6; Table 7; Table 8; Table 9).

Adverse events

The included studies did not report any serious adverse events that we considered to be related to study treatments. Most non‐serious adverse events and side effects were mild or moderate in severity and were in children receiving pharmaceutical interventions, although some were also reported in placebo arms. These included common pharmaceutical side effects such as nausea, abdominal pain, dry mouth, drowsiness and headache. One non‐serious adverse event was severe faecal incontinence in one solifenacin‐treated child. In stimulation interventions, there was one unspecified adverse reaction to the electrodes and one child with increased urinary incontinence during the treatment period.

Overall completeness and applicability of evidence

A 2003 systematic review of interventions for children with daytime urinary incontinence highlighted the lack of evidence and little has changed 15 years later (Sureshkumar 2003). Of the 27 included studies, most had small numbers of participants and in many, findings were inconclusive. The included studies reported few of the review's intended outcomes, with the majority only reporting one or two relevant outcomes with inconsistent subjective and objective outcome measurements. Heterogeneity in inclusion criteria, definitions of daytime urinary incontinence, the duration and content of interventions and outcome measurements meant that few data were available for meta‐analyses. Some studies evaluated interventions or elements of interventions that may no longer be widely in use. For these reasons, published clinical study evidence relating to the treatment of functional daytime urinary incontinence in children cannot be said to form a sound or applicable evidence base.

Quality of the evidence

Overall, the quality of the evidence identified by the review was poor. Studies were mostly small, often poorly designed and frequently not well reported. In many studies, especially those assessing behavioural interventions, the interventions were not well described. Some studies combined interventions and compared them in ways that do not facilitate the assessment of the effectiveness of individual interventions or components thereof. GRADE evaluation of certainty of evidence determined that all pooled analyses were of low‐ or very low‐certainty, commonly because of imprecision due to small sample size and wide confidence intervals, inconsistency due to heterogeneity in study populations, interventions and effect estimates, and considerable risk of bias.

Potential biases in the review process

While we made all efforts to identify relevant published and unpublished data through electronic database and registry searches, and contacting study authors, there remains a possibility that the review process has missed relevant studies. By searching clinical trials registry entries alongside published database searches, we have attempted to minimise the risk of missed studies, though relevant data may remain unregistered or unpublished, especially from studies conducted before registration became a widespread research requirement.

Agreements and disagreements with other studies or reviews

The results of this systematic review are in agreement with previous reviews of interventions for children with daytime urinary incontinence, which also reported sparseness of evidence, inconclusive results, heterogeneity in populations, interventions and outcome measurement and the need for well‐designed clinical studies. The reviews considered all treatments for daytime urinary incontinence (Sureshkumar 2003), electrical neurostimulation for lower urinary tract dysfunctions including incontinence, neurostimulation for non‐neurogenic overactive bladder and associated incontinence (Fernandez 2017), and biofeedback in the management of children with non‐neuropathic voiding disorders and associated incontinence (Fazeli 2015).

Authors' conclusions

Implications for practice.

The studies identified for the review were generally very small, and variation in the design and delivery of interventions and in the outcome measured means that they offer little evidence that can inform practice. A broad range of physical, stimulation, feedback and behavioural interventions are currently considered for the treatment of these children and, while it seems some of the interventions may help some children, the evidence base offers little help in determining which, if any, may be most appropriate for which child.

Although TENS may be superior to placebo in children with overactive bladder and urinary incontinence, and urotherapy with use of a timer watch may achieve better results than urotherapy alone, we consider the evidence to be of low‐ and very low‐certainty respectively.

Many of the interventions considered may involve considerable investment in healthcare equipment and human resources, as well as commitment from the child and family. It is important that such factors be taken into account in treatment choices, especially where there is uncertainty over the effectiveness of interventions. In this context, treatment choices should be informed not only by diligent history taking and timely comprehensive assessment and review in order to accurately diagnose the child’s bladder dysfunction but also by some assessment of children's and carers' opportunities, capacities and commitment, and family and other supports available to the child. In lower‐resource settings, it may be of interest that urotherapy alone may be equivalent to urotherapy with more intensive voiding education and technology‐dependent feedback, although the certainty of the evidence is low (Kajbafzadeh 2011; Klijn 2006; Van Gool 2014; Analysis 3.1).

It is regrettable that the review found little that can guide evidence‐based practice in the treatment of a stressful and burdensome health problem. Children affected, their parents and the clinicians working with them are faced with very little evidence to help them to make treatment decisions. In this scenario, the clinical experience of individual clinicians and the support of parents may be the most valuable resources.

Implications for research.

Despite the prevalence and potential impact of daytime urinary incontinence in children, it is a clinical area that has been largely neglected by health research. Given that many of the interventions currently considered for treating the condition can involve considerable investment of time, commitment and resources from healthcare providers, children and carers, research is urgently needed to identify which are the most effective.

Larger studies are needed with sufficient power, well‐defined interventions, representative populations and consistent outcome measurements to assess the effectiveness of interventions. The development and use of a core outcome set for research in childhood continence would be of benefit. Care should be taken to ensure that interventions are chosen, defined and compared in ways that facilitate the assessment of individual interventions or their components. Future studies should evaluate children's and carers' views of interventions to consider whether benefits outweigh harms or inconvenience and their effect on quality of life, socioeconomic and psychosocial outcomes. Ultimately, an evidence‐based clinical decision support tool would be helpful in an clinical area in which the range of interventions is so varied.

Although the reliability of the evidence was judged to be of low certainty, TENS in children with overactive bladder and daytime urinary incontinence and PFMT in the context of dysfunctional voiding seem to warrant further investigation. The use of timers on watches or mobile phones to remind children of a voiding schedule seems to be a simple intervention with promise.

Appendices

Appendix 1. Cochrane Incontinence Specialised Register search terms

The terms that will be used to search the Cochrane Incontinence Specialised Register are given below:

(({DESIGN.CCT*} OR {DESIGN.RCT*})

AND

({TOPIC.URINE.INCON.diurnal.} OR {TOPIC.URINE.ENURESIS.daytimewetting.} OR {TOPIC.URINE.OVERACTIVEbladder.children.} OR {topic.urine.incon.urge.children.} OR {TOPIC.URINE.NEUROGENIC.children.} OR {TOPIC.URINE.INCON.children*})

(All searches will be of the keyword field of Reference Manager 2012).

Data and analyses

Comparison 1. Transcutaneous electrical nerve stimulation (TENS) versus sham (placebo) TENS.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Resolution of daytime urinary incontinence at treatment completion	3	93	Risk Ratio (M‐H, Random, 95% CI)	4.89 [1.68, 14.21]

Comparison 2. Pelvic floor muscle training and urotherapy versus urotherapy alone.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Resolution of daytime urinary incontinence at 12 months	3	91	Risk Ratio (M‐H, Random, 95% CI)	2.36 [0.65, 8.53]

Comparison 3. Voiding education with feedback alongside standard urotherapy versus standard urotherapy alone.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Resolution of daytime urinary incontinence at 12 months	3	151	Risk Ratio (M‐H, Random, 95% CI)	1.13 [0.87, 1.45]

Comparison 4. Pelvic floor muscle training versus anticholinergics.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Resolution of daytime urinary incontinence	2	86	Risk Ratio (M‐H, Random, 95% CI)	1.92 [1.17, 3.15]

Comparison 5. Transcutaneous electrical nerve stimulation (TENS) versus anticholinergics.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Resolution of daytime urinary incontinence	2	72	Risk Ratio (M‐H, Random, 95% CI)	0.81 [0.05, 12.50]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Ayan 2007.

Methods	Study design: 3‐arm RCT
Participants	Setting: not reported Country: Turkey Age: 4‐12 years of age Number randomised: 72 Inclusion criteria: voiding dysfunction; incontinence, frequency, urgency, obstructive symptoms with or without recurrent nonfebrile UTI Exclusion criteria: anatomical or neurogenic disease – lesion or abnormal voiding pattern (no further detail provided), abnormality of renal ultrasound, large PVR, febrile UTI, failed therapy
Interventions	Group 1 (n = 20): behavioural modification (3 months; timed voiding, double voiding, relaxation of PF during voiding) only; 13 boys, 7 girls; mean age 8.65 years; mean DVSS 9.6 Group 2 (n = 32): behavioural modification and tolterodine (1 mg twice daily); 15 boys, 16 girls; 1 dropout due to oedema within 3 days of medication not included in analyses; mean age 9 years; mean DVSS 9.5 Group III (n = 20): behavioural modification and placebo (0.5 mg glucose in 20 mL water once daily); 8 boys, 12 girls; mean age 8.25 years; mean DVSS 10.7
Outcomes	DVSS scores at 1 month and 3 months' treatment Scores from subset DVSS questions 1, 2, 6, 7 relating to instability at 1 month of treatment Adverse events
Notes	Source of funding: none declared
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No information on sequence generation
Allocation concealment (selection bias)	Unclear risk	No information provided about allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No information on who was blinded or how
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No information on blinding of outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	No reason to suspect incomplete reporting of outcomes
Selective reporting (reporting bias)	Low risk	No reason to suspect selective reporting
Other bias	Low risk	No reason to suspect any other source of bias

Borch 2017.

Methods	Study design: 3‐arm RCT
Participants	Setting: outpatient clinics at 2 sites Country: Denmark Ages: 5‐14 years Number randomised: 66 Inclusion criteria: urge incontinence > 2 times/week; > 4 voids/day; UI not responded to > 2 months of standard urotherapy Exclusion criteria: neurogenic bladder; GU anomalies or GI tract; previous surgery to kidney or bladder; ongoing or previous treatment with anticholinergics; PVR > 20 mL on 2 ultrasounds; constipation as determined by Rome III; recurrent UTI
Interventions	Group 1 (n = 22): active TENS (sacral S2‐S3 TENS 2 h continuously daily) and oxybutynin (5 mg twice daily); 11 boys, 7 girls completed study; mean age 7.65 years Group 2 (n = 21): active TENS and placebo; 19 boys, 0 girls completed study; mean age 7.47 years Group 3 (n = 23): oxybutynin and sham TENS; 10 boys, 4 girls completed study; mean age 7.48 years
Outcomes	Dry after 10 weeks of treatment Response after 10 weeks of treatment ‐ defined according to ICCS guidelines as "no response" (< 50% symptom reduction), "response" (50%‐99% reduction) and "complete response" (dry, 100% reduction)
Notes	Source of funding: monetary or material support for the clinical study provided by Aarhus University Hospital
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Sequence generated by computer
Allocation concealment (selection bias)	Unclear risk	No information on allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants and investigators blinded
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcomes self‐reported by blinded participants/parents
Incomplete outcome data (attrition bias) All outcomes	High risk	High rate of attrition: 15/66 did not complete the study, with uneven numbers lost between groups
Selective reporting (reporting bias)	High risk	Some outcomes are not reported fully, with data only presented graphically or for significant results only (for example, significant mean differences for number of wet days/week, severity of UI and frequency only reported for comparison of active TENS and oxybutynin vs TENS and placebo).
Other bias	Low risk	No reason to suspect any other bias

Brownrigg 2015.

Methods	Study design: individually RCT
Participants	Setting: outpatient urology clinics at tertiary university hospital Country: Canada Ages: 6‐10 years Number randomised: 60 Inclusion criteria: non‐neurogenic LUT dysfunction and dysfunctional elimination syndrome (BBD) Exclusion criteria: high‐grade hydronephrosis, VUR, learning disabilities
Interventions	Group 1 (n = 30): individual urotherapy (1 x 15‐min individual urotherapy session). 23 girls, 2 boys completed; 5 lost to follow‐up. Mean age 7.3. Daytime UI 17 (68%); night‐time NE 13 (52%); recurrent UTI 15 (60%); voiding problems 9 (36%) Group 2 (n = 30): group urotherapy (1 x 1‐h group urotherapy session); 20 girls, 4 boys completed; 4 did not attend, 2 lost to follow‐up. Mean age 7.3. Daytime UI 11 (46%); night‐time NE 11 (46%); recurrent UTI 9 (38%); voiding problems 12 (50%). Urotherapy curriculum (common to both groups): anatomy and physiology of the urinary system; voiding education. Progressive relaxation; psychosocial aspects of BBD; checklist and patient handouts used to standardise urotherapy content
Outcomes	Mean (SD) BBD symptom score (Vancouver questionnaire) Mean (SD) change in BBD symptom score Mean (SD) QoL score (PinQ) Follow‐up data collected at a median of 14 weeks (range 9‐35) after intervention
Notes	Source of funding: none declared BBD symptoms such as incontinence, urgency, frequency, dysuria, hesitancy, straining and constipation were measured using the Vancouver Questionnaire. This questionnaire is a 14‐item, 5‐point Likert scale instrument shown to be valid and reliable for children 4‐16 years old (median 8) with BBD (Afshar 2009). PinQ was used to measure 6 domains of QoL i.e. social, self‐esteem, family, body image, independence and mental health. The tool has shown high reliability under test‐retest conditions in children ≥ 6 years (Bower 2006).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Sequence from tables of random number generated independently
Allocation concealment (selection bias)	Low risk	Quote: "Randomization sequence was generated by an independent biostatistician. A third party research associate, not involved in the trial, was responsible for randomizing each patient and reporting allocation to the trial research staff after eligibility was confirmed and informed consent obtained." Comment: allocation concealed until after study entry
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding of participants was not possible due to differences between interventions. Study personnel were not blinded.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: “Outcome assessors were blinded to treatment allocation. The results of the questionnaires were not included in the medical record and remained unknown to the clinician(s)/researcher(s).” Comment: outcome assessors blinded to allocation
Incomplete outcome data (attrition bias) All outcomes	Low risk	Data reported for all who completed study. Patient flow and attrition reported
Selective reporting (reporting bias)	Low risk	No reason to suspect selective reporting. Study methods, outcomes and limitations fully reported
Other bias	Low risk	No reason to suspect any other source of bias

Campos 2013.

Methods	Study design: individually RCT
Participants	Setting: Division of Urology, university hospital (single centre) Country: Brazil Number randomised: 47 Ages: 5‐10 years Inclusion criteria: daytime UI, absence of prior treatment for UI Exclusion criteria: neurological disease; anatomical abnormalities; UTI
Interventions	Group 1 (n = 21): oxybutynin (at dose of 0.2 mg/kg, split into 2 x/day, for 3 months) with urotherapy. 13 girls, 8 boys. Median age 9 years. Median (range) daily UI episodes/week 3 (2‐6) Group 2 (n = 26): PFMT (1 x /week with physiotherapist, 2 x/week at home for 3 months) with urotherapy. 16 girls, 10 boys. Median age 8.5. Median (range) daily UI episodes/week 3 (1‐12)
Outcomes	Dry day and night at 4 months (parental self‐report 1 month after completion) Days/month dry day and night at end of each month’s treatment Adverse events/side effects
Notes	Source of funding: none declared Both groups received similar urotherapy: scheduled voiding; advice on fluid intake and appropriate timing and type of drinks; use of 500 mL bottle to facilitate monitoring for fluid intake; voiding posture instruction
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Standard manual sequence generation using opaque and sealed envelopes sequentially numbered
Allocation concealment (selection bias)	Unclear risk	No information on allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding of participants and personnel not possible as interventions were fundamentally different
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No information on blinding of outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Unclear reporting regarding attrition
Selective reporting (reporting bias)	Unclear risk	Data are not reported for outcomes as described in the text. However, the study authors provided clarifications and fuller data when contacted.
Other bias	Low risk	No reason to suspect other sources of bias

de Paula 2017.

Methods	Study design: individually RCT
Participants	Setting: not reported Country: Brazil Number randomised: 16 Ages: 3‐18 years Inclusion criteria: OAB diagnosis, no treatment for 6 months, urinary urge with or without incontinence Exclusion criteria: UTI, abnormal uroflow (not bell), high PVR, LUT dysfunction, using nappies (diapers), neurological disease, anticholinergics or tricyclics
Interventions	Group 1 (n = 8): parasacral TENS with urotherapy. 3 boys, 5 girls, all with daytime UI at baseline. Median age 6.5 years Group 2 (n = 8): placebo (sham) TENS with urotherapy. 3 boys, 5 girls, all with daytime UI at baseline. Median age 8.5 years
Outcomes	Presence of incontinence Presence of urgency Parents’ perception of improvement (OAB and UI): VAS from 0‐10, with 0 meaning “no improvement in voiding” and 10 meaning “complete resolution of symptoms.” Evaluated at baseline, after 20 weekly sessions (over 140 days) and 60 days after completion
Notes	Source of funding: none Both groups received similar urotherapy: advice on voiding habit, voiding before bed and upon waking, voiding schedule every 3‐4 h, no delay in micturition, toilet position and posture, increase drink volumes, no caffeine, avoid drinking 2 h before bed and increased fibre
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Sequence generated by computer programme
Allocation concealment (selection bias)	Unclear risk	No information on allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Patients were blinded but TENS therapist could not be blinded and there is no information on blinding of any other study personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No information on blinding of any study personnel
Incomplete outcome data (attrition bias) All outcomes	High risk	Only 2 participants lost to follow‐up but this represents 12.5%
Selective reporting (reporting bias)	Low risk	All outcomes described in the methods section were reported. No reason to suspect selective reporting
Other bias	Low risk	No reason to suspect any other source of bias

Fielding 1980.

Methods	Study design: individually RCT
Participants	Setting: 2 specialist enuresis clinics Country: UK Number randomised: 75, of whom 30 with daytime UI and nocturnal enuresis: 24 girls, mean age 7 years; 6 boys, mean age 8 years (45 with nocturnal enuresis only) Ages: 5‐15 years Inclusion criteria: daytime UI and NE or NE only (daytime UI defined as wetting after age of 5 years ≥ 1/week and of sufficient severity to concern parents. NE defined as nocturnal enuresis after age of 5 years ≥ 1/week. Wetting was damp underclothes to more obvious voiding causing wetting of top clothes). Exclusion criteria: UTI, organic pathology causative of UI, previous treatment
Interventions	Group 1: (baseline randomised number unclear; n = 24 overall with outcome data; n = 8 with daytime UI). NE alarm for 14 weeks and bladder training (RCT) 1 session/day for 4 weeks: voiding followed by drinking 500 mL fluid, timing of first urge to void, encouragement to postpone, timing and measurement of voiding, aimed to extend postponement duration by 1‐2 min each session, parental encouragement Group 2: (baseline randomised number unclear; n = 25 overall with outcome data; n = 8 with daytime UI). NE alarm for 14 weeks
Outcomes	Mean number of days with UI in 4‐week period
Notes	Source of funding: none declared Only subgroup of children with daytime UI who completed the study considered by the review
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No information on sequence generation
Allocation concealment (selection bias)	Unclear risk	No information on allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding not possible due to inherent difference between treatments
Blinding of outcome assessment (detection bias) All outcomes	High risk	Self‐reported outcome data by unblinded participants/parents
Incomplete outcome data (attrition bias) All outcomes	High risk	Very high attrition rates are poorly explained. Outcome data are not reported clearly
Selective reporting (reporting bias)	Unclear risk	Data collection and results are not reported clearly by current standards. It is not clear whether all outcomes are reported.
Other bias	Low risk	No reason to suspect any other source of bias

Golli 2013.

Methods	Study design: individually RCT
Participants	Setting: Department of Paediatric Nephrology, referred from paediatric clinics Country: Slovenia Number randomised: 64 (24 boys, 40 girls); mean age 7.25 years. LUT conditions (defined by ICCS): 36 OAB and/or urgency UI; 25 voiding postponement; 2 underactive bladder; 1 stress UI (39 also enuresis, 28 also constipation) Ages: 4‐15 years Inclusion criteria: LUT conditions Exclusion criteria: neurological disorders, anatomic anomalies of LUT (except VUR), acute UTI
Interventions	Group 1 (n = 32): voiding re‐education programme. Mean age 6.92 years (29 completed) Group 2 (n = 32): no treatment (delayed programme, with 30 completing programme after 3‐month delay), mean age 7.57 years Voiding re‐education programme (3 months): 5‐day inpatient programme (lead by nephrologist and delivered by trained nurses and a daily psychologist): behavioural modification ‐ fluid intake, regular voiding habits, preventing and treating constipation, PFMT psychologist 1‐1 sessions on positive reinforcement of programme and support of self‐image Outpatient programme: monthly Transfer of inpatient techniques to home setting Bladder diary for record of fluid intake, voiding, bowel habits, UI, urgency, exercises – voiding and bowel postponement, PF strengthening and relaxing Diaries reviewed at outpatient clinic appointments
Outcomes	Completely dry (self‐reported by carer or child) Successful treatment of DV defined as normalisation of DVSS scores (below 6 for girls, 9 for boys) Followed up at 6, 12, 36 and 48 months after completion of programme
Notes	Source of funding: none declared
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No information on sequence generation
Allocation concealment (selection bias)	Unclear risk	No information on allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding of participants and personnel was not possible: 1 group had the intervention while the other was on a waiting list control
Blinding of outcome assessment (detection bias) All outcomes	High risk	Blinding of outcome assessment not possible: self‐report was unblinded; 1 group had the intervention while the other was on a waiting list control
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low level of attrition
Selective reporting (reporting bias)	High risk	Only results favourable to the intervention are reported and with insufficiently detailed data to understand its effect on specific DV symptoms. In terms of the effect of the intervention on daytime UI, reporting is opaque but infers success. There is no indication in the paper of the proportions of children in each of the 2 treatment groups that experienced daytime UI prior to the programme.
Other bias	Low risk	No reason to suspect any other source of bias

Hagstroem 2009.

Methods	Study design: individually RCT
Participants	Setting: outpatient clinics at the Center for Child Incontinence, Aarhus (single centre) Country: Denmark Number randomised: 25 Ages: 5‐14 years Inclusion criteria: daytime UI at least 2 days/week; urgency, refractory to a minimum of 12 months of urotherapy and 3 months of anticholinergics (oxybutynin), normal urinalysis, unremarkable urinary tract ultrasound and normal physical examination. Exclusion criteria: ongoing faecal problems according to Rome III criteria, previous attempted electrostimulation, previous LUT surgery
Interventions	Group 1 (n = 13): sacral TENS (2 h daily for 4 weeks; neurostimulation with digital TENS stimulator, 10 Hz frequency with a 200 sec pulse duration and biphasic waveform with self‐adhesive surface electrodes at the level of the S2‐S3 roots). 7 boys, 6 girls; mean age 8.7 years Group 2 (n = 12): sham sacral TENS (2 h daily for 4 weeks; sham neurostimulation with identical electrodes and digital TENS stimulator modified to produce no stimulation). 3 boys, 9 girls; mean age 8.5 years
Outcomes	1‐week incontinence score Change in 1‐week incontinence score Wet days/week Change in number of wet days/week Daytime incontinence episodes Change in number of UI episodes/day Voiding frequency Change in number of voids/day (Children completed dry pie forms and 48‐h frequency‐volume charts before intervention and during week 4)
Notes	Source of funding: supported by grants from the University of Aarhus Research Foundation and Karen Elise Jensen Foundation Withdrawal from anticholinergics 2 weeks prior to and during study. On study entry, children admitted to dept of paediatrics for 3 days for ambulatory 24‐h natural fill urodynamics and conventional videourodynamics to rule out infravesical obstruction and document detrusor overactivity. During the week preceding hospitalisation, children completed daily records of severity of incontinence using the dry pie tool and 48‐h frequency‐volume charts. On the day of hospitalisation, children completed an urgency VAS before any urodynamic procedure.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was performed by a computer.
Allocation concealment (selection bias)	Low risk	Participants and personnel unable to tell which TENS units were active and which sham. Each stimulator had a number tied to the randomisation computer, which ensured blinding. Half of the stimulators were modified by the manufacturer not to deliver any electrical current for placebo. The exteriors of the active and placebo stimulators were identical.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants and personnel unable to tell which TENS units were active and which sham. Children and parents were informed that they would not be able to determine by sensation whether sham or active TENS was done.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Participants and personnel were blinded to allocation until after outcome assessment.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Data are reported for all outcomes and withdrawals are accounted for.
Selective reporting (reporting bias)	Low risk	There is no reason to suspect selective reporting.
Other bias	Low risk	There is no reason to suspect any other source of bias.

Hagstroem 2010.

Methods	Study design: individually RCT
Participants	Setting: children’s continence clinic at university hospital (single centre) Country: Denmark Number randomised: 58 Ages: 5‐14 years Inclusion criteria: ≥ 1 episode of daytime UI/week, voiding frequency ≥ 6 times/day, OAB (urgency), unremarkable kidney/LUT ultrasonography, normal clinical exam, no indication of bladder underactivity or obstruction on uroflowmetry Exclusion criteria: previous treatment with timer‐assisted urotherapy, history or present use of anticholinergics or alpha blockers
Interventions	Group 1 (n = 28): standard urotherapy (12 weeks). 12 girls, 16 boys; mean age 7.65 years. Primary UI 23; daytime UI alone 4; daytime UI + NE 24; history of FI 6; mean UI episodes/week pretreatment 8.0. Group 2 (n = 30): standard urotherapy and timer watch (with 7 alarms set to remind of voiding schedule). 9 girls, 21 boys; mean age 7.48 years. Primary UI 26; daytime UI alone 4; daytime UI + NE 26; history of FI 10; mean UI episodes/week pretreatment 9.9. Standard urotherapy: advice on daily fluid intake timed voiding at 2‐h intervals until bedtime 48‐h bladder diaries weeks 1, 11 + 12 (fluid intake, timing, voiding frequency, UI)
Outcomes	Change in mean dry days/week at 12 weeks Median wet days per week Response to treatment at 12 weeks by ICCS standards (0%‐49% reduction in wet days = no response, 50%‐89% = partial response, ≥ 90% = response) Up to 12 months after end of treatment period
Notes	Source of funding: supported by grants from University of Aarhus Research Foundation, Arberg Time and Karen Elise Jensen Foundation
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No information on sequence generation
Allocation concealment (selection bias)	Unclear risk	No information on allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not possible to blind participants as the interventions were different. No information on blinding of personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No information on blinding of outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	Data are reported fully for all outcomes.
Selective reporting (reporting bias)	Low risk	No reason to suspect selective reporting
Other bias	Unclear risk	57/58 participants had previously tried urotherapy without success. Thus the lack of success in the urotherapy‐alone group in this study may not be surprising. In particular, scheduled voiding as advised in urotherapy was not adhered to without the reminder provided by the alarm watch; in the urotherapy‐alone group, only 9/28 were assessed to be compliant with timed voiding.

Halliday 1987.

Methods	Study design: individually RCT
Participants	Setting: paediatric outpatient clinic (single centre) Country: UK Number randomised: 44 (35 girls, 9 boys); mean age 8.5; range 5‐13. Daytime UI present continuously from birth 34; nocturnal enuresis 20; FI 5 Ages: 5‐15 years Inclusion criteria: daytime UI sufficiently problematic to be principal complaint of carer at consultation (for most underwear sufficiently wet to soak outer clothing) Exclusion criteria: severe LUT abnormality, severe social or psychological problems
Interventions	Group 1 (n = 22): wetting alarm: buzzer alarm activated when a pad worn inside underwear became wet (may help child to recognise when bladder was full). Up to 3 months of treatment with alarm (ceased when dry for 6 weeks, considered success). 2 dropouts during week 1 Group 2 (n = 22): random alarm: superficially identical alarm with pad worn inside underwear, but alarm sounded randomly approximately every 2 h whether or not the sensor was wet (this may remind child to use the toilet and establish regular voiding). Up to 3 months of treatment with alarm (ceased when dry for 6 weeks, considered success). 3 dropouts during week 1
Outcomes	Success rate (dry for 6 weeks) at 3 months
Notes	Source of funding: The Mental Health Foundation
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No information on sequence generation
Allocation concealment (selection bias)	Low risk	Allocation was performed by a person not involved in the study subsequently and arrangements seem adequate to prevent personnel from having knowledge of treatment allocation in advance of, or during, enrolment.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Personnel blinded. Parents and children could not be blinded to allocation but had the potential benefits of both alarm types explained, such that lay bias was unlikely.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No specific information on blinding of outcome assessment but inferred by “the psychologist and paediatrician remained in ignorance of the type of alarm allocated.”
Incomplete outcome data (attrition bias) All outcomes	Low risk	Relatively low level of attrition
Selective reporting (reporting bias)	Low risk	The focus of the report is on psychological disturbance amongst children and parents. UI outcomes are reported sufficiently to provide background for these considerations and sufficiently to understand the shorter‐term impact of the treatments between the two groups.
Other bias	Low risk	No reason to suspect any other source of bias

Hellstrom 1989.

Methods	Study design: multi‐arm RCT with cross‐over element
Participants	Setting: outpatient department, Departments of Paediatric Surgery and Paediatrics, University of Gothenburg Country: Sweden Number randomised: 61 randomised (3 withdrawn during study: 2 with newly detected unco‐ordinated micturition, 1 with newly detected neurogenic bladder disturbance). 58 included in analyses: 22 boys (mean age 9.4), 36 girls (mean age 8.6). 55/58 had history of daytime UI (50 primary, 5 secondary with dry periods > 6 months). At baseline 46/58 daytime UI at least daily, 8/58 once or several times/week, 1 once or twice/month; 3/58 had urgency or frequency without UI Ages: 6‐14 years Inclusion criteria: urgency with/without urge incontinence, instability of the bladder defined as detrusor contractions during filling or premature micturition at small bladder volume Exclusion criteria: neurogenic bladder, urinary tract obstruction, unco‐ordinated micturition according to history or flow rate and EMG sphincter measurements, bacteriuria in last 3 months, constipation, encopresis, other drug treatments that effect bladder
Interventions	Group 1 (n = 26): placebo with cross‐over to terodiline: placebo (4 weeks), terodiline (12.5 mg, 1/day for 1 week, 2/day for 3 weeks) Group 2 (n = 18): terodiline with cross‐over to placebo: terodiline (12.5 mg, 1/day for 1 week, 2/day for 3 weeks), placebo (4 weeks) Group 3 (n = 14): terodiline followed by terodiline: terodiline (12.5 mg, 1/day for 1 week, 2/day for 3 weeks), repeated On study completion, the children were grouped in to 2 groups (terodiline and placebo) for analysis by time period. In all groups information and training was provided on: how to gain bladder control, how to initiate voiding, void 4‐6 x/day, recording of time of each void and incontinence episodes, conduct of 12‐h pad test (hospital measure volume voided)
Outcomes	Daytime UI episodes/day Mean change in UI episodes/day ‐ baseline to 4 weeks (1st medication period, before cross‐over) Number of daytime micturitions/day Mean change in daytime micturitions/day ‐ baseline to 4 weeks (1st medication period, before cross‐over) Mean change in dry days/week ‐ baseline to 4 weeks (1st medication period, before cross‐over) Subjectively reported improvement in overall bladder symptoms Adverse events
Notes	Source of funding: KabiVitrum AB, Stockholm; First of May Flower Annual Campaign for Children's Health All groups had a 2‐week run‐in period with no study medication to gather baseline data and familiarise with study protocols
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Standard random number table used
Allocation concealment (selection bias)	Low risk	Allocation concealment adequate, using off‐site arrangements
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants and study personnel blinded by similar terodiline and placebo products, remote allocation and concealment, similar medication schedules in all treatment groups and periods. More participant‐reported side effects on placebo treatment than drug treatment
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Double‐blind study with allocation remotely done by manufacturers' preparation and allocation of active treatment and placebo according to a random‐number table, which was revealed at the end of the study.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All recruited and completed participants are accounted for in data. Only 3/61 withdrawn, with reasons provided
Selective reporting (reporting bias)	Unclear risk	Little reason to suspect selective reporting. All outcomes reported for first treatment period before cross‐over from placebo to terodiline treatment. However, outcomes for second treatment period reported only narratively. Acknowledgement by study authors that data from second period were not reported fully due to long half‐life of drug and no washout period at cross‐over
Other bias	Low risk	No reason to suspect any other source of bias

Kajbafzadeh 2011.

Methods	Study design: individually RCT
Participants	Setting: Pediatric Urology Research Center and Department of Physical Therapy, Children’s Hospital Medical Center Country: Iran Number randomised: 80 Ages: > 5 years Inclusion criteria: voiding dysfunction with either constipation or FI (voiding dysfunction defined as habitual contraction of urethral sphincter during voiding phase as shown in 2 uroflow EMGs, showing staccato patterns and increased EMG activity while voiding or constant EMG with plateau pattern); normal VCUG. Exclusion criteria: physical examination, KUB or history suggestive of neurological defect. (Excluded and MRI requested. Readmission to study possible if normal.)
Interventions	Group 1 (n = 40): PFMT/biofeedback + behavioural modification. Physiotherapist‐supervised PFMT with feedback presented graphically to children via animated computer game displays; 2 sessions/week – minimum of 6 sessions, maximum of 12; EMG beginning and end of each session; sessions stopped once EMG normal; mean number of sessions was 9.6 (range 6‐12); PFMT at home for at least 15 min twice daily. 8 boys, 32 girls; mean age 8.5 years (5‐16 years). 10/40 (25%) daytime UI; 11/40 (27.5%) nocturnal enuresis; 17/40 (42.5%) urgency; 25/40 (62.5%) constipation Group 2 (n = 40): behavioural modification. Monthly visits encouraged for 6 months to reinforce behavioural modification education and compliance. 10 boys, 30 girls; mean age 9 years (6‐16 years). 11/40 (27.5%) daytime UI; 9/40 (22.5%) nocturnal enuresis; 16/40 (40%) urgency; 20/40 (50%) constipation
Outcomes	Daytime UI resolved NE resolved Urgency resolved Staccato resolved No longer flat void pattern on EMG EMG activity during voiding resolved
Notes	Source of funding: none declared Behavioural modification education for both groups: scheduled voiding, hydration and high‐fibre diet. Patient/guardian education about bladder function and GI tract
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Standard random number chart used
Allocation concealment (selection bias)	Unclear risk	No information on allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding not possible as 1 group received PFMT with feedback, whereas the other did not
Blinding of outcome assessment (detection bias) All outcomes	High risk	Symptom outcomes were self‐reported by unblinded participants and guardians. No information on blinding of outcome assessors for uroflowmetry outcomes
Incomplete outcome data (attrition bias) All outcomes	Low risk	Complete follow‐up data
Selective reporting (reporting bias)	Low risk	No reason to suspect selective reporting. All outcomes reported
Other bias	Low risk	No reason to suspect any other source of bias

Klijn 2006.

Methods	Study design: multi‐arm RCT
Participants	Setting: paediatric urology unit, university hospital (single centre) Country: Netherlands Number randomised: 143 randomised; 130 completed study Ages: 6‐16 years Inclusion criteria: clear peaks and declines in uroflowmetry, PVR ≥ 10%, history of recurrent UTI with or without daytime UI, sufficient cognitive capacity to read clock and to understand urinary problems Exclusion criteria: not reported
Interventions	Group 1 (n = 44): standard urotherapy alone (weeks 1‐24). All girls. Mean age 8.3 years. Daytime UI 30/44 (68%); constipation 24/44 (55%); symptoms > 1 year 27/44 (61%); reflux 9/44 (20%) Group 2 (n = 46): standard urotherapy with video. Daily viewing of videotaped urotherapy instructions presented by popular children’s TV host, with additional personalised section recorded during first visit (weeks 1‐8) followed by standard urotherapy (weeks 9‐24). All girls. Mean age 8.3 years. Daytime UI 32/46 (70%); constipation 29/46 (63%); symptoms > 1 year 31/46 (67%); reflux 13/46 (28%) Group 3 (n = 53): standard urotherapy plus home uroflow biofeedback. Uroflow units with palmtop displays were developed that could be used in the home and children instructed on their use and what shape of curve to aim for. Used at least 4 times daily (weeks 1‐8) followed by standard urotherapy (weeks 9‐24). 50 girls, 3 boys. Mean age 8.2 years. Daytime UI 33/53 (62%); constipation 31/53 (58%); symptoms > 1 year 38/53 (72%); reflux 12/53 (23%) Standard urotherapy/behavioural modification (outpatient visits at baseline, 4 and 8 weeks with telephone support between visits; weeks 9‐24 reduced visits and telephone support): instruction on bladder and sphincter function, advice on voiding scheduling and posture, voiding charts, advice on defecation scheduling with or without laxatives, prophylactic antibiotics given
Outcomes	Cure of daytime UI at 12 months (28 weeks after treatment period) Adverse events
Notes	Source of funding: ZonMW (a national Dutch institute supporting preventive medical research)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated sequence
Allocation concealment (selection bias)	Low risk	Allocation concealment specifically described
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding of participants and personnel not possible due to different natures of interventions
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No information on blinding of outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	Outcome data are reported for both ITT and PP analyses. Low attrition, with all losses and withdrawals accounted for
Selective reporting (reporting bias)	Low risk	All outcomes are reported and discussed. Results are largely not significant. No reason to suspect selective reporting
Other bias	Low risk	No reason to suspect any other source of bias

Kroll 2006.

Methods	Study design: individually RCT
Participants	Setting: not reported Country: Poland Number randomised: 60 (57 completed) Ages: 5‐17 years Inclusion criteria: detrusor‐sphincter discoordination (Dx by uroflowmetry), history of UTIs, irregular micturitions with or without daytime UI in bladder diaries, normal upper urinary tract Exclusion criteria: not reported
Interventions	Group 1 (n = 30): doxazosin. (6 weeks. Selective alpha1‐blocker in age‐related dosage: patients < 20 kg body weight 0.5‐1 mg/day; > 20‐40 kg body weight 1‐2 mg/day; > 20 kg body weight 2 mg/day in 1 evening dose. Daytime UI 19/30 Group 2 (n = 30): behavioural modification (6 weeks. Timed voiding, maintaining a voiding diary). Daytime UI 20/30
Outcomes	Cure of daytime UI at 6 weeks Adverse effects
Notes	Source of funding: none declared Both groups had a 2‐week run‐in period with bladder diary and were provided instruction on voiding posture and PF relaxation. Outcome data for children affected by daytime UI provided by the study authors
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No information on sequence generation
Allocation concealment (selection bias)	Unclear risk	No information on allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding not possible due to differences in the interventions
Blinding of outcome assessment (detection bias) All outcomes	High risk	Outcomes self‐reported by unblinded participants and parents
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low attrition. Data for all expected outcomes
Selective reporting (reporting bias)	Low risk	Data reported for outcomes described in the methods
Other bias	Low risk	No reason to suspect any other source of bias

Ladi Seyedian 2014.

Methods	Study design: individually RCT
Participants	Setting: Pediatric Urology Research Center and Department of Pediatric Urology (single centre) Country: Iran Number randomised: 60 Ages: ≥ 5 years old Inclusion criteria: LUT symptoms/DV (urgency, frequency, daytime UI, enuresis or recurrent episodes of febrile UTI with no response to medical treatment); abnormal flow curve shape (staccato) and positive EMG activity via perineal electrodes during voiding Exclusion criteria: anatomical defects; neurological or mental disorders
Interventions	Group 1 (n = 30): PFMT with behavioural urotherapy. 12 sessions (2 x 30‐min sessions per week for 6 weeks) of PFMT exercises (with and without Swiss ball) alongside behavioural therapy (12 months). 25 girls, 5 boys. Median age 7 years (5‐14). Daytime UI 17, urgency 14 Group 2 (n = 30): behavioural urotherapy alone (12 months). 22 girls, 8 boys. Median age 9 years (5‐14). Daytime UI 11, urgency 11 Behavioural urotherapy: parents and children in both groups were trained about the urinary and gastrointestinal tract function: normal drinking habit, scheduled voiding, the normal mechanism of toilet training, and the effects of high‐fibre diet. Bimonthly visits during 12‐month study to support training programme and enhance compliance
Outcomes	Daytime UI cured at 1 year Voided volume
Notes	Source of funding: none declared Anticholinergic medications were discontinued a week prior to study commencement. Prophylactic antibiotics were continued in participants with recurrent UTIs or in the presence of VUR. Constipation was managed conservatively, by a high‐fibre diet and mild laxative.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No information on sequence generation
Allocation concealment (selection bias)	Unclear risk	No information on allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding not possible due to difference between the interventions received by participants/delivered by personnel
Blinding of outcome assessment (detection bias) All outcomes	High risk	Principal outcomes self‐reported by unblinded participants/parents. No information on blinding of objective outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	Data are complete for all outcomes. Low attrition rate (1 participant)
Selective reporting (reporting bias)	Low risk	All expected outcomes per methods section reported
Other bias	Low risk	No reason to suspect any other source of bias

Lordelo 2010.

Methods	Study design: individually RCT
Participants	Setting: section of paediatric urology, Bahiana School of Medicine, Salvador (single centre) Country: Brazil Number randomised: 37 Ages: > 4 years Inclusion criteria: OAB defined as the presence of characteristic symptoms of urgency (with or without daytime incontinence), no elevated post‐void residual urine and bell‐shaped curve on uroflowmetry Exclusion criteria: LUT symptoms secondary to anatomical anomaly such as posterior urethral valves, ureterocele or ectopic ureter, neurogenic bladder, non‐resident status (children who lived outside the city where the study took place), inability to comply with treatment requirements
Interventions	Group 1 (n = 21): parasacral TENS (20 x 20‐min sessions, 3 x/week over 7 weeks). 8 boys, 13 girls; mean age 7.5 years. Daytime UI 16/21 (76.2%); holding manoeuvres 15/21 (71.4%); nocturnal enuresis 16/21 (76.2%); history of UTI 13 (61.9%); pyelonephritis 8/21 (38.1%) Group 2 (n = 16): sham TENS (20 x 20‐min sessions, 3 x/week over 7 weeks). 4 boys, 12 girls; mean age 7.4 years. Daytime UI 14/16 (87.5%); holding manoeuvres 12/16 (75%); nocturnal enuresis 13/16 (81.3%); history of UTI 11/16 (68.8%); pyelonephritis 7/16 (43.8%) TENS stimulation: frequency was 10 Hz with a generated pulse of 700 s. Current intensity was increased to maximum level tolerated. 2 superficial 3.5 cm electrodes placed on each side of S3 and S2 and electrodes also placed in the scapular area 3 cm above the inferior scapular edge. In the treatment group, only the sacral electrodes were activated while in the sham group the stimulation was at the scapular electrodes.
Outcomes	Cure: parental report of complete resolution of symptoms after 7 weeks of treatment Improvement: parental report of significant improvement of symptoms after 7 weeks of treatment
Notes	Source of funding: supported by grants from Fundação de Amparo a Pesquisa do Estado da Bahia (from the government of Bahia State, Brazil) Both groups also underwent urotherapy reinforced by a booklet: voiding before sleeping, increasing volume of liquid ingested daily, eating foods rich in fibre, refraining from postponement of voiding when experiencing symptoms of urgency, postural advice on micturition and toilet seat adapters and foot supports as needed
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Sealed, opaque envelopes suggests true randomisation
Allocation concealment (selection bias)	Unclear risk	No statement on concealment of allocation
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Study described as single‐blinded. Participants blinded. No information on blinding of treating personnel
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Self‐reported outcomes by participants blinded to allocation. Personnel‐assessed outcomes also blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	Withdrawals (2/37) are reported with reasons. Other than outcome data for these, all data are reported.
Selective reporting (reporting bias)	Low risk	All expected outcomes are reported.
Other bias	Low risk	No reason to suspect any other source of bias

Nemett 2008.

Methods	Study design: individually RCT
Participants	Setting: new referrals to an interdisciplinary paediatric urology clinic/physical therapy centre (single‐centre study in 2 clinics) Country: USA Number randomised: 32 (11 did not complete: 8 non‐attendance, 1 neurological exclusion after enrolment, 1 ureteral surgery scheduled, 1 erroneously recruited, no VUR or UI) Ages: 4‐11 years Inclusion criteria: DV with daytime UI and/or VUR, treated by paediatric urologist ≥ 6 months prior to inclusion as appropriate without resolution of symptoms, symptoms of stable severity ≥ 6 months prior to inclusion Exclusion criteria: neurological, spinal or urogenital structural anomalies, known history of child abuse, girls progressed to Tanner stage IV pubertal development
Interventions	Group 1 (n = 10 completed): osteopathy plus standard care. 4 x 1‐h individually customised manual physical therapy regimen using osteopathy approach over 10‐12 weeks. Included mobilisation of bodily tissues to relieve movement restrictions, achieve balanced alignment and mobility and postural symmetry. Visits scheduled to coincide with 4 fortnightly urology visits. 6 girls, 4 boys; mean age 6.5 years. Daytime UI 6/10 Group 2 (n = 11 completed): standard care alone (urology visits) over 10‐12 weeks. 8 girls, 3 boys; mean age 7.1 years. Daytime UI 8/11 Standard care: 4 x 1‐h urology appointments at 2‐week intervals, providing (as indicated) medication, behavioural modification (advice on timed voiding and evacuation schedules, dietary modification), PFMT, PFMT with feedback, treatment of constipation
Outcomes	Carer and child reporting of “improvement” in daytime UI at 3 months (end of treatment) The proportion of outcomes in which there was improvement
Notes	Source of funding: supported in part by grants from the National Kidney Foundation of Maryland
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No information on sequence generation
Allocation concealment (selection bias)	Unclear risk	No information on allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	High risk	As 1 group received the osteopathy intervention and the others did not, blinding was not possible.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Carer and child self‐reporting was unblinded. No information provided on blinding of assessment of other objective parameters
Incomplete outcome data (attrition bias) All outcomes	High risk	High (34%) attrition. Data reported fully for the rest
Selective reporting (reporting bias)	High risk	No prespecified primary outcome reported. Results reported in the text focus on the proportion of all outcomes (VUR, UI, PVR, UTI, dissynergic voiding) in which there were undefined improvements and other outcomes that favour the intervention.
Other bias	Unclear risk	Children in both groups are also attending unregulated standard care in which they can receive treatments including medications on the advice of paediatric urologists.

Newgreen 2017.

Methods	Study design: multi‐arm RCT
Participants	Country: 16 countries worldwide Number randomised: 189: 148 children (2 did not commence study), 41 adolescents Ages: children 5‐< 12 years; adolescent 12‐< 18 years Inclusion criteria: ICCS diagnosis of OAB, ≥ 4 episodes of daytime incontinence during a 7‐day prebaseline diary period, height and weight within normal percentiles for age according to the Centers for Disease Control and Prevention growth charts Exclusion criteria: daily voiding frequency < 5, extraordinary daytime urinary frequency (voiding frequency ≥ 1/h with a mean voided volume < 50% of EBC and normal nocturnal bladder behaviour for age), LUT pathologies other than OAB, any condition or treatment that could cause urinary symptoms or interfere with assessment of efficacy parameters (e.g. monosymptomatic enuresis, polyuria (> 75 mL/kg/body weight/24 h), central or nephrogenic diabetes insipidus, DV, congenital abnormalities affecting LUT function, constipation (Rome III criteria), chronic UTI or > 3 UTIs 2 months prior to screening, catheterisation within 2 weeks prior to screening, haematuria ≥++ on dipstick test unless urological or kidney disease ruled out, kidney or bladder stones, partial/complete gastrointestinal obstruction, decreased gastrointestinal motility (e.g. paralytic ileus), risk of gastric retention, mean voided volume (excluding morning volume) > EBC for age ((age + 1) x 30) mL or > 390 m; PVR volume > 20 mL
Interventions	Group 1: solifenacin (once‐daily solifenacin oral suspension) with urotherapy (children) 65/73 completed. 29 boys, 44 girls; mean age 7.6 years. Baseline mean UI episodes/day = 2.5 Group 2: placebo with urotherapy (children) 66/75 completed. 38 boys, 35 girls; mean age 7.4 years. Baseline mean UI episodes/day = 3.0 Group 3: solifenacin (once‐daily solifenacin oral suspension) with urotherapy (adolescents) 17/22 completed. 5 boys, 17 girls; mean age 14.2 years. Baseline mean UI episodes/day = 1.8 Group 4: placebo with urotherapy (adolescents) 16/19 completed. 3 boys, 16 girls; mean age 14.4 years. Baseline mean UI episodes/day = 2.8 Urotherapy included: information and demystification, with explanation about normal LUT function and how the patient deviates from this; instruction about what can be done, including regular voiding habits, voiding posture, avoiding holding manoeuvres; lifestyle advice concerning fluid intake and preventing constipation; use of bladder diaries to document symptoms and voiding habits; support and encouragement by the caregiver. Timer watch provided to promote 7 voids in 24 h
Outcomes	Mean (SE) daytime UI episodes per 24 h in children 5‐12 years Mean (SE) night‐time UI episodes per 24 h in children 5‐12 years Mean (SE) number dry days/week in children 5‐12 years Mean (SE) number dry nights/week in children 5‐12 years Adjusted mean change in number of incontinence episodes per 24 h Adjusted mean change in voided volume from start to end of treatment (mL) Adjusted mean change in mean micturitions per 24 h from start to end of treatment Adverse events
Notes	Source of funding: Astellas Pharma
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No information on sequence generation
Allocation concealment (selection bias)	Low risk	Allocation was done remotely off‐site
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Run‐in (treatment washout) period was participant‐blinded in that the investigators knew no active drug was involved at that stage. Investigators were additionally blinded in the 'double‐blinded; phase.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Unclear who collected data and whether blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data reported, including losses to follow‐up and insufficiency of adolescent participants for analyses
Selective reporting (reporting bias)	Low risk	Data are reported for outcomes defined in methods. No reason to suspect selective reporting
Other bias	Unclear risk	May have been variation across 16 countries in delivery of urotherapy: country was identified as a significant random effect. Astellas Pharma funded the study.

Patidar 2015.

Methods	Study design: individually RCT
Participants	Setting: single‐centre hospital study, Lucknow Country: India Number randomised: 40 (24 boys, 16 girls) Ages: age criteria not reported (mean age 7.71 in intervention group, 8.38 years in the control group) Inclusion criteria: non‐neurogenic OAB unresponsive to behavioral therapy and at least 6 months of anticholinergic medication, anticholinergic drugs discontinued 2 weeks before the commencement of therapy Exclusion criteria: neurogenic bladder, DV, defecation disorders, LUT surgery, UTI, LUT symptoms secondary to anatomical anomalies such as posterior urethral valves, ureterocele, or ectopic ureter
Interventions	Group 1 (n = 22): tibial TENS. 21/22 completed. Mean age 7.7 years. 7/21 mild UI (some drops on pants); 8/21 moderate UI (underclothes wet); 6/21 severe UI (use of pads) Group 2 (n = 18): sham TENS. 16/18 completed. Mean age 8.4 years. 5/16 mild UI (some drops on pants); 7/16 moderate UI (underclothes wet); 4/16 severe UI (use of pads) Posterior tibial nerve stimulation: 1 x weekly 30‐min session for 12 weeks; neuromodulation using a TENS III stimulator (indigenous machine); 2 self‐adhesive surface electrodes placed in the medial region of the foot, approximately 3‐4 cm cephalad to the medial malleolus between the posterior margin of the tibia and the soleus muscle. Adjustable voltage pulse intensity of 0‐10 mA, a fixed pulse width of 200 ms and a frequency of 20 Hz Sham TENS followed the same schedule but patch electrodes were applied above the medial malleolus to simulate the test group but posterior tibial nerve stimulation settings were not applied and no stimulation was given
Outcomes	Cure: no UI post‐treatment (after 12 weeks of treatment) Self‐reported complete resolution of OAB symptoms (after 12 weeks of treatment)
Notes	Source of funding: none
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No information on random sequence generation
Allocation concealment (selection bias)	Unclear risk	No information on allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No information on blinding of personnel
Blinding of outcome assessment (detection bias) All outcomes	Low risk	UI and OAB improvement outcomes self‐reported by blinded participants
Incomplete outcome data (attrition bias) All outcomes	Low risk	Losses to follow‐up reported and all remaining data reported
Selective reporting (reporting bias)	Low risk	All expected outcomes reported in full
Other bias	Low risk	No reason to suspect any other source of bias

Quintiliano 2015.

Methods	Study design: individually RCT
Participants	Setting: urology departments at two university hospitals Country: Brazil Number randomised: 28 Ages: 4‐17 years Inclusion criteria: urinary urgency at least 3 x/week, bell‐shaped uroflowmetry curve, PVR < 10% EBC for age, no previous treatment, DVSS greater than normal (6 in boys and 9 in girls) Exclusion criteria: neurological disease, urinary tract anatomical anomalies
Interventions	Group 1 (n = 12): TENS (3 x/week) with placebo (2 x daily) with urotherapy. 4 boys, 9 girls; mean age 6.3 years. Urge UI 11; frequency 9; nocturia 3; enuresis 9 Group 2 (n = 15): oxybutynin (0.3 mg/kg 2 x daily) with sham TENS (3 x/week) with urotherapy. 5 boys, 10 girls; mean age 6.5 years. Urge UI 14; frequency 10; nocturia 2; enuresis 10 TENS: 20 sessions, 20 min/session, 3 x/week = 7 weeks. Surface electrodes placed symmetrically on parasacral region (and 2 unused placed on 1 scapula). Biphasic square current pulse, 10 Hz, pulse width 700 milliseconds, intensity increased to threshold Sham TENS followed same schedule but only scapular electrodes were active. Standard urotherapy: scheduled voiding every 3 h/not allowing 4 h to pass without voiding; not retaining urine when there was urinary urgency; urinating before bedtime; avoiding coffee, tea, soda, chocolate and citrus fruits during treatment; ingesting a greater amount of liquid during the day; eating high‐fibre foods; toilet seat reducer/footrest as necessary
Outcomes	Cure: “Complete resolution of symptoms” 3 months after start of treatment DVSS score 3 months after start of treatment Voiding frequency 3 months after start of treatment Adverse events
Notes	Source of funding: none declared
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Remote and computer‐generated random sequence
Allocation concealment (selection bias)	Unclear risk	There is no statement regarding allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Participants blinded but therapists were not blinded
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Evaluation by interviewers blinded to treatment allocation
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Withdrawals and reasons reported in oxybutynin group. However, participants are missing from flowchart in TENS group
Selective reporting (reporting bias)	Low risk	All expected outcomes are reported
Other bias	Low risk	No reason to suspect any other source of bias

Reis 2014.

Methods	Study design: individually RCT
Participants	Setting: hospital outpatient clinic (single centre) Country: Brazil Number randomised: 78 Ages: 5‐16 years Inclusion criteria: daytime UI and nocturnal enuresis Exclusion criteria: monosymptomatic nocturnal enuresis, unwilling to participate, unable to understand, neurological or psychiatric diseases, or anatomical changes to urinary tract
Interventions	Group 1 (n = 39): PFMT with EMG biofeedback (2 x 20‐min sessions/week, maximum 20; mean 10.9) and urotherapy. 31/39 completed (10 boys, 21 girls; mean age 9.2 years). Mean UI episodes/day 1.94; OAB 16/31; DV 10/31 Group 2 (n = 39): TENS (2 x 20‐min sessions/week, maximum 20; mean 18.1) and urotherapy. 33/39 completed (11 boys, 22 girls; mean age 9.6 years). Mean UI episodes/day 1.85, OAB 25/33, DV 4/33 PFMT: in physiotherapy clinic. Re‐education of PF muscles was based on sustained contractions followed by prolonged relaxation, isolating accessory muscle activity such as abdominal, gluteal and adductors. EMG biofeedback with child able to see graphic trace on screen TENS: in physiotherapy clinic. Sacral (S2‐S4) with silicone electrodes. Current with frequency of 10 Hz, pulse width 700 μs and intensity at the sensitive threshold. Urotherapy: advice on timing of voiding, avoiding delay or restraint manoeuvres, voiding posture, fluid intake and dietary advice
Outcomes	Mean UI episodes/day post‐treatment Mean voiding frequency/day post‐treatment Mean functional bladder capacity (highest bladder volume) post‐treatment DVSS post‐treatment QoL: the AUQUEI (Autoquestionnaire Qualité de Vie Enfant Imagé) questionnaire) (Assumpcao 2000)
Notes	Source of funding: none
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Children described as being divided randomly by lottery
Allocation concealment (selection bias)	Unclear risk	No information provided about allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding of participants or personnel not possible as interventions were evidently different
Blinding of outcome assessment (detection bias) All outcomes	High risk	Outcomes assessed by unblinded self‐(carer)‐report. No information provided about blinding of analyses
Incomplete outcome data (attrition bias) All outcomes	High risk	Relatively high proportions of each group did not complete the study, with no reasons given
Selective reporting (reporting bias)	Low risk	Data are reported for outcomes defined in methods. No reason to suspect selective reporting
Other bias	Low risk	No reason to suspect any other source of bias

Rhodes 2008.

Methods	Study design: individually RCT
Participants	Setting: paediatric outpatient clinic Country: UK Number randomised: 20 (14 girls, 6 boys); Mean age 7.4 years Ages: 6‐10 years Inclusion criteria: daytime UI with or without NE Exclusion criteria: underlying serious renal or urological abnormality, previous treatment, management or investigations for daytime UI, history of UTI
Interventions	Group 1 (n = 10): home bladder education workbook (providing a fortnightly lesson plan for bladder rehabilitation, including delayed voiding and advice on fluid intake). Mean age 7.8 years. 4/10 daytime UI only, 6/10 daytime UI and NE Group 2 (n = 10): normal care/verbal bladder education (verbal urotherapy advice only, analogous to the contents of the workbook, provided by a urology nurse). Mean age 7.1 years. 6/10 daytime UI only, 4/10 daytime UI and NE
Outcomes	Days wet/week at weeks 2 and 4 – mean (median)
Notes	Source of funding: none declared At weeks 2 and 4 parents in both groups received phone calls to offer further advice on bladder control, encourage adherence with advice and collect outcome data. At baseline participants were asked how many times wet per day rather than how many days wet per week, so not possible to determine difference between baseline and end of study
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No information on sequence generation
Allocation concealment (selection bias)	Unclear risk	No information on allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not possible to blind participants or personnel due to the different nature of the interventions
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Appears to have been some effort to blind nurse carrying out follow‐up advice and data collection calls but uncertain how effective this could have been as parents were likely to mention the workbook.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study and data are reported for all.
Selective reporting (reporting bias)	Unclear risk	Reporting and analysis of methods and data was limited: no standard deviations are provided; description of advice delivered by the workbook or verbally was not sufficient to fully understand the interventions.
Other bias	Unclear risk	A convenience sample over a 3‐month period and no data about how many chose not to participate

Sillen 2014.

Methods	Study design: individually RCT
Participants	Setting: clinic setting Country: Sweden Number randomised: 62 Ages: 5‐12 years Inclusion criteria: incontinence episodes ≥ 1/week, micturition frequency ≥ 5/day (at start of study) Exclusion criteria: high bladder capacity (> 150% of ((age+1) x 30) mL); PVR > 10 mL; DV (by flow curve), concomitant drug treatment for OAB, untreated UTI or constipation
Interventions	Group 1 (n = 32): urotherapy alone (12 weeks). 19 boys, 11 girls; mean age 8.0 years. Mean UI episodes/day 2.2, mean frequency per day 7.7, mean maximum voided volume 144 mL Group 2 (n = 30): TENS and urotherapy (12 weeks). 16 boys, 16 girls; mean age 8.0 years. Mean UI episodes/day 1.9, mean frequency/day 7.3, mean maximum voided volume 158 mL TENS: 2 x daily for 20 min; low frequency (10 Hz); self‐adhesive surface electrodes placed bilaterally at sacral region (S2‐S3); current intensity increased to maximum tolerated (to a maximum of 40 Ma) Standard urotherapy (both groups): behavioral therapy and detailed lifestyle advice in order to normalise bladder function. Explanation of normal body and LUT function. Explanation of voiding position, timed voiding (before urgency attacks), drinking habits, quality of beverages and prevention of constipation
Outcomes	Number of participants completely dry (complete UI cure) at 12 weeks Number of participants with decreased number of UI episodes at 12 weeks Mean (SD) difference in number of UI episodes/day at 12 weeks Number of participants with decreased voiding frequency at 12 weeks Mean (SD) difference in voiding frequency/day at 12 weeks Mean (SD) difference in maximal voiding volume at 12 weeks Adverse events
Notes	Source of funding: grants from the Region Västra Götaland
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computerised randomisation
Allocation concealment (selection bias)	Unclear risk	There is no explicit statement regarding allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	High risk	1 group received TENS, the other did not. No sham TENS was used to attempt blinding of participants.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Outcomes were self‐reported by unblinded participants. No mention of blinding of study personnel extracting or analysing data
Incomplete outcome data (attrition bias) All outcomes	Low risk	Data for all completing participants reported. Dropouts, including reasons, fully reported
Selective reporting (reporting bias)	Low risk	All outcomes mentioned in methods are reported. No reason to suspect selective reporting
Other bias	Low risk	No reason to suspect other forms of bias

Van Gool 2014.

Methods	Study design: multi‐arm RCT Two branches: branch 1 for OAB/DV (storage symptoms), branch 2 for DV symptoms Only Branch 1 included in review
Participants	Setting: multi‐centre study in hospital outpatient settings Country: Belgium, Germany, the Netherlands, Sweden Number randomised: 202 in whole study (97 in branch 1) Ages: 6‐12 years Inclusion criteria: urge incontinence, uninhibitable urge to void and numerous small voidings per day (OAB), with or without a history of UTI, with or without VUR Exclusion criteria: renal or urinary tract disease other than UTI with/without VUR, preceding urinary tract surgery (including urethral dilatation), monosymptomatic NE, known neurological disorder, informed consent refused, inability to follow protocol, known adverse reactions to anticholinergics
Interventions	Group 1 (n = 33): urotherapy and placebo (blinded against oxybutynin; 6‐month treatment period; sugar‐free syrup, in a dosage of 0.3 mg/kg bodyweight/day, in 3 doses/day). 20 girls, 13 boys, 24 completed Group 2 (n = 30): urotherapy and oxybutynin (double‐blinded against placebo; 6‐month treatment period; oxybutynin chloride 1 mg/mL sugar‐free syrup, in a dosage of 0.3 mg/kg bodyweight per day, in 3 doses per day). 22 girls, 8 boys, 25 completed Group 3 (n = 34): urotherapy and voiding education (with uroflowmetry + feedback), 6‐12 sessions/child over 6 months. 28 girls, 6 boys, 31 completed Standard urotherapy to children and parents: instruction on fluid intake, voiding habits, hygiene, verbal and given to family to take home; bladder diaries; education on urgency, voided volumes and voiding frequency; 3 urotherapy visits during 6‐month treatment period
Outcomes	Cure of daytime UI at 12 months (6 months after treatment period) Adverse effects
Notes	Source of funding: BYK Nederland NV made special batches of oxybutynin chloride and placebo free of charge and SCA Hygiene provided materials for pad tests free of charge.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No information on sequence generation
Allocation concealment (selection bias)	Low risk	Allocation determined remotely from study sites
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Although double‐blinding was in place for placebo and oxybutynin arms, it was not possible for the voiding education arm.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No information on blinding of outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	Outcomes are reported for both ITT and PP analyses and losses to follow‐up reported with reasons
Selective reporting (reporting bias)	High risk	A large proportion of baseline and outcome measurements (diary and questionnaire data, test data) described in the methods section are not reported
Other bias	Low risk	No reason to suspect any other source of bias

Vasconcelos 2007.

Methods	Study design: individually RCT
Participants	Setting: paediatric nephro‐urology unit, university hospital (single centre) Country: Brazil Number randomised: 59 (3 did not complete) Ages: age criteria not reported (mean age 10.8 years in group 1, 10.3 years in group 2) Inclusion criteria: child or adolescent, dysfunctional elimination syndrome (an abnormal pattern of elimination for the child’s age characterised by UI and FI and withholding), ≥ 6 months prior treatment in outpatient clinic without any improvement of their UI, urgency or urge incontinence, and/or constipation Exclusion criteria: occult neurological lesions; complex urinary tract abnormalities (other than VUR), attention deficit hyperactivity disorder, primary monosymptomatic nocturnal enuresis
Interventions	Group 1 (n = 26): PFMT (24 x 60‐min sessions of PF exercises over a 3‐month period). Sessions consisted of a series of 7 PF musculature contraction and relaxation exercises) and behavioural modification. 9 boys, 17 girls; mean age 10.8 years, range 7.0‐14.9 years. Daytime UI 21; nocturnal incontinence 24; urge/urge incontinence 23 Only those with daytime UI included in review analyses Group 2 (n = 30): PFMT with feedback (16 x 60‐min sessions of PF exercises with biofeedback over a 2‐month period. Commercially available biofeedback system that generated EMG tracing during the contraction and relaxation phase) and behavioural modification. 10 boys, 20 girls; mean age 10.3 years, range 5.9‐15.2 years. Daytime UI 20; nocturnal incontinence 24; urge/urge incontinence 25. Only those with daytime UI included in review analyses Behavioural modification: scheduled voiding and drinking; training in adequate toilet posture; reinforcement using voiding diaries
Outcomes	Daytime UI cured (no wetting episodes over a 4‐week period) Daytime UI improved (≥ 50% reduction in wetting episodes over a 4‐week period)
Notes	Source of funding: partially supported by the Brazilian National Research Council
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer generation of randomisation sequence
Allocation concealment (selection bias)	Unclear risk	No information on allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding was not possible because the interventions were different in nature and schedule for each group.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Outcomes were self‐reported by unblinded participants or parents. However, interventions being compared were only different in technical ways that are unlikely to have influenced lay perceptions.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data are reported fully for all outcomes and low attrition fully reported
Selective reporting (reporting bias)	Low risk	All outcomes are reported. No reason to suspect selective reporting in either the substance or focus of reporting
Other bias	Low risk	No reason to suspect any other source of bias

Zeng 2012.

Methods	Study design: individually RCT
Participants	Setting: urology department at a university hospital (single centre) Country: China Number randomised: 39 Ages: 8‐14 years Inclusion criteria: ≥ 5 episodes of postmicturition dribbling in month prior to enrolment, normal urinary tract, normal urinalysis, DA‐C defined as a contraction ≥ 15 cm water after cessation of the voiding contraction Exclusion criteria: overt or suspected neurological problems, other urological problems, UTI, constipation, vaginal voiding, other medical treatment
Interventions	Group 1 (n = 21): PFMT with biofeedback (2‐3, 20‐30‐min training sessions/week for 12 weeks. Patients taught to perceive relaxation and contraction of the anal sphincter, and instructed to contract while simultaneously maintaining stable abdominal pressure). 15 girls, 6 boys; mean age 10.5, 20 completed Group 2 (n = 18): tolterodine (1 mg, orally, 2 x daily for 12 weeks). 13 girls, 5 boys; mean age 11.1 years, 16 completed
Outcomes	Complete cure of post‐micturition dribbling after 12 weeks of treatment (no postmicturition dribbling episodes in the month immediately following treatment) Significant improvement of postmicturition dribbling after 12 weeks of treatment (only 1 episode in month after treatment)
Notes	Source of funding: none declared
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated sequence generation
Allocation concealment (selection bias)	Unclear risk	No information on allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding of participants and personnel not possible due to inherent differences in the treatments
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	The report states that a physician blinded to allocation assessed the outcomes but since outcomes were self‐reported by participants it is unclear what affect blinding the physician assessor had.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low attrition, with all cases accounted for
Selective reporting (reporting bias)	Low risk	All expected outcomes reported. No reason to suspect selective reporting
Other bias	Low risk	No reason to suspect any other source of bias

Zivkovic 2010.

Methods	Study design: individually RCT
Participants	Setting: physical and rehabilitation medicine clinic (single centre) Country: Serbia Number randomised: 86 Ages: 3‐13 years Inclusion criteria: DV (UI, hesitancy, straining, weak stream intermittency), staccato/interrupted uroflowmetry with positive PF EMG activity on voiding, previous unsuccessful treatment ≥ 3 months by paediatrician with timed voiding, fluid intake advice and constipation management Exclusion criteria: neurological/psychiatric disorders, structural abnormality of the LUT
Interventions	Group 1 (n = 43): PF muscle/diaphragmatic exercises and urotherapy. 28 girls, 15 boys; mean age 7.5 years. Daytime UI 24/43, NE 21/43. Only those with daytime UI included in review analyses Group 2 (n = 32): urotherapy alone. 23 girls, 9 boys; mean age 6.7 years. Daytime UI 18/32, NE 15/32. Only those with daytime UI included in review analyses. Monthly therapeutic sessions with total number determined by clinical results (mean 10 sessions over 1 year). Diaphragmatic breathing exercises: performed in lying and sitting positions. Children were asked to inhale through the nose, bulge the abdomen outward as much as possible, hold their breath for a few seconds, and then exhale slowly through pursed lips. Children were instructed to watch the anterior abdominal wall movement during inspiration and to repeat the process while seated on the toilet to initiate voiding. Aimed at achieving abdominal muscle relaxation. Asked to perform exercises daily at home PF muscle exercises: 3‐second contraction, 30‐second relaxation; 30 repetitions daily; asked to perform exercises daily at home Standard urotherapy: education about urinary tract function; fluid intake advice (200 mL 5‐6 x/day); regular voiding; genital hygiene; voiding posture instruction; treatment of constipation
Outcomes	Daytime UI completely resolved at 1 year after treatment commencement
Notes	Source of funding: none declared Ongoing pharmacological treatments were not stopped prior to study
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was performed by children drawing a sealed envelope with an enclosed assignment.
Allocation concealment (selection bias)	Unclear risk	No information on allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding of participants and personnel not possible due to differences between the interventions
Blinding of outcome assessment (detection bias) All outcomes	High risk	Outcome self‐reported by unblinded children/parents
Incomplete outcome data (attrition bias) All outcomes	Low risk	All data are reported for participants that completed the study and losses to follow‐up reported
Selective reporting (reporting bias)	Low risk	No reason to suspect selective reporting. All expected outcomes reported
Other bias	Low risk	No reason to suspect any other source of bias

BBD: bladder‐bowel dysfunction; DA‐C: detrusor after‐contraction; DV: dysfunctional voiding; DVSS: dysfunctional voiding symptom score; EBC: expected bladder capacity; EMG: electromyography; FI: faecal incontinence; GI: gastrointestinal; GU: genitourinary; ICCS: International Children’s Continence Society; ITT: intention‐to‐treat; KUB: kidney, ureter and bladder (radiograph); LUT: lower urinary tract; MRI: magnetic resonance imaging; NE: nocturnal enuresis; OAB: overactive bladder; PF: pelvic floor; PFMT: pelvic floor muscle training; PP: per protocol; PVR: postvoid residual; QoL: quality of life; RCT: randomised controlled trial; SD: standard deviation; SE: standard error; TENS: transcutaneous electrical nerve stimulation; UI: urinary incontinence; UTI: urinary tract infection; VAS: visual analogue scale; VCUG: voiding cystourethrogram; VUR: vesicoureteral reflux

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Alcantara 2015	No comparator arm
Barroso 2006	No randomisation. Observational study
Barroso 2011	Review. No references not already identified by search
Benninga 2008	Not completed due to poor recruitment (correspondence from study author)
Boudaoud 2015	No usable data on daytime urinary incontinence outcomes. A study‐specific composite symptom score was reported.
Bower 2004	Systematic review. Only abstract available so not possible to check included studies for possibly eligible comparator arms
Deng 2011	Study compares tolterodine, oxybutynin and placebo
Egger 1992	Not an original study. Secondary analysis of subgroups with night‐time enuresis or day‐ and night‐time enuresis from 2 previous studies of diet modification on children with migraine and hyperkinetic behavior. Only a subgroup of participants had daytime UI. In randomised cross‐over element of study it is not possible to determine which participants, if any, had daytime UI.
Fan 2014	Invasive procedure (acupuncture) plus TENS‐like stimulation versus oxybutynin, and no UI‐specific outcomes reported
Fan 2016	No indication of how many children had daytime UI, and no UI‐specific outcomes reported
Fazeli 2015	Systematic review. No references not already identified by search
Fernandez 2017	Meta‐analysis
Ferreira 2014	The majority of participants (age range 16‐26) were outside review age range
Hagglund 1965	No randomisation
Hou 2010	Observational study. No randomisation
Kajiwara 2010	Focus is on NE that persists after cure of daytime OAB symptoms
Kramer 2005	Comparison is doxazosin vs placebo for DV and incontinence. Reported that dietary changes and stool softeners were used for any children with coexisting history of constipation but numbers are not reported.
Landgraf 2007	Assessment/validation of QoL scale in 2 studies and a community sample. Details of studies not reported
Lebedev 1995	Intervention is aimed at treatment of NE
Leclair 2010	Narrative review. No references not identified by search
Liu 1996	Invasive procedure (acupuncture) in children with incontinence with neuropathic aetiology
Marschall‐Kehrel 2009	Pharmaceutical intervention vs placebo. Conservative treatment prior to start of intervention (urotherapy) same in both groups. Report states: "The trial design did not allow for separate evaluation of the effect of urotherapy prior to medical treatment."
Matson 1977	Population 20‐26 months old
Minni 1990	Acupuncture with or without electrical stimulation. No randomisation reported
NCT00518804	Intense behavioural therapy aimed at global skills development in autistic children aged 3‐6 years. Incontinence not reported as a baseline factor or an outcome
NCT01655069	Single (pharmaceutical) arm extension of a previous pharmaceutical study
NCT01978210	Pilot study for NCT02369445. Children aged 3‐6
NCT02097121	Invasive procedure. 3‐armed study comparing doses of onabotulinum toxin A injected into the bladder wall
NCT02369445	Principal cause of functional UI in intended participants is autism (autistic children aged 3‐10). Comparator is autism‐specific toilet training
NCT02614482	Single (pharmaceutical) arm extension of a previous pharmaceutical study
Ni 2016	All participants received surgical interventions
Oktar 2018	No eligible daytime UI outcomes
Pu 2011	Observational study. No randomisation
Rashid 2011	Population was children with Hinmans syndrome. Sizeable proportion (43%) of participants aged 1‐6 in both groups. All participants received pharmacotherapy
Rodríguez Sacristán 1983	Majority of study population have developmental/cognitive reasons for continence problems
Steinbok 2005	Surgery vs medical treatment
Sureshkumar 2003	Systematic review. No references not already identified by search
Tang 1997	Study compares Chinese manufactured and imported oxybutynin
Trsinar 1996	Invasive procedure. No report of any randomisation and uncertainty as to equivalence of the groups
Trsinar 2012	Invasive procedure. No report of any randomisation and uncertainty as to equivalence of the groups
Voorham 2015	The majority of participants were outside review age range. Mean age was 56.13 years (range 16.24 to 72.96 years)
Xing 2015	Observational study. No randomisation
Yao 2016	Invasive treatment. Botulinum toxin A injected in detrusor compared with botulinum toxin A injected in both detrusor and trigone
Ye 1999	Observational study. No randomisation
You 2008	Observational study. No randomisation
Yousefichaijan 2017	No eligible daytime UI outcomes
Zhang 2010	Observational study. No randomisation
Zhang 2011	UI of neurological aetiology (neuropathic acontractile sphincter incontinence due to damage to central or peripheral nerves)
Zhu 2013	Observational study. No randomisation
Zhu 2016	Systematic review. No references not already identified by search
Zivkovic 2017	No randomisation. Controlled, prospective clinical study

DV: dysfunctional voiding; NE: nocturnal enuresis; OAB: overactive bladder; QoL: quality of life; TENS: transcutaneous electrical nerve stimulation; UI: urinary incontinence

Characteristics of studies awaiting assessment [ordered by study ID]

Braga 2017.

Methods	Study design: RCT
Participants	150 children aged 5‐10 years with score of ≥ 11 on Vancouver global BBD scale. Only a small proportion of children have daytime UI daily (17/75 and 16/75 per group), not reported how many have UI less frequently than daily
Interventions	Animated voiding education video vs standard urotherapy
Outcomes	Mean change in Vancouver global BBD scale, with data not reported for daytime UI
Notes	We emailed study authors for information on children's daytime UI baseline status and outcomes.

Kueter 2000.

Methods	Study design: RCT
Participants	Children aged 3‐14 years with DV
Interventions	Pharmacological treatment (ditropan with or without antibiotic prophylaxis) vs 5‐step multi‐modality treatment approach
Outcomes	Change in DVSS
Notes	Abstract with insufficient information on methods and procedures and minimal outcome data. We emailed study authors

Mourani 2012.

Methods	Study design: RCT
Participants	52 girls aged 4‐12 years with OAB dysfunction
Interventions	PFMT with biofeedback plus behavioural modification (voiding and drinking schedule, instruction and toilet behaviour) vs anticholinergic drugs only
Outcomes	Daytime UI, bladder capacity and incidence of UTI
Notes	Abstract only. We emailed study authors for information on children's daytime UI baseline status and outcomes

Stauber 2007.

Methods	Study design: RCT
Participants	Boys aged 8‐12 years with functional UI
Interventions	Inpatient education programme involving cognitive‐behavioural stress management training vs more knowledge‐oriented education programme without stress management
Outcomes	Daytime UI compared using non‐parametric tests, with no usable data
Notes	We emailed study authors for further information and data

Weber 2011.

Methods	Study design: RCT
Participants	Children with DV. Unclear how many with daytime UI
Interventions	Biofeedback vs none
Outcomes	DVSS scores
Notes	Abstracts with insufficient information on methods and procedures and minimal outcome data. We emailed study authors

DV: dysfunctional voiding; BBD: bladder and bowel dysfunction; DVSS: dysfunctional voiding symptom score; OAB: overactive bladder; PFMT: pelvic floor muscle training; RCT: randomised controlled trial; UI: urinary incontinence; UTI: urinary tract infection

Characteristics of ongoing studies [ordered by study ID]

ACTRN12611000828921.

Trial name or title	A randomised controlled trial comparing the proportion of patients with daytime incontinence cured at 3 months by the personal alarm watch versus a conventional watch to aid timed voiding
Methods	RCT
Participants	Children aged 3‐15 years with daytime UI ≥ 2 episodes/week in previous 2 weeks
Interventions	Alarm watch vs watch without alarms
Outcomes	Proportion of children cured at 3 months measured by a bladder diary Time to achieve cure measured by a bladder diary Change in daytime wetting (compared to baseline) Completeness of bladder emptying (compared to baseline) Proportion relapsed by 6 months (to assess whether treatment success is sustained) QoL
Starting date	October 2011
Contact information	Dr Patrina Caldwell, Centre for Kidney Research, The Children's Hospital at Westmead, Locked Bag 4001, Westmead NSW 2145, Australia. patrina.caldwell@health.nsw.gov.au
Notes	At time of contact with the researchers recruitment had finished but some participants were still being followed up, so the study had not been unblinded.

NCT02336906.

Trial name or title	A controlled randomized, physician blinded study to assess isolated urotherapy vs. urotherapy with constipation treatment for patients with lower urinary tract dysfunction (LUTD) not meeting ROME‐III constipation criteria
Methods	RCT
Participants	Children aged 5‐17 years with LUTD, daytime/daytime and night‐time UI
Interventions	Urotherapy with constipation treatment vs urotherapy alone
Outcomes	Complete (Vancouver score ≤ 90% of screening value) or partial (Vancouver score ≤ 50% screening value) response at the end of intervention Sustained response Adverse events
Starting date	January 2020
Contact information	Noam Zevit, Pediatric Gastroenterologist, Rabin Medical Center, Israel
Notes

NCT03478813.

Trial name or title	Voiding school as a treatment of children's functional incontinence ‐ evaluation and implementation of the intervention
Methods	RCT
Participants	Children aged 5‐6 years with daytime incontinence or enuresis
Interventions	Voiding school (children educated in groups of 4‐6 children with child‐oriented methods highlighting learning by doing in order to achieve better bladder control) compared with usual care (individual advice concerning voiding habits and general life‐style advice)
Outcomes	Changes in wetting episodes Changes in symptoms associated with incontinence Changes in quality of life
Starting date	January 2018
Contact information	Anneli Saarikoski, MNSc, Helsinki University Central Hospital. anpasaa@utu.fi
Notes

LUTD: lower urinary tract dysfunction; QoL: quality of life; RCT: randomised controlled trial; UI: urinary incontinence

Differences between protocol and review

Random‐effects model

Our protocol stated that we would use fixed‐effects models in meta‐analysis unless we identified significant heterogeneity between studies, in which case we would use a random‐effects model (Buckley 2016). Where only small numbers of small studies are available for meta‐analysis as in this review, the I² statistic cannot measure relative heterogeneity accurately and the Q‐test has been criticised for having low power to detect heterogeneity (Borenstein 2017; Hoaglin 2016; von Hippel 2015). For this reason and because some studies employed dissimilar treatment protocols, we considered a random‐effects model appropriate for all meta‐analyses.

Title

To comply with Cochrane recommendations, the title has been changed to 'Conservative interventions for treating functional daytime urinary incontinence in children'.

Contributions of authors

BSB: conducted screening, data extraction and analysis and drafted the review text, with statistical, clinical and interpretation inputs from all authors appropriate to their speciality.
 CDS: conducted screening, data extraction and analysis.
 LS: conducted data extraction and analysis.
 QD: approved the final manuscript.
 JSWK: approved the final manuscript.

Sources of support

Internal sources

• No sources of support supplied

External sources

• National Institute for Health Research, UK.

  This project was supported by the National Institute for Health Research (NIHR), via Cochrane Infrastructure funding to Cochrane Incontinence. The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the Systematic Reviews Programme, NIHR, National Health Service or the Department of Health. The NIHR is the largest single funder of Cochrane Incontinence.

Declarations of interest

BSB: none declared
 CDS: none declared
 LS: none declared
 QD: none declared
 JSWK: none declared

New