Tricyclic and related drugs for nocturnal enuresis in children

Abstract

Background

Enuresis (bedwetting) affects up to 20% of five year‐olds and 2% of adults. Although spontaneous remission often occurs, the social, emotional and psychological costs can be great. Tricyclics have been used to treat enuresis since the 1960s.

Objectives

To assess the effects of tricyclic and related drugs compared with other interventions for treating children with enuresis.

Search methods

We searched the Cochrane Incontinence Group Specialised Trials Register (containing trials identified from the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, MEDLINE in process, ClinicalTrials.gov, WHO ICTRP and handsearching of journals and conference proceedings), on 30 November 2015, and reference lists of relevant articles.

Selection criteria

We included all randomised and quasi‐randomised trials comparing a tricyclic or related drug with another intervention for treating enuresis. We also included combination therapies that included tricyclics. We excluded trials for treating daytime wetting.

Data collection and analysis

Two review authors independently assessed the quality of the eligible trials, and extracted data. We settled differences by discussion with a third review author.

Main results

Sixty‐four trials met the inclusion criteria, involving 4071 children. The quality of many trials was poor, with comparisons addressed by single studies. Minor adverse effects were common, and reported in 30 trials. These included dizziness, headache, mood changes, gastrointestinal discomforts and neutropenia. More serious side‐effects can occur but were not reported. Seven trials reported no adverse effects.

Tricyclics are more effective than placebo, particularly for short‐term outcomes. Compared to placebo, imipramine resulted in one fewer wet nights per week (mean difference (MD) ‐0.95, 95% confidence interval (CI) ‐1.40 to ‐0.50; 4 trials, 347 children), with fewer failing to achieve 14 consecutive dry nights (78% versus 95% for placebo, RR 0.74, 95% CI 0.61 to 0.90; 12 trials, 831 children). Amitriptyline and desipramine were more effective than placebo, but nortriptyline and mianserin showed no difference. Most tricyclics did not have a sustained effect after ceasing treatment, with 96% wetting at follow‐up for imipramine versus 97% for placebo.

Imipramine combined with oxybutynin is also more effective than placebo, with 33% failing to achieve 14 consecutive dry nights at the end of treatment versus 78% for placebo (RR 0.43, 95% CI 0.23 to 0.78; 1 trial, 47 children) and 45% wetting at follow‐up versus 79% for placebo (RR 0.58, 95% CI 0.34 to 0.99; 1 trial, 36 children).

There was insufficient evidence to judge the effect between different doses of tricyclics, and between different tricyclics. Treatment outcomes between tricyclic and desmopressin were similar, but were mixed when tricyclic was compared with an anticholinergic. However, when imipramine was compared with desmopressin plus oxybutynin (1 trial, 45 children), the combination therapy was more effective, with one fewer wet nights per week (MD 1.07, 95% CI 0.06 to 2.08) and 36% failing to achieve 14 consecutive dry nights versus 87% for imipramine (RR 2.39, 95% CI 1.35 to 4.25). Tricyclics were also more effective or showed no difference in response when compared to other drugs which are no longer used for enuresis.

Tricyclics were less effective than alarms. Although there was no difference in the number of wet nights, 67% failed to achieve 14 consecutive dry nights for imipramine versus only 17% for alarms (RR 4.00, 95% CI 1.06 to 15.08; 1 trial, 24 children). Alarm therapy also had a more sustained effect after ceasing treatment with 100% on imipramine versus 58% on alarms wetting at follow‐up (RR 1.67, 95% CI 1.03 to 2.69; 1 trial, 24 children).

Imipramine was more effective than simple behavioural therapies during treatment, with one fewer wet nights per week compared with star chart plus placebo (MD ‐0.80, 95% CI ‐1.33 to ‐0.27; 1 trial, 250 children). At follow‐up 40% were wet with imipramine versus 80% with fluids and avoiding punishment (RR 0.50, 95% CI 0.28 to 0.89; 1 trial, 40 children). However, imipramine was less effective than complex behavioural therapies, with 61% failing to achieve 14 consecutive dry nights for imipramine versus 33% for the three‐step programme (RR 1.83, 95% CI 1.08 to 3.12; 1 trial, 72 children) and 16% for the three‐step programme combined with motivational therapy and computer‐led education (RR 3.91, 95% CI 2.30 to 6.66; 1 trial, 132 children) at the end of treatment, with similar results at follow‐up.

Tricyclics were more effective than restricted diet, with 99% failing to achieve 14 consecutive dry nights versus 84% for imipramine (RR 0.84, 95% CI 0.75 to 0.93; 1 trial, 147 children).There was insufficient evidence to judge the effect of tricyclics compared to the other miscellaneous interventions studied.

At the end of treatment there were about two fewer wet nights for imipramine plus oxybutynin compared with imipramine monotherapy (MD ‐2.10, 95% CI ‐2.99 to ‐1.21; 1 trial, 63 children) and 48% on imipramine plus oxybutynin failed to achieve 14 consecutive dry nights compared with 74% on imipramine monotherapy (RR 0.68, 95% CI 0.50 to 0.92; 2 trials, 101 children). At follow‐up, 45% on imipramine plus oxybutynin were wetting versus 83% on imipramine monotherapy (RR 0.55, 95% CI 0.32 to 0.92; 1 trial, 36 children).

When imipramine combined with desmopressin was compared with imipramine monotherapy, there was no difference in outcomes. However, when imipramine plus desmopressin was compared with desmopressin monotherapy, the combination was more effective, with 15% not achieving 14 consecutive dry nights at the end of treatment for imipramine plus desmopressin versus 40% for desmopressin monotherapy (RR 0.38, 95% CI 0.17 to 0.83; 1 trial, 86 children). Tricyclics combined with alarm therapy were not more effective than alarm monotherapy, alarm combined with desmopressin or alarm combined with nortriptyline. The addition of a tricyclic to other behavioural therapies did not alter treatment response.

Authors' conclusions

There was evidence that tricyclics are effective at reducing the number of wet nights during treatment, but do not have a sustained effect after treatment stops, with most children relapsing. In contrast, there was evidence that alarm therapy has better short‐ and long‐term outcomes. There was some evidence that tricyclics combined with anticholinergics may be more effective that tricyclic monotherapy.

Plain language summary

Tricyclics and related drugs for treating bedwetting in children

Background:
 Bedwetting (nocturnal enuresis or enuresis) is the involuntary loss of urine during sleep without being caused by a physical disorder. It can result in social, emotional and psychological problems and lower quality of life. It affects around 15% to 20% of five‐year‐olds, and can persist in up to 2% of adults. Many different types of drugs have been used to treat bedwetting, both as a single treatment and in combination with other treatments. Tricyclics are antidepressants, which have been used for treating bedwetting since the 1960s. They probably work by their antispasmodic effect on the bladder. This review examines 64 trials of tricyclic drugs, involving 4071 children.

Main findings:
 The most commonly used tricyclic for treating bedwetting is imipramine. Compared to placebo, tricyclics reduce bedwetting by about one night per week during treatment and about a fifth of the children achieve 14 dry nights. However, most wet again once the tricyclics are stopped.

Adverse effects: 
 Tricyclics are just as effective as the other commonly‐used drug, desmopressin. However, they have more side effects, such as seizures, effects on the heart, liver and blood including low white cell counts, particularly if overdosed, which can be serious. The available evidence suggests that bedwetting alarms are a more effective treatment for bedwetting than tricyclics, and do not have the side effects. They also have a more sustained effect after stopping, but they require more effort by the child and more support by their families to be effective.

Limitations:
 Most of the studies included in the review are old (latest 2008), small and of low quality, with insufficient information to judge the risk of bias. The evidence in this review is current up to 30 November 2015.

Summary of findings

Summary of findings for the main comparison. Tricyclics versus placebo.

Tricyclics versus placebo
Patient or population: children with nocturnal enuresis Setting: Outpatients Intervention: TRICYCLIC DRUG Comparison: PLACEBO
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	Number of participants (studies)	Quality of the evidence (GRADE)	Comments
	Risk with PLACEBO	Risk with TRICYCLIC DRUG
Number of wet nights per week at the end of treatment ‐ imipramine vs placebo	The mean number of wet nights per week at the end of treatment ‐ imipramine vs placebo was 0	The mean number of wet nights per week at the end of treatment ‐ imipramine vs placebo in the intervention group was 0.95 fewer wet nights (1.4 fewer to 0.5 fewer)	‐	347 (4 RCTs)	⊕⊝⊝⊝ VERY LOW 1, 2
Number not achieving 14 consecutive dry nights at the end of treatment ‐ imipramine vs placebo	Study population		RR 0.74 (0.61 to 0.90)	831 (12 RCTs)	⊕⊝⊝⊝ VERY LOW 2, 3, 4
	945 per 1000	700 per 1000 (577 to 851)
	Moderate
	940 per 1000	695 per 1000 (573 to 846)
Number not achieving 14 consecutive dry nights at the end of treatment ‐ amitriptyline vs placebo	Study population		RR 0.83 (0.72 to 0.97)	98 (2 RCTs)	⊕⊕⊝⊝ LOW 1, 2
	898 per 1000	745 per 1000 (647 to 871)
	Moderate
	808 per 1000	670 per 1000 (582 to 783)
Number of wet nights per week at follow up ‐ imipramine vs placebo	The mean number of wet nights per week at follow up ‐ imipramine vs placebo was 0	The mean number of wet nights per week at follow up ‐ imipramine vs placebo in the intervention group was 1.5 fewer wet nights (4.85 fewer to 1.85 more)	‐	21 (1 RCT)	⊕⊝⊝⊝ VERY LOW 2, 4, 5
Number not achieving 14 consecutive dry nights or relapsing at follow up ‐ imipramine vs placebo	Study population		RR 0.99 (0.95 to 1.03)	416 (5 RCTs)	⊕⊕⊝⊝ LOW 4, 5
	968 per 1000	958 per 1000 (919 to 997)
	Moderate
	988 per 1000	979 per 1000 (939 to 1000)
Adverse events assessed as reported in paper			‐	56 (18 RCTs)	⊕⊝⊝⊝ VERY LOW 3, 4, 6	Adverse effects commonly reported for tricyclics include dizziness, gastrointestinal symptoms (nausea, vomiting, abdominal pain, diarrhoea, dry mouth), headache, lethargy, sleep disturbance. Rare adverse effects include neutropenia, epistaxis, depression and weight loss. Adverse effects were reported for imipramine, nortriptyline, amitriptyline and viloxazine. Similar effects were reported for placebo, but were reported less commonly. Some studies did not identify the arm in which adverse effects occurred.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio
GRADE Working Group grades of evidence High quality: We are very confident that the true effect lies close to that of the estimate of the effect Moderate quality: 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 quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect Outcomes were downgraded one level for serious imprecision (most studies were small and had large confidence intervals), and two levels for very serious study limitations (high risk of bias associated with sequence generation, allocation concealment and high level of drop out).

¹Most of the studies were cross‐over trials, with inadequate washout period. Downgraded one level for study limitations (data taken from first arm of cross‐over trial and analysed as parallel‐group trial).
 ²Most studies were small and had large confidence intervals.
 ³Studies were heterogeneous.
 ⁴Unclear or high risk of risk of bias for random sequence generation and allocation concealment.
 ⁵Large number of dropouts.
 ⁶Not all studies differentiated in which arm adverse effects occured.

Summary of findings 2. Tricyclics versus another drug.

Tricyclics versus another drug
Patient or population: Children with nocturnal enuresis Setting: Outpatients Intervention: TRICYCLIC DRUG Comparison: ANOTHER DRUG
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	Number of participants (studies)	Quality of the evidence (GRADE)	Comments
	Risk with ANOTHER DRUG	Risk with TRICYCLIC DRUG
Number of wet nights per week at the end of treatment ‐ imipramine vs desmopressin	The mean number of wet nights per week at the end of treatment ‐ imipramine vs desmopressin was 0	The mean number of wet nights per week at the end of treatment ‐ imipramine vs desmopressin in the intervention group was 0.11 more wet nights (0.41 fewer to 0.62 more)	‐	300 (3 RCTs)	⊕⊝⊝⊝ VERY LOW 1, 2, 3
Number of wet nights per week at the end of treatment ‐ amitriptyline vs desmopressin	The mean number of wet nights per week at the end of treatment ‐ amitriptyline vs desmopressin was 0	The mean number of wet nights per week at the end of treatment ‐ amitriptyline vs desmopressin in the intervention group was 1.4 fewer wet nights (2.68 fewer to 0.12 fewer)	‐	31 (1 RCT)	⊕⊝⊝⊝ VERY LOW 1, 4, 5, 6
Number not achieving 14 consecutive dry nights at the end of treatment ‐ imipramine vs desmopressin	Study population		RR 0.88 (0.64 to 1.20)	222 (4 RCTs)	⊕⊝⊝⊝ VERY LOW 1, 2, 3, 5, 7
	426 per 1000	375 per 1000 (273 to 511)
	Moderate
	405 per 1000	356 per 1000 (259 to 486)
Number not achieving 14 consecutive dry nights at the end of treatment ‐ imipramine vs desmopressin + oxybutynin	Study population		RR 2.39 (1.35 to 4.25)	45 (1 RCT)	⊕⊝⊝⊝ VERY LOW 1, 3, 5, 7
	364 per 1000	869 per 1000 (491 to 1000)
	Moderate
	364 per 1000	869 per 1000 (491 to 1000)
Adverse events assessed as reported in paper			‐	(9 RCTs)	⊕⊝⊝⊝ VERY LOW 2, 3, 8	Tricyclics were compared with other tricyclics (viloxazine, nortriptyline), anticholinergics (oxybutynin, emepronium), desmopressin, and desmopressin combined with imipramine or oxybutynin. Adverse effects of rash, headache, back pain and nasal inflammation was found for desmopressin. Most studies did not differentiate which treatment caused which adverse effects, and included common adverse effects such as dizziness, gastrointerstinal symptoms, headache and lethargy.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio
GRADE Working Group grades of evidence High quality: We are very confident that the true effect lies close to that of the estimate of the effect Moderate quality: 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 quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect Outcomes were downgraded two levels for very serious study limitations (high risk of bias associated with inadequate blinding, sequence generation, allocation concealment and high level of drop out).

¹Most studies were small and had large confidence intervals.
 ²Studies heterogeneous with large variation in effects.
 ³High or unclear risk of bias for random sequence generation and allocation concealment.
 ⁴Premature termination of trial.
 ⁵Large numbers of dropouts.
 ⁶No reporting of side effects.
 ⁷No or inadequate blinding.
 ⁸Not all studies differentiated in which arm adverse effects occured.

Summary of findings 3. Tricyclics versus behavioural interventions.

Tricyclics versus behavioural interventions
Patient or population: children with nocturnal enuresis Setting: Outpatients Intervention: DRUGS Comparison: BEHAVIOURAL INTERVENTIONS
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	Number of participants (studies)	Quality of the evidence (GRADE)	Comments
	Risk with BEHAVIOURAL INTERVENTIONS	Risk with DRUGS
Number not achieving 14 consecutive dry nights at the end of treatment ‐ imipramine vs alarm	Study population		RR 4.00 (1.06 to 15.08)	24 (1 RCT)	⊕⊝⊝⊝ VERY LOW 1, 2, 3, 4
	167 per 1000	667 per 1000 (177 to 1000)
	Moderate
	167 per 1000	667 per 1000 (177 to 1000)
Number not achieving 14 consecutive dry nights at the end of treatment ‐ imipramine vs alarm (Mozes detector)	Study population		RR 1.28 (1.07 to 1.53)	184 (2 RCTs)	⊕⊝⊝⊝ VERY LOW 1, 2, 3, 4, 5
	634 per 1000	812 per 1000 (679 to 971)
	Moderate
	541 per 1000	693 per 1000 (579 to 828)
Number not achieving 14 consecutive dry nights at the end of treatment ‐ imipramine vs 3 step therapy (reassurance, retention control training + wakening, parental involvement)	Study population		RR 1.83 (1.08 to 3.12)	72 (1 RCT)	⊕⊕⊝⊝ LOW 1, 2, 4
	333 per 1000	610 per 1000 (360 to 1000)
	Moderate
	333 per 1000	610 per 1000 (360 to 1000)
Number not achieving 14 consecutive dry nights at the end of treatment ‐ imipramine vs 3 step therapy + motivation/counselling + education	Study population		RR 3.91 (2.30 to 6.66)	132 (1 RCT)	⊕⊕⊝⊝ LOW 1, 2, 4
	156 per 1000	611 per 1000 (359 to 1000)
	Moderate
	156 per 1000	611 per 1000 (359 to 1000)
Number not achieving 14 consecutive dry nights or relapsing at follow up ‐ imipramine vs alarm	Study population		RR 1.67 (1.03 to 2.69)	24 (1 RCT)	⊕⊝⊝⊝ VERY LOW 1, 2, 3, 4
	583 per 1000	974 per 1000 (601 to 1000)
	Moderate
	583 per 1000	974 per 1000 (601 to 1000)
Number not achieving 14 consecutive dry nights or relapsing at follow up ‐ imipramine vs 3 step therapy + motivation/counselling + education	Study population		RR 3.56 (2.21 to 5.72)	132 (1 RCT)	⊕⊕⊝⊝ LOW 1, 2, 4
	188 per 1000	668 per 1000 (414 to 1000)
	Moderate
	188 per 1000	668 per 1000 (414 to 1000)
Adverse events assessed as reported in paper			‐	(4 RCTs)	⊕⊝⊝⊝ VERY LOW 1, 2, 5, 6	Drowsiness reported for amitriptyline, and headache, fatigue and gastrointestinal symptoms mentioned for imipramine. Fear, skin damage (erythema, burn, ulceration) noted for electric shock alarm
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio
GRADE Working Group grades of evidence High quality: We are very confident that the true effect lies close to that of the estimate of the effect Moderate quality: 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 quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect Outcomes were downgraded one level for serious imprecision (most studies were small and had large confidence intervals), and two levels for very serious study limitations (high risk of bias associated with sequence generation, allocation concealment and high level of drop out)

¹No or inadequate blinding.
 ²High or unclear risk of bias for random sequence generation and allocation concealment.
 ³Large numbers of dropouts.
 ⁴Most studies were small and had large confidence intevals.
 ⁵Studies heterogeneous with large variations in effect.
 ⁶Not all studies differentiated in which arm adverse effects occurred.

Summary of findings 4. Combination therapy involving tricyclics.

Combination therapy involving tricyclics
Patient or population: children with nocturnal enuresis Setting: Outpatients Intervention: COMBINATION THERAPY INVOLVING TRICYCLICS Comparison: OTHER
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	Number of participants (studies)	Quality of the evidence (GRADE)	Comments
	Risk with OTHER	Risk with COMBINATION THERAPY INVOLVING TRICYCLICS
Number of wet nights per week at the end of treatment ‐ imipramine + oxybutynin vs imipramine	The mean number of wet nights per week at the end of treatment ‐ imipramine + oxybutynin vs imipramine was 0	The mean number of wet nights per week at the end of treatment ‐ imipramine + oxybutynin vs imipramine in the intervention group was 2.1 fewer wet nights (2.99 fewer to 1.21 fewer)	‐	63 (1 RCT)	⊕⊝⊝⊝ VERY LOW 1, 2, 3, 4
Number of wet nights per week at the end of treatment ‐ Amitriptyline + desmopressin vs amitriptyline	The mean number of wet nights per week at the end of treatment ‐ Amitriptyline + desmopressin vs amitriptyline was 0	The mean number of wet nights per week at the end of treatment ‐ Amitriptyline + desmopressin vs amitriptyline in the intervention group was 0 wet nights (1.64 fewer to 1.64 more)	‐	28 (1 RCT)	⊕⊝⊝⊝ VERY LOW 1, 4, 5, 6
Number not achieving 14 consecutive dry nights at the end of treatment ‐ imipramine + oxybutynin vs placebo	Study population		RR 0.43 (0.23 to 0.78)	47 (1 RCT)	⊕⊕⊕⊝ MODERATE 2, 4, 6
	783 per 1000	337 per 1000 (180 to 610)
	Moderate
	783 per 1000	337 per 1000 (180 to 610)
Adverse events assessed as reported in paper			‐	(2 RCTs)	⊕⊕⊕⊝ MODERATE 2, 3	Gastrointestinal effects were reported for imipramine, desmopressin and oxybutynin. However, frequency of adverse effects was higher for the combination therapy
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio
GRADE Working Group grades of evidence High quality: We are very confident that the true effect lies close to that of the estimate of the effect Moderate quality: 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 quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect Outcomes were downgraded one level for serious imprecision (most studies were small and had large confidence intervals), and two levels for very serious study limitations (high risk of bias associated with sequence generation, allocation concealment and high level of drop out).

¹Large numbers of dropouts.
 ²High or unclear risk of bias for random sequence generation and allocation concealment.
 ³No or inadequate blinding.
 ⁴Small sample size.
 ⁵Premature termination of trial.
 ⁶No side effects mentioned.

Background

This is one of seven reviews of interventions for nocturnal enuresis (enuresis or bedwetting). The others focus on:

• desmopressin (Glazener 2002),

• drugs other than desmopressin or tricyclics (Deshpande 2012),

• enuresis alarm therapy (Glazener 2005),

• simple behavioural interventions (Caldwell 2013),

• complex behavioural and educational interventions (Glazener 2004)

• complementary and miscellaneous therapy (Huang 2011).

This review is an update of previous updated reviews by Glazener (Glazener 2000; Glazener 2003) based on the original review by Lister‐Sharp and colleagues at the NHS Centre for Reviews and Dissemination at the University of York, UK (Lister‐Sharp 1997).

Description of the condition

Nocturnal enuresis is the involuntary loss of urine at night, in the absence of organic disease, at an age when a child could reasonably be expected to be dry (by consensus, at a developmental age of five years) (APA 2013; Austin 2014; WHO 1992). Enuresis can affect the child's quality of life and self esteem, and impact on their social, emotional and psychological well being (Deshpande 2011; Hagglof 1998).

With improved understanding about the pathophysiology of nocturnal enuresis and the promotion by the International Children's Continence Society (ICCS) for using standardised nomenclature in paediatric continence research (Austin 2014), enuresis is now subclassified as either monosymptomatic (bedwetting only) or non‐monosymptomatic (bedwetting with lower urinary tract symptoms in the presence or absence of daytime urinary incontinence), as the pathophysiology and treatment for both are different. However, many of the older studies did not distinguish between these groups.

The ICCS has also defined treatment outcomes in terms of initial change in frequency of wetting (non‐response is less than 50% reduction in frequency of wetting, partial response is 50% to 99% reduction, and complete response is 100% reduction), relapse (more than one symptom recurrence per month after cessation of treatment) and duration of treatment success (continued success is no relapse six months after cessation of treatment, and complete success is no relapse after two years) (Austin 2014). Unfortunately, most of the older studies have not assessed these outcomes or were unclear about their definition of success. As this is an update of previous systematic reviews where treatment outcomes were measured in predefined ways which did not use standardised ICCS definitions, we have continued to use the same outcomes for this update that were used in the previous reviews. There were no studies that included two‐year follow‐up and we therefore could not assess long‐term complete success.

Prevalences and causes

Nocturnal enuresis is common in children. Estimating the prevalence of nocturnal enuresis is difficult because there is variation in the diagnostic criteria, with very few studies using ICCS definitions (Austin 2014; De Jonge 1973; Krantz 1994). The generally quoted prevalence rates are 15% to 20% of five‐year‐olds, 7% of seven‐year‐olds, 5% of 10‐year‐olds, 2% to 3% of teenagers and 0.5% to 2% of adults, with a yearly spontaneous remission rate of 14% until adolescence (Blackwell 1989; Forsythe 1974; Rutter 1973). The prevalence of nocturnal enuresis is higher amongst children in residential care (Morgan 1970) and is more persistent in those who have severe enuresis (Yeung 2006).

Butler 2000 proposed the conceptual model that nocturnal enuresis occurs when there is a defective arousal response during sleep to the sensation of a full bladder, a lack of inhibition of bladder emptying during sleep or excessive nocturnal urine production. There are also many known risk factors for nocturnal enuresis, including:

• genetic influences (APA 1980; Bakwin 1971; Bakwin 1973; Eiberg 1995)

• delay in attaining bladder control (Järvelin 1989; Koff 1996)

• physiological factors (Djurhuus 1992; Nørgaard 1993) such as constipation, faecal incontinence, daytime urinary incontinence, caffeine consumption, nocturnal detrusor overactivity (Blackwell 1989; Kawauchi 2003; Loening‐Baucke 1997; McGrath 2008; O'Regan 1986; Sureshkumar 2009)

• sleep apnoea and upper airway obstructive symptoms (Maizels 1993)

• psychological factors (Devlin 1991; Moffatt 1989; Rutter 1973; Shaffer 1977)

It is thought that these risk factors impact on sleep arousal, inhibition of bladder emptying or bladder contraction, and overnight urine production.

Description of the intervention

Interventions

Pharmacological, behavioural and a variety of other interventions have been used to treat enuresis (Neveus 2010). Our review is restricted to interventions that were compared with a tricyclic or a drug related to tricyclics.

Tricyclic drugs and drugs related to tricyclics used for enuresis which we include in this review were amitriptyline, clomipramine, desipramine, imipramine, mianserin, nortriptyline and viloxazine, although some of these are now contraindicated. Tricyclics were compared with placebo, with different doses of the same tricyclic, with another tricyclic or a related drug (clomipramine, mianserin, nortriptyline and viloxazine) or drugs other then tricyclics, such as amphetamine, chlordiazepoxide, desmopressin, diazepam, emepronium, ephedrine sulphate, furosemide, hydroxyzine, meprobamate, oxybutynin, piracetam and tolterodine.

Behavioural interventions that were compared with a tricyclic drug include alarm training, star charts, random waking, three‐step therapy (which includes reassurance, retention control training and waking), counselling and education, shaming, and fluid restriction with avoidance of punishment. Tricyclics were also compared with hypnotherapy, psychotherapy and a restricted diet.

Tricyclic drugs

Tricyclic drugs have been used for treating nocturnal enuresis for over 50 years. They are now considered a third‐line therapy for enuresis.

The tricyclic drug most commonly used for enuresis is imipramine. The recommended age‐/weight‐related dosages range from 25 mg for six‐year‐olds (weight 20 to 25 kg) to 50 to 75 mg for those over 11 years. Imipramine should not be given to children under six years of age and the dose should not exceed 75 mg daily. The maximum period of treatment should not exceed three months and withdrawal of medication needs to be gradual. Electrocardiograph (ECG) monitoring is recommended and a full physical examination should be given before a further course is prescribed (BNF 2002).

Tricyclics have significant adverse effects, including cardiotoxicity, convulsions, and hepatic and haematological reactions, particularly in overdose (Fitzwater 1992; Parkin 1972; Rushton 1993). Serious adverse effects were not reported in any of the included trials except for neutropenia. Minor side effects related to their anticholinergic actions include postural hypotension, dry mouth, constipation, perspiration, tachycardia, nausea, lethargy and insomnia. Other tricyclic drugs (amitriptyline, dothiepin, doxepin and trimipramine) have the same associated risks as imipramine, but have a more sedative effect (BNF 2002). Viloxazine is reported to have a less sedative effect than imipramine, with fewer and milder anticholinergic and cardiovascular side effects (BNF 2002).

Other drugs related to tricyclics

Related drugs included in this review have similar actions to the tricyclics but fewer side effects. Viloxazine is a bicyclic antidepressant and mianserin is a tetracyclic antidepressant. They all inhibit reuptake of noradrenaline, thus increasing functional noradrenaline availability. They have fewer anticholinergic, sedative and cardiotoxic effects than imipramine.

How the intervention might work

Tricyclic drugs act on the central nervous system by blocking the synaptic alpha receptor reuptake of noradrenaline and serotonin into the neurons, thus prolonging their effects. The usual clinical use of tricyclics is as an antidepressant. They also have anticholinergic or antispasmodic effects on the bladder and other parts of the body, local anaesthetic properties and effects on the sleep centre in the brain, which may improve arousability (Warzak 1994). The beneficial effect of imipramine was initially attributed to its anticholinergic effects, but it has been recently shown that imipramine can be helpful even in children who have not responded to anticholinergic therapy (Gepertz 2004), suggesting that another mechanism of action may also contribute to its effectiveness.

Why it is important to do this review

Although enuresis in itself is pathologically benign and has a high rate of spontaneous remission, it can affect the child's quality of life and self esteem, and can impact on their social, emotional and psychological well‐being (Deshpande 2012; Hagglof 1998; Joinson 2007; Von Gontard 2011), persisting into adulthood with long‐term effects (Fitzwater 1992; Yeung 2004). Enuresis affects the whole family, and children with enuresis may experience parental disapproval, sibling teasing and be at risk of increased emotional and physical abuse (Warzak 1992). Consequently it is important to identify effective interventions for treating enuresis, as effective treatment can improve self esteem (Hagglof 1998).

It is also important to assess the effectiveness of tricyclics against other known treatments, both as monotherapy or in combination with other treatments, given the potential risk of side effects for tricyclics compared with other effective treatments.

Objectives

To assess the effects of tricyclic and related drugs compared with other interventions for the treatment of children with nocturnal enuresis.

Methods

Criteria for considering studies for this review

Types of studies

Randomised or quasi‐randomised trials of tricyclics or related drugs versus any treatments for nocturnal enuresis.

Types of participants

Children (as defined by the trialists, usually up to age 16) with nocturnal enuresis.

We excluded studies of adults with enuresis or where participants only had daytime urinary incontinence or had an organic cause for their enuresis. We included studies which included children with daytime urinary incontinence only if the primary focus of the study was enuresis. As it was often difficult to differentiate monosymptomatic and non‐monosymptomatic enuresis, we include both types of enuresis, with an analysis of the subset of those with monosymptomatic enuresis if possible.

Types of interventions

Any trial which used a tricyclic (imipramine, amitriptyline, clomipramine, desipramine, nortripryline, trimipramine) or a related drug (mianserin, viloxazine) in at least one arm of the study. This includes tricyclics combined with another therapy.

Comparison interventions included placebo, different doses of the same tricyclics, other tricyclics or tricyclic‐related drugs, other drugs (amphetamine, desmopressin, diclofenac sodium, diazepam, emepronium, ephedrine sulphate, furosemide, oxybutynin, tolterodine), behavioural interventions (e.g. alarms, random wakening, reward systems such as star charts, bladder exercises, restricting fluids), combinations of other treatments (desmopressin combined with oxybutynin, desmopressin combined with alarm therapy, meprobamate combined with hydrozine, three‐step behavioural programmes combined with motivation) and complementary and miscellaneous therapies (hypnosis, psychotherapy, shaming, restricted diet).

We made the following comparisons:

1. Tricyclic or related drugs versus no treatment or placebo treatment.

2. A higher dose of tricyclic versus a lower dose of the same drug.

3. Tricyclic or related drugs versus treatment with another drug (including desmopressin, a different tricyclic, or another drug).

4. Tricyclic or related drugs versus behavioural interventions (including enuresis alarm therapy).

5. Tricyclic or related drugs versus complementary or miscellaneous interventions.

6. Combination therapy involving tricyclic drugs versus another treatment.

Types of outcome measures

We based the outcomes considered in this update on outcomes used in previous versions of the review.

Primary outcomes

• mean number of wet nights per week (over the previous two weeks) at the end of treatment;

• number of participants failing to attain 14 consecutive dry nights by the end of treatment (which is equivalent to the number failing to achieve a complete response).

Secondary outcomes

• mean number of wet nights per week (over the previous two weeks) when participants were followed up after treatment had ceased;

• number failing to attain 14 consecutive dry nights or subsequently relapsing after cessation of treatment (which is equivalent to the number failing to achieve a complete response or relapsing at follow‐up);

• adverse effects.

Search methods for identification of studies

We did not impose any language or other restrictions on any of the searches.

Electronic searches

This review drew on the search strategy developed for the Cochrane Incontinence Group. We identified relevant trials from the Cochrane Incontinence Group Specialised Trials Register. For more details of the search methods used to build the Specialised Register please see the Group's module in the Cochrane Library. The Register contains trials identified from the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, and MEDLINE in process, ClinicalTrials.gov, WHO ICTRP and handsearching of journals and conference proceedings. Most of the trials in the Cochrane Incontinence Group Specialised Register are also contained in CENTRAL. The date of the last search was: 30 November 2015.

The terms used to search the Incontinence Group Specialised Register are given below:
 (TOPIC.URINE.ENURESIS*)
 AND
 ({DESIGN.CCT*} OR {DESIGN.RCT*})

All searches were of the keyword field of Reference Manager 2012.

Searching other resources

The review authors searched the reference lists of relevant articles. We also searched the reference list of a previous systematic review of enuresis treatments (Lister‐Sharp 1997).

Data collection and analysis

We assessed the studies for this review using the methods described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

Selection of studies

At least two review authors checked the titles and, where possible, abstracts of all studies located by the searches, to identify those likely to be evaluations of the effects of interventions for nocturnal enuresis. We then obtained full papers and assessed them to identify studies that met the inclusion criteria.

Data extraction and management

We extracted the data using a standard form, and processed the included trial data as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). Where appropriate, we converted the results to:

• the mean and standard deviation (SD) of the number of wet nights per week, or

• the number of children failing to achieve complete response during treatment, defined as 14 consecutive dry nights at the end of the treatment period, or

• the mean number of wet nights per week at follow‐up (after stopping treatment), or

• the number who did not achieve complete response at the end of treatment, plus those who relapsed after stopping active treatment at follow‐up (i.e. those who were wet at follow‐up after completion of treatment).

Where a mean value was reported with no standard deviation, and it was not possible to use the other available information to estimate the SD, we entered the data into into 'Other Data' tables'.

Assessment of risk of bias in included studies

We assessed standard 'Risk of bias' items according to the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) for each study. These include:

• random sequence generation;

• allocation concealment;

• blinding of study participants;

• blinding of study personnel;

• blinding of outcome assessors;

• incomplete outcome data addressed;

• freedom from selective reporting;

• freedom from other bias (e.g. whether children with daytime wetting or organic causes of their enuresis were specifically excluded, whether there were baseline differences between groups, whether there was a washout period in cross‐over trials, and whether appropriate statistical techniques were used).

Measures of treatment effect

We intended, where possible, to calculate standardised effect sizes and 95% confidence intervals (CIs); mean differences (MDs) where outcomes were continuous variables and risk ratios (RRs) where they were binary. We calculated the mean differences by the inverse of the variance, and gave them as differences in the number of wet nights per week. Negative values indicate fewer wet nights in the tricyclic group.

Unit of analysis issues

The unit of allocation was the individual participant.

We have indicated cross‐over trials by a suffix hash (#) after the trial ID. For cross‐over trials, we would have calculated data from paired comparisons to provide information for generic inverse variance analysis, but none of the trials provided suitable data. Rather than put the data available in 'Other Data' tables we chose to tabulate the data as if they were parallel groups. This approach leads to wider confidence intervals; hence statistically significant results using a parallel‐group approach would be significant if correctly analysed as cross‐over studies, but nonsignificant results might hide a significant difference. We conducted a sensitivity analysis excluding them, to check whether their inclusion materially changed the findings. However, if data were provided from the first arm of the trial, we included them as if the data came from parallel groups; in this case we did not use the suffix hash and did not exclude them in the sensitivity analysis.

Dealing with missing data

When intention‐to‐treat analysis was not mentioned in the paper, or if some participants were not analysed in the group to which they were randomised, we looked for sufficient information to restore them to the correct group. If we could not analyse participants in their allocated groups, we reported this in the text ('Risk of bias' table and in the 'Risk of bias' section of the text).

Where there were missing outcome data in the included studies, the primary analysis used the number of participants with available data as the denominator (i.e. 'available case' analysis). We also attempted to contact the authors. If there was evidence of differential dropouts from the groups which may have been caused by adverse effects of the interventions or ineffectiveness, we mentioned this in the 'Risk of bias' table. We derived standard deviations from standard errors, the CI, t values or P values that related to the differences between means in two groups, according to the Cochrane Handbook for Systematic Reviews of Interventions Section 7.7.3.3 (Higgins 2011).

Assessment of heterogeneity

We further investigated differences between trials when we found statistically significant heterogeneity (at the 10% probability level or using the I² statistic > 30% (Higgins 2003)) or when it appeared obvious from visual inspection of the results. If there was no obvious reason for the heterogeneity, or if it persisted despite the removal of outlying trials, we preferred a random‐effects model.

Assessment of reporting biases

We investigated the possibility of publication bias and related biases if we conducted any meta‐analyses with 10 or more studies.

Data synthesis

We performed meta‐analysis according to the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) if there were enough trials for each comparison. We used a fixed‐effect model to calculate the pooled estimates and the 95% CIs (Berlin 1989). If there was substantial heterogeneity (e.g. I² > 50%), assuming treatment effects differed between trials, we used the random‐effects method to produce an overall summary estimate.

Subgroup analysis and investigation of heterogeneity

We had planned to conduct an analysis of the subset of participants with monosymptomatic enuresis, if data were available.

If there was heterogeneity, we had planned to investigate it using subgroup analysis restricted to primary outcomes, and sensitivity analysis.

Sensitivity analysis

If necessary, we analysed data using both random‐effects and fixed‐effect models. Where appropriate, we explored the possible effects of missing data in a sensitivity analysis (for example, assuming the 'worst‐case' and 'best‐case' scenarios for all participants with missing outcome data).

Results

Description of studies

Results of the search

The literature search produced 446 abstracts and titles to assess. We assessed 145 full‐text articles: there were 73 reports of 64 included studies, and 65 reports of 58 studies that we excluded. Additionally, seven studies are awaiting classification ). The flow of literature through this assessment process is shown in the PRISMA flowchart (Figure 1).

1.

PRISMA study flow diagram

Included studies

Design

This review includes 64 randomised or quasi‐randomised trials of tricyclics or related drugs versus any comparable treatments for nocturnal enuresis. Double‐blind cross‐over studies are flagged with a hash suffix (#) in the study ID. Cross‐over studies which had inadequate washout (and were therefore analysed as parallel‐group trials) were not flagged with a hash.

Sample size

The sample size of the studies ranged from nine (Treffert 1964#) to 298 (Forsythe 1969) participants. There were 30 studies with 50 or fewer participants, 20 with 51 to 99 participants and 13 studies with 100 or more participants, plus one study in Korean where we are awaiting feedback from the authors about the number of participants in the study (Kang 2003).

Setting

Seven trials were based in institutional settings (Alderton 1967#; Alderton 1970#; Harrison 1970; Khorana 1972; Roy 1970; Thomsen 1967#; Treffert 1964#), one of which included some children with low IQs (Alderton 1970#). One other trial involved children with learning difficulties (Kumazawa 1990). Two trials were known to have been conducted in the community (Danquah 1975; Lake 1968); one admitted children to hospital for trial treatment (Haegglund 1964), and the remainder were based on referrals to hospital outpatient clinics or gave no information about setting. There were insufficient studies providing useable data of children from institutional settings or with learning difficulties to do a subgroup analysis.

Participants

This review includes 4071 children (mean study size n = 66, range 7 to 298), in which 2269 children received a tricyclic. Most of the trials included children aged between five and 16. Two trials included some older participants (Seo 2001; Liederman 1969) and nine trials included a few children under the age of five (Forsythe 1969; Forsythe 1972a; Friday 1966; Haegglund 1964; Lines 1968#; Miyazaki 1973; Moltke 1979; Schröder 1971; Shah 1971#). Only one trial (which did not have useable data) included children aged three to five, while the rest included children with a wide range of ages, and therefore subgroup analysis by age was not possible.

Interventions

Eight different types of tricyclics (imipramine, amitriptyline, clomipramine, desipramine, nortripryline, trimipramine) or related drugs (mianserin, viloxazine) were used in studies included in this review. They were compared with placebos, with other drugs (including a different dose of the same tricyclic, other tricyclics or related drugs, or other classes of drug), behavioural interventions, and complementary and other miscellaneous treatments.

Tricyclic and related drugs

The most common tricyclic tested was imipramine:

• 51 studies compared imipramine with another treatment (Agarwala 1968#; Alderton 1967#; Alderton 1970#; Attenburrow 1984; Banerjee 1993; Batislam 1995; Bhatia 1990; Bindelglas 1968; Ciotti 1983; Drew 1966; Esmaeili 2008; Forsythe 1969; Fournier 1987; Friday 1966; Haegglund 1964; Harrison 1970; Hoashi 1995; Hodes 1973; Holt 1986; Iester 1991; Ingle 1968; Kang 2003; Khorana 1972; Kolvin 1972; Kumazawa 1990; Kunin 1970#; Lee 2005; Manhas 1967; Martin 1971#; Maxwell 1971#; McKendry 1975; Moltke 1979; Motavalli 1994; Naitoh 2005; Netley 1984; Neveus 2008#; Petersen 1974#; Poussaint 1965a; Roy 1970; Schröder 1971; Seo 2001; Shaffer 1968#; Smellie 1996; Tahmaz 2000; Thomsen 1967#; Treffert 1964#; Vertucci 1997; Wagner 1982; Ye 2001; Yurdakok 1986; Yurdakok 1987).

Other tricyclics studied included:

• amitriptyline (Burke 1995; Danquah 1975; Forsythe 1972b; Lines 1968#; Mehrotra 1980; Miyazaki 1973; Poussaint 1966a; Poussaint 1966b; Shah 1971#);

• nortriptyline (Forsythe 1969; Lake 1968; Scholander 1968);

• clomipramine (Motavalli 1994);

• desipramine (Liederman 1969);

• trimipramine (Forsythe 1972a).

Drugs related to tricyclics studied included mianserin (Smellie 1996) and viloxazine (Attenburrow 1984; Yurdakok 1987).

Combination therapy involving tricyclics

Combination treatments involving tricyclics in this study were

• imipramine and oxybutynin (Esmaeili 2008; Tahmaz 2000);

• imipramine and desmopressin (Seo 2001),

• amitriptyline and desmopressin (Burke 1995);

• imipramine and diclofenac sodium (Batislam 1995);

• amitriptyline and chlordiazepoxide (Forsythe 1972b);

• imipramine and alarm training (Fournier 1987; Naitoh 2005);

• nortriptyline and alarm training (Scholander 1968);

• imipramine and simple behavioural therapy (Bhatia 1990; Kumazawa 1990);

• amitriptyline and simple behavioural therapy (Mehrotra 1980).

Comparisons

Control

A tricyclic or tricyclic‐related drug was compared with placebo (Agarwala 1968#; Alderton 1967#; Alderton 1970#; Attenburrow 1984; Batislam 1995; Bindelglas 1968; Drew 1966; Forsythe 1969; Forsythe 1972a; Forsythe 1972b; Fournier 1987; Friday 1966; Harrison 1970; Hodes 1973; Khorana 1972; Kolvin 1972; Lake 1968; Liederman 1969; Lines 1968#; Manhas 1967; Martin 1971#; Maxwell 1971#; Miyazaki 1973; Moltke 1979; Neveus 2008#; Petersen 1974#; Poussaint 1965a; Poussaint 1966a; Poussaint 1966b; Roy 1970; Schröder 1971; Shaffer 1968#; Shah 1971#; Smellie 1996; Tahmaz 2000; Thomsen 1967#; Treffert 1964#) or with no treatment (Haegglund 1964; Wagner 1982) in 39 studies.

Other drugs

Other drugs which were compared with tricyclics included:

• desmopressin (Burke 1995; Hoashi 1995; Holt 1986; Kang 2003; Lee 2005; Naitoh 2005; Seo 2001; Vertucci 1997; Ye 2001);

• the anticholinergics oxybutynin (Esmaeili 2008; Lee 2005; Tahmaz 2000), tolterodine (Neveus 2008#) and emepronium (Petersen 1974#);

• psychoactive drugs such as chlordiazepoxide (Yurdakok 1986; Forsythe 1972b), meprobamate with hydroxyzine (Ingle 1968) and piracetam (Yurdakok 1986);

• other drugs such as diclofenac sodium (an anti‐inflammatory drug) (Batislam 1995); ephedrine (a stimulant) (Kunin 1970#) and furosemide (a diuretic) (Moltke 1979).

Behavioural interventions

• Behavioral interventions compared with a tricyclic or related drug include alarm therapy and other behavioural therapies. Of the alarm comparisons: pad‐and‐bell (bed) alarms were used in three trials (Kolvin 1972; Scholander 1968; Wagner 1982);

• body‐worn alarms in two trials (Fournier 1987; Naitoh 2005);

• electric shock alarms in another two trials (McKendry 1975; Netley 1984).

The other behavioral therapies studied were

• simple behavioural interventions such as restricting fluids and avoiding punishments (Bhatia 1990);

• random or alarm‐clock wakening (Fournier 1987; Iester 1991; Mehrotra 1980);

• bladder training (retention control) (Iester 1991; Kumazawa 1990);

• motivational counselling or rewards (Iester 1991; Maxwell 1971#; Mehrotra 1980);

• computer‐assisted education (Iester 1991).

Complementary and miscellaneous interventions

Complementary and miscellaneous interventions included:

• restricted diet (McKendry 1975);

• hypnosis (Banerjee 1993);

• psychotherapy (Ciotti 1983);

• traditional ritual shaming (in a trial based in a Ghanaian fishing community (Danquah 1975)).

New studies found for this update

This is the fourth update of this review. The original version (Lister‐Sharp 1997) included 23 studies of tricyclics. Ten new trials were included in the first update (Glazener 2003) (Alderton 1967#; Alderton 1970#; Bhatia 1990; Forsythe 1969; Haegglund 1964; Hodes 1973; Scholander 1968; Tahmaz 2000; Vertucci 1997; Wagner 1982), two new trials were added for the second (minor) update (Ciotti 1983; Hoashi 1995) and another two for the third (minor) update (Lee 2005; Miyazaki 1973), with the addition of data from another 21 trials which were initially included only in a sensitivity analysis in the first update.

In this fourth update, we have added six new trials (Esmaeili 2008; Naitoh 2005; Neveus 2008#; Seo 2001; Kang 2003; Ye 2001). In this update we have also assessed the effect of combination therapy involving tricyclics (Batislam 1995; Bhatia 1990; Burke 1995; Esmaeili 2008; Forsythe 1972b; Fournier 1987; Kumazawa 1990; Mehrotra 1980; Naitoh 2005; Scholander 1968; Seo 2001; Tahmaz 2000). Details are given in the Characteristics of included studies tables.

Excluded studies

We excluded 58 studies. Of those:

• 46 were non‐randomised (Abrams 1963; Alison 1973; Arai 1971; Butler 2001; Cupalova‐Naglova 1970; D'Hollander 1967; De Jonge 1972; Esperanca 1969; Fisher 1963; Fritz 1994; Heising 1978; Höfler 1978; Jorgensen 1980; Kales 1977; Kardash 1968; Korczyn 1979; Kurokawa 1963; Lake 1979; Laybourne 1968; Libert 1991; Manglick 1992; Mariuz 1963; Meijer 1965; Miller 1968; Mishra 1980; Monda 1995; Nigam 1973; Noack 1964; Philpott 1970; Polak 1981; Porot 1970; Poussaint 1965b; Rapoport 1980; Rett 1968; Russell 1992; Saeedalzakerin 2000; Schjetne 1970; Simeon 1981; Site 1974; Smith 1967; Soulayrol 1970; Steinicke 1971; Tamburello 1971; Werry 1975; Werry 1977; Yamanishi 1988), including one which was both not an RCT and not about enuresis (Lake 1979);

• three other studies which were not about nocturnal enuresis (Al‐Waili 2000; Meadow 1982; Valentine 1968);

• seven were in adults (Alessi 1992; Blackman 1964; Dorison 1962; Hicks 1964; Kelly 1974; Milner 1968; Tong 2001);

• two reported unreliable data because children were transferred between groups (Kim 2001; McConaghy 1969).

A full list and details of the excluded studies are given in the Characteristics of excluded studies tables.

Studies awaiting classification

There are seven studies awaiting classification:

• one report of one study was unobtainable (Cheng 2008);

• four studies gave insufficient information to determine eligibility; we are hoping to obtain further information from the authors (Chertin 2007; GP Research Grp 1969; Hoseinzadeh 1997 NEW; Ice 1966);

• one study is awaiting translation (Cupalova‐Naglova 1968);

• one study is unpublished, with details being sought from the trialists (Houts 2003).

Risk of bias in included studies

There was insufficient information to judge the risk of bias in most studies, particularly in the older studies (Figure 2; Figure 3).

2.

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

3.

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

Allocation

Adequate sequence generation

Four studies demonstrated adequate random sequence generation (Agarwala 1968#; Lines 1968#; Martin 1971#; Shaffer 1968#) and seven studies were quasi‐randomised and therefore had inadequate sequence generation (Banerjee 1993; Esmaeili 2008; Haegglund 1964; Ingle 1968; Manhas 1967; Naitoh 2005; Poussaint 1965a). There was insufficient information in the remaining studies to judge whether random sequence generation occurred.

Concealment of allocation to groups

We judged seven studies to have used adequate methods of concealment of allocation (Burke 1995; Kunin 1970#; Liederman 1969; Poussaint 1966a; Poussaint 1966b; Shah 1971#; Smellie 1996) and six studies had inadequate concealment of allocation to groups (Banerjee 1993; Haegglund 1964; Ingle 1968; Manhas 1967; Naitoh 2005; Poussaint 1965a).There was insufficient information in the remaining studies to judge the adequacy of concealment of allocation to treatment groups.

Blinding

Blinding of participants

We judged 39 studies to be adequately blinded, including 34 which used placebo, with participants not aware of which intervention they received (Agarwala 1968#; Alderton 1967#; Alderton 1970#; Attenburrow 1984; Bindelglas 1968; Burke 1995; Drew 1966; Forsythe 1969; Forsythe 1972a; Forsythe 1972b; Fournier 1987; Friday 1966; Harrison 1970; Hoashi 1995; Hodes 1973; Holt 1986; Khorana 1972; Kumazawa 1990; Kunin 1970#; Lake 1968; Liederman 1969; Lines 1968#; Manhas 1967; Martin 1971#; Maxwell 1971#; Miyazaki 1973; Moltke 1979; Neveus 2008#; Petersen 1974#; Poussaint 1965a; Poussaint 1966a; Poussaint 1966b; Scholander 1968; Schröder 1971; Shaffer 1968#; Shah 1971#; Smellie 1996; Thomsen 1967#; Treffert 1964#).

Participants were not blinded in 17 of the studies (Banerjee 1993; Bhatia 1990; Ciotti 1983; Danquah 1975; Esmaeili 2008; Haegglund 1964; Iester 1991; Lee 2005; McKendry 1975; Mehrotra 1980; Motavalli 1994; Naitoh 2005; Roy 1970; Vertucci 1997; Wagner 1982; Ye 2001; Yurdakok 1986), and there was insufficient information to judge whether study participants were blinded in the remaining eight studies.

Blinding of study personnel

We judged 37 studies to have study personnel blinded to treatment allocation, including studies which reported "double blinding" (Agarwala 1968#; Alderton 1967#; Alderton 1970#; Attenburrow 1984; Bindelglas 1968; Burke 1995; Drew 1966; Forsythe 1969; Forsythe 1972a; Forsythe 1972b; Fournier 1987; Friday 1966; Harrison 1970; Hoashi 1995; Hodes 1973; Holt 1986; Kunin 1970#; Lake 1968; Liederman 1969; Lines 1968#; Manhas 1967; Martin 1971#; Maxwell 1971#; Miyazaki 1973; Moltke 1979; Neveus 2008#; Petersen 1974#; Poussaint 1965a; Poussaint 1966a; Poussaint 1966b; Scholander 1968; Schröder 1971; Shaffer 1968#; Shah 1971#; Smellie 1996; Thomsen 1967#; Treffert 1964#).

Study personnel were not blinded in 17 of the studies (Banerjee 1993; Bhatia 1990; Ciotti 1983; Danquah 1975; Esmaeili 2008; Haegglund 1964; Iester 1991; Lee 2005; McKendry 1975; Motavalli 1994; Naitoh 2005; Roy 1970; Tahmaz 2000; Vertucci 1997; Wagner 1982; Ye 2001; Yurdakok 1986) and there was insufficient information to judge whether study personnel were blinded in the remaining 10 studies.

Blinding of outcome assessors

When not stated, we assumed outcome assessors were parents who recorded the child's treatment response. We judged 35 studies to have adequate blinding of outcome assessors to treatment allocation (often because the participants were blinded) (Agarwala 1968#; Alderton 1967#; Alderton 1970#; Attenburrow 1984; Bindelglas 1968; Burke 1995; Drew 1966; Forsythe 1969; Forsythe 1972a; Forsythe 1972b; Fournier 1987; Friday 1966; Harrison 1970; Hoashi 1995; Hodes 1973; Holt 1986; Kunin 1970#; Lake 1968; Lines 1968#; Martin 1971#; Maxwell 1971#; Miyazaki 1973; Moltke 1979; Neveus 2008#; Petersen 1974#; Poussaint 1965a; Poussaint 1966a; Poussaint 1966b; Scholander 1968; Schröder 1971; Shaffer 1968#; Shah 1971#; Smellie 1996; Thomsen 1967#; Treffert 1964#).

Outcome assessors were not blinded in 16 of the studies (Banerjee 1993; Ciotti 1983; Danquah 1975; Esmaeili 2008; Haegglund 1964; Iester 1991; Lee 2005; McKendry 1975; Motavalli 1994; Naitoh 2005; Roy 1970; Tahmaz 2000; Vertucci 1997; Wagner 1982; Ye 2001; Yurdakok 1986) and there was insufficient information to judge whether outcome assessors were blinded in the remaining 13 studies.

Incomplete outcome data

Dropouts

In 26 trials incomplete outcome data were adequately addressed with results given for all randomised participants (Agarwala 1968#; Batislam 1995; Ciotti 1983; Drew 1966; Hodes 1973; Iester 1991; Kumazawa 1990; Martin 1971#; Mehrotra 1980; Naitoh 2005; Scholander 1968; Seo 2001; Treffert 1964#; Yurdakok 1986; Yurdakok 1987) or the number and reason for dropouts reported (Forsythe 1969; Forsythe 1972a; Forsythe 1972b; Friday 1966; Haegglund 1964; Harrison 1970; Lines 1968#; Neveus 2008#; Petersen 1974#; Poussaint 1965a; Vertucci 1997).

In most trials, there did not seem to be differential dropout rates, with dropouts mostly due to children failing to meet the inclusion criteria after baseline monitoring or due to lack of follow‐up. Some studies reported that the children who dropped out were comparable between the groups.

However, in 21 trials, there was high risk of attrition bias with dropouts not reported or reasons for dropping not given (Alderton 1970#; Attenburrow 1984; Bhatia 1990; Bindelglas 1968; Danquah 1975; Esmaeili 2008; Kang 2003; Lee 2005; Liederman 1969; Maxwell 1971#; Miyazaki 1973; Moltke 1979; Netley 1984; Poussaint 1966a; Poussaint 1966b; Roy 1970; Schröder 1971; Shaffer 1968#; Shah 1971#; Tahmaz 2000), or there were clearly differential dropouts between groups resulting from allocation to an unacceptable treatment (diet and electric shock alarm (McKendry 1975)). For this last trial, dropouts were counted as failures in the analysis. In the remaining 17 trials there was insufficient information to judge whether incomplete outcome data were adequately addressed.

Where possible, we have reported data according to known outcomes.

Follow‐up

Only 18 of the trials followed the participants for at least three months after completion of treatment (Banerjee 1993; Batislam 1995; Bhatia 1990; Bindelglas 1968; Burke 1995; Danquah 1975; Fournier 1987; Haegglund 1964; Iester 1991; Kolvin 1972; Martin 1971#; McKendry 1975; Mehrotra 1980; Moltke 1979; Scholander 1968; Seo 2001; Tahmaz 2000; Ye 2001). Follow‐up is not appropriate for the 20 cross‐over trials because the children were exposed to both interventions, so follow‐up outcomes could not be ascribed to any one treatment.

Selective reporting

We assessed for selective reporting bias in studies by examining whether systematic baseline measurements of enuresis were conducted prior to the start of treatment, whether adverse events were mentioned, whether organic causes of enuresis or daytime urinary incontinence were excluded, whether full data sets were provided, whether study protocols were provided, and whether all prespecified outcomes were reported.

• Only 13 studies had low risk of selective reporting bias with adequate reporting of outcomes (Attenburrow 1984; Forsythe 1969; Fournier 1987; Harrison 1970; Hodes 1973; Kunin 1970#; Lake 1968; Miyazaki 1973; Moltke 1979; Naitoh 2005; Tahmaz 2000; Treffert 1964#; Vertucci 1997).

• Ten studies had high risk of selective reporting bias (Batislam 1995; Burke 1995; Danquah 1975; Friday 1966; Haegglund 1964; Kang 2003; Lee 2005; McKendry 1975; Mehrotra 1980; Schröder 1971), either because prespecified outcomes were not reported or systematic baseline measurements of wetting, adverse events, exclusion of organic causes or daytime urinary incontinence or full data sets were not provided.

• In the remaining 41 studies there was insufficient information to judge whether there was any selective reporting bias.

Baseline comparability of groups

Amongst the trials which were not cross‐overs, two reported that there were baseline differences between the groups (Netley 1984; Roy 1970). Neither gave details of the method of concealment of allocation to groups. The remaining trials either reported that the groups were comparable at baseline or did not report this information.

Adverse effects

Adverse effects in the studies included: dizziness, headache, sweating, drowsiness/lethargy, sleep disturbance/restlessness, apathy/depression, sullenness, tremor, flu, upset stomach/abdominal pain, vomiting/diarrhoea, constipation, dry mouth, rash, nausea, neutropenia, burning sensation, weight loss/anorexia, anxiety, scleral injection and cold extremities. Adverse events were reported in 30 trials (Agarwala 1968#; Attenburrow 1984; Batislam 1995; Ciotti 1983; Danquah 1975; Forsythe 1969; Friday 1966; Hoashi 1995; Hodes 1973; Holt 1986; Khorana 1972; Lake 1968; Lee 2005; Liederman 1969; Manhas 1967; Martin 1971#; Maxwell 1971#; McKendry 1975; Mehrotra 1980Miyazaki 1973; Petersen 1974#; Poussaint 1965a; Poussaint 1966a; Poussaint 1966b; Scholander 1968; Schröder 1971; Seo 2001; Shaffer 1968#; Tahmaz 2000; Vertucci 1997), but none of the side effects were of major consequence.

Some trials did not give details of the nature, severity or quantity of the side effects, some noted that they occurred on placebo treatment and some failed to indicate in which treatment arm the side effects occurred. Others noted that behavioural disturbances, in particular, might have been present before treatment began, but that treatment appeared to exacerbate symptoms.

Seven trials reported that there were no adverse effects (Bindelglas 1968; Burke 1995; Drew 1966; Forsythe 1972a; Kumazawa 1990; Thomsen 1967#; Ye 2001). The remaining trials gave no information about side effects.

Daytime wetting, organic causes and baseline recording of wetting

Children with daytime wetting were specifically excluded in only 10 trials (Ciotti 1983; Esmaeili 2008; Hoashi 1995; Holt 1986; Kunin 1970#; Naitoh 2005; Neveus 2008#; Tahmaz 2000; Wagner 1982; Haegglund 1964). In one trial (Lee 2005), although children with daytime wetting were included in the study, the author was able to provide us with data for the monosymptomatic group (night wetting only) which we used in this review. Eight trials stated that they included some children with daytime wetting (Attenburrow 1984; Bhatia 1990; Friday 1966; Lines 1968#; McKendry 1975; Miyazaki 1973; Petersen 1974#; Shaffer 1968#), and the remaining trials did not mention whether or not children with daytime wetting were included.

Eleven studies included children with organic causes for their enuresis (Alderton 1967#; Alderton 1970#; Ciotti 1983; Drew 1966; Forsythe 1969; Harrison 1970; Hodes 1973; Lake 1968; Petersen 1974#; Shaffer 1968#; Treffert 1964#) and 51 studies stated that they excluded children with organic causes for their enuresis. Two trials did not mention whether or not organic causes were excluded (Kang 2003; Mehrotra 1980).

Quality of data

The reports of 10 trials did not provide useable data (Banerjee 1993; Batislam 1995; Bindelglas 1968; Ciotti 1983; Friday 1966; Shaffer 1968#; Shah 1971#; Thomsen 1967#; Yurdakok 1986; Yurdakok 1987) and we therefore only included their information in the Characteristics of included studies tables. Twenty‐one trials which provided continuous data failed to provide SDs (Alderton 1967#; Alderton 1970#; Danquah 1975; Drew 1966; Fournier 1987; Harrison 1970; Kolvin 1972; Lake 1968; Lines 1968#; Martin 1971#; Moltke 1979; Naitoh 2005; Petersen 1974#; Poussaint 1966a; Poussaint 1966b; Roy 1970; Smellie 1996; Seo 2001; Treffert 1964#; Wagner 1982; Treffert 1964#) and we therefore reported their data without standard deviations in the 'Other Data' tables.

Other potential sources of bias

Cross‐over trials

For cross‐over trials, we would have calculated data from paired comparisons to provide information for generic inverse variance analysis, but none of the trials provided suitable data. We chose instead to tabulate the data as if they were parallel groups. This approach leads to wider confidence intervals; hence statistically significant results using a parallel‐group approach would be significant if correctly analysed as cross‐over studies, but nonsignificant results might hide a significant difference. There were 20 cross‐over studies. Thirteen trials were reported to be double‐blind randomised cross‐over trials, and are identified by a suffix hash (#) after the trial ID (Agarwala 1968#; Alderton 1967#; Alderton 1970#; Kunin 1970#; Lines 1968#; Martin 1971#; Maxwell 1971#; Neveus 2008#; Petersen 1974#; Shaffer 1968#; Shah 1971#; Thomsen 1967#; Treffert 1964#).

• Only one of these trials had an adequate washout phase between the two arms of the trial (Kunin 1970#).

• In another two trials, interspersed placebo arms could be regarded as washout phases (Alderton 1967#; Alderton 1970#).

• In one study, some children stopped treatment early during the second phase, mostly due to significant improvement on amitriptyline (Lines 1968#), which potentially biased the results.

The remaining seven cross‐over trials had inadequate washout between interventions, and we therefore used data from the first arm of the trial before crossing over to the alternative arm, and analysed them as parallel‐group trials, but with only half the participant numbers (Drew 1966; Harrison 1970; Khorana 1972; Miyazaki 1973; Poussaint 1965a; Poussaint 1966a; Vertucci 1997). In this group we did not apply the suffix hash to the study ID.

Previous treatment

In four trials, children had not received previous treatment for their enuresis (Batislam 1995; Iester 1991; Shah 1971#; Wagner 1982). In 16 trials some or all of the children had failed previously with other methods (Agarwala 1968#; Friday 1966; Ingle 1968; Kumazawa 1990; Lee 2005; Lines 1968#; Miyazaki 1973; Naitoh 2005; Neveus 2008#; Petersen 1974#; Poussaint 1965a; Scholander 1968; Schröder 1971; Thomsen 1967#; Yurdakok 1986; Yurdakok 1987). Information about previous treatment was not given in the remainding trials.

Effects of interventions

See: Table 1; Table 2; Table 3; Table 4

We describe the effects of the interventions under six categories:

1. Tricyclics versus placebo

2. Different doses of the same tricyclics

3. Tricyclics versus another drug

4. Tricyclics versus behavioural interventions (including alarm therapy)

5. Tricyclics versus complementary or miscellaneous interventions

6. Combination therapy involving a tricyclic versus another intervention

All reported results are fixed‐effect model unless explicitly stated otherwise.

1. Tricyclic and related drugs compared with placebo (Analysis 1, 'Other Data' tables 1)

We identified 32 trials which provided data comparing a tricyclic drug with placebo. However, data from 16 of these failed to provide SDs and we have therefore reported them in 'Other Data' tables only, and five others failed to provide useable numerical data.

The following tricyclics were used:

• Imipramine was compared with placebo in 28 trials (Agarwala 1968#; Alderton 1967#; Alderton 1970#; Attenburrow 1984; Drew 1966; Forsythe 1969; Forsythe 1972a; Fournier 1987; Haegglund 1964; Harrison 1970; Hodes 1973; Kolvin 1972; Khorana 1972; Manhas 1967; Martin 1971#; Maxwell 1971#; Moltke 1979; Neveus 2008#; Petersen 1974#; Poussaint 1965a; Poussaint 1966a; Poussaint 1966b; Roy 1970; Schröder 1971; Smellie 1996; Tahmaz 2000;Treffert 1964#; Wagner 1982), and compared to no treatment in one trial (Roy 1970). Imipramine‐N‐oxide (a metabolite of imipramine) was compared to placebo in one trial (Petersen 1974#);

• Amitriptyline was compared to placebo in four trials (Lines 1968#; Miyazaki 1973; Poussaint 1966a; Poussaint 1966b);

• Nortriptyline was compared to placebo in two trials (Forsythe 1969; Lake 1968);

• Desipramine was compared to placebo in one trial (Liederman 1969);

• Drugs related to tricyclics (viloxazine (Attenburrow 1984) and mianserin (Smellie 1996)) were compared to placebo in one trial each

The mean number of wet nights per week at the end of treatment was generally lower for children on tricyclics compared to placebo.

During treatment

Mean wet nights

Most studies compared imipramine with placebo, and showed about one fewer wet nights per week for imipramine (MD ‐0.95, 95% CI ‐1.40 to ‐0.50; participants = 347; studies = 4; I² = 0%, Analysis 1.1.1). Three of the four were cross‐over studies (Agarwala 1968#; Maxwell 1971#; Neveus 2008#) and excluding them left one study (Attenburrow 1984) where the difference in wet nights did not reach statistical significance (MD ‐2.50, 95% CI ‐5.74 to 0.74). There was also a single study which showed viloxazine to be superior to placebo with about three fewer wet nights for viloxazine (MD ‐3.10, 95% CI ‐5.64 to ‐0.56; participants = 24; Analysis 1.1.2).

1.1. Analysis.

Comparison 1 TRICYCLIC DRUG vs PLACEBO, Outcome 1 Number of wet nights per week at the end of treatment.

Sixteen studies of imipramine compared to placebo reported mean wet nights at the end of treatment, but did not provide standard deviations. These are included in the 'Other Data' tables (Analysis 1.2.1).

1.2. Analysis.

Comparison 1 TRICYCLIC DRUG vs PLACEBO, Outcome 2 Number of wet nights per week at the end of treatment (no SDs).

Number of wet nights per week at the end of treatment (no SDs)
Study	Intervention 1	Intervention 2
imipramine vs placebo
Alderton 1967#	Mean = 3.43, n = 12	Mean = 4.53, n = 12
Alderton 1970#	Mean = 2.17, n = 7	Mean = 4.22, n = 7
Drew 1966	Mean = 2.38, n = 11	Mean = 3.79, n = 11
Fournier 1987	Mean = 1.9, n = 8	Mean = 5, n = 8
Harrison 1970	Mean = 2.52, n=30	Mean = 3.3, n = 32
Kolvin 1972	Mean = 2.3, n = 35	Mean = 2.7, n = 27
Martin 1971#	25 mg dose: Mean = 2.83, n = 57 10 mg dose: mean = 3.69, n = 57	Mean = 4.52, n = 57
Moltke 1979	Mean = 3.43, n = 43	Mean = 4.69, n = 44
Petersen 1974#	Mean = 2.4, n = 61	Mean = 4.37, n = 61
Poussaint 1965a	Mean = 2.9, n = 13	Mean = 4.8, n = 13
Poussaint 1966a	Mean = 3.1, n = 16	Mean = 4.6, n = 16
Poussaint 1966b	Mean = 4.1, n = 9	Mean = 5.5, n = 9
Roy 1970	Mean = 1.8, n = 14	Mean = 2, n = 6
Smellie 1996	Mean = 2, n = 25	Mean = 4.5, n = 29
Treffert 1964#	Mean = 1.86, n = 9	Mean = 2.36, n = 9
Wagner 1982	Mean = 2.75, n = 12	Mean = 4.26, n = 12
imipramine vs no treatment control
Roy 1970	Mean = 1.8, n = 14	Mean = 2.6, n = 6
imipramine‐N‐oxide vs placebo
Petersen 1974#	Mean = 3.24, n = 61	Mean = 4.37, n = 61
amitriptyline vs placebo
Lines 1968#	Mean = 2.33, n = 36	Mean = 3.51, n = 36
Poussaint 1966a	Mean = 3.1, n = 16	Mean = 4.6, n = 16
Poussaint 1966b	Mean = 4.1, n = 9	Mean = 5.5, n = 9
nortriptyline vs placebo
Lake 1968	Mean = 3.56, n = 25	Mean = 4.39, n = 29
mianserin vs placebo
Smellie 1996	Mean = 4.5, n = 26

Complete response

More children achieved a complete response at the end of treatment with imipramine compared to placebo (e.g. 78% failed to achieve 14 dry nights on imipramine versus 95% on placebo, RR 0.82, 95% CI 0.78 to 0.87; participants = 831; studies = 12; I² = 98%, Analysis 1.3.1, fixed‐effect). There was evidence of publication bias with an asymmetric funnel plot suggesting that the true effect may closer to RR 1.0 (Figure 4). There was significant heterogeneity. Using a random‐effects model, significantly more children still achieved 14 consecutive dry nights on imipramine (RR 0.74, 95% CI 0.61 to 0.90; participants = 831; studies = 12; I² = 98%, Analysis 1.3.1). A sensitivity analysis excluding the one cross‐over trial (Agarwala 1968#) and one trial conducted amongst institutionalised psychiatric inpatients (Khorana 1972) failed to alter the heterogeneity but still showed a significant difference in favour of the imipramine groups. In the analysis of the subset of children with monosymptomatic enuresis, there was less heterogeneity but results were similar: 54% failed to achieve 14 consecutive dry nights on imipramine versus 84% on placebo (RR 0.63, 95% CI 0.46 to 0.88; participants = 79; studies = 3; I² = 0%, fixed‐effect).

1.3. Analysis.

Comparison 1 TRICYCLIC DRUG vs PLACEBO, Outcome 3 Number not achieving 14 consecutive dry nights at the end of treatment.

4.

Funnel plot of comparison: 1 TRICYCLIC DRUG vs PLACEBO, outcome: 1.3 Number not achieving 14 consecutive dry nights at the end of treatment.

When compared to placebo, more children achieved a complete response at the end of treatment on amitriptyline (73% failed to achieve 14 consecutive dry nights versus 90% for placebo, RR 0.82, 95% CI 0.69 to 0.98; participants = 98; studies = 2; I² = 0%, Analysis 1.3.2, random‐effects). Using desipramine, 77% failed to achieve 14 consecutive dry nights for desipramine versus 94% for placebo (RR 0.83, 95% CI 0.70 to 0.97; participants = 100; Analysis 1.3.4, Liederman 1969, random‐effects). Studies comparing nortriptyline (RR 1.00, 95% CI 0.98 to 1.02; participants = 175; Analysis 1.3.3, Forsythe 1969, random‐effects) and mianserin (RR 0.86, 95% CI 0.61 to 1.22; participants = 55; Analysis 1.3.5, Smellie 1996, random‐effects) with placebo failed to demonstrate a significant difference.

After treatment stopped

Mean wet nights

Only seven studies of tricyclics compared to placebo followed participants after completion of treatment. One study showed no statistically significant difference in the mean number of wet nights after stopping imipramine (MD ‐1.50, 95% CI ‐4.85 to 1.85; participants = 21; Analysis 1.4.1, Attenburrow 1984), but had significantly fewer wet nights after stopping viloxazine, with three fewer wet nights per week compared with placebo (MD ‐2.80, 95% CI ‐4.78 to ‐0.82; participants = 24; Analysis 1.4.2, Attenburrow 1984). One study reported mean wet nights at follow‐up but did not provide standard deviations (Analysis 1.5,1, Kolvin 1972).

1.4. Analysis.

Comparison 1 TRICYCLIC DRUG vs PLACEBO, Outcome 4 Number of wet nights per week at follow‐up.

1.5. Analysis.

Comparison 1 TRICYCLIC DRUG vs PLACEBO, Outcome 5 Number of wet nights per week at follow‐up (no SDs).

Number of wet nights per week at follow‐up (no SDs)
Study	Intervention 1	Intervention 2
imipramine vs placebo
Kolvin 1972	Mean = 3.35, n = 35	Mean = 2.83, n = 27

Complete response

Five small trials provided follow‐up data as relapse rates. There was no evidence to suggest that the modest advantages of imipramine during treatment were sustained after treatment was stopped: almost all the children in both groups who had been cured during treatment relapsed (e.g. 96% on imipramine did not achieve 14 consecutive dry nights or relapsed at follow‐up versus 97% after placebo was stopped, RR 0.99, 95% CI 0.95 to 1.03; participants = 416; studies = 5; I² = 0% , Analysis 1.6.1). In the analysis of the subset of children with monosymptomatic enuresis, 84% of those on imipramine did not achieve 14 consecutive dry nights or relapsed at follow‐up versus 91% after placebo was stopped (RR 0.91, 95% CI 0.75 to 1.12; participants = 65; studies = 3; I² = 46%).

1.6. Analysis.

Comparison 1 TRICYCLIC DRUG vs PLACEBO, Outcome 6 Number not achieving 14 consecutive dry nights or relapsing at follow‐up.

There was no difference in the number not achieving 14 consecutive dry nights or relapsing at follow‐up after cessation of nortriptyline compared with placebo in one study (RR 1.00, 95% CI 0.97 to 1.03; participants = 175; Analysis 1.6.2, Forsythe 1969).

Adverse effects

There were 18 studies comparing a tricyclic with placebo that mentioned adverse effects. Adverse effects were reported for imipramine, nortriptyline, amitriptyline and viloxazine. Adverse effects commonly reported for tricyclics include dizziness, gastrointestinal symptoms (nausea, vomiting, abdominal pain, diarrhoea, dry mouth), headache, lethargy, sleep disturbance. Rare adverse effects include neutropenia, epistaxis, depression and weight loss. Similar effects were reported for placebo, but were reported less commonly. Some studies did not identify the arm in which adverse effects occurred.

Trials with no useable data

A further five trials provided no useable data: one stated that imipramine performed significantly better than placebo (Bindelglas 1968); another reported that more children improved on imipramine than on placebo (Friday 1966); another found that 15 of the 17 participants showed the 'appropriate rise and fall of dry nights' with imipramine (Shaffer 1968#); One reported that amitriptyline was better than placebo (Shah 1971#); and in the last, the trialists concluded that although imipramine reduced the frequency of bedwetting during treatment, this effect was not maintained afterwards in a group of 'unmotivated' older boys (Thomsen 1967#).

2. Higher versus lower doses of tricyclic and related drugs (Analysis 2, 'Other data' tables)

Two cross‐over trials compared different doses of imipramine (Shaffer 1968#; Martin 1971#). One trial stated that there was no difference between high and low doses (Shaffer 1968#) but provided no data. In the other trial, the mean number of wet nights at the end of treatment was reported without standard deviations (Analysis 2.1.1, Martin 1971#).

2.1. Analysis.

Comparison 2 COMPARING DOSES OF TRICYCLIC DRUGS, Outcome 1 Number of wet nights per week at the end of treatment (no SDs).

Number of wet nights per week at the end of treatment (no SDs)
Study	Intervention 1	Intervention 2
imipramine higher dose vs lower dose
Martin 1971#	25 mg dose: Mean = 2.83, n = 57	10 mg dose: mean = 3.69, n = 57

3. Tricyclic and related drugs compared with other drugs (Analysis 3, 'Other data' tables)

Twenty trials compared a tricyclic or similar drug with another tricyclic or a drug of another class (Attenburrow 1984; Burke 1995; Forsythe 1969; Hoashi 1995; Holt 1986; Ingle 1968; Kunin 1970#; Lee 2005; Moltke 1979; Motavalli 1994; Petersen 1974#; Smellie 1996; Tahmaz 2000; Vertucci 1997; Esmaeili 2008; Naitoh 2005; Neveus 2008#; Kang 2003; Ye 2001; Seo 2001). Most studies were small and were compared with another drug in single trials. Another three studies did not provide useable data (Batislam 1995; Yurdakok 1986; Yurdakok 1987).

Comparisons between tricyclics

Imipramine was compared with three other tricyclics in single studies.

• When imipramine was compared with viloxazine, there was no difference in the mean number of wet nights at the end of treatment (MD 0.60, 95% CI ‐3.04 to 4.24; participants = 21; Analysis 3.1.1, Attenburrow 1984) and at follow‐up after cessation of medication (MD 1.30, 95% CI ‐2.21 to 4.81; participants = 21; Analysis 3.4.1).

• When imipramine was compared with mianserin in a single study, more children achieved a complete response at the end of treatment with imipramine (56% failed to achieve 14 consecutive dry nights on imipramine versus 88% on mianserin, RR 0.63, 95% CI 0.44 to 0.92; participants = 51; Analysis 3.3.1, Smellie 1996). The mean number of wet nights at the end of treatment was reported without standard deviations (Analysis 3.2.1, Smellie 1996).

• When comparing imipramine with nortriptyline (Forsythe 1969), there was no difference in the number achieving a complete response at the end of treatment (RR 0.99, 95% CI 0.95 to 1.02; participants = 166; Analysis 3.3.2) and not achieving 14 consecutive dry nights or relapsing after cessation of treatment at follow‐up (RR 0.99, 95% CI 0.94 to 1.03; participants = 166; Analysis 3.5.1) between the two tricyclics.

3.1. Analysis.

Comparison 3 DRUG‐DRUG COMPARISONS, Outcome 1 Number of wet nights per week at the end of treatment.

3.4. Analysis.

Comparison 3 DRUG‐DRUG COMPARISONS, Outcome 4 Number of wet nights per week at follow‐up.

3.3. Analysis.

Comparison 3 DRUG‐DRUG COMPARISONS, Outcome 3 Number not achieving 14 consecutive dry nights at the end of treatment.

3.2. Analysis.

Comparison 3 DRUG‐DRUG COMPARISONS, Outcome 2 Number of wet nights per week at the end of treatment (no SDs).

Number of wet nights per week at the end of treatment (no SDs)
Study	Intervention 1	Intervention 2
imipramine vs mianserin
Smellie 1996	Mean = 2, n = 25	Mean = 4.5, n = 26
imipramine vs desmopressin
Vertucci 1997	Mean = 2.5, n = 28	Mean = 1, n = 29
Ye 2001	Mean = 3.4, n = 44	Mean = 2.3, n = 39
imipramine vs furosemide
Moltke 1979	Mean = 3.43, n = 43	Mean = 5.11, n = 41
imipramine vs emepronium (Ceteprin)
Petersen 1974#	Mean = 2.4, n = 61	Mean = 4.36, n = 61
imipramine‐N‐oxide vs emepronium (Ceteprin)
Petersen 1974#	Mean = 3.24, n = 61	Mean = 4.36, n = 61

3.5. Analysis.

Comparison 3 DRUG‐DRUG COMPARISONS, Outcome 5 Number not achieving 14 consecutive dry nights or relapsing at follow‐up.

In trials with no useable data, imipramine was reported to be better than amitriptyline on a therapeutic index score (Yurdakok 1986); and imipramine and viloxazine were reported to be similar (Yurdakok 1987).

Comparison with desmopressin

The tricyclics imipramine and amitriptyline were compared with desmopressin.

• When imipramine monotherapy was compared with desmopressin monotherapy, there was no statistically significant difference in the mean number of wet nights at the end of treatment (MD 0.11, 95% CI ‐0.41 to 0.62; participants = 300; studies = 3; I² = 79%, Analysis 3.1.2), in the number achieving a complete response at the end of treatment (RR 0.88, 95% CI 0.64 to 1.20; participants = 222; studies = 4; I² = 84%, Analysis 3.3.3) and in the mean number of wet nights at follow‐up after cessation of treatment (MD ‐0.20, 95% CI ‐1.60 to 1.20; participants = 36; studies = 1; Analysis 3.4.2).

Two studies reported the mean number of wet nights at the end of treatment without providing standard deviations (Analysis 3.2.2).

• When imipramine monotherapy was compared with desmopressin combined with oxybutynin in a single study (Lee 2005), the combination therapy appeared superior with about one fewer wet nights per week for the combination therapy (MD 1.07, 95% CI 0.06 to 2.08; participants = 45; Analysis 3.1.3) and more achieving a complete response at the end of treatment (87% failing to achieve 14 consecutive dry nights for imipramine versus 36% for desmopressin and oxybutynin, RR 2.39, 95% CI 1.35 to 4.25; participants = 45, Analysis 3.3.4).

• In one small study that compared amitriptyline with desmopressin (Burke 1995), there was about one fewer wet nights at the end of treatment for amitriptyline (MD ‐1.40, 95% CI ‐2.68 to ‐0.12; participants = 31; Analysis 3.1.4); however, there was no statistically significant difference in the number achieving a full response at the end of treatment (RR 0.83, 95% CI 0.62 to 1.12; participants = 31; Analysis 3.3.5) or at follow‐up after cessation of treatment (RR 1.00, 95% CI 0.88 to 1.13; participants = 31; Analysis 3.5.2), and no difference in the mean number of wet night at follow‐up when treatment ceased (MD 0.10, 95% CI ‐2.35 to 2.55; participants = 15; Analysis 3.4.3).

Comparison with other drugs

Tricyclics were compared with the anticholinergics oxybutynin and tolterodine.

• In one study there was one fewer wet nights at the end of treatment for oxybutynin compared to imipramine (Esmaeili 2008) (MD 1.00, 95% CI 0.02 to 1.98; participants = 55; Analysis 3.1.5), but there was no statistically significant difference in the number achieving a complete response at the end of treatment (RR 1.06, 95% CI 0.82 to 1.37; participants = 85; studies = 2; I² = 4%, Analysis 3.3.6) and in the number not achieving 14 consecutive dry nights or relapsing at follow‐up (RR 0.94, 95% CI 0.67 to 1.32; participants = 21; studies = 1; Analysis 3.5.3).

• When imipramine was compared to tolterodine in a single small study, there was no difference in the mean number of wet nights at the end of treatment between the two drugs (MD ‐1.30, 95% CI ‐3.44 to 0.84; participants = 18; Analysis 3.1.6) but more children achieved a full response at the end of treatment when using imipramine (44% failed to achieve 14 consecutive dry nights on imipramine versus 100% for tolterodine, RR 0.47, 95% CI 0.24 to 0.95; participants = 18; Analysis 3.3.7).

Tricyclics were also compared with other drugs such as chlordiazepoxide, diclofenac sodium, emepronium, ephedrine, furosemide, meprobamate plus hydrozine and piracetam, which are now no longer used for treating enuresis. In single small studies the mean number of wet nights at the end of treatment was lower for imipramine with two fewer wet nights compared to ephedrine (MD ‐2.05, 95% CI ‐3.08 to ‐1.02; participants = 36; Analysis 3.1.8, Kunin 1970#), and four fewer wet nights compared to a mixture of meprobamate plus hydroxyzine (MD ‐4.60, 95% CI ‐5.96 to ‐3.24; participants = 25, Analysis 3.1.9, Ingle 1968), but was similar when compared with chlorpromazine (MD ‐1.20, 95% CI ‐3.20 to 0.80; participants = 19; Analysis 3.1.7, Motavalli 1994). The number achieving a full response at the end of treatment was also higher for imipramine when compared with the meprobamate plus hydroxyzine mixture (58% on imipramine failed to achieve 14 consecutive dry nights versus 100% on meprobamate plus hydroxyzine, RR 0.60, 95% CI 0.37 to 0.96; participants = 25, Analysis 3.3.8). The mean number of wet nights at the end of treatment was reported for furosemide (Analysis 3.2.3, Moltke 1979), and emepronium (Analysis 3.2.4, Petersen 1974#) without providing SDs. In a trial with no useable data, imipramine was reported to be better than chlordiazepoxide or piracetam on a therapeutic index score (Yurdakok 1986).

Adverse effects

Nine studies comparing a tricyclic with other tricyclics (viloxazine, nortriptyline), anticholinergics (oxybutynin, emepronium), desmopressin, and desmopressin combined with imipramine or oxybutynin mentioned adverse effects. Adverse effects of rash, headache, back pain and nasal inflammation were reported for desmopressin. Most studies did not differentiate which treatment caused which adverse effects, including common effects such as dizziness, gastrointerstinal symptoms, headache and lethargy.

4. Tricyclic and related drugs compared with behavioural interventions (Analysis 4, 'Other data' tables)

Twelve trials compared a tricyclic or similar drug with a behavioural intervention (Bhatia 1990; Danquah 1975; Fournier 1987; Iester 1991; Kolvin 1972; Maxwell 1971#; McKendry 1975; Mehrotra 1980; Motavalli 1994; Netley 1984; Wagner 1982), and Kumazawa 1990, which compared a combination therapy involving a tricyclic (imipramine combined with motivation and bladder exercises) which is mentioned in Analysis 6. Almost all of the comparisons or outcomes were addressed in single, small trials.

Comparisons with alarms

Seven studies compared a tricyclic or related drug with alarm therapy (Danquah 1975; Fournier 1987; Kolvin 1972; McKendry 1975; Motavalli 1994; Netley 1984; Wagner 1982).

Comparing imipramine with alarm therapy in a single small study, there was no statistically significant difference in the mean number of wet nights at the end of treatment (MD 0.65, 95% CI ‐0.91 to 2.21; participants = 20; Analysis 4.1.1, Motavalli 1994). However, the number achieving a complete response at the end of treatment was higher for standard alarm therapy (67% for imipramine failed to achieve 14 consecutive dry nights versus 17% for alarm therapy, RR 4.00, 95% CI 1.06 to 15.08; participants = 24; Analysis 4.3.1, Wagner 1982). In the two trials evaluating the Mozes electric shock alarm, 81% failed to achieve 14 consecutive dry nights for imipramine versus 63% for Mozes alarms, RR 1.28, 95% CI 1.07 to 1.53; participants = 184; studies = 2; I² = 58%, Analysis 4.3.2), but the electric shock method was unacceptable to children and their parents (McKendry 1975; Netley 1984), and is no longer in current use.

4.1. Analysis.

Comparison 4 DRUGS vs BEHAVIOURAL INTERVENTIONS, Outcome 1 Number of wet nights per week at the end of treatment.

4.3. Analysis.

Comparison 4 DRUGS vs BEHAVIOURAL INTERVENTIONS, Outcome 3 Number not achieving 14 consecutive dry nights at the end of treatment.

At follow‐up, the number not achieving 14 consecutive dry nights or relapsing after cessation of treatment was also lower for alarm therapy (100% after imipramine versus 58% after alarms, RR 1.67, 95% CI 1.03 to 2.69; participants = 24; Analysis 4.5.1, Wagner 1982). The mean number of wet nights at the end of treatment for imipramine compared with alarm therapy was reported for three studies (Fournier 1987; Kolvin 1972; Wagner 1982, Analysis 4.2.1) and at follow‐up for one study (Analysis 4.4.1, Kolvin 1972) without providing standard deviations.

4.5. Analysis.

Comparison 4 DRUGS vs BEHAVIOURAL INTERVENTIONS, Outcome 5 Number not achieving 14 consecutive dry nights or relapsing at follow‐up.

4.2. Analysis.

Comparison 4 DRUGS vs BEHAVIOURAL INTERVENTIONS, Outcome 2 Number of wet nights per week at the end of treatment (no SDs).

Number of wet nights per week at the end of treatment (no SDs)
Study	Intervention 1	Intervention 2
imipramine vs alarm
Fournier 1987	Mean = 1.9, n = 8	Mean = 2.5, n = 8
Kolvin 1972	Mean = 2.32, n = 35	Mean = 2.28, n = 32
Wagner 1982	Mean = 2.75, n = 12	Mean = 0.58, n = 12
imipramine vs random wakening
Fournier 1987	Mean = 1.9, n = 8	Mean = 3.3, n = 8

4.4. Analysis.

Comparison 4 DRUGS vs BEHAVIOURAL INTERVENTIONS, Outcome 4 Number of wet nights per week at follow‐up (no SDs).

Number of wet nights per week at follow‐up (no SDs)
Study	Intervention 1	Intervention 2
imipramine vs alarm
Kolvin 1972	Mean = 3.4, n = 35	Mean = 2.3, n = 32
amitriptyline vs alarm
Danquah 1975	Mean = 4, n = 10	Mean = 3.2, n = 10
amitriptyline vs shaming
Danquah 1975	Mean = 4, n = 10	Mean = 5.6, n = 10

In a single small study, there was also no statistically significant difference in the mean number of wet nights at the end of treatment between clomipramine and alarm therapy (MD 1.90, 95% CI ‐0.34 to 4.14; participants = 19; Analysis 4.1.2). The mean number of wet nights at follow‐up was reported for amitriptyline and alarm therapy without providing SDs (Analysis 4.4.2).

Comparisons with simple behavioural methods

Three studies compared a tricyclic with a simple behavioural method. One of these studies compared the additive effect of imipramine to star charts and is discussed under combination therapy involving tricyclics (Maxwell 1971#). The number not achieving 14 consecutive dry nights or relapsed at follow‐up was lower for imipramine compared with fluid restriction and avoiding punishment in a single small study (40% for imipramine compared with 80% for the behavioural therapy, RR 0.50, 95% CI 0.28 to 0.89; participants = 40; Analysis 4.5.2, Bhatia 1990).The mean number of wet nights at the end of treatment was reported for imipramine compared to random waking (Analysis 4.2.2, Fournier 1987) and at follow‐up after cessation of treatment for amitriptyline compared with shaming (Analysis 4.4.3, Danquah 1975) without providing SDs.

Comparisons with complicated or complex behavioural methods

In one relatively large study (Iester 1991), the number achieving a full response at the end of treatment was lower for imipramine compared with:

• the three‐step programme (parental reassurance, retention control training and wakening using an alarm clock and parental involvement): 61% failed to achieve 14 consecutive dry nights for imipramine versus 33% for the complex behaviour method (RR 1.83, 95% CI 1.08 to 3.12; participants = 72; Analysis 4.3.3)

• the three‐step programme combined with motivational therapy and computer‐led education: 61% failed to achieve 14 consecutive dry nights for imipramine versus 16% for the complex behaviour method: (RR 3.91, 95% CI 2.30 to 6.66; participants = 132; Analysis 4.3.4).

At follow‐up, the number not achieving 14 consecutive dry nights or relapsing after cessation of treatment was also lower for;

• the three‐step programme (parental reassurance, retention control training and wakening using an alarm clock and parental involvement): 67% failed to achieve 14 consecutive dry nights or relapsed for imipramine versus 39% for the complex behaviour method (RR 1.71, 95% CI 1.07 to 2.74; participants = 72; Analysis 4.5.3)

• the three‐step programme combined with motivational therapy and computer‐led education: 67% failed to achieve 14 consecutive dry nights or relapsed for imipramine versus 19% for the complex behaviour method (RR 3.56, 95% CI 2.21 to 5.72; participants = 132; Analysis 4.5.4).

Adverse effects

Four studies comparing tricyclics with a behavioural intervention mentioned adverse effects. Drowsiness was reported for amitriptyline, and headache, fatigue and gastrointestinal symptoms for imipramine. Fear and skin damage (erythema, burn, ulceration) were noted for the electric shock enuresis alarm.

5. Tricyclic and related drugs compared with complementary or miscellaneous interventions (Analysis 5)

Three small trials compared a tricyclic or related drug with the complementary or miscellaneous interventions: hypnosis (Banerjee 1993); psychotherapy (Ciotti 1983); and restricted diet (McKendry 1975). In the study comparing imipramine with a restricted diet (McKendry 1975), the number achieving a full response at the end of treatment was higher for those using imipramine (84% failed to achieve 14 consecutive dry nights for imipramine versus 99% for restricted diet, RR 0.84, 95% CI 0.75 to 0.93; participants = 147; Analysis 5.1.1), whether or not dropouts were counted as failures. However, there was differential dropout from the groups, with most families requesting transfer to another treatment after one to two months because of the low success rate and the restrictive nature of the diet.

5.1. Analysis.

Comparison 5 DRUGS VS COMPLEMENTARY / MISCELLANEOUS INTERVENTIONS, Outcome 1 Number not achieving 14 dry night at the end of treatment.

Two studies had no useable data (Banerjee 1993; Ciotti 1983), although the authors stated there were lower relapse rates in the hypnosis group (Banerjee 1993).

Adverse effects

Two studies comparing tricyclics with a complementary or miscellaneous intervention mentioned adverse effects. Although adverse effects were reported infrequently for the tricyclics, there were no adverse effects reported for psychotherapy; however, children with the restricted diet had aggressive behaviour,

6. Combination therapy involving tricyclics vs other (Analysis 6, 'Other data' tables)

Nine trials assessed the effects of tricyclics in combination with another therapy (Bhatia 1990; Burke 1995; Esmaeili 2008; Forsythe 1972b; Fournier 1987; Kumazawa 1990; Naitoh 2005; Seo 2001; Tahmaz 2000).

Tricyclics combined with another medication

Tricyclics were combined with anticholinergics in two trials (Esmaeili 2008; Tahmaz 2000).

Tricyclics plus anticholinergics versus placebo

In one study comparing imipramine and oxybutynin with placebo (Tahmaz 2000):

• the number not achieving 14 consecutive dry nights at the end of treatment (33% for imipramine and oxybutynin versus 78% for placebo: RR 0.43, 95% CI 0.23 to 0.78; participants = 47; Analysis 6.3.1) was lower for the combination therapy.

• the number not achieving 14 consecutive dry nights or relapsing at follow‐up (45% for imipramine combined with oxybutynin versus 79% for placebo: RR 0.58, 95% CI 0.34 to 0.99; participants = 36; Analysis 6.5.1) was lower for the combination therapy.

6.3. Analysis.

Comparison 6 COMBINATION THERAPY INVOLVING TRICYCLICS VS OTHER, Outcome 3 Number not achieving 14 consecutive dry nights at the end of treatment.

6.5. Analysis.

Comparison 6 COMBINATION THERAPY INVOLVING TRICYCLICS VS OTHER, Outcome 5 Number not achieving 14 consecutive dry nights or relapsing at follow‐up.

Tricyclics plus anticholinergics versus imipramine monotherapy

Imipramine and oxybutynin was also superior to imipramine monotherapy at the end of treatment:

• with about two fewer wet nights for imipramine combined with oxybutynin (MD ‐2.10, 95% CI ‐2.99 to ‐1.21; participants = 63, Analysis 6.1.1)

• with more achieving a full response (48% in the combination therapy group failed to achieve 14 consecutive dry nights at the end of treatment versus 74% in the imipramine monotherapy group: RR 0.68, 95% CI 0.50 to 0.92; participants = 101; studies = 2; I² = 0%, Analysis 6.3.2).

• At follow‐up, the number not achieving 14 consecutive dry nights or relapsing after cessation of treatment was also lower in the combination therapy group (45% in the combination group versus 83% on imipramine monotherapy, RR 0.55, 95% CI 0.32 to 0.92; participants = 34; Analysis 6.5.2, Tahmaz 2000).

6.1. Analysis.

Comparison 6 COMBINATION THERAPY INVOLVING TRICYCLICS VS OTHER, Outcome 1 Number of wet nights per week at the end of treatment.

Tricyclics plus desmopressin versus imipramine monotherapy

Tricyclics combined with desmopressin were compared to tricyclic monotherapy in two trials (Burke 1995; Seo 2001). In the small study comparing the combination of amitriptyline and desmopressin with amitriptyline monotherapy (Burke 1995), there was no difference in:

• mean wet nights (MD 0.00, 95% CI ‐1.64 to 1.64; participants = 28; Analysis 6.1.2) and the number achieving a complete response (RR 0.82, 95% CI 0.51 to 1.32; participants = 28; Analysis 6.3.5) at the end of treatment.

• mean number of wet nights (MD 1.20, 95% CI ‐1.65 to 4.05; participants = 18; Analysis 6.4.1) and the number not achieving 14 consecutive dry nights or relapsing (RR 0.93, 95% CI 0.77 to 1.13; participants = 28; Analysis 6.5.3) at follow‐up.

6.4. Analysis.

Comparison 6 COMBINATION THERAPY INVOLVING TRICYCLICS VS OTHER, Outcome 4 Number of wet nights per week at follow‐up.

In one other study that compared imipramine and desmopressin with imipramine monotherapy (Seo 2001), there was also no difference in the number achieving a complete response at the end of treatment (RR 0.82, 95% CI 0.32 to 2.06; participants = 89; Analysis 6.3.3).

Tricyclics plus desmopressin versus desmopressin monotherapy

Imipramine combined with desmopressin was compared to desmopressin monotherapy in one study (Seo 2001). The number not achieving 14 consecutive dry nights at the end of treatment was lower for the combination group (15% in the combination group versus 40% in the desmopressin group, RR 0.38, 95% CI 0.17 to 0.83; participants = 86; Analysis 6.3.4).

Tricyclics combined with other drugs

Imipramine combined with chlordiazepoxide (a benzodiazepine) was compared with placebo in a large study (Forsythe 1972b) showing no difference in the number achieving a complete response at the end of treatment (RR 0.98, 95% CI 0.95 to 1.01; participants = 215; Analysis 6.3.6) and the number not achieving 14 consecutive dry nights or relapsing at follow‐up (RR 1.00, 95% CI 0.97 to 1.03; participants = 200; Analysis 6.5.4).

Tricyclics combined with a behavioural therapy

Alarms

There were three studies that combined a tricyclic with alarm therapy (imipramine: Fournier 1987; Naitoh 2005; nortriptyline: Scholander 1968). In one study (Naitoh 2005) there was no difference in the number achieving a complete response at the end of treatment when imipramine and alarm therapy were compared with alarm monotherapy (RR 1.12, 95% CI 0.85 to 1.48; participants = 70; Analysis 6.3.7) or between nortriptyline combined with alarm therapy and alarm combined with placebo (RR 1.50, 95% CI 0.71 to 3.16; participants = 30; Analysis 6.3.9, Scholander 1968). There was also no difference between imipramine combined with alarm therapy and alarm combined with desmopressin (RR 0.98, 95% CI 0.77 to 1.26; participants = 68; Analysis 6.3.8). The mean number of wet nights at the end of treatment for imipramine and alarm combination compared with placebo (Analysis 6.2.1), imipramine monotherapy (Analysis 6.2.2) and random waking (Analysis 6.2.3) were reported without SDs.

6.2. Analysis.

Comparison 6 COMBINATION THERAPY INVOLVING TRICYCLICS VS OTHER, Outcome 2 Number of wet nights per week at the end of treatment (no SDs).

Number of wet nights per week at the end of treatment (no SDs)
Study	Intervention 1	Intervention 2
imipramine + alarm vs placebo
Fournier 1987	Mean = 1, n = 8	Mean = 5, n = 8
imipramine + alarm vs imipramine
Fournier 1987	Mean = 1, n = 8	Mean = 1.9, n = 8
Imipramine + alarm vs random waking
Fournier 1987	Mean = 1, n = 8	Mean = 3.3, n = 8

Motivation/bladder exercises

In a single large trial, there was about one fewer wet night at the end of treatment for imipramine plus star chart compared with placebo plus star chart (MD ‐0.80, 95% CI ‐1.33 to ‐0.27; participants = 250; Analysis 6.1.4, Maxwell 1971#). In one small study (Kumazawa 1990), imipramine combined with motivation and bladder exercises was compared with placebo combined with motivation and bladder exercises, showing no statistically significant difference in the mean number of wet nights at the end of treatment (MD ‐1.13, 95% CI ‐2.84 to 0.58; participants = 20; Analysis 6.1.3), suggesting that the addition of imipramine to motivation and bladder exercises did not improve the outcome.

Other interventions

One small study (Bhatia 1990) showed no difference at follow‐up in the number not achieving 14 consecutive dry nights or relapsing between imipramine combined with fluid restriction and avoidance of punishment and imipramine monotherapy (RR 0.25, 95% CI 0.06 to 1.03; participants = 40; Analysis 6.5.5). However, when compared with placebo combined with fluid restriction and avoidance of punishment, those in the imipramine combination group had lower numbers not achieving 14 consecutive dry nights or relapsing at follow‐up (RR 0.13, 95% CI 0.03 to 0.47; participants = 40; Analysis 6.5.6), suggesting imipramine was the effective component of the therapy.

Adverse effects

Two studies mentioned adverse effects when a tricyclic was combined with another intervention. Although gastrointestinal effects were reported for imipramine, desmopressin and oxybutynin, the frequency of adverse effects was higher when these drugs were combined compared to monotherapy.

Discussion

Summary of main results

Tricyclics are more effective than placebo or no treatment, particularly for short‐term responses. The tricyclics imipramine, amitriptyline and desipramine were more effective than placebo, but we found no difference for nortriptyline and mianserin. Most tricyclics did not have a sustained effect and participants relapsed at follow‐up after cessation of treatment. We did not find any studies that included two‐year follow‐up and therefore could not assess long‐term complete success.

There was insufficient evidence to determine whether there was any difference in treatment response between the different doses of tricyclics. When comparing one tricyclic with another, there was no difference in treatment effects between imipramine and nortriptyline, nor between imipramine and viloxazine, but imipramine appears to be superior to mianserin.

There was not much difference between tricyclics and desmopressin, although there was about one fewer wet nights for amitriptyline compared with desmopressin. However, desmopressin combined with oxybutynin was superior to imipramine monotherapy for short‐term responses. There were mixed outcomes when comparing tricyclic monotherapy with an anticholinergic monotherapy for short‐term responses, with fewer wet nights at the end of treatment for oxybutynin compared to imipramine, but more children achieving 14 consecutive dry nights with imipramine compared to tolterodine.

Alarm therapy was superior to tricyclics, with significantly more children achieving 14 consecutive dry nights, although there was no statistically significant difference in the mean number of wet nights demonstrated. Alarm therapy also had a more sustained response, with 42% dry after cessation of treatment, compared with none after cessation of imipramine.

Overall, tricyclics were more effective than simple behavioural therapies, but less effective than complex behavioural therapies. A restrictive diet appears ineffective and there was insufficient evidence to judge the effect of tricyclics compared to the other complementary and miscellaneous interventions.

When tricyclics were combined with other treatments, the combination of a tricyclic with an anticholinergic was superior to both placebo and to tricyclic monotherapy, in both the short and long term. Tricyclics combined with desmopressin were not more effective than tricyclic monotherapy, but were more effective than desmopressin monotherapy. There was no additional benefit of adding a tricyclic to alarm therapy, with no statistically significant difference demonstrated when compared to alarm monotherapy (with or without placebo), alarm combined with desmopressin or alarm combined with nortriptyline.

There were mixed responses when tricyclics were combined with other behavioural therapies, with no additive effect of tricyclics to motivation and bladder exercises, but tricyclics combined with fluid restriction and avoidance of punishment was superior to fluid restriction and avoidance of punishment without tricyclics. There was also no additive effect of fluid restriction and avoidance of punishment to tricyclic therapy.

Adverse effects were mentioned in 30 trials, with seven trials reporting no adverse events. Reported adverse events were usually of minor consequences, but some trials did not give details of the nature or severity of the side effects, and others failed to indicate the treatment arm in which the adverse effects occurred. The frequency of adverse effects for placebo was less than for tricyclics. Adverse effects were reported more frequently for combination therapy when tricyclics were combined with another drug such as desmopressin or oxybutynin. An important adverse effect was the skin damage, including rashes, burns and ulceration caused by the electric shock alarm.

Overall completeness and applicability of evidence

Enuresis is a common health condition faced by children, and tricyclics were one of the first therapies offered for treating it (Neveus 2010). This review includes 4071 children, with most participants aged between five and 16. The studies were predominantly in an outpatient setting, where children with enuresis are commonly seen. However, seven trials were based in institutional settings and two in community settings.

Many of the studies are quite old, and include some therapies which are now no longer used. We have also studied the effects of combination therapy involving tricyclics as a separate comparison in this review, as combination therapy is increasingly used for treating therapy‐resistant enuresis.

Quality of the evidence

The 64 trials (including 20 cross‐over trials) included in this review were mostly small and of poor quality, with 30 studies having 50 or fewer participants. Most comparisons or outcomes were addressed only by single trials, precluding meta‐analysis. The small sample sizes resulted in wide confidence intervals and imprecision which could obscure or overestimate underlying treatment effects; in general the evidence was too poor to identify whether any treatment or combination of treatments was better than any other. Most of the trials had insufficient follow‐up, with only 18 studies following participants for at least three months after completion of treatment.

The quality of the trials was often low, with 10 trials not providing any useable data, and 21 trials failing to provide standard deviations for their continuous variables. Because none of the cross‐over studies provided suitable data for generic inverse variance analysis, we incorporated cross‐over studies and analysed them as parallel‐group trials, which carries the risk of a unit‐of‐analysis error. None of the trials met all the criteria for a good‐quality study at low risk of bias. The risk of bias in many studies was difficult to judge because of insufficient details of methodology.

Potential biases in the review process

Despite our best efforts, we may have missed some published literature. The quality assessment was based on the published study reports. This may have unintentionally downgraded the perceived quality of blinding and allocation concealment if no information was provided. Given the absence of trial registration for most of the trials included in this review, bias due to selective reporting of outcomes was difficult to ascertain.

Agreements and disagreements with other studies or reviews

The findings from this review are consistent with previous reviews, showing that although tricyclics are effective in reducing enuresis during treatment, most children relapse after cessation of therapy.

Authors' conclusions

Implications for practice.

Although tricyclics are effective in reducing enuresis in the short term, they do not have a sustained effect when treatment ceases. Although only minor adverse effects were reported in the included studies, decisions to use these drugs should take account of the uncertainty over potentially serious adverse effects of tricyclics in this population, arising from the insufficent level of detail provided by many of the studies in this review. There is limited, very low‐quality evidence that combining tricyclics with other therapies such as anticholinergics may be effective.

Implications for research.

As the benefits of tricyclics were relatively small and transient, and there are concerns about safety, further research into their effects is likely to be of limited value. The use of tricyclics as an adjunct to other therapies may warrant further exploration, particularly for children who are resistant to other treatment. Further research is needed to determine which interventions are appropriate for which groups and in which circumstances, in order to guide choice of treatment.

Further research on alternative drugs for enuresis must incorporate comparisons with proven effective treatments such as desmopressin and alarm therapy. Placebo‐controlled drug trials without incorporating comparisons with established therapies should be discouraged when proven effective therapies are available. The methodological quality of the trials also needs to be improved, with a focus on adequate sample sizes, length of follow‐up, reporting of adverse events and relapse rates and the use uniform outcome measures as defined by the International Children's Continence Society, making a conscious efforts to adopt the ICCS nomenclature guidelines.

Patient‐reported outcome measures such as quality of life and other important outcomes such as cost should be reported, to determine the impact of the condition and of the treatment on the child and family, their care providers and on health services. Children with nocturnal enuresis should be included in studies focusing on treatment of night wetting, irrespective of whether they have monosymptomatic or non‐monosymptomatic nocturnal enuresis, which will increase the broad applicability of results, but the two groups should be analysed separately. Subgroup analysis of monosymptomatic and non‐monosymptomatic groups may provide further knowledge about the different responses to treatment of these subgroups.

What's new

Date	Event	Description
12 January 2016	New citation required and conclusions have changed	Following new trials were added in this update Esmaeili 2008; Kang 2003; Naitoh 2005; Neveus 2008#; Seo 2001; Ye 2001. Risk of bias assessment performed on all trials in accordance with the current methodology. Quality of evidence was assessed by adopting the GRADE approach. This review also assessed the effect of combination therapy involving tricyclics
12 January 2016	New search has been performed	We added the following new trials to this update Esmaeili 2008; Kang 2003; Naitoh 2005; Neveus 2008#; Seo 2001; Ye 2001. We conducted 'Risk of bias' assessment on all trials in accordance with the current methodology. We assessed the quality of evidence by adopting the GRADE approach. This update also assesses the effect of combination therapy involving tricyclics.

History

Protocol first published: Issue 2, 2000
 Review first published: Issue 2, 2000

Date	Event	Description
10 October 2008	Amended	Converted to new review format.
25 July 2007	New search has been performed	Minor Update Issue 4, 2007. Two further new trials (Lee 2005 and Miyazaki 1972) added. There were no new findings or conclusions.
15 May 2005	New search has been performed	Minor Update Issue 3 2005. Two new included trials (Hoashi 1995 and Ciotti 1983), and four new excluded trials. There were no new findings or conclusions.
28 May 2003	New citation required and conclusions have changed	Substantive Update Issue 3 2003. 31 new trials added whose data were not included previously. The evidence in favour of tricyclics was less than in the previous version, and alarms appeared to have longer lasting effects.

Data and analyses

Comparison 1. TRICYCLIC DRUG vs PLACEBO.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of wet nights per week at the end of treatment	4		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
1.1 imipramine vs placebo	4	347	Mean Difference (IV, Fixed, 95% CI)	‐0.95 [‐1.40, ‐0.50]
1.2 viloxazine vs placebo	1	24	Mean Difference (IV, Fixed, 95% CI)	‐3.1 [‐5.64, ‐0.56]
2 Number of wet nights per week at the end of treatment (no SDs)			Other data	No numeric data
2.1 imipramine vs placebo			Other data	No numeric data
2.2 imipramine vs no treatment control			Other data	No numeric data
2.3 imipramine‐N‐oxide vs placebo			Other data	No numeric data
2.4 amitriptyline vs placebo			Other data	No numeric data
2.5 nortriptyline vs placebo			Other data	No numeric data
2.6 mianserin vs placebo			Other data	No numeric data
3 Number not achieving 14 consecutive dry nights at the end of treatment	15		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
3.1 imipramine vs placebo	12	831	Risk Ratio (M‐H, Random, 95% CI)	0.74 [0.61, 0.90]
3.2 amitriptyline vs placebo	2	98	Risk Ratio (M‐H, Random, 95% CI)	0.83 [0.72, 0.97]
3.3 nortriptyline vs placebo	1	175	Risk Ratio (M‐H, Random, 95% CI)	1.0 [0.98, 1.02]
3.4 desipramine vs placebo	1	100	Risk Ratio (M‐H, Random, 95% CI)	0.83 [0.70, 0.97]
3.5 mianserin vs placebo	1	55	Risk Ratio (M‐H, Random, 95% CI)	0.86 [0.61, 1.22]
4 Number of wet nights per week at follow‐up	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
4.1 imipramine vs placebo	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.2 viloxazine vs placebo	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
5 Number of wet nights per week at follow‐up (no SDs)			Other data	No numeric data
5.1 imipramine vs placebo			Other data	No numeric data
6 Number not achieving 14 consecutive dry nights or relapsing at follow‐up	5		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
6.1 imipramine vs placebo	5	416	Risk Ratio (M‐H, Fixed, 95% CI)	0.99 [0.95, 1.03]
6.2 nortriptyline vs placebo	1	175	Risk Ratio (M‐H, Fixed, 95% CI)	1.00 [0.97, 1.03]

Comparison 2. COMPARING DOSES OF TRICYCLIC DRUGS.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of wet nights per week at the end of treatment (no SDs)			Other data	No numeric data
1.1 imipramine higher dose vs lower dose			Other data	No numeric data

Comparison 3. DRUG‐DRUG COMPARISONS.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of wet nights per week at the end of treatment	10		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
1.1 imipramine vs viloxazine	1	21	Mean Difference (IV, Fixed, 95% CI)	0.60 [‐3.04, 4.24]
1.2 imipramine vs desmopressin	3	300	Mean Difference (IV, Fixed, 95% CI)	0.11 [‐0.41, 0.62]
1.3 imipramine vs desmopressin + oxybutynin	1	45	Mean Difference (IV, Fixed, 95% CI)	1.07 [0.06, 2.08]
1.4 amitriptyline vs desmopressin	1	31	Mean Difference (IV, Fixed, 95% CI)	‐1.40 [‐2.68, ‐0.12]
1.5 imipramine vs oxybutynin	1	55	Mean Difference (IV, Fixed, 95% CI)	1.0 [0.02, 1.98]
1.6 imipramine vs tolterodine	1	18	Mean Difference (IV, Fixed, 95% CI)	‐1.30 [‐3.44, 0.84]
1.7 imipramine vs clomipramine	1	19	Mean Difference (IV, Fixed, 95% CI)	‐1.20 [‐3.20, 0.80]
1.8 imipramine vs ephedrine sulphate	1	36	Mean Difference (IV, Fixed, 95% CI)	‐2.05 [‐3.08, ‐1.02]
1.9 imipramine vs meprobamate + hydroxyzine	1	25	Mean Difference (IV, Fixed, 95% CI)	‐4.6 [‐5.96, ‐3.24]
2 Number of wet nights per week at the end of treatment (no SDs)			Other data	No numeric data
2.1 imipramine vs mianserin			Other data	No numeric data
2.2 imipramine vs desmopressin			Other data	No numeric data
2.3 imipramine vs furosemide			Other data	No numeric data
2.4 imipramine vs emepronium (Ceteprin)			Other data	No numeric data
2.5 imipramine‐N‐oxide vs emepronium (Ceteprin)			Other data	No numeric data
3 Number not achieving 14 consecutive dry nights at the end of treatment	11		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
3.1 imipramine vs mianserin	1	51	Risk Ratio (M‐H, Fixed, 95% CI)	0.63 [0.44, 0.92]
3.2 imipramine vs nortriptyline	1	166	Risk Ratio (M‐H, Fixed, 95% CI)	0.99 [0.95, 1.02]
3.3 imipramine vs desmopressin	4	222	Risk Ratio (M‐H, Fixed, 95% CI)	0.88 [0.64, 1.20]
3.4 imipramine vs desmopressin + oxybutynin	1	45	Risk Ratio (M‐H, Fixed, 95% CI)	2.39 [1.35, 4.25]
3.5 amitriptyline vs desmopressin	1	31	Risk Ratio (M‐H, Fixed, 95% CI)	0.83 [0.62, 1.12]
3.6 imipramine vs oxybutynin	2	85	Risk Ratio (M‐H, Fixed, 95% CI)	1.02 [0.79, 1.32]
3.7 imipramine vs tolterodine	1	18	Risk Ratio (M‐H, Fixed, 95% CI)	0.47 [0.24, 0.95]
3.8 imipramine vs meprobamate + hydroxyzine	1	25	Risk Ratio (M‐H, Fixed, 95% CI)	0.60 [0.37, 0.96]
4 Number of wet nights per week at follow‐up	3		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
4.1 imipramine vs viloxazine	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.2 imipramine vs desmopressin	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.3 amitriptyline vs desmopressin	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
5 Number not achieving 14 consecutive dry nights or relapsing at follow‐up	3		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
5.1 imipramine vs nortriptyline	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
5.2 amitriptyline vs desmopressin	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
5.3 imipramine vs oxybutynin	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]

Comparison 4. DRUGS vs BEHAVIOURAL INTERVENTIONS.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of wet nights per week at the end of treatment	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.1 imipramine vs alarm	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.2 clomipramine vs alarm	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
2 Number of wet nights per week at the end of treatment (no SDs)			Other data	No numeric data
2.1 imipramine vs alarm			Other data	No numeric data
2.2 imipramine vs random wakening			Other data	No numeric data
3 Number not achieving 14 consecutive dry nights at the end of treatment	4		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
3.1 imipramine vs alarm	1	24	Risk Ratio (M‐H, Fixed, 95% CI)	4.0 [1.06, 15.08]
3.2 imipramine vs alarm (Mozes detector)	2	184	Risk Ratio (M‐H, Fixed, 95% CI)	1.28 [1.07, 1.53]
3.3 imipramine vs 3‐step therapy (reassurance, retention control training + wakening, parental involvement)	1	72	Risk Ratio (M‐H, Fixed, 95% CI)	1.83 [1.08, 3.12]
3.4 imipramine vs 3‐step therapy + motivation/counselling + education	1	132	Risk Ratio (M‐H, Fixed, 95% CI)	3.91 [2.30, 6.66]
4 Number of wet nights per week at follow‐up (no SDs)			Other data	No numeric data
4.1 imipramine vs alarm			Other data	No numeric data
4.2 amitriptyline vs alarm			Other data	No numeric data
4.3 amitriptyline vs shaming			Other data	No numeric data
5 Number not achieving 14 consecutive dry nights or relapsing at follow‐up	3		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
5.1 imipramine vs alarm	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
5.2 imipramine vs simple behavioural intervention (restrict fluids, avoid punishment) + placebo	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
5.3 imipramine vs 3 step therapy (reassurance, retention control training + wakening, parental involvement)	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
5.4 imipramine vs 3 step therapy + motivation/counselling + education	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]

Comparison 5. DRUGS VS COMPLEMENTARY / MISCELLANEOUS INTERVENTIONS.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number not achieving 14 dry night at the end of treatment	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
1.1 imipramine vs restricted diet	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]

Comparison 6. COMBINATION THERAPY INVOLVING TRICYCLICS VS OTHER.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of wet nights per week at the end of treatment	4		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.1 imipramine + oxybutynin vs imipramine	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.2 Amitriptyline + desmopressin vs amitriptyline	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.3 imipramine + motivation + bladder exercises vs placebo + motivation + bladder exercises	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.4 imipramine + star chart vs placebo + star chart	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
2 Number of wet nights per week at the end of treatment (no SDs)			Other data	No numeric data
2.1 imipramine + alarm vs placebo			Other data	No numeric data
2.2 imipramine + alarm vs imipramine			Other data	No numeric data
2.3 Imipramine + alarm vs random waking			Other data	No numeric data
3 Number not achieving 14 consecutive dry nights at the end of treatment	7		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
3.1 imipramine + oxybutynin vs placebo	1	47	Risk Ratio (M‐H, Fixed, 95% CI)	0.43 [0.23, 0.78]
3.2 imipramine + oxybutynin vs imipramine	2	101	Risk Ratio (M‐H, Fixed, 95% CI)	0.68 [0.50, 0.92]
3.3 imipramine + desmopressin vs imipramine	1	89	Risk Ratio (M‐H, Fixed, 95% CI)	0.82 [0.32, 2.06]
3.4 imipramine + desmopressin vs desmopressin	1	86	Risk Ratio (M‐H, Fixed, 95% CI)	0.38 [0.17, 0.83]
3.5 amitriptyline + desmopressin vs amitriptyline	1	28	Risk Ratio (M‐H, Fixed, 95% CI)	0.82 [0.51, 1.32]
3.6 amitriptyline + chlordiazepoxide vs placebo	1	215	Risk Ratio (M‐H, Fixed, 95% CI)	0.98 [0.95, 1.01]
3.7 imipramine + alarm vs alarm	1	70	Risk Ratio (M‐H, Fixed, 95% CI)	1.12 [0.85, 1.48]
3.8 imipramine + alarm vs desmopressin + alarm	1	68	Risk Ratio (M‐H, Fixed, 95% CI)	0.98 [0.77, 1.26]
3.9 nortriptyline + alarm vs alarm + placebo	1	30	Risk Ratio (M‐H, Fixed, 95% CI)	1.5 [0.71, 3.16]
4 Number of wet nights per week at follow‐up	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
4.1 amitriptyline + desmopressin vs amitriptyline	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
5 Number not achieving 14 consecutive dry nights or relapsing at follow‐up	4		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
5.1 imipramine + oxybutynin vs placebo	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
5.2 imipramine + oxybutynin vs imipramine	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
5.3 amitriptyline + desmopressin vs amitriptyline	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
5.4 imipramine + chlordiazepoxide vs placebo	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
5.5 imipramine + simple behavioural intervention (restrict fluids, avoid punishment) vs imipramine	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
5.6 imipramine + simple behavioural intervention (restrict fluids, avoid punishment) vs placebo + simple behaviour	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Agarwala 1968#.

Methods	Design: RCT (double‐ blind cross‐over) using random number list and stratified for sex and previous imipramine treatment Systematic baseline measure of wetting: No Organic causes excluded: Yes Daytime wetting excluded: Not mentioned
Participants	No. of children (boys): 29 (15) Inclusion: parents considered enuresis to be a problem; at least 6 years; baseline wetting 6 to 7 times per week Exclusion: mental retardation; organic causes Previous treatment: some had imipramine Age: 6 to 12 years
Interventions	A (29): imipramine 1 tablet for 2 weeks, doubled for next 2 weeks if no response B (29): matching placebo as above Duration of treatment: 2 weeks, but 4 weeks if no response
Outcomes	Wet nights in 2 weeks on treatment: A: 5.5 (SD 3.3), B: 7.8 (4) No. not achieving 14 dry nights during trial: A: 27/29, B: 29/29 Side effects: A: 1 (dizzy on 50 mg)
Notes	No wash‐out Order effect
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Tablets of imipramine and identical inert placebo were used in a double‐blind cross‐over study. The order of treatment was pre‐determined by a random number list stratified for sex and whether or not the child had previously been treated with Imipramine.
Allocation concealment (selection bias)	Unclear risk	Not stated. Used a random number list stratified for sex and whether or not child previously been treated with imipramine.
Blinding (performance bias and detection bias) participants All outcomes	Low risk	"Throughout the study, only the pharmacists knew which preparation each child was having. The code was broken at the end of the study."
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	"Throughout the study, only the pharmacists knew which preparation each child was having. The code was broken at the end of the study."
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Mother. "Throughout the study, only the pharmacists knew which preparation each child was having. The code was broken at the end of the study."
Incomplete outcome data (attrition bias) All outcomes	Low risk	All 29 accounted for.
Selective reporting (reporting bias)	Unclear risk	Study protocol not to hand.
Other bias	Unclear risk	None stated

Alderton 1967#.

Methods	RCT (double‐ blind placebo‐controlled cross‐over, children randomly assigned to each of 6 possible drug condition sequences) Systematic baseline measure of wetting: Yes Organic causes excluded: No Daytime wetting excluded: Not mentioned Setting: child care institution, Toronto, Canada
Participants	No. of children (boys): 12 (12) Dropouts: 0 Inclusion: boys with character disorders, 1 borderline schizophrenic Ages: 7 to 10 years, median 8 years 10 months
Interventions	A (12): imipramine 25 mg at night B (12): imipramine 25 mg in afternoon C (12): imipramine 25 mg at night and 25 mg in afternoon Each arm preceded by matching placebo phase for 2 weeks, and final arm followed by matching placebo phase Duration of treatment: 28 weeks (including 4 week observation phases before and after trial) Follow‐up: none
Outcomes	No. of wet nights aggregated: active arms (A + B + C): 423/863 (= 3.43 wet nights per week) placebo arms (pre‐A + pre‐B + pre‐C + final): 379/586 (= 4.53 wet nights per week)
Notes	No SDs Days when medication given late, or nights away from institution, not counted Placebo phases considered as equivalent to wash‐out phases
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Children were randomly assigned to each of the 6 possbile drug condition sequences
Allocation concealment (selection bias)	Unclear risk	Unclear. Not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	"Double blind. Imipramine and matching placebo”.
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	"Double blind. Imipramine and matching placebo”
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Childcare staff. "Double blind. Imipramine and matching placebo”
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	12 accounted for in the data but that was 12 that started. However, they mentioned early in the paper that additional participants were added after the trial started. No mention was made of this in the rest of the paper.
Selective reporting (reporting bias)	Unclear risk	Protocol not to hand
Other bias	High risk	Geigy supplied drug and placebo? Anything else to the relationship?

Alderton 1970#.

Methods	Design: RCT (double‐blind placebo‐controlled cross‐over, children randomly assigned to each of 6 possible drug condition sequences) Systematic baseline measure of wetting: Yes Organic causes excluded: No Daytime wetting excluded: Not mentioned Setting: child care institution, Toronto, Canada
Participants	No. of children (boys): 9 (8) Dropouts: 2 Inclusion: residential treatment for behaviour disturbances, 4 normal IQ, 4 low average IQ, 1 borderline IQ Ages: 7 to 10 years (mean 8 years 11 months)
Interventions	A (7): imipramine 25 mg in the afternoon and 25 mg at night B (7): imipramine 50 mg in afternoon C (7): imipramine 50 mg at night Each arm preceded by matching placebo phase for 4 weeks, and final arm followed by matching placebo phase Duration of treatment: 36 weeks (including 4‐week observation phases before and after trial) Follow‐up: none
Outcomes	No. of wet nights aggregated: active arms (A + B + C): 212/683 (= 2.17 wet nights per week) placebo arms (pre‐A + pre‐B + pre‐C + final): 491/814 (= 4.22 wet nights per week) P < 0.001
Notes	No SDs Unclear whether dropouts counted in outcome data Days when medication given late, or nights away from institution, not counted Placebo phases considered as equivalent to wash‐out phases
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Participants were randomly assigned to the 6 different drug condition sequences
Allocation concealment (selection bias)	Unclear risk	Unclear. Not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	"double‐blind investigation conducting using...imipramine tablets and matching placebo"
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	"double‐blind investigation conducting using...imipramine tablets and matching placebo"
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Child‐care staff. "double‐blind investigation conducting using...imipramine tablets and matching placebo"
Incomplete outcome data (attrition bias) All outcomes	High risk	2 did not complete trial. “One left before the final no medication period and one discharged after the 2nd drug condition.” No explanation given. No further information given regarding the treatment regime for the 2 participants
Selective reporting (reporting bias)	Unclear risk	Protocol not available
Other bias	High risk	? sponsored by drug company Geigy

Attenburrow 1984.

Methods	Design: RCT (double‐blind) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: No
Participants	No. of children (boys) 33 (11) Dropouts: 13 Inclusion: suitable for drug therapy; parental consent; no abnormalities in blood or urine Daytime wetting included Most had received simple preliminary treatments e.g. lifting and fluid restriction Median age 7 years (range 5 to 13) Baseline wetting: mean number of dry nights in week A: 2.8, B: 2.4, C: 1.3
Interventions	A (12): viloxazine (100 mg for under‐10s; 150 mg for over‐10s) B (9): imipramine (50 mg for under‐10s; 75 mg for over‐10s) C (12): placebo Duration of treatment: 7 weeks Follow‐up: 2 weeks
Outcomes	Mean number (SD) of dry nights in final week A: 4.4 (3.8), B: 3.8 (4.5), C: 1.3 (2.4) Wet A: 2.6 (3.8), B: 3.2 (4.5), C: 5.7 (2.4) Follow‐up: mean number (SD) of dry nights A: 4.1 (2.8), B: 2.8 (4.8), C: 1.3 (2.1) Wet A: 2.9 (2.8), B: 4.2 (4.8), C: 5.7 (2.1) Side effects: A: dizziness and flu with sinus blockage (1); headache (1); lethargy (1) B: lethargy (4); constipation (3); upset stomach (2); vomiting, sweating and shakiness (1); vomiting and drowsiness leading to withdrawal (1); dizziness and dry mouth (1); anorexia (1) C: rash (2); nightmares (1)
Notes	Includes some children with diurnal enuresis and encopresis Few details about dropouts (appear to have been more from imipramine group) Groups well‐matched for age and social class and no significant difference in baseline wetting Not intention‐to‐treat analysis
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	“children were randomly assigned under double‐blind conditions to treatment with...”
Allocation concealment (selection bias)	Unclear risk	Unclear. Not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	"...children were randomly assigned under double‐blind condition..."
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	"...children were randomly assigned under double‐blind condition..."
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Parents assessors. "...children were randomly assigned under double‐blind condition..."
Incomplete outcome data (attrition bias) All outcomes	High risk	46 included. 13 withdrawn, 2 permission revoked, 6 defaulted, 4 UTI, 1 side effects. Unsure which treatment groups these belonged to and whether data excluded or included these withdrawn participants. No intention‐to‐treat analysis.
Selective reporting (reporting bias)	Low risk	Number of dry nights
Other bias	Unclear risk	None stated

Banerjee 1993.

Methods	Design: CCT (alternate allocation) Systematic baseline measure of wetting: No Organic causes excluded: Yes Daytime wetting excluded: Not mentioned
Participants	No. of children (boys): 50 (30) Inclusion: nocturnal enuresis every night Exclusion: medical and surgical causes for enuresis Age: 5 to 16 years Baseline wetting: 7 times per week
Interventions	A (25): hypnosis, 2 x 30‐min sessions in 1st week, 1 in 2nd week, then every 1 to 2 weeks for 3 months B (25): imipramine (25 mg every night, increased weekly if no response) for 3 months Follow‐up at 1, 2, 3 and 6 months
Outcomes	Dry or improved rates at 3 months: A: 18/25, B: 19/25 No. relapsing: A: 1, B: 13 Failure to improve or relapse rate: A: 8/25, B: 19/25 Maximum effect for A by 2nd week, B by 4th week
Notes	No useable data Low relapse in hypnosis group attributed to continuing self hypnosis A less effective if age < 7 years
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Participants were assigned alternately on admission to hypnosis and imipramine
Allocation concealment (selection bias)	High risk	No concealment
Blinding (performance bias and detection bias) participants All outcomes	High risk	Not blinded
Blinding (performance bias and detection bias) study personnel All outcomes	High risk	Not blinded
Blinding (performance bias and detection bias) outcome assessors All outcomes	High risk	Not blinded
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	4 dropouts from the hypnosis group, 5 from the imipramine group. No further information
Selective reporting (reporting bias)	Unclear risk	Protocol not stated
Other bias	Unclear risk	Not stated

Batislam 1995.

Methods	Design: RCT Systematic baseline measure of wetting: No Organic causes excluded: Yes Daytime wetting excluded: Not mentioned
Participants	No. of children (boys): 78 (48) Inclusion: primary nocturnal enuresis; at least 3 wet nights/week Exclusion: organic causes Previous treatment: none Age: 6 to 18 years
Interventions	A (16): imipramine mg/kg/d B (20): diclofenac sodium C (30): imipramine + diclofenac D (12): placebo Duration of treatment 4 weeks Follow‐up at 1 and 3 months
Outcomes	No. improving >5 0%: A: 2/16, B: 4/20, C: 2/30, D: 6/12 No. relapsing: A: 6, B: 0, C: 4, D: 0 Failure to improve or relapse rate: A: 16/16, B: 16/20, C: 26/30, D: 6/12 Adverse events: 8 had mild gastrointestinal symptoms, but not clear which groups
Notes	No useable data Unexplained difference in size of groups at allocation
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Patients were divided at random into 4 groups.
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	Unclear risk	Not stated
Blinding (performance bias and detection bias) study personnel All outcomes	Unclear risk	Not stated
Blinding (performance bias and detection bias) outcome assessors All outcomes	Unclear risk	Parents. Not stated
Incomplete outcome data (attrition bias) All outcomes	Low risk	No attrition
Selective reporting (reporting bias)	High risk	No protocol. No systematic baseline measurement of wetting. Adverse events mentioned but treatment groups not mentioned
Other bias	High risk	Unexplained difference in size of groups at allocation

Bhatia 1990.

Methods	Design: RCT (matched on age and sex) Systematic baseline measure of wetting: No Organic causes excluded: Yes Daytime wetting excluded: No Setting: Child Guidance Clinic, New Delhi, India
Participants	No. of children (boys): 60 (39) Dropouts: 22 due to irregular follow‐up Inclusion: nocturnal enuresis (52), diurnal enuresis (24), daytime wetting (6); behaviour problems Age: 4 to 12 years
Interventions	A (20): placebo + simple behavioural intervention (restrict fluids, avoid punishment) B (20): imipramine 10 mg (age 3 to 6); 25 mg (> 6 years), doubled after 2 weeks if no improvement, stopped 4 weeks after cure C (20): imipramine + simple behavioural intervention Duration of treatment: 6 weeks Follow‐ up: 6 months
Outcomes	No. not achieving 14 dry nights or relapsing after: A: 16/20, B: 8/20, C: 2/20
Notes	High dropout rate due to inadequate follow‐ up
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"60 cases were divided into three groups (age and sex matched) randomly irrespective of type and frequency of enuresis". No further information
Allocation concealment (selection bias)	Unclear risk	Unclear. No information
Blinding (performance bias and detection bias) participants All outcomes	High risk	Treatment groups would have been aware of the treatment arms based on the treatment given No blinding noted
Blinding (performance bias and detection bias) study personnel All outcomes	High risk	As above
Blinding (performance bias and detection bias) outcome assessors All outcomes	Unclear risk	Not stated. Outcome assessors not stated
Incomplete outcome data (attrition bias) All outcomes	High risk	High dropout rate. 22 excluded due to irregular follow‐up out of 82. Not enough information regarding the dropouts
Selective reporting (reporting bias)	Unclear risk	Protocol not available
Other bias	Unclear risk	Not stated

Bindelglas 1968.

Methods	Design: RCT (double‐blind) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned
Participants	No. of children (boys): 63 (59) Mean age 11 years 6 months Baseline wetting: unclear
Interventions	A : imipramine (25 mg for 11 years and under, 50 mg for those over 11 years) B : placebo Duration of treatment: 1 month All participants given drug after 1 month then followed up for up to 24 months Also 10‐year follow‐up
Outcomes	No significant difference was found between placebo and baseline Drug performed significantly better than placebo or baseline during first month At 10‐year follow‐up (all children treated), no negative effects noted in 29 children, age range then 16 to 25
Notes	No useable data Not reported if comparable groups No details for dropouts No explicit entry criteria No details previous treatment
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"a random assignment of the drug or placebo was continued for one month". No further information
Allocation concealment (selection bias)	Unclear risk	Unclear. Not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	"a double blind study..."
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	as above
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Parents. as above
Incomplete outcome data (attrition bias) All outcomes	High risk	48 of 100 followed up. No information given regarding those not followed up
Selective reporting (reporting bias)	Unclear risk	Only one group of results were reported
Other bias	Unclear risk	Not stated

Burke 1995.

Methods	Design: RCT (multicentre, double blind) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned
Participants	No. of children: 45 (boys: A: 11, B: 10, C: 9) Dropouts: A: 0, B: 3, C: 3 Inclusion: age 6 to 17 years; at least 3 wet nights per week for preceding 3‐month period and not dry for more than 6 months Exclusion: enuresis treatment in preceding 6 months; nocturnal enuresis of neurogenic origin; UTI; abnormal urinalysis haematology or blood biochemistry; concomitant medication known to interfere with study medication Mean (SD) age: A: 8.6 (2.4), B: 8.9 (2.5), C: 8.9 (2.4) (range 6 to 14) Baseline wetting: mean (SD) number of wet nights per week: A: 5.8 (0.9), B: 6.0 (0.9), C: 6.3 (0.9)
Interventions	A (14): amitriptyline hydrochloride (25 mg or 50 mg) B (17): desmopressin (20 µg) intranasally C (14): desmopressin + amitriptyline Duration of treatment: 16 weeks Follow‐up: 12 weeks
Outcomes	Mean (SD) number of wet nights per week: A: 3.3 (1.9), B: 4.7 (1.7), C: 3.3 (2.5) Number attaining cure: A: 3, B: 1, C: 5 7 out of 8 children who were cured relapsed. The exception was treated with amitriptyline + desmopressin Follow‐up mean (SD) number of wet nights per week: A: (n = 10) 3.9 (2.9), B: (n = 5) 3.8 (1.9), C: (n = 8) 5.1 (3.2) Side effects: none reported Most parents said all the drugs were easy to use
Notes	No significant difference between groups in terms of number, age, height and weight Trial prematurely halted due to 1 drug ceasing to be available Not stated if intention‐to‐treat Not full quota of participants
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The study was designed as a randomised double‐blind parallel group comparison. To randomise patients at each centre, sealed envelopes used in sequence were opened by the pharmacy department.
Allocation concealment (selection bias)	Low risk	"To randomise patients at each centre, sealed envelopes used in sequence were opened by the pharmacy department."
Blinding (performance bias and detection bias) participants All outcomes	Low risk	“Amitriptyline...and placebo were administered as one tablet at bed time...or two tablets...” (depending on age).  “Desmopressin and placebo were supplied as...dropper bottles with a...”
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	“Amitriptyline...and placebo were administered as one tablet at bed time...or two tablets...” (depending on age).  “Desmopressin and placebo were supplied as...dropper bottles with a...”
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Parent. “Amitriptyline...and placebo were administered as one tablet at bed time...or two tablets...” (depending on age).  “Desmopressin and placebo were supplied as...dropper bottles with a...”
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Group 1 amitriptyline, all 13 completed. Group 2 demospressin, 3 withdrew from total of 17. No further information given. Group 3 combined, 3 were withdrawn for poor compliance from total 14
Selective reporting (reporting bias)	High risk	Side effects not reported, ITT not mentioned
Other bias	Unclear risk	Premature termination of trial. Not sure whether this may have impacted on the results.

Ciotti 1983.

Methods	Design: RCT (divided at random into 2 groups) Systematic baseline measure of wetting: No Organic causes excluded: No Daytime wetting excluded: Yes Setting: Paediatric Day Hospital, Cesena, Italy
Participants	Number of children (boys): 21 (12) Inclusion: nocturnal primary enuresis Exclusion: diurnal enuresis Age: 6 to 14 years Baseline wetting: at least 20/30 nights
Interventions	A (12): Imipramine 25 mg at night B (9): Psychotherapy (monthly) for 6 months
Outcomes	Improved after 3 months (< 5 wet nights per month): A: 3/12, B: 1/9 (P < 0.05) Adverse effects: A: 1 (agitation and depression), B: 0
Notes	No useable data Focus of paper was on relationship between enuresis and psychological/behavioural disorders
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not stated
Allocation concealment (selection bias)	Unclear risk	Unclear. Not stated
Blinding (performance bias and detection bias) participants All outcomes	High risk	No blinding
Blinding (performance bias and detection bias) study personnel All outcomes	High risk	As above
Blinding (performance bias and detection bias) outcome assessors All outcomes	High risk	As above
Incomplete outcome data (attrition bias) All outcomes	Low risk	No dropouts
Selective reporting (reporting bias)	Unclear risk	No information
Other bias	Unclear risk	No information

Danquah 1975.

Methods	Design: RCT (but mention of matching) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned Setting: Ghanaian fishing community
Participants	No. of children (boys): 30 (all boys) Exclusion: more than a week of traditional treatment Mean age: 10.4 years Mean frequency of wetting at baseline: A: 5.6 B: 4.00 C: 3.20 Groups comparable in age and intelligence
Interventions	A (10): traditional shaming B (10): amitriptyline hydrochloride C (10): alarm Duration of treatment: 7 weeks Follow‐up after 3 months
Outcomes	Mean frequency of wetting after treatment A: 5.6, B: 4.0, C: 3.2 Participants in traditional shaming seemed depressed and evidence of loss of self esteem and of participants isolating themselves from friends Drug treatment was said to cause drowsiness at first Parents not disturbed by alarm because they slept outside
Notes	No details of dropouts No SDs No details of previous treatment Traditional shaming consisted of being carried from home by a singing mob and being thrown into the lagoon
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	All the 30 children were matched as closely as possible in age and measured intelligence.  The 3 groups of 10 each were then allocated randomly to the 3 forms of treatment.
Allocation concealment (selection bias)	Unclear risk	Unclear. Not stated
Blinding (performance bias and detection bias) participants All outcomes	High risk	Not blinded study. All knew which arm they were in
Blinding (performance bias and detection bias) study personnel All outcomes	High risk	Not blinded study. All knew which arm they were in
Blinding (performance bias and detection bias) outcome assessors All outcomes	High risk	Parents. Not blinded study. All knew which arm they were in
Incomplete outcome data (attrition bias) All outcomes	High risk	Attrition was not mentioned
Selective reporting (reporting bias)	High risk	“response to treatment defined as 21 consecutive dry nights…relapse…defined as a return to a wetting frequency of once or more per week after initial response to treatment…” Results for these were not presented. Rather only frequency of wetting was presented
Other bias	Unclear risk	Not stated

Drew 1966.

Methods	Design: RCT (divided into 2 nearly equal groups on a random basis; double‐blind cross‐over trial) Systematic baseline measure of wetting: Yes Organic causes excluded: No Daytime wetting excluded: Not mentioned Setting: Children's homes in Melbourne, Australia
Participants	No. of children: 28 (only 11 in cross‐over trial) Inclusion: at least 3 wet beds/week Exclusion: none Ages: 5 to 15 years Baseline wetting: A: 65.8% wet nights, B: 64.7%
Interventions	A (11): placebo then imipramine B (11): imipramine then placebo Dose increased from 2 to 4 tabs if any wet beds in first week Duration of treatment: 4 weeks on each treatment, then continued on imipramine or placebo
Outcomes	Wet nights during first arm of trial: A: 56.9% (3.79 nights), B: 39.2% (2.38 nights) Adverse events: None
Notes	Data from first arm reported No SDs No wash‐out
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	11 of these children who were in the 1 institution were divided into 2 nearly equal groups on a random basis. (During the last 2 weeks of the trial), “ observation revealed the presence of a further 17 enuretic children in the institutions. The 28 enuretics were then divided in two new groups on a random basis."
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	"Each child in Group 1 was given two tablets of placebo (supplied as tablet B)... At the end of the four‐week period, each child in Group 1 was given imipramine (tablet A, identical in appearance to tablet B) for four weeks... Each child in Group 2 was treated in exactly the same way as those in Group 1, except that imipramine therapy was followed by, rather than being preceded by, the giving of the placebo." "The experimental procedure was such that it was not known at any given time, until all observations had been made, which tablets the patients were receiving. The identity of imipramine with tablet A was not revealed until all results and calculations were completed."
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	As above
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	staff at institution, not study personnel. As above
Incomplete outcome data (attrition bias) All outcomes	Low risk	No dropouts
Selective reporting (reporting bias)	Unclear risk	Study protocol not given
Other bias	High risk	Acknowledgement was made for the support of Geigy Pharmaceuticals. ? Further relationship.

Esmaeili 2008.

Methods	Design: RCT Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting specifically excluded: Yes Setting/recruitment: University hospital clinic
Participants	Number of participants (boys):112 (69) Number of dropouts: 23 (17 boys, 6 girls) Inclusion criteria: PNE Exclusion criteria: Prior drug therapy for PNE, secondary PE, abnormal U/A, urologic and neurologic abnormalities Ages: 6 to 14 (8.9) Baseline wetting: mean wet night/week : group A (4.9 ± 1.12), B (5 ± 1.1), C (5.4 ± 1)
Interventions	A   (29) Imipramine: 10 mg participants weighing < 30 kg and 25 mg > 30 kg B   (26) Oxybutynin: 3.75 participants weighing < 30 kg and 5 mg > 30 kg C   (34) Imipramine ± Oxybutinin: as above Duration of treatment: 1 month Duration of follow‐up; 1 month
Outcomes	Mean wet nights at end of trial: A (3.5 ± 2), B (2.5 ± 1.7), C (1.4 ± 1.5) No. not achieving 14 dry nights at end of trial: A (25), B(20), C(6) No. not achieving 14 dry nights at end of trial or relapsing after: A (13), B (9), C (6) Adverse events (numbers and details): None
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	The children were divided into 3 groups randomly according to their referral sequence number of the clinic
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	High risk	Not blinded
Blinding (performance bias and detection bias) study personnel All outcomes	High risk	Not blinded
Blinding (performance bias and detection bias) outcome assessors All outcomes	High risk	Parents. Not blinded
Incomplete outcome data (attrition bias) All outcomes	High risk	17 lost to follow‐up in analysis of relapse at 1 month follow‐up. No further information given Otherwise exclusions dealt with. 21% dropout rate
Selective reporting (reporting bias)	Unclear risk	No protocol available. Wet nights pretreatment compared with during treatment and cure and relapse rate data presented
Other bias	Unclear risk	None stated

Forsythe 1969.

Methods	Design: RCT (allotted at random, stratified by sex) Systematic baseline measure of wetting: No Organic causes excluded: No Daytime wetting excluded: Not mentioned Setting: Royal Belfast Hospital for Sick Children, Northern Ireland
Participants	No. of children (boys): 298 (190) Dropouts: 51 (A: 22, B: 11, C: 12 plus another 2 from each group due to 'toxic reactions') Inclusion: persistent bedwetters (95% primary enuresis), no UTI Ages: 5 to 15 years + 6 children just under 5 years Baseline wetting: > 6 times per week for previous year
Interventions	A (78): imipramine 50 mg + placebo of nortriptyline B (87): placebo of imipramine + placebo of nortriptyline C (88): placebo of imipramine + nortriptyline 25 mg Duration of treatment: 8 weeks Follow‐up: 8 weeks
Outcomes	No. not achieving 14 dry nights during treatment (defined as no wet nights in 56): A: 77/78, B: 87/87, C: 88/88 No. not achieving 14 dry nights after treatment (defined as no wet nights in 56): A: 76/78, B: 86/87, C: 87/88 Adverse events: 'toxic reactions' causing dropout: A: 2/78 (1 abdominal pain, 1 apathy + vomiting), B: 2/87 (1 wetting worse, 1 depressed + crying), C: 2/86 (1 rash, 1 nausea)
Notes	Groups comparable at baseline on age and sex Reasons for dropout specified and not related to groups Stricter definition of cure Older children more likely to improve
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"A randomisation scheme was prepared and a medical secretary was responsible for the scheme and distribution of tablets. Mothers were told that these were new tablets which had not been tried before"
Allocation concealment (selection bias)	Unclear risk	"Children were allotted at random to one of the (treatment) regimes..."
Blinding (performance bias and detection bias) participants All outcomes	Low risk	"...was not possible to have identical Tofranil and Allegron tablets, it was decided to use double placebos."
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	."...was not possible to have identical Tofranil and Allegron tablets, it was decided to use double placebos."
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Asssessors not stated...."...was not possible to have identical Tofranil and Allegron tablets, it was decided to use double placebos."
Incomplete outcome data (attrition bias) All outcomes	Low risk	"...overall omission rate...was high 51 (17.1%) but...the percentage number of omissions did not differ significantly...between the three treatment regimes." Group A impramine: 18 did not reply, 2 went abroad, 2 refused to cooperate, 2 developed side effects – 1 severe pain, 1 vomiting. Group B placebo: 9 did not reply, 2 refused to cooperate, 2 developed side effects – 1 claimed wetting became worse, 1 depressed and cried ++. Group C amitriptyline: 11 did not reply, 1 went abroad, 2 developed side effects –1 rash and drowsy, 1 nausea.
Selective reporting (reporting bias)	Low risk	Results average no of wet nights and treatment results as delineated by authors were reported in tables.
Other bias	High risk	Supply of tablets from drug companies acknowledged.

Forsythe 1972a.

Methods	Design: RCT Systematic baseline measure of wetting: No Organic causes excluded: Yes Daytime wetting excluded: Not mentioned Setting: Royal Belfast Hospital for Sick Children, Northern Ireland
Participants	No. of children (boys): 240 (131) Dropouts: A: 31, B: 23 Inclusion: at least 6 wet nights per week for at least 1 year (98% had primary enuresis) Exclusion: UTI Age: 4 to 14 (but only 1 child aged 4 years 11 months)
Interventions	A (90): 25 mg trimipramine, increasing by 1 tab per week if no response B (96): identical placebo in corresponding doses Duration of treatment: 8 weeks Follow‐up: 8 weeks
Outcomes	No. not achieving 14 dry nights during treatment (failed): A: 90/90, B: 96/96 Failure + relapse after end of trial: A: 88/90, B: 93/96 Adverse events: None
Notes	Groups and dropouts comparable
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"A randomisation scheme was prepared for each sex, and a medical secretary was responsible for the scheme and distribution of the tablets".
Allocation concealment (selection bias)	Unclear risk	Unclear
Blinding (performance bias and detection bias) participants All outcomes	Low risk	The trimipramine tablets (25mg) and placebo were identical in size, shape and colour.
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	As above
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	As above. Unsure who assessors are.
Incomplete outcome data (attrition bias) All outcomes	Low risk	186/240 completed. Table that shows number of participants omitted from trial has heading “reason for omission” but no reasons were given. However, no significant differences in percentage omissions between placebo and treatment
Selective reporting (reporting bias)	Unclear risk	Protocol not stated
Other bias	High risk	May & Baker Ltd supplied the drugs.

Forsythe 1972b.

Methods	RCT (double‐blind placebo‐controlled, stratified by sex) Systematic baseline measure of wetting: No Organic causes excluded: Yes Daytime wetting excluded: Not mentioned Setting: Belfast Hospital for Sick Children, Northern Ireland
Participants	No. of children (boys): 241 (140) Dropouts before start of study: A: 10, B: 16 Dropouts before end of Follow‐up: A: 6, B: 9 Inclusion: nocturnal enuresis at least 6 times per week in previous year Ages: 5 to 14 years
Interventions	A (121): chlordiazepoxide (5 mg) + amitriptyline (12.5 mg) B (120): Placebo Duration of treatment: 8 weeks Follow‐up: 8 weeks
Outcomes	No. not achieving 14 dry nights: A: 109/111, B: 104/104 No. improving by 50%: A: 50/111, B: 88/104 No. not achieving 14 dry nights or relapsing after: A: 103/104, B: 95/96 Improved (at least 50% dry nights): A: 69/104, B: 82/96
Notes	Chlordiazepoxide used to suppress emotional stimuli Groups and dropouts comparable
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	A medical secretary was responsible for the randomisation scheme for each sex and distribution of the tablets.
Allocation concealment (selection bias)	Unclear risk	For each sex separately the children were allotted at random to one of the regimens.
Blinding (performance bias and detection bias) participants All outcomes	Low risk	Identical drug and placebo tablets were provided.
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	Identical drug and placebo tablets were provided.
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Identical drug and placebo tablets were provided.
Incomplete outcome data (attrition bias) All outcomes	Low risk	The percentage number of omissions did not differ significantly (P < 0.05) between the 2 treatment regimens. 26 omissions. Active group: 8 did not reply, 1 went abroad, 1 had N&V stopped treatment. Placebo group: 12 did not reply, 1 went abroad, 3 refused to swallow tablets. Dropouts comparable
Selective reporting (reporting bias)	Unclear risk	Protocol not stated
Other bias	Unclear risk	None stated

Fournier 1987.

Methods	Design: RCT (double‐blind) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting: Not mentioned Differences in baseline severity of wetting ‐ MANOVA used
Participants	No. of children (boys): 64 (47) completed the study 5 extra children dropped out Inclusion: no treatment in past 3 months Mean age: 8 years 5 months Baseline wetting: mean number of wet nights in week 2: A: 5.3, B: 6, C: 4.5, D: 4.2, E: 4.5
Interventions	A: (8) imipramine B: (8) enuresis alarm C: (8) placebo D: (8) random awakening E: (8) alarm + imipramine F: (8) Alarm + placebo G: (8) random awakening + placebo H: (8) imipramine + random awakening Duration of treatment: 6 weeks Follow‐up: 3 months but some children continued on treatments
Outcomes	Mean number of wet nights per week: A: 1.9, B: 2.5, C: 5, D: 3.3, E: 1 No results for F, G or H
Notes	Parallel groups 4 boys dropped out because of side effects or noncompliance, 1 girl with UTI No SDs
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomly assigned.
Allocation concealment (selection bias)	Unclear risk	Unclear. Not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	Children, parents and the psychiatrist interviewing the children were blind to the treatment the child received. Following a double‐blind procedure, a child received active medication of 25mg tablets of Tofranil…The placebo was an identical nonlactose tablet. Only the tablets could be blinded. Alarm and random awakening couldn’t be blinded.
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	Children, parents and the psychiatrist interviewing the children were blind to the treatment the child received. Following a double‐blind procedure, a child received active medication of 25mg tablets of Tofranil…The placebo was an identical nonlactose tablet. Only the tablets could be blinded. Alarm and random awakening couldn’t be blinded.
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Parents. Children, parents and the psychiatrist interviewing the children were blind to the treatment the child received. Following a double‐blind procedure, a child received active medication of 25mg tablets of Tofranil…The placebo was an identical nonlactose tablet. Only the tablets could be blinded. Alarm and random awakening couldn’t be blinded.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	64 completed study. 59 at 3‐month follow‐up. No information provided
Selective reporting (reporting bias)	Low risk	Results were related to dry/wet nights over period of study
Other bias	Unclear risk	Adverse events not mentioned

Friday 1966.

Methods	Design: RCT (double‐blind) Systematic baseline measure of wetting: No Organic causes excluded: Yes Daytime wetting: No (25% had daytime wetting) Setting: outpatient clinic
Participants	No. of children (boys): 51 (27) Inclusion: failed other treatment, 80% primary enuresis, 25% daytime wetting, most had "nervous symptoms" Exclusion: Infection, renal anomaly, obstructive uropathy Previous treatment: all had failed with some of: fluid restriction, bribes, punishment, scheduled wakening, psychotherapy or drugs Ages: 4 to 15 years
Interventions	A (22): imipramine B (29): placebo Duration of treatment: 2 weeks Follow‐up: 2 weeks
Outcomes	No. failing to improve (defined as 50% or more decrease in wet nights): A: 4/22, B: 16/29 Side effects: A: 1 (neutropenia, rash)
Notes	No useable data Includes 25% children with daytime wetting Groups comparable at baseline
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Either imipramine or a matched placebo in identical packaging were dispensed from a random group of coded bottles containing 14 tablets, with the physician not knowing which was being given.
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	Either imipramine or a matched placebo in identical packaging were dispensed from a random group of coded bottles containing 14 tablets, with the physician not knowing which was being given.
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	Either imipramine or a matched placebo in identical packaging were dispensed from a random group of coded bottles containing 14 tablets, with the physician not knowing which was being given.
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Parents. Either imipramine or a matched placebo in identical packaging were dispensed from a random group of coded bottles containing 14 tablets, with the physician not knowing which was being given.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No attrition or exclusion
Selective reporting (reporting bias)	High risk	Study protocol not available.No systematic baseline measure of wetting
Other bias	Unclear risk	Not stated

Haegglund 1964.

Methods	Design: CCT (every 3rd child to A, every 4th child to B) Systematic baseline measure of wetting: No Organic causes excluded: Yes Daytime wetting excluded: Yes
Participants	No. of children (boys): 34 (27) 19 children with daytime wetting excluded from results Dropouts: 0 Inclusion: Normal IQ, wetting every night Exclusion: UTI, other urographic abnormality, abnormal EEG Previous treatment: not mentioned Ages: 4 to 14 years
Interventions	A (7): imipramine (age < 7 years 30 mg, > 7 years 50 mg) B (8): no drug treatment All children also received supportive psychiatric management Duration of treatment: 3 months Follow‐up: 3 to 8 months
Outcomes	No. failing to achieve 14 dry nights or relapsing afterwards: A: 4/7, B: 8/8 Side effects: not mentioned
Notes	Data presented only from 15 children with monosymptomatic nocturnal enuresis No fluid restriction or nightly waking Bladder capacity increased more in group A than B
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Every 4th enuretic child admitted was placed in the treatment group and the next following in the control group
Allocation concealment (selection bias)	High risk	No allocation concealment: knew exactly who was where
Blinding (performance bias and detection bias) participants All outcomes	High risk	Those in the medication group got medication while those in the control group had no medication. Not a blinded trial
Blinding (performance bias and detection bias) study personnel All outcomes	High risk	Those in the medication group got medication while those in the control group had no medication. Not a blinded trial
Blinding (performance bias and detection bias) outcome assessors All outcomes	High risk	Unclear assessors. Those in the medication group got medication while those in the control group had no medication. Not a blinded trial
Incomplete outcome data (attrition bias) All outcomes	Low risk	All 34 accounted for. Exclusion discussed
Selective reporting (reporting bias)	High risk	Protocol not available. No systematic baseline measure of wetting
Other bias	Unclear risk	Not stated

Harrison 1970.

Methods	Design: RCT (cross‐over double‐blind) Systematic baseline measure of wetting: Yes Organic causes excluded: No Daytime wetting: Not mentioned Setting: single sex orphanages
Participants	No. of children (boys): 62 (14) Dropouts: A: 2 Inclusion: institutionalised children Ages: 6 to 18 Baseline wetting: A: 62%, B: 66%
Interventions	A (30): imipramine 25 mg (age < 12 years), 50 mg (> 12 years) B (32): placebo Duration treatment: 20 nights each
Outcomes	No. wet nights in first arm: A: 36% (mean 2.52), B: 47.2% (3.3) Wet nights after cross‐over to second arm of trial: A: 55.6%, B: 56.3%
Notes	Data from first arm of trial used Carry‐over effect in second arms No SDs
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	2 groups of equal numbers were selected randomly from lists of bed‐wetters supplied by the institution
Allocation concealment (selection bias)	Unclear risk	not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	"...double‐blind testing, using a placebo and the drug, with the added control of a cross‐over design..."
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	"...double‐blind testing, using a placebo and the drug, with the added control of a cross‐over design..."
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	? orphanage workers. "...double‐blind testing, using a placebo and the drug, with the added control of a cross‐over design..."
Incomplete outcome data (attrition bias) All outcomes	Low risk	2 cases had to be dropped from the trial (1 was hospitalised, and 1 was transferred to another province), both from group A
Selective reporting (reporting bias)	Low risk	Incidence of wet nights reported
Other bias	High risk	? sponsored by Geigy, drug manufacturer

Hoashi 1995.

Methods	Design: RCT (randomly allocated in blocks of 4) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting: Yes Setting: Day clinics at drug company, Japan
Participants	No. of children (boys): 231 (167) Dropouts: 7 not wet at baseline, 1 moved, 2 daytime wetting, 4 delayed treatment, 2 noncompliance Inclusion: wet 10/14 nights Exclusion: daytime wetting, organic causes, family disruption Age: at least 6 years Baseline wetting: mean 12.8 wet nights in 2 weeks
Interventions	A (112): desmopressin 10 mcg nasal drops + placebo tablet B (112): imipramine tablet (25 mg) + placebo nasal drops Duration of treatment: 4 weeks Follow‐up: none
Outcomes	Wet nights during treatment (N, mean, SE) At 2 weeks: A: (111) 9.5 (SE 0.5), B: (111) 9 (0.5) At 4 weeks: A: (109) 8.7 (SE 0.5), B: (109) 8 (0.5) Adverse effects causing stopping: A: 0, B: 1 Other adverse effects: oedema, headache, sleepiness, insomnia, sleep disorder, dizziness, appetite loss, nausea, vomiting, diarrhoea, abdominal pain,eyelid swelling, rash, red eyelid, nose symptoms (itchy nose, bleeding nose, blocked nose, runny nose), fever, tiredness, shaking head, thirsty, dry lips Total number of side effects: A: 29/120, B: 29/118 Other outcomes: wetting score, side effect score, satisfaction score, overall score
Notes	Japanese language Data measured from graphs In this update, 2‐week treatment data not used No follow‐up
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not stated
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	Double‐dummy method, 20 mcg of intranasal desmopressin or placebo spray and 25 mg imipramine tablet or placebo tablet.
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	Double‐dummy method, 20 mcg of intranasal desmopressin or placebo spray and 25 mg imipramine tablet or placebo tablet.
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Unclear assessors. Double‐dummy method, 20 mcg of intranasal desmopressin or placebo spray and 25 mg imipramine tablet or placebo tablet.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	4 completely excluded from the trial: no explanation
Selective reporting (reporting bias)	Unclear risk	No information
Other bias	Unclear risk	No information

Hodes 1973.

Methods	Design: RCT (double‐blind) Systematic baseline measure of wetting: Yes (but not reported) Organic causes excluded: No Daytime wetting: Not mentioned Setting: GP practice, London, UK
Participants	No. of children: 74 Inclusion: postal questionnaire returned or follow‐up by health visitor Ages: 5 to 15 years
Interventions	A (36): imipramine 25 mg (50 mg if age > 6 years) B (38): placebo Duration treatment: 30 days, repeated for up to 5½ months until cured or family decided to stop Follow‐up: none
Outcomes	No. not achieving 14 dry nights: A: 29/36, B: 32/38 Adverse events: macular rash but unclear on which treatment
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not stated. "...on first attenance were randomly selected to be given tablet A or B."
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	The child was asked to keep a chart of wet and dry nights, the mother was encouraged... The tablets were either imipramine 25mg or a placebo ‐ the identification was unknown to the general practitioner until the end of the study. Treatment was offered until cure resulted or treatment was stopped by the patient.
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	The child was asked to keep a chart of wet and dry nights, the mother was encouraged... The tablets were either imipramine 25mg or a placebo ‐ the identification was unknown to the general practitioner until the end of the study. Treatment was offered until cure resulted or treatment was stopped by the patient.
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Mother. The child was asked to keep a chart of wet and dry nights, the mother was encouraged... The tablets were either imipramine 25mg or a placebo ‐ the identification was unknown to the general practitioner until the end of the study. Treatment was offered until cure resulted or treatment was stopped by the patient.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All 74 accounted for
Selective reporting (reporting bias)	Low risk	Seems to be complete. Results given in table form of differences in wet nights and also time for treatment: aim from the Methods.
Other bias	Unclear risk	Not stated

Holt 1986.

Methods	Design: RCT (method not given, double‐blind) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Yes
Participants	No. of children: 36 Inclusion: 2 or more wet nights per week; age 8 to 12; Exclusion: daytime wetting; diabetes, insipidus or other chronic illness where need daily medication; other treatment for bedwetting. Comparable in terms of sex, age and weight Mean age (years): A: 9.8, B: 9.5 (range 8 to 12) Baseline wetting: wet bed 2 or more times per week
Interventions	A (19): imipramine 50 mg and placebo nasal spray B (17): intranasal desmopressin 20 µg and placebo tablets Duration of treatment: 4 weeks Follow‐up: 6 weeks
Outcomes	Results first 2 weeks of treatment: % reduction in wet nights: A: 48%, B: 45% Results final 2 weeks of treatment: A: 54%, B: 32% Mean (SD) number of wet nights per first 14 days: A: 4.9 (4.3), B: 4.8 (4.0) Mean (SD) number of wet nights per last 14 days: A: 4.5 (3.7), B: 6.0 (4.4) Mean (SD) number of wet nights per 14 days at follow‐up: A: 7.7 (3.9), B: 7.3 (4.5) Side effects: B: rash (1)
Notes	Norwegian translation Direct comparison of imipramine and desmopressin In this update, first 2‐week treatment data not used No details of previous treatment
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not stated.
Allocation concealment (selection bias)	Unclear risk	Unclear Not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	“randomised to treatment group, double blind with double placebo”
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	As above
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	As above
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Not stated.
Selective reporting (reporting bias)	Unclear risk	No information
Other bias	Unclear risk	No information

Iester 1991.

Methods	Design: RCT (randomly divided with computer) Systematic baseline measure of wetting: No Organic causes excluded: Yes Daytime wetting excluded: Not mentioned
Participants	No. of children: 168 Inclusions: functional enuresis Exclusion: organic causes, emotional disturbance Previous treatment: none Ages: 6 to 11 years
Interventions	A (36): 6 weeks with imipramine 0.9 ‐ 1.5 mg/kg, max 50 mg B (36): 3‐step programme: a) reassurance to parents; b) bladder retention training and wakening with alarm clock before micturition; c) parental involvement C (96): Motivational therapy (counselling) + education (computer programme) + 3‐step therapy Duration treatment: 6 months Follow up: 12 months
Outcomes	Failure rates: A: 22/36, B: 12/36, C: 15/96 Relapse: A: 2, B: 2, C: 3 Failure + relapse rates: A: 24/36, B: 14/36, C: 18/96 Compliance with treatment: A: 14/36, B: 24/36, C: 81/96
Notes	Also in complex review
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Participants were "randomly divided into 3 groups"
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	High risk	No blinding
Blinding (performance bias and detection bias) study personnel All outcomes	High risk	No blinding
Blinding (performance bias and detection bias) outcome assessors All outcomes	High risk	No blinding
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants accounted for.
Selective reporting (reporting bias)	Unclear risk	Baseline measurements, full data set and adverse effects not mentioned
Other bias	Unclear risk	Not stated

Ingle 1968.

Methods	Design: CCT (alternate allocation to groups) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned
Participants	No. of children: 25 (boys A: 6, B: 5) Inclusions: wetting consistently over 3 years Most had previous treatment from other practitioner without any relief Mean age A: 8.9 years, B: 8 years Baseline wetting: mean frequency of wetting A: 8.8, B: 9
Interventions	A (13): tranquillisers ‐ meprobamate (400 mg daily and hydroxyzine 1 mg/kg daily) B (12): imipramine 25 mg daily increased to 50 mg in some cases Duration of treatment: 6 weeks Follow‐up: 1 week
Outcomes	Mean (SD) frequency of wetting: A: 6.5 (1.19), B: 1.9 (2.11) Number totally dry (not defined) A: 0, B: 5 (2 in A showed some improvement) Nearly 60% who responded relapsed after discontinuation of the drug When tranquilliser group given imipramine 5 became totally dry and 6 showed improvement 2 complained of burning sensation
Notes	No details of blinding No details of dropouts Age, sex distribution and frequency of bed wetting comparable in groups Very short follow‐up No statistical analysis No details how progress monitored or when measurements taken
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Participants was given Schedule A (tranquilisers) or Schedule B (imipramine) for a period of 6 weeks. After the initial period children from Group A were given imipramine and the results compared with the previous one
Allocation concealment (selection bias)	High risk	No concealment
Blinding (performance bias and detection bias) participants All outcomes	Unclear risk	Not stated
Blinding (performance bias and detection bias) study personnel All outcomes	Unclear risk	Not stated
Blinding (performance bias and detection bias) outcome assessors All outcomes	Unclear risk	Not stated
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Not stated
Selective reporting (reporting bias)	Unclear risk	Protocol not at hand
Other bias	High risk	Authors thanked Geigy Limited for their liberal supply of Tofranil and help. What sort of help not defined.

Kang 2003.

Methods	Design: Randomised prospective study
Participants	Children with mean age of 7.9 years
Interventions	Imipramine vs oral desmopressin
Outcomes	Number not achieving 14 dry nights at 12 weeks
Notes	Intention‐to‐treat analysis not performed. In Korean, brief translation received.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not available
Allocation concealment (selection bias)	Unclear risk	Not available
Blinding (performance bias and detection bias) participants All outcomes	Unclear risk	Not available
Blinding (performance bias and detection bias) study personnel All outcomes	Unclear risk	Not available
Blinding (performance bias and detection bias) outcome assessors All outcomes	Unclear risk	Not available
Incomplete outcome data (attrition bias) All outcomes	High risk	Excluded those not completing study from analysis
Selective reporting (reporting bias)	High risk	Not all prespecified outcome measures were mentioned ‐ Organic causes and daytime wetting not mentioned. Full data set not given
Other bias	Unclear risk	Not stated

Khorana 1972.

Methods	Design: RCT (single‐blind cross‐over, random allocation to first arm) Systematic baseline measure of wetting: No Organic causes excluded: Yes Daytime wetting: Not mentioned Setting: psychiatric inpatients in India
Participants	No. of children (boys): 100 (74) Dropouts: A: 8, B: 16 Inclusion: consecutive children with primary enuresis Exclusion: physical or neurological disorder, severe mental retardation, unco‐operative Ages: 5 to 15 years (mean 8.2 years)
Interventions	A (42): imipramine hydrochloride (25 mg raised to 50 mg if age < 10 years and no response, 50 mg raised to 75 mg at age > 10 years) B (34): placebo Duration treatment: 12 weeks Follow‐up: none
Outcomes	No. not achieving 14 dry nights in first arm of trial: A: 23/42, B: 34/34 Adverse events: negligible, not requiring treatment
Notes	Results from first arm used Groups comparable at baseline on age, sex, precipitants, duration of illness, intelligence Therapist knew which treatment was given
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	First 100 patients who satisfied the selection critieria comprised the material. sequence generation not stated
Allocation concealment (selection bias)	Unclear risk	Unclear. Not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	"...either 25mg of imipramine...or the placebo. There was no obvious difference between the two preparations." If no response after 2 weeks, dose was increased to a maximum of 2 or 3 tablets respectively. Both drug and placebo had similar increases.Same dosing used when treatment was crossed over after 12 weeks
Blinding (performance bias and detection bias) study personnel All outcomes	Unclear risk	"...either 25mg of imipramine...or the placebo. There was no obvious difference between the two preparations." If no response after 2 weeks, dose was increased to a maximum of 2 or 3 tablets respectively. Both drug and placebo had similar increases.Same dosing used when treatment was crossed over after 12 weeks.
Blinding (performance bias and detection bias) outcome assessors All outcomes	Unclear risk	No information re assessor.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No information
Selective reporting (reporting bias)	Unclear risk	No information
Other bias	Unclear risk	No information. Incomplete paper

Kolvin 1972.

Methods	Design: RCT Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned
Participants	No. of children (boys): 94 (56) 2 dropouts Inclusion: wetting at least 3 nights a week; age range (not stated); not receiving treatment elsewhere Previous treatment: no details Mean age: 9 years 4 months (range 8 to 10) Baseline wetting: mean number of wet nights per month A: 22.7 B: 22.0 C: 20.9
Interventions	A (35): imipramine B (32): pad and buzzer alarm C (27): placebo Duration of treatment: 2 months Follow‐up after 4 months
Outcomes	Mean number of wet night in final month (% improvement): A: 9.3 (64), B: 9.1 (62), C: 11.0 (53) At follow‐up mean number of wet nights per month (% improvement): A: 13.4 (43), B: 9.3 (64), C: 11.3 (54)
Notes	No details of blinding Not reported if comparable groups Not intention‐to‐treat No SDs
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The children were randomly divided in 3 groups for treatment
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	Unclear risk	Not stated
Blinding (performance bias and detection bias) study personnel All outcomes	Unclear risk	Not stated
Blinding (performance bias and detection bias) outcome assessors All outcomes	Unclear risk	Not stated
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	10 lost at follow‐up of 4 months. 1 from placebo, 4 from buzzer and 5 from imipramine. There was an increase in the no. of wet nights in the imipramine 4‐month group. The group also showed a higher loss to follow‐up than the others. This was not addressed. May have been that those lost to follow‐up were cured, leading to misleadingly high mean no. of wet nights
Selective reporting (reporting bias)	Unclear risk	Protocol not available
Other bias	Unclear risk	Not stated

Kumazawa 1990.

Methods	Design: RCT ("blind") Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned
Participants	No. of children 20 (boys A: 7, B: 9) Dropouts: none Inclusion: aged 6 to 16 years; 1 wet night per month; no previous treatment; no organic causes; no urinary infection; parental consent Previous treatment: various punishments and cold water baths Mean age: 8 years Baseline wetting: mean (SD) wet nights per month A: 13.2 (9.7), B: 16.6 (7.8)
Interventions	A (10): motivational reinforcement and bladder exercises then placebo B (10): motivational reinforcement and bladder exercises then 25 mg imipramine Duration of treatment: 6 months Follow‐up: none
Outcomes	Mean number of wet nights per month (SD): A: 3.7 (7.15), B: 8.1 (8.3) By end of study, number achieving 80% reduction in wet nights: A: 7, B: 5 Side effects: none
Notes	Groups comparable on all baseline values Unusual participants (children with learning difficulties) Small sample groups Severity of enuresis ‐ 1 wet night during last 3 months No follow‐up Foreign language
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No information. Foreign language
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	RCT blind with placebo arm
Blinding (performance bias and detection bias) study personnel All outcomes	Unclear risk	RCT blind with placebo arm
Blinding (performance bias and detection bias) outcome assessors All outcomes	Unclear risk	outcome assessors? RCT blind with placebo arm
Incomplete outcome data (attrition bias) All outcomes	Low risk	No dropouts
Selective reporting (reporting bias)	Unclear risk	No information. Foreign language
Other bias	Unclear risk	No information. Foreign language

Kunin 1970#.

Methods	Design: RCT (double‐blind randomised cross‐over with 1 week wash‐out) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Yes
Participants	No. of children (boys): 18 (8) Inclusion: average of 3 wet nights per week; no diurnal wetting Previous treatment: no details Mean age: 7.7 years (range 5 to 11) Severity at baseline: mean number of wet nights over 28 days: 19.8
Interventions	A (18): imipramine hydrochloride (25 or 50 mg) B (18): ephedrine sulphate (7.5 or 15 mg) Duration of treatment: 28 days in each condition Follow‐up: not specified
Outcomes	Mean number (SD) of wet nights over 28 days: A: 9.2 (6.72), B: 17.4 (5.8) 5 children became completely dry but all required medication. 8 children were dry > 90% of the time with 2 children needing no further medication
Notes	Unclear if intention‐to‐treat Only part of trial (children with no organic causes) included here
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	“each patient received a coded medication assigned according to a randomised list maintained by an administrative officer who did not participate in the evaluation.”
Allocation concealment (selection bias)	Low risk	“each patient received a coded medication assigned according to a randomised list maintained by an administrative officer who did not participate in the evaluation.”
Blinding (performance bias and detection bias) participants All outcomes	Low risk	Both drugs identical in apperance Cross‐over with wash‐out
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	Both drugs identical in apperance Cross‐over with wash‐out
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Parents. Both drugs identical in apperance Cross‐over with wash‐out
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Did not state how many participants started the trial. No mention of attrition/exclusion
Selective reporting (reporting bias)	Low risk	Data seemed to be complete with wet nights for each child given
Other bias	Unclear risk	Not stated

Lake 1968.

Methods	Design: RCT (double‐blind cross‐over), allocated at random, stratified by doctor Systematic baseline measure of wetting: Yes Organic causes excluded: No Daytime wetting excluded: Not mentioned Setting: community (GP practices)
Participants	No. of children (boys): 54 (37) Inclusion: age 5 to 12; wet at least twice in 2 weeks Baseline wetting: A: 216/350 (62%), B: 291/406 (72%)
Interventions	A (25 + 29): nortriptyline B (29 + 25): placebo Duration of treatment: 16 weeks (2 weeks on each arm with 2 weeks wash‐outs, 2 periods of A and B each) Follow‐up: none
Outcomes	1st arm only: No. wet nights during Rx: A: 769/1512 in 54 children (3.56/week), B: 935/1512 in 54 children (4.39/week) Side effects: A: 1 (headache, sore tummy), B: 3 (headache, drowsy, vomiting, dry mouth, sweating
Notes	Data from all arms aggregated No SDs 2 week wash‐outs between each of 4 treatment periods (2 of A, 2 of B)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The children...were then allocated randomly to order A or order B of treatment, so that half of each doctor's group of children was on either order. The medication was given on a double‐blind basis. Additionally, the participating doctors did not know these orders of treatment.
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	The medication was given on a double‐blind basis. Additionally, the participating doctors did not know these orders of treatment.
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	The medication was given on a double‐blind basis. Additionally, the participating doctors did not know these orders of treatment.
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Parents. The medication was given on a double‐blind basis. Additionally, the participating doctors did not know these orders of treatment.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	54 completed results were available. No mention of attrition. No mention of the number starting
Selective reporting (reporting bias)	Low risk	Expected wet nights reported
Other bias	High risk	Dista Products supplied the drug and placebo and a grant‐in‐aid for the study

Lee 2005.

Methods	Design: RCT (randomly assigned to 1 of 3 groups by dividing a number from a random numbers table by 3) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: 53% also had daytime wetting but results available separately Setting: 2 hospitals, 2003 to 2004
Participants	Number of children (boys): 145 (100) Monosymptomatic enuresis (Trial 1): 68 Polysymptomatic (+ daytime wetting, Trial 2): 77 Dropouts: 13 (drug side effects or outcome unknown: A: 3, B: 3, C: 7) Inclusion criteria: at least 3 wet nights/week Exclusion criteria: drug treatment in 14 days prior to start of study Age: 7.8 years (SD 2.5) (range 5 to 15) Baseline wetting: 6.36/week (SD 1.5)
Interventions	Trial 1 (numbers for monosymptomatic enuresis subgroup) A (22): desmopressin 0.1 or 0.2 mg, oxybutynin 5 mg B (23): desmopressin 0.2 mg increased to 0.4 mg if no response C (23): imipramine 25 mg Given orally before bedtime Duration of treatment: 6 months Follow‐up: none
Outcomes	Trial 1 (monosymptomatic enuresis) Wet nights at 6 months, N mean (SD): A: (22) 0.93 (1.35), B: (23) 0.7 (0.95), C: (23) 2.0 (2.05) No. children cured (excellent response = 0 ‐ 1 wet night per month): A: 14/22, B: 14/23, C: 3/23 No. children failed at end of treatment (not excellent): A: 8/22, B: 9/23, C: 20/23 Trial 2 (daytime wetting) Wet nights at 6 months, N mean (SD): A: (26) 1.2 (1.55), B: (26) 1.23 (0.88), C: (25) 2.63 (2) No. children cured (excellent response = 0 ‐ 1 wet night per month): A: 9/26, B: 7/26, C: 3/25 No. children failed (not excellent): A: 17/26, B: 19/26, C: 22/25 Adverse effects (both trials, including dropouts): A: 0/51, B: 2/52, C: 22/55 A: no adverse effects, B: severe headache C: nausea, decreased appetite and mood changes
Notes	Data obtained from author enabling results to be given separately for enuresis and daytime wetting groups. Only data from enuresis subgroup were included in this review Groups comparable at baseline on age, sex, disease type, baseline wetting Higher side effect and drop‐out rate in imipramine group
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	At study entry the children were randomly assigned to 1 of 3 regimens
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	High risk	In the desmopressin only group Group B dosage was adjusted by increments of 0.2 mg at 2 weeks if the response rate was not significant. Group A had 2 meds. Group C is the group with 1 med that was not adjusted
Blinding (performance bias and detection bias) study personnel All outcomes	High risk	In the desmopressin only group Group B dosage was adjusted by increments of 0.2 mg at 2 weeks if the response rate was not significant. Group A had 2 meds. Group C is the group with 1 med that was not adjusted
Blinding (performance bias and detection bias) outcome assessors All outcomes	High risk	Patient/parent. In the desmopressin only group Group B dosage was adjusted by increments of 0.2 mg at 2 weeks if the response rate was not significant. Group A had 2 meds. Group C is the group with 1 med that was not adjusted
Incomplete outcome data (attrition bias) All outcomes	High risk	Total of 13 patients did not adhere to the study protocol completely. As the causes were related not to disease prognosis but to drug side effects, and as the outcomes were unknown, we did not apply the intent to treat principle to this group.
Selective reporting (reporting bias)	High risk	Only reported those who completed 6 months of treatment
Other bias	High risk	Varying doses of drugs, doses titrated for each participant depending on response

Liederman 1969.

Methods	Design: RCT (double‐blind) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned
Participants	No. of children (boys): 109 (71) Dropouts: 9 Inclusion: diagnosis of functional enuresis Age range 6 to 22 years (mainly < 12 years) Severity of wetting range from 5 to 80 bedwetting incidents per month with at least 20 per month for most participants
Interventions	A (53): desipramine (dosage depends on age, usually 50 to 75 mg) B (47): placebo Duration of treatment: 60 days Follow‐up: none
Outcomes	Number (%) achieving 50% decrease in wetting after 1 month A: 27 (51), B: 13 (28) Number (%) with considerable improvement after 2 months A: 32 (60), B: 21 (45) Number completely dry after 2 months A: 12, B: 3 Side effects: 4 children and 1 adult (postural hypotension, mild abdominal cramps and headaches)
Notes	Comparability of groups not reported Not intention‐to‐treat analysis Some adults
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not stated
Allocation concealment (selection bias)	Low risk	Randomly dispensed by the pharmacist in code‐labelled bottles. Central allocation
Blinding (performance bias and detection bias) participants All outcomes	Low risk	Received either desipramine or placebo, randomly dispensed by the pharmacist in code‐labelled bottles. Neither the physician nor the participant knew which bottle contained the active drug.
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	Received either desipramine or placebo, randomly dispensed by the pharmacist in code‐labelled bottles. Neither the physician nor the participant knew which bottle contained the active drug.
Blinding (performance bias and detection bias) outcome assessors All outcomes	Unclear risk	Unsure who assessor is. Received either desipramine or placebo, randomly dispensed by the pharmacist in code‐labelled bottles. Neither the physician nor the participant knew which bottle contained the active drug.
Incomplete outcome data (attrition bias) All outcomes	High risk	9 exclusions from study because they did not complete the 60‐day course. No reasons or analysis given. No ITT mentioned
Selective reporting (reporting bias)	Unclear risk	Outcomes described in the Methods section are reported. Protocol not to hand
Other bias	Unclear risk	Not stated

Lines 1968#.

Methods	Design: RCT (double‐blind cross‐over trial, from preselected randomised list) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: No Setting: Dept. Child Health, Adelaide, Australia
Participants	No. of children: 36 Dropouts: 10 from 2nd arm (9 due to significant improvement, of whom 8 in A) Inclusions: enuresis at least 2 wet nights/week, no response to lifting Exclusion: organic causes Previous treatment: almost all, + lifting for 4 weeks before trial Ages: 3 years 11 months to 13 years month (mean 7 years 7 months)
Interventions	A (36): amitriptyline B (36): placebo Duration treatment: 3 months each arm
Outcomes	Mean wet nights/week: A: 2.33, B: 3.51 No. not achieving 14 dry nights: A: 30/36, B: 36/36
Notes	No SDs No wash‐out Occasional child moved to 2nd arm early if no improvement
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Which tablet was given first was determined by a nursing sister from a preselected randomised list. "Each patient was given a three month supply of teither amitriptyline tablets...or a placebo of identical appearance...At the end of the three months, cross‐over was affected and the second tablet taken for a further three months".
Allocation concealment (selection bias)	Unclear risk	Not stated.
Blinding (performance bias and detection bias) participants All outcomes	Low risk	Neither the patient nor the writer (who examined all patients) knew the identity of the tablet.
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	Neither the patient nor the writer (who examined all patients) knew the identity of the tablet.
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Author. Neither the patient nor the writer (who examined all patients) knew the identity of the tablet.
Incomplete outcome data (attrition bias) All outcomes	Low risk	10 failed to complete the course. In all but one case, the reason given was that the response was so satisfactory that the tablets had been stopped to see if the improvement was maintained, which it was. "Eight of these nine...were taking amitriptyline..."
Selective reporting (reporting bias)	Unclear risk	Study protocol not available
Other bias	High risk	Merck Sharp and Dohme supplied the amitriptyline and placebo

Manhas 1967.

Methods	Design: CCT (alternate allocation, double‐blind) part cross‐over Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned
Participants	No. of children (boys): 72 (31) Inclusion: regular and consistent bed wetters Previous treatment: no details Age range 5 to 15 years
Interventions	A (29): imipramine (25 mg for under‐12s and 50 mg for over‐12s) B (27): placebo C (8): placebo then imipramine D (8): imipramine then placebo Duration of treatment: 4 weeks each condition Follow‐up: none
Outcomes	Number attaining: complete relief A: 19, B: 1 part relief A: 6, B: 3 no relief A: 4, B: 23 Side effects: abdominal pain (3); giddiness (3) ‐ 1 with placebo; dryness of mouth (1); headache (1); abdominal pain and epistaxis (1)
Notes	Data from parallel groups reported No baseline results Comparability of groups not reported No details of dropout No details of inclusion/exclusion criteria
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Not stated
Allocation concealment (selection bias)	High risk	Inadequate
Blinding (performance bias and detection bias) participants All outcomes	Low risk	"Double blind"
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	"Double blind"
Blinding (performance bias and detection bias) outcome assessors All outcomes	Unclear risk	Unsure who outcome assessors are
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No details given
Selective reporting (reporting bias)	Unclear risk	No details given
Other bias	Unclear risk	Not stated

Martin 1971#.

Methods	Design: RCT (double‐blind randomised cross‐over, using Latin square design) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned Follow‐up after 3 months
Participants	No. of children (boys): 57 (42) Dropouts: 0 Inclusion: 3 wet nights per week for a period of > 6 months Exclusion: organic heart disease; hyperthyroid; glaucoma; diabetes; kidney or liver disease; those taking thyroid, MAO inhibitors or anticholinergics Previous treatment: no details Age range: 5 to 15 years Baseline wetting: mean number of wet nights in 26: 20.7
Interventions	A (57): imipramine pamoate 10 mg (suspension) B (57): imipramine pamoate 25 mg (suspension) C (57): placebo Given orally 1 hour before bedtime Duration of treatment: 26 days in each condition
Outcomes	Mean number of wet nights in 26 days: A: 13.7, B: 10.5, C: 16.8 Side effects: anxiety reaction; constipation; sleep disturbance; abdominal pain; headache; weight loss
Notes	No wash‐out No SDs
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Each patient acted as his own control, and the order in which these treatments were administered was determined by a Latin square rotation.
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	Two doses of imipramine pamoate (10 and 25 mg) and placebo compared in double‐blind cross‐over trial of 26 days on each treatment Liquids of identical appearance and taste were offered for each treatment.
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	As above
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Parents. As above
Incomplete outcome data (attrition bias) All outcomes	Low risk	No dropouts after entering
Selective reporting (reporting bias)	Unclear risk	No study protocol
Other bias	Unclear risk	Not stated

Maxwell 1971#.

Methods	Design: RCT (multicentre double‐blind randomised cross‐over) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned
Participants	No. of children (boys): 135 (84) Dropouts: 10 Details of 125 (84 boys) Inclusion: age 5 to 12; normal except for enuresis; wetting 3+ times per week; home environment guaranteed stable for 8 weeks Exclusion: organic disease; MAO inhibitors within previous 2 weeks Previous treatment: no details Age range: 5 to 12 years Baseline wetting: mean (SD) number of dry nights per 28: 7.0 (7.0)
Interventions	A (125): imipramine (25 mg age 5 to 7 years; 50 mg 8 to 12 years) + star chart B (125): placebo + star chart Duration of treatment: 4 weeks in each condition Follow‐up: none
Outcomes	Mean (SD) number of dry nights per month: A: 16.6 (8.7), B: 13.2 (8.5) Regardless of treatment, results in 2nd month better than first; not carry‐over from drugs, probably due to star charts Side effects A: anorexia (2); diarrhoea (1); constipation (1); depression (1) 77 participants preferred imipramine and 22 preferred placebo Outcomes better (data not shown) if child had not been dry for > a week previously, did not sleep for too long, talked early and was not yet at school
Notes	Groups comparable at baseline No wash‐out Not intention‐to‐treat Cannot see order effect ‐ cross‐over results combined Cannot see role of star chart Very short baseline
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not stated
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	“Bottle labelled syrup‐50 or syrup‐25 depending on age". (“neither patient nor doctor knew whether it contained imipramine or placebo”)
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	“Bottle labelled syrup‐50 or syrup‐25 depending on age". (“neither patient nor doctor knew whether it contained imipramine or placebo”)
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Unknown assessors. “Bottle labelled syrup‐50 or syrup‐25 depending on age". (“neither patient nor doctor knew whether it contained imipramine or placebo”)
Incomplete outcome data (attrition bias) All outcomes	High risk	10 participants failed to complete the 2nd period. 6 placebo and 4 imipramine. No further explanation given. ? nature of participants. No intention‐to‐treat.
Selective reporting (reporting bias)	Unclear risk	No protocol
Other bias	High risk	Authors: one was a medical advisor and the other a statistician for Geigy pharmaceuticals.

McKendry 1975.

Methods	Design: RCT Systematic baseline measure of wetting: No Organic causes excluded: Yes Daytime wetting excluded: No Setting: Paediatric outpatients, Toronto, Canada
Participants	No. of children (boys): 222 (151) No. of dropouts: 53 (A: 9, B: 12, C: 32) Inclusion: primary nocturnal enuresis, "a few" had diurnal wetting Exclusion: organic causes Ages: Mean 9 (range 5 to 17) Baseline wetting (self reported): A: 83.4%, B: 82.3%, C: 87.4%
Interventions	A (73): restricted diet (no dairy, eggs, citrus, tomato, chocolate) B (74): imipramine 10 mg at bedtime, increased to max 40 mg for age 5 ‐ 9, up to 60 mg for 10+ C (75): Mozes Detector (body‐worn detector, sounds alarm + delivers electric shock when a few drops of urine pass) Duration of treatment: 2 months Follow‐up A: 3 month, B: 19 month, C: 14 month
Outcomes	No. not achieving 14 dry nights: A: 63/64, B: 49/62, C: 20/43 (+ 32 dropouts) ITT (including dropouts as failures): A: 72/73, B: 61/74, C: 52/75 A: most parents requested transfer to another treatment within 1 to 2 months Adverse events: A: 2/12 children became aggressive; B: 3/16 had headaches, abdominal pain or fatigue, C: 10/16 showed fear or anxiety about the machine. For C, electric shocks resulted in skin erythema, discolouration, painless cold burns and ulceration
Notes	High dropout/transfer rate from A due to parents finding diet unsuccessful and restricting High dropout rate from C due to parents refusing to allow their child to use the Mozes detector, or finding it too expensive, or children fearing it especially if < age 8 years
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	“Each was assigned to one of three treatments by random allocation.”
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	High risk	Not blinded. Difficulty with blinding given treatment modalities
Blinding (performance bias and detection bias) study personnel All outcomes	High risk	Not blinded. Difficulty with blinding given treatment modalities
Blinding (performance bias and detection bias) outcome assessors All outcomes	High risk	Not blinded. Difficulty with blinding given treatment modalities
Incomplete outcome data (attrition bias) All outcomes	High risk	of the 222 participants who entered the study, 169 completed treatment, 73 were assessed more fully at follow‐up. “...the duration of treatment varied.” Not adequately explained. high dropout rate
Selective reporting (reporting bias)	High risk	Protocol not available but data given for outcomes stated in Methods. Daytime wetting not excluded and no systematic baseline measure of wetting.
Other bias	Unclear risk	Not stated

Mehrotra 1980.

Methods	Design: RCT (stratified by age, sex and socioeconomic status) Systematic baseline measure of wetting: no Organic causes excluded: not mentioned Daytime wetting excluded: not mentioned
Participants	Number of children: 60 Dropouts: 0 Inclusion criteria: children attending outpatients for enuresis Baseline data not provided for individual interventions
Interventions	A: (20) placebo tablets (B complex) and behavioural intervention (waking and lifting 2 hours after bedtime, stars and rewards for dry nights) B: (20) amitriptyline (10 mg ages 3 ‐ 6, 25 mg > 6 years; dose doubled if no improvement after 3 weeks) C: (20) amitriptyline and behavioural intervention Duration: 5 weeks Follow‐up: 4 to 5 months
Outcomes	Failed (i.e. not cured): A: 17/20; B: 6/20; C: 11/20 Failed or relapsed at follow‐up: A: 15/20; B: 10/20; C: 15/20 Adverse events: 2 drowsy with amitriptyline (B or C)
Notes	Parallel groups
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"Divided equally into three groups randomly"
Allocation concealment (selection bias)	Unclear risk	Not mentioned
Blinding (performance bias and detection bias) participants All outcomes	High risk	Not possible to blind all arms
Blinding (performance bias and detection bias) study personnel All outcomes	Unclear risk	Not mentioned
Blinding (performance bias and detection bias) outcome assessors All outcomes	Unclear risk	Not possible to blind all arms
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up
Selective reporting (reporting bias)	High risk	Systematic baseline measure of wetting: No Organic causes excluded: Not mentioned Daytime wetting excluded: Not mentioned
Other bias	Unclear risk	Not mentioned

Miyazaki 1973.

Methods	Design: RCT (double‐blind cross‐over) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: No (10/38 children had daytime wetting) Setting: 6 hospital outpatient departments (1972)
Participants	No. of children (boys): 38 (27) Dropouts: 8/38 Inclusion: age > 3 years, > 1 wet night/week Exclusion: organic disease and disability, epilepsy, no toilet training, if on concurrent treatment (antidepressants, multivitamin B, neurological drugs) Other treatment (continued): lifestyle (reprimands and fluid restriction, 14 children only) Baseline wetting: every night, 15/38 children
Interventions	A (29): amitriptyline 25 mg yellow tablets for 1 week, 1 ‐ 2 tablets for 2 weeks if no response B (29): matching yellow placebo tablets Initial control period: blue placebo tablets x 1 week Duration of treatment: 3 weeks each arm Follow‐up: none
Outcomes	No. not achieving 14 dry nights: A: 13/29, B: 22/29 First arm of trial, no. not achieving 14 dry nights: A: 6/13, B: 8/13 Adverse effects: A: 4/36, B: 5/35 (drowsiness A: 4, B: 4; dry mouth A: 2, B: 2; urinary retention B: 1; deep sleep B: 1)
Notes	No wash‐out No SDs (graphical data for wet nights only) Data available from first arm of trial Study in Japanese
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not stated
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	Double‐blind cross‐over. matching yellow placebo tablets
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	Double‐blind cross‐over. matching yellow placebo tablets
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Double‐blind cross‐over. matching yellow placebo tablets
Incomplete outcome data (attrition bias) All outcomes	High risk	21% dropout
Selective reporting (reporting bias)	Low risk	Adverse events mentioned, systematic baseline measure of wetting mentioned
Other bias	Unclear risk	Not mentioned

Moltke 1979.

Methods	Design: RCT (double‐blind) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned
Participants	No. of children 88 (boys 62%) Dropouts: 47 Previous treatment: no details Age range: 4 to 13 years Average age: 8 years Severity of wetting at baseline: mean percentage of bedwetting frequency A: 77%, B: 81%, C: 79%
Interventions	A (41): furosemide (Lasix) 40 mg B (43): imipramine 25 mg C (44): placebo Doses increased if no response Duration of treatment: 4 weeks Follow‐up: 1 year
Outcomes	At end of study mean percentage wet nights A: 73%, B: 49%, C: 67% When followed up after 1 year, 59 children had increased frequency of wetting to baseline level or worse
Notes	Danish paper No SDs No details of reasons for dropouts
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not mentioned
Allocation concealment (selection bias)	Unclear risk	Not mentioned
Blinding (performance bias and detection bias) participants All outcomes	Low risk	"Double blind"
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	"Double blind"
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	"Double blind"
Incomplete outcome data (attrition bias) All outcomes	High risk	No loss to follow‐up at end of treatment. High dropout rate
Selective reporting (reporting bias)	Low risk	All outcomes reported
Other bias	Unclear risk	Not stated

Motavalli 1994.

Methods	Design: RCT Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned
Participants	No. of children (boys): 29 (A: 6 B: 4 C: 4) Inclusion: age 5 to 14; no organic causes; normal intelligence; wetting 2+ times a week; no treatment in previous 2 months Mean age A: 9.1 year B: 9.2 year C: 8.3 year Baseline wetting: mean (SD) number of wet nights in 15 days: A: 9.1 (4.1), B: 11.2 (3.8), C: 10.9 (3.3)
Interventions	A (10): imipramine ‐ dose depended on age B (9): clomipramine C (10): alarm Duration of treatment: 8 weeks Follow‐up: none
Outcomes	Mean (SD) frequency of wetting during final two weeks of treatment: A: 4.1 (2.6), B: 6.6 (5.5), C: 2.8 (4.3)
Notes	Foreign language (translation) No significant difference between groups in terms of age or IQ Not blinded Unclear if intention‐to‐treat
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not stated "randomly divided into three groups"
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	High risk	Not blinded
Blinding (performance bias and detection bias) study personnel All outcomes	High risk	Not blinded
Blinding (performance bias and detection bias) outcome assessors All outcomes	High risk	Not blinded
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Not stated, Unclear ITT
Selective reporting (reporting bias)	Unclear risk	Not stated
Other bias	Unclear risk	Not stated

Naitoh 2005.

Methods	Design: controlled trial Systematic baseline measure of wetting :Yes Organic causes excluded:Yes Daytime wetting specifically excluded:Yes Setting/recruitment: Clinic (hospital)
Participants	Number of participants (boys);105 (76) Number of dropouts: 0 Inclusion criteria: Monosymptomatic enuresis Exclusion Criteria: Daytime symptom and incontinency, frequency ≧ 8 times/day and urologic anomalities Ages range: 6 to 13 (mean 9.4 years) Baseline wetting: Alarm group 13 days/2 weeks, desmopressin 12.6/2 weeks , Imipiramin 12.3/2 weeks
Interventions	A  37 (alarm) B  35 ( alarm + desmopressin) C  33 (alarm + imipramine) Duration of treatment: 3 ‐ 6 months Duration of follow‐up:no follow‐up beyond end of treatment
Outcomes	Mean wet nights at end of trial: Not mentioned Only data mentioned: No. not achieving 14 dry nights at end of trial (6 months): Group A: 26/37, B 28/35, C26/33 Adverse events (numbers and details): not mentioned
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	"Patients without previous treatment were enrolled randomly into the three groups, and the patients who had had previous treatment were not given the same treatment again."
Allocation concealment (selection bias)	High risk	As above.
Blinding (performance bias and detection bias) participants All outcomes	High risk	Not blinded
Blinding (performance bias and detection bias) study personnel All outcomes	High risk	Not blinded
Blinding (performance bias and detection bias) outcome assessors All outcomes	High risk	Not blinded. Parents
Incomplete outcome data (attrition bias) All outcomes	Low risk	No dropouts
Selective reporting (reporting bias)	Low risk	All outcomes detailed in Methods section reported
Other bias	Unclear risk	Not stated

Netley 1984.

Methods	Design: RCT Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned Setting: Hospital for Sick Children Enuresis Clinic, Toronto, Canada
Participants	No. of children: 62 No. of dropouts: 27 Inclusion: primary nocturnal enuresis, age 6 to 12 Ages: mean A: 9 years, B: 10.7 years
Interventions	A (31): imipramine B (31): Mozes detector (buzzer + electric shock to abdominal wall on wetting) Duration of treatment unclear, ? till dry for 2 months?
Outcomes	Final outcome, after 2 dry months: A: 13/17 failed, B: 7/18 failed ITT (including dropouts as failures): A: 27/31, B: 20/31
Notes	High dropout rate (44%) Groups not comparable on age at baseline Younger children (< 8 years) apprehensive about detector Authors conclude Mozes detector is suitable for children > 8 years
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The children were randomly assigned to 1 of 2 groups
Allocation concealment (selection bias)	Unclear risk	Unclear. Not stated
Blinding (performance bias and detection bias) participants All outcomes	Unclear risk	Not stated
Blinding (performance bias and detection bias) study personnel All outcomes	Unclear risk	Not stated
Blinding (performance bias and detection bias) outcome assessors All outcomes	Unclear risk	Not stated
Incomplete outcome data (attrition bias) All outcomes	High risk	High dropout rate. 17/31 in group 1 (imipramine) and 18/31 in group 2 (device) were tested on all 3 occasions. Most of the participants who dropped out of the study had failed to keep their appointments for the final assessment. No further information given regarding those who were lost to follow‐up
Selective reporting (reporting bias)	Unclear risk	Protocol not available
Other bias	Unclear risk	Not stated

Neveus 2008#.

Methods	Design: RCT (cross‐over design) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting specifically excluded: Yes Setting/recruitment: University hospital
Participants	Number of participants (boys):27 (22 boys) Number of dropouts: 2 Inclusion criteria: severe primary monosymptomatic NE (at least 7 wet nights in 2 weeks;unresponsive to both alarm and desmopressin, previous daytime incontinency but dry for 1 year Exclusion criteria: concomitant cardiac,nephrologic, metabolic or neurologic diseases, urinary tract infection or previous treatment with anticholinergic or tricyclic drugs Ages: 6 ‐ 13 year (9.4 ± 2.1) Baseline wetting: mean wet night/week : placebo (11.0 ± 3.9) tolterodine (10.4 ± 3.9) imipramine (7.8 ± 5.1)
Interventions	Placebo (9): Tolterodine (9): 1 ‐ 2 mg (higher dose for children > 8 years) Imipramine (9): 25 ‐ 50 mg (higher dose for children > 8 years) Duration of treatment: 6 weeks (cross‐over fashion) Duration of follow up: not mentioned
Outcomes	Mean wet nights at end of trial: placebo (11.0 ± 3.9), tolterodine (10.4 ± 3.9), imipramine (7.8 ± 5.1) No. not achieving 14 dry nights at end of trial: placebo (8), tolterodine (9), imipramine (4) No. not achieving 14 dry nights at end of trial or relapsing after:
Notes	Adverse events (numbers and details): no serious adverse effect, 9 participants in imipramine group (3 slight mood change, 2 insomnia, 1 palpitation, 2 slight nausea and 1 child interrupted medication because of intense nausea, 1 child had slight mood change by tolterodine)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	All children were given all three treatments but in a randomised, double‐blinded crossover fashion.
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	Double‐blind with placebo arm
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	Double‐blind with placebo arm
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Not stated but presumably parents. "the number of wet and dry nights...was assessed during the last 2 weeks of each treatment period." Double‐blind with placebo arm
Incomplete outcome data (attrition bias) All outcomes	Low risk	1 cured. 1 dropped out due to side effects. 25/27 analysed
Selective reporting (reporting bias)	Unclear risk	Outcome was no. of wet nights during the last 2 treatment weeks. Only this outcome was reported. No protocol available
Other bias	Unclear risk	Not stated

Petersen 1974#.

Methods	Design: RCT (randomised double‐blind cross‐over) Systematic baseline measure of wetting: No Organic causes excluded: No Daytime wetting excluded: No Setting: outpatient clinic
Participants	No. of children: 69 Dropouts: 7 Inclusion: primary and secondary enuresis, some diurnal wetting, organic causes or behavioural disturbances Exclusion: none Previous treatment: almost all (waking, alarms, drugs, hospital admission) Ages: 5 to 15 years (mean for boys 8 years 6 months, girls 8 years 3 months)
Interventions	A (61): imipramine B (61): imipramine‐N‐oxide C (61): emperonium (Cetiprin) D (61): placebo (all matching preparations) Duration of treatment: 4 weeks on each drug Follow‐up: children continued on 1 successful drug for 3 months: A: 32, B: 10, C: 2, D: 4
Outcomes	Mean wet nights per week in last 3 weeks of each drug: A: 2.4, B: 3.24, C: 4.36, D: 4.37 Adverse events: A: 13, B: 7, C: 10, D: 11 (nausea, decreased appetite, sleep disturbance, rash, headaches, sweating, fatigue, sullenness, tremor)
Notes	No SDs No wash‐outs All children had fluid restriction and wakening once a night
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Drugs were given in randomised sequence. No further information
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	Drugs were administered as identical appearing capsules in bottles marked by a 6‐ciphered code number for period, patient, and drug. The code was accessible in such a form that it could be broken without yielding any knowledge of the general code or permitting any conclusion concerning the treatment of other patients. Parents were told that 4 drugs were to be tried but nothing was said about the nature of drugs or placebo.
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	As above
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Parents. As above
Incomplete outcome data (attrition bias) All outcomes	Low risk	7 dropped out: lost to follow‐up for > 2 weeks and 1 was completely dry
Selective reporting (reporting bias)	Unclear risk	Assumption of carry‐over effect led to correction of data where one week’s data were omitted for analysis was provided together with the data for the full 4‐week period. Notwithstanding, protocol not stated
Other bias	Unclear risk	Not stated

Poussaint 1965a.

Methods	Design: RCT (double‐blind, cross‐over in some cases, assigned to treatment in rotation) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned
Participants	No. of children (boys): 47 (36) Dropouts: 7 Inclusion: high‐frequency enuresis ‐ > one wet night per week Previous treatment: 2 children had psychotherapy for at least 1 year Age range: 5 to 16 years Baseline wetting: average number of wet nights per week: A:5.2, B:5.9, C:5.7, D:5.6
Interventions	A (13): imipramine (4 weeks) then placebo (4 weeks) B (13): placebo (4 weeks) then imipramine (4 weeks) C (10): imipramine (4 weeks) then imipramine (4 weeks) D (11): placebo (4 weeks) then placebo (4 weeks) Duration of treatment: 4 or 8 weeks Follow‐up: 2 months
Outcomes	In cross‐over trial, drug better than placebo in 69%, equal in 23% and placebo better than drug in 8% In cross‐over, mean number of wet nights in 4th week of treatment: A: 2.9, B: 4.8 In non‐crossover, mean number of wet nights in final week of treatment: C:2.4, D:4.2 Number of children totally dry: C:6, D:1 (no relapses) Only relapses were when medication abruptly withdrawn ‐ all had medication restored Side effects: more irritable (8); dizziness (1); dry mouth (1); decreased appetite (1) Similar complaints noted in placebo children In follow‐up, 24% of children 'cured' by imipramine
Notes	Results from first arm of trial used (as for parallel groups) Urinalysis by own physician Dubious baseline Not stated if comparable groups No wash‐out phase Not intention‐to‐treat Short follow‐up Results from graphs No SDs
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	"The coded drug and placebo were dispensed by the hospital pharmacy for the patients who were assigned to one or the other treatment in rotation."
Allocation concealment (selection bias)	High risk	Inadequate. No information
Blinding (performance bias and detection bias) participants All outcomes	Low risk	The method of assignment was unknown to the investigators until the data were analysed. The study design included a double‐blind placebo control with a cross‐over for some participants midway in the 8‐week treatment period
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	The method of assignment was unknown to the investigators until the data were analysed. The study design included a double‐blind placebo control with a cross‐over for some participants midway in the 8‐week treatment period
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Parents. The method of assignment was unknown to the investigators until the data were analysed. The study design included a double‐blind placebo control with a cross‐over for some participants midway in the 8‐week treatment period
Incomplete outcome data (attrition bias) All outcomes	Low risk	7 excluded ( 5 withdrew, could not swallow tablets, feared the drugs; 1 had bladder infection on placebo, 1 appeared to be psychotic)
Selective reporting (reporting bias)	Unclear risk	There is a rating scale outlined but there are no data presented using the rating scale Protocol not presented
Other bias	Unclear risk	Not stated

Poussaint 1966a.

Methods	Design: RCT (double‐blind cross‐over, children assigned at random by pharmacy) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned Setting: referred by physicians
Participants	No. of children: 50 Dropouts: 10 (1 UTI, 6 non‐compliance with protocol, 3 no FU) Inclusion: enuresis at least once per week Exclusion: organic causes, learning difficulty Ages: 5 to 15
Interventions	Trial 1 A (16): amitriptyline 25 mg age < 12 years, 50 mg > 12 years B (16): placebo Duration of treatment: 4 weeks Follow‐up: none
Outcomes	Trial 1 No. of wet nights in 4th week (No, mean): A: 16, 3.1, B: 16, 4.6 Side effects: (n = 40) minimal, A: irritable (7), calmer (2), nocturia (10), drowsy (3), headache (2), lower appetite (1), fatigue (1), stomach ache (1), scleral injection (1); B: irritable (5), stomach ache (1), fatigue (1), lower appetite (1)
Notes	Data from first arm of trial used (as for parallel groups) Data from graphs only No SDs
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The design included a double‐blind placebo control with cross‐over for most patients midway in the 8 week treatment period.   The coded drug and placebo were dispensed by the hospital pharmacy to the participants who were assigned to treatment groups at random
Allocation concealment (selection bias)	Low risk	Central allocation from pharmacy
Blinding (performance bias and detection bias) participants All outcomes	Low risk	"...data were tabulated and rated...before codes were  broken. In addition, there was a cross‐over design."
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	"...data were tabulated and rated...before codes were broken. In addition, there was a cross‐over design."
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Parent. "...data were tabulated and rated...before codes were broken. In addition, there was a cross‐over design."
Incomplete outcome data (attrition bias) All outcomes	High risk	Of the 60 children who started the study, 56 were treated for 8 weeks. Although they cooperated fully with the treatment program, 6 were excluded from the statistical evaluation because medication was not properly taken. Three others were excluded because they did not keep appointments, and another was omitted because of a bladder infection. Total 10 exclusions. Not clear which groups the excluded children belonged to. High dropout rate.
Selective reporting (reporting bias)	Unclear risk	Not enough information
Other bias	High risk	Amitriptyline was supplied by Merck Sharp & Dohme. Was that the extent of their involvement?

Poussaint 1966b.

Methods	Design: RCT (double‐blind children assigned at random by pharmacy) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned Setting: referred by physicians
Participants	No. of children: 50 Dropouts: 10 (1 UTI, 6 non‐compliance with protocol, 3 no FU) Inclusion: enuresis at least once per week Exclusion: organic causes, learning difficulty Ages: 5 to 15
Interventions	Trial 2 C (9): amitriptyline 25 mg age < 12 years, 50 mg > 12 years D (9): placebo Duration of treatment: 8 weeks Follow‐up: none
Outcomes	Trial 2 No. of wet nights in 8th week (No,. mean): C: 9, 4.1, D: 9, 5.5 Side effects: (n = 40) minimal, A: irritable (7), calmer (2), nocturia (10), drowsy (3), headache (2), lower appetite (1), fatigue (1), stomach ache (1), scleral injection (1); B: irritable (5), stomach ache (1), fatigue (1), lower appetite (1)
Notes	Parallel groups Data from graphs only No SDs
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	As above
Allocation concealment (selection bias)	Low risk	As above
Blinding (performance bias and detection bias) participants All outcomes	Low risk	As above
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	As above
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	As above
Incomplete outcome data (attrition bias) All outcomes	High risk	As above
Selective reporting (reporting bias)	Unclear risk	As above
Other bias	Unclear risk	As above

Roy 1970.

Methods	Design: RCT Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned Setting: Deaf and dumb boys attending a specialised boarding school
Participants	No. of children: 26 (all boys) Inclusion: age 6+; normal urinalysis; wetting bed 1+ times a week; no other medication Mean age (years) 11.4 (range 7 to 17) Baseline wetting: mean number of wet nights per week A: 4.2, B: 3.1, C: 2.5
Interventions	A (14): 25 mg imipramine B (6): corresponding placebo C (6): no treatment control Follow‐up: 2 weeks
Outcomes	Mean number of wet nights per week: A: 1.8, B: 2.0, C: 2.6 Adjusted means: A: 1.6, B: 3.2, C: 2.9 3 psychosocial factors had a significant negative correlation with percentage improvement: age of child; number of years resident in the institution; depression score
Notes	Foreign language Sketchy medical Not comparable groups (groups comparable in age but significant difference between groups in terms of baseline wetness) Not stated if intention‐to‐treat No details of previous treatment No SDs
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not stated
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	High risk	Only imipramine and placebo blinded. Control not blinded
Blinding (performance bias and detection bias) study personnel All outcomes	High risk	Only imipramine and placebo blinded. Control not blinded
Blinding (performance bias and detection bias) outcome assessors All outcomes	High risk	Only imipramine and placebo blinded. Control not blinded
Incomplete outcome data (attrition bias) All outcomes	High risk	Differential timing of outcomes reported
Selective reporting (reporting bias)	Unclear risk	Not known
Other bias	Unclear risk	Not stated

Scholander 1968.

Methods	Design: RCT (double‐blind) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned
Participants	No. of children (boys): 30 (23) No dropouts Previous treatment: all had received imipramine, amitriptyline or nortriptyline Age range: 7 to 17 years Severity of wetting at baseline: wet bed between 2 and 12 times a week
Interventions	First week: A (15): enuresis alarm but switched off B (15): enuresis alarm but switched off Next 2 weeks: A (15): enuresis alarm + nortriptyline B (15): enuresis alarm + placebo Duration of treatment: 2 weeks on tablets Follow‐up after 6 to 12 months
Outcomes	Number with no wet nights in final week of drug/placebo treatment: A: 9/15, B: 6/15 Side effects: dry mouth, troubled sleep (groups not specified)
Notes	Swedish paper No details of inclusion/exclusion criteria Groups comparable in age and frequency of wet nights
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not stated
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	Double‐blind
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	Double‐blind
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Double‐blind
Incomplete outcome data (attrition bias) All outcomes	Low risk	No dropouts
Selective reporting (reporting bias)	Unclear risk	Not stated
Other bias	Unclear risk	Not stated

Schröder 1971.

Methods	Design: RCT (double‐blind) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned
Participants	No. of children: 62 Dropouts: 34 Inclusion: age 4 to 10 years; resistant to previous therapy; no secondary symptoms; no side effects; sufficient information from GPs Previous treatment: had various different treatments Age: 3.5 to 11 years Baseline wetting: no details
Interventions	A (35): imipramine 30 mg twice a day B (27): placebo Duration of treatment 25 days
Outcomes	A: 28 improved and 7 showed no change or worsened B: 7 improved and 20 showed no change or worsened No. cured: A: 0/35, B: 1/27 Side effects only observed in children under 7 years Side effects found in both groups
Notes	German paper No results for severity No details of comparability at baseline Not intention‐to‐treat No means or SDs
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not stated
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	Double‐blind
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	Double‐blind
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Double‐blind
Incomplete outcome data (attrition bias) All outcomes	High risk	High dropout rate. No ITT
Selective reporting (reporting bias)	High risk	Mentioned no side effects
Other bias	Unclear risk	Not stated

Seo 2001.

Methods	Design: RCT Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: not mentioned Setting: University Hospital, April 1997 to July 2000.
Participants	Number of participants: 129 Number of dropouts: 0 Inclusion criteria: Monosymptomatic nocturnal enuresis Exclusion criteria: urinary tract dysfunction, neurogenic disorder and urinary reflux Ages: imipramine: Boys 5 ‐ 15 years old (mean 8.4), Girls: 5 ‐ 18 years (mean 7), total : 5 ‐ 18 years (mean 8.9) Desmopressn: Boys: 5 ‐ 13 years (mean 8.1); Girls: 5 ‐ 17 years (mean 9.0); Total 5 ‐ 17 (mean 8.3) Imipramine + desmopressin: Boys: 5 ‐ 14years (mean 8.0); Girls: 5 ‐ 14 years (mean 7. 9 years); Total 5 ‐ 14 (mean 8.0) Baseline wetting: Imipramne: 6.3 enuretic episodes/week Desmopressin: 6.2 enuretic episodes/week Imipramne + desmopressn: 6.0 episodes per week
Interventions	Imipramine + desmopressin‐ 46 participants. (Imipramine 12.5 mg and desmopressin 0.2 mg) Imipramine ‐ 43 participants. Imipramine 25 mg Desmopresin ‐ 40 participants. Desmopressin 0.2 mg Duration of treatment: Imipramine: 4 ‐ 34 weeks (mean 16.5 weeks) Desmopressin: 4 ‐ 35 weeks (mean 18.1 weeks) Imipramine + desmopressn: 3 ‐ 36 weeks (mean 17.2 weeks) Duration of follow‐up: 3 months
Outcomes	Excellent: 0 ‐ 1 enuretic episodes/week Good: > 50% reduction of enuretic episodes Poor: < 50% reduction of enuretic episodes No. not achieving 14 consecutive dry nights (derived from number not achieving excellent response. This may be an overestimation) Imipramine + desmopressn: 7/46 Desmopressn: Excellent: 16/40 Imipramine: 8/43 Wet nights after end of treatment: not mentioned No. failed or relapsed after end of treatment: Recurrent participants: unable to calculate because numbers as relapse did not match those who achieved excellent response Wet nights on treatment. No SD. Not mentioned Adverse events: loss appetite in 1 on imipramine and 6 on combined, constipation in 2 on imipramine, in 6 on desmopressin and 8 on combined, abdominal pain in 1 on combined
Notes	Korean paper
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomised to 3 groups, not mentioned
Allocation concealment (selection bias)	Unclear risk	Not mentioned
Blinding (performance bias and detection bias) participants All outcomes	Unclear risk	Not mentioned
Blinding (performance bias and detection bias) study personnel All outcomes	Unclear risk	Not mentioned
Blinding (performance bias and detection bias) outcome assessors All outcomes	Unclear risk	Not mentioned
Incomplete outcome data (attrition bias) All outcomes	Low risk	No dropouts
Selective reporting (reporting bias)	Unclear risk	Adverse events mentioned.
Other bias	Unclear risk	Insufficient information

Shaffer 1968#.

Methods	Design: RCT (double‐blind, double‐cross‐over with stratified randomisation consisting of sets of Latin squares, catering for age, sex and previous psychiatric treatment) Allocation to abrupt versus gradual withdrawal also by random plan Systematic baseline measure of wetting: Yes Organic causes excluded: No Daytime wetting excluded: No
Participants	No. of children (boys): 62 (49) 14 others excluded because of UTI or improvement before treatment 5 others stopped attending before treatment started Of the 62 that entered the trial, 3 gave no results Inclusion: any child of school age with a history of nocturnal enuresis > twice a week provisionally accepted if a) investigations revealed no disease or abnormality b) wetting 3+ times a fortnight 29 were day wetters Previous treatment: none Severity: wetting at least twice a week 30 said to have had a month of consecutive wet nights and in preceding month, only 7 said to have 14+ dry nights 40 children aged < 8 years
Interventions	A (59): high‐dose imipramine (75 mg) B (59): low‐dose imipramine (50 mg) C (59): placebo Duration of treatment: 1 month each condition then each group split in 2 ‐ half had treatment abruptly stopped, half had treatment reduced over 4 weeks
Outcomes	Of 17 participants, 15 showed appropriate rise and fall of dry nights with drug No statistics Within participants: comparing doses, no evidence that any difference between high and low (looking at patterns for 11 participants ‐ why not 18?) Between participants: significant differences found between conditions only in period 4 for placebo and low and placebo and high Side effects: restless, irritable, tearful and fidgety, difficulty in concentrating (3); sleep disturbances (6) Those showing behaviour disturbances all considered disturbed before but treatment appeared to exacerbate symptoms
Notes	No useable data Daytime wetting not excluded No significant differences between groups No wash‐out Not intention‐to‐treat analysis Urine examination was abnormal in 21 children 11 children had associated faecal soiling Some groups inexplicably not included in analysis
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were allocated by a stratified randomisation consisting of sets of Latin squares.
Allocation concealment (selection bias)	Unclear risk	(C ‐ Inadequate). The allocation to treatment group was not known by any of the clinic staff seeing the child. No information about how, so unclear how allocation was concealed from participants
Blinding (performance bias and detection bias) participants All outcomes	Low risk	Active drug and a placebo were used in a double cross‐over design with 9 treatment groups, and each strength or placebo was dispensed in uniform packs of 16 white tablets
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	Active drug and a placebo were used in a double cross‐over design with 9 treatment groups, and each strength or placebo was dispensed in uniform packs of 16 white tablets.
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Parents. Active drug and a placebo were used in a double cross‐over design with 9 treatment groups, and each strength or placebo was dispensed in uniform packs of 16 white tablets
Incomplete outcome data (attrition bias) All outcomes	High risk	When analysing within participants for high dose versus low dose, they looked at 11 participats whereas there should have been 18 ‐ 19
Selective reporting (reporting bias)	Unclear risk	Difficult to ascertain with information given. Study protocol not available. 9 treatment groups, no further information about the 9 treatment groups
Other bias	High risk	"This trial could not have been conducted without the generous help of Geigy..."

Shah 1971#.

Methods	Design: RCT (double‐blind cross‐over) Systematic baseline measure of wetting: No Organic causes excluded: Yes Daytime wetting excluded: Not mentioned Setting: Child Guidance Clinic, Bombay, India
Participants	No. of children (boys): 20 (7) Dropouts: 8 (6 recovered/improved on A) Inclusion: enuresis + behaviour problems, learning difficulties (6) Previous treatment: none for enuresis Ages: 3 to 5
Interventions	A (20): amitriptyline 25 mg age < 9 years (increased after 2 weeks if no response), 50 mg age > 9 years B (20): placebo Duration of treatment: 6 weeks each arm
Outcomes	Data not given but A stated to be better than placebo
Notes	No useable data No wash‐out Method of analysis questionable
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Double blind study...each child acted as his own control,...the distribution (of treatment) being randomised.
Allocation concealment (selection bias)	Low risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	Treatment was given orally as identical looking tablets of drug and placebo. Drug and placebo were coded and the code was kept in a sealed envelope opened at termination.
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	Treatment was given orally as identical looking tablets of drug and placebo. Drug and placebo were coded and the code was kept in a sealed envelope opened at termination.
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Parents. Treatment was given orally as identical looking tablets of drug and placebo. Drug and placebo were coded and the code was kept in a sealed envelope opened at termination.
Incomplete outcome data (attrition bias) All outcomes	High risk	8 of 20 did not take both treatments. 6 were drug alone (4 complete recovery, 2 marked recovery). 2 were placebo alone (no improvement). High dropout.
Selective reporting (reporting bias)	Unclear risk	“Open sequential method of analysis” where the statistician terminated the study as soon as a significant difference between the 2 treatments was observed. No protocol given
Other bias	High risk	The supply of amitryptiline and placebo tablets used to conduct the trial was from M/ss. Kembiotic Collaborators.

Smellie 1996.

Methods	Design: RCT (double‐blind) Randomisation carried out centrally according to standard drug procedures Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned
Participants	No. of children: 80 (boys A: 19, B: 22, C: 24) Dropouts during treatment: 0 Lost to follow‐up: 4 Previous treatment: no details Age range 5 to 13 years Baseline wetting: mean number of dry nights A: 1.6, B: 1.6, C: 1
Interventions	A (25): imipramine 25 mg B (26): mianserin 10 mg C (29): placebo Duration of treatment: 8 weeks Follow‐up: 4 weeks
Outcomes	Mean number of dry nights at week 6: A: 5, B: 2.5, C: 2.5 Mean wetness score A: 3, B: 5.6, C: 6 Number of children achieving 7 consecutive dry nights A: 21, B: 9, C: 7 After 4 weeks without treatment % showing some improvement A: 74, B: 54, C: 59
Notes	Data from graphs No SDs Not reported if comparable groups No details of previous treatment Follow‐up not intention to treat
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomisation was carried out centrally according to standard drug trial procedures and the key to randomisation was held by the local hospital pharmacist, in case of emergency.
Allocation concealment (selection bias)	Low risk	Sounds like centrally allocated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	"...issued with a 30 days' supply of tablets identical in appearance and packaging containing either imipramine 25mg, mianserin 10mg, or placebo..."
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	"...issued with a 30 days' supply of tablets identical in appearance and packaging containing either imipramine 25mg, mianserin 10mg, or placebo..."
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	? parents."...issued with a 30 days' supply of tablets identical in appearance and packaging containing either imipramine 25mg, mianserin 10mg, or placebo..."
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	All 80 children completed the preliminary four weeks' assessment and eight weeks' treatment, but no record was available for the final four weeks' observation without treatment in four children (two allocated to placebo, two to imipramine)
Selective reporting (reporting bias)	Unclear risk	Protocol not available
Other bias	High risk	Tablets were prepared and supplied by Organon and the study was supervised and co‐ordinated by Mr Peter Wright. Did Peter Wright work for Organon?

Tahmaz 2000.

Methods	Design: RCT (divided at random into 4 groups) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Yes Setting: Dept Urology, Military Medical Faculty, Turkey
Participants	No. of children (boys): 77 (48) Dropouts: A: 2, B: 7, C: 2, D: 9 (not explained) Inclusion: primary nocturnal enuresis, ≥ 3 wet nights/week, no current treatment Exclusion: organic causes, daytime wetting, UTI Ages: 6 to 14 years (9.44 ± 2.17)
Interventions	A (14): imipramine 0.9 ‐ 1.5 mg/kg/day B (16): oxybutynin 5 mg 3 x /day C (24): imipramine + oxybutynin D (23): placebo (not described) If cured, treatment tapered off gradually Duration of treatment: 3 months Follow‐up: 6 months
Outcomes	Cure defined as > 90% reduction in wet nights Failed: A: 7/14, B: 10/16, C: 8/24, D: 18/23 Failed or relapsed: A: 10/12, B: 8/9, C: 10/22, D: 11/14 Side effects: A: 3/14 (dry mouth, nausea), B: 4/16 (dry mouth, nausea), C: 7/24, D: 4/23
Notes	Groups comparable at baseline on age, sex and disease severity but groups of unequal sizes No details of previous treatment No details of nature of placebo treatment Cure not defined as complete dryness
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Patients were divided "at random" into 4 groups.
Allocation concealment (selection bias)	Unclear risk	Unclear. Not stated
Blinding (performance bias and detection bias) participants All outcomes	Unclear risk	Patients were treated for 3 months. If the patients had completely dry nights, the medication was tapered, usually first by decreasing oxybutynin, then imipramine. This suggests that the outcome assessors and/or personnel knew the identity of the drugs and which group participants belonged to
Blinding (performance bias and detection bias) study personnel All outcomes	High risk	Patients were treated for 3 months. If the patients had completely dry nights, the medication was tapered, usually first by decreasing oxybutynin, then imipramine. This suggests that the outcome assessors and/or personnel knew the identity of the drugs and which group participants belonged to
Blinding (performance bias and detection bias) outcome assessors All outcomes	High risk	Parents. Patients were treated for 3 months. If the patients had completely dry nights, the medication was tapered, usually first by decreasing oxybutynin, then imipramine. This suggests that the outcome assessors and/or personnel knew the identity of the drugs and which group participants belonged to
Incomplete outcome data (attrition bias) All outcomes	High risk	20/77 lost to follow‐up at  6 months but not addressed in the paper
Selective reporting (reporting bias)	Low risk	All outcomes reported of improvement, cure vs no response
Other bias	Low risk	Not stated

Thomsen 1967#.

Methods	Design: RCT (double‐blind, randomised cross‐over then both groups taking active drug after control period) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned Setting: Residential homes for dependent, delinquent and neglected boys
Participants	Number of children: initially 30 boys (11 dropouts) Inclusion: Wetting beds regularly (at least twice a week) Mean age: 12 years (range 7 to 16) Previous treatment: reduction of fluid intake and getting up during the night
Interventions	A: imipramine (25 mg for under 12 years old; 50 mg for over 12 years) B: placebo Duration of treatment: 4 weeks in each condition Follow‐up: 4 weeks
Outcomes	Results of only one treatment regimen given but numbers look as though both groups combined Side effects: none
Notes	No useable data Outcomes not clearly defined Comparability of groups not reported No wash‐out Not intention‐to‐treat Short follow‐up Institutional setting Placebo not used in analysis All boys
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The subjects were randomly assigned (using the double‐blind technique) to either a placebo or drug group.
Allocation concealment (selection bias)	Unclear risk	The medication was supplied to the adult in charge and there was no further contact between the boys and the physicians.
Blinding (performance bias and detection bias) participants All outcomes	Low risk	double‐blind with placebo arm
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	double‐blind with placebo arm
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	(adults in charge at the residential homes). double‐ blind with placebo arm
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	30 started. 19 completed. Attrition due to return to their family home, sent to other institutions, etc. No clear information about the characteristics of > 33% of their samples
Selective reporting (reporting bias)	Unclear risk	Results tabulated do not break down between the placebo group and the treatment groups. Not enough information
Other bias	High risk	Geigy....supplied imipramine and placebo and in part supported this study.

Treffert 1964#.

Methods	Design: RCT (double‐blind cross‐over) Systematic baseline measure of wetting: Yes Organic causes excluded: No Daytime wetting excluded: Not mentioned Setting: Children in Winebago State Hospital (hospital for neurotic, psychotic and brain‐injured boys)
Participants	No. of children (boys): 9 (9) Inclusion: nocturnal enuresis Baseline wetting: mean 2.9 times per week
Interventions	A (9): imipramine 25 mg age < 12 years, 50 mg > 12 years B (9): placebo Duration of treatment: 4 weeks each arm Follow‐up: 4 weeks
Outcomes	No. of wet nights during trial: A: 1.86, B: 2.36 Enuresis returned to baseline levels after end of trial
Notes	No SDs No wash‐out
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	This present study is a double‐blind procedure... "Each of... (the)...9 patients was then given, in random order, a 28‐day trial on a placebo and a 28‐day trial on imipramine."
Allocation concealment (selection bias)	Unclear risk	No information stated
Blinding (performance bias and detection bias) participants All outcomes	Low risk	Double‐blind with placebo arm
Blinding (performance bias and detection bias) study personnel All outcomes	Low risk	Double‐blind with placebo arm
Blinding (performance bias and detection bias) outcome assessors All outcomes	Low risk	Hospital staff (participants were hospitalised). Double‐blind with placebo arm
Incomplete outcome data (attrition bias) All outcomes	Low risk	No dropouts or attrition
Selective reporting (reporting bias)	Low risk	Bed wetting in 28 day periods with/without intervention
Other bias	High risk	Imipramine and placebo were supplied by Geigy Pharmaceuticals

Vertucci 1997.

Methods	Design: RCT (double‐blind cross‐over) Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Not mentioned Setting: Child Neuropsychiatry Clinics, Italy
Participants	No. of children: 57 Dropouts 5 Inclusion: primary nocturnal enuresis, age > 5 years, baseline wetting at least 3 nights/week Previous treatment: not mentioned Age range: 6 to 15
Interventions	A (29): desmopressin 30 mcg intranasal then imipramine B (28): Imipramine 0.9 mg/kg then desmopressin Duration of treatment: 3 weeks each Follow‐up: 2 weeks
Outcomes	Mean wet nights during first arm of trial: A: 1; B: 2.5 No. not achieving 14 dry nights: A: 4/29; B: 9/28 Mean wet nights after both drugs: A: 3.5; B: 2.8 Side effects: desmopressin: 1 back pain, 1 inflamed nasal mucosa; imipramine: 1 pallor, restlessness and cold extremities
Notes	Data estimated from graphs Data only given from first arm of trial here No wash‐out SDs not available Blinding not possible due to different routes of administration
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Patients were randomised into 2 groups A desmopressin then impramine; B imipramine then desmopressin
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding (performance bias and detection bias) participants All outcomes	High risk	Just says one group took one drug then the other. If there was no attempt at blinding then it would have been obvious since intranasal versus oral
Blinding (performance bias and detection bias) study personnel All outcomes	High risk	Just says one group took one drug then the other. If there was no attempt at blinding then it would have been obvious since intranasal versus oral
Blinding (performance bias and detection bias) outcome assessors All outcomes	High risk	Participants (parents) were assessors. Just says one group took one drug then the other. If there was no attempt at blinding then it would have been obvious since intranasal versus oral
Incomplete outcome data (attrition bias) All outcomes	Low risk	5 participants (out of 62) dropped out (8.18%): 2 did not present for testing, 2 did not fulfil entry criteria, 1 stopped desmopressin treatment because of pain in the lumbosacral region
Selective reporting (reporting bias)	Low risk	Outcomes outlined in Methods were reported
Other bias	Low risk	Not stated

Wagner 1982.

Methods	Design: RCT Systematic baseline measure of wetting: Yes Organic causes excluded: Yes Daytime wetting excluded: Yes Setting: referrals from paediatric clinics, doctors, schools and newspaper advertisements
Participants	No. of children (boys): 49 (40) Dropouts: 13 Inclusion: age 6 to 16 years, IQ > 70, primary nocturnal enuresis Exclusion: daytime wetting, physical or neurological disorders, treatment with drugs or alarms in previous year Ages: 6 to 16 Baseline wetting: min 3 times per week, A: 75%, B: 77%, C: 64%
Interventions	A (12): alarm (pad‐and‐bell/buzzer) B (12): imipramine (if < 32 kg, 25 mg/day, if > 32 kg, 50 mg/day) C (12): waiting list Duration of treatment: 14 weeks or until dry for 14 nights Follow‐up: max 44 days
Outcomes	% wet nights in 14th week: A: 8.25%, B: 39.25%, C: 60.83% (A significantly better than B or C) No. not achieving 14 dry nights: A: 2/12, B: 8/12, C: 11/12 No. not achieving 14 dry nights or relapsing: A: 7/12, B: 12/12, C: 12/12 Time from cure to relapse: A: 37.8 d (range 25 to 44), B: 17 d (3 to 27)
Notes	Groups comparable on baseline wetting No SDs or means
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Subjects were then randomly assigned to one of three conditions, conditioning treatment, pharmacotherapy, or a clinical waiting list.
Allocation concealment (selection bias)	Unclear risk	Unclear. Not stated
Blinding (performance bias and detection bias) participants All outcomes	High risk	Not blinded
Blinding (performance bias and detection bias) study personnel All outcomes	High risk	Not blinded
Blinding (performance bias and detection bias) outcome assessors All outcomes	High risk	(Conditioning – study personnel, other groups – parents). Not blinded
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	36 completed the treatment programmes, 12 in each group. 13 dropped out but no information given about them
Selective reporting (reporting bias)	Unclear risk	Wetting frequency data were obtained. Study protocol not available
Other bias	Unclear risk	Not stated

Ye 2001.

Methods	Design: RCT Systematic baseline measure of wetting: Yes (outcome measured as a reduction in wet nights, therefore baseline wetting must have been measured in both groups) Organic causes excluded: Yes Daytime wetting excluded: Not mentioned Setting: Single centre in Korea, University Hospital
Participants	Number of participants (boys): 83 (60) Number of dropouts: Inclusion criteria: monosymptomatic ‐ primary and secondary Exclusion criteria: organic causes Ages: A: 9.3 (range 7 ‐ 17), B: 9.6 ( range 5 ‐ 17) Baseline wetting: A: 6.1 nights per week, B: 6.4 nights per week
Interventions	A:(44) imipramine B:(39) desmopressin Duration of treatment: not mentioned Duration of follow‐up: 3 months
Outcomes	Number not achieving 14 consecutive dry nights: A: 9/44; B: 16/39 (P value given) Number of wet nights per week end of treatment (no SD): A: 3.4 nights per week, B: 2.3 nights per week Adverse events: "no serious adverse events were observed in both groups"
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"randomised", insufficient information about sequence generatio to permit judgement
Allocation concealment (selection bias)	Unclear risk	Not mentioned, therefore insufficient informatio to permit judgement
Blinding (performance bias and detection bias) participants All outcomes	High risk	No blinding, and the outcome likely to be influenced by lack of blinding
Blinding (performance bias and detection bias) study personnel All outcomes	High risk	No blinding, and the outcome likely to be influenced by lack of blinding
Blinding (performance bias and detection bias) outcome assessors All outcomes	High risk	No blinding, and the outcome likely to be influenced by lack of blinding
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Insufficient reporting of attrition/exclusions to permit judgement
Selective reporting (reporting bias)	Unclear risk	Insufficient informatio to permit judgement
Other bias	Unclear risk	Insufficient informaton to assess whether an important risk of bias exists

Yurdakok 1986.

Methods	Design: RCT (groups randomly assigned) Systematic baseline measure of wetting: No Organic causes excluded: Yes Daytime wetting excluded: Not mentioned Setting: Paediatric outpatient department, Ankara, Turkey
Participants	No. of children (boys): 41 (25) Inclusion: age at least 6 years, suitable for drug treatment, nocturnal enuresis Exclusion: urine abnormality Previous treatment: most had failed with simple methods such as waking or fluid restriction Ages: 6 to 16 years (mean 11 years)
Interventions	A (14) imipramine B (8) amitriptyline C (10) chlordiazepoxide‐clidinium D (9) piracetam
Outcomes	Imipramine significantly better than amitriptyline, chlordiazepoxide clinidium and piracetam on therapeutic index score: A: 1.57, B: 1.25, C: 1.3, D: 0.33
Notes	No useable data Groups comparable at baseline on age and social class
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	“Groups were chosen at random but...were well matched for age and social class.” “...children were randomly assigned to treatment with...”
Allocation concealment (selection bias)	Unclear risk	Unclear. Not stated
Blinding (performance bias and detection bias) participants All outcomes	High risk	Not blinded
Blinding (performance bias and detection bias) study personnel All outcomes	High risk	Not blinded
Blinding (performance bias and detection bias) outcome assessors All outcomes	High risk	Not blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants accounted for that were included in the study
Selective reporting (reporting bias)	Unclear risk	Outcomes discussed for both treatment arms. No study protocol. Unclear constituent of therapeutic index score
Other bias	Unclear risk	Not stated

Yurdakok 1987.

Methods	Design: RCT (grouped at random) Systematic baseline measure of wetting: No Organic causes excluded: Yes Daytime wetting excluded: Not mentioned Setting: Paediatric Dept, Ulus Social Security Hospital
Participants	No. of children (boys): 37 (25) Inclusion: age at least 6 years Exclusion: urine abnormalities Previous treatment: wakening, fluid restriction Ages: 6 to 17 years (mean 10 years)
Interventions	A (16): imipramine 25 mg before bedtime B (21): viloxazine 50 mg before bedtime (doses doubled if age over 10)
Outcomes	No difference between A and B on therapeutic index score: A: 1.55, B: 1.29
Notes	No useable data Groups comparable at baseline on age and social class
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	“children were randomly assigned treatment”
Allocation concealment (selection bias)	Unclear risk	Unclear. No information
Blinding (performance bias and detection bias) participants All outcomes	Unclear risk	No information
Blinding (performance bias and detection bias) study personnel All outcomes	Unclear risk	No information
Blinding (performance bias and detection bias) outcome assessors All outcomes	Unclear risk	No information. Unknown who assessors were
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants accounted for that were included in the study. Those excluded had abnormality in the urine cultures or did not come in at all
Selective reporting (reporting bias)	Unclear risk	Outcomes discussed for both treatment arms. No study protocol. Unclear constituent of therapeutic index score
Other bias	Unclear risk	Not stated

cc = cubic centilitres;
 EEG = electroencephalogram;
 ITT: intention‐to‐treat;
 kg = kilogram;
 MAO: monoamine oxidase
 mg = milligram;
 No. = number;
 PNE: primary nocturnal enuresis:
 SD = standard deviation;
 UTI = urinary tract infection

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Abrams 1963	RCT: No Comparison group: Yes Organic causes excluded: Yes Systematic baseline measurement of wetting: No Systematic outcome measures: Yes Interventions: Imipramine
Al‐Waili 2000	RCT: Yes but excluded as all children had detrusor instability Comparison group: Yes Organic causes excluded: Yes Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes Intervention: carbamazepine vs placebo
Alessi 1992	RCT: Yes Excluded as population adult Army recruits (age 21 ‐ 39) Intervention: Imipramine
Alison 1973	RCT: No Comparison group: Yes Organic causes excluded: No Systematic baseline measurement of wetting: No Systematic outcome measures: Yes Interventions: Amitriptyline (perphenazine)
Arai 1971	RCT: No Comparison group: Yes Organic causes excluded: Yes Systematic baseline measurement of wetting: No Systematic outcome measures: Yes Interventions: Imipramine
Blackman 1964	RCT: Yes Excluded as population adult Army recruits (age 17 ‐ 25) Intervention: Imipramine
Butler 2001	RCT: No Comparison group: No Intervention: Withdrawal from desmopressin or imipramine treatment, use of alarms optional
Cupalova‐Naglova 1970	not a RCT, academic dissertation insufficient data
D'Hollander 1967	RCT: No (cross‐over design but randomisation not mentioned) Comparison group: Yes Organic causes excluded: No Systematic baseline measurement of wetting: No Systematic outcome measures: Yes Interventions: Imipramine and amitriptyline
De Jonge 1972	RCT: No (parallel‐ group design but randomisation not mentioned) Comparison group: No Organic causes excluded: Yes Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes Interventions: Imipramine
Dorison 1962	RCT: Yes, but in adults Comparison group: Yes Organic causes excluded: No Systematic baseline measurement of wetting: No Systematic outcome measures: Yes Interventions: Imipramine
Esperanca 1969	RCT: No (cross‐over design but randomisation not mentioned) Comparison group: Yes Organic causes excluded: Yes Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes Interventions: Imipramine + diet
Fisher 1963	RCT: No (cross‐over design but randomisation not mentioned) Comparison group: Yes Organic causes excluded: Yes Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes Interventions: Imipramine
Fritz 1994	RCT: No Comparison group: No Organic causes excluded: Yes Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes Interventions: Imipramine
Heising 1978	RCT: No (parents unaware of treatment but also unaware that placebo might be used. No mention of how children were allocated to groups) Intervention: Clomipramine vs placebo
Hicks 1964	RCT: Yes, but in adults Comparison group: Yes Organic causes excluded: No Systematic baseline measurement of wetting: No Systematic outcome measures: Yes Interventions: Imipramine
Höfler 1978	RCT: No Comparison group: No Organic causes excluded: Yes Systematic baseline measurement of wetting: Yes Systematic outcome measures: No Interventions: Amitriptyline
Jorgensen 1980	RCT: No Comparison group: Yes Organic causes excluded: Systematic baseline measurement of wetting: No Systematic outcome measures: Yes Interventions: Imipramine
Kales 1977	RCT: No Comparison group: No Organic causes excluded: Yes Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes Interventions: Imipramine
Kardash 1968	RCT: No (randomisation not mentioned) Comparison group: Yes Organic causes excluded: Yes Systematic baseline measurement of wetting: No Systematic outcome measures: Yes Interventions: Imipramine
Kelly 1974	RCT: Yes (double‐ blind cross‐over) Organic causes excluded: No Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes Interventions: Lentizol (sustained releasee amitriptyline) Excluded as boys were young adults (mean age 14.36 y, range 6 ‐ 31) from a Male Residential Special School for special needs
Kim 2001	RCT: Yes Comparison group: Yes Organic causes excluded: No Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes (but improvement rates only) Interventions: Imipramine, desmopressin 2 children switched from imipramine to desmopressin due to side effects but data not given according to allocated groups therefore not useable
Korczyn 1979	RCT: No (cross‐over design but randomisation not mentioned) Comparison group: Yes Organic causes excluded: No Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes Interventions: Imipramine
Kurokawa 1963	RCT: No Comparison group: No Organic causes excluded: Yes Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes Interventions: Amitriptyline
Lake 1979	The outcomes measured in this study (BP and plasma norepinephrine levels) are not related to nocturnal enuresis. This is part of larger study (reference 16, but not provided with full references)
Laybourne 1968	RCT: No (cross‐over design but randomisation not mentioned) Comparison group: Yes Organic causes excluded: Yes Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes Interventions: Imipramine
Libert 1991	RCT: No Comparison group: No Organic causes excluded: Yes Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes Interventions: Viloxazine
Manglick 1992	RCT: No (cross‐over design but randomisation not mentioned) Comparison group: Yes Organic causes excluded: No Systematic baseline measurement of wetting: No Systematic outcome measures: Yes Interventions: Imipramine
Mariuz 1963	RCT: No (cross‐over design but randomisation not mentioned) Comparison group: Yes Organic causes excluded: Yes Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes Interventions: Imipramine
McConaghy 1969	RCT: Yes Comparison group: Yes Organic causes excluded: No Systematic baseline measurement of wetting: No Systematic outcome measures: Yes Interventions: Imipramine, amphetamine, pad + bell and random wakening Excluded because children were moved between trial arms, data therefore unreliable
Meadow 1982	RCT: Yes Diurnal enuresis Interventions: Imipramine
Meijer 1965	RCT: No Comparison group: No Organic causes excluded: Yes Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes Interventions: Imipramine
Miller 1968	RCT: No (cross‐over design but randomisation not mentioned) Comparison group: Yes Organic causes excluded: Yes Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes Interventions: Imipramine
Milner 1968	RCT: Yes Interventions: Desipramine, imipramine, nortriptyline Setting: Adult mental hospital Excluded as participants were adults
Mishra 1980	RCT: No (parallel‐ group design but randomisation not mentioned) Comparison group: Yes Organic causes excluded: Yes Systematic baseline measurement of wetting: No Systematic outcome measures: No Interventions: Amitriptyline
Monda 1995	RCT: No Comparison group: Yes Organic causes excluded: Yes Systematic baseline measurement of wetting: No Systematic outcome measures: Yes Interventions: Imipramine, observation, desmopressin and alarms
Nigam 1973	RCT: No Interventions: Imipramine, supportive psychotherapy and placebo Setting: Outpatients, District Hospital, Hamirpur, India
Noack 1964	RCT: No Comparison group: Yes Organic causes excluded: No Systematic baseline measurement of wetting: No Systematic outcome measures: Yes Interventions: Imipramine
Philpott 1970	RCT: No Comparison group: No Organic causes excluded: Yes Systematic baseline measurement of wetting: No Systematic outcome measures: Yes Interventions: Imipramine, alarms
Polak 1981	RCT: No (parallel‐ group design but randomisation not mentioned) Comparison group: Yes Organic causes excluded: No Systematic baseline measurement of wetting: No Systematic outcome measures: No Interventions: Syndocarb, imipramine, methylphenidate, psychotherapy
Porot 1970	RCT: No Comparison group: Yes Organic causes excluded: No Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes Interventions: Amitriptyline
Poussaint 1965b	RCT: No Comparison group: no Systematic baseline: yes Systematic outcome measure: yes Organic causes excluded: yes Intervention: Protriptyline
Rapoport 1980	RCT : No 2 studies, using cross‐over design but not randomised Comparison group: Yes Organic causes excluded: Yes Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes Intervention: Tricyclics (imipramine, desmethylimipramine); methscopolamine
Rett 1968	RCT: No (triple‐ blind to patients nurses and researchers, but no mention of how children were allocated to groups, and the lead researcher was aware of allocation) Intervention: Trimethoprim vs placebo
Russell 1992	RCT: No (before‐ and‐ after study strengthened by Latin square design) Intervention: clinical standard setting and audit aimed at general practitioner trainers. Tricyclics for enuresis used as one of four exemplar conditions
Saeedalzakerin 2000	Main study outcome is parents' behaviour. Not a RCT. Abstract only.
Schjetne 1970	RCT: No Comparison group: Yes Organic causes excluded: Yes Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes Interventions: Imipramine (Tofrinil)
Simeon 1981	RCT: No Comparison group: No Organic causes excluded: No Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes Interventions: Maprotiline (Ludiomil)
Site 1974	RCT: No Comparison group: Yes Organic causes excluded: Yes Systematic baseline measurement of wetting: No Systematic outcome measures: Yes Interventions: Imipramine, antispasmodic, sedative
Smith 1967	RCT: No (single‐ blind cross‐over with respect to carers and participants but allocation stated not to be concealed from the psychiatrist performing randomisation) Intervention: Nortriptyline, placebo Setting: residential institute for 'mentally retarded' youths
Soulayrol 1970	RCT: No Comparison group: Yes Organic causes excluded: No Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes Interventions: Afranil, Dolibrax
Steinicke 1971	RCT: No (randomisation not mentioned) Diurnal and nocturnal enuresis Intervention: Imipramine, placebo
Tamburello 1971	RCT: Unclear (double blind administration, but unclear when placebo was administered) Intervention: Butriptyline, placebo
Tong 2001	RCT: Yes Intervention: Amitriptyline, placebo Excluded because participants were adult psychiatric patients treated with clozapine suffering from clozapine‐induced enuresis
Valentine 1968	RCT: Yes (double‐ blind cross‐over, order determined by random number tables) Comparison group: Yes (cross‐over) Organic causes excluded: No Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes Interventions: Imipramine, placebo Setting: Residential 'mentally handicapped' children in hospital, Leicester, UK Excluded as all children had diurnal wetting
Werry 1975	RCT: No Comparison group: Yes Organic causes excluded: No Systematic baseline measurement of wetting: Yes Systematic outcome measures: Yes Interventions: Imipramine
Werry 1977	RCT: No Comparison group: Yes Organic causes excluded: No Systematic baseline measurement of wetting: No Systematic outcome measures: Yes Interventions: Imipramine and chlordiazepoxide
Yamanishi 1988	RCT: No Comparison group: Yes Organic causes excluded: No Systematic baseline measurement of wetting: Yes Systematic outcome measures: No Interventions: Imipramine and bladder training

Characteristics of studies awaiting assessment [ordered by study ID]

Cheng 2008.

Methods
Participants
Interventions	Vitamin B injection on acupoints vs imipramine
Outcomes
Notes	Still trying to obtain the full text of this article. More appropriate in acupuncture‐related reviews

Chertin 2007.

Methods	RCT
Participants	Number of participants (boys) 54 (34) Inclusion criteria: participants with ADHD and NE
Interventions	A: (27) Desmopressin and oxybutynin B: (27) Imipramine
Outcomes	Only data available: Incidence of NE (DVSS score q2) 0.9 ± 0.2 vs 2.9 ± 0.2 P < 0.0001
Notes	Authors contacted by email; no response Systematic baseline measure of wetting: yes for DVSS score but not broken down to DUI and NE Organic causes excluded?: Yes Daytime wetting specifically excluded?: No Setting/recruitment: referred participants with ADHD and NE (? Clinic) Number of dropouts – not given

Cupalova‐Naglova 1968.

Methods
Participants
Interventions	Imipramine, 'chloridzaepoxide' (chlordiazepoxide?) and placebo
Outcomes
Notes	The paper was written In Czech. We are awaiting translation

GP Research Grp 1969.

Methods	RCT
Participants
Interventions	Imipramine
Outcomes
Notes	Comparison group: Yes Organic causes excluded: No Systematic baseline measurement of wetting: No Systematic outcome measures: Yes Randomisation was for length of treatment (6, 8 or 10 weeks) and for stopping treatment abruptly at the end of that time versus tailing off gradually, therefore data not useable in review outcomes. Need to contact authors.

Hoseinzadeh 1997 NEW.

Methods	Unclear
Participants	44 participants
Interventions	Imipramine versus alarms
Outcomes
Notes	Even after translation study design is unclear ‐ need to contact authors for further clarification.

Houts 2003.

Methods	RCT
Participants	Enuresis
Interventions	Full‐spectrum home training which has 4 components: alarm; cleanliness training; retention control training; and overlearning compared to imipramine and oxybutynin
Outcomes
Notes	Book chapter by Houts from 2003 which mentions (on page 398) an unpublished trial from 1991, probably by Houts and colleagues, Need to contact Houts to find out more

Ice 1966.

Methods	RCT: Yes (double‐blind placebo‐controlled cross‐over)
Participants	19 participants
Interventions	Imipramine vs placebo for 6 weeks each
Outcomes
Notes	Setting: institution for dependent and delinquent youths Reported in abstract only, no data provided (enuresis stated to have improved on imipramine without changes in anxiety or subjective motivation). Need to contact authors

Differences between protocol and review

For this update, we added a new category of outcome which assessed combination therapies involving tricyclics, and moved all combination therapies previously in the other outcomes to this new outcome category. We also assessed for selective reporting bias in studies by examining whether systematic baseline measurements of enuresis were conducted prior to the start of treatment, whether adverse events were mentioned, whether organic causes of enuresis or daytime urinary incontinence were excluded, whether full data sets were provided, where study protocols were provided, and whether all prespecified outcomes were reported.

Contributions of authors

Two of the review authors (from PS, PC, and WW) independently assessed each study identified by the updated searches and extracted data from the included studies. PS and PC also reviewed the included studies from the previous review and updated the 'Risk of bias' tables. PS and PC analysed the data and wrote the updated review, and PC also provided a clinical perspective and interpretation.

Sources of support

Internal sources

• Chief Scientist Office, Scottish Executive Health Department, UK.

External sources

• National Institute for Health Research (NIHR), UK.

  This project was supported by the National Institute for Health Research, via Cochrane Infrastructure, Cochrane Programme Grant or Cochrane Incentive funding to the Incontinence Group. The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the Systematic Reviews Programme, NIHR, NHS or the Department of Health.

Declarations of interest

Patrina HY Caldwell has a non‐personal indirect financial conflict of interest. She invented an enuresis alarm in year 2009. The Children's Hospital at Westmead and the University of Sydney jointly own the patent and rights, and have sole control and benefits from any income if it were to successfully commercialise. PC will not benefit financially. Commercialisation is not assured.

Premala Sureshkumar ‐ none known

Wicky CF Wong ‐ none known

New search for studies and content updated (conclusions changed)