Pelvic floor muscle training added to another active treatment versus the same active treatment alone for urinary incontinence in women

  • Review
  • Intervention

Authors


Abstract

Background

Pelvic floor muscle training (PFMT) is a first-line conservative treatment for urinary incontinence in women. Other active treatments include: physical therapies (e.g. vaginal cones); behavioural therapies (e.g. bladder training); electrical or magnetic stimulation; mechanical devices (e.g. continence pessaries); drug therapies (e.g. anticholinergics (solifenacin, oxybutynin, etc.) and duloxetine); and surgical interventions including sling procedures and colposuspension. This systematic review evaluated the effects of adding PFMT to any other active treatment for urinary incontinence in women

Objectives

To compare the effects of pelvic floor muscle training combined with another active treatment versus the same active treatment alone in the management of women with urinary incontinence.

Search methods

We searched the Cochrane Incontinence Group Specialised Register, which contains trials identified from the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, MEDLINE in process, and handsearching of journals and conference proceedings (searched 28 February 2013), EMBASE (January 1947 to 2013 Week 9), CINAHL (January 1982 to 5 March 2013), ClinicalTrials.gov (searched 30 May 2013), WHO ICTRP (searched 3 June 2013) and the reference lists of relevant articles.

Selection criteria

We included randomised or quasi-randomised trials with two or more arms in women with clinical or urodynamic evidence of stress urinary incontinence, urgency urinary incontinence or mixed urinary incontinence. One arm of the trial included PFMT added to another active treatment; the other arm included the same active treatment alone.

Data collection and analysis

Two review authors independently assessed trials for eligibility and methodological quality and resolved any disagreement by discussion or consultation with a third party. We extracted and processed data in accordance with the Cochrane Handbook for Systematic Reviews of Interventions. Other potential sources of bias we incorporated into the 'Risk of bias' tables were ethical approval, conflict of interest and funding source.

Main results

Eleven trials met the eligibility criteria for inclusion, comprising women with stress urinary incontinence (SUI), urgency urinary incontinence (UUI) or mixed urinary incontinence (MUI), and they compared PFMT added to another active treatment (494 women) with the same active treatment alone (490 women). The pre-specified comparisons were reported by single trials except electrical stimulation which was reported by two trials. However, the two trials reporting electrical stimulation could not be pooled as one of the trials did not report any relevant data. We considered the included trials to be at unclear risk of bias for most of the domains, predominantly due to the lack of adequate information in a number of trials. This affected our rating of the quality of evidence. 

The majority of the trials did not report the primary outcomes specified in the review (cure/improvement, quality of life) or measured the outcomes in different ways. Effect estimates from small, single trials across a number of comparisons were indeterminate for key outcomes relating to symptoms and we rated the quality of evidence, using the GRADE approach, as either low or very low. There was moderate-quality evidence from a single trial investigating women with SUI, UUI or MUI that a higher proportion of women who received a combination of PFMT and heat and steam generating sheet reported cure compared to those who received the sheet alone: 19/37 (51%) versus 8/37 (22%) with a risk ratio (RR) of 2.38, 95% confidence interval (CI) 1.19 to 4.73). More women reported cure or improvement of incontinence in another trial comparing PFMT added to vaginal cones to vaginal cones alone: 14/15 (93%) versus 14/19 (75%), but this was not statistically significant (RR 1.27, 95% CI 0.94 to 1.71). We judged the quality of the evidence to be very low. Only one trial evaluating PFMT when added to drug therapy provided information about adverse events (RR 0.84, 95% CI 0.45 to 1.60; very low-quality evidence).

With regard to condition-specific quality of life, there were no statistically significant differences between women (with SUI, UUI or MUI) who received PFMT added to bladder training and those who received bladder training alone at three months after treatment either on the Incontinence Impact Questionnaire-Revised scale (mean difference (MD) -5.90, 95% CI -35.53 to 23.73) or on the Urogenital Distress Inventory scale (MD -18.90, 95% CI -37.92 to 0.12). A similar pattern of results was observed between women with SUI who received PFMT plus either a continence pessary or duloxetine and those who received the continence pessary or duloxetine alone. In all these comparisons, the quality of the evidence for the reported critical outcomes ranged from moderate to very low.

Authors' conclusions

This systematic review found insufficient evidence to state whether or not there were additional effects of adding PFMT to other active treatment when compared with the same active treatment alone for urinary incontinence (SUI, UUI or MUI) in women. These results should be interpreted with caution as most of the comparisons were investigated in small, single trials. None of the trials in this review were large enough to provide reliable evidence. Also, none of the included trials reported data on adverse events associated with the PFMT regimen, thereby making it very difficult to evaluate the safety of PFMT.

Résumé scientifique

Entraînement des muscles du plancher pelvien ajouté à un autre traitement actif par rapport au même traitement actif seul pour l'incontinence urinaire chez la femme

Contexte

L’entraînement des muscles du plancher pelvien (EMPP) est un traitement conservateur de première ligne dans l'incontinence urinaire chez la femme. D'autres traitements actifs inclus : les thérapies physiques (telles que les cônes vaginaux), les thérapies comportementales (telles que la rééducation de la vessie), les stimulations électriques ou magnétiques, les dispositifs mécaniques (tels que les pessaires de continence), les traitements médicamenteux (tels que les anticholinergiques (la solifénacine, l'oxybutynine, etc.) et la duloxétine) et les interventions chirurgicales, y compris les procédures de fronde et la colposuspension. Cette revue systématique a évalué les effets de l'ajout de l’EMPP à tout autre traitement actif pour l'incontinence urinaire chez la femme

Objectifs

Comparer les effets de l'entraînement des muscles du plancher pelvien combinée à un autre traitement actif par rapport au même traitement actif seul dans la prise en charge des femmes souffrant d'incontinence urinaire.

Stratégie de recherche documentaire

Nous avons effectué des recherches dans le registre spécialisé du groupe Cochrane sur l'incontinence, qui contient des essais identifiés dans le registre Cochrane des essais contrôlés (CENTRAL), MEDLINE, processus en cours de MEDLINE et des recherches manuelles dans des journaux et actes de conférence (recherche du 28 février 2013), EMBASE (de janvier 1947 à la semaine 9 de 2013), CINAHL (de janvier 1982 au 5 mars 2013), ClinicalTrials.gov (recherche effectuée le 30 mai 2013), WHO ICTRP (recherche effectuée le 3 juin 2013) et les listes bibliographiques des articles pertinents.

Critères de sélection

Nous avons inclus les essais randomisés ou quasi-randomisés avec deux ou plusieurs groupes chez les femmes avec des preuves cliniques ou urodynamiques d'incontinence urinaire liée au stress, d'incontinence urinaire d'urgence ou d'incontinence urinaire mixte. Un groupe de l'essai inclus l’EMPP ajouté à un autre traitement actif; l'autre groupe inclus le même traitement actif seul.

Recueil et analyse des données

Deux auteurs de la revue ont évalué l'éligibilité et la qualité méthodologique des essais et ont résolu les désaccords par discussion ou consultation avec un tiers. Nous avons extrait et traité les données en conformité avec le guide d’examen systématique des interventions Cochrane. D'autres sources potentielles de biais que nous avons inclus dans l'évaluation du «risque de biais» étaient l'approbation éthique, le conflit d'intérêt et la source de financement.

Résultats principaux

Onze essais remplissaient les critères d'éligibilité à l'inclusion, comprenant les femmes souffrant d'incontinence urinaire liée au stress (IUS), d'incontinence urinaire d’urgence (IUI) ou d'incontinence urinaire mixte (IUM) et ils comparaient l’EMPP ajouté à un autre traitement actif (494 femmes) avec le même traitement actif seul (490 femmes). Les comparaisons préalablement spécifiées ont été signalées par des essais uniques à l'exception de la stimulation électrique qui était rapportée dans deux essais. Cependant, les deux essais rendant compte de la stimulation électrique n'ont pas pu être combinés, l'un des essais n'ayant pas rapporté de données pertinentes. Les essais inclus étaient à risque de biais incertain pour la plupart des domaines, principalement en raison du manque d'information adéquate dans un certain nombre d'essais. Ce qui a affecté notre évaluation de la qualité des preuves. 

La majorité des essais ne rapportaient pas les principaux critères de jugement spécifiés dans la revue (guérison/amélioration, qualité de vie) ou ils mesuraient les critères de jugement de différentes manières. Les effets estimés, provenant d'essais uniques de petite taille à travers un certain nombre de comparaisons, n’ont pas été déterminants pour les critères de jugement principaux concernant les symptômes et nous avons noté la qualité des preuves, en utilisant l'approche GRADE, comme faible ou très faible. Il y avait des preuves de qualité moyenne issues d'un seul essai étudiant les femmes souffrant d'IUS, d’IUI, ou d’IUM, montrant qu' une proportion plus importante de femmes ayant reçu une combinaison d’EMPP et de fiches générant de la chaleur et de la vapeur rapportaient une guérison par rapport à celles qui recevaient les fiches seules : 19/37 (51%) versus 8/37 (22%) avec un risque relatif (RR) de 2,38, intervalle de confiance (IC) à 95% de 1,19 à 4,73). Davantage de femmes avaient rendu compte de guérison ou d'amélioration de l'incontinence dans un autre essai comparant l’EMPP utilisé en complément avec les cônes vaginaux avec les cônes vaginaux seuls : 14/15 (93%) versus 14/19 (75%), mais cela n'était pas statistiquement significatif (RR 1,27, IC à 95% de 0,94 à 1,71). Nous avons estimé que la qualité des preuves était très faible. Seul un essai évaluant l’EMPP ajouté à un traitement médicamenteux a fourni des informations concernant les effets indésirables (RR de 0,84, IC à 95% de 0,45 à 1,60; preuves de très faible qualité).

En ce qui concerne les conditions spécifiques de la qualité de vie, il n'y avait aucune différence statistiquement significative entre les femmes (avec l'IUS, l’IUI, ou l’IUM) ayant suivi un EMPP en complément de la rééducation de la vessie et celles ayant reçu la rééducation vésicale seule, ceci trois mois après le traitement, soit sur l’échelle révisée du questionnaire sur l’impact de l'incontinence (différence moyenne (DM) de -5,90, IC à 95% -de 35,53 à 23,73), soit sur l’échelle de l’inventaire sur la détresse urogénitale (DM -18.90, IC à 95% - de 37,92 à 0,12). Un résultat similaire était observé entre les femmes souffrant d'IUS et ayant suivi un EMPP associé au pessaire de continence ou à la duloxétine et celles ayant reçu uniquement le pessaire de continence ou la duloxétine. Dans l'ensemble de ces comparaisons, la qualité des preuves pour les critères de jugement variait de modérée à très faible.

Conclusions des auteurs

Cette revue systématique n'a pas trouvé suffisamment de preuves pour affirmer si oui ou non il y avait des effets en ajoutant l’EMPP à un autre traitement actif par rapport au même traitement actif seul pour l'incontinence urinaire (IUS, IUI, ou IUM) chez les femmes. Ces résultats doivent être interprétés avec prudence car la plupart des comparaisons ont été étudiées dans des essais uniques de petite taille. Aucun des essais dans cette revue n’était de taille suffisante pour fournir des preuves fiables. Par ailleurs, aucun des essais inclus n’a rapporté de données sur les effets indésirables associés à l’EMPP, ce qui rend la sécurité de l’EMPP très difficile à évaluer.

Plain language summary

Pelvic floor muscle training added to another active treatment versus the same active treatment alone for urinary incontinence in women

Involuntary leakage of urine (urinary incontinence) affects women of all ages, particularly older women who live in residential care such as nursing homes. Some women leak urine during exercise or when they cough or sneeze (stress urinary incontinence) and this may occur as a result of weakness of the pelvic floor muscles such as damage during childbirth. Other women leak urine before going to the toilet when there is a sudden and compelling need to pass urine (urgency urinary incontinence) and this may be caused by involuntary contraction of the bladder muscle. Mixed urinary incontinence is the combination of both stress and urgency urinary incontinence. Pelvic floor muscle training is a supervised treatment and it involves muscle-clenching exercises to strengthen the pelvic floor muscles. It is a common treatment used by women to stop urine leakage. Other treatments are also available which can either be used alone or in combination with pelvic floor muscle training.

In this review, the combination of pelvic floor muscle training with another active treatment was compared with the same active treatment alone for the treatment of all types of urine leakage. There was not enough evidence say whether or not the addition of pelvic floor muscle training to another active treatment results in more benefits when compared to the same active treatment alone. There was also insufficient evidence to evaluate the adverse events associated with the addition of PFMT to other active treatment.

Résumé simplifié

Entraînement des muscles du plancher pelvien ajouté à un autre traitement actif par rapport au même traitement actif seul pour l'incontinence urinaire chez la femme

Les fuites involontaires d'urine (incontinence urinaire) affectent les femmes de tout âge, en particulier chez les femmes plus âgées vivant dans des institutions spécialisées telles que des maisons de retraite médicalisées. Certaines femmes souffrent d'incontinence urinaire pendant l'exercice ou lorsqu' elles toussent ou éternuent (l'incontinence urinaire liée au stress) et cela peut survenir en raison de la faiblesse des muscles du plancher pelvien, telle que des lésions pendant l'accouchement. D'autres femmes souffrent d'incontinence urinaire avant de se rendre aux toilettes, lorsqu'elles ont un besoin soudain d'uriner (incontinence urinaire d'urgence) et cela peut être causé par une contraction involontaire des muscles de la vessie. L'incontinence urinaire mixte est la combinaison des deux, l'incontinence urinaire liée au stress et d'urgence. L’entraînement des muscles du plancher pelvien (EMPP) est un traitement supervisé qui implique des exercices de contraction musculaire pour renforcer les muscles pelviens. Ce traitement est couramment utilisé chez les femmes pour arrêter les fuites urinaires. D'autres traitements sont également disponibles et peuvent être utilisés seuls ou en combinaison avec un entraînement des muscles pelviens.

Dans cette revue, la combinaison d'entraînement des muscles du plancher pelvien avec un autre traitement actif a été comparée au même traitement actif seul pour le traitement de tous les types de fuites urinaires. Il n'y avait pas suffisamment de preuves pour déterminer si oui ou non l'ajout d'entraînement des muscles du plancher pelvien à un autre traitement actif apportait plus de bénéfices par rapport au même traitement actif seul. Les preuves étaient également insuffisantes pour comparer les effets indésirables associés à l'ajout de la formation PFMT par rapport à un autre traitement actif.

Notes de traduction

Traduit par: French Cochrane Centre 1st November, 2013
Traduction financée par: Financeurs pour le Canada : Instituts de Recherche en Santé du Canada, Ministère de la Santé et des Services Sociaux du Québec, Fonds de recherche du Québec-Santé et Institut National d'Excellence en Santé et en Services Sociaux; pour la France : Ministère en charge de la Santé

Summary of findings(Explanation)

Summary of findings for the main comparison. PFMT added to vaginal cones versus vaginal cones alone for urinary incontinence in women
  1. 1Random sequence generation and allocation concealment unclear.
    2Confidence interval is very wide (0.94 to 1.71).

PFMT added to vaginal cones versus vaginal cones alone for urinary incontinence in women
Patient or population: patients with urinary incontinence in women
Settings:
Intervention: PFMT added to vaginal cones versus vaginal cones alone
OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
Control PFMT added to vaginal cones versus vaginal cones alone
Number of women cured or improved (subjective) - not reportedSee commentSee commentNot estimable-Not reported 
Number of women reporting incontinence at 1 year or more after treatment (subjective) - not reportedSee commentSee commentNot estimable-Not reported 
Objective measure of urine leakage (pad test) Study population RR 1.27
(0.94 to 1.71)
34
(1 study)
⊕⊝⊝⊝
very low 1,2
 
737 per 1000 936 per 1000
(693 to 1000)
Moderate
737 per 1000 936 per 1000
(693 to 1000)
Number of women experiencing pain - not reportedSee commentSee commentNot estimable-Not reported 
Condition-specific quality of life assessed by patient questionnaire such as Incontinence Impact Questionnaire (IIQ), King's Health Questionnaire (KHQ) - not reportedSee commentSee commentNot estimable-Not reported 
General health status evaluation e.g. Short Form (SF)-36 - not reportedSee commentSee commentNot estimable-Not reported 
Number of women requiring further treatment such as surgery, drugs, mechanical devices (relapse) - not reportedSee commentSee commentNot estimable-Not reported 
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (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; PFMT: pelvic floor muscle training; RR: risk ratio
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Summary of findings 2 PFMT added to lifestyle intervention versus lifestyle intervention alone for urinary incontinence in women

Summary of findings 2. PFMT added to lifestyle intervention versus lifestyle intervention alone for urinary incontinence in women
PFMT added to lifestyle intervention versus lifestyle intervention alone for urinary incontinence in women
Patient or population: patients with urinary incontinence in women
Settings:
Intervention: PFMT added to lifestyle intervention versus lifestyle intervention alone
OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
Control PFMT added to lifestyle intervention versus lifestyle intervention alone
Number of women cured or improved (subjective) - not reportedSee commentSee commentNot estimable-Not reported 
Number of women reporting incontinence at 1 year or more after treatment (subjective) - not reportedSee commentSee commentNot estimable-Not reported 
Objective measure of urine leakage (e.g. pad test) - not reportedSee commentSee commentNot estimable-Not reported 
Number of women reporting adverse events - not reportedSee commentSee commentNot estimable-Not reported 
Condition-specific quality of life - not reportedSee commentSee commentNot estimable-Not reported 
General health status evaluation e.g. Short Form (SF)-36 - not reportedSee commentSee commentNot estimable-Not reported 
Number of women requiring further treatment such as surgery, drugs, mechanical devices (relapse) - not reportedSee commentSee commentNot estimable-Not reported 
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (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; PFMT: pelvic floor muscle training
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Summary of findings 3 PFMT added to bladder training versus bladder training alone for urinary incontinence in women

Summary of findings 3. PFMT added to bladder training versus bladder training alone for urinary incontinence in women
  1. 1Random sequence generation and allocation concealment is unclear.
    2Confidence interval is very wide (0.84 to 3.46).
    3Confidence interval is very wide (-35.53 to 23.73).
    4Confidence interval is very wide (0.57 to 1.57).

PFMT added to bladder training versus bladder training alone for urinary incontinence in women
Patient or population: patients with urinary incontinence in women
Settings:
Intervention: PFMT added to bladder training versus bladder training alone
OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
Control PFMT added to bladder training versus bladder training alone
Number of women cured - 3 months after treatment Study population RR 1.71
(0.84 to 3.46)
122
(1 study)
⊕⊝⊝⊝
very low 1,2
 
159 per 1000 271 per 1000
(133 to 549)
Moderate
159 per 1000 272 per 1000
(134 to 550)
Number of women reporting incontinence at 1 year or more after treatment (subjective) - not reportedSee commentSee commentNot estimable-Not reported 
Objective measure of urine leakage (e.g. pad test) - not reportedSee commentSee commentNot estimable-Not reported 
Number of women experiencing pain - not reportedSee commentSee commentNot estimable-Not reported 
Condition-specific quality of life - 3 months after treatment
Incontinence Impact Questionnaire- Revised (IIQ-R)
 The mean condition-specific quality of life - 3 months after treatment in the intervention groups was
5.9 lower
(35.53 lower to 23.73 higher)
 118
(1 study)
⊕⊝⊝⊝
very low 1,3
 
General health status evaluation e.g. Short Form (SF)-36 - not reportedSee commentSee commentNot estimable-Not reported 
Number of women requiring further treatment such as surgery, drugs, mechanical devices (relapse) 396 per 1000 376 per 1000
(226 to 621)
RR 0.95
(0.57 to 1.57)
96
(1 study)
⊕⊝⊝⊝
very low 1,4
 
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (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; PFMT: pelvic floor muscle training; RR: risk ratio
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Summary of findings 4 PFMT added to electrical stimulation versus electrical stimulation alone (excluding implanted electrodes) for urinary incontinence in women

Summary of findings 4. PFMT added to electrical stimulation versus electrical stimulation alone (excluding implanted electrodes) for urinary incontinence in women
  1. 1Random sequence generation and allocation concealment unclear.
    2Confidence interval very wide (0.37 to 24.58).

PFMT added to electrical stimulation versus electrical stimulation alone (excluding implanted electrodes) for urinary incontinence in women
Patient or population: patients with urinary incontinence in women
Settings:
Intervention: PFMT added to electrical stimulation versus electrical stimulation alone (excluding implanted electrodes)
OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
Control PFMT added to electrical stimulation versus electrical stimulation alone (excluding implanted electrodes)
Number of women cured Study population RR 3
(0.37 to 24.58)
22
(1 study)
⊕⊝⊝⊝
very low 1,2
 
91 per 1000 273 per 1000
(34 to 1000)
Moderate
91 per 1000 273 per 1000
(34 to 1000)
Number of women reporting incontinence at 1 year or more after treatment (subjective) - not reportedSee commentSee commentNot estimable-Not reported 
Objective measure of urine leakage (e.g. pad test) - not reportedSee commentSee commentNot estimable-Not reported 
Number of women experiencing pain - not reportedSee commentSee commentNot estimable-Not reported 
Condition-specific quality of life assessed by patient questionnaire such as Incontinence Impact Questionnaire (IIQ), King's Health Questionnaire (KHQ) - not reportedSee commentSee commentNot estimable-Not reported 
General health status evaluation e.g. Short Form (SF)-36 - not reportedSee commentSee commentNot estimable-Not reported 
Number of women requiring further treatment such as surgery, drugs, mechanical devices (relapse) - not reportedSee commentSee commentNot estimable-Not reported 
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (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; PFMT: pelvic floor muscle training; RR: risk ratio
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Summary of findings 5 PFMT added to magnetic stimulation versus magnetic stimulation alone for urinary incontinence in women

Summary of findings 5. PFMT added to magnetic stimulation versus magnetic stimulation alone for urinary incontinence in women
PFMT added to magnetic stimulation versus magnetic stimulation alone for urinary incontinence in women
Patient or population: patients with urinary incontinence in women
Settings:
Intervention: PFMT added to magnetic stimulation versus magnetic stimulation alone
OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
Control PFMT added to magnetic stimulation versus magnetic stimulation alone
Number of women cured or improved (subjective) - not reportedSee commentSee commentNot estimable-Not reported 
Number of women reporting incontinence at 1 year or more after treatment (subjective) - not reportedSee commentSee commentNot estimable-Not reported 
Objective measure of urine leakage (e.g. pad test) - not reportedSee commentSee commentNot estimable-Not reported 
Number of women reporting adverse events - not reportedSee commentSee commentNot estimable-Not reported 
Condition-specific quality of life assessed by patient questionnaire such as Incontinence Impact Questionnaire (IIQ), King's Health Questionnaire (KHQ) - not reportedSee commentSee commentNot estimable-Not reported 
General health status evaluation e.g. Short Form (SF)-36 - not reportedSee commentSee commentNot estimable-Not reported 
Number of women requiring further treatment such as surgery, drugs, mechanical devices (relapse) - not reportedSee commentSee commentNot estimable-Not reported 
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (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; PFMT: pelvic floor muscle training
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Summary of findings 6 PFMT added to continence pessary versus continence pessary alone for urinary incontinence in women

Summary of findings 6. PFMT added to continence pessary versus continence pessary alone for urinary incontinence in women
  1. 1Wide confidence interval (0.67 to 1.16).
    2Confidence interval is wide (0.62 to 1.08).

PFMT added to continence pessary versus continence pessary alone for urinary incontinence in women
Patient or population: patients with urinary incontinence in women
Settings:
Intervention: PFMT added to continence pessary versus continence pessary alone
OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
Control PFMT added to continence pessary versus continence pessary alone
Number of women cured or improved (subjective) at 12 months Study population RR 0.88
(0.67 to 1.16)
207
(1 study)
⊕⊕⊕⊝
moderate 1
 
531 per 1000 468 per 1000
(356 to 616)
Moderate
579 per 1000 510 per 1000
(388 to 672)
Number of women reporting incontinence at 1 year or more after treatment (subjective) - not reportedSee commentSee commentNot estimable-Not reported 
Objective measure of urine leakage (e.g. pad test) - not reportedSee commentSee commentNot estimable-Not reported 
Number of women reporting adverse events - not reportedSee commentSee commentNot estimable-Not reported 
Condition-specific quality of life at 12 months
Urogenital Distress Inventory (UDI)
Study population RR 0.81
(0.62 to 1.08)
207
(1 study)
⊕⊕⊕⊝
moderate 2
 
542 per 1000 439 per 1000
(336 to 585)
Moderate
494 per 1000 400 per 1000
(306 to 534)
General health status evaluation e.g. Short Form (SF)-36 - not reportedSee commentSee commentNot estimable-Not reported 
Number of women requiring further treatment such as surgery, drugs, mechanical devices (relapse) - not reportedSee commentSee commentNot estimable-Not reported 
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (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; PFMT: pelvic floor muscle training; RR: risk ratio
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Summary of findings 7 PFMT added to drug therapy versus drug therapy alone for urinary incontinence in women

Summary of findings 7. PFMT added to drug therapy versus drug therapy alone for urinary incontinence in women
  1. 1Random sequence generation and allocation concealment is unclear.
    2Confidence interval is very wide (0.83 to 1.63).
    3Confidence interval is very wide (0.45 to 1.50).
    4Confidence interval is very wide (-2.08 to 13.76).

PFMT added to drug therapy versus drug therapy alone for urinary incontinence in women
Patient or population: patients with urinary incontinence in women
Settings:
Intervention: PFMT added to drug therapy versus drug therapy alone
OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
Control PFMT added to drug therapy versus drug therapy alone
Number of women cured - PFMT + clenbuterol versus clenbuterol Study population RR 1.16
(0.83 to 1.63)
32
(1 study)
⊕⊝⊝⊝
very low 1,2
 
769 per 1000 892 per 1000
(638 to 1000)
Moderate
769 per 1000 892 per 1000
(638 to 1000)
Number of women reporting incontinence at 1 year or more after treatment (subjective) - not reportedSee commentSee commentNot estimable-Not reported 
Objective measure of urine leakage (e.g. pad test) - not reportedSee commentSee commentNot estimable-Not reported 
Number of women reporting adverse events 207 per 1000 174 per 1000
(1000 to 332)
RR 0.84
(45 to 1.60)
162
(1 study)
⊕⊝⊝⊝
very low 1,3
 
Condition-specific quality of life on I-QoL Questionnaire - PFMT + duloxetine versus duloxetine
Incontinence Quality of Life questionnaire
 The mean condition-specific quality of life on I-QoL questionnaire - PFMT + duloxetine versus duloxetine in the intervention groups was
5.84 higher
(2.08 lower to 13.76 higher)
 101
(1 study)
⊕⊕⊝⊝
low 4
 
General health status evaluation e.g. Short Form (SF)-36 - not reportedSee commentSee commentNot estimable-Not reported 
Number of women requiring further treatment such as surgery, drugs, mechanical devices (relapse) - not reportedSee commentSee commentNot estimable-Not reported 
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (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; PFMT: pelvic floor muscle training; RR: risk ratio
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Summary of findings 8 PFMT prior to surgical intervention versus surgical intervention alone for urinary incontinence in women

Summary of findings 8. PFMT prior to surgical intervention versus surgical intervention alone for urinary incontinence in women
PFMT prior to surgical intervention versus surgical intervention alone for urinary incontinence in women
Patient or population: patients with urinary incontinence in women
Settings:
Intervention: PFMT prior to surgical intervention versus surgical intervention alone
OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
Control PFMT prior to surgical intervention versus surgical intervention alone
Number of women cured or improved (subjective) - not reportedSee commentSee commentNot estimable-Not reported 
Number of women reporting incontinence at 1 year or more after treatment (subjective) - not reportedSee commentSee commentNot estimable-Not reported 
Objective measure of urine leakage (e.g. pad test) - not reportedSee commentSee commentNot estimable-Not reported 
Number of women reporting adverse events - not reportedSee commentSee commentNot estimable-Not reported 
Condition-specific quality of life assessed by patient questionnaire such as Incontinence Impact Questionnaire (IIQ), King's Health Questionnaire (KHQ) - not reportedSee commentSee commentNot estimable-Not reported 
General health status evaluation e.g. Short Form (SF)-36 - not reportedSee commentSee commentNot estimable-Not reported 
Number of women requiring further treatment such as surgery, drugs, mechanical devices (relapse) - not reportedSee commentSee commentNot estimable-Not reported 
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (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; PFMT: pelvic floor muscle training
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Summary of findings 9 PFMT added to HSGS versus HSGS alone for urinary incontinence in women

Summary of findings 9. PFMT added to HSGS versus HSGS alone for urinary incontinence in women
  1. 1Allocation concealment unclear.

PFMT added to HSGS versus HSGS alone for urinary incontinence in women
Patient or population: patients with urinary incontinence in women
Settings:
Intervention: PFMT added to other versus other treatment alone
OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
Control PFMT added to other versus other treatment alone
Number of women cured Study population RR 2.38
(1.19 to 4.73)
74
(1 study)
⊕⊕⊕⊝
moderate 1
 
216 per 1000 515 per 1000
(257 to 1000)
Moderate
216 per 1000 514 per 1000
(257 to 1000)
Number of women reporting incontinence at 1 year or more after treatment (subjective) - not reportedSee commentSee commentNot estimable-Not reported 
Objective measure of urine leakage (e.g. pad test) - not reportedSee commentSee commentNot estimable-Not reported 
Number of women experiencing pain - not reportedSee commentSee commentNot estimable-Not reported 
Condition-specific quality of life assessed by patient questionnaire such as Incontinence Impact Questionnaire (IIQ), King's Health Questionnaire (KHQ) - not reportedSee commentSee commentNot estimable-Not reported 
General health status evaluation e.g. Short Form (SF)-36 - not reportedSee commentSee commentNot estimable-Not reported 
Number of women requiring further treatment such as surgery, drugs, mechanical devices (relapse) - not reportedSee commentSee commentNot estimable-Not reported 
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (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; HSGS: heat and steam generating sheet; PFMT: pelvic floor muscle training; RR: risk ratio
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Background

Different treatment options are currently available for the management of urinary incontinence in women. Conservative interventions include:

Drug therapies include anticholinergics (Madhuvrata 2012; Nabi 2006), duloxetine (Mariappan 2005), local vaginal oestrogens (Cody 2012) and intravesical botulinum toxin (Duthie 2011). Surgical interventions include sling procedures (Ogah 2009; Rehman 2011), colposuspension (Dean 2006; Lapitan 2012) and injection of peri-urethral bulking agents (Kirchin 2012).

The focus of this review is to determine the benefits of adding PFMT to any of the treatments above for the management of urinary incontinence in women. There is a separate Cochrane review dealing with the conservative treatment of postprostatectomy urinary incontinence in men (Campbell 2012).

Description of the condition

Urinary incontinence or loss of bladder control, according to International Continence Society (ICS), is defined as the complaint of any involuntary loss of urine (Abrams 2009). It is a common problem that may affect women of all ages with a wide range of severity and a variety of symptoms; however, it is more prevalent in older women particularly amongst those in institutionalised care (Milsom 2009).

The prevalence of urinary incontinence varies, depending on the age of the study population, the study methods and settings, as well as the definition of the problem (Culligan 2000). In the general population, the estimated prevalence of urinary incontinence in middle-aged and older women ranges from 30% to 60% and increases with advancing age: the prevalence of daily urinary incontinence ranges from 5% to 15%, and is over 15% in institutionalised women who are over the age of 70 (Milsom 2009). Nonetheless, these figures may not actually reflect the true nature, size and scope of this problem for it is usually under-diagnosed and under-reported due to its embarrassing nature and associated stigmatisation (Shaw 2001a).

Urinary incontinence has an impact on many aspects of a woman's life (Grimby 1993; Hunskaar 1991; Sinclair 2011). Women with urinary incontinence have a significant reduction in their quality of life (Shaw 2001b). It significantly affects couples' relationships (Nilsson 2009): it is reported that 25% to 50% of incontinent women experience sexual dysfunction (Barber 2002). Evidence has also shown that women with urinary incontinence have coexisting psychiatric illness. Melville et al reported that major depression was three times more common in incontinent women compared to their continent counterparts (6.1% versus 2.2%) (Melville 2002). The financial impact of urinary incontinence is enormous: the estimated annual direct cost of treating urinary incontinence in women in the USA was put at USD 12.4 billion in 2001 (Wilson 2001). In the UK, the annual NHS cost of treating clinically significant storage symptoms in women was estimated to be GBP 233 million (Turner 2004). With an increasingly ageing population, these cost are likely to increase in the future.

Types of urinary incontinence

There are three main types of urinary incontinence.

Stress urinary incontinence (SUI)

This is defined by the ICS and the International Urogynecological Association (IUGA) as the complaint of involuntary leakage of urine with coughing, sneezing or physical exertion (Haylen 2010). The term urodynamic stress incontinence (USI) is used to describe involuntary leakage of urine with increased intra-abdominal pressure in the absence of detrusor contraction during urodynamic evaluation (Abrams 2009). Stress urinary incontinence is the most common type of urinary incontinence, affecting an estimated 50% (half) of all incontinent women (Milsom 2009). It is more prevalent in young and middle-aged women, particularly white, non-Hispanic (Milsom 2009). It is often associated with weakness of pelvic floor support (muscles and collagen-dependent tissues) (Long 2008), damage to the bladder sphincter mechanism, or both, resulting in bladder neck hypermobility and rotational descent of the proximal urethra with associated intrinsic sphincter deficiency (Schorge 2008). This results in reduction of urethral closure pressure and consequently urine leakage during exertion or physical exercise.

Risk factors for SUI in women include pregnancy, vaginal delivery, increasing parity, advancing age, post-menopausal state, obesity (MacArthur 2006; MacLennan 2000) and gynaecological procedures such as hysterectomy (Allahdin 2008). The aim of treatment is to strengthen the pelvic floor support or restore the normal function of the sphincter mechanism, or both.

Urgency urinary incontinence (UUI)

This is defined by the IUGA and ICS as the complaint of involuntary leakage of urine associated with urgency (Haylen 2010). Urgency is a sudden and compelling desire to void urine which is difficult to defer (Abrams 2009). Overactive bladder (OAB) is the presence of urinary urgency usually associated with frequency and nocturia with UUI (OAB-wet) or without UUI (OAB-dry) in the absence of urinary tract infection (UTI) or other pathology (Haylen 2010). Urinary frequency is defined as passing urine more than eight times in 24 hours (Fitzgerald 2003; Fitzgerald 2002) while nocturia is waking up from sleep more than once per night to urinate (van Kerrebroeck 2002). In patients with detrusor overactivity (DO), a spontaneous or induced detrusor contraction is observed during urodynamic testing (Abrams 2009). UUI is more prevalent in older women and accounts for a small proportion of women with urinary incontinence (Milsom 2009). In continent individuals, reflex (involuntary) contraction of the pelvic floor muscles and the striated muscle of the urethra occurs during the filling (storage) phase of the bladder (Morrison 1995). This in turn leads to increased intra-urethral pressure and reflex inhibition of detrusor contraction thereby preventing urine leakage and urgency. Thus any abnormality of the pelvic floor muscles (structural or neural) which disrupts this reflex inhibition of detrusor during the filing phase may result in urgency urinary incontinence.

The cause of urgency urinary incontinence in some cases is idiopathic (unknown cause). Other causes include neurogenic (multiple sclerosis, Alzheimer's or Parkinson's disease), stroke, tumour of the bladder and bladder pain syndrome (interstitial cystitis), defined by the ICS as "an unpleasant sensation (pain, pressure, discomfort) perceived to be related to the urinary bladder associated with lower urinary tract symptom(s) of more than 6 weeks duration, in the absence of infection or other identifiable causes" (Abrams 2009). The aim of treatment is to reduce the symptoms of OAB or UUI.

Mixed urinary incontinence (MUI)

This is the complaint of involuntary leakage of urine associated with urgency and also with exertion, effort, sneezing and coughing (Abrams 2009). The prevalence of MUI increases with age. It has been suggested that mixed urinary incontinence should be managed conservatively or with drugs at first so as to reduce the need for surgical intervention (Karram 1989). However, if symptoms persist without significant evidence of detrusor overactivity on urodynamics, surgery may be performed.

Description of the intervention

Pelvic floor muscle training (PFMT)

Pelvic floor muscle training (PFMT) was popularised by Arnold Kegel (Kegel 1948) for the management of urinary incontinence and has since remained a first-line conservative measure. It is commonly recommended for the treatment of patients with stress or mixed urinary incontinence (Dumoulin 2010). Less commonly, it can be used for urgency urinary incontinence.

The main aim of PFMT is to improve the function of the pelvic floor muscles in terms of strength, endurance and co-ordination, thereby providing maximum support to the pelvic organs (particularly, the bladder neck and the proximal urethra), before and during an increase in intra-abdominal pressure so as to prevent urine leakage. There are different ways through which PFMT appears to work (Bø 2004):

  • Patients can learn how to use conscious pelvic floor muscle pre-contraction before and during exertion to prevent urine leakage (co-ordination).

  • Pelvic floor muscle strength training increases long-lasting muscle volume thereby providing structural support to the pelvic organs (strengthening).

The reported cure rates of PFMT vary (Bernstein 1997; Bø 1999; Kegel 1948), depending on a number of factors. These factors include the type and severity of incontinence, type of instruction and follow-up, patients' adherence and the outcome measures used. Structured, supervised and more intensive programmes have been associated with more success than simple verbal instructions (Bø 1990; Dumoulin 2010).

How the intervention might work

Strong, fast and well-timed voluntary pelvic floor muscle contractions have the effect of pressing the urethra against the posterior aspect of the symphysis pubis, thereby producing a mechanical increase in intra-urethral pressure (DeLancey 1988). Thus a positive urethral closure pressure is maintained during increase in intra-abdominal pressure, resulting in correction of the negative closure pressure usually observed in patients with stress incontinence.

Pelvic floor muscle strength training also aims to provide more support to the bladder neck and proximal urethra, which are observed to be poorly supported in some patients with urinary incontinence, by raising the position of the levator ani muscle through increased muscle volume (hypertrophy) and muscle stiffness (Bø 2004). The overall effect of this is to raise urethral closure pressure at rest and during increase in intra-abdominal pressure.

In urgency urinary incontinence, there is inability to inhibit detrusor contractions, leading to abnormally high detrusor pressures. Reflex inhibition of detrusor activity has been shown to follow electrical stimulation of pelvic floor muscles (Godec 1975) and may also accompany repeated and conscious pelvic floor muscle contraction (Polden 1990) thereby controlling UUI. However, the timing, number, intensity and duration of the pelvic floor muscle contraction considered adequate to inhibit detrusor contraction are unknown (Dumoulin 2010).

It is possible that adding other active treatments to basic PFMT may enhance its effectiveness, particularly if those treatments are effective in their own right.

Why it is important to do this review

To date, there is no sufficient evidence-based rationale indicating that PFMT in combination with another active treatment is a better treatment of choice than the active treatment alone for urinary incontinence in women. Adding a treatment such as PFMT might be time consuming, increase resource use and decrease adherence. Therefore if adding PFMT does not improve outcome over and above the other treatment, then there is no point incurring extra cost (both direct and indirect) for no added benefit. Thus a considerable doubt exists about the real and potential therapeutic effectiveness, cost-effectiveness and risks of PFMT added to another active treatment in comparison with the active treatment alone for the treatment of women with urinary incontinence. Therefore, there is a compelling need for a systematic review of the existing trial-based evidence. The outcome of this review will complement what is already known about the effectiveness of PFMT (Boyle 2012; Dumoulin 2010; Hay-Smith 2011; Herderschee 2011).

Objectives

To compare the effects of pelvic floor muscle training combined with another active treatment versus the same active treatment alone in the management of women with urinary incontinence.

Methods

Criteria for considering studies for this review

Types of studies

Randomised controlled trials and quasi-randomised trials (for example allocation by alternation) of pelvic floor muscle training added to an active treatment versus the active treatment alone for urinary incontinence in women were included. Trials using more than two arms of interventions were also included, providing one of the arms involved the use of PFMT plus an active treatment and another arm involved the same active treatment alone. Other forms of clinical trials were excluded.

Types of participants

Adult women with stress urinary incontinence (SUI), urgency urinary incontinence (UUI) or mixed urinary incontinence (MUI).

This review considered all modes of diagnosis of incontinence (symptoms, signs and/or urodynamic evaluation) for inclusion. This is because many patients are referred for PFMT on the basis of symptoms and/or clinical signs alone as there is no consensus yet on the need for urodynamic testing before PFMT is performed (Glazener 2012; Thuroff 2011). Also, the outcome of a conservative management of urinary incontinence has been shown to be no different with respect to the mode of diagnosis (Elser 1999). Trials that recruited men and women were considered for inclusion providing demographic and outcome data were reported separately for women.

Studies of women with urinary incontinence whose symptoms were due to significant external factors, for example cognitive impairment, neurological disorders and lack of independent mobility, which are considered to be outside the urinary tract, were excluded. Studies that recruited women with nocturnal enuresis were also excluded.

Studies which specifically investigated antenatal or postnatal women (up to three months after delivery) were excluded from this review. The effect of PFMT might differ in this group of women given the physiological changes that occur during pregnancy and the postpartum period. These women have been considered in another Cochrane review (Boyle 2012).

Studies which recruited women in long-term care facilities were also excluded. Urinary incontinence in this category of women is often associated with other co-morbid conditions such as dementia, depression, lack of independent mobility, etc. (Milsom 2009) which might influence the outcome of PFMT or their ability to comply with treatment.

Types of interventions

One arm of the trial used pelvic floor muscle training (PFMT) added to another active treatment. The comparison was the same active treatment alone.

In this review, PFMT was counted as a programme of repeated voluntary pelvic floor muscle contractions taught or supervised (or both) by healthcare professionals. All types of PFMT programmes were considered for inclusion, for example variations in timing and purpose of PFMT (such as PFMT for strengthening, PFMT for urge suppression), ways of teaching PFMT, and types and number of contractions. If biofeedback was used once or more in the teaching or delivery of PFMT we called this a PFMT intervention, and clearly labelled any trial that used biofeedback as a PFMT + biofeedback trial to recognise the potential additional effect of biofeedback. Trials in which PFMT was combined with advice on frequency/urgency strategies (but without a scheduled voiding regimen characteristic of bladder training) or other lifestyle advice (such as weight reduction) with leaflets or verbal instructions only was considered as 'pure PFMT'.

The comparisons were:

A Physical

1. PFMT added to vaginal cones versus vaginal cones alone

B Behavioural

2. PFMT added to lifestyle intervention (e.g. weight reduction) versus lifestyle intervention alone (lifestyle intervention must be structured or supervised)

3. PFMT added to bladder training versus bladder training alone (bladder training must include scheduled voiding regimen)

C Electrical/magnetic

4. PFMT added to electrical stimulation versus electrical stimulation alone (excluding implanted electrodes)

5. PFMT added to magnetic stimulation versus magnetic stimulation alone

D Mechanical

6. PFMT added to continence pessaries versus continence pessaries alone

E Drugs

7. PFMT added to drug therapy (e.g. tolterodine, duloxetine) versus drug therapy alone

F Surgery

8. PFMT prior to surgical intervention (e.g. tension-free vaginal tape (TVT)) versus surgical intervention alone

G Other

9. PFMT added to any other standalone active treatment versus the same standalone active treatment.

Types of outcome measures

The Standardisation Committee of the International Continence Society recommended that research looking into the effects of therapeutic interventions for women with urinary incontinence should take into consideration the following five outcome domains: patient's observations with respect to the symptoms of urinary incontinence, quantification of patient's symptoms, clinician's observations (functional and anatomical), patient's quality of life and socio-economic implication of treatment (Lose 1998). For this review, one or more outcomes of interest were considered from each domain.

Primary outcomes
Women's observations
  • Number of women cured of symptoms of urinary incontinence (within first year, as reported by the participants and not the clinicians)

  • Number of women cured or improved (as reported by the participants and not the clinicians)

  • Symptom and condition-specific quality of life assessed by various measures such as the Urinary Incontinence Quality of Life (I-QoL) scale, King's Health Questionnaire, the Incontinence Impact Questionnaire (IIQ), the Social Activity Index, the Leicester Impact Scale, etc.

  • Number of women improved on patient global impression of improvement in the first three months after the end of treatment

Secondary outcomes
1. Quantification of symptoms
  • Number of women reporting incontinence at one year or more after treatment (subjective)

  • Number of micturitions during the day

  • Number of micturitions during the night

  • Urine loss (measured on pad or paper towel weight tests)

  • Other quantification of symptoms reported by individual trials

2. Clinician's observations
  • Objective measurement of incontinence such as observation of urine leakage during cough test

  • Measurement of pelvic floor muscle function such as electromyography, vaginal squeeze pressure, pelvic floor muscle force and morphological measurements (dynamometry, ultrasound)

3. Generic quality of life
  • General health status evaluation e.g. Short Form (SF)-36, Norwegian version of the Quality of Life Scale (QoLS-N), etc.

  • Other quality of life measures as reported by individual trials

4. Economic analysis
  • Costs of intervention, resource implications of differences in outcomes and overall cost utility and cost-effectiveness

5. Adverse effects
  • Number of women reporting adverse events

  • Pain or discomfort

  • Other adverse outcomes as reported by individual trials

6. Other outcomes
  • Sexual function

  • Pelvic organ prolapse

  • Number of women requiring further treatment such as surgery, drugs, mechanical devices (relapse)

  • Treatment adherence evaluation using for example a self administered treatment adherence questionnaire

  • Patient satisfaction with treatment assessed using for example the validated Patient Satisfaction Questionnaire

  • Other outcomes not pre-specified but considered to be important during the review, e.g. long-term follow-up

Quality of evidence

We assessed the quality of evidence by adopting the GRADE approach. The following factors were considered for assessing the quality of evidence:

  1. Limitations in the study design

  2. Inconsistency of results

  3. Indirectness of evidence

  4. Imprecision

  5. Publication bias

The review authors classified primary and secondary outcomes, as defined above, as 'critical', 'important' or 'not important' for decision making from the woman's perspective. The GRADE working group strongly recommends including up to seven critical outcomes in a systematic review (Guyatt 2011a; Guyatt 2011b).

In this systematic review, the seven critical outcomes for assessing the quality of evidence were as follows:

  • Number of women cured or improved (subjective)

  • Condition-specific quality of life assessed by patient questionnaire such as Incontinence Impact Questionnaire (IIQ), King's Health Questionnaire (KHQ)

  • Number of women reporting incontinence at one year or more after treatment (subjective)

  • Objective measure of urine leakage (e.g. pad test)

  • Number of women reporting adverse events

  • General health status evaluation e.g. Short Form (SF-36)

  • Number of women requiring further treatment such as surgery, drugs, mechanical devices

Search methods for identification of studies

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

Electronic searches

This review drew on the search strategy developed for the Cochrane Incontinence Group. We identified relevant trials from the Cochrane Incontinence Group Specialised Register of trials. 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, CINAHL and MEDLINE in Process, 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 of the Specialised Register was: 28 February 2013.

The terms used to search the Incontinence Group Specialised Register are given below:

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

{TOPIC.URINE.INCON*} AND

({INTVENT.PHYS.PFMT*} OR {INTVENT.PHYS.BIOFEED*}) AND

({INTVENT.SURG*} OR {INTVENT.CHEM.DRUG*} OR {INTVENT.PSYCH*} OR {INTVENT.LIFESTYLE*} OR {INTVENT.MECH*} OR {INTVENT.ELECTSTIM*} OR {INTVENT.CONES*})

(All searches were of the keyword field of Reference Manager 12, Thomson Reuters).

For this review we also specifically searched the following databases; the search strategies are given in Appendix 1:

  • EMBASE Classic and EMBASE (on OVID SP) covering 1947 to 2013 Week 9. Date of last search: 7 March 2013.

  • CINAHL on EBSCO Host covering January 1982 to 5 March 2013. Date of last search: 5 March 2013.

  • ClinicalTrials.gov. Date of last search: 30 May 2013.

  • WHO ICTRP. Date of last search: 3 June 2013.

Searching other resources

We searched the references lists of relevant articles and the included and excluded studies in other relevant Cochrane reviews.

Data collection and analysis

Selection of studies

Only randomised and quasi-randomised controlled trials were included. Two review authors independently screened the list of titles and abstracts generated by the search. We retrieved full-text articles of potentially relevant studies. Two review authors independently assessed the full-text articles for eligibility. Any differences of opinion were resolved through discussion or by involving a third party. We listed studies formally considered for the review but excluded, with reasons given for their exclusion.

Data extraction and management

Two of the review authors performed extraction of data from the included studies independently using a standardised form. Any disagreement was resolved by discussion or by consulting a third party. Where there was insufficient information regarding the outcomes or other relevant aspects of the published reports, we contacted study authors. For data entry, we used Review Manager software (RevMan 2012). Processing of data from the included trials was done according to the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

Assessment of risk of bias in included studies

We assessed the risk of bias in the included studies using the Cochrane 'Risk of bias' assessment tool (Higgins 2011). This included:

  • sequence generation;

  • allocation concealment;

  • blinding of participants or therapists;

  • blinding of outcome assessors;

  • completeness of outcome data;

  • selective outcome reporting;

Other potential sources of bias we incorporated into the 'Risk of bias' tables were ethical approval, conflict of interest and funding source. Some of these additional domains are also used in other systematic review (Omar 2013). Two review authors independently assessed the above mentioned domains. Any differences of opinion were resolved through consensus or by consulting a third party.

Measures of treatment effect

Analyses were based on available data from all included trials relevant to the comparisons and outcomes of interest. For trials with multiple publications, only the most up-to-date of the trials or those with complete data for each outcome were included. We planned to undertake a meta-analysis but this could not be done because each of the pre-specified comparisons (except electrical stimulation) was addressed by single trials. For categorical outcomes we related the numbers reporting an outcome to the numbers at risk in each group to calculate a risk ratio (RR) with 95% confidence intervals (CI). For continuous variables we used means and standard deviations to calculate a mean difference (MD) with 95% CI. Where data required for us to calculate RRs or MDs were not given, we utilised the most detailed numerical data available (e.g. test statistics, P values) to calculate the actual numbers or mean and standard deviations.

Unit of analysis issues

The primary analysis was per woman randomised. Initially, we planned to analyse two-period, two-intervention cross-over trials with continuous outcomes by determining the mean person difference between the two treatment periods and the standard error of this mean to obtain the effect estimates for inclusion in a meta-analysis, where possible by using the generic inverse variance method (Higgins 2011). However, cross-over trials were not identified for inclusion in this review. Similarly, we intend to analyse cluster-randomised trials by reducing them to their effective sample size (that is original sample size divided by design effect; design effect = 1 + (M - 1) x ICC where M is the average cluster size and ICC is the intra-cluster correlation coefficient) (Higgins 2011) and then combine the data obtained (dichotomous or continuous) in a meta-analysis where possible. In the end, no cluster-randomised trial was included in this review.

Dealing with missing data

We analysed the data on an intention-to-treat basis as far as possible. By intention-to-treat analysis, we mean that: 1. outcome data must be measured on all participants; 2. all randomised participants must be included in the analysis; and 3. participants must be retained in the intervention groups to which they were assigned (Higgins 2011). However, for this review, the criterion set for intention-to-treat analysis was that participants be retained and analysed in the intervention groups to which they were assigned. Where this was not the case, we considered whether the trial should be excluded. We made attempts to obtain missing data from the original trialists. However, where this was not possible, data were reported as given in the trials, except where there was evidence of differential loss to follow-up between the intervention groups. In that case, the use of imputation of missing data was considered. 

Assessment of heterogeneity

We intended to assess heterogeneity between studies by visual inspection of plots of the data, the Chi2 test for heterogeneity and the I2 statistic (Higgins 2003). We also planned to use the thresholds for interpretation of the I2 statistic as defined by the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). However, in the end, this was not done because none of the pre-specified comparisons were investigated by more than one trial.

Assessment of reporting biases

In view of the difficulty in detecting and correcting for publication bias and other reporting biases, we planned to minimise their potential impact by ensuring a comprehensive search for eligible studies and by watching out for duplication of data. However, as it turned out, each comparison was addressed by a single trial.

Data synthesis

We intended to combine trials with similar interventions in a meta-analysis using a fixed-effect model approach unless there was evidence of heterogeneity across studies. Where significant heterogeneity existed, we planned to use a random-effects approach. However, we could not combine trials because they all addressed different interventions.

Subgroup analysis and investigation of heterogeneity

We intended to do subgroup data analysis where possible by the type of underlying urinary incontinence or lower urinary tract symptoms:

  • stress urinary incontinence;

  • urgency urinary incontinence;

  • mixed urinary incontinence (both stress and urgency urinary incontinence);

  • 'unclear' if there was no clear cut diagnosis with respect to the type of urinary incontinence.

Ultimately, we could not perform subgroup analysis because there were few trials with each addressing different interventions.

Where heterogeneity between trials was found to be substantive, we planned to conduct an investigation to identify its cause(s). The investigation of heterogeneity was meant to address populations and interventions in the individual trials. The investigation could also include subgroup analyses, meta-regression and sensitivity analyses. Where heterogeneity persisted after appropriate investigation and possible removal of outlying trials, a random-effects model could be used in the meta-analysis. In the end, there was no need to investigate heterogeneity as the included trials tested different comparisons.

Sensitivity analysis

We planned to perform sensitivity analysis by including or excluding trials at high risk of bias. However, this was not applicable as meta-analysis could not be performed.

Results

Description of studies

Results of the search

The search produced a total of 641 titles and abstracts out of which we considered 132 full-text articles for further assessment. Twenty-nine reports of 11 trials met the eligibility criteria for inclusion in this review, while 103 reports of 84 studies were excluded (reasons for exclusion are stated in the Characteristics of excluded studies table). The PRISMA flow chart in Figure 1 illustrates the flow of literature through the search and assessment process.

Figure 1.

PRISMA study flow diagram.

Included studies

Nine of the included trials (Ghoniem 2005; Hofbauer 1990; Ishiko 2000; Jeyaseelan 2002; Jin 2012; Kim 2011; Richter 2010; Wise 1993; Wyman 1998) contained more than two intervention arms: description and data were provided for all the arms in this review. The trials included a total of 984 women of whom 494 received some form of PFMT added to another active treatment and 490 received comparator treatments which were the other active treatment alone.

In terms of number of participants per comparison group:

Two trials (Richter 2010; Wyman 1998) reported an a priori power calculation; another one (Burgio 2010a) used an a priori power calculation at an early stage of the trial but later decided to use a conditional power calculation (based on available participants) due to slow accrual of participants.

Sample characteristics
Mode of diagnosis of urinary incontinence

The trials based the diagnosis of urinary incontinence on:

Types of urinary incontinence

The trials recruited women with:

Age

The included trials recruited women aged:

Four trials (Burgio 2010a; Chen 2008; Hofbauer 1990; Jin 2012) did not set age limits (either a lower or an upper limit) while two trials (Jeyaseelan 2002; Wise 1993) did not present any data on the age of the included women.

Frequency of urinary incontinence episodes

Five trials used frequency of incontinence episodes as one of the inclusion criteria:

Duration of urinary incontinence symptoms

In five trials (Burgio 2010a; Ishiko 2000; Jin 2012; Kim 2011; Wyman 1998), duration of UI was reported as one of the baseline characteristics with none using this as an inclusion criterion. The reported mean or median duration of symptoms varied between 2.1 and 8.6 years.

Other characteristics

Exclusion criteria were reported by seven of the included trials (Ghoniem 2005; Hofbauer 1990; Ishiko 2000; Jin 2012; Kim 2011; Richter 2010; Wyman 1998). Common reasons for excluding participants across trials were: presence of uncontrolled diabetes mellitus, persistent urinary tract infection, disease of the nervous system, impaired mental state, advanced pelvic organ prolapse and post-void residual volume more than a specified amount.

Interventions
Pelvic floor muscle training (PFMT)

Detailed descriptions of the PFMT programmes of the included trials are given in the Characteristics of included studies table. The purpose of this review was to examine the additional effects of adding PFMT to another active treatment. Therefore, the review authors were particularly interested in the effectiveness of PFMT with respect to the confirmation of a correct voluntary pelvic floor muscle contraction, duration of PFMT and PFMT 'dose'. Additionally, we were interested in whether the 'experimental group' received any additional intervention to enhance the effectiveness of PFMT.

Confirmation of a correct pelvic floor muscle contraction

Only two trials (Ghoniem 2005; Wise 1993) reported that the correct type of voluntary pelvic floor muscle contraction was confirmed but full details about the mode of confirmation were not reported.

Duration of PFMT

This varied between four and 12 weeks among the trials:

'Dose' of PFMT

A PFMT programme may be prescribed to:

  1. increase strength (i.e. the maximum force generated in a single contraction by a muscle), characterised by low numbers of repetitions with high 'loads'('loads' can be increased by increasing the amount of voluntary efforts with each contraction);

  2. increase endurance (i.e. the ability to contract repetitively or sustain a single contraction over time), characterised by high numbers of repetitions or prolonged contractions with low to moderate 'loads';

  3. co-ordinate muscle activity by using voluntary pelvic floor muscle contraction to either minimise urine leakage (with increased intra-abdominal pressure) or suppress urge (suppression of detrusor contraction) (behavioural training); or

  4. a combination of these. In this review, the trials included the following programmes:

    1. one trial targeted each of: endurance training (Chen 2008), a combination of strength and endurance training (Kim 2011), or a combination of endurance and co-ordination training (Burgio 2010a);

    2. two trials (Ghoniem 2005; Wyman 1998) used a combination of strength, endurance and co-ordination training programmes.

In six trials (Hofbauer 1990; Ishiko 2000; Jeyaseelan 2002; Jin 2012; Richter 2010; Wise 1993), it was difficult to characterise the PFMT programme (contraction effort, frequency, number and duration) because full details were not provided about the key training parameters such as amount and duration of voluntary contractions.

Additional intervention to enhance PFMT effectiveness

Some trials added extra interventions to the PFMT regimen in order to increase its effects:

Comparators

The active and concomitant comparators were:

Further details about the participants, interventions and comparators are provided in the Characteristics of included studies table.

Outcome measures

The choice of outcome measures varied considerably among trials and this made it impossible to combine results from the majority of individual trials. Only the outcomes reported at endpoints (at the end of or shortly after the end of the interventions) were used in the analysis due to the assumption that at that time the maximum benefits could be expected to have been gained. One trial (Jeyaseelan 2002) reported all its outcomes in medians and ranges and was therefore not included in the analysis of data.

Excluded studies

We excluded 103 reports of 84 trials and reasons for their exclusion are given in the Characteristics of excluded studies table. Most trials were excluded either because the interventions or the comparators were not relevant. For example, Millard and colleagues (Millard 2004) administered PFMT via a two-page written instruction sheet; another trial (Berghmans 2000) added bladder training, another active treatment, to PFMT as a component of lower urinary tract exercise (i.e. the exercise did not contain 'pure' PFMT and this combination was added to another active treatment) while Fitzgerald and colleagues (BE-DRI 2008) used behavioural therapy which included PFMT and timed voiding: the latter is an active treatment on its own (Ostaszkiewicz 2004).

Risk of bias in included studies

Figure 2 and Figure 3 summarise the risk of bias of the included trials. Four trials (Chen 2008; Jeyaseelan 2002; Jin 2012; Wise 1993) were published as conference abstracts and it was therefore difficult to assess the risk of bias with most domains being assessed as 'unclear' risk.

Figure 2.

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

Figure 3.

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

Allocation

Sequence generation

Three trials (Ghoniem 2005; Kim 2011; Richter 2010) provided sufficient details about the methods used in random sequence generation to be sure this was genuine and adequate. Therefore, we considered these trials to be at low risk. For the remaining trials, the risk of bias was unclear because they did not provide enough details about the methods used in sequence generation.

Allocation concealment

Two trials (Ghoniem 2005; Richter 2010) gave enough details to be sure there was adequate allocation concealment and we thus considered them to be at low risk. Other trials did not give clear and sufficient information about allocation concealment and thus the risk of bias was unclear.

Blinding

It was decided that given the nature of an intervention such as PFMT, blinding of women as well as therapists was not practicable. Though one trial (Ghoniem 2005) attempted this by blinding the participants (gave 'sham' PFMT to one of the treatment groups), the adequacy and genuineness of such a blinding process was unclear, thus we categorised this trial as unclear with regard to performance bias. We categorised the remaining trials as being at high risk with respect to performance bias.

In the domain of detection bias, only one trial (Richter 2010) clearly stated that outcome assessors were blinded and we therefore categorised it as being at low risk. The remaining trials did not provide sufficient or any information about outcome assessment and we thus categorised them as unclear.

Incomplete outcome data

Description of dropout and withdrawal

Four trials (Chen 2008; Hofbauer 1990; Jeyaseelan 2002; Jin 2012) did not clearly state whether there was loss to follow-up or not, though in three of these trials (Hofbauer 1990; Jeyaseelan 2002; Jin 2012) it appeared there were no dropouts. In the remaining trials, the proportion of losses to follow-up for all the treatment groups was as follows:

One trial (Wyman 1998) did not report the number of withdrawals by treatment group. In another trial (Kim 2011), there were no dropouts in the experimental and control groups. In two trials (Burgio 2010a; Wise 1993) the proportion of losses to follow-up was higher in the PFMT plus the active treatment group than the active treatment group while in another two trials (Ishiko 2000; Richter 2010) more women dropped out in the active treatment group than the PFMT plus the active treatment group. In the remaining trial (Ghoniem 2005) the proportion of dropouts did not differ significantly between the experimental and the control groups.

Analysis by full intention-to-treat (ITT) principle

Trials were required to retain and analyse participants in the group to which they were randomly assigned. Only three trials (Burgio 2010a; Ghoniem 2005; Richter 2010) clearly reported that the primary analysis was by intention-to-treat. However, it was difficult to ascertain if any of these trials actually met the above criterion for intention-to-treat analysis.

We therefore categorised trials as being at low risk of bias if the proportion of loss to follow-up was 10% or less and there was no evidence of differential loss to follow-up between the comparison groups of interest. In this regard, we rated one trial (Kim 2011) as being at low risk; we categorised two trials (Richter 2010; Wise 1993) as being at high risk while the remaining trials were unclear.

Selective reporting

It was difficult to assess whether the included trials selectively reported their outcomes or not as the protocols for these trials were not available for review. In some of the trials, there was incomplete data reporting with data not made available for one or more of the outcomes specified in the methods section. Therefore, we rated all the trials as unclear for this domain of risk of bias.

Other potential sources of bias

Ethical approval

In four trials (Hofbauer 1990; Jeyaseelan 2002; Kim 2011; Wise 1993), it was neither stated that ethical approval was obtained nor that informed consent was sought from participants.

Source of funding or financial assistance

Two trials (Burgio 2010a; Wyman 1998) received funding or support from public sources; another two were funded either by pharmaceutical companies (Ghoniem 2005) or a private organisation (Richter 2010). Two trials (Chen 2008; Jin 2012) stated that no funding or financial assistance was received. The remaining trials (Hofbauer 1990; Ishiko 2000; Jeyaseelan 2002; Kim 2011; Wise 1993) did not give any report on their source of funding or financial support.

Conflict of interest

Three trials (Burgio 2010a; Ghoniem 2005; Richter 2010) clearly made conflict of interest statements in which some of the authors had financial and/other relationships with some pharmaceutical companies; in one of them (Ghoniem 2005) some of the authors had financial interests and/or other relationships with one of the organisations that supported the trial. One trial (Kim 2011) stated that the authors had no conflict of interest. The remaining trials (Chen 2008; Hofbauer 1990; Ishiko 2000; Jeyaseelan 2002; Jin 2012; Wise 1993; Wyman 1998) did not make any statement with respect to their conflict of interest.

Effects of interventions

See: Summary of findings for the main comparison PFMT added to vaginal cones versus vaginal cones alone for urinary incontinence in women; Summary of findings 2 PFMT added to lifestyle intervention versus lifestyle intervention alone for urinary incontinence in women; Summary of findings 3 PFMT added to bladder training versus bladder training alone for urinary incontinence in women; Summary of findings 4 PFMT added to electrical stimulation versus electrical stimulation alone (excluding implanted electrodes) for urinary incontinence in women; Summary of findings 5 PFMT added to magnetic stimulation versus magnetic stimulation alone for urinary incontinence in women; Summary of findings 6 PFMT added to continence pessary versus continence pessary alone for urinary incontinence in women; Summary of findings 7 PFMT added to drug therapy versus drug therapy alone for urinary incontinence in women; Summary of findings 8 PFMT prior to surgical intervention versus surgical intervention alone for urinary incontinence in women; Summary of findings 9 PFMT added to HSGS versus HSGS alone for urinary incontinence in women

The 11 included trials compared PFMT added to another active treatment (494 women) with the same active treatment alone (490 women). Seven trials reported data on at least one or more of the pre-specified primary outcomes while eight trials contained data on at least one or more of the pre-specified secondary outcomes. None of the trials reported any data on socio-economic outcomes.

The following comparisons were addressed:

A Physical interventions

1. PFMT added to vaginal cones versus vaginal cones alone

One small trial (Wise 1993) compared the effects of a combined PFMT and vaginal cones treatment with vaginal cones treatment alone for women with SUI.

Secondary outcome measures

Number of women cured or improved (objective assessment)

A number of outcomes were reported but only one contained usable data, that is the number of women cured or improved on pad testing (objective assessment of cure or improvement). There were no statistically significant differences in the estimated size of treatment effect between the two intervention groups at endpoint (RR 1.27, 95% CI 0.94 to1.71, Analysis 1.1).

B Behavioural interventions

2. PFMT added to lifestyle intervention (e.g. weight reduction) versus lifestyle intervention alone (lifestyle intervention must be structured or supervised)

None of the trials addressed this comparison.

PFMT added to bladder training versus bladder training alone (bladder training must include scheduled voiding regimen)

For this comparison only one trial (Wyman 1998) with 204 participants contributed data. The trial compared the effects of interventions in women with SUI, UUI or MUI.

Primary outcome measures
Number of women 'cured' or 'improved' (as reported by the women)

Cure rate was assessed immediately after and at three months after treatment using a standardised diary and cure was defined as complete cessation (100% reduction) of incontinence. Immediately after treatment, women who received combined PFMT and bladder training were more likely to be cured than those who received bladder training alone (19/61 versus 12/67), but this difference was not statistically significant (RR 1.74, 95% CI 0.92 to 3.28; Analysis 3.1). At three months after treatment, there was also no statistically significant difference in the estimated size of treatment effect between the two intervention groups (16/59 versus 10/63; RR 1.71, 95% CI 0.84 to 3.46; Analysis 3.1).

Cure or improvement was defined as the proportion of women who had 50% or greater reduction in incontinence episodes on a standardised diary. More women who received a combination of PFMT and bladder training reported cure or improvement immediately after treatment compared to those who were treated with bladder training alone (43/61 versus 35/67; RR 1.35, 95% CI 1.02 to 1.79; Analysis 3.2.1) but there was no statistically significant difference between the two intervention groups at three months after intervention (35/59 versus 28/61; RR 1.29, 95% CI 0.92 to 1.82; Analysis 3.2.2).

Symptom and condition-specific quality of life

The impact of urinary incontinence on quality of life was assessed by two validated scales: the Incontinence Impact Questionnaire-Revised (IIQ-R) and the Urogenital Distress Inventory (UDI) scale. Both instruments have established validity and reliability for assessing the impact of urinary incontinence on the quality of life of women (Shumaker 1994). On these scales, lower scores imply lower impact of incontinence on quality of life and vice versa. Assessment was carried out immediately and at three months after treatment.

Data analysis indicated that immediately after treatment, the addition of PFMT to bladder training resulted in statistically significantly lower (better) impact on the quality of life than with bladder training alone on both scales (IIQ-R: MD -25.50, 95% CI -49.95 to -1.05; Analysis 3.3.1, UDI: MD -31.10, 95% CI -48.94 to -13.26; Analysis 3.4.1). However, this difference did not persist at three months after treatment on both scales (IIQ-R: MD -5.90, 95% CI -35.53 to 23.73; Analysis 3.3.2, UDI: MD -18.90, 95% CI -37.92 to 0.12; Analysis 3.4.2)

Patient global impression of improvement

The instrument used in making this assessment was not specified. Therefore, this outcome was not used for this comparison. However, the pattern of more women improved immediately after treatment, but not three months later, was repeated and details are available in Analysis 3.5.

Secondary outcome measures

Frequency of incontinence episodes per week was assessed from the records in a standardised diary immediately after the intervention (Analysis 3.6). While women had fewer episodes of incontinence in the combined treatment group compared with bladder training alone group (6.8 versus 10.6), this result was not statistically significant (MD -3.80, 95% CI -8.51 to 0.91, Analysis 3.6).

Other outcome measures
Patient satisfaction with treatment outcome

The instrument used in assessing the level of satisfaction of the women with treatment outcome was not reported. However, the pattern of more women improved immediately after treatment, but not three months later, was repeated and details are available in Analysis 3.7).

Number of women requiring further treatment (relapse)

After completion of the 12-week treatment, women were followed up for approximately three years. A similar number of women had sought further treatment such as surgical intervention or drug therapy among those who received PFMT in combination with bladder training: 18/48 (38%) and the control (bladder training alone): 19/48 (40%). No statistically significant difference was found in the estimated size of treatment effect (RR 0.95, 95% CI 0.57 to 1.57; Analysis 3.8)

C Electrical/magnetic interventions

PFMT added to electrical stimulation versus electrical stimulation alone (excluding implanted electrodes)

Two very small trials (Hofbauer 1990 (N = 43); Jeyaseelan 2002 (N = 19)) investigated the effects of this comparison in women with SUI. However, Jeyaseelan 2002 provided no useable data, thus the results could not be pooled and no meta-analysis could be performed.

Hofbauer 1990 reported the following outcomes of interest:

Primary outcome measures
Number of women 'cured' or 'improved' (as reported by the women)

Cure was self reported by the women and was defined as the proportion of women who became continent (free of symptoms of urinary incontinence) at a specified point after treatment onset. The trial was too small to detect statistically significant differences in cure rates between women who received PFMT added to electrical stimulation and those who were given electrical stimulation alone (3/11 versus 1/11; RR 3.00, 95% CI 0.37 to 24.58; Analysis 4.1).

Cure or improvement was also self reported but the success threshold was not defined. Again, there were no statistically significant difference in the estimated size of treatment effect between the two intervention groups (4/11 versus 2/11; RR 2.00, 95% CI 0.46 to 8.76; Analysis 4.2).

Jeyaseelan 2002 reported the following outcomes of interest:

Primary outcome measures

Condition-specific quality of life

Condition-specific quality of life was assessed at endpoint using two scales: Incontinence Impact Questionnaire (IIQ) and Urogenital Distress Inventory (UDI). Further details about these tools as well as the interpretation of scores were not given and data were reported in medians and ranges. Women who received PFMT added to electrical stimulation had lower (better) median scores than those who received electrical stimulation alone on both instruments: IIQ: -27 (-63 to 0) versus 7 (-50 to 150); UDI: -32 (-50 to 18) versus -28 (-86 to 22).

Secondary outcome measures

Objective assessment of improvement on pad test

Women who received PFMT added to electrical stimulation had lower median pad weights compared to those who received electrical stimulation alone: -53 (-77 to -23) versus 39 (-39 to 29), implying less urine loss.

Frequency of incontinence episodes

Details about how this outcome was measured were not reported. However, women who received a combination of PFMT and electrical stimulation had fewer median episodes of urine leakage compared to those who were treated with electrical stimulation alone: -58 (-100 to -50) versus -36 (-58 to 166).

PFMT added to magnetic stimulation versus magnetic stimulation alone

This comparison was not investigated by any of the included trials.

D Mechanical interventions

PFMT added to pessaries versus pessaries alone

Only one trial (Richter 2010) with 446 participants reported a number of outcomes on the effects adding PFMT to continence pessary versus continence pessary alone for women with SUI.

Primary outcome measures
Number of women cured or improved (as reported by the women)

Cure or improvement rate was assessed using the seven-day bladder diary and success (improvement) was defined as the proportion of women who had 75% or greater reduction in frequency of incontinence episodes per week. Assessment was carried out at three, six and 12 months after the start of treatment (but data were only available at three and 12 months). The result indicated that there were no statistically significant differences in cure or improvement rates between women who received PFMT added to continence pessaries and those who were treated with pessaries alone either at six months (80/132 versus 69/110; RR 0.97, 95% CI 0.79 to 1.18, Analysis 6.1.1) or at 12 months (52/111 versus 51/96; RR 0.88, 95% CI 0.67 to 1.16, Analysis 6.1.2) after the start of treatment.

Symptom and condition-specific quality of life

This outcome was assessed at three and 12 months after the onset of intervention. The instrument used was the Urogenital Distress Inventory stress incontinence sub-scale of the Pelvic Floor Distress Inventory, a validated tool that measures the impact of pelvic floor disorders on the quality of life of women (Barber 2001). On this scale success was defined as the proportion of women without 'bothersome' stress incontinence symptoms. For more details see the Characteristics of included studies table. There were no statistically significant differences in the estimated size of treatment effect between the two intervention groups either at three months (RR 1.12, 95% CI 0.86 to 1.47, Analysis 6.2.1) or at 12 months (RR 0.81, 95% CI 0.62 to 1.08, Analysis 6.2.2) post-randomisation.

Patient global impression of improvement

This outcome was assessed at three, six and 12 months post-randomisation using the validated Patient Global Impression of Improvement (PGI-I) Questionnaire. The validity and reliability of this instrument have been established by Yalcin and colleague (Yalcin 2003). Success was defined as the proportion of women with a response of 'much better' or 'very much better' on this scale. There were no statistically significant differences in the women's global impression of improvement between the two intervention groups at any of the endpoints: three months (RR 1.13, 95% CI 0.91 to 1.41, Analysis 6.3.1), six months (RR 1.00, 95% CI 0.78 to 1.30, Analysis 6.3.2) or 12 months (RR 0.90, 95% CI 0.67 to 1.21, Analysis 6.3.3).

Other outcome measure
Patient satisfaction with treatment outcome

This was assessed at three, six and 12 months after the start of treatment using the Patient Satisfaction Question which has been found to be valid and reliable in assessing the extent to which women were satisfied with treatment (Burgio 2006). Success criteria were not reported. Analysis of data showed that there were no statistically significant differences in satisfaction between women who were treated with PFMT added to continence pessary and those who received continence pessary alone with approximately equal proportions of women in each treatment group reporting the same level of satisfaction at each time point: three months (118/132 versus 94/110; RR 1.05, 95% CI 0.95 to 1.15, Analysis 6.4.1); six months (104/123 versus 87/102; RR 0.99, 95% CI 0.89 to 1.11, Analysis 6.4.2); and 12 months (81/111 versus 75/96; RR 0.93, 95% CI 0.80 to 1.09, Analysis 6.4.3).

E Drug interventions

Each drug was tested only in single trials.

Duloxetine

One trial (Ghoniem 2005) with 201 participants reported a number of outcomes on the benefits of adding PFMT to duloxetine therapy for women with SUI. The trial was too small to assess differences in outcomes reliably, and the confidence intervals were wide.

Primary outcome measures
Number of women cured or improved (as reported by the women)

Cure or improvement was assessed from the paper diaries completed by the women at the endpoint of treatment. Success was defined as the proportion of women who had 50% or greater reduction in the frequency of incontinence episodes per week. There were no statistically significant differences in the estimated size of treatment effect between women who were treated with PFMT added to duloxetine and those who received duloxetine alone (RR 1.09, 95% CI 0.77 to 1.53, Analysis 7.2.1).

Symptom and condition-specific quality of life

This outcome was assessed at the endpoint of treatment using Incontinence Quality of Life (I-QoL) Questionnaire. The validity of this instrument has been established by Patrick and colleagues (Patrick 1999). Scores were assigned to different domains of the questionnaire and mean (SD) scores calculated. Higher scores mean less symptom impact (better) on the quality of life. The result indicated than there were no statistically significant differences in this outcome between the two intervention groups (MD 5.84, 95% CI -2.08 to 13.76, Analysis 7.3.1).

Patient global impression of improvement

Patient global impression of improvement was determined within the first three months after randomisation using the validated Patient Global Impression of Improvement (PGI-I) Questionnaire (Yalcin 2003). Success was defined as the number of women with a PGI-I score in one of the three 'better' categories, that is 'very much better', 'much better' or 'a little better'. The estimated size of treatment effect was not statistically significant between the women who received a combined PFMT and duloxetine and those who received duloxetine alone (RR 1.31, 95% CI 0.96 to 1.78, Analysis 7.4.1).

Secondary outcome measures
Frequency of incontinence episodes per week

This outcome was determined in first three months after randomisation using paper diaries completed by the women. No statistically significant differences in outcome were detected between the two intervention groups (MD 0.31, 95% CI -3.55 to 4.17, Analysis 7.5.1).

Number of continence pads used per week

This outcome was computed for each intervention group at the endpoint of treatment. There were no statistically significant differences in number of continence pads used between women who received a combination of PFMT and duloxetine and those who were given duloxetine alone (MD 0.61, 95% CI -2.18 to 3.40, Analysis 7.9).

Oxybutynin

One small trial (Burgio 2010a) investigated the effects of adding PFMT to oxybutynin treatment for women with urgency predominant urinary incontinence. The following outcomes were reported and contributed data to the comparison.

Primary outcome measures
Patient global impression of improvement

One trial addressed this outcome (Burgio 2010a) within the first three months after randomisation using the validated Patient Global Impression of Improvement (PGI-I) Questionnaire (Burgio 2006; Yalcin 2003). Success was defined as the proportion of women who felt 'much better' at endpoint. Analysis of data showed that there was no statistically significant difference in the estimated size of treatment effect between the two intervention groups (RR 0.86, 95% CI 0.68 to 1.09, Analysis 7.4.2).

Secondary outcome measures
Frequency of incontinence episodes per week

This outcome was measured within the first three months and at 12 months post-randomisation using the seven-day bladder diary. Women were more likely to be incontinent with PFMT in combination with oxybutynin versus those who were treated with oxybutynin alone, but this did not reach statistical significance either within the first three months (MD 0.40, 95% CI -2.52 to 3.32; Analysis 7.5.2) or at 12 months (MD 2.80, 95% CI -2.19 to 7.79; Analysis 7.6.1) post-randomisation.

Frequency of micturitions per 24 hours

Although women emptied their bladders more often in the combined treatment group, this difference did not reach statistical significance at the end of treatment (MD 0.20, 95% CI -1.11 to 1.51, Analysis 7.7.1).

Volume of urine per micturition

For this outcome, higher volumes of urine per void means better treatment effect. Women tended to have higher volumes on the drug alone, but this did not differ significantly between the two intervention groups when subjected to statistical analysis and the difference was only 16 ml (MD -16.30, 95% CI -73.77 to 41.17, Analysis 7.8.1).

Other outcome measures
Patient satisfaction with treatment outcome

The validated Patient Satisfaction Questionnaire (Burgio 2006) was used to assess the level of satisfaction of each woman with treatment outcome at endpoint. The number of women who were 'completely satisfied' with treatment outcome was determined. Analysis of data showed that although more women were satisfied with the drug alone, there were no statistically significant differences in the estimated size of treatment effect between the two intervention groups (RR 0.89, 95% CI 0.70 to 1.14, Analysis 7.11.1).

Solifenacin

One trial (Jin 2012) contributed data towards analysis of the effects of adding PFMT to solifenacin treatment for women with over-active bladder. Only one of the reported outcomes had usable data, that is treatment adverse effects, a secondary outcome measure (Analysis 7.10). Adverse effects were assessed with respect to the side effects of solifenacin, a treatment taken by both intervention groups. No statistically significant differences were found in adverse effects due to treatment between women who were treated with combined PFMT and solifenacin treatment and those who received solifenacin treatment alone (RR 0.84, 95% CI 0.45 to 1.60, Analysis 7.10.1).

Clenbuterol

One small trial (Ishiko 2000) compared the effects of combined PFMT and clenbuterol treatment with clenbuterol treatment alone for women with SUI.

Primary outcome measure
Number of women cured (as reported by the women)

Cure rate was defined as the proportion of women who reported 100% reduction in symptoms of urinary incontinence at the end of treatment. There were no statistically significant differences in the per cent of women with self reported cure between the two intervention groups (RR 1.16, 95% CI 0.83 to 1.63, Analysis 7.1.1).

Other outcome measures
Women's satisfaction with treatment outcome

The scale used in measuring this outcome was not specified. The trial was too small to identify significant differences in the number of women who were satisfied with either treatment (Analysis 7.11.2).

Other drugs (unspecified)

One very small trial (Chen 2008) tested the effects of adding PFMT to an unspecified drug therapy for women with overactive bladder.

Other outcome measures

Treatment benefits

Treatment benefits were assessed using the Benefit Questionnaire (further detail was not reported). More women reported that they benefited from combined treatment in the intervention group compared to the control group treated with the drug alone: 11/15, 73% versus 4/14, 29%. This result was statistically significant (RR 2.57, 95% CI 1.06 to 6.20, Analysis 7.12.1).

F Surgical interventions

PFMT prior to surgical intervention (e.g. tension-free vaginal tape (TVT)) versus surgical intervention alone

This comparison was not tested by any of the included trials.

G Other interventions

PFMT + heat and steam generating sheet versus heat and steam generating sheet alone

This comparison was tested by one trial (Kim 2011) in women with UUI or MUI. The trialists hypothesised that the heat and steam generating sheet (HSGS) would reduce incontinence episodes by heating the abdominal and lower back which in turn might result in positive effects on renal function, such as suppression of the activity of renal sympathetic nerves and promotion of bladder emptying. Details about the heat and steam generating sheet are available in the Characteristics of included studies table.

Primary outcome measures

Number of women cured (as reported by the women)

Cure was assessed by interview and success was defined as the proportion of women with complete cessation of urine loss episodes at the end of treatment. Analysis of data indicated that more women were cured in the intervention group (PFMT added to HSGS) compared to the comparison group (HSGS alone): 19/37, 51% versus 8/37, 22%. This result was statistically significant (RR 2.38, 95% CI 1.19 to 4.73, Analysis 9.1.1).

Discussion

This is the first Cochrane systematic review on the effects of adding pelvic floor muscle training (PFMT) to other active treatment versus the same active treatment alone for urinary incontinence in women and should be considered in the context of the other Cochrane reviews on pelvic floor muscle training (Boyle 2012; Dumoulin 2010; Hay-Smith 2011; Herbison 2013; Herderschee 2011). The review examines whether the addition of PFMT to other active treatment is more beneficial than the same active treatment alone for the treatment of women with urinary incontinence.

Summary of main results

Is PFMT added to another active treatment more effective than the same active treatment alone?

This question was addressed by 11 trials (Burgio 2010a; Chen 2008; Ghoniem 2005;Hofbauer 1990; Ishiko 2000; Jeyaseelan 2002; Jin 2012; Kim 2011; Richter 2010; Wise 1993; Wyman 1998). We classified the primary and secondary outcomes, as defined earlier, as 'critical', 'important' or 'not important' for decision making from the woman's perspective. Seven outcomes from the comparisons were considered to be 'critical'; we applied the GRADE approach to determine the quality of evidence associated with these outcomes. The results are presented in the 'Summary of findings' tables.

Vaginal cones

The additional effects of adding PFMT to vaginal cones was examined by Wise and colleagues (Wise 1993) in women with stress urinary incontinence (SUI). Although more women were either cured or improved with combined PFMT and vaginal cones than the control on objective assessment of urine leakage, this difference was not statistically significant (Analysis 1.1). The 12-week PFMT was potentially too short and not adequately described to decide whether the exercise dose was theoretically sufficient. Adherence was not reported. Further, we considered the quality of the evidence (effect estimate) for the objective measure of urine leakage (pad test) to be very low when adopting the GRADE approach and none of the other outcomes which we considered critical for decision making were reported (Summary of findings for the main comparison).

Bladder training

The addition of PFMT to bladder training (Wyman 1998) in women with SUI, urgency urinary incontinence (UUI) or mixed urinary incontinence (MUI) did not result in a statistically significant difference in the number of women cured either immediately after treatment (three months after randomisation) (Analysis 3.1.1) or at three months after intervention (six months after randomisation) (Analysis 3.1.2). There was no statistically significant benefit from adding PFMT to bladder training on the quality of life at three months after intervention (Analysis 3.4.2). A similar number of women (around 40%) required further treatments at approximately three years after treatment in those who received PFMT added to bladder training versus bladder training alone. The description of the 12-week PFMT programme suggested it could theoretically increase strength, endurance and co-ordination although it was probably of insufficient duration for muscle strengthening. Training adherence was not reported. We judged the quality of evidence for the three reported critical outcomes to be very low when adopting the GRADE approach and none of the other four outcomes which we considered critical for decision making from the patient's perspective were reported (Summary of findings 3).

Electrical stimulation

Adding PFMT to electrical stimulation as reported by Hofbauer and colleagues in women with SUI (Hofbauer 1990) did not result in a statistically significant difference in the cure rates. The content of PFMT was not described and at six weeks duration was probably insufficient to maximise any possible training effect. Adherence was not reported. Further, we considered the quality of the evidence (effect estimate) for the number of women cured to be very low when adopting the GRADE approach and none of the other outcomes which we considered critical for decision making were reported (Summary of findings 4).

Continence pessary

The addition of PFMT to a continence pessary (Richter 2010) did not result in statistically significant benefits in women with SUI in terms of the number of women cured or improved (Analysis 6.1.2) and the impact of urinary incontinence on the quality of life (Analysis 6.2.2) at 12 months after treatment. At eight weeks duration the incompletely described PFMT programme (that included stress and urgency strategies) was probably insufficient to maximise any possible training effect and adherence was not reported. We considered the quality of the evidence (effect estimate) for the number of women cured or improved (subjective) at 12 months and condition-specific quality of life at 12 months to be moderate when adopting the GRADE approach and none of the other outcomes which we considered critical for decision making were reported (Summary of findings 6).

Drug treatment

The benefits of adding PFMT to drug treatment did not show any statistically significant difference between the experimental (PFMT plus drug) and the control (drug alone) groups. There was no statistically significant difference in the number of women cured when adding PFMT to clenbuterol (Ishiko 2000). The 12-week PFMT programme was potentially too short and not adequately described to decide whether the exercise dose was theoretically sufficient, and adherence was not reported. Similarly, the addition of PFMT to duloxetine (Ghoniem 2005) did not result in a statistically significant better outcome in the domain of symptom and condition-specific quality of life. The description of the 12-week PFMT programme suggested it could theoretically increase strength, endurance and co-ordination although was probably of insufficient duration for muscle strengthening. In addition, training adherence was not reported. Further, the quality of the evidence for condition-specific quality of life on the I-QoL Questionnaire was low; we also considered the quality of evidence for the number of women cured and the number of women reporting adverse events to be very low when adopting the GRADE methodology and none of the other outcomes which we considered critical for decision making were reported (Summary of findings 7).

Other active treatment

The additional benefits of PFMT over and above a heat and steam generating sheet (HSGS) were investigated by Kim and colleagues (Kim 2011). More women were cured in the combined PFMT and HSGS group compared to the HSGS alone group and this result was statistically significant (Analysis 9.1.1). The description of the 12-week PFMT programme suggested it could theoretically increase strength and endurance, but this duration was probably insufficient for muscle strengthening. Training adherence was not reported. We considered the effect estimate to be of moderate quality when adopting the GRADE approach (Summary of findings 9).

Problems with pelvic floor muscle training regimens

The absence of additional effects of PFMT over and above the active treatment alone in most of the included trials might have been due to a number of factors. Of particular concern was the difficulty in evaluating the potential effectiveness of the PFMT intervention and/or the obvious inadequacy of the exercise dose offered to women. All but three of the trials gave insufficient detail of the PFMT programme, or the PFMT programme was too short for muscle strengthening, to think that the exercise dose could be sufficient for treatment effect. Wyman, Kim and Ghoniem described PFMT programmes that might theoretically strengthen pelvic floor muscles (although in all three trials the treatment duration of 12 weeks was probably too short to establish muscle hypertrophy). It is also worth considering whether participants might have regarded the addition of PFMT as an additional treatment burden and thus either carried it out suboptimally or abandoned it altogether. None of the trials reported training adherence, which further compromises the ability to appraise the potential effectiveness of the PFMT intervention.

However, a likely explanation is that all the trials were too small (and hence underpowered) to detect statistically significant differences between the interventions.

Overall completeness and applicability of evidence

Three of the pre-specified objectives (interventions) were not investigated by any of the included trials (PFMT/lifestyle intervention versus lifestyle intervention alone, PFMT/magnetic stimulation versus magnetic stimulation alone and PFMT prior to surgical intervention versus surgical intervention alone). The remaining pre-specified comparisons were each addressed by single trials except PFMT plus electrical stimulation versus electrical stimulation which was investigated by two trials (Hofbauer 1990; Jeyaseelan 2002).

Jeyaseelan 2002 reported results using median and range and therefore the results could not be pooled and there was no meta-analysis. Therefore it was not possible to improve the power for any of the comparisons. Some of the trialists used combination of interventions with no regard to the types of urinary incontinence and this might have influenced the results of the reported outcomes. For example, combining PFMT and anticholinergics (without the appropriate 'dose' or 'doses') for women with UUI or urgency predominant urinary incontinence when PFMT has been shown to work better for women with SUI or MUI (Dumoulin 2010).

None of the included trials reported any data on socio-economic implications of the intervention, while only one trial (Wyman 1998) reported data in a usable form on long-term follow-up. Also, only one trial (Jin 2012), evaluating PFMT when added to drug therapy, provided information about treatment adverse events with respect to the side effects of the drug therapy; none of the included trials reported data on adverse events associated with the PFMT regimen, thereby making it very difficult for us to evaluate the safety of PFMT. Treatment adherence which might impact on other outcome measures was not reported or analysed in a usable form by any of the included trials.

Quality of the evidence

Trial quality and methodological assessment

Methodological assessment plays a crucial role in determining the quality of the estimated size of treatment effects of any intervention. In this review, we assessed the methodological flaws of the included trials using the reports of the trials. Therefore, our judgement of methodological quality and hence the quality of effect estimates was influenced by the quality of reporting.

Four trials (Chen 2008; Jeyaseelan 2002; Jin 2012; Wise 1993) were published as conference abstracts with few details on study designs, methods or data, thereby making it very difficult to assess their methodological quality. Of the 11 included trials, only two (Ghoniem 2005; Richter 2010) gave detailed descriptions of the randomisation process for the review authors to be sure there was adequate sequence generation and allocation concealment. Thus we judged them to be at low risk with respect to selection bias. A key intervention in this review is that given the nature of PFMT, it was difficult to blind the participants or treatment providers to group allocation (performance bias). With regard to detection bias, outcome assessors were adequately blinded in only one of the included trials (Richter 2010).

In the domain of attrition bias, the rates of withdrawals and losses to follow-up were high in some of the included trials but with small differences in rates within treatment groups. In terms of size, most of the included trials were small, meaning that a high attrition rate would result in under-powering of the trials and hence the occurrence of type II error (false negative results). A common problem with most of the included trials was incomplete reporting particularly with respect to the trial methods and data. Thus we assessed some domains of the risk of bias as 'unclear' due to incomplete reporting of methods.

In this review, the estimated sizes of treatment effects were generally small and therefore none of them were upgraded. However, the quality of the body of evidence was downgraded if we considered the randomisation process (sequence generation and allocation concealment) of the trial to be inadequate and/or if the effect estimate crossed the line of 'no effect' by 25% or 50% on either side (that is, effect estimate with a wide confidence interval).

Potential biases in the review process

We searched all the important databases and imposed no language restriction in the course of the search. However, we were mindful of the fact that these databases might not have contained all the potentially eligible trials.

Agreements and disagreements with other studies or reviews

We identified one systematic review in the 5th edition of the International Consultation on Incontinence (ICI) (Moore 2013) which addressed the effects of adding PFMT to other active treatment versus the same active treatment alone. Moore and colleagues included eight trials (Burgio 2008; Burgio 2010a; Ghoniem 2005; Hofbauer 1990; Ishiko 2000; Wilson 1998; Wise 1993; Wyman 1998), six of which were also included in this review (Burgio 2010a; Ghoniem 2005; Hofbauer 1990; Ishiko 2000; Wise 1993; Wyman 1998). One trial (Wilson 1998) was not included in this review as the participants were postpartum women; whereas the other trial (Burgio 2008) was excluded because the behavioural intervention consisted of PFMT in addition to delayed voiding to increase voiding intervals and individualised fluid management which we considered as active treatments on their own. Additionally, Moore's review did not include five trials (Chen 2008; Jeyaseelan 2002; Jin 2012; Kim 2011; Richter 2010) which were included in this review. Overall, the findings of the review conducted by Moore and colleagues (Moore 2013) are in agreement with those of this review.

Authors' conclusions

Implications for practice

All the identified trials randomised participants with between six (Jeyaseelan 2002) and 150 (Richter 2010) per treatment arm and thus were not powered to detect any significant difference in the primary outcomes of interest in the review. In addition, they all addressed different participants and interventions with disparate and few outcome data. This limited our ability to combine data in a pooled analysis for any of the comparisons. Therefore, our confidence in the estimated size of treatment effects for most (if not all) of the comparisons is uncertain. Based on these findings, this review did not find sufficient evidence as to whether or not there were additional effects of adding pelvic floor muscle training (PFMT) to other active treatment when compared to the same active treatment alone for urinary incontinence in women. These results should be interpreted with caution as most of the comparisons were investigated by single trials which were small and none of the trials in this review were large enough to answer the questions they were designed to answer.

Implications for research

This review has demonstrated that there is insufficient evidence to conclude whether or not adding PFMT to another active treatment is more beneficial either in the short or the long term than the same active treatment alone. This was partly due to either scanty or no trials on the various interventions postulated. Additionally, the methodological quality of some of the included trials fell short of the recommendations and principles set out in the CONSORT statement as illustrated in Figure 2; Figure 3. Moreover, majority of the trials did not report the required information for making decisions and we judged them as 'unclear'.

Therefore, there is a need for more research on the effect of PFMT when added to other treatments. For example, this review could not identify any trial which investigated the additional effects of PFMT over and above common active treatments such as surgical intervention (PFMT prior to surgical intervention versus surgical intervention alone) or structured lifestyle intervention (PFMT added to structured lifestyle intervention versus strictured lifestyle intervention alone).

Furthermore, future research should take into consideration the synergistic effects of combined PFMT and another active treatment in relation to the types of urinary incontinence before combination: for example, PFMT plus duloxetine versus duloxetine for stress urinary incontinence; or PFMT plus anticholinergic versus anticholinergic for urgency urinary incontinence or overactive bladder.

In addition, future research should equally focus on quality of life, socio-economic implications, long-term effects as well as adverse events associated with combining PFMT with other active treatments. Above all, future research should be conducted in accordance with the recommendations and principles outlined in the CONSORT statement for improving the reporting of trials.

Acknowledgements

The review authors would like to acknowledge the team of the Cochrane Incontinence Group for their help.

Data and analyses

Download statistical data

Comparison 1. PFMT added to vaginal cones versus vaginal cones alone
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Number of women cured or improved (objective assessment)1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
Analysis 1.1.

Comparison 1 PFMT added to vaginal cones versus vaginal cones alone, Outcome 1 Number of women cured or improved (objective assessment).

Comparison 3. PFMT added to bladder training versus bladder training alone
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Number of women cured1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
1.1 Immediately after treatment1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
1.2 3 months after treatment1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
2 Number of women cured or improved1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
2.1 Immediately after treatment1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
2.2 3 months after treatment1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
3 Condition-specific quality of life on IIQ-R1 Mean Difference (IV, Fixed, 95% CI)Totals not selected
3.1 Immediately after treatment1 Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]
3.2 3 months after treatment1 Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]
4 Condition-specific quality of life on UDI1 Mean Difference (IV, Fixed, 95% CI)Totals not selected
4.1 Immediately after treatment1 Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]
4.2 3 months after treatment1 Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]
5 Number of women improved using patient global impression of improvement1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
5.1 Immediately after treatment1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
5.2 3 months after treatment1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
6 Incontinence episode per week1 Mean Difference (IV, Fixed, 95% CI)Totals not selected
7 Patient satisfaction with treatment outcome1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
7.1 Immediately after treatment1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
7.2 3 months after treatment1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
8 Number of women requiring further treatment (relapse)1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
Analysis 3.1.

Comparison 3 PFMT added to bladder training versus bladder training alone, Outcome 1 Number of women cured.

Analysis 3.2.

Comparison 3 PFMT added to bladder training versus bladder training alone, Outcome 2 Number of women cured or improved.

Analysis 3.3.

Comparison 3 PFMT added to bladder training versus bladder training alone, Outcome 3 Condition-specific quality of life on IIQ-R.

Analysis 3.4.

Comparison 3 PFMT added to bladder training versus bladder training alone, Outcome 4 Condition-specific quality of life on UDI.

Analysis 3.5.

Comparison 3 PFMT added to bladder training versus bladder training alone, Outcome 5 Number of women improved using patient global impression of improvement.

Analysis 3.6.

Comparison 3 PFMT added to bladder training versus bladder training alone, Outcome 6 Incontinence episode per week.

Analysis 3.7.

Comparison 3 PFMT added to bladder training versus bladder training alone, Outcome 7 Patient satisfaction with treatment outcome.

Analysis 3.8.

Comparison 3 PFMT added to bladder training versus bladder training alone, Outcome 8 Number of women requiring further treatment (relapse).

Comparison 4. PFMT added to electrical stimulation versus electrical stimulation alone (excluding implanted electrodes)
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Number of women cured1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
2 Number of women cured or improved1 Risk Ratio (M-H, Fixed, 95% CI)Subtotals only
Analysis 4.1.

Comparison 4 PFMT added to electrical stimulation versus electrical stimulation alone (excluding implanted electrodes), Outcome 1 Number of women cured.

Analysis 4.2.

Comparison 4 PFMT added to electrical stimulation versus electrical stimulation alone (excluding implanted electrodes), Outcome 2 Number of women cured or improved.

Comparison 6. PFMT added to continence pessary versus continence pessary alone
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Number of women cured or improved1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
1.1 At 3 months1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
1.2 At 12 months1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
2 Condition-specific quality of life on UDI1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
2.1 At 3 months1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
2.2 At 12 months1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
3 Number of women improved using patient global impression of improvement1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
3.1 At 3 months1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
3.2 At 6 months1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
3.3 At 12 months1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
4 Patient satisfaction with treatment outcome1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
4.1 At 3 months1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
4.2 At 6 months1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
4.3 At 12 months1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
Analysis 6.1.

Comparison 6 PFMT added to continence pessary versus continence pessary alone, Outcome 1 Number of women cured or improved.

Analysis 6.2.

Comparison 6 PFMT added to continence pessary versus continence pessary alone, Outcome 2 Condition-specific quality of life on UDI.

Analysis 6.3.

Comparison 6 PFMT added to continence pessary versus continence pessary alone, Outcome 3 Number of women improved using patient global impression of improvement.

Analysis 6.4.

Comparison 6 PFMT added to continence pessary versus continence pessary alone, Outcome 4 Patient satisfaction with treatment outcome.

Comparison 7. PFMT added to drug therapy versus drug therapy alone
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Number of women cured1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
1.1 PFMT + clenbuterol vs clenbuterol1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
2 Number of women cured or improved1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
2.1 PFMT + duloxetine vs duloxetine1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
3 Condition-specific quality of life on I-QoL questionnaire1 Mean Difference (IV, Fixed, 95% CI)Totals not selected
3.1 PFMT + duloxetine vs duloxetine1 Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]
4 Number of women improved on patient global impression of improvement in first 3 months2 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
4.1 PFMT + duloxetine vs duloxetine1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
4.2 PFMT + oxybutynin vs oxybutynin1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
5 Frequency of incontinence episodes per week in first 3 months2 Mean Difference (IV, Fixed, 95% CI)Totals not selected
5.1 PFMT + duloxetine vs duloxetine1 Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]
5.2 PFMT + oxybutynin vs oxybutynin1 Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]
6 Frequency of incontinence episodes per week at 12 months1 Mean Difference (IV, Fixed, 95% CI)Totals not selected
6.1 PFMT + oxybutynin vs oxybutynin1 Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]
7 Frequency of micturitions per 24 hours1 Mean Difference (IV, Fixed, 95% CI)Totals not selected
7.1 PFMT + oxybutynin vs oxybutynin1 Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]
8 Volumes of urine per micturition1 Mean Difference (IV, Fixed, 95% CI)Totals not selected
8.1 PFMT + oxybutynin vs oxybutynin1 Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]
9 Number of continence pads used per week1 Mean Difference (IV, Fixed, 95% CI)Totals not selected
9.1 PFMT + duloxetine vs duloxetine1 Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]
10 Treatment adverse events1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
10.1 PFMT + solifenacin vs solifenacin1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
11 Patient satisfaction with treatment outcome in first 3 months2 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
11.1 PFMT + oxybutynin vs oxybutynin1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
11.2 PFMT + clenbuterol vs clenbuterol1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
12 Treatment benefit1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
12.1 PFMT + ?drug vs ?drug (drug name not reported)1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
Analysis 7.1.

Comparison 7 PFMT added to drug therapy versus drug therapy alone, Outcome 1 Number of women cured.

Analysis 7.2.

Comparison 7 PFMT added to drug therapy versus drug therapy alone, Outcome 2 Number of women cured or improved.

Analysis 7.3.

Comparison 7 PFMT added to drug therapy versus drug therapy alone, Outcome 3 Condition-specific quality of life on I-QoL questionnaire.

Analysis 7.4.

Comparison 7 PFMT added to drug therapy versus drug therapy alone, Outcome 4 Number of women improved on patient global impression of improvement in first 3 months.

Analysis 7.5.

Comparison 7 PFMT added to drug therapy versus drug therapy alone, Outcome 5 Frequency of incontinence episodes per week in first 3 months.

Analysis 7.6.

Comparison 7 PFMT added to drug therapy versus drug therapy alone, Outcome 6 Frequency of incontinence episodes per week at 12 months.

Analysis 7.7.

Comparison 7 PFMT added to drug therapy versus drug therapy alone, Outcome 7 Frequency of micturitions per 24 hours.

Analysis 7.8.

Comparison 7 PFMT added to drug therapy versus drug therapy alone, Outcome 8 Volumes of urine per micturition.

Analysis 7.9.

Comparison 7 PFMT added to drug therapy versus drug therapy alone, Outcome 9 Number of continence pads used per week.

Analysis 7.10.

Comparison 7 PFMT added to drug therapy versus drug therapy alone, Outcome 10 Treatment adverse events.

Analysis 7.11.

Comparison 7 PFMT added to drug therapy versus drug therapy alone, Outcome 11 Patient satisfaction with treatment outcome in first 3 months.

Analysis 7.12.

Comparison 7 PFMT added to drug therapy versus drug therapy alone, Outcome 12 Treatment benefit.

Comparison 9. PFMT added to other treatment versus other treatment alone
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Number of women cured1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
1.1 PFMT + heat and steam generating sheet versus heat and steam generating sheet alone1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
Analysis 9.1.

Comparison 9 PFMT added to other treatment versus other treatment alone, Outcome 1 Number of women cured.

Appendices

Appendix 1. Search strategies

Other electronic searches performed specifically for this review are detailed below.

EMBASE Classic and EMBASE (on OVID SP) covering 1947 to 2013 Week 9.  Date of last search: 7 March 2013. The search strategy is given below.

1. Randomized Controlled Trial/

2. crossover procedure/ or double blind procedure/ or parallel design/ or single blind procedure/

3. Placebo/

4. placebo$.tw,ot.

5. random$.tw,ot.

6. ((singl$ or doubl$ or trebl$ or tripl$) adj25 (blind$ or mask$)).tw,ot.

7. crossover.tw,ot.

8. cross over$.tw,ot.

9. allocat$.tw,ot.

10. trial.ti.

11. parallel design/

12. triple blind procedure/

13. or/1-12

14. exp animals/ or exp invertebrate/ or animal experiment/ or animal model/ or animal tissue/ or animal cell/

15. exp human/ or exp "human tissue, cells or cell components"/

16. 14 and 15

17. 14 not 16

18. 13 not 17

19. pelvic floor muscle training/

20. exp feedback system/

21. kegel*.tw.

22. (pelvi* adj4 (exercis* or train* or muscle*)).tw.

23. PFMT.tw.

24. 19 or 20 or 21 or 22 or 23

25. incontinence/ or mixed incontinence/ or stress incontinence/ or urge incontinence/ or urine incontinence/

26. continence/

27. overactive bladder/

28. micturition disorder/ or lower urinary tract symptom/ or pollakisuria/

29. urinary dysfunction/ or bladder instability/ or detrusor dyssynergia/ or neurogenic bladder/ or urinary urgency/ or urine extravasation/

30. (incontinen$ or continen$).tw.

31. ((bladder or detrusor or vesic$) adj5 (instab$ or stab$ or unstab* or irritab$ or hyperreflexi$ or dys?ynerg$ or dyskinesi$ or irritat$)).tw.

32. (urin$ adj2 leak$).tw.

33. ((bladder or detrusor or vesic$) adj2 (hyper$ or overactiv$)).tw.

34. (bladder$ adj2 (neuropath$ or neurogen* or neurolog$)).tw.

35. (nervous adj pollakisur$).tw.

36. or/25-35

37. 18 and 24 and 36

38. (2011* or 2012* or 2013*).em.

39. 37 and 38

The EMBASE search was limited by entry month to 2011, 2012 and 2013 to cover those years that are not currently included in the EMBASE search that is searched by The Cochrane Collaboration and incorporated into the CENTRAL database.

Key: / = EMTREE term; .tw. = text word search; .ot. = original title (for non-English titles); $ = truncation; adjn = within n words of other word in any word order.

CINAHL on EBSCO Host (covering January 1982 to 5 March 2013). Date of last search: 5 March 2013. The search strategy is given below.

#Query
S39S31 AND S38
S38S32 OR S33 OR S34 OR S35 OR S36 OR S37
S37TI ( PFMT OR PFE ) OR AB ( PFMT OR PFE )
S36TI pelvi* N5 floor* OR AB pelvi* N5 floor*
S35TI pelvi* N5 muscle* OR AB pelvi* N5 muscle*
S34TI kegel* OR AB kegel*
S33(MM "Pelvic Floor Muscles")
S32(MM "Kegel Exercises") OR (MH "Therapeutic Exercise+") OR (MH "Muscle Strengthening+")
S31S23 AND S30
S30S24 OR S25 OR S26 OR S27 OR S28 OR S29
S29TI overactiv* N3 bladder* OR AB overactiv* N3 bladder*
S28TI urin* N3 leak* OR AB urin* N3 leak*
S27TI ( incontinen* OR continen* ) OR AB ( incontinen* OR continen* )
S26MH incontinence
S25MH overactive bladder
S24MH Urinary incontinence+
S23S1 or S2 or S3 or S4 or S5 or S6 or S7 or S8 or S9 or S10 or S11 or S12 or S13 or S14 or S15 or S16 or S17 or S18 or S19 or S20 or S21 or S22
S22TI ( singl* N25 blind* OR singl* N25 mask* OR doubl* N25 blind* or doubl* N25 mask* OR trebl* N25 blind* OR trebl* N25 mask*OR tripl* N25 blind* OR tripl* N25 mask* ) or AB ( singl* N25 blind* OR singl* N25 mask* OR doubl* N25 blind* or doubl* N25 mask* OR trebl* N25 blind* OR trebl* N25 mask*OR tripl* N25 blind* OR tripl* N25 mask* )
S21(MH "Comparative Studies")
S20(MH "Clinical Research+")
S19(MH "Static Group Comparison")
S18(MH "Quantitative Studies")
S17(MH "Crossover Design") or (MH "Solomon Four-Group Design")
S16(MH "Factorial Design")
S15(MH "Community Trials")
S14(MH "Random Sample")
S13TI balance* N2 block* or AB balance* N2 block*
S12TI "latin square" or AB "latin square"
S11TI factorial or AB factorial
S10TI clin* N25 trial* or AB clin* N25 trial*
S9(MH "Study Design")
S8(AB random*) OR (TI random*)
S7(AB placebo*) OR (TI placebo*)
S6(MH "Placebos")
S5PT Clinical Trial
S4(MH "Clinical Trials+")
S3MH (random assignment) OR (crossover design)
S2cross-over
S1crossover

Key: MH = exact CINAHL subject heading; + = exploded CINAHL heading; MM = exact major CINAHL subject heading; N = within n words of the other word, in any order; PT = publication type; AB = abstract; TI = title.

Searching for ongoing trials

ClinicalTrials.gov (date of last search: 30 May 2013). The search terms used are given below.

  • Pelvic training

  • Pelvic exercise

  • Pelvic exercises

WHO ICTRP (date of last search: 3 June 2013). The search terms used are given below.

  • Pelvic floor muscle training

  • Pelvic floor muscle exercise*

  • Pelvic floor exercise*

  • Pelvic exercise*

  • Pelvic training

Key: * indicates truncation

Contributions of authors

Reuben Olugbenga Ayeleke (ROA), E. Jean C Hay-Smith (JHS) and Muhammad Imran Omar (MIO) were responsible for the conception and writing of the protocol. ROA and MIO performed abstract screening, full-text screening, data extraction, 'Risk of bias' assessment and quality of evidence assessment. All review authors contributed in the analysis of data and writing the manuscript of the review.

Declarations of interest

None known.

Sources of support

Internal sources

  • No sources of support supplied

External sources

  • NIHR, UK.

    The Cochrane Incontinence Review Group is funded by NIHR UK

Differences between protocol and review

There was no difference between the protocol and review.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Burgio 2010a

Methods2-arm randomised controlled trial, parallel design
Participants64 women with urgency predominant incontinence
Interventions

A: Drug therapy alone group (n = 32). Individuals in this group received oxybutynin 5 mg daily with dose gradually increased during visits to the maximum level the individual could tolerate (dose range: 5 to 30 mg)

B: Behavioural therapy + drug therapy (n = 32): participants in this group received drug therapy as described above and behavioural therapy. Behavioural therapy included PFMT and urge suppression strategies. PFMT consisted of 3 sessions of 15 exercises daily (total of 45 exercises). During each session, participants were instructed to contract for 10 seconds and relax for another 10 seconds (maximum duration of 10 seconds was achieved on a gradual basis). They were also taught the skills on urge suppression strategies

Outcomes

1. Patient global perception of improvement: this was measured using the Global Perception of Improvement rating. Success was defined as the proportion of participants who felt 'much better' at the end of the treatment

At 8 weeks:

A: 28/31; B: 21/27 

2. Condition-specific quality of life: assessed using the Incontinence Impact Questionnaire and Urogenital Distress Inventory (reported as mean score and SD; details of data not reported)

3. Patient satisfaction with treatment outcome: success was defined as the proportion of participants who were completely satisfied with the treatment outcome. It was assessed using the Patient Satisfaction Questionnaire

At 8 weeks:

A: 27/31; B: 21/27

4. Frequency of incontinence episodes per week: mean (SD) of incontinence episodes frequency was assessed at endpoint using the 7-day bladder diary

At 8 weeks:

A: 2.0 (4.9), n = 30: B: 2.4 (6.2), n = 27

At 12 months:

A: 1.7 (3.9), n = 27; B: 4.5 (11.4), n = 22

5. Frequency of micturition per 24 hours (in mean and SD)

At 8 weeks:

A: 8.2 (1.9), n = 31; B: 8.4 (3.0), n = 27

6. Volumes of urine voided per 24 hours (in mean and SD)

At 8 weeks:

A: 256.7 (86.7), n = 31; B: 240.4 (129.1), n = 27

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskStated as "stratified block randomisation". Exact process not specified
Allocation concealment (selection bias)Unclear riskIt was not stated whether or not the allocations were concealed
Blinding of participants and personnel (performance bias)
All outcomes
High riskBlinding of participants not possible
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskCompleted questionnaires were submitted in sealed envelopes and given to the nurses who administered the intervention. However, it is not specified whether the same or different nurses assessed the outcomes
Incomplete outcome data (attrition bias)
All outcomes
Unclear risk5/64 dropped out of the trial: A 1/32; B 4/32. Reasons not specified
Selective reporting (reporting bias)Unclear riskProtocol not available
Ethical approvalLow riskApproved by the institutional review board
Source of funding or supportLow riskStated (received grants from public institutions)
Conflict of interestUnclear riskSome of the authors had financial and other relationships with some pharmaceutical companies

Chen 2008

Methods2-arm randomised controlled trial, parallel design
Participants29 women with over-active bladder
Interventions

A: Drug alone (n = 14): details of drug including name and dose not stated

B: PFMT + drug (n = 15). PFMT was assisted by perineal surface electromyography and was taught inially. Participants were then instructed to perform 3 sets of PFMT per day, 15 contractions per set, continuously at home for 8 weeks. Drug regimen: as stated above

Outcomes

1. Urgency episodes per 24 hours: this was determined at baseline and endpoint using the 3-day voiding diary and mean percentage change was calculated for the 2 groups (no useable data)

2. Daytime frequency per 24 hours: this was obtained before and after treatment using the 3-day voiding diary and mean percentage change calculated (no useable data)

3. Treatment benefit: this was determined 4 weeks post-treatment using the 'Benefit Questionnaire' and proportion of participants with perceived benefits calculated for each group

A: 4/14; B: 11/15

4. Symptom bothersome: scores were obtained before and after treatment and mean percentage change in bothersome scores was obtained for the 2 groups (no useable data)

5. Quality of life: total scores were calculated for different domains of the quality of life (such as sleeping, concern and coping) before and after treatment. Mean percentage increase was calculated for the 2 groups (no useable data)

NotesDropouts: not reported, only the number of participants who completed the trial was stated
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskProcess involved in randomisation was not reported
Allocation concealment (selection bias)Unclear riskProcess involved in allocation concealment was not stated
Blinding of participants and personnel (performance bias)
All outcomes
High riskBlinding of participants not possible (assumed not done)
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot specified
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNot reported, only the number of participants who completed the trial was stated
Selective reporting (reporting bias)Unclear riskProtocol not available
Ethical approvalLow riskApproved by the Ethics committee
Source of funding or supportLow riskStated "none" according to the authors
Conflict of interestUnclear riskNot declared

Ghoniem 2005

Methods4-arm randomised controlled trial, parallel design
Participants201 women with predominant symptoms of stress urinary incontinence (SUI)
Interventions

A: No active treatment (n = 47). Received placebo plus imitation (sham) PFMT for 12 weeks. Imitation PFMT consisted of initial therapist-supervised instructions on how to train the hip abductors. Participants were then given written instructions and a training log with the recommendation of 3 sets of 10 long and 2 sets of 10 rapid contractions 4 days weekly. However, no instructions were given to the participants to contract the pelvic floor muscles with physical activities associated with urine leakage (skill training)

B. PFMT only (n = 50). Received placebo plus PFMT for 12 weeks. PFMT comprised 30 minutes of initial therapist supervised instructions on how to contract the pelvic floor muscles. The correct type of contraction was confirmed by pelvic examination. Then participants received instructions to perform 3 sets of 10 long (6 to 8 seconds) and 2 sets of 10 rapid (1 to 2 seconds) contractions 4 days weekly (total of 200 contractions per week). At 4 and 8 weeks, participants received 15 minutes of re-instruction and manual feedback and a training log was completed. Finally, skill training was giving by instructing participants to contract the pelvic floor muscles with physical events usually associated with urine loss

C: Duloxetine + sham PFMT (n = 52). This group received duloxetine and sham PFMT. Duloxetine was given at a dose of 40 mg twice daily for 12 weeks. Sham PFMT (as described above)

D: PFMT + duloxetine (n = 52). This is the combined group. Participants in this group received PFMT and duloxetine as described above

For this review comparison D versus C is relevant

Outcomes

1. Incontinence episode frequency (IEF) per week. This was computed from participant completed paper diaries at each visit. Mean (SD) weekly IEF at the endpoint was calculated for each treatment group

A: 18.50 (17.10), n = 44; B: 20.93 (16.26), n = 46; C: 10.96 (8.53), n = 46; D: 11.27 (10.06), n = 44

2. Improvement (IEF responder rate): this was defined as the proportion of participants who had a 50% or greater decrease in IEF with treatment as computed from the paper diaries

A: 11/44; B: 12/46; C: 26/46; D: 27/44

3. Number of continence pads used. Mean (SD) pads per week was calculated for each treatment group at endpoint

A: 10.22 (7.56), n = 44; B: 11.48 (8.36), n = 46; C: 7.23 (5.98), n = 46; D: 7.84 (7.41), n = 44

4. Condition-specific quality of life: this was assessed at endpoint using the Incontinence Quality of Life (I-QoL) score questionnaire and mean (SD) score was obtained for each group

A: 69.34 (20.69), n = 45; B: 68.76 (22.70), n = 49; C: 68.23 (20.87), n = 50; D: 74.07 (19.70), n = 51

5. Patient Global Impression of Improvement (PGI-I): this was defined as the proportion of participants with a PGI-I score in one of the following 3 categories: 1. 'very much better', 2. 'much better' or 3. 'a little better'. This was obtained using the validated PGI-I questionnaire

A: 19/45; B: 32/49; C: 27/50; D: 36/51

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk"...treatments were assigned using a centralised computer voice response"
Allocation concealment (selection bias)Low risk"...treatments were assigned using a centralised computer voice response"
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskDuloxetine and placebo were given in double-blind fashion. However, it is not specified who exactly was blinded. Participants were blinded to PFMT or sham PFMT
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot stated whether or not outcome assessors were blinded
Incomplete outcome data (attrition bias)
All outcomes
Unclear risk

Dropouts: all: 56/201; A: 10/47; B: 10/50; C: 19/52; D: 17/52

No differential loss to follow-up between group C and D. However, there is excessive loss to follow-up as 56/201 participants were dropped-out.

Selective reporting (reporting bias)Unclear riskTrial protocol not available
Ethical approvalLow riskApproved by the ethics committee
Source of funding or supportLow riskStated, supported by private organisations
Conflict of interestUnclear riskStated but some of the authors had financial and other relationships with one of the organisations which supported the trial

Hofbauer 1990

Methods4-arm randomised controlled trial, parallel design
Participants43 women with urodynamic evidence of stress urinary incontinence (SUI)
Interventions

A. PFMT + electrical stimulation (ES) (n = 11): participants in this group received both PFMT and ES. PFMT was part of an exercise programme which also included abdominal and hip exercise and was administered twice weekly for 20 minutes by a therapist in addition to a daily home exercise programme. Electrical stimulation consisted of vaginal and lumbar electrodes which were administered for 10 minutes, 3 times weekly for a total of 6 weeks. Output was increased until noticeable contraction was achieved and participant then added voluntary effort

B. PFMT alone (n = 11): as described above

C. Electrical stimulation (ES) alone (n = 11): as described above

D. Sham electrical stimulation (n = 10): as for ES above but current was so low that no effect (contraction) was possible

For this review comparison A versus C is relevant

Outcomes

1. Cure: this is the proportion of participants who became continent (free of symptoms of incontinence) at the end of the treatment as reported by the participants

At 10 to 12 weeks from the onset of treatment:

A: 3/11; B: 6/11; C: 1/11; D: 0/11

2. Improvement: proportion of participants who reported improvement in the symptoms of incontinence; success threshold not defined

At 10 to 12 weeks from the onset of treatment:

A: 4/11; B: 1/11; C: 2/11; D: 0/11

3. Success rate: this is the proportion of participants who reported cure of or significant improvement in the symptoms of incontinence

At 10 to 12 weeks from the onset of treatment:

A: 7/11; B:7/11; C: 3/11; D: 0/11

NotesDropouts: not stated
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskFurther translation required
Allocation concealment (selection bias)Unclear riskAs stated above
Blinding of participants and personnel (performance bias)
All outcomes
High riskBlinding of participants undergoing PFMT not possible
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskAs stated above
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskDropouts: not reported
Selective reporting (reporting bias)Unclear riskAs stated above
Ethical approvalUnclear riskAs stated above
Source of funding or supportUnclear riskAs stated above
Conflict of interestUnclear riskAs stated above

Ishiko 2000

Methods3-arm randomised controlled trial, parallel design
Participants61 women with symptoms of stress, urinary incontinence
Interventions

A. Drug therapy (DT) group (n = 18). Participants in this group received clenbuterol tablets 20 µg twice daily

B. PFMT group (n = 20). Participants in this group received instructions on PFMT from gynaecologic specialists until they understood the technique. They were then instructed to perform the exercise for 10 minutes daily (other details not reported). Video tapes that demonstrated the proper method of performing PFMT were also given to the participants

C. PFMT + DT group (n = 23): participants in this group received both clenbuterol and PFMT as described above

For this review comparison C versus A is relevant

Outcomes

1. Cure: as reported by participants and is the proportion of participants who reported 100% reduction in symptoms of incontinence (i.e. no incontinence at all)

A: 10/13; B: 10/19; C: 17/19

2. Patient satisfaction with treatment outcome: defined as the proportion of participants who were completely satisfied with treatment outcome. Scale used for the assessment not stated

A: 11/13; B: 6/19; C: 13/19

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskThe envelope method was used to randomise participants to treatment groups; not stated whether envelopes were sequentially numbered, opaque and sealed
Allocation concealment (selection bias)Unclear riskThe envelope method was used to randomise participants to treatment groups; not stated whether envelopes were sequentially numbered, opaque and sealed
Blinding of participants and personnel (performance bias)
All outcomes
High riskBlinding of participants undergoing PFMT not possible
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot reported
Incomplete outcome data (attrition bias)
All outcomes
Unclear risk

Dropouts: all: 10/61; A: 5/18, B: 1/20; C: 4/23

Differential loss to follow-up: not fully reported (2 and 3 participants withdrew from groups A and C respectively due to adverse drug effects; other reasons for withdrawal not reported)

Selective reporting (reporting bias)Unclear riskTrial protocol not available
Ethical approvalLow riskApproved by the ethics committee
Source of funding or supportUnclear riskNot stated
Conflict of interestUnclear riskNot stated

Jeyaseelan 2002

Methods3-arm randomised controlled trial, parallel design
Participants16 women with stress incontinence
Interventions

A. Electrical stimulation (ES) group (n = 6): participants in this group used electrical stimulator for 1 hour a day every day (except when menstruating)

B. Pelvic floor muscle training (PFMT) alone (n = 7): PFMT consisted of individualised exercise regimen with instruction to the participants to carry out a minimum of 3 exercises per day with progression over the treatment period. Biofeedback was provided by means of a Periform probe. Other details not given

C. PFMT + ES (combined) (n = 6). Participants in this group received both PFMT and ES as described above

For this review comparison C versus A is relevant

Outcomes

1. Severity of incontinence assessed using 24-hour pad test and 3-day voiding diary

2. Condition-specific quality of life assessed using Incontinence Impact Questionnaire (IIQ) and Urogenital Distress Inventory (UDI).

NotesNo useable data were reported in the trial (data reported in median and range)
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNot reported
Allocation concealment (selection bias)Unclear riskNot reported
Blinding of participants and personnel (performance bias)
All outcomes
High riskBlinding of participants not possible. Assumed not done
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot reported
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNot explicitly reported
Selective reporting (reporting bias)Unclear riskNot reported
Ethical approvalUnclear riskNot reported
Source of funding or supportUnclear riskNot reported
Conflict of interestUnclear riskNot reported

Jin 2012

Methods3-arm randomised controlled trial, parallel design
Participants242 women with urodynamic evidence of over-active bladder
Interventions

A. Drug alone (n = 82). Participants in this group received oral solifenacin 5 mg once daily  

B. PFMT alone  (n = 80). Participants in this group performed PFMT once daily; other details were not given

C. PFMT + drug (n = 80). Participants in this group received both PFMT and drug as stated above

For this review comparison C versus A is relevant

Outcomes

1. Frequency of micturition per 24 hours (no useable data)

2. Number of episodes of over-active bladder in 24 hours (no useable data)

3. Volume of urine voided per micturition in 24 hour (no useable data)

4. Adverse events: proportion (%) of participants who reported adverse events (mainly dry mouth) with solifenacin

A: 17/82; B: 0/80; C: 14/80

Notes

Dropouts: not reported

All participants randomised at baseline included in analysis

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNot reported
Allocation concealment (selection bias)Unclear riskNot reported
Blinding of participants and personnel (performance bias)
All outcomes
High riskBlinding of participants not possible for PFMT
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot reported
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNot reported
Selective reporting (reporting bias)Unclear riskTrial protocol not available
Ethical approvalLow risk

"ethics not required" according to the authors

Trial was conducted in accordance with Helsinki declaration

Informed consent was obtained from participants

Source of funding or supportLow riskNo funding source according to the authors
Conflict of interestUnclear riskNot stated

Kim 2011

Methods4-arm randomised controlled trial
Participants147 women with stress, urge or mixed UI from a single centre in Tokyo
Interventions

A. General education (GE) group (n = 36): general education classes were held (topics including cognitive function, osteoporosis and oral hygiene) once a month, a total of 3 times

B. Heat and steam generating sheet (HSGS) group (n = 37): participants in this group received HSGS, a thin, flexible, filmed sheet that generated heat and steam. When placed on the skin surface. It raises the temperature to 38 to 40°C by generating heat and steam continuously for up to 5 hours. Participants were asked to place the HSGS on their lower back once daily immediately after waking period, taking note of the time they started and ended

C. Exercise (Ex) group (n = 37): this group received stretching exercise, fitness exercise and PFM exercise. Participants were initially instructed to perform 10 fast contractions (3 seconds) with a 5-second rest and 10 sustained contractions (8 to 10 seconds) with a 10-second rest between the contractions

D. Ex + HSGS group (n = 37): participants in this group received both exercise and HSGS as described above

For this review comparison D versus B is relevant

Outcomes

Cure of urine loss episodes (assessed by interview, with cure defined as the proportion of participants with complete cessation of urine loss episodes)

At 3 months:

A: 1/34; B: 8/37; C: 12/35; D: 19/37

NotesChanges in frequency of urine loss episodes: assessed based on changes on a 5-point scale obtained in the interviews conducted at baseline (before treatment) and at 3 months after treatment. Data not available, only graphical presentation  
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskUsed "computer-generated random numbers"
Allocation concealment (selection bias)Unclear riskUsed "computer generated random numbers", however, no further information provided
Blinding of participants and personnel (performance bias)
All outcomes
High riskBlinding of the participants not possible
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot stated
Incomplete outcome data (attrition bias)
All outcomes
Low risk4/147 dropped out of the trial. However, there were no dropouts in the comparison of interest and there was no differential loss to follow-up in the other 1 groups
Selective reporting (reporting bias)Unclear riskProtocol not available
Ethical approvalUnclear riskNot stated
Source of funding or supportUnclear riskNot stated
Conflict of interestLow riskDeclared (no conflict of interest)

Richter 2010

Methods3-arm randomised controlled trial, parallel design
Participants446 women with symptoms of stress urinary incontinence
Interventions

A. Continence pessary alone group (n = 149). Individuals in this group were fitted with a continence ring or dish either by a physician or a nurse. Most participants were fitted successfully in 1 clinic visit while up to 3 visits at 1 to 2-week intervals were allowed for others to achieve optimal fitting. At the end of the 8-week treatment period, participants were encouraged to continue to use the pessary

B. Behavioural therapy (PFMT + continence strategies) (n = 146). Intervention in this group consisted of pelvic floor muscle training (PFMT) and exercise and additional skills and strategies on the use of muscles to prevent urgency and stress incontinence. Treatment was administered by registered nurses, nurse practitioners and physical therapists and was implemented in 4 visits at 2-week intervals. During each visit, participants received instructions on PFMT and exercise and also acquired additional skills and strategies on stress urge incontinence prevention. They were then given individualised prescriptions for daily PFM exercise and practice. At the end of the 8-week treatment period, participants received an individualised home maintenance programme to enable them sustain their skills and muscle strength

C. Continence pessary + behavioural therapy (combined) (n = 150). Treatment regimen was as described for both pessary and behavioural therapy groups. In addition, participants in this group could continue in the trial with only 1 of the therapies at the end of the 8-week treatment period

For this review comparison C versus A is relevant

Outcomes

1. The patient global impression of improvement (PGI-I) was assessed for the 3 groups using a validated PGI-I questionnaire with success defined as the proportion of participants with a response of 'much better' or 'very much better'

At 3 months:

A: 59/110; B: 72/124; C: 80/132

At 6 months:

A: 52/102; B: 59/116; C: 63/123

At 12 months:

A: 47/96; B: 48/99; C: 49/111

2. Condition-specific quality of life (in form of the Pelvic Floor Distress inventory): this was assessed using the Urogenital Distress Inventory - stress incontinence sub-scale with success defined as the proportion of participants with absence of bothersome stress incontinence symptoms (indicated by an answer of 'no' to all 6 items on the sub-scale or a response of 'yes' but with a bother of 'not at all' or 'somewhat'

At 3 months:

A: 49/110; B: 71/124; C: 66/132

At 6 months: data not reported

At 12 months:

A: 52/96; B: 59/99; C: 49/111

3. Frequency of incontinence episodes per week (self reported improvement) assessed by using the 7-day diary with success defined as the proportion of women with 75% or more reduction in frequency of incontinence episodes

At 3 months:

A: 69/110; B: 68/124; C: 80/132

At 6 months: data not reported

At 12 months:

A: 51/96; B: 54/99; C: 52/111

4. Patient satisfaction with treatment: this was assessed using the validated Patient Satisfaction Questionnaire

At 3 months:

A: 94/110; B: 110/124; C: 118/132

At 6 months:

A: 87/102; B: 95/116; C: 104/123

At 12 months:

A: 75/96; B: 79/99; C: 81/111

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskPermuted block randomisation schedule was used
Allocation concealment (selection bias)Low riskAllocation contained in sealed envelopes, opened by the interventionist only after the participants met all the inclusion/exclusion criteria
Blinding of participants and personnel (performance bias)
All outcomes
High riskBlinding of participants not possible especially for PFMT
Blinding of outcome assessment (detection bias)
All outcomes
Low riskAll outcome assessors were blinded to the treatment group assignment
Incomplete outcome data (attrition bias)
All outcomes
High risk

Dropouts: at 3 months: all 79/445, C: 18/150, B: 22/146, A: 39/149; at 6 months: all 104/445, C: 27/150, B: 30/146, A: 47/149; at 12 months: all: 139/445; C: 39/150; B: 47/146; A: 53/149

"After randomization, dropout patterns differed among the three treatment groups (P = 0.015) with the pessary only group having the highest attrition rate ..."                                                        

Selective reporting (reporting bias)Unclear riskTrial protocol not available
Ethical approvalLow riskApproved by the ethics committee
Source of funding or supportLow riskDisclosed, funded by "Eunice Kennedy Shriver"
Conflict of interestUnclear riskDeclared some of the authors were associated with a major pharmaceutical company

Wise 1993

Methods3-arm randomised controlled trial, parallel design
Participants62 women with urodynamically proven genuine stress, urinary incontinence (GSI)
Interventions

A. Maximal electrical stimulation alone (n = 20). Participants in this group received a battery-powered vaginal stimulator (impulse frequency: 20 Hz; duration: 0.75 ms; current intensity: 0 to 90 mA) at home daily for 20 minutes

B. Vaginal cones alone (n = 21). Participants in this group were instructed to use cones twice daily for 15 minutes and to increase the weight of the cones when successful on 2 occasions. They did not undergo vaginal examination. It was not reported whether participants were instructed to contract PFMs in order to hold the cones

C. Kegel exercise + vaginal cones (n = 21). Participants in this group received vaginal cones as stated above. In addition, they were taught by vaginal examination to voluntarily contract their pelvic floor muscles and carried out 10 sessions of 10 contractions daily. No further details were reported

For this review comparison C versus B is relevant

Outcomes

1. Improvement: threshold not defined, unclear whether self reported, detailed data not reported, only the level of significance was given for each treatment group

2. Reduction in urine leakage: this was assessed objectively (using pad testing); success threshold was not defined, details of data not reported, only P values were given

3. Decrease in pad weight: only the P values were reported, other details not given

4. Improvement on pad testing: objective assessment of improvement using pad testing; only proportions of participants were reported, success threshold was not defined

A: 12/16; B: 14/19; C: 14/15 

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNot reported
Allocation concealment (selection bias)Unclear riskNot reported
Blinding of participants and personnel (performance bias)
All outcomes
High riskBlinding of participants not possible, especially to PFMT
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot reported
Incomplete outcome data (attrition bias)
All outcomes
High risk

Dropouts: all 12/62; C: 6/21, B: 2/21; A: 4/20

There is differential loss to follow-up

Selective reporting (reporting bias)Unclear riskTrial protocol not available
Ethical approvalUnclear riskNot stated
Source of funding or supportUnclear riskNot disclosed
Conflict of interestUnclear riskNot disclosed

Wyman 1998

  1. a

    BT: bladder training
    DT: drug therapy
    ES: electrical stimulation
    Ex: exercise
    GE: general education
    HSGS: heat and steam generating sheet
    IEF: incontinence episode frequency
    PFM: pelvic floor muscle
    PFMT: pelvic floor muscle training
    SD: standard deviation
    UI: urinary incontinence

Methods3-arm randomised controlled trial, parallel design
Participants204 women with urodynamic evidence of stress urinary incontinence (GSI), detrusor instability (DI) or both (mixed incontinence).
Interventions

A. Bladder training (BT) group (n = 68): involved a progressive voiding schedule that was altered every week for the first 6 weeks of the programme but remained unchanged for the last 6 weeks. The voiding interval was initially set at 30 or 60 minutes, depending on the baseline voiding diary and increased by 30 minutes each week if there was reduction in episodes of incontinence 

B. Pelvic floor muscle training (PFMT) alone (n = 69). PFMT was also structured and it consisted of an initial teaching session (which also included instructions on continence strategies) followed by a graded home exercise with audio cassette practice tapes and 4 office biofeedback sessions. In all, 10 fast (3-second) contractions and 40 sustained (10-second) contractions (a total of 50 contractions) with 10-second rest periods between contractions were performed daily by the third week. Patients received 4 weekly 30-minute sessions of visual and verbal biofeedback. Visual biofeedback was provided via a strip-chart recorder demonstrating vaginal and abdominal pressures as measured by vaginal balloons

 C. PFMT + BT (combined) (n = 67). Treatment regimen was as described for the BT and PFMT groups. BT was implemented initially while PFMT was added during the third week, including instructions on continence strategies (urge inhibition and preventive contractions)

For this review comparison C versus A is relevant

Outcomes

1. Incontinence episodes per week (mean (SD)): this was assessed at endpoint using the records in a standardised diary

Immediately after treatment:

A: 10.6 (16.3), n = 68; B: 9.6 (10.8), n = 64; C: 6.8 (10.7), n = 61

3 months after treatment: data not reported

2. Cure rates: cure was defined as the proportion of participants who had 100% reduction in incontinence episodes, assessed using the standardised diary

Immediately after treatment:

A: 12/67; B: 8/62; C: 19/61

3 months after treatment:

A: 10/63; B: 13/65; C: 16/59

3. Improvement rates: improvement was defined as the proportion of participants who had 50% or greater reduction in incontinence episodes, assessed using the standardised diary

Immediately after treatment:

A: 35/67; B: 36/63; C: 43/61

3 months after treatment:

A: 28/61; B: 36/64; C: 35/59

4. Patient perceived improvement: instrument used in assessment not stated, success threshold was not defined but will be taken as the proportion of participants who were 'much better' or 'somewhat better' for the purpose of this review

Immediately after treatment:

A: 43/66; B: 48/63; C: 55/61

3 months after treatment:

A: 37/60; B: 45/64; C: 44/59

5. Patient satisfaction with treatment outcome: instrument used in assessment not stated, success threshold was not defined but will be taken as the proportion of participants who were 'very satisfied' or 'slightly satisfied' with treatment outcome for the purpose of this review

Immediately after treatment:

A: 48/66; B: 56/63; C: 57/61

3 months after treatment:

A: 47/60; B: 53/64; C: 51/58

6. Condition-specific quality of life assessed at endpoint using:

   i. Urogenital Distress Inventory (UDI); reported as mean (SD):

Immediately after treatment:

A: 95.5 (54.4), n = 67; B: 90.8 (52.0), n = 63; C: 64.4 (48.6), n = 61

3 months after treatment:

A: 91.7 (55.0), n = 60; B: 85.0 (52.4), n = 64; C: 72.8 (50.4), n = 58

   ii. Incontinence Impact Questionnaire-Revised (IIQ-R); reported as mean (SD):

Immediately after treatment:

A: 72.1 (75.2), n = 66; B: 56.8 (61.4), n = 63; C: 46.6 (65.3), n = 61

3 months after treatment:

A: 65.7 (80.2), n = 60; B: 59.3 (67.7), n = 64; C: 59.8 (83.9), n = 58

7. Treatment adherence: this was defined as the proportion of participants adhering to the voiding schedule; assessed using treatment logs or standardised questionnaire (no useable data were: only percentages, without the actual proportions, were reported)

8. Number of women requiring further treatment (relapse): women were followed up for a mean time of 3.2 years and the overall number of women requiring additional treatment such as surgery, drug, etc. was determined for each treatment group

A: 19/48; B: 29/52: C: 18/48

NotesDropouts in each treatment group were not reported
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNot reported
Allocation concealment (selection bias)Unclear riskNot reported
Blinding of participants and personnel (performance bias)
All outcomes
High riskBlinding of participants not possible especially to PFMT
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot reported
Incomplete outcome data (attrition bias)
All outcomes
Unclear risk

Dropouts: immediately after treatment 9/204; 3 months after treatment 16/204

Differential loss to follow-up: not reported

Selective reporting (reporting bias)High riskOne pre-specified outcome (pad weight) was eventually not reported due to large number of missing data according to the authors
Ethical approvalLow riskApproved by the ethics committee
Source of funding or supportLow riskDisclosed (public institutions)
Conflict of interestUnclear riskNot stated

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
  1. a

    PFMT: pelvic floor muscle training

Alewijnse 2003Intervention not relevant
Aslan 2008Intervention not relevant
Barber 2008Participants and intervention not relevant
Bawden 1992Intervention not relevant
BE-DRI 2008Intervention not relevant
Berghmans 2000Intervention not relevant
Berghmans 2000aIntervention not relevant
Berghmans 2001aIntervention not relevant
Berghmans 2002Intervention not relevant
Beuttenmuller 2010Intervention not relevant
Bidmead 2002Intervention not relevant
Bo 2002Intervention not relevant
Bo 2012Systematic review
Burgio 1998Intervention not relevant
Burgio 2001aIntervention not relevant
Burgio 2007Intervention not relevant
Cammu 1996Intervention not relevant
Capobianco 2012Participants not relevant
Chancellor 2008Intervention not relevant
Crothers 2003Intervention not relevant
de Jong 2006Intervention not relevant
Dowell 1997Design not relevant
Driusso 2008Intervention not relevant
Dumoulin 2011Intervention not relevant
Fonda 1995Intervention not relevant
Goode 2003Intervention not relevant
Goode 2011aPost-prostatectomy patients
Greer 2012Systematic review
Gunthorpe 1994Intervention not relevant
Ha 2008Intervention not relevant
Hahn 1991Intervention not relevant
Haken 1991Intervention not relevant
Henalla 1989Intervention not relevant
Herschorn 2004Intervention not relevant
Huang 2006Design not relevant
Huang 2012Intervention not relevant
Kafri 2007Intervention not relevant
Kangchai 2002The study is about the efficacy of a self management promotion programme and the participants were not relevant
Kaya 2011Intervention not relevant
Kim 2001Intervention not relevant
Kim 2006Design not relevant
Kim 2007Intervention not relevant
Kim 2009Intervention not relevant
Kincade 2007Intervention not relevant
Kirschner-Hermanns 1995Intervention not relevant
Kobayashi 2009Intervention not relevant
Lagro-Janssen 1991Intervention not relevant
Laycock 1988Intervention not relevant
Laycock 1993Intervention not relevant
Laycock 1995Intervention not relevant
Laycock 2001Intervention not relevant
Lee 2005Intervention not relevant
Madersbacher 2003Intervention not relevant
Madersbacher 2004Intervention not relevant and recruited both men and women with no separate data for women
Maher 2009Intervention not relevant
McCormack 2004Design not relevant
Millard 2003Participants were provided with a leaflet and were not under a structured PFMT programme and included both men and women (no separate data for women)
Millard 2003aParticipants were provided with a leaflet and were not under a structured PFMT programme and included both men and women (no separate data for women)
Millard 2003bParticipants were provided with a leaflet and were not under a structured PFMT programme and included both men and women (no separate data for women)
Millard 2004Participants were provided with a leaflet and were not under a structured PFMT programme and included both men and women (no separate data for women)
Morkved 2002Intervention not relevant
O'Brien 1996Intervention not relevant
Oldham 2010Intervention not relevant
PRIDE 2004Intervention not relevant
Rutledge 2012Intervention not relevant
Sampselle 2003Design not relevant
Sanchez 2008Intervention not relevant
Savage 2005Design not relevant
Scott 1979Randomisation was not done for intervention (incontinence versus no incontinence)
Smith 1994Intervention not relevant
Sran 2011Intervention not relevant
Suzuki 2003Intervention not relevant
Tapp 1987Intervention not relevant
Terry 1996Intervention not relevant
Van Hespen 2006Participants and intervention not relevant
Viereck 2011Intervention not relevant
Voigt 1996Intervention not relevant
von der Heide 2003Intervention not relevant
Waterfield 2007Participant and intervention not relevant
Wells 1999Intervention not relevant
Wilson 1984Intervention not relevant
Yamanishi 2006Intervention not relevant
Yoon 1999Intervention not relevant
Zhao 2000Intervention not relevant

Ancillary