Pelvic-floor muscle training after birth made the pelvic floor stronger and more enduring but did not significantly reduce urinary incontinence or improve quality of life: in a meta-analysis of 19 randomised trials of training for postpartum urinary incontinence, strength rose (SMD 0.61, 95% CI 0.10-1.12) and endurance rose (SMD 0.68, 0.06-1.30), while incontinence incidence (RR 0.87, 0.73-1.04; p = 0.12) and quality of life (SMD −0.46, −1.18 to 0.27; p = 0.22) showed no significant difference; supervised programmes longer than 8 weeks gave the largest strength gains
Main caveat: The authors': limited data and substantial heterogeneity across trials, so the effect on symptoms is uncertain.
Chen and Zhang asked whether pelvic-floor muscle training treats urinary incontinence after childbirth. They searched Embase, PubMed and the Cochrane Library for randomised trials published before October 2024, screened and extracted in pairs, rated risk of bias with RoB 2.0 and pooled with random-effects models. Nineteen trials were included.
Training improved pelvic-floor muscle strength (SMD 0.61) and endurance (SMD 0.68), both significant. It did not significantly change urinary incontinence incidence (RR 0.87, 95% CI 0.73-1.04) or quality of life (SMD −0.46, −1.18 to 0.27). In subgroup analyses, supervised programmes lasting more than 8 weeks were the most effective for strength.
The authors conclude that training clearly improves pelvic-floor muscle function in postpartum women with incontinence, in line with its mechanism, but that its direct effect on symptoms, incontinence incidence and quality of life, is less certain because of limited data and substantial heterogeneity, and should be read with caution.
Participants
19 RCTs of postpartum women; the total number of women, their age, parity, mode of delivery, incontinence type and severity at entry, and timing after birth are not in the abstract.
Protocol
Search: Embase, PubMed, Cochrane Library; studies published before October 2024.
Selection and appraisal: two reviewers independently screened, extracted and rated risk of bias with the Cochrane Risk of Bias Tool 2.0.
Analysis: random-effects meta-analysis; risk ratios for incontinence incidence, SMDs for strength, endurance and quality of life; subgroup analyses (by supervision and duration, among others not named).
Not in the abstract: the comparators (usual care, no training, other), training dose, how strength and endurance were measured, the quality-of-life instruments and which direction of the SMD means better, the number of trials per outcome, I² values, risk-of-bias results, registration and funding.
Key Results
| Outcome (PFMT vs comparator) | Effect (95% CI) | p | Significant |
|---|---|---|---|
| Pelvic-floor muscle strength | SMD 0.61 (0.10 to 1.12) | 0.02 | yes |
| Pelvic-floor muscle endurance | SMD 0.68 (0.06 to 1.30) | 0.03 | yes |
| Urinary incontinence incidence | RR 0.87 (0.73 to 1.04) | 0.12 | no |
| Quality of life | SMD −0.46 (−1.18 to 0.27) | 0.22 | no |
Subgroup: supervised PFMT lasting more than 8 weeks was the most effective approach for improving strength (no effect size in the abstract).
The abstract gives the number of trials and women for no individual outcome.
Practical Application
Pelvic-floor training after birth strengthens the pelvic floor; whether it stops leaking is less certain. Across 19 trials (per-outcome counts not in the abstract) it raised strength (SMD 0.61) and endurance (SMD 0.68), but incontinence incidence (RR 0.87, 0.73-1.04, not significant: from 27% lower to 4% higher) and quality of life (SMD −0.46, −1.18 to 0.27, not significant) showed no significant difference.
What this changed in [SYNTH-95]: the guide's pelvic-floor line used to say without qualification that after birth pelvic-floor training "lowered the odds of leaking by 37%", from REF-1398 (7 RCTs, OR 0.63, 0.41-0.97, moderate certainty).This second meta-analysis (19 trials in all; how many fed the incontinence outcome is not in the abstract) did not find a significant reduction in incontinence (point estimate 13% lower risk, interval including no effect) and found no significant gain in quality of life.So the guide now states the reduction in leaking as found in one meta-analysis and not confirmed in another, not as a settled 37%, and lists a promise that the training will stop leaking among the things not to do; REF-1398 itself also found no difference in severity.The two are not a flat contradiction: they use different measures (odds ratio against risk ratio), possibly different trial sets and populations, and their intervals overlap.
What it does not change: pelvic-floor training stays in the guide from pregnancy on and after birth. Both meta-analyses point the same way on leaking (fewer cases, significant in one), this one shows the muscles respond, and the supervised-and-longer-than-8-weeks finding for strength fits the guide's rule that the programme belongs to a pelvic-health physiotherapist. Nothing in this abstract shows harm, and harms are not reported.
How the assistant says it (inference, not tested here): "Pelvic-floor training makes the muscles stronger, and in one large review fewer women leaked after it; another review found the drop in leaking too uncertain to call. It is worth doing, ideally supervised by a pelvic-health physio; if you still leak, tell them."
Not treatment advice: leaking, heaviness or pelvic pain after birth goes to the obstetric provider or a pelvic-health physiotherapist, who decides on assessment and the programme (SYNTH-95).
Limitations
The authors': limited data and substantial heterogeneity across trials, so the effect on symptoms is uncertain.
Not in the abstract: the number of women, trials per outcome, I² values, the risk-of-bias results, certainty, comparators, training doses, outcome instruments and the direction of the quality-of-life scale, and funding.
Population unclear: framed as treatment of existing incontinence, yet the main symptom outcome is "incidence"; whether prevention trials in continent women were mixed in is not stated.
Marker-adjacent outcomes: strength and endurance of the pelvic floor are muscle-function measures; the patient-relevant outcomes, leaking and quality of life, are the ones that did not reach significance.
Written from the abstract and the PubMed record; full text not read.
Revision 2026-11-10 (decision): the lines on SYNTH-95 now describe the guide as it stands — the 37% is one meta-analysis not confirmed by this one, and promising that leaking will stop is listed under What NOT to do.
How this entry was checked and changed
Why this level
a meta-analysis of randomised trials, risk of bias not read (abstract only): the abstract names RoB 2.0 but does not report what it found and gives no certainty rating; the authors also report substantial heterogeneity, with no I² values in the abstract