PORTRET: Surgery versus Physiotherapy for Stress Urinary Incontinence
PORTRET (Physiotherapy OR TVT Randomised Efficacy Trial) was a multicenter, open-label randomized trial in 460 women with moderate-to-severe stress urinary incontinence in the Netherlands that compared initial midurethral sling surgery with initial supervised pelvic floor muscle physiotherapy, with crossover allowed and 12-month Patient Global Impression of Improvement (PGI-I) as the primary endpoint. Subjective improvement was reported by 90.8% of surgery-group women and 64.4% of physiotherapy-group women with outcome data (absolute difference 26.4 percentage points, 95% CI 18.1 to 34.5), and 49.0% of the physiotherapy group crossed over to surgery.[1]
- Design and population
- Multicenter pragmatic open-label randomized superiority trial. Women aged 35 to 80 years with moderate or severe stress urinary incontinence (stress predominant if mixed) at 4 university and 19 general hospitals in the Netherlands.
- Arms
- Initial midurethral sling surgery versus initial supervised pelvic floor muscle physiotherapy, with crossover to surgery allowed
- Report and follow-up
- 2013 · 12 months; outcome data for 196/215 (91.2%) surgery and 174/202 (86.1%) physiotherapy women analyzed
| Endpoint | Results | Effect | Finding |
|---|---|---|---|
| 12-month subjective improvement (PGI-I much or very much better); absolute difference in percentage points, surgery minus physiotherapy | 177/195 (90.8%) vs 112/174 (64.4%) | difference 26.4; CI 18.1–34.5; P <0.001 | Met |
| 12-month subjective cure (no leakage with physical activity, coughing or sneezing on the UDI item); absolute difference in percentage points | 167/196 (85.2%) vs 93/174 (53.4%) | difference 31.8; CI 22.6–40.3; P <0.001 | Met |
| 12-month objective cure (negative cough stress test at bladder volume of at least 300 mL); absolute difference in percentage points | 140/183 (76.5%) vs 94/160 (58.8%) | difference 17.8; CI 7.9–27.3; P 0.001 | Met |
Harms: Bladder perforation: 6/215 (2.8%) vs 0/202 (0.0%); Vaginal epithelial perforation: 8/215 (3.7%) vs 2/202 (1.0%); Reoperation for tape exposure: 5/215 (2.3%) vs 1/202 (0.5%)
Netherlands Trial Register NTR1248 · Report DOI
Design
Women aged 35 to 80 years were referred to an outpatient gynecology or urology clinic with stress incontinence graded moderate or severe on the Sandvik severity index. Mixed incontinence was eligible when stress episodes outnumbered urge episodes on the Urogenital Distress Inventory (UDI). Leakage on straining or coughing at a bladder volume of at least 300 mL established the diagnosis; urodynamics was not required. Previous incontinence surgery and prolapse of stage 2 or higher excluded enrollment. Women could be untreated or have had physiotherapy more than 6 months earlier (34/215 and 38/202 analyzed women in the two groups).[1]
Of 656 women invited, 460 consented and were randomized 1:1 by a central computerized system in blocks of four per center, stratified by incontinence severity. Treatment assignments were not concealed. Enrollment ran from March 2008 to May 2010 at 4 university and 19 general hospitals, and manufacturers of slings took no part in the trial.[1]
The primary endpoint was PGI-I at 12 months, dichotomized at "much better" or "very much better." Subjective cure was a negative answer to the UDI question on leakage with physical activity, coughing or sneezing, and objective cure was a negative cough stress test at a bladder volume of at least 300 mL. The sample-size calculation assumed 80% improvement with surgery and 65% with physiotherapy and required 197 women per group for 90% power; 460 were planned to allow for 15% loss to follow-up. The main analysis was a modified intention-to-treat analysis that kept crossovers in their assigned group, repeated with multiple imputation as a sensitivity analysis. A per-protocol comparison of three groups (physiotherapy only, surgery after physiotherapy, initial surgery) was added post hoc.[1]
Results
Analyses included 215 surgery-group and 202 physiotherapy-group women. In the surgery group 15 women withdrew consent before treatment; in the physiotherapy group 28 did not start treatment (27 withdrew consent, 1 did not meet inclusion criteria). At 12 months, outcome data were available for 196 (91.2%) and 174 (86.1%) of them.[1]
PGI-I improvement, the primary endpoint, was reported by 177/195 (90.8%) after initial surgery and 112/174 (64.4%) after initial physiotherapy (difference 26.4 percentage points, 95% CI 18.1 to 34.5; P<0.001), which met the primary hypothesis; with imputed data the difference was 26.5 points (95% CI 18.5 to 34.2). Subjective cure was 167/196 (85.2%) versus 93/174 (53.4%) (difference 31.8 points, 95% CI 22.6 to 40.3; P<0.001), and objective cure was 140/183 (76.5%) versus 94/160 (58.8%) (difference 17.8 points, 95% CI 7.9 to 27.3; P=0.001). The UDI urinary incontinence domain fell by 37.3 points after surgery and 26.9 points after physiotherapy (P<0.001), a moderate effect size.[1]
In the post hoc per-protocol analysis, PGI-I improvement was 26/82 (31.7%) in women who had physiotherapy only, 86/92 (93.5%) in women who crossed over to surgery and 177/195 (90.8%) after initial surgery. Surgery after physiotherapy and initial surgery did not differ (P=0.68), and physiotherapy only was lower than either (P<0.001). Because the crossover decision was made by the woman, this comparison is not randomized.[1]
A questionnaire follow-up at a mean of 11.9 years (range 10.5 to 14.8) reached 184 of 386 included women (47.7%). Among responders, PGI-I improvement did not differ significantly by initial assignment: 87/100 (87.0%) after initial surgery and 66/83 (79.5%) after initial physiotherapy (P=0.229), and 73/84 (86.9%) of responders assigned to physiotherapy reported having had surgery. When non-responders' 12-month values were carried forward, improvement was 163/188 (86.7%) versus 108/164 (65.9%) (difference 20.8 points, 95% CI 12.1 to 29.6; P<0.001), and subjective cure was 142/180 (78.9%) versus 86/156 (55.1%). In the post hoc grouping, responders who had physiotherapy only reported improvement in 4/11 (36.4%), against 87.0% after initial surgery and 86.1% (62/72) after surgery following physiotherapy.[2]
Surgery and treatment delivery
Forty-nine gynecologists and urologists placed the slings, each with at least 20 previous procedures; both retropubic and transobturator routes were allowed. Physiotherapy followed the Dutch guideline and was delivered by 83 of the 478 certified pelvic physiotherapists in the Netherlands: education on pelvic floor function, a coordinated contraction before increases in abdominal pressure, and a supervised program building to 8 to 12 maximal contractions three times daily. Sessions were intended to number nine over 9 to 18 weeks, and biofeedback or functional electrostimulation could be added.[1]
A woman dissatisfied with her assigned treatment could cross over, as in usual practice. In the physiotherapy group, 99/202 (49.0%) had surgery after a mean of 31.7 (SD 12.7) weeks, and 90 of these 99 had reported no improvement on the last PGI-I before surgery. Women who did not cross over attended 9.1 (SD 4.9) sessions and those who did cross over attended 7.4 (SD 4.4). Twenty-two women (11.2% as reported) received additional physiotherapy after surgery to relax the pelvic floor for obstructive voiding symptoms.[1]
Adverse events
Sixty-five adverse events occurred in 41 of 417 women (9.8%), all related to surgery; every event in the physiotherapy group occurred in a woman who crossed over. Among 215 women in the surgery group, bladder perforation occurred in 6 (2.8%), vaginal epithelial perforation in 8 (3.7%), hematoma in 16 (7.4%), blood loss of 500 mL or more in 2 (0.9%) and new urge incontinence in 13 (6.0%). Bladder and vaginal epithelial perforations were repaired during surgery without further clinical consequence. Six women had reoperation for tape exposure (5 in the surgery group, 1 in the crossover group) and one needed the sling loosened for persistent voiding problems.[1]
At 12 years, 8/173 responders who had received a sling (4.6%) reported a urogenital reoperation: 5 for a surgical complication, 2 for recurrent incontinence and 1 for recurrence combined with a complication.[2]
Limitations
Assignments were not concealed, and the primary endpoint and subjective cure were patient reported. Women who preferred surgery may have been more likely to enroll, because Dutch guidelines would otherwise have directed them to physiotherapy first. The 49.0% crossover rate means the 12-month intention-to-treat result compares a strategy of physiotherapy first, with surgery for nearly half, against sling surgery; the per-protocol analysis separating physiotherapy-only women was post hoc. Objective cure rested on a cough stress test available for 183 surgery-group and 160 physiotherapy-group women; in physiotherapy-only women subjective cure was 13/82 (15.9%) but objective cure 33/75 (44.0%), which the authors attribute to women controlling their pelvic floor muscles during the provocative test yet leaking with unexpected events in daily life.[1]
The 12-year follow-up had a 47.7% response rate, carried 12-month values forward for non-responders (which the authors state may underestimate crossover or overestimate outcomes), recorded patient-reported outcomes only and did not measure long-term adherence to pelvic floor training.[2]
In conversation
- Confirms ESTEEM (2019). Physiotherapy before sling added nothing at 12 years; ESTEEM likewise found perioperative pelvic floor therapy gave no clinically significant symptom benefit over sling alone.
Same setting ATLAS (2010)
The 12-year report cites ESTEEM, in which adding behavioral and pelvic floor muscle therapy to sling surgery in women with mixed incontinence gave no clinically significant symptom difference at 12 months, and reads it as consistent with the finding that physiotherapy before surgery added nothing to initial surgery.[2]
Guidelines and approvals
- EAU Guidelines on Non-neurogenic Female LUTS (limited update March 2026). The pelvic floor muscle training chapter describes PORTRET's 12-month results, and its summary of evidence states that midurethral sling surgery is superior to pelvic floor muscle training in women with moderate-to-severe SUI (level of evidence 1b). Its recommendations say to offer supervised intensive pelvic floor muscle training lasting at least three months as first-line therapy to all women with SUI or mixed incontinence , and to balance the efficacy and lack of adverse events of training against the expected effect and possible complications of invasive surgery for SUI .[3]
Report history
| Published | Report | Cutoff | Headline results | Citation |
|---|---|---|---|---|
| 2013-09-19 | 12-month randomized report · primary | — | 12-month subjective improvement (PGI-I much or very much better); absolute difference in percentage points, surgery minus physiotherapy: 177/195 (90.8%) vs 112/174 (64.4%); difference 26.4; CI 18.1–34.5; P <0.001 Met 12-month subjective cure (no leakage with physical activity, coughing or sneezing on the UDI item); absolute difference in percentage points: 167/196 (85.2%) vs 93/174 (53.4%); difference 31.8; CI 22.6–40.3; P <0.001 Met 12-month objective cure (negative cough stress test at bladder volume of at least 300 mL); absolute difference in percentage points: 140/183 (76.5%) vs 94/160 (58.8%); difference 17.8; CI 7.9–27.3; P 0.001 Met 12-month PGI-S no symptoms; absolute difference in percentage points: 167/195 (85.6%) vs 114/174 (65.5%); difference 20.1; CI 11.4–28.6; P ≤0.001 Mean change in UDI urinary incontinence domain score at 12 months (0 to 100; lower is better): −37.3 (SD 20.9) vs −26.9 (SD 20.6); P <0.001 12-month PGI-I improvement, post hoc per-protocol groups: initial surgery versus physiotherapy only; absolute difference in percentage points: 177/195 (90.8%) vs 26/82 (31.7%); difference 59.1 | Labrie J, Berghmans BL, Fischer K, et al. Surgery versus physiotherapy for stress urinary incontinence. *N Engl J Med.* 2013;369(12):1124-1133. DOI |
| 2025-02-11 | 12-year questionnaire follow-up · updated | — | Long-term PGI-I improvement, intention-to-treat, responders only; absolute difference in percentage points as printed in Table 1: 87/100 (87.0%) vs 66/83 (79.5%); difference -7.5; CI -18.5–3.6; P 0.229 Not met Long-term PGI-I improvement, intention-to-treat, all women with 12-month values carried forward for non-responders; absolute difference in percentage points: 163/188 (86.7%) vs 108/164 (65.9%); difference 20.8; CI 12.1–29.6; P <0.001 Met Long-term subjective cure, all women; absolute difference in percentage points: 142/180 (78.9%) vs 86/156 (55.1%); difference 23.8; CI 13.9–33.7; P <0.001 Met Long-term PGI-S no symptoms, all women; absolute difference in percentage points: 144/188 (76.6%) vs 105/165 (63.6%); difference 13; CI 3.4–22.5; P 0.010 Long-term PGI-I improvement, post hoc: initial surgery versus physiotherapy only, responders only; absolute difference in percentage points: 87/100 (87.0%) vs 4/11 (36.4%); difference 50.6; CI 28.2–73.1 Long-term PGI-I improvement, post hoc: initial surgery versus physiotherapy only, all women; absolute difference in percentage points: 163/188 (86.7%) vs 14/57 (24.6%); difference 62.1; CI 49.7–74.6 | van Oorschot HFC, Tijsseling D, Labrie J, van der Vaart CH. Twelve-year follow-up of a randomised controlled trial comparing the effectiveness of pelvic floor muscle training versus mid-urethral sling surgery for female moderate to severe urinary incontinence. *BJOG.* 2025;132(6):826-833. DOI |
References
1. Labrie J, Berghmans BLCM, Fischer K, et al. "Surgery versus Physiotherapy for Stress Urinary Incontinence." N Engl J Med. 2013;369(12):1124-1133. doi:10.1056/NEJMoa1210627
2. van Oorschot HFC, Tijsseling D, Labrie J, et al. "Twelve-Year Follow-Up of a Randomised Controlled Trial Comparing the Effectiveness of Pelvic Floor Muscle Training Versus Mid-Urethral Sling Surgery for Female Moderate to Severe Urinary Incontinence." BJOG. 2025;132(6):826-833. doi:10.1111/1471-0528.18092
3. European Association of Urology. EAU Guidelines on Management of Non-Neurogenic Female Lower Urinary Tract Symptoms (LUTS) (Limited Update March 2026), section 4.2.4.a.3.g. Guideline.