Stress Urinary Incontinence: What Pelvic Floor Training Does for Leaking

Stress urinary incontinence is leaking on effort, when you cough, sneeze, laugh, run or lift, with no warning urge beforehand, and the first-line treatment for it in the guidelines is a supervised programme of pelvic floor muscle training of at least 3 months. It is the condition on which pelvic health physiotherapy has its strongest evidence: in a Cochrane review, 56% of women were cured with training against 6% of controls¹, and NICE makes the supervised course an offer rather than a suggestion².

This is my condition, so I should declare the bias at the top. Mine started about six months after my second child, showed up first on a hill sprint in a PE lesson, and got named three years later by the fourth clinician I mentioned it to. Sixteen weeks of prescribed work took it from something that shaped my running routes to something that occasionally reminds me it exists. That is a good outcome and it is not a cure, and I would rather say that than write a recovery story. Before anything else here, read red flags and when to stop and get checked.

What stress urinary incontinence is, precisely

It is involuntary leakage on effort or exertion, or on sneezing or coughing. The mechanism is a mismatch: abdominal pressure rises faster than the urethra can be held closed, so a small volume escapes. There is no urgency beforehand, which is the clinical dividing line from urge incontinence and overactive bladder, a different problem that gets offered a lot of the same advice. When both patterns are present it is called mixed urinary incontinence, and the dominant symptom is treated first.

The proportions are worth knowing because they change what your appointment should cover. In the Norwegian EPINCONT survey of 27,936 community-dwelling women aged 20 and over, of the women with incontinence about half had stress type, 36% mixed and 11% urgency³.

How common it is, and why the numbers disagree

Population studies from many countries report the prevalence of any urinary incontinence in women as ranging from about 5% to 70%, with most reporting 25% to 45%, and more than 40% of women aged 70 and over affected⁴. The Norwegian survey found 25% of women had leakage of any kind, and nearly 7% had significant incontinence, defined as moderate or severe and experienced as bothersome³. A nationally representative US sample of 1,961 women found 15.7%, with a 95% confidence interval of 13.2 to 18.2, when the definition was tightened to moderate to severe leakage on a validated severity index⁵.

Those figures are not in conflict and Norwegian bladders are not different from American ones. The 25% counts any leakage and the 15.7% counts moderate to severe leakage, and the gap between them is entirely definitional. Which is worth holding onto, because it is exactly the mechanism by which you can be told your problem is both extremely common and not worth treating in the same sentence.

What the evidence says training does

The Cochrane review compares pelvic floor muscle training against no treatment, placebo, sham or another inactive control across 31 trials and 1,817 women. For stress urinary incontinence specifically¹:

• Cure: 56% with training against 6% of controls, a risk ratio of 8.38 with a 95% confidence interval of 3.68 to 19.07, from 4 trials and 165 women, graded high certainty.
• Cure or improvement: 74% against 11%, a risk ratio of 6.33 with a 95% confidence interval of 3.88 to 10.33, from 3 trials and 242 women, moderate certainty.
• Leakage episodes: 1.23 fewer per 24 hours, with a 95% confidence interval of 1.78 to 0.68 fewer, from 7 trials and 432 women, moderate certainty.

Three things must travel with those numbers. The high-certainty cure result comes from 165 women, which is a very small base for an eightfold effect. Quality of life was not pooled in the review, is graded low certainty and should never be quoted as a single effect size. And the authors themselves note that limited follow-up beyond the end of treatment means long-term outcomes and cost-effectiveness remain uncertain.

A separate review puts the direction of travel plainly: there is Level 1, Grade A evidence that pelvic floor muscle training is effective in treating stress urinary incontinence, with short-term cure rates measured as less than 2 g of leakage on pad testing varying between 35% and 80%, and supervised, more intensive training being more effective than unsupervised training⁶. That 35% to 80% spread is not sloppiness. It is what happens when different trials define cure differently.

What the guideline actually says

NICE NG123 recommendation 1.4.4 says to offer a trial of supervised pelvic floor muscle training of at least 3 months’ duration as first-line treatment to women with stress or mixed urinary incontinence, and 1.4.7 says to continue the programme if it is beneficial². NG210 repeats the offer for women including pregnant women, and adds recommendation 1.6.18: at least one review to assess progress during the programme, and one review at the end⁷.

On dose, NG123 states that programmes should comprise at least 8 contractions performed 3 times per day. Read the tag before you treat that as current evidence: that recommendation is marked [2006] in the guideline, meaning it was carried forward unchanged from the previous guidance and was not re-reviewed in the 2019 update². It is a minimum for a programme, not a personal prescription, and how a real prescription is built is covered in pelvic floor exercise programmes explained.

Why supervision is the word that matters

Because the movement is easy to get wrong, and wrong in a specific way. In 47 women referred for urodynamic evaluation of urinary incontinence, only 49% achieved an ideal contraction after brief standardised verbal instruction, and 25% performed a technique that could potentially promote incontinence, with the authors concluding that simple verbal or written instruction does not represent adequate preparation for starting a programme⁸. That figure needs its companion, because on its own it gets repeated as “half of all women cannot do a pelvic floor contraction”, which is not what it says. In 779 women attending ordinary community primary care, correct contraction on the first attempt ranged from 68.6% in women with neither prolapse nor stress incontinence down to 85.8% in women with both, and of the 120 who got it wrong at first, 78% learned after brief instruction⁹. The difference between the two studies is the denominator: Bump looked at 47 women already referred with a problem, Henderson at 779 walking in off the street. The defensible claim is narrower and more useful than the scary one: a meaningful minority get it wrong, most of them learn quickly once someone shows them, and the error rate is highest among exactly the people who already have symptoms and are therefore reading this page.

That was a referred, symptomatic population in 1991, and it is not a statement about all women. But the population it describes is the population reading this page. Getting the contraction checked by someone who can feel it is the difference between three months of training and three months of bracing, and it is covered in how to do a pelvic floor contraction correctly. If squeezing harder makes your symptoms worse, read hypertonic pelvic floor and downtraining before you do another set.

What else is in the programme

Bladder habit and fluids, because volume and caffeine change how much a weak closure has to hold: see bladder training and fluid advice. Bowel habit, because straining loads the same structures daily: see bowel habit and defaecation technique. Load and impact, because the leak usually appears at a specific point in a specific activity and that point can be trained towards: see returning to running and lifting.

The single most useful thing I brought to my first appointment was a three-day bladder diary I nearly did not bother filling in. It converted “I leak when I run” into a pattern with times, volumes and triggers on it, and the assessment moved much faster because of it.

What happens after discharge

Adherence falls, and this site does not pretend otherwise. A systematic review of 19 studies following 1,141 women for between 1 and 15 years found long-term adherence to training ranging from 10% to 70%, long-term success among the original responders ranging from 41% to 85%, and surgery rates at long-term follow-up ranging from 4.9% to 58%¹⁰.

The finding people skip is in the same review: short-term outcomes could be maintained at long-term follow-up without incentives for continued training. So the honest position is that most people stop doing the full programme, and that the benefit does not collapse the moment they do. Keeping it up after discharge is the practical version of that.

When surgery enters the conversation

After a fair trial, or after an informed refusal of one. NICE recommendation 1.5.2 sets out the options when non-surgical management has failed and the woman wishes to consider surgery: colposuspension, open or laparoscopic, or an autologous rectus fascial sling, with the option of a retropubic mid-urethral mesh sling only subject to a further set of conditions².

That list looks the way it does because of what happened before it. The Cumberlege review, published in July 2020, heard from women who underwent mesh surgery for relatively minor stress urinary incontinence without having first had, or having been offered, conservative treatment, and stated that conservative measures must be offered to women before surgery¹¹. In England a period of restricted practice on vaginal mesh for stress incontinence began in July 2018 and, at the most recent authoritative statements traced, remained in place; NHS information notes that it may not currently be possible to have vaginal mesh surgery for urinary incontinence on the NHS¹². It was a restriction and a high vigilance period, not a ban.

For what a fair trial means and how you know it is over, see when physiotherapy is not enough. For the treatment itself in full, see pelvic floor physiotherapy, and for the mesh history, mesh and what changed after the inquiries.