Swimmer's shoulder: the stroke faults that load the cuff

Shoulder pain is the reason most adult swimmers stop swimming. It rarely arrives as an injury — there is no moment, no pop, nothing to tell your doctor about. It shows up as a dull ache at the front of the shoulder that appears forty minutes into a session, then thirty, then ten, until the whole thing stops being fun.

Search for the cause and you will be told, confidently, that you are crossing your midline. Maybe you are. But the honest version of this topic starts somewhere less satisfying: the best available synthesis of the research could not find strong evidence for any single stroke fault as a cause. What it did find points somewhere else entirely, and that changes what you should do about it.

What "swimmer's shoulder" actually means

Swimmer's shoulder is not a diagnosis. It is an umbrella term for any shoulder pain caused by swimming, and it covers several different problems: rotator cuff tendinopathy, shoulder impingement, bursitis, a SLAP tear of the labrum, or instability. Each of those is managed differently, which is exactly why a generic internet answer cannot tell you what yours is. What they share is a mechanism — repeated load on tissue that has not been given enough time to adapt to it.

The most common single finding, at least in swimmers fast enough to be studied, is tendinopathy of the supraspinatus — the rotator cuff muscle that runs across the top of the shoulder joint. In a study of 80 elite swimmers aged 13 to 25 by Sein and colleagues, published in the British Journal of Sports Medicine, 73 of the 80 reported shoulder pain, 84% had a positive impingement sign, and 69% of those scanned had supraspinatus tendinopathy visible on MRI.

The uncomfortable part: the evidence for any single stroke fault is thin

In 2023, McKenzie and colleagues published a systematic review in the Scandinavian Journal of Medicine & Science in Sports that assessed 80 candidate risk factors for shoulder pain and injury in competitive swimmers across 22 studies. Their conclusion, stated plainly: there was no strong evidence supporting or refuting the association between any of those 80 variables and shoulder injury or pain.

Several things swim coaching treats as settled came out with moderate evidence against them: internal and external range of motion, training frequency, specialty stroke, height and weight, sex, and age. That does not mean these things are irrelevant. It means the studies that exist have not been able to show they predict who gets hurt.

The same pattern shows up in the Sein data. The number of hours swum per week correlated with supraspinatus tendinopathy, and so did weekly mileage. Stroke preference did not. Butterfly swimmers were not the casualties; high-volume swimmers were.

What the evidence does point at: load, and your shoulders' capacity to hold it

Two findings survive the scrutiny reasonably well, and both are about load rather than shape.

The first is volume. Feijen and colleagues reviewed 12 studies covering 1,460 competitive swimmers for the Journal of Athletic Training and found the highest rates of shoulder pain in adolescents — 91.3%, against a range of 19.4% to 70.3% in the other age groups — and it was the adolescents who trained most, at roughly 17 hours a week. Weekly training volume and years in the sport both correlated with supraspinatus tendon thickness, and every swimmer in whom the tendon had thickened had shoulder pain.

The second is the shape of the load over time. The highest-quality study in the McKenzie review, with 201 swimmers, found that a high acute-to-chronic workload ratio — a sharp jump in this week's training relative to what you have been doing for the past month — and reduced posterior shoulder strength endurance were associated with injury. McKenzie's own conclusion was that shoulder strength-endurance may be the single most clinically relevant modifiable factor we have.

Put those together and the picture is not "you have a bad stroke". It is closer to: too much, too soon, on shoulders that fatigue too early. That is good news, because both halves of that are things you control.

How many arm strokes you are actually asking for

The scale of this is easy to underestimate, so count it. A 3,000 m session in a 25 m pool is 120 lengths. At 18 strokes a length — a normal count for a competent adult freestyler — that is 2,160 arm pulls, about 1,080 per shoulder. Swim four times a week and each shoulder does roughly 4,300 pulls a week, or over 200,000 in a training year.

The useful consequence: anything that reduces your stroke count reduces the number of loading cycles for the same distance swum. A swimmer who drops from 22 strokes a length to 18 has removed 480 pulls from a 3,000 m session without swimming less. That is why body position work is a shoulder intervention as much as a speed one, and why fixing legs that sink often quiets a shoulder that nobody touched.

The four stroke faults worth fixing anyway

These are the ones U.S. Masters Swimming's published guidance keeps returning to, and the mechanism for each is at least coherent. Treat them as coaching consensus with a plausible rationale, which is what they are.

Crossing your midline

This happens in two places and they are different faults. At entry, the hand lands past the centre line of your body, so the arm starts the pull in a position of more internal rotation than it should. During the pull, the hand sweeps across under your chest. Both put the shoulder in a narrower, less mechanically favourable position and, as Alex Ewart writes for USMS, both require more effort to hold the same speed — which means the surrounding muscles fatigue sooner, and a fatigued shoulder takes more of the load through structures that were not meant to carry it.

The tell is timing: pain that is absent for the first twenty minutes and reliable after forty is a fatigue pattern, not a structural one.

Pulling too wide

Less discussed and, in adult swimmers, at least as common. A wide pull is the same problem as trying to push yourself out of the pool with your hands set far apart — the leverage is worse, so you work harder for the same result. The fix is not to force the hands narrow; it is usually to restore the body roll that made the wide path feel necessary.

Dropping your elbow

The elbow sinks below the hand at the catch, so instead of pressing back against the water with the whole forearm, you push down with the hand. USMS describes this as one of the most common errors among masters swimmers: it increases the external rotation the shoulder is driven into during the pull, loads the rotator cuff more, and takes the latissimus dorsi — the strongest muscle available for the job — out of the stroke.

Two cues from the same source are worth trying. Press with the forearm, not the hand. And start the pull deeper: many adults begin with the hand within a couple of inches of the surface, where a high-elbow catch is nearly impossible; six to eight inches down, fingertips angled slightly toward the bottom, makes it available.

Swimming flat

Without rotation, the recovering arm has to reach further back behind the body, and the pulling arm has to travel wider. USMS puts the workable range at roughly 20 to 30 degrees of body roll, while noting it varies between swimmers. If your pain lives in the recovery rather than the pull, this is the first thing to test. Rotation is also the thing that bilateral breathing tends to even out, which is a reason to train it even though it will not make you faster.

Butterfly asks for all four at once, because both arms recover over the water on every cycle and there is no side to rotate towards. Coaches deal with that through timing rather than strength — letting the undulation carry the recovery, which is what the two-kick timing is about. That is consensus reasoning, not trial evidence, and it is another argument for keeping butterfly volume modest.

Where the pain sits tells you which fault to look at

This is the most useful piece of practical framing in the USMS material: the phase of the stroke in which it hurts narrows the list. Three swimmers can have pain in the same spot at the front of the shoulder for three different reasons.

Film yourself. Every one of these faults is invisible from inside the stroke and obvious from the side of the pool, and a phone in a cheap waterproof case propped on the lane rope beats any amount of reading.

Paddles, bands and the dryland question

Hand paddles increase the surface area your shoulder pulls against, which is the point of them and also the risk. USMS's guidance is unremarkable and sensible: do not use them every session, and keep most of your volume paddle-free. What paddles and fins actually train sets out when they earn their place in a set. The same logic applies to bands and to any drill that makes the arms do the work the legs were doing. Kick sets are the usual substitute when the arms need a quiet week, with one caveat: a board held out in front holds the shoulders in extension for the whole set, so a board-free kick on your side is the kinder version. What kickboard work is actually for goes through both sides of it.

Dryland training cuts both ways. Too little leaves you without the posterior shoulder strength endurance that the research keeps flagging; too much, or badly balanced — all pressing, no pulling — adds load to the same tissue you are already loading in the pool. If you are going to add strength work specifically for this, get it prescribed by someone who can watch you do it.

The 25 m and 50 m pool difference nobody mentions

Same distance, different number of arm pulls. Every push-off carries you several metres without a stroke, and a 25 m pool gives you roughly twice as many push-offs as a 50 m pool for the same swim. A 3,000 m session short course includes about 119 walls; long course, about 59. If each push-off and glide covers five metres before you take a stroke, the short-course session asks for around 300 m less stroking — about a tenth of the swim.

That is an estimate built on an assumed glide distance, not a measurement, so treat the size of the number loosely. The direction is solid, and it is worth knowing if you have just moved from a 25 m pool to a 50 m one and your shoulder started complaining for no visible reason. Nothing about your stroke changed; the number of times you performed it went up. The course conversion tool shows the same arithmetic from the timing side.

The adult swimmer's version of this problem

Most of the research above is on competitive swimmers training 15 to 25 hours a week. You are almost certainly not. Tate and colleagues studied 236 female swimmers aged 8 to 77 and found shoulder pain with disability in about a fifth of every age group, masters included — but what separated the swimmers with pain from those without, across multiple age groups, was greater swimming exposure, a previous traumatic injury, self-rated instability, and doing no other sport.

For an adult swimming three or four times a week, the realistic failure mode is not chronic overload. It is the spike: signing up for an event, coming back after a layoff, joining a masters squad and matching the session that everyone else has built up to over two years. That is the acute-to-chronic workload ratio in plain language, and it is entirely avoidable.

Changing your training without guessing

The practical rules are boring and they work.

• Know your normal. Keep a rolling four-week average of weekly distance. A week much bigger than that average is the week you are most exposed, and a common coaching guideline is to keep increases in the region of 10% — consensus rather than an evidence-based threshold, since the research has not established a cut-off.
• Add intensity or volume, not both. The week you introduce a hard threshold set is not the week to add a fourth session.
• Set your intervals from a measurement. The critical swim speed calculator gives you a threshold pace from two time trials, so your hard sets are hard relative to you. Training at somebody else's pace is one of the reliable ways to overreach.
• Watch stroke count, not just pace. Stroke count creeping up at the same pace across a set is fatigue arriving. It is also the moment technique degrades and the shoulder starts taking load it was not taking on rep one. The pace calculator and the SWOLF numbers together tell you whether you are getting faster or just thrashing.
• Build the volume before the speed. If you are coming back from a break, rebuilding endurance first and choosing a sane session length matters more than any drill.
• Do not spend every metre on freestyle. Backstroke moves the arm through a different arc, and mixing strokes spreads the repetitions across more tissue — coaching consensus rather than trial evidence, but it costs nothing. If you have avoided it because you cannot hold a lane, that is a solvable problem: swimming backstroke straight.

When the technique fixes do not work

Give a stroke change two or three weeks of consistent work. If pain is unchanged, or it is present at rest, at night, or when you lift your arm out of the water, technique is not your problem and no article can be.

Where to start this week

Film one length from the side. Note which phase of the stroke hurts and check the matching row in the table above. Take one fault, not four, and give it a fortnight of easy swimming with fist drill and single-arm work in the warm-up. Write down your weekly distance so you know what your normal is. And if you were about to add a session because a race is coming, add it next month instead.

The rest of the technique section covers the stroke mechanics underneath all of this, and none of it is worth swimming through pain to reach.