Shoulder external rotation: it is the movement that almost every programme for the shoulder puts first, usually accompanied by a clear promise: your external rotators are weak, strengthen them with a band and you will never hurt yourself again.
The part about the band is reasonable. The promise, much less so. The link between weak external rotators and injury has really been measured, and what emerged is narrower and more fragile than it is usually presented: it applies mainly to people who use their arm above their head in sport, the evidence is described as "limited" by the very authors who gathered it, and for the general population, people who go to the gym or sit all day, there are still no studies that have examined the question.
This does not mean throwing the exercise away. It means understanding what it is really for, what to expect, and where it is worth spending the rest of your training time. That is exactly what you will find below: the full technique, the progressions, and the evidence behind them with its stated limits.
⚠ Important: read this before continuing
The exercises and advice in this article are general guidance and may not suit your situation. Every person has a unique clinical history: what works for others could be counterproductive for you.
Want a professional's opinion? Start with the questionnaire: you sign up and see who will follow you, a physiotherapist registered with the Register (Albo) chosen for your case, whom you can write to in chat. If you would rather hear their voice, the first consultation is free and we look at your situation together.
First of all: when the shoulder needs assessing, not training
Most shoulder pain is benign and responds to movement. Some situations, however, are not a loading problem, and with exercise they get worse or you simply lose precious time. If you recognise yourself in one of these, get assessed before starting any programme:
- Arm or shoulder pain together with chest tightness, shortness of breath, cold sweating or nausea. It must be treated as an emergency, without waiting to see whether it passes.
- After a trauma or a fall: you cannot actively raise the arm, the shoulder looks deformed, or you feel the joint is "out of place".
- A hot, swollen and very painful shoulder, especially with fever or general malaise: it needs to be seen promptly.
- Night pain that does not change whatever position you adopt, particularly if it comes with weight loss, night sweats, or if you have a history of cancer.
- Tingling, loss of strength or persistent numbness in the hand, or pain running down the arm with the hand becoming clumsy: the problem may originate in the neck or along the course of the nerves, not in the rotator cuff.
- Sudden loss of strength when lifting the arm after a clean tear, perhaps while lifting a weight: this may be an acute tendon injury.
None of these things is resolved with a band. Shoulder external rotation exercises make sense when you know you are training a loading problem, not something else.
What this movement actually does
Shoulder external rotation is the action of rotating the forearm outwards while keeping the elbow bent: the movement you make to open a hinged door by pulling it towards you, or to bring your hand behind your head before throwing.
It is produced mainly by two muscles of the rotator cuff, the infraspinatus and the teres minor, with a contribution from the posterior deltoid. They are small muscles with short lever arms: in practice they bear loads far lower than the large motors of the shoulder, which is why in the gym they are trained with low resistance and controlled movements. It is a practical, common-sense choice, not a rule demonstrated by a study.
The real point is that the cuff is not only for rotating. It works as a control system: it keeps the head of the humerus centred in its socket while the large muscles (deltoid, pectoralis major, latissimus dorsi) generate the force of the movement. It is a stabilising task more than a powerful one, and this changes how it makes sense to train it.
What the studies say about the link with injuries
This is the part that is usually skipped. The claim "weak external rotators equals a shoulder that gets injured" has been tested prospectively, by measuring strength before the season and counting injuries afterwards, and the results need to be read in full.
A study of 138 young elite handball athletes (70 boys and 68 girls, average age 14.1 years) measured rotational strength, mobility and scapular dyskinesis before the season, then recorded overuse symptoms throughout the year. Reduced external rotation strength was a risk factor for overuse shoulder injury, both in absolute terms (OR 10.70; 95% CI 1.2-95.6; p = 0.034) and normalised (OR 1.2; 95% CI 1.0-1.4; p = 0.015), as was the ratio between external and internal rotation (OR 1.2; 95% CI 1.1-1.5; p = 0.012). It is worth looking at that enormous confidence interval on the first figure, from 1.2 to 95.6: the signal is there, but its size is anything but precise. In the same study, scapular dyskinesis was not associated with risk, and the average questionnaire response rate was 63%.
A systematic review then pooled six prospective studies for a total of 523 trained overhead athletes, followed for a season. Those who did not get injured had, before the season, significantly greater eccentric strength in both internal rotation (p < 0.01) and external rotation (p < 0.05). External rotation mobility below 106° was a risk factor (OR 1.12; p < 0.001). The imbalance between external and internal rotation appeared as a risk factor in two studies, with a relative risk of 2.57 (p < 0.05) in one of them.
The authors' conclusions, however, are measured: the evidence supporting weakness in external and internal rotation is limited, that supporting the imbalance between the two is very limited. And there is one sentence that changes the picture considerably for most readers: no study has investigated the general population on this topic.
A more recent review of overhead sports confirms the direction with the same caution: after excluding studies of insufficient methodological quality, nine papers remain, only one of high quality and eight of acceptable quality, and it finds moderate evidence for three modifiable factors associated with the risk of shoulder injury: rotational strength, scapular dyskinesis, and the presence of a prevention programme. In the same review the evidence is moderate also for two factors you cannot change: playing position and sex. And there is a point on which the two reviews disagree: on rotational mobility, this more recent review finds moderate evidence of no association with risk, whereas the earlier one listed it among the risk factors. On strength, by contrast, the direction is the same.
To sum up honestly: if you play handball, volleyball, tennis or baseball, external rotation strength is a documented risk factor, with limited but consistent evidence. If you are someone who goes to the gym three times a week or spends the day at a computer, the same claim is sold to you with the same confidence, but it has never been tested in your population.
Does training it really prevent injuries?
Knowing that a factor is associated with risk does not yet mean that changing it reduces injuries. These are two different questions, and the second has been studied separately.
The most solid study is a cluster randomised trial of 45 elite handball teams, 660 players, followed through a seven-month competitive season. The intervention group carried out a prevention programme three times a week within the warm-up, led by coaches and captains, which included internal rotation mobility, external rotation strength, strength of the scapular muscles, kinetic chain work, and thoracic mobility.
The result: average prevalence of shoulder problems of 17% (95% CI 16-19) in the intervention group against 23% (95% CI 21-26) in controls, with a 28% lower risk (OR 0.72; 95% CI 0.52-0.98; p = 0.038). A real effect, obtained with a few minutes of warm-up.
But the details matter. For substantial problems, those that really reduce participation and performance, prevalence was 5% against 8%, with a 22% reduction that did not reach statistical significance (OR 0.78; 95% CI 0.53-1.16; p = 0.23). And the programme that worked was not "external rotation with a band": it was a multi-component package. We do not know which piece did the work.
Looking more broadly, a systematic review of shoulder prevention programmes in overhead athletes found seven studies. Two reported a reduction in injuries (OR 0.72; 95% CI 0.52-0.98 and OR 0.22; 95% CI 0.06-0.75) and one a benefit from stretching the posterior capsule (HR 0.36; 95% CI 0.13-0.95). Only three studies showed a favourable effect, and only one was at low risk of bias. The authors' conclusion is clear: at present, no conclusions can be drawn about the effectiveness of shoulder prevention programmes in the overhead athlete.
So: it probably helps, the measured effect is modest, and the certainty is low. Anyone who promises you a shoulder that is injury-proof in five minutes a day is going well beyond the data.
The impingement story, and why to take it with a pinch of salt
There is a narrative that always accompanies these exercises: weak external rotators let the head of the humerus ride up, the space under the acromion shrinks, the tendons get pinched, and impingement develops. Strengthening shoulder external rotation lowers the humeral head again and the problem goes away.
It is an elegant model, and it is the reasoning on which subacromial decompression surgery was built for decades: if the problem is mechanical, you remove the excess bone and the pain goes away.
That reasoning has been put to the test. A multicentre British placebo-controlled trial randomised 313 patients into three groups: arthroscopic decompression, exploratory arthroscopy alone (a surgical placebo, with no removal of bone or soft tissue), or no treatment, which in the study meant a single review appointment with the specialist at three months, without surgery. It is worth knowing who these patients were: they had subacromial pain for at least three months, an intact cuff, and had already completed a non-surgical pathway that included therapeutic exercise and at least one cortisone injection. They were therefore not people new to exercise: they were people for whom exercise had not been enough.
At six months, the Oxford Shoulder Score was 32.7 (SD 11.6) after decompression and 34.2 (SD 9.2) after arthroscopy alone: a mean difference of -1.3 points (95% CI -3.9 to 1.3; p = 0.3141). Both surgical groups did slightly better than the no-treatment group (29.4; SD 11.9), but the authors define those differences as not clinically important.
If physically removing the presumed obstacle does no better than a fake incision, the story of mechanical compression as the cause of pain does not hold up as it was told. And as a consequence, the explanation "your band recreates space under the acromion" does not hold up either.
This does not take away the value of the exercise. It does, however, shift the explanation: it is unlikely that the benefit of the band lies in "recreating space", and more likely that it comes from something else: tendon tolerance to load, motor control, general strength, confidence in movement. None of these mechanisms has been isolated and demonstrated in a study of external rotation: they are reasonable hypotheses, not conclusions. It is a more modest motivation, but an honest one.
What still stands, in practice
Stripping away the promises, what remains in favour of shoulder external rotation is still enough to justify five minutes in the warm-up:
- It costs little, and the side effects, when they occur, are mild. Not "zero risk": in the only high-quality trial collected by Cochrane, side effects with manual therapy and exercise were more frequent than with placebo (31% against 8%), but these were short-lived pain after the session.
- It is a documented risk factor in overhead sports, with limited evidence but pointing in the same direction across several independent studies.
- It is part of the programmes that have shown an effect, although within a broader package that included mobility and the kinetic chain.
- It is part of the rehabilitation protocols for painful shoulder that have reduced recourse to surgery, although, as you will see later, against a credible placebo the differences remain small.
And what does not still stand is the idea that it is the missing piece, that it rebalances the mechanics of the joint, or that it can replace managing the overall training load. If your shoulder hurts in the gym, the first lever is not to add external rotations: it is to look at the volume, frequency and progression of what you are already doing. Shoulder pain in the gym arises much more often from there.
An important tip: before trying these exercises, make sure you have received a correct diagnosis. People often think they have one problem when the real cause of the pain is another. Doing the wrong exercises can make things worse instead of better.
The basic exercises
Three variants, all doable with a band or a cable. Precision matters more than load here.
External rotation with the elbow fixed. Standing or seated, elbow bent to 90 degrees and against your side, band anchored to the side at elbow height. Rotate the forearm outwards keeping the elbow still against your body, then return slowly. It is the reference exercise for shoulder external rotation, and the one to learn first. A rolled towel between elbow and side helps you tell whether you are cheating.
External rotation lying on your side. Lie on the side opposite to the one you are training, the elbow of the upper arm bent to 90 degrees and resting on your side, a light dumbbell in hand. Rotate the forearm upwards and lower it back down. Here gravity provides the resistance: it is harder to compensate, and for this reason it is an excellent control exercise.
External rotation at 90 degrees of abduction. Arm abducted to 90 degrees with the elbow bent, band in front of you: rotate the forearm upwards against the resistance. It reproduces the position of throwing and smashing, so it is the most specific variant for those who play overhead sports, and also the most demanding. It makes sense when the first two are clean and pain-free.
Progressions
When the basic versions become easy, the next step is not to load the same movement more heavily, but to integrate external rotation into the work of the scapula and the shoulder girdle.
Face pulls. Cable or band at face height: you pull towards the face, opening the arms and externally rotating the shoulders at the end of the movement. It combines external rotation and work for the scapular muscles in a single action, and it is the easiest to add to a gym programme.
Cuban rotation. You start with the elbows high and the forearms pointing downwards, then rotate externally, bringing the forearms upwards. It requires coordination and mobility: with too much load it turns into a shrug.
YTWL. Prone on a bench or standing bent forward, you trace the letters Y, T, W and L with your arms. It is not an external rotation exercise in the strict sense, but it places it within a complete work for the cuff and scapula, and it is useful for seeing where you really have a gap.
If you want to see how these movements fit into a complete programme for the scapular girdle, you will find everything in strengthening exercises for the shoulder.
How to fit them into the week
On the dosages of shoulder external rotation there are very precise numbers with no study behind them. It is better to start from how the exercise was delivered in the works that showed an effect.
If your goal is prevention in overhead sport, the handball trial programme was carried out three times a week as part of the warm-up, led by the coach. Not at the end of the session, not on rest days: before training, consistently, for the whole season. It is a replicable model, and regularity seems to matter more than the specific choice of exercise.
If you are working on a painful shoulder, in the randomised study with the best results, patients did the exercises at home once or twice a day for 12 weeks, with five or six individual guided sessions spread over the same period. The home volume was high, but it was not left to chance: there was periodic supervision that adjusted the approach.
For the rest, the practical criterion is quality: stop the set when the elbow starts to drift away from the side or the shoulder rises towards the ear, not when you reach a predetermined number. And if pain increases during the exercise or is still higher the next day, you have used too much resistance.
The mistakes I see most often
The elbow moving away from the side. In the basic version it is the number one error: as soon as the elbow detaches, the movement becomes a disguised abduction and the external rotators stop being the protagonists.
Too much resistance. It is the error that comes from thinking of this as a strength exercise. With a stiff band you compensate with the trunk, the trapezius and momentum, and you train everything except what you intended.
Going fast. The controlled return, the phase in which the muscle brakes as it lengthens, is what was prescribed in the randomised study that reduced recourse to surgery: eccentric exercises for the cuff. But be careful not to read more into this than there is: the Cochrane review concludes that one type of exercise has rarely been shown to be superior to another, so eccentric work is not proven better in itself. It remains a good practical reason to slow down: if you let the band snap back, for half of the movement you are not working anything.
Doing it only when it hurts. As a risk factor, strength is measured before the season. As an intervention, the programmes that worked lasted months.
If you are recovering from a shoulder problem
The ground here is different, and in part more solid. In the treatment of persistent painful shoulder, cuff exercise is the backbone of the protocols, not an accessory.
A randomised study of 102 patients with persistent subacromial impingement syndrome lasting more than six months, in whom previous conservative treatment had failed, compared a specific strategy (eccentric exercises for the cuff and concentric/eccentric exercises for the scapular stabilisers, combined with manual mobilisation performed by the physiotherapist) against generic neck and shoulder exercises. At 12 weeks, the Constant-Murley score had improved by 24 points (95% CI 19-28) in the specific group against 9 points (5-13) in the control group. Above all, by the end of the study's 12 weeks, 20% of patients in the specific group (10 out of 51) had chosen surgery, compared with 63% of the control group (29 out of 46).
Three necessary qualifications, because they make the difference between citing a study and using it well. The first: that population was chronic and selected within orthopaedics, with pain lasting more than six months, recruited by orthopaedic specialists, with previous conservative treatment already failed, not people with a two-week niggle. The second: the intervention was not just do-it-yourself exercise: it included manual mobilisation performed by the physiotherapist. The third, which is often told badly: the five or six individual guided sessions were received by both groups, so they do not explain the difference; what distinguished the two arms was the type of exercise and the manual therapy. The measured difference is between two ways of doing physiotherapy, not between doing physiotherapy and not doing it.
The other side of the coin should also be said. A Cochrane review of 60 studies and 3,620 participants found only one high-quality trial comparing manual therapy and exercise with a credible placebo (inactive ultrasound), in 120 people with chronic rotator cuff disease: at 22 weeks, pain improved by 24.8 points out of 100 against 17.3 for placebo (adjusted mean difference 6.8; 95% CI -0.70 to 14.30), and function by 22.4 against 15.6 (difference 7.1; 95% CI 0.30-13.90), differences the authors judge not clinically important. And a conclusion that deserves repeating: one type of exercise has rarely been shown to be superior to another.
Taking it all together: doing shoulder exercise is worthwhile and can keep you out of the operating theatre, but the exact choice of exercise matters less than we think. Progression, dosage and consistency matter more. This also holds after major events such as a shoulder dislocation, where the pathway has to be built in stages.
For those who work at a desk: what to expect
The popular version says that shoulders rolled forward shorten the internal rotators, and therefore that it is enough to stretch the "short" ones and strengthen the "weak" ones to stand up straight. This recipe, stretch the short, strengthen the weak, was tested by a meta-analysis of 23 studies and 969 healthy people. Stretching did not change posture, either acutely (d = 0.01; p = 0.97) or chronically (d = -0.19; p = 0.16). Strengthening weak muscles, on the other hand, produced large improvements (d = -0.83; p = 0.01) and was superior to stretching (d = 0.81; p = 0.004), with a large effect in the thoracic and cervical region (d = -1.04; p = 0.005) and none at the lumbar and lumbopelvic level (d = -0.23; p = 0.25).
Here, however, honesty is needed about what that study did and did not measure, because it is the point at which this source is most often stretched beyond its evidence. It did not study shoulder external rotation: it studied strengthening in general, and the outcomes were measures of spinal and pelvic posture: thoracic kyphosis, head tilt, lumbar lordosis, pelvic tilt. It contains neither the rolled-forward shoulder nor the scapula. And the outcome was posture, not pain, in healthy people. That strengthening external rotation straightens the shoulders, and that more open shoulders mean less pain, are two extra steps that this meta-analysis does not take.
What can be drawn from it, without forcing it: the "stretch the tight pectorals" part of the popular recipe did not hold up under testing, while strength work is the component that showed something. Use it as a reason to train, not as a diagnosis of the cause of your discomfort.
Why telerehabilitation works well for this topic
Shoulder external rotation exercises are easy to perform badly and hard to correct on your own: the elbow drifting, the shoulder rising, the wrong resistance are errors you do not feel while doing them. They are, however, perfectly visible on a video call, because there is no need to touch anything to spot them.
With Ri-Hub, a physiotherapist watches how you perform the exercise, measures where you start, sets the resistance on the control you have today and updates it over time, at 29,90€ per session. If, as the Cochrane review suggests, one exercise rarely beats another, what remains to be treated is how you do it and for how long you keep it up: being followed for weeks from home removes exactly the friction that makes people give up on day ten.
If you play a sport overhead, it makes sense to build the work into the season and not at the end of it: in pathways for those who play volleyball, rotational strengthening fits into the warm-up, as in the protocols that have shown an effect.
In summary
Shoulder external rotation deserves a place in your week: it costs little, the drawbacks are mild, and it is one of three modifiable factors for which the most recent review of overhead sports finds moderate evidence, together with scapular dyskinesis and simply following a prevention programme. It deserves, however, honest expectations. The link with injuries is documented in a limited way and almost only in overhead sport; prevention programmes show modest effects with low certainty; and the mechanical explanation of impingement that usually accompanies these exercises has not survived the surgical placebo test.
The practical message is less spectacular but more useful: do the rotational work regularly and with good technique, and put the remaining energy where it matters most, in managing how much and how you load your shoulder in the rest of your training.
You might also be interested in:
Muscle Strain: Symptoms, Treatment and Recovery Times Runner Mobility: An Essential Routine for Running Better Sports Injury Prevention: The Complete Guide for Athletes
Stay in touch with us Join our dedicated Facebook group: we answer all patients, share practical advice and clear up the most common questions.
Sources
- Decreased external rotation strength is a risk factor for overuse shoulder injury in youth elite handball athletes, Knee Surgery, Sports Traumatology, Arthroscopy, 2020
- Are muscle weakness and stiffness risk factors of the development of rotator cuff tendinopathy in overhead athletes: a systematic review, Therapeutic Advances in Chronic Disease, 2021
- Risk factors and prevention strategies for shoulder injuries in overhead sports: an updated systematic review, Journal of Experimental Orthopaedics, 2022
- Preventing overuse shoulder injuries among throwing athletes: a cluster-randomised controlled trial in 660 elite handball players, British Journal of Sports Medicine, 2017
- Effectiveness of shoulder injury prevention programs in an overhead athletic population: A systematic review, Physical Therapy in Sport, 2021
- Arthroscopic subacromial decompression for subacromial shoulder pain (CSAW): a multicentre, pragmatic, parallel group, placebo-controlled, three-group, randomised surgical trial, The Lancet, 2018
- Effect of specific exercise strategy on need for surgery in patients with subacromial impingement syndrome: randomised controlled study, BMJ, 2012
- Manual therapy and exercise for rotator cuff disease, Cochrane Database of Systematic Reviews, 2016
- Effects of Stretching or Strengthening Exercise on Spinal and Lumbopelvic Posture: A Systematic Review with Meta-Analysis, Sports Medicine - Open, 2024
Last check of the sources: August 2026.
Need a consultation?
Start with the Ri-Hub questionnaire: a few questions about your case, you sign up, and you see who will follow you: a physiotherapist registered with the Register (Albo), whom you can write to in chat whenever you like. If you would rather talk it through by voice, the first consultation is free, lasts twenty minutes and commits you to nothing: they arrive having already read what you wrote, and together you decide whether and how to continue.
