Posture guide
Rounded shoulders means your shoulder blades sit further forward around your ribcage, so your shoulders look hunched from the side. It's very common, it often comes with a forward head and a rounded upper back, and on its own it isn't a reliable cause of shoulder pain.
In trials, programmes built on strengthening the muscles around the shoulder blades improved measured shoulder position over 3 to 16 weeks, while stretching the chest on its own didn't lengthen the muscle or change posture. Here are the moves with 3D demos, what rows can and can't do, and when shoulder symptoms need a doctor.
Your shoulder blades sit further forward around the ribcage and tip forward, so your shoulders look hunched from the side. In one study, about 7 in 10 healthy adults had rounded shoulders. It often goes with a forward head and a rounded upper back, but each can occur alone.
Shoulder-blade position in people with shoulder impingement is inconsistent across studies, upper-back rounding doesn't differ between people with and without shoulder pain, and the length of the small chest muscle wasn't linked to shoulder pain. Sitting tall does let you lift your arm a little higher in the moment.
Across 18 trials, corrective exercise programmes improved forward-shoulder measures, though the trials were small and varied. Eight weeks in swimmers and 16 weeks in teenagers improved head and shoulder posture.
Six weeks of daily stretching didn't lengthen the small chest muscle (pectoralis minor), and adding chest stretches to a 12-week shoulder programme added nothing. Strengthening programmes do change upper-body posture.
Rows, reverse flys and Y-T-W raises are in the programmes that worked, and the push-up plus works the muscle that holds the shoulder blade against the ribs. But strength itself doesn't match where the shoulder blades sit at rest, so use them as a mix, not a single fix.
Rounded shoulders (forward shoulder posture, in clinical terms) means the shoulder blade sits further forward and outward around the ribcage (protraction), tips forward and turns inward.
Trials measure it from a side photo, as an angle or as how far the point of the shoulder (the acromion) sits forward. Clinicians also measure how far the back of the shoulder rests from the table when you lie on your back: this gives the same answer each time but is poor at spotting a short chest muscle. The length of the pectoralis minor can be measured between bony landmarks with a tape or calipers.
There's no validated normal value, so use any of these to track your own change, not to diagnose yourself.
The pectoralis minor runs from the upper ribs to a knob at the front of the shoulder blade (the coracoid). In healthy people, a shorter pec minor went with more upper-back rounding and an inward-turned shoulder blade. One study linked a short pec minor to shoulder-blade movement like that seen in impingement; another found no difference.
A review found the three can occur alone or in any combination, and in 160 adults a forward head wasn't linked to the curve of the upper back. "Upper crossed syndrome" is a label for all three together, explained as tight front muscles and weak back muscles: treat it as a description, not a diagnosis.
In a 1992 study of 88 healthy adults, 73% had a rounded right shoulder and 66% a rounded left one, judged against a plumb line. It's a common shape.
Source 2
In 10 studies of subacromial impingement (pain at the top of the shoulder when you lift your arm), shoulder-blade position was inconsistent: some found more forward tilt, others less, others no difference, and the authors said restoring an "ideal" shoulder-blade posture isn't supported.
Moderate evidence from 10 studies found no difference in upper-back rounding between people with and without shoulder pain. In 144 people, pec minor length wasn't linked to shoulder pain or disability, and in office workers rounded shoulders didn't differ between people with and without neck pain.
There's strong evidence that the arm lifts further when you sit upright than when you slouch. Taping 60 people with impingement into a more upright posture let them lift the arm higher straight away, though the pain itself was no less intense.
Across 18 trials, corrective exercise programmes reduced forward-shoulder measures by a large amount on average (moderate certainty), though the trials varied widely.
In 28 elite swimmers, 8 weeks of exercise reduced forward head and shoulder posture and increased shoulder strength. In 130 teenagers, 16 weeks of resistance and stretching exercise added to school physical education lessons improved head and shoulder angles, though shoulder pain didn't change. In 60 young women, adding lower-trapezius strengthening to chest stretching improved a lying test of rounded shoulders more than stretching alone over 3 weeks.
A review of 23 controlled trials found stretching programmes didn't change posture, while strengthening programmes did, most clearly in the upper back and neck (moderate certainty).
Source 12
A one-arm stretch lengthened the pec minor more than two other stretches, stretching with the arm out to the side at shoulder height or higher stretched it most, and a doorway stretch lengthened it straight away.
But 6 weeks of daily stretching didn't change its resting length, static stretching alone had no effect on its length across 6 trials with 263 people, and adding pec minor stretches to a 12-week exercise programme for shoulder pain gave no extra benefit in 80 people.
Prone Y and T raises work the middle and lower trapezius hardest. The push-up plus works the serratus anterior (the muscle that holds the shoulder blade against the ribs) hardest with the least upper trapezius. Band pull-aparts with palms up favour the lower trapezius. These studies measured muscle activity, not posture.
Older reviews found little reliable evidence that exercise corrects shoulder-blade posture, and strength didn't match resting shoulder-blade position. In 20 people, 6 weeks of stretching and strengthening made the upper trunk more upright, but where the shoulder blades sat at rest didn't change.
In 8 trials of 387 people with subacromial pain, shoulder-blade stabilising exercise reduced pain more than standard physiotherapy. Across 7 studies, treatments improved disability without changing shoulder-blade movement, so the benefit isn't explained by better posture.
Rows, reverse flys, face pulls, band pull-aparts and Y-T-W raises, 3 days a week with a rest day between. Our editorial starting point: Y-T-W raises, 5 to 8 of each letter for 2 rounds; band pull-aparts with palms up and face pulls, 2 to 3 sets of 12 to 15; band rows, 3 sets of 10 to 15; chest-supported dumbbell rows, 3 sets of 8 to 12.
The push-up plus (pushing your shoulder blades apart at the top of a push-up) works the serratus anterior with little upper trapezius. The wall version shifts more work to the upper trapezius, so move to a bench, then the floor. Our editorial starting point: 2 to 3 sets of 10, plus standing band external rotation, 2 to 3 sets of 12 to 15.
Source 14
If you like stretching, a one-arm doorway stretch with the elbow at shoulder height stretches the pec minor most. Our editorial starting point: 3 holds of 30 seconds each side. The effect is short-lived and hasn't been shown to change posture.
Trials ran 3 to 16 weeks; 8 weeks changed posture in swimmers and 16 weeks in teenagers. Expect weeks rather than days, and keep training to keep the change (our editorial suggestion).
All rounded-shoulder measures improved while 24 people with impingement wore a shoulder brace, but we found no study showing a lasting change, and a brace worn for 30 minutes of typing didn't change pain, fatigue or neck alignment. Kinesiology tape lengthened the pec minor in pooled trials, with no long-term posture data.
At a desk, follow the Health and Safety Executive's set-up (screen at eye level, keyboard just below elbow height) and take short breaks often. We found no study of sleeping positions or pillows for rounded shoulders.
Three days a week with a rest day between: Y-T-W raises, 5 to 8 of each letter for 2 rounds; band pull-aparts (palms up) and face pulls, 2 to 3 sets of 12 to 15; band rows, 3 sets of 10 to 15; dumbbell rows, 3 sets of 8 to 12; reverse flys, 2 to 3 sets of 10 to 12; wall angels, 8 slow slides (our editorial starting points).
3D demoProne Y-T-W raisesBeginner · No equipment
3D demoBand pull-apartIntermediate · A band
3D demoBand rowBeginner · A band + An anchor point
3D demoChest-supported dumbbell rowBeginner · A dumbbell + A bench or box
3D demoIncline reverse flyBeginner · A dumbbell + A bench or box
3D demoFace pullIntermediate · A band
3D demoWall angelBeginner · A wallPush-up plus on a bench before the floor, 2 to 3 sets of 10; shoulder-blade punches, 2 sets of 10 to 12; band external rotation, 2 to 3 sets of 12 to 15 (our editorial starting points).
For comfort, not correction: stretching on its own hasn't changed posture. Doorway stretch, 3 holds of 30 seconds each side with the elbow at shoulder height; elbows back, 2 holds of 20 to 30 seconds; thoracic rotations, 6 each side; open books, 8 to 10 each side (our editorial starting points).
The upper-back posture routine is the closest fit. The upper-body workout at home covers rows, face pulls and Y-T-W raises with a band; the back, pull-day and shoulder workouts do the same in a gym; the push-up progression includes the push-up plus.
A 12-minute routine for a stiff upper back and rounded shoulders: chin tucks, wall angels, Y-T-W raises and open books.
12 minBeginner8 movesA foam roller, A wallAn upper body workout at home with no gym: push-ups, band rows, pike push-ups, face pulls, close-grip push-ups and Y-T-W raises, with dumbbell options.
30 minIntermediate13 movesA band, An anchor pointA gym back workout: lat pulldowns, cable rows, dumbbell rows, face pulls, back extensions and shrugs.
45 minBeginner8 movesPull-up bar, A bandA pull day workout for back and biceps: pull-ups, barbell rows, chest-supported rows, face pulls and curls, with sets, reps and rest.
55 minIntermediate10 movesPull-up bar, A barbellA shoulder workout for all three heads: overhead press, lateral raises, reverse flyes, face pulls and rotations.
40 minIntermediate8 movesA band, A barbellWork up to full push-ups: wall, incline and knee push-ups, then the floor push-up and harder variations.
15 minBeginner8 movesA wall, A bench or boxGet emergency help for chest pain that feels tight or like squeezing, or that spreads to your arms, neck or jaw: pain from the heart can spread to the arm.
Source 37
Get advice soon (in the US, your doctor or urgent care) for any of these:
Source 38
See a GP if shoulder pain is getting worse, hasn't improved after 2 weeks, or makes it very difficult to move your arm or shoulder. The shoulder pain and shoulder impingement guides cover what helps the pain.
Source 38
Stretching programmes didn't change posture, 6 weeks of daily pec minor stretching didn't lengthen it, and adding chest stretches to a shoulder programme added nothing. Strengthening is what changed posture.
Shoulder-blade position in impingement is inconsistent, rounding doesn't differ with or without shoulder pain, and exercise helps pain without changing shoulder-blade movement.
Braces change your position while you wear them; we found no study showing a lasting change.
It's a descriptive label. Programmes based on it change posture measures but not pain or function consistently, and strength didn't match shoulder-blade position.
We found no trial of rows on their own, and strength didn't match where the shoulder blades sat at rest. Rows belong in a mix.
A review found no reliable evidence that exercise causes postural problems. Keep pressing, and balance it with pulling (our editorial suggestion).
Source 29
Measured shoulder position improves with exercise. Across 18 trials, corrective programmes improved forward-shoulder measures, though the trials were small, varied and short. Eight weeks in swimmers and 16 weeks in teenagers reduced forward head and shoulder posture.
Strength work for the muscles that draw the shoulder blades back and hold them against the ribs: rows, reverse flys, face pulls, band pull-aparts, Y-T-W raises and the push-up plus. Stretching programmes on their own didn't change posture in trials.
Rows are in most of the programmes that improved shoulder posture, and strengthening changed upper-body posture in a review. But we found no trial of rows on their own, and one 6-week stretch-and-strengthen programme made the upper back more upright without changing where the shoulder blades sat at rest. Use them as part of a mix.
It's optional. A one-arm doorway stretch lengthens it most, especially with the arm at shoulder height or higher, but 6 weeks of daily stretching didn't change its resting length. Adding pec stretches to a 12-week shoulder programme gave no extra benefit.
They're not a reliable cause. Reviews found no consistent shoulder-blade position in impingement, and no difference in upper-back rounding between people with and without shoulder pain. Sitting tall lets you lift your arm a little higher in the moment, and shoulder-blade exercises help pain, though not by changing shoulder-blade movement.
Trials ran 3 to 16 weeks: 8 weeks changed posture in swimmers and 16 weeks in teenagers. Expect weeks, not days, and keep training to keep the change.
Rounded-shoulder measures improve while you wear one. A brace worn for 30 minutes of typing didn't change pain, fatigue or neck alignment, and we found no study showing a lasting change.
Rounded shoulders is about where the shoulder blades sit; kyphosis is the curve of the upper-back bones. They often go together but can occur alone. If standing tall doesn't flatten your upper back, or the curve is new, see the kyphosis guide.
A label for a forward head, rounded shoulders and a rounded upper back together, explained as tight front muscles and weak back muscles. Programmes based on it improve posture measures but not reliably pain or function. Treat it as a description, not a diagnosis.
The same programmes treat both: in swimmers and teenagers, exercise improved head and shoulder posture together. Combine shoulder-blade strength with deep neck work such as the lying chin tuck, which the forward head posture guide covers.
A review found no reliable evidence that exercise causes postural problems. Keep pressing, and balance it with pulling work such as rows.
We found no study testing sleeping positions or pillows for rounded shoulders. Sleep in whatever position is comfortable.
Exercise education, not a diagnosis or treatment plan, and clinical review of these guides is pending. Get a curve or hunch that is new or getting worse checked, above all in a growing child or after a fall, and get urgent help for numbness or weakness in the legs or a change in bladder or bowel control. Health disclaimer.
