Yoga for rounded shoulders: the upper back and chest, poses to try, desk habits, what the evidence shows, and a class

Common, and usually harmless

In 88 healthy adults, 66% to 73% had a rounded shoulder, and in 275 teenagers 58% had forward-set shoulders. How rounded people were wasn't related to how bad or how frequent their pain was.

Sources 1, 2

Only loosely linked to pain

Upper-back rounding didn't differ between people with and without shoulder pain. Adults with neck pain hold the head a few degrees further forward in snapshot studies, but sitting posture at 17 didn't predict neck pain at 22, and bending the neck over a phone didn't predict new neck pain over a year.

Sources 3, 4, 5, 6

Strengthening changed posture in trials; stretching didn't

Across 23 trials, stretching had no effect on posture while strengthening improved the upper back and neck. Across 28 trials of corrective exercise, mostly 6 to 12 weeks of three sessions a week, head and shoulder angles improved, but pain and function didn't consistently.

Sources 7, 8

No yoga trial has tested rounded shoulders

Our searches found no trial. A registered trial comparing yoga with Pilates finished in 2023 with no results posted, and a 2025 study of cyclic yoga in 42 women reported a difference in rounded shoulders without giving any numbers.

Sources 9, 10, 11

Choose the poses that build the upper back

Our reading of one lab study of muscle activity in 20 women: locust with the arms reaching forward worked the middle and lower trapezius hardest of 15 poses. That measured muscle activity, not posture, so pair the back lifts with prone raises, wall angels and rows, and keep chest stretches for comfort.

Sources 12, 7

A new hunch needs a GP, not a posture class

A new or increasing hunch or height loss in an older adult can be a spinal fracture, and a stiff curve in a teenager that doesn't straighten when standing tall isn't a slouch. The warning signs are below.

Sources 13, 14

Forward-set shoulder blades

The shoulder blades sit further forward around the ribcage, so the shoulders look hunched from the side. They often come with a head that pokes forward and a rounder upper back, but "Forward head posture, rounded shoulders, and increased thoracic kyphosis can exist alone or in any combination", and in 160 people aged 17 to 83, "a forward head was not associated with increased thoracic curvature".

Sources 15, 16

Upper crossed syndrome is a label

It names a forward head, rounded shoulders and a rounded upper back together, explained as tight muscles at the front and weak ones at the back. Trials use it to choose who takes part; treat it as a description, not a diagnosis (our reading).

Source 8

How common

In 88 healthy adults aged 20 to 50, a forward head was found in 66%, a rounded right shoulder in 73% and a rounded left shoulder in 66%, and "No relationship was found between the severity of postural abnormality and the severity and frequency of pain". In 275 teenagers, 68% had a forward head and 58% forward-set shoulders. In healthy women aged 65 and over, the head sat further forward in each older age group: a normal change with age.

Sources 1, 2, 17

How it's measured

A phone app measuring the head-neck angle was repeatable, but "The minimum detectable change was 4.96° for intra-rater and 5.52° for inter-rater reliability", so smaller changes can be measurement noise. Photographs have the strongest evidence among methods that don't use X-rays. Tape and wall measures of the shoulders give the same answer when repeated but haven't been checked against X-rays. A mirror isn't a measurement.

Sources 18, 19, 20

A rounded upper back is a different question

A slouch is flexible and straightens when you stand tall. Scheuermann's kyphosis is a stiff curve with wedge-shaped bones in the spine that "will not passively correct with extension", and in older adults a growing curve can come with spinal fractures. Our yoga for kyphosis guide covers the curve.

Sources 13, 21, 14

Shoulder pain: no consistent linkStrong evidence

"There is a moderate level of evidence of no significant difference in thoracic kyphosis between groups with and without shoulder pain" (10 studies). In 56 volleyball players, forward shoulder angle didn't differ between those with and without shoulder pain.

Sources 3, 22

Neck pain: a small difference in snapshot studiesMixed evidence

In 10 snapshot studies, adults with neck pain had a slightly more forward head (about 5 degrees on average), but teenagers didn't, and age confused the link; a 2023 review of 26 studies found about 3 degrees on one measure and no difference on two others. Differences this size are near the error of the measuring tools, and snapshot studies can't show which came first. In 1,108 Australian 17-year-olds, "There was no significant difference in the odds of neck pain or headache across the clusters" of posture.

Sources 4, 23, 24, 18

Posture didn't predict later neck painEmerging evidence

Followed from 17 to 22, sitting posture didn't predict persistent neck pain, and young women who sat slumped had lower odds than those who sat upright; the authors wrote that "generic public health messages to sit up straight to prevent neck pain needs rethinking". In 396 adults followed for a year, neck bending while texting didn't raise the odds of new neck pain, while poor sleep and too little physical activity did. On X-ray in people over 45, the neck curve didn't differ with and without neck pain.

Sources 5, 6, 25

What sitting tall does changeStrong evidence

You can lift your arm further when sitting upright than when slumped, in the moment. Taping people with shoulder impingement into a more upright posture raised the point where the arm hurt, but not how much it hurt, and in pain-free people a forward head and rounded shoulders changed how the shoulder blades moved. These are short-term mechanics, not proof that posture causes pain.

Sources 3, 26, 27

Why the idea lasts

A 2019 viewpoint notes that "Despite the absence of strong evidence to support these common beliefs, a large posture industry has flourished".

Source 28

Strengthening changes posture; stretching doesn'tStrong evidence

Across 23 controlled trials in 969 people, stretching programmes had no effect on posture, while strengthening was "effective in the thoracic and cervical spine" (moderate certainty). A 2025 consensus of 20 stretching researchers agreed that stretch training doesn't improve posture.

Sources 7, 29

Corrective programmes: posture improved, pain didn't consistentlyEmerging evidence

Across 28 trials (901 people), corrective exercise improved forward head, forward shoulder and upper-back measures, with results that varied a lot between trials (rated moderate certainty for posture), but neck pain didn't change significantly and "the effects on muscle activation, function, balance, and pain were inconclusive". Most trials were small, and an earlier review relied partly on studies found only through Google Scholar; another found most of 22 trials at high risk of bias.

Sources 8, 30, 31

People with neck pain: about 3 degreesMixed evidence

In 10 trials (550 people), 3 to 8 weeks of exercise improved the head-neck angle by 3.4 degrees (moderate certainty) and pain by about 1.4 points out of 10 (low certainty). A second 2026 review of 13 trials rated its evidence low to very low. A 3-degree average is less than the 5-degree change one phone app can reliably detect.

Sources 32, 33, 18

Single trials in teenagers and swimmersEmerging evidence

In 130 teenagers, 16 and 32 weeks of strength and stretching added to PE lessons improved head and shoulder angles, but "For the shoulder pain scores in both groups, there were no significant changes". The posture gains were still there 16 weeks after the programme stopped: the authors found the break "was not sufficient to reduce the overall training effects". In 28 elite swimmers, 8 weeks of exercise reduced forward head and shoulder measures.

Sources 34, 35

No yoga trial for rounded shouldersStrong evidence

Our PubMed search for yoga with rounded or forward shoulders, forward head posture, the head-neck angle or upper crossed syndrome found 2 records, neither a trial. A registered trial in 30 women compared 6 Pilates exercises with 6 yoga poses (warrior I, triangle, child's pose, cow face, bound angle and half bow); it was completed in June 2023 with no results posted. A study of cyclic yoga in 42 middle-aged women reported a difference in rounded shoulders against a control group, but its summary gives no numbers, measuring method or details of how women were allocated, so it can't be weighed.

Sources 9, 10, 11

The yoga kyphosis trialEmerging evidence

In 118 adults aged 60 and over with a rounded upper back, hour-long classes 3 times a week for 24 weeks didn't change the main measure of the curve, while a secondary measure improved by 4.4% (under 2 degrees). Half the yoga group came to at least 80% of classes, and the authors called it "developmental, not definitive". A follow-up of 21 people from the yoga group, with no comparison group, found the shoulder blades sat 2.9% less far forward; it appeared in a journal from a publisher widely criticised as predatory.

Sources 36, 37

Which poses work the upper-back musclesEmerging evidence

In 20 women with yoga experience, "locust arms forward" produced the highest activity of 15 poses in the trapezius muscles between and below the shoulder blades, while dancer's pose, reverse tabletop and warrior II produced little. Outside yoga, the push-up plus works the serratus anterior hardest. These studies measured muscle activity, not posture: choosing these poses for posture is our reading, alongside the strengthening trials above.

Sources 12, 38

Yoga for neck and shoulder achesMixed evidence

For long-lasting neck pain, yoga eased pain and disability more than usual care in 3 small trials (188 people), short term only, and a review of 10 trials drew "a very cautious conclusion". In 150 office workers, hatha yoga was "tentatively found to be non-inferior to stretching exercise". For office workers, strengthening the neck, shoulder and shoulder-blade muscles eased neck pain at low certainty. None of these measured posture; our yoga for neck pain guide covers the pain.

Sources 39, 40, 41, 42

Move often

The Health and Safety Executive (HSE): "Take short breaks often, rather than longer ones less often", and breaks "should allow users to get up from their workstations and move around, or at least stretch and change posture". The Chartered Society of Physiotherapy: "There is no perfect posture or right or wrong desk set up", and "The key is to keep moving. Your best posture is your next posture".

Sources 43, 44

Set-up helps comfort more than painEmerging evidence

The HSE suggests the "Top of screen level with eyes, about an arm's length away". In Cochrane reviews, workstation adjustment and sit-stand desks "did not have an effect on upper limb pain or discomfort" in single trials, extra breaks reduced neck discomfort a little at very low quality, and ergonomic changes "may result in little to no difference in pain". In office workers followed over time, low task variety and a keyboard close to the body were among the risk factors for new neck pain.

Sources 45, 46, 47, 48

A few poses at your desk

Chin tucks, upper-back turns, the seated upper-back lift and a few breaths with your arms reaching up fit a short break. Our desk yoga guide has a 10-minute routine (our suggestion).

Source 43

Two or three times a week

Our editorial starting point: the 25-minute yoga for posture class 2 or 3 times a week, with the upper back routine on other days, for at least 8 weeks. Most corrective programmes ran three sessions a week. The UK osteoporosis consensus says "Yoga and Pilates and similar exercise programmes should be considered to help with posture and pain", and suggests repeating back-strengthening exercises 3 to 5 times, held for 3 to 5 seconds, at least twice a week.

Sources 8, 49

Keep some of it going

Trials ran from 3 weeks to 8 months and changes were a few degrees. In the teenage trial the gains held for 16 weeks after the programme stopped, but no trial has followed people for longer, so keep some of the work in your week, as with any strength (our suggestion).

Sources 34, 32

Chest stretches are for comfort

Stretching on its own didn't change posture in the trials. Keep the chest stretches you enjoy, in the middle of your range, and pair them with upper-back strength.

Source 7

Call 999 or go to A&E

Call 999: shoulder or arm pain with chest pain that feels tight or squeezing, or chest pain that spreads to your arms, neck or jaw, especially with shortness of breath, sweating or feeling sick. It could be a heart attack.

Call 999 or go to A&E: new problems walking, loss of bladder or bowel control, or a sudden loss of coordination, such as difficulty doing up buttons. These can be signs of pressure on the spinal cord.

Sources 50, 51, 52

Get help the same day

Same day (urgent GP appointment or NHS 111): severe neck pain that starts suddenly or gets worse quickly; heaviness or weakness in the arms or legs; pins and needles in an arm along with neck or arm pain; or pain between the shoulder blades.

Same day: clumsiness in your hands, or numbness, pins and needles or weakness in one or both arms, that is getting worse; problems with balance or walking since the neck pain started; or neck pain with a severe headache.

Same day (GP or NHS 111): sudden or very bad shoulder pain; you cannot move your arm; the arm or shoulder is badly swollen or has changed shape; pins and needles that do not go away; no feeling in the arm; the arm is hot or cold to the touch; pain that started after an injury such as a fall; or severe pain in both shoulders.

Same day (GP or NHS 111): the shoulder or arm is hot, swollen or red, or you have a high temperature (although not everyone has one) or feel unwell. A joint infection needs same-day care.

Sources 52, 50

A new hunch, height loss or sudden back pain

• A new or increasing hunch, or height loss.
• New back pain after a minor fall, bend, lift or cough: it can be a spinal fracture from osteoporosis, and not every spinal fracture hurts.
• Nerve problems with a spinal fracture need urgent specialist care.

Sources 53, 14, 54

A child, teenager or young adult

See a GP if you think your child might have kyphosis. A curve can get worse until growth stops, and a stiff, sharp or painful curve that doesn't straighten when standing tall isn't a simple slouch. Rarely, a curve that keeps growing causes weakness, numbness or tingling in the legs, bladder or bowel changes, or difficulty breathing.

Back pain and stiffness that persists in a teenager or young adult. Ankylosing spondylitis tends to start at that age and can affect the whole spine.

Sources 13, 55

See a GP or physiotherapist

• Neck pain or stiffness that has not eased after a few weeks, or after 8 weeks of looking after it yourself; painkillers are not helping; or you are worried. In many areas you can refer yourself to NHS physiotherapy.
• Neck pain with feeling generally unwell, or that started while you were ill with another condition such as rheumatoid arthritis or cancer; or neck pain with blurred vision, ringing in the ears or dizziness.
• Severe pain that is constant or disturbs your sleep, a fever or unexplained weight loss, or a history of cancer, inflammatory arthritis, tuberculosis (TB), a weakened immune system, neck surgery or a violent injury. Serious causes are very rare, but these need checking.
• Pain at the top of the back, between the shoulders, rather than in the lower back.

Sources 52, 53

During yoga

NHS inform's neck exercise advice: rate your pain from 0 to 10. During exercise, 0 to 3 is minimal and 4 to 5 is acceptable; 6 or more is too much. If it goes higher, do fewer repetitions, move more slowly or rest longer between movements.

Exercise should not make your neck worse overall. Soreness from a new exercise should ease quickly, and your neck should be no worse the next morning. Stop any exercise that makes your symptoms worse or causes new pain.

Arm symptoms are the exception. If an exercise sends pain, tingling or numbness further down your arm, or the extra symptoms last more than an hour afterwards, make the movement smaller or leave it out. (A hospital leaflet on nerve glides allows a short-lived rise in symptoms but not one lasting over an hour; the rest of this rule is this site's suggestion.)

Stop and get checked for dizziness, double vision, slurred speech, trouble swallowing or any of the warning signs above.

A joint pain that is sharp, gets worse during a session, lingers into the next day or changes how you move is a reason to use an easier version, and to see a GP or physiotherapist if it keeps happening (the rule in our plans). Muscle soreness for a day or two after something new is normal. In a hot class, stop if you feel dizzy, sick or unwell; our hot yoga guide lists the signs of heat illness.

You can say no to any hands-on adjustment, and a teacher registered with Yoga Alliance must get your explicit consent each time.

Sources 56, 57, 58, 59

Chest pain or fainting while exercising: call 999

Call 999 if:

• You get sudden pain or discomfort in your chest that doesn't go away; it can feel like squeezing or pressure, burning or indigestion.
• The pain spreads to your left or right arm, or your neck, jaw, stomach or back.
• You have chest pain and feel sweaty, sick, light-headed or short of breath.
• Someone faints while exercising, can't be woken within 1 minute, or has chest pain or a pounding, fluttering or irregular heartbeat.

Sources 60, 61

Unusual breathlessness

Call 999 or go to A&E (accident and emergency) straight away for severe difficulty breathing (gasping, choking or not able to get words out), a chest that feels tight or heavy, pain that spreads to your arms, back, neck and jaw, lips or skin turning very pale, blue or grey, or sudden confusion. If you're more short of breath than usual, or your heart is racing, fluttering or skipping beats, contact NHS 111.

Source 62

Head-down poses, bones and pregnancy

With glaucoma, or high blood pressure that isn't under control, skip headstands, shoulder stands and long head-down holds. With osteoporosis, swap deep forward folds for a flat-back hinge from the hips, and twist smoothly. In pregnancy, our yoga in pregnancy guide covers the changes. NCCIH advises beginners to avoid headstands, shoulder stands, the lotus position and forceful breathing.

Sources 63, 58

Before you start

The NHS advises getting advice from a healthcare professional first if you have a health problem, an injury or symptoms, or have recently had a heart attack or an operation.

"Rounded shoulders cause neck and shoulder pain"Strong evidence

Upper-back rounding didn't differ in people with and without shoulder pain, the severity of posture wasn't related to pain in healthy adults, and posture at 17 didn't predict neck pain at 22.

Sources 3, 1, 5

"Open your chest and your shoulders will go back"Strong evidence

Stretching programmes didn't change posture in 23 trials, and stretching researchers agree. Strengthening the upper back is what changed shoulder angles.

Sources 7, 29

"Yoga will fix rounded shoulders"Strong evidence

No yoga trial has tested it. Strength programmes changed measured posture a little, but pain and function didn't consistently improve.

Sources 9, 8

"There's one perfect posture to hold"Emerging evidence

Young women who sat slumped at 17 had lower odds of neck pain at 22 than those who sat upright. The Chartered Society of Physiotherapy says "There is no perfect posture or right or wrong desk set up".

Sources 5, 44

"A few weeks of yoga will pull your shoulders back for good"Emerging evidence

Trials ran from 3 weeks to 8 months and changes were a few degrees. Gains held for 16 weeks after stopping in one trial of teenagers, and none followed people for longer.

Sources 32, 34

"Deep backbends and handstands fix a hunch"Mixed evidence

They weren't part of the yoga kyphosis trial, which used gentle back lifts and standing poses, and some load the neck or are head-down. The poses that work the upper-back muscles hardest are gentle face-down lifts.

Sources 36, 12