Upper back tightness is the commonest complaint a screen produces, and the usual explanation for it is backwards. The muscles between your shoulder blades are not short. They are long, lightly loaded, and being asked to hold a position for six hours without a break.
Stretching a muscle that is tired from holding is treating the sensation. It works for about twenty minutes, which is roughly how long the relief lasts, and that is the clue.
The short answer
- The ache between the shoulder blades is usually a load complaint, not a length complaint.
- Those muscles sit long in a slumped position, not short. Stretching a long muscle is an odd prescription.
- There is a two-minute check that separates a range restriction from a fatigue ache. It is below.
- How stiff a back feels does not match how stiff it measures. That is the single most useful finding here.
- The best-supported intervention in this region is two minutes a day of resistance training, not a stretch.
Upper back tightness is a load sign
Sit at a screen and picture what happens between the shoulder blades.
The trunk rounds forward. The shoulder blades slide away from the spine and tip forward over the rib cage. The muscles that would hold them back — the lower and middle trapezius, the rhomboids, the mid-back extensors — are now lengthened and working at a low level, continuously, to stop your upper body folding any further. They are not resting. They are not shortening. They are doing an isometric hold for the length of your working day, at maybe five per cent of what they can produce, which is exactly the loading pattern that produces discomfort fastest.
That is why the sensation is so familiar and so specific: a dull, spreading ache in a band across the mid-back, worse late in the day, better within a minute of standing up and worse again within twenty minutes of sitting down.
Sustained low-level muscular effort of this kind produces discomfort in laboratory conditions in people with no complaints at all, and the discomfort resolves when the position changes. That is a well-described response to a held position rather than a disease of the tissue. What it means practically is that you changed the exposure, not the muscle — which is also why the relief does not last.
Length or load, and how to tell
This is the part worth two minutes of your day, because the two problems get run together constantly and they want different responses.
The test is whether the range appears when you ask for it.
- Sit on a hard chair, feet flat, knees together, arms folded across the chest. Turn as far as you can to the left without the pelvis following, then to the right. Note where each side stops against something fixed in the room.
- Kneel, put your hands on a chair seat or a low table, sit your hips back toward your heels and let the chest sink between your arms. That is extension of the same segment.
- Stand tall, reach one arm overhead and lean away from it. Both sides.
If all three directions are available on demand — not comfortable, not pretty, but available — then the segment is not short and a stretch is not the missing ingredient. You have found the range you were told you had lost, which means the ache is about the hours, not the length.
If a direction genuinely will not go, that is a range finding and it is worth working. Thoracic spine mobility sets out what the segment does and the nine movements the app files to it.
Most people who take this check find their range intact. That is not a disappointing result. It rules out the explanation you were about to spend six weeks acting on.
Feeling stiff is not being stiff
The evidence for that distinction is better than most people expect.
Stanton and colleagues, in Scientific Reports in 2017, measured back stiffness two ways in the same people: what they reported feeling, and what an instrument recorded mechanically. The two did not match. People who described their backs as stiff were not reliably stiffer on measurement, and the sensation could be manipulated by changing what people expected rather than by changing the back.
Take that seriously and the whole vocabulary of this topic shifts. "Tight" is a sensation your nervous system produces. It is a real experience and it is information, but it is not a direct readout of the length of anything, and it is not a diagnosis you can act on with a stretch.
Grade: reasonably solid for the mismatch between felt and measured stiffness. One well-designed study rather than a literature, but it is a direct test of the assumption everything else in this area rests on, and it came back negative.
The shape story fails on the same ground. Christensen and Hartvigsen reviewed the relationship between spinal curves and health in the Journal of Manipulative and Physiological Therapeutics in 2008 and found no association. A rounded upper back is a shape. Shapes vary enormously in people who feel nothing at all.
The finding that goes against us
This site sells mobility work, so here is the result that argues with it, printed at full strength before any caveat.
Andersen and colleagues, in Pain in 2011, gave office workers with frequent neck and shoulder pain two minutes a day of progressive resistance training and measured a real reduction in pain over ten weeks. Two minutes. A randomised trial with a clinical outcome in exactly the population this page is written for.
No stretching programme for the upper back and neck has produced a result like that. Not one. The same group's earlier work found specific strength training for the painful muscles outperformed general fitness training, and a separate one-year trial reached the same conclusion. The Cochrane reviews of exercise for mechanical neck disorders point the same way and are more cautious about it: some benefit for specific strengthening work, on low-quality evidence.
Grade: reasonably solid for a small daily dose of resistance training. Weak to absent for stretching as a treatment for this complaint.
There is a second intervention with support, and it is even duller. Waongenngarm and colleagues ran a year-long trial in office workers at high risk of neck and low back pain, testing active breaks and deliberate changes of position against a control group, and found fewer people went on to develop pain. Changing position often is not a stretching programme either. The micro-breaks research grades that literature in full.
Pain that persists, wakes you at night, radiates into a limb, comes with numbness or weakness, or followed a fall or an injury is a matter for a clinician who can examine you. Nothing here diagnoses anything and nothing here is treatment.
Where the movements do fit
None of the above says do not stretch. It says be accurate about why.
Movement is a break from a held position, and that is worth having on its own. Anything that takes the segment out of the shape it has been in for two hours qualifies. It does not have to be a stretch, it does not have to be long, and it should be frequent rather than heavy.
If the range check found a direction that will not go, work that direction. Rotation and extension, from the app's own catalogue: M043 Seated Tall Rotation, a 40-second hold each side, no kit; M039 Lying Open-Book, dynamic, 45 seconds each side; M041 Kneeling Thoracic Extension, a 45-second hold; M044 Supported Backbend Hold, 45 seconds over a roller. Dynamic work runs one repetition per breath cycle, six to twelve repetitions, inhaling four seconds and exhaling six. Holds run 20 to 60 seconds at a mild stretch, never a sharp one.
Add a small amount of resistance work for the same region. That is where the trial evidence is, and two minutes a day was the dose that produced it.
Do not chase a shape. The front-of-chest work that usually accompanies this complaint is pleasant and its claim on your posture is weak — chest opener stretches grades it honestly. If the neck is involved too, neck stretches is written on the same footing.
What we cannot read
An admission, since this page has spent its length grading other people's claims.
Limber cannot tell you whether your upper back is stiff. The thoracic area of the score is fed by one test at full weight — a seated rotation — and that test is not built, in the app or here. The Spine group is fed by that area alone, so it is the one station of five that never gets a reading. Four of five is the ceiling of the whole battery, and this is the station missing.
What you can take is the shoulder mobility test, the Apley scratch, both sides, five bands. It reads the reach chain the upper back sits inside — hands more than two hand-widths apart reads 10, one to two hand-widths reads 30, under one reads 55, fingertips touching 75, fingers overlapping 95. It sees the chain and it cannot say which link was short, which is a real limit and is printed on the instrument itself.
And the Range Score composes whatever you have measured into one 0–100 reading, then draws the unmeasured ground as unmeasured rather than filling it in with an average. An area with no test has no score. A group with no scored area has no score. That is why a partial survey here reads honestly instead of flatteringly.
The rest of the region sits under shoulders and upper back.
Questions
Why does my upper back feel tight all the time?
Most often because the muscles between the shoulder blades are holding a position rather than being short. Sitting lengthens them and asks them to work continuously at a low level, which is the loading pattern that produces an ache fastest. The clue is the timing: relief within a minute of standing, return within twenty minutes of sitting.
Should I stretch a tight upper back?
Only if a check shows the range is genuinely missing, and most of the time it is not. Sit tall and rotate each way with the pelvis still, then kneel and let the chest sink between your arms. If those directions go when you ask for them, the segment is not short, and a stretch is treating the sensation rather than the cause.
What is the difference between tight and weak?
Tight describes a sensation, and it turns out not to match measured stiffness — people who report stiff backs are not reliably stiffer on instrumented measurement. Weak describes a capacity, and it is the one with better evidence attached in this region: a small daily dose of resistance training reduced neck and shoulder pain in office workers where no stretching programme has.
Can Limber measure my upper back?
No. The seated thoracic rotation test that would read the segment is designed and not built, in the app or on this site, so the Spine group of the Range Score never gets a score. The nearest reading available is the shoulder reach, which covers the chain the upper back contributes to without being able to isolate it.
Why does the ache come back so fast after stretching?
Because you changed the exposure rather than the tissue. Acute range gained from a stretch is largely back at baseline within a few minutes, and the load that produced the ache resumes the moment you sit down again. Frequent short changes of position address the exposure directly, which is why they outperform one long stretching session for this particular complaint.
Is pain between the shoulder blades serious?
Usually not — it is one of the commonest complaints in desk work and it settles when the position changes. What is worth examining is pain that radiates into an arm, comes with numbness, weakness or pins and needles, wakes you at night, followed a fall, or arrives with fever, unexplained weight loss or breathlessness. Any of those is a reason to see a clinician who can examine you.