The Rym Space
← All conditions

Shoulder · treated under Musculoskeletal Physiotherapy

Frozen Shoulder

Also called Adhesive capsulitis, Stiff shoulder.

Pain, then stiffness, from a shoulder capsule that has thickened and tightened. Strongly linked to diabetes, it runs a long course, and the right treatment changes with the stage.

Written by , BPT, MPT · MIAP License No. 69386

Updated · Awaiting clinical review

A physiotherapist in scrubs guiding a patient's bent arm across her body to test shoulder movement

What it is

Frozen shoulder is a stiff, painful shoulder caused by the joint capsule itself. The capsule is the sleeve of tissue that seals the ball and socket. In frozen shoulder it becomes inflamed, then thickens and contracts, so the ball can no longer glide and the arm loses movement in every direction, most of all turning outwards [2]. The medical name is adhesive capsulitis. It usually arrives without an injury, peaks in the mid fifties, and affects women slightly more often than men [2][4].

It is described in three overlapping stages. A painful stage, in which the shoulder aches at rest and at night and movement shrinks week by week, lasting roughly ten to thirty-six weeks; a stiff stage, in which the pain eases but the range stays restricted, usually from four to twelve months after it starts; and a thawing stage, in which movement gradually returns and which can itself take anywhere from one to three and a half years [2]. The stages are a useful map, not a promise. A systematic review found no good evidence that the shoulder reliably passes through them to full recovery on its own, and in the trials it examined most of the improvement came early, not late [3]. That is the case for treating it early, and treating it right for its stage.

How it presents

  • A deep ache in the shoulder and outer upper arm that came on without an injury, building over weeks
  • Pain at night and difficulty lying on that side, worst in the early months
  • A sharp catch when the arm is jarred or moved suddenly: reaching for a seat belt, a back pocket, or a bra strap
  • Steadily shrinking movement: reaching behind the back, into a sleeve or up to a high shelf gets harder every few weeks
  • The arm cannot be turned outwards, even when someone else moves it gently for you, which is the sign that separates the capsule from a tendon [2]
  • Usually one shoulder. The other becomes affected in 6% to 17% of people, generally within five years and after the first has settled [2]
  • The stiffness outlasts the pain: in the middle stage the shoulder hurts less but moves no better

The mix of pain and stiffness shifts with the stage, and that shift is what we track. Early on, pain leads and stiffness follows; later the pain fades and stiffness is what limits you. Which stage you are in changes what treatment should be doing, and it is the first thing the assessment settles.

Who gets it and why

The trigger is not known. What is known is what happens inside the joint: the lining of the capsule becomes inflamed, and the capsule then thickens with scar-like tissue and contracts, so that a sleeve that should be loose and folded becomes tight and short. The front of the capsule, which controls turning outwards, is the part most affected, which is why external rotation goes first and returns last. It can also follow a period of not using the arm, after a fracture, surgery or a stroke, when it is called secondary frozen shoulder.

The strongest association is with diabetes. In a meta-analysis of 18 studies, people with diabetes were five times more likely than others to have a frozen shoulder, about 13% of people with diabetes had one, and 30% of people presenting with a frozen shoulder had diabetes [5]. Thyroid disease matters too: pooled data from ten case-control studies found the odds of thyroid disease nearly doubled in people with frozen shoulder, with the link strongest for an underactive thyroid, including the mild form that causes no symptoms [6]. A frozen shoulder in someone who has never been tested for either is a reason to test.

Risk factors

  • Age, most often the fifties, peaking around 56; it is unusual before 40 [2]
  • Diabetes, which raises the likelihood about fivefold [5]
  • An underactive thyroid, including the subclinical form found only on a blood test [6]
  • Being female, slightly: women were affected 1.6 times as often as men in the largest long-term cohort [4]
  • A period of immobility after a shoulder fracture, surgery or stroke
  • Having had it on the other side, since the second shoulder follows in 6% to 17% of people [2]

How it is diagnosed

Frozen shoulder is diagnosed in the room, from the story and one decisive finding. With your arm at your side and elbow bent, we turn the forearm outwards for you, with your muscles relaxed. In a frozen shoulder that movement stops early and hard, at a fraction of the other side; the trial that compared surgery with physiotherapy used a loss of at least half of this passive external rotation as its entry criterion [14]. A rotator cuff problem does not do that: the arm turns fully when someone else moves it, even if lifting it yourself hurts or is weak [2]. Arthritis of the shoulder can stiffen the joint in the same way, which is where the X-ray comes in.

The rest of the examination maps the loss of movement in each direction, grades how irritable the shoulder is (how easily it is provoked and how long it takes to settle), checks the neck and the nerves, and looks for the conditions that mimic it. The physiotherapy guideline describes the pattern to recognise: a gradual, progressive onset of pain with loss of both active and passive movement, in lifting and in rotation [1]. Irritability matters more than it sounds, because it decides how hard the shoulder can be worked.

When imaging earns its place

A plain X-ray is worth having once, and it is expected to be normal. Its job is to rule out the conditions that stiffen a shoulder the same way and are managed differently: osteoarthritis of the joint and, rarely, a dislocation that was never recognised. Ultrasound and MRI do not diagnose frozen shoulder and do not change the treatment. They earn their place when there was an injury, when there is weakness, or when the story does not fit.

What the evidence says

Treatment by treatment, with the studies behind each verdict numbered to the sources at the end of the page.

  • Corticosteroid injection, early

    Core treatment

    The one treatment with clear short-term superiority. Across 65 randomised trials with 4097 participants, an injection of corticosteroid into the joint was the only treatment that beat both placebo and physiotherapy for pain and function in the first twelve weeks, and it worked best when given within the first year [9]. In a placebo-controlled trial of 93 people, the injected groups were clearly better than placebo at six weeks and three months, and all groups had caught up by twelve months [10]. It is a treatment for the painful stage, and is not indicated once only stiffness remains [2].

  • Stretching matched to the stage

    Core treatment

    Stretching restores the capsule's length, but the dose has to match how irritable the shoulder is: gentle and within comfort while it is painful, firmer and held longer once it is merely stiff [1]. In a prospective study of 75 people in the stiff stage treated with a four-direction stretching programme, 90% reported a satisfactory outcome and only 7% went on to manipulation or surgical release [12]. A stretch that leaves the shoulder aching hours later was too much for the stage you are in.

  • A home programme, little and often

    Core treatment

    Whatever else is done, the home programme carries the result. The network meta-analysis found that adding a simple home programme of exercises and stretches to an injection gave added benefit in the months that followed [9], and in the placebo-controlled trial every group, placebo included, was taught a home programme and had improved by twelve months [10]. Several short bouts a day beat one long session, and it is the part of treatment that runs for the whole course.

  • Joint mobilisation and hands-on treatment

    Helps alongside

    Gliding the ball in its socket, held short of pain, to ease the capsule and win a few degrees to stretch into. The Cochrane review of 32 trials found the evidence for manual therapy and exercise on their own mostly of low quality and could not say which technique is best; the combination was less effective than a steroid injection at seven weeks, with no clinically important difference between them at six and twelve months [7]. The guideline lists it as an option [1]. We use it as a means to the exercise, not instead of it.

  • Machines: ultrasound, TENS, laser, heat

    Use with caution

    Widely offered, thinly supported. The Cochrane review of 19 trials of electrotherapy found the evidence for ultrasound, shortwave and TENS too poor to say whether they help at all; only low-level laser had moderate-quality support, from a single trial, for less pain over the first four weeks when added to exercise [8]. Heat before you stretch is reasonable for comfort. A course made up of machines alone is weeks lost in the period when improvement comes most easily [3].

  • Hydrodilatation (distension injection)

    Second line

    Fluid, usually saline with steroid, injected into the joint under imaging to stretch the capsule from the inside. The Cochrane review found short-term benefit in pain, movement and function over placebo, from the one trial at low risk of bias, and could not say whether it beats an ordinary steroid injection [13]. Reasonable for a shoulder that has stayed stiff despite a proper course of stretching and an injection, not a first step.

  • Manipulation under anaesthesia

    Second line

    The surgeon frees the tight capsule by moving the shoulder through its range while you are asleep. In the UK FROST trial, 503 people referred to hospital were randomised to manipulation, keyhole release, or early structured physiotherapy with a steroid injection. At twelve months the shoulder scores were 38.3, 40.3 and 37.2 out of 48, differences below the level set in advance as clinically meaningful, so none was superior [14]. It has a place for a shoulder still badly stiff after months of proper treatment, not before.

  • Arthroscopic capsular release

    Second line

    Keyhole surgery to divide the contracted front of the capsule, followed by manipulation and physiotherapy. In UK FROST it produced the highest score at a year, but by a margin too small to be clinically meaningful, and it carried the most serious adverse events: eight, against two after manipulation [14]. The trial's reading was that it carries higher risk with no clear advantage, and that early structured physiotherapy plus an injection gets most people to the same place [14].

  • Aggressive stretching in the painful stage

    Avoid

    Forcing an irritable shoulder to its end of range provokes more inflammation and more guarding. In a two-year study of 77 people, those treated with intensive stretching and manipulation did worse than those given supportive care and exercises within their pain limits: 63% against 89% reached a normal or near-normal shoulder by two years [11]. Pain that lingers for hours after a session is the signal that the dose was wrong.

  • Waiting for it to thaw on its own

    Use with caution

    The old teaching was that frozen shoulder cures itself in two to three years. The evidence does not bear that out. A systematic review found only low-quality evidence that an untreated shoulder improves, and none that it recovers fully; most of the recovery in the trials it examined came early, not late [3]. In a cohort of 223 people followed for an average of 4.4 years, 41% still had some symptoms, though for most these were mild [4]. Waiting is a choice, not a plan.

Read as a whole: an injection early, while the shoulder is painful, buys the fastest relief; a stretching programme matched to the stage, done at home several times a day, does the long work; hands-on treatment helps the programme along; and the invasive options, tested head to head against early physiotherapy plus an injection, did not get people further at a year [14]. The stage is the plan. Settle the irritability first, then work the range, and change course rather than repeat it if a proper course has not moved the shoulder.

How we treat it

At The Rym Space, Bagalur

  1. 1

    Assessment

    The story, then the movement map: each direction measured against the other side, with passive external rotation the deciding test. We grade how irritable the shoulder is, check the neck, and ask about diabetes and thyroid. You leave knowing which stage you are in and what that means.

  2. 2

    A plan you start that day

    Positions that let you sleep, how to dress and reach without provoking it, and a first set of gentle movements within comfort. If the shoulder is in its painful stage we talk through an injection, and who to see for it, so that the window for it is not missed.

  3. 3

    Week 1: settle the pain

    3 to 5 sessions close together. Hands-on treatment and gentle joint glides held short of pain, heat before movement, and a short home routine repeated several times a day. The aim this week is a calmer shoulder and better nights, not more range.

  4. 4

    Week 2 onwards: work the range

    As irritability drops, the stretches get firmer and are held longer, in the four directions the capsule restricts. Sessions space out as the home programme carries the work, and each visit re-measures the range so that the dose stays right for the stage.

  5. 5

    The thawing stage: rebuild the arm

    Once the range is returning, the rotator cuff and shoulder blade muscles, weakened by months of restriction, are strengthened, and everyday reaching, lifting and sport are rebuilt on a graded plan. Stretching continues, because the last degrees of external rotation are the slowest to come back.

  6. 6

    Discharge, or the next tier

    Done when you can reach, dress and sleep without thinking about it, and the range is close to the other side. A shoulder that has not moved despite a proper course and an injection is talked through honestly: hydrodilatation, or a surgical opinion, with the evidence in front of you.

What to expect, and when

Ranges, not promises. Where a figure comes from a study, the study is numbered.

First weeks
Pain settles before movement does. A steroid injection given in the painful stage brings the fastest relief, with clear differences over placebo at six weeks and three months [10]. Sleep usually improves first.
By 3 to 6 months
The stiff stage, which typically runs from four to twelve months after onset: the pain fades but the range stays restricted [2]. This is when firmer stretching is tolerated and does its work. The trials with longitudinal data show that most of the improvement comes early, not late [3].
By 12 months
In the placebo-controlled trial every group, injected or not, had improved to a similar degree by twelve months [10], and in the stretching study 90% of people treated in the stiff stage were satisfied at a mean follow-up of twenty-two months [12]. Improved is not full: the last of the external rotation returns slowest.
The long view
Frozen shoulder often runs a course of two to three years [13]. In a cohort of 223 people followed for an average of 4.4 years, 59% had a normal or near-normal shoulder, 41% had some ongoing symptom, mild in 94% of them, and 6% had severe pain and loss of function [4].

Two of those figures look contradictory and are not. Most people are much better within a year of proper treatment; a minority keep a mild restriction or ache that never fully goes [2][4]. The people who did worst in the long term were those with the most severe symptoms at the start [4], which is one more reason not to wait.

When to be seen sooner

Signs that this is something else, or that it needs looking at now rather than after a few weeks of self-care.

  • A stiff shoulder after a fall, a seizure or an electric shock, until an X-ray has excluded a dislocation that was not put back or a fracture
  • Marked weakness rather than stiffness, or an arm that cannot be lifted at all after an injury, which points to a rotator cuff tear
  • A shoulder that is hot, red or swollen, or fever with the pain
  • Pain that is constant and unrelated to movement, particularly with unexplained weight loss or a history of cancer
  • Numbness, tingling or weakness in the hand, or pain that travels below the elbow, which point to the neck rather than the shoulder
  • Both shoulders freezing at the same time in someone who has not been tested for diabetes or thyroid disease
  • A frozen shoulder under the age of 40, which is unusual enough to deserve a second look [2]

Common questions

Will it get better on its own?
Some, yes; fully, not reliably. A systematic review found only low-quality evidence that an untreated frozen shoulder improves at all, and none that it recovers fully, and most of the improvement in the trials came early rather than late [3]. In a long-term cohort, 41% still had some symptom years later, mostly mild [4]. Treating it early is what shifts the course.
Should I have an injection?
In the painful stage, probably. It was the only treatment that clearly beat both placebo and physiotherapy in the short term across 65 trials, and it works best within the first year [9]. Once the pain has gone and only stiffness is left, it is not indicated [2]. We will say which stage you are in and point you to the right doctor if it fits.
Do I need an MRI?
No. Frozen shoulder is diagnosed by examination, and the finding that settles it is a loss of passive external rotation, which is felt, not scanned [2]. An X-ray is worth having once, to exclude arthritis. A scan earns its place if there was an injury, if there is weakness, or if the picture does not fit.
Should I push through the pain to keep it moving?
No. In the two-year comparison, people given intensive stretching and manipulation while the shoulder was irritable did worse than those who exercised within their pain limits: 63% against 89% recovered [11]. Movement within comfort, several times a day, is the rule until the pain settles; then the stretches get firmer.
Will I need surgery?
Most people do not. In the UK FROST trial, manipulation under anaesthesia and keyhole release were no better at a year than early physiotherapy with an injection, and the keyhole surgery carried more serious complications [14]. Surgery is for a shoulder still badly stiff after a proper course of treatment, and the choice is made with you, not for you.
Why is it linked to diabetes?
The mechanism is not settled, but the link is strong: about 30% of people who present with a frozen shoulder have diabetes, and people with diabetes are five times more likely to develop one [5]. If you have never been tested, a frozen shoulder is a reason to be [5].

Further reading

Sources

Guidelines, systematic reviews and trials only. Each number in the text links to the entry here.

  1. 1.Kelley MJ, Shaffer MA, Kuhn JE, et al. Shoulder pain and mobility deficits: adhesive capsulitis. Clinical practice guidelines. J Orthop Sports Phys Ther. 2013;43(5):A1-A31. pubmed.ncbi.nlm.nih.gov ↗
  2. 2.Dias R, Cutts S, Massoud S. Frozen shoulder. BMJ. 2005;331(7530):1453-1456. pubmed.ncbi.nlm.nih.gov ↗
  3. 3.Wong CK, Levine WN, Deo K, et al. Natural history of frozen shoulder: fact or fiction? A systematic review. Physiotherapy. 2017;103(1):40-47. pubmed.ncbi.nlm.nih.gov ↗
  4. 4.Hand C, Clipsham K, Rees JL, Carr AJ. Long-term outcome of frozen shoulder. J Shoulder Elbow Surg. 2008;17(2):231-236. pubmed.ncbi.nlm.nih.gov ↗
  5. 5.Zreik NH, Malik RA, Charalambous CP. Adhesive capsulitis of the shoulder and diabetes: a meta-analysis of prevalence. Muscles Ligaments Tendons J. 2016;6(1):26-34. pubmed.ncbi.nlm.nih.gov ↗
  6. 6.Chuang SH, Chen YP, Huang SW, Kuo YJ. Association between adhesive capsulitis and thyroid disease: a meta-analysis. J Shoulder Elbow Surg. 2023;32(6):1314-1322. pubmed.ncbi.nlm.nih.gov ↗
  7. 7.Page MJ, Green S, Kramer S, et al. Manual therapy and exercise for adhesive capsulitis (frozen shoulder). Cochrane Database Syst Rev. 2014;8:CD011275. pubmed.ncbi.nlm.nih.gov ↗
  8. 8.Page MJ, Green S, Kramer S, et al. Electrotherapy modalities for adhesive capsulitis (frozen shoulder). Cochrane Database Syst Rev. 2014;10:CD011324. pubmed.ncbi.nlm.nih.gov ↗
  9. 9.Challoumas D, Biddle M, McLean M, Millar NL. Comparison of treatments for frozen shoulder: a systematic review and meta-analysis. JAMA Netw Open. 2020;3(12):e2029581. pubmed.ncbi.nlm.nih.gov ↗
  10. 10.Carette S, Moffet H, Tardif J, et al. Intraarticular corticosteroids, supervised physiotherapy, or a combination of the two in the treatment of adhesive capsulitis of the shoulder: a placebo-controlled trial. Arthritis Rheum. 2003;48(3):829-838. pubmed.ncbi.nlm.nih.gov ↗
  11. 11.Diercks RL, Stevens M. Gentle thawing of the frozen shoulder: a prospective study of supervised neglect versus intensive physical therapy in seventy-seven patients with frozen shoulder syndrome followed up for two years. J Shoulder Elbow Surg. 2004;13(5):499-502. pubmed.ncbi.nlm.nih.gov ↗
  12. 12.Griggs SM, Ahn A, Green A. Idiopathic adhesive capsulitis. A prospective functional outcome study of nonoperative treatment. J Bone Joint Surg Am. 2000;82(10):1398-1407. pubmed.ncbi.nlm.nih.gov ↗
  13. 13.Buchbinder R, Green S, Youd JM, Johnston RV, Cumpston M. Arthrographic distension for adhesive capsulitis (frozen shoulder). Cochrane Database Syst Rev. 2008;1:CD007005. pubmed.ncbi.nlm.nih.gov ↗
  14. 14.Rangan A, Brealey SD, Keding A, et al. Management of adults with primary frozen shoulder in secondary care (UK FROST): a multicentre, pragmatic, three-arm, superiority randomised clinical trial. Lancet. 2020;396(10256):977-989. pubmed.ncbi.nlm.nih.gov ↗

Book an appointment

Start with a frozen shoulder assessment

Your first appointment is a full assessment: history, hands on examination and movement testing, followed by an explanation of what we found and the plan to fix it. You leave with something to do that day, not just a follow up booking.

WhatsApp Call +91 88008 96463
Aerospace Park, Airport RoadBagalur Colony, BagalurBengaluru, Karnataka 562149

Monday to Saturday: 9:00 am to 1:00 pm, 4:00 pm to 8:00 pmSunday: By appointment

Other conditions