A 54-year-old woman comes in eight months into a shoulder that started aching for no reason. She cannot fasten a bra behind her back, reach the seatbelt, or sleep on that side. An orthopedist gave her a subacromial cortisone shot that did nothing. A physical therapist stretched the shoulder hard for six weeks and the night pain got worse. The MRI report says “mild rotator cuff tendinosis,” which describes most shoulders over fifty. When Dr. Movshis moves her arm for her, it stops at about 20 degrees of external rotation with a firm, rubbery end point, and she winces before it gets there. That is frozen shoulder, and the injection she received went into the wrong compartment.
Dr. Alex Movshis treats frozen shoulder at Modal Pain Management, 369 Lexington Avenue in Midtown Manhattan, with ultrasound-guided injection into the glenohumeral joint, hydrodilatation when the restriction is the dominant problem, and a suprascapular nerve block when night pain is what is stopping recovery. This page covers what the condition is, how it is told apart from a rotator cuff problem and from a pinched nerve, what the evidence says about each treatment, and what to expect here.
What frozen shoulder is, and why the name is being retired
The glenohumeral joint is held by a fibrous sleeve, the capsule, which is loose enough at rest to let the arm rotate through a wide arc. In frozen shoulder that capsule inflames, thickens with scar-like collagen, and contracts. The roomy fold under the joint, the axillary recess, closes. The coracohumeral ligament and the rotator interval at the front of the joint tighten. Biopsies of the capsule show a dense population of fibroblasts and myofibroblasts, the same contractile cells that pull a healing wound closed, and shoulders that had received a corticosteroid injection showed fewer of them (Hettrich 2016). The result is a joint that will not move even when someone else moves it, which is the single finding that separates this condition from every other cause of shoulder pain.
The 2011 consensus definition from the American Shoulder and Elbow Surgeons describes frozen shoulder as functional restriction of both active and passive motion with radiographs that are essentially normal (Zuckerman 2011). “Adhesive capsulitis” is the older name and is misleading, because there are no adhesions to break: the capsule is contracted, not stuck to anything. Lewis proposed “frozen shoulder contracture syndrome” as the more accurate label (Lewis 2015), and it is the model we use, because it points at the right target. The treatment problem is a thickened, shortened capsule and the pain it generates, not inflammation of a bursa and not a torn tendon, which is why the routine subacromial injection so often fails these patients.
Prefer to watch? A 45-second narrated animation of the mechanism — the capsule that inflames, thickens, and tightens, and why the stiffness persists on passive motion — with the full transcript and PubMed-verified references.
Who gets it
Frozen shoulder affects roughly 2–5% of the general population at some point, peaks between 40 and 65, and affects women more often than men (Kelley 2013). About half of cases are primary, meaning nothing triggers them. The rest are secondary to something identifiable, and the 2011 consensus splits those into three groups (Zuckerman 2011). Systemic means diabetes or thyroid disease. Extrinsic means something outside the shoulder made the arm stop moving, such as a fracture, breast or chest surgery, a stroke, or a pinched nerve in the neck. Intrinsic means a rotator cuff tendon or calcific deposit in the same shoulder started the process.
Diabetes is the risk factor that changes management. A meta-analysis found frozen shoulder in 13.4% of people with diabetes, and diabetes in about 30% of people presenting with frozen shoulder (Zreik 2016). Diabetic frozen shoulder tends to be more severe, lasts longer, is more often bilateral, and responds less completely to every treatment. Every patient with frozen shoulder and no known diabetes gets a hemoglobin A1c, and a surprising number of new diagnoses are made that way.
The three phases, and what they are actually for
The freezing phase is pain-dominant. A deep, poorly localized ache builds over weeks, night pain wakes the patient and makes lying on that side impossible, and stiffness is only beginning. It usually lasts two to nine months. The frozen phase is stiffness-dominant: the pain settles, but external rotation, overhead reach, and reaching behind the back are markedly restricted, typically for four to twelve months. The thawing phase is the slow return of motion over months to years.
Two things about the phases matter more than the timeline. The first is that they overlap, and the boundaries are not sharp, and the natural-history review found no evidence that untreated shoulders progress through the phases to full resolution (Wong 2017). The second is that the phase determines the treatment. In the freezing phase the capsule is too irritable to stretch, and the job is pain control and an injection. In the frozen phase the job is mobilization, with hydrodilatation if the restriction is the dominant problem. Treating the frozen phase with rest, or the freezing phase with aggressive stretching, is how patients lose a year.
The natural-history claim that everyone recovers on their own is weaker than it sounds. A systematic review of the studies behind it found no good evidence that frozen shoulder resolves fully in all untreated patients (Wong 2017). In the largest long-term follow-up, 269 shoulders reviewed an average of 4.4 years after onset, 59% were normal or near-normal and 41% still had symptoms, 94% of them mild with pain the commonest complaint, and 6% of them severe with pain and functional loss (Hand 2008). Those numbers argue for treating early rather than waiting.
How frozen shoulder is diagnosed
The diagnosis is made by examination, and the key test takes ten seconds. With the elbow at the side and bent to 90 degrees, the examiner rotates the forearm outward. In a rotator cuff problem the arm goes to 60–80 degrees and hurts at the top. In frozen shoulder it stops early, often at 10–30 degrees, against a firm end point, and the loss is the same whether the patient moves the arm or the examiner does. Passive loss of external rotation is the signature. Abduction and internal rotation (reaching behind the back) are lost next. Strength is preserved once the pain is controlled, which separates this from a full-thickness cuff tear.
Ultrasound in the exam room does two jobs. It shows the thickened coracohumeral ligament and the rotator interval at the front of the joint, and it excludes a full-thickness rotator cuff tear, calcific tendinitis, and a large effusion, each of which changes the plan. Plain radiographs are ordered to exclude glenohumeral arthritis, which produces a stiff shoulder with a different treatment, and an unrecognized fracture or dislocation. MRI is rarely needed for the diagnosis and is ordered when the exam does not fit or a surgical opinion is being sought. In a patient over fifty the MRI will usually show tendinosis, and it takes an examiner who has felt the capsular end point not to be led astray by the report.
The neck is examined in every case. A pinched nerve in the neck can produce shoulder pain and, by making the arm hurt to move, a secondary frozen shoulder. If neck extension and rotation toward the painful side reproduce the pain, or there is numbness or weakness in the arm, the neck is treated as well as the shoulder.
What the evidence says about each treatment
Intra-articular corticosteroid. The injection with the strongest short-term evidence. A meta-analysis of eight randomized trials with 416 patients found better pain, motion, and function than control at 4–6 weeks and again at 12–16 weeks, with a smaller, less certain advantage still present at 24–26 weeks (Sun 2017). The 2020 systematic review of 65 trials and 4,097 patients found intra-articular corticosteroid was the only intervention with both statistical and clinical superiority for short-term pain and function, and that adding a home exercise program and physiotherapy to the injection carried the benefit into the mid-term (Challoumas 2020). The injection does not shorten the disease by itself. It creates the window in which the capsule can be mobilized, and that is what shortens the course.
Hydrodilatation. A larger volume of saline and local anesthetic injected with the steroid to distend the contracted capsule from the inside. The meta-analysis of the randomized trials is sobering: a small effect on pain and range of motion, no significant effect on disability, no relationship between the volume injected and the result, and an average benefit the authors judged clinically insignificant (Saltychev 2018). It is therefore not routine here. It is offered in the frozen phase, when restriction rather than pain is the main problem and the steroid alone has not restored motion, as a low-risk addition at the same visit.
Suprascapular nerve block. The suprascapular nerve carries most of the sensation from the capsule and the back and top of the joint. Blocking it under ultrasound at the suprascapular notch gives hours to weeks of pain relief and, more usefully, breaks the cycle of night pain that prevents both sleep and rehabilitation. In a 30-patient randomized trial in primary care, a single suprascapular nerve block produced faster and more complete relief of pain and restoration of motion than a course of intra-articular injections (Jones 1999), and a 2021 systematic review of nine studies found suprascapular nerve block associated with significant improvements in pain and range of motion, while calling for larger trials against intra-articular injection (Jump 2021). We use it for the patient whose freezing phase is dominated by pain, and repeat it if needed. When a block works well but wears off, radiofrequency ablation of the same nerve extends the relief.
Physical therapy. Necessary, but the timing and the intensity are what matter. The Cochrane review found moderate-quality evidence that six weeks of manual therapy and exercise alone produced less improvement at seven weeks than a corticosteroid injection (Page 2014), and the network meta-analysis found the combination of injection and exercise carried benefit furthest (Challoumas 2020). Stretching an irritable capsule hard makes it worse. Once the pain is controlled, long-hold stretching within tolerable discomfort, assisted range of motion, and scapular strengthening are the core of the program, and it runs for months.
Surgery. Manipulation under anesthesia and arthroscopic capsular release exist for the patient who is still restricted after a year of correct treatment. The UK FROST trial randomized 503 patients with primary frozen shoulder to manipulation, arthroscopic release, or early structured physiotherapy with a steroid injection, and at 12 months the three groups were within about three points of each other on the Oxford Shoulder Score, every difference below the five-point target, with eight serious adverse events after capsular release against two after manipulation, and manipulation the most cost-effective of the three (Rangan 2020). Surgery is the exception, not the plan.
How it is treated at Modal Pain Management
The first procedure for a confirmed frozen shoulder is a corticosteroid injection into the glenohumeral joint under ultrasound, placed from the back of the shoulder so the needle path avoids the biceps tendon and the axillary nerve, with the needle watched into the joint and the spread of the injectate confirmed on screen. Landmark-guided glenohumeral injections miss the joint often enough that a “cortisone shot that did nothing” is more often a shot that never reached the capsule than a shot that failed. Patients with diabetes are told to expect higher glucose readings for two to four days and how to adjust.
If the shoulder is in the frozen phase, the restriction is the main problem, and the steroid alone has not restored motion, hydrodilatation can be added at a follow-up visit: 20–40 mL of saline and local anesthetic distends the capsule under real-time ultrasound, and the patient begins assisted range of motion immediately afterward, while the joint is numb. The evidence for it is modest and it is offered as a selective addition, not a routine step.
If night pain dominates and the patient cannot start rehabilitation, an ultrasound-guided suprascapular nerve block is done first or alongside the joint injection. Relief is usually immediate and lasts days to weeks, which is long enough to sleep and to begin the mobility program. A block that works and wears off can be repeated or converted to radiofrequency ablation.
Physical therapy is arranged at the first visit and starts once the pain allows, usually within one to two weeks. The program is built for the phase: pendulum and assisted range of motion first, long-hold external rotation and cross-body stretches next, scapular and rotator cuff strengthening as motion returns. The secondary muscle pain in the upper trapezius and levator scapulae, which nearly every frozen shoulder patient has after months of guarding, responds to trigger point injections when it is limiting the program.
Progress is measured, not guessed. External rotation and overhead reach are recorded at every visit, and reassessment is at three-month intervals. A patient who has had a correctly placed injection, a proper mobility program, and pain control, and who is still functionally restricted at nine to twelve months, is referred to a shoulder surgeon we work with for a manipulation or release opinion. That is a small minority.
What is not frozen shoulder
Three conditions are routinely mislabeled as frozen shoulder, and each has a different treatment. Rotator cuff tendinopathy and subacromial bursitis hurt in the same place and disturb sleep the same way, but passive motion is full or nearly full and the painful arc is between 60 and 120 degrees of abduction. Glenohumeral osteoarthritis is stiff in every direction like frozen shoulder, but the radiograph shows joint-space loss and osteophytes, and the treatment is different. A pinched nerve in the neck produces shoulder pain with numbness or tingling below the elbow, is reproduced by neck movement, and leaves the shoulder itself moving normally. A stiff shoulder after a fracture or dislocation is a post-traumatic contracture and is managed on the same principles as frozen shoulder but with the surgeon who treated the injury.
The wider picture of a shoulder that started hurting without an injury, and how the five common causes are told apart, is on the shoulder pain page.
What to expect at Modal Pain Management
The first visit is a 45-minute consultation. Bring any imaging and the list of what has been tried, including which compartment any previous injection went into if you know it. Dr. Movshis examines the shoulder and the neck, measures range of motion, scans the shoulder with ultrasound in the room, and, when the diagnosis is frozen shoulder, performs the joint injection the same day. Hydrodilatation or a suprascapular nerve block is added by phase. The physical therapy referral is written before you leave. You drive yourself home and use the arm normally that evening.
If you cannot reach behind your back, and someone else cannot move your arm there either, that is the visit to book.
References
- Kelley MJ, Shaffer MA, Kuhn JE, et al. Shoulder pain and mobility deficits: adhesive capsulitis. J Orthop Sports Phys Ther. 2013;43(5):A1–A31. doi:10.2519/jospt.2013.0302
- Zuckerman JD, Rokito A. Frozen shoulder: a consensus definition. J Shoulder Elbow Surg. 2011;20(2):322–325. doi:10.1016/j.jse.2010.07.008
- Lewis J. Frozen shoulder contracture syndrome — aetiology, diagnosis and management. Man Ther. 2015;20(1):2–9. doi:10.1016/j.math.2014.07.006
- Hettrich CM, DiCarlo EF, Faryniarz D, et al. The effect of myofibroblasts and corticosteroid injections in adhesive capsulitis. J Shoulder Elbow Surg. 2016;25(8):1274–1279. doi:10.1016/j.jse.2016.01.012
- 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. doi:10.11138/mltj/2016.6.1.026
- Hand C, Clipsham K, Rees JL, Carr AJ. Long-term outcome of frozen shoulder. J Shoulder Elbow Surg. 2008;17(2):231–236. doi:10.1016/j.jse.2007.05.009
- 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. doi:10.1016/j.physio.2016.05.009
- Sun Y, Zhang P, Liu S, et al. Intra-articular steroid injection for frozen shoulder: a systematic review and meta-analysis of randomized controlled trials with trial sequential analysis. Am J Sports Med. 2017;45(9):2171–2179. doi:10.1177/0363546516669944
- 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. doi:10.1001/jamanetworkopen.2020.29581
- 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. doi:10.1016/S0140-6736(20)31965-6
- Saltychev M, Laimi K, Virolainen P, Fredericson M. Effectiveness of hydrodilatation in adhesive capsulitis of shoulder: a systematic review and meta-analysis. Scand J Surg. 2018;107(4):285–293. doi:10.1177/1457496918772367
- Page MJ, Green S, Kramer S, et al. Manual therapy and exercise for adhesive capsulitis (frozen shoulder). Cochrane Database Syst Rev. 2014;(8):CD011275. doi:10.1002/14651858.CD011275
- Jones DS, Chattopadhyay C. Suprascapular nerve block for the treatment of frozen shoulder in primary care: a randomized trial. Br J Gen Pract. 1999;49(438):39–41. PubMed 10622015
- Jump CM, Waghmare A, Mati W, et al. The impact of suprascapular nerve interventions in patients with frozen shoulder: a systematic review and meta-analysis. JBJS Rev. 2021;9(12):e21.00042. doi:10.2106/JBJS.RVW.21.00042
