A note on this column: This article is general medical information, not a description of a service offered at this practice. Dr. Ishiguro's surgical practice is focused on hip and knee conditions — he does not treat AC joint injuries or perform shoulder surgery. This column is written to help readers understand a common sports injury and know when and how to seek the right specialist, typically a shoulder or sports orthopaedic surgeon.
Key points
- A "separated shoulder" is an injury to the AC joint at the top of the shoulder — it is different from a shoulder dislocation, which involves the ball-and-socket joint
- It usually follows a direct fall onto the point of the shoulder, most often in contact sports and cycling
- Injuries are graded I to VI (the Rockwood classification) according to which ligaments are torn and how far the collarbone has shifted
- Grades I and II are almost always treated without surgery, and grades IV–VI are commonly considered for surgery
- For grade III, randomised trials have generally found similar long-term function with or without surgery, with faster early recovery after non-operative care
- A visible bump often remains after non-operative treatment; it is usually a cosmetic issue rather than a functional one, but a minority of people stay symptomatic
What the AC joint is
The acromioclavicular (AC) joint is the small joint at the top of the shoulder where the outer end of the collarbone (clavicle) meets the acromion, a bony projection of the shoulder blade. You can usually feel it as a small ridge on top of the shoulder. Two sets of ligaments hold it together: the AC ligaments, which wrap around the joint itself, and the stronger coracoclavicular (CC) ligaments, which run from the underside of the collarbone down to the coracoid process of the shoulder blade and provide most of the joint's vertical stability.
When someone lands on the point of the shoulder with the arm at the side, the shoulder blade is driven downward while the collarbone is held in place. The AC ligaments tear first; with more force, the CC ligaments tear as well, and the shoulder drops away from the collarbone. The end of the collarbone then appears to stick up — which is why people often describe a "step" or "bump". In Japanese this injury is called kensa kansetsu dakkyū (肩鎖関節脱臼).
It is worth separating this from a shoulder dislocation, where the ball of the upper arm comes out of the socket. The two are often confused because both are called "shoulder" injuries, but they involve different joints and are managed very differently.
Who gets it
AC joint injuries are among the most common shoulder injuries in active young adults, particularly men who play contact sports.
What the evidence shows: A cohort study at the United States Military Academy followed students over five years (17,606 person-years) and recorded 162 AC joint injuries — an incidence of 9.2 per 1,000 person-years. Most (89%) were low-grade sprains. Men had about twice the injury rate of women, and men's rugby, wrestling and ice hockey carried the highest risk. Low-grade injuries cost an average of about 10 days away from sport, and high-grade injuries about 64 days (Pallis et al., American Journal of Sports Medicine, 2012). This was a highly active military-academy population, so the figures are higher than would be expected in the general public.
In everyday life in Tokyo, falls from bicycles, judo and rugby, skiing and snowboarding trips, and falls on stairs or icy pavements are typical causes.
Symptoms and how the diagnosis is made
Typical symptoms are pain and tenderness directly on top of the shoulder, swelling, and pain when lifting the arm overhead or reaching across the body. In higher-grade injuries the outer end of the collarbone is visibly prominent. A doctor will usually press over the AC joint, check whether the collarbone can be pushed back down, and perform a cross-body adduction test (bringing the arm across the chest), which typically provokes pain at the joint. It is also important to check for other injuries — a fracture of the collarbone, a rotator cuff tear, or nerve symptoms.
X-rays are the main test. Images of both shoulders are often taken so the injured side can be compared with the normal side, measuring the coracoclavicular distance (the gap between the collarbone and the coracoid). An additional view from above (axillary view) helps show whether the collarbone has been pushed backwards. MRI or ultrasound is not needed in most cases, but may be used when the grade is uncertain or another injury is suspected.
The grades of AC joint injury
The most widely used grading system is the Rockwood classification, which describes which ligaments are torn and in which direction and how far the collarbone has moved.
| Grade | What is injured | Usual approach |
|---|---|---|
| Type I | AC ligaments sprained; joint not displaced | Non-operative |
| Type II | AC ligaments torn, CC ligaments sprained; collarbone slightly raised | Non-operative |
| Type III | AC and CC ligaments torn; collarbone clearly elevated | Usually non-operative first; surgery debated |
| Type IV | Collarbone pushed backwards into the trapezius muscle | Surgery commonly considered |
| Type V | Severe upward displacement with tearing of the muscle attachments over the collarbone | Surgery commonly considered; some trials included non-operative care |
| Type VI | Collarbone displaced downward beneath the acromion or coracoid (rare) | Surgery |
Because grade III is the "grey zone", an international expert group has proposed splitting it into two subtypes.
What the evidence shows: The ISAKOS Upper Extremity Committee proposed dividing Rockwood type III injuries into type IIIA — a stable joint without overriding of the collarbone on a cross-body X-ray and without significant scapular dysfunction — and type IIIB, an unstable joint with an overriding collarbone on that view and shoulder-blade dysfunction that does not respond to rehabilitation (Beitzel et al., Arthroscopy, 2014). This is a consensus statement rather than a trial, but it is commonly used to frame which grade III patients might be considered for surgery after a period of rehabilitation.
Non-operative treatment
For grades I and II, and for most grade III injuries at first, treatment is non-operative. It typically involves a sling for comfort for a short period (often days to a few weeks, depending on pain), ice and pain relief, and then a gradual rehabilitation programme focusing on shoulder range of motion, shoulder-blade (scapular) control, and later strengthening. Heavy lifting and contact sport are usually avoided until pain has settled and strength has returned.
A natural question is whether a special brace that pushes the collarbone down helps the joint heal in a better position.
What the evidence shows: A small two-centre randomised trial of 35 patients with Rockwood type III injuries compared a simple sling with a brace that applies downward pressure to the end of the collarbone. Over 12 weeks there was no significant difference between groups in collarbone position on X-ray or in patient-reported shoulder scores (Finsterwald et al., JSES International, 2023). The trial was small and short, but it did not find an advantage for the brace.
After non-operative treatment, the bump at the top of the shoulder usually remains. For most people this does not interfere with function, but it is worth knowing in advance.
Surgery versus non-operative care: what the trials show
Surgery for AC joint separation aims to put the collarbone back in position and hold it there while the ligaments heal, or to reconstruct the coracoclavicular ligaments. Methods have included hook plates, suspension devices passed through bone tunnels ("tightrope"-type devices), screws, wires, and ligament grafts, done through open or arthroscopic approaches. Several randomised trials have compared surgery with non-operative care, mostly for grade III (and some grade IV–V) injuries.
What the evidence shows: In a multicentre Canadian trial of 83 patients with complete (grade III–V) injuries, those treated without surgery had better disability scores at 6 weeks and 3 months, while there was no difference at 6 months, 1 year or 2 years. Hook-plate fixation gave better alignment on X-ray but a higher reoperation rate (Canadian Orthopaedic Trauma Society, Journal of Orthopaedic Trauma, 2015). The ACORN trial from Edinburgh randomised 60 patients aged 16–35 with grade III or IV injuries to a tunnelled suspension device or non-operative care: function at 1 year was similar, recovery was faster without surgery, and surgery cost more — but 5 non-operatively treated patients remained dissatisfied and went on to have delayed reconstruction (Murray et al., Journal of Bone and Joint Surgery, 2018).
Longer-term follow-up and pooled analyses point in the same direction.
What the evidence shows: A Cochrane review of six randomised or quasi-randomised trials (357 mainly young male adults) found low-quality evidence of no difference in shoulder function at one year or longer, earlier recovery after conservative treatment, and more adverse events after surgery — largely hardware problems and infection, many involving older devices no longer widely used (Tamaoki et al., Cochrane Database of Systematic Reviews, 2019). A 2026 meta-analysis of five randomised trials (367 patients, Rockwood III–V) likewise found comparable function at two years or more, early scores favouring non-operative care, more minor complications with surgery, and similar satisfaction and return-to-work and sport (Lameire et al., Orthopaedic Journal of Sports Medicine, 2026). A ten-year follow-up of a randomised trial of hook-plate fixation versus non-operative care for type III injuries found no significant difference in function at one or ten years (Michael et al., Orthopaedics & Traumatology: Surgery & Research, 2025).
These trials have limitations: they are small, cannot be blinded, often lose patients to follow-up, and mostly enrolled young men. Several used fixation methods that have since been refined. So the evidence does not show that surgery is never useful — rather that, for a typical grade III injury, starting with non-operative care is a reasonable, well-supported option, and that a minority of people who stay painful or unstable after rehabilitation may later be considered for reconstruction.
When surgery is more commonly discussed
Surgery tends to be considered for grade IV and VI injuries, many grade V injuries, open injuries or those threatening the skin, and grade III injuries that remain painful or unstable (type IIIB) after a genuine period of rehabilitation. Some overhead athletes and manual workers also discuss earlier surgery with their surgeon. The decision is individual and depends on the injury, the person's work and sport, and their own priorities — including how they feel about the cosmetic bump.
Recovery and returning to activity
Recovery time varies with the grade. Low-grade sprains often allow a return to sport within a couple of weeks, while higher-grade injuries take considerably longer, whether or not surgery is done. Some people develop late pain from wear in the AC joint itself years after the injury. Persistent pain on top of the shoulder, clicking, or weakness with overhead activity several months after injury is a reason for review by a shoulder specialist.
When to see a doctor
Anyone with significant pain on top of the shoulder after a fall, a visible step at the end of the collarbone, or difficulty lifting the arm should be assessed, because a collarbone fracture or other shoulder injury can look similar. Urgent assessment is needed if there is numbness or weakness in the arm, the skin over the collarbone is tented or broken, or there is difficulty breathing.
References (PubMed / Journal)
Pallis M, Cameron KL, Svoboda SJ, Owens BD. Epidemiology of acromioclavicular joint injury in young athletes. Am J Sports Med. 2012;40(9):2072-2077. DOI: 10.1177/0363546512450162
Beitzel K, Mazzocca AD, Bak K, et al. ISAKOS upper extremity committee consensus statement on the need for diversification of the Rockwood classification for acromioclavicular joint injuries. Arthroscopy. 2014;30(2):271-278. DOI: 10.1016/j.arthro.2013.11.005
Finsterwald M, Dao Trong ML, Hollo D, et al. Conservative treatment of Rockwood type III acromioclavicular joint separation: a randomized controlled trial sling vs. brace. JSES Int. 2023. DOI: 10.1016/j.jseint.2023.02.017
Canadian Orthopaedic Trauma Society. Multicenter randomized clinical trial of nonoperative versus operative treatment of acute acromio-clavicular joint dislocation. J Orthop Trauma. 2015;29(11):479-487. DOI: 10.1097/BOT.0000000000000437
Murray IR, Robinson PG, Goudie EB, Duckworth AD, Clark K, Robinson CM. Open reduction and tunneled suspensory device fixation compared with nonoperative treatment for type-III and type-IV acromioclavicular joint dislocations: the ACORN prospective, randomized controlled trial. J Bone Joint Surg Am. 2018;100(22):1912-1918. DOI: 10.2106/JBJS.18.00412
Tamaoki MJS, Lenza M, Matsunaga FT, Belloti JC, Matsumoto MH, Faloppa F. Surgical versus conservative interventions for treating acromioclavicular dislocation of the shoulder in adults. Cochrane Database Syst Rev. 2019;10(10):CD007429. DOI: 10.1002/14651858.CD007429.pub3
Michael R, Sinclair K, Bédard L, et al. A ten-year prospective randomized trial comparing non-operative treatment to hook plate fixation for Rockwood III acromio-clavicular dislocation. Orthop Traumatol Surg Res. 2025:104159. DOI: 10.1016/j.otsr.2025.104159
Lameire DL, Cecere M, Chan A, et al. Outcomes after operative versus nonoperative management of type III to V acromioclavicular joint dislocations: a systematic review and meta-analysis of randomized controlled trials. Orthop J Sports Med. 2026;14(7):23259671261440812. DOI: 10.1177/23259671261440812