Frozen Shoulder Manipulation Under Anesthesia

Manipulation under anesthesia (MUA) is a procedure in which a surgeon forcibly moves a frozen shoulder through its full range of motion while the patient is sedated or under general anesthesia, physically breaking apart the thickened, contracted joint capsule that restricts movement. It has been used for decades and remains one of the most common interventions for frozen shoulder that has not responded to conservative treatment. The procedure typically takes only a few minutes, but the science behind when it works best, what it does to the joint, and how it compares to surgical alternatives is more layered than the simplicity of the technique might suggest.

What Frozen Shoulder Actually Is

Frozen shoulder, formally called adhesive capsulitis, is not just stiffness from disuse or a sore rotator cuff. The joint capsule itself becomes inflamed and then progressively fibrotic, shrinking and tightening until the shoulder loses much of its range of motion. Research on capsular tissue from frozen shoulders shows densely packed collagen fibers and an overactive population of fibroblasts, the cells responsible for producing connective tissue, along with elevated markers of inflammation and cell activation compared to normal shoulder tissue.¹ The process involves inflammatory cytokines, immune cells including lymphocytes and macrophages, and fibrotic growth factors. The pattern of fibroblast proliferation is similar to what occurs in Dupuytren’s contracture, a condition that thickens and shortens tissue in the hand.²

This fibrotic capsule is what MUA targets. The condition classically progresses through a painful “freezing” stage, a “frozen” stage with severe stiffness but sometimes less pain, and a gradual “thawing” stage where motion slowly returns. Many people recover on their own, but recovery can take one to three years, and some never fully regain motion without intervention. When physical therapy, steroid injections, and time have not worked, MUA is one of the main options on the table.

What Happens During the Procedure

Once the patient is fully anesthetized and the shoulder muscles are relaxed, the surgeon stabilizes the shoulder blade with one hand and uses the other to move the arm through a specific sequence of motions: forward flexion (raising the arm overhead), abduction (moving it out to the side), and external rotation (rotating it outward). These controlled movements put tension on different sections of the joint capsule, stretching and tearing the fibrotic tissue that has contracted around the joint. Different positions stress different parts of the capsule. Flexion and abduction tighten the front and bottom portions, while internal rotation with the arm pulled across the body stresses the back of the capsule.³

The word “manipulation” sounds gentle, but what happens inside the joint is significant. Arthroscopic studies performed immediately after MUA reveal capsular tears in predictable locations. One study examining 30 shoulders after manipulation found that the capsule ruptured superiorly in about a third of patients, anteriorly in the majority, and posteriorly in roughly half. Beyond capsular tears, the study also found labral detachments, partial tears of the subscapularis tendon (one of the rotator cuff muscles), and damage to the middle glenohumeral ligament in a minority of cases.⁴ This is an important point: MUA works by causing controlled damage to the pathological tissue. It is effective precisely because it ruptures the contracted capsule, but the tradeoff is that the procedure is not perfectly selective in what it breaks.

How Well Does MUA Work

The short answer is that for most patients, it works well and the improvement lasts. A review of the evidence found that MUA leads to substantial increases in range of motion and shoulder function scores, significant pain reduction, and roughly 85% patient satisfaction. The long-term data showed that results continued to improve beyond the first year after the procedure.⁵

One of the longest follow-up studies available tracked patients for an average of 15 years after MUA. Forward elevation improved from about 104 degrees before the procedure to 168 degrees afterward, and external rotation jumped from about 23 degrees to 67 degrees. At final follow-up, the large majority reported no pain or only slight pain, no fractures or dislocations had occurred, and only one of the 19 shoulders required further surgery.⁶ Those numbers are encouraging, though it is worth noting the study was small. The overall body of evidence points to MUA as a durable intervention, not a temporary fix.

Recurrence appears to be uncommon. A large study of 240 patients found that only about 2% reported recurrence more than five years after MUA, and less than 1% needed a second manipulation.⁷ That said, the re-intervention rate across the broader literature sits higher, around 14%, when you include patients who needed additional treatment of any kind, not just a repeat MUA.⁸

MUA Versus Arthroscopic Capsular Release

The main surgical alternative to MUA is arthroscopic capsular release (ACR), in which a surgeon uses a camera and instruments inserted through small incisions to cut the contracted capsule under direct visualization rather than tearing it blindly through manipulation. Intuitively, you might expect the precision of arthroscopy to produce better outcomes, and some surgeons do prefer it. But the comparative evidence tells a more nuanced story.

Two recent meta-analyses reached broadly similar conclusions. One found that ACR produced slightly larger gains in forward flexion at three and six months compared to MUA, but by 12 months the difference disappeared. Pain scores, external rotation, and complication rates were comparable between the two approaches.⁹ The other meta-analysis found that ACR had a small statistical advantage in pain reduction beyond 12 months, but the difference did not reach the threshold considered clinically meaningful. More notably, this analysis found that the ACR group had a higher rate of severe complications compared to MUA.¹⁰

A prospective randomized study confirmed this picture, finding that both MUA and ACR produced significant improvements in pain, range of motion, and function, with no meaningful long-term differences between groups. The MUA group showed slightly faster early gains.¹¹ In practice, the choice between MUA and ACR often depends on surgeon preference, local expertise, and the specific characteristics of the patient’s frozen shoulder rather than a clear clinical superiority of one technique over the other.

What Can Go Wrong

MUA is generally safe, with an overall complication rate reported at around 0.4% in one large review.¹² However, this figure likely understates the true rate because most studies evaluating MUA were not specifically designed to monitor for complications. The known risks fall into a few categories.

Fractures are the most feared complication, especially in patients with osteoporosis. In one series, a fracture of the glenoid rim occurred in a 72-year-old patient, though it healed on its own without lasting problems.¹³ Humeral shaft fractures are possible but rare in clinical practice. Beyond fractures, MRI studies performed after manipulation have documented bone bruises, capsular tears in various patterns, and labral tears.¹⁴ While many of these findings are expected consequences of the procedure and resolve over time, they do exist.

The most serious complications tend to happen outside controlled medical settings. A case report described a patient who sustained a subscapularis tendon tear, biceps tendon subluxation, and shoulder subluxation after an overzealous manipulation performed by an untrained professional.¹⁵ The key takeaway here is that MUA should be performed by an experienced orthopedic surgeon in a controlled environment with proper anesthesia, not by anyone who simply grabs the arm and forces it to move.

Does Adding a Steroid Injection Help

Many surgeons inject a corticosteroid into the joint immediately after manipulation, hoping the anti-inflammatory effect will help maintain the gains in range of motion. The evidence on this is mixed and somewhat underwhelming. A systematic review and meta-analysis comparing MUA with and without corticosteroid injection found no significant differences in pain, forward flexion, abduction, or internal rotation. The one exception was external rotation, which improved by roughly 7 degrees more when steroids were added.¹⁶

Another study reached a similar conclusion, finding that steroid injection after manipulation offered no advantage over manipulation alone, except again for external rotation movements.¹⁷ A separate retrospective study, however, found that patients who received both MUA and a steroid injection showed greater early improvements in pain scores and passive range of motion at two weeks compared to those who had MUA alone.¹⁸ The takeaway is that a steroid injection probably offers a modest short-term boost, particularly for external rotation, but it does not fundamentally change long-term outcomes. Since the injection itself is low risk, most surgeons continue to include it as a routine part of the procedure.

What About People With Diabetes

Frozen shoulder is two to four times more common in people with diabetes than in the general population, and it tends to be more stubborn. This raises a natural question: does diabetes make MUA less effective? The answer is reassuring but comes with a caveat.

Studies comparing outcomes in patients with and without diabetes have generally found that MUA produces equivalent improvements in shoulder scores and range of motion for both groups, whether measured at short-term or long-term follow-up.¹⁹ The caveat is that patients with diabetes are more likely to need the procedure repeated. One study found that 36% of diabetic patients required a repeat procedure compared to 15% of non-diabetic patients, even though final functional scores were comparable between the groups.²⁰ So if you have diabetes and are considering MUA, the procedure should still work, but there is a higher chance you will need a second round to get there.

Interestingly, when researchers have looked at arthroscopic capsular release in diabetic patients, the picture is slightly less optimistic. One study found that patients with diabetes had higher disability scores in the short term after ACR compared to non-diabetic patients.²¹ The underlying biology probably explains this: the metabolic environment in diabetes promotes inflammation and fibrosis, which is the very process driving frozen shoulder. Breaking up the capsule once may not be enough if the body’s tendency to lay down scar tissue is heightened.

Does It Matter How Long You Have Been Frozen

A common concern is that waiting too long before having MUA will lead to worse results. Some surgeons advocate early intervention, reasoning that the capsule becomes harder and more fibrotic with time, making it more difficult to manipulate safely. The evidence, however, does not clearly support this worry. A regression analysis found no correlation between the duration of symptoms before MUA and outcomes at either early or long-term follow-up. Patients who had been symptomatic for longer fared just as well as those treated sooner.²²

That said, this finding does not mean there is no reason to avoid unnecessary delay. While final outcomes may be similar regardless of timing, every additional month spent with a frozen shoulder is a month of restricted function, sleep disruption, and reduced quality of life. The evidence simply suggests that if you have been dealing with a frozen shoulder for a long time and are worried it is “too late” for MUA, it probably is not.

Rehabilitation After the Procedure

MUA is not a standalone event. What happens in the days and weeks after manipulation is critical to maintaining the range of motion gained during the procedure. The capsule has just been torn, and if the shoulder is not moved aggressively in the recovery period, the body’s healing response will lay down new scar tissue and the gains will be lost. Early and intensive physiotherapy is a standard part of the post-MUA protocol.²³

The challenge is that moving a recently manipulated shoulder is painful. Some centers begin physical therapy within 24 hours of the procedure.²⁴ To help patients tolerate this, various pain-control strategies are used. One approach that has gained interest is the use of a continuous nerve block, where local anesthetic is infused through a catheter near the brachial plexus nerves that supply the shoulder. The goal is to keep the shoulder numb enough that the patient can perform range-of-motion exercises without being limited by pain.²⁵ Without adequate pain control, patients tend to guard the shoulder and avoid movement, which defeats the purpose of the manipulation.

Some practitioners have explored performing the manipulation itself under a regional nerve block instead of general anesthesia. One study compared translational manipulation under interscalene block to a standard physical therapy approach and found faster early improvement in the manipulation group, with fewer activity limitations four years later.²⁶ Regional anesthesia for MUA has the advantage of avoiding the risks of general anesthesia and potentially allowing the patient to begin exercises sooner, though it is not universally adopted.

Cost and Practical Considerations

For patients weighing their options, cost can be a deciding factor. A prospective study comparing manipulation under a brachial plexus block to continued physiotherapy found that the manipulation group achieved better quality-of-life outcomes at roughly half the total cost of the physiotherapy group.²⁷ This makes intuitive sense: a single procedure followed by a recovery period is less expensive than months of ongoing therapy visits. The cost comparison with arthroscopic capsular release depends on the healthcare system, but ACR generally involves operating room time, specialized equipment, and more post-operative care, all of which add up.

From a practical standpoint, MUA is typically an outpatient procedure. You go in, get put under, the surgeon manipulates the shoulder in a matter of minutes, and you go home the same day. Recovery involves daily home exercises and usually several weeks of formal physiotherapy. Most patients notice a dramatic improvement in motion immediately after the procedure, though there is often some regression in the first few weeks as swelling and pain set in. The trajectory is generally a big initial gain, a partial dip, and then a gradual climb back toward the full range achieved under anesthesia.

What MRI Reveals About Healing After MUA

Advanced imaging has given researchers a window into what happens to the shoulder in the months after manipulation. One study performed MRI scans before manipulation, one week after, and six months after. At one week, most patients showed bone bruises, capsular tears, and in some cases labral tears. By six months, these findings had changed in various ways: some healed completely, others persisted.²⁸ The patients with more dramatic MRI findings did not necessarily have worse clinical outcomes, which raises an interesting question: the structural damage visible on imaging after MUA may not correlate neatly with how the shoulder actually feels and functions.

This disconnect between imaging and symptoms is common across musculoskeletal medicine. Many people with labral tears or rotator cuff changes on MRI have perfectly functional, painless shoulders. In the context of MUA, the structural changes visible on post-procedure MRI are largely a sign that the manipulation successfully disrupted the pathological capsule. They are expected consequences, not necessarily complications. The clinical outcome seems to depend more on the rehabilitation that follows than on the exact pattern of tissue disruption seen on imaging.

How Patients Experience the Decision

The UK FROST trial, one of the larger randomized trials for frozen shoulder treatment, included a qualitative component exploring how patients felt about their treatment. Participants who received MUA found the treatment acceptable, were satisfied with the results, and said they would choose the same treatment again.²⁹ Patients across all treatment arms experienced improvements in pain, movement, and function. What stands out in qualitative research like this is that patients with frozen shoulder are often desperate for relief by the time they reach the point of considering a procedure. Months of limited sleep, inability to dress normally, and constant aching wear people down. The speed of improvement after MUA, even if incomplete, is often the most valued aspect of the procedure from the patient’s perspective.

For people currently deliberating, the evidence supports MUA as a safe, effective, and cost-efficient option for frozen shoulder that has not improved adequately with conservative measures. It works about as well as arthroscopic capsular release in the long run, carries a low complication rate when performed by a trained surgeon, and produces durable gains that hold up over many years. The main prerequisite for a good outcome is a commitment to aggressive rehabilitation in the weeks that follow.