Shockwave therapy probably reduces pain from frozen shoulder for a short time, and a procedure called manipulation under anaesthesia probably helps people raise the arm further forward, according to an analysis of 189 randomised trials published in The BMJ. The researchers rated the evidence for both findings as of moderate certainty. For most of the 59 treatments they compared, and for side effects, the results remain uncertain.
Frozen shoulder, known in medicine as adhesive capsulitis, makes the shoulder painful and stiff. Many different treatments are used for it. The researchers set out to compare their benefits and harms: their effect on pain, on everyday function, on how far the shoulder can move, and their side effects.
How the review was done
The team looked for every randomised controlled trial in adults with frozen shoulder, that is, every trial in which chance decided who got which treatment. The search ran up to May 2025 and was updated in November 2025. They then combined the results in a network meta-analysis, a method that pools many trials and can also compare treatments that were never tested against each other directly.
In all, 189 trials with 12,988 participants and 59 different treatments were included. Most of the trials behind the main results had a high risk of bias, meaning flaws in the way they were run that could distort their findings: 71% of them for pain, 72% for function and 71% for range of movement. The researchers graded how certain each result was, on a scale that runs down to very low.
Less pain and better movement, for a short time
In the short term, shockwave therapy, in which pressure waves are sent into the shoulder from outside the body, probably reduced pain compared with usual care. On a pain scale from 0 to 10, the difference was 1.97 points. The evidence for this was of moderate certainty. Shockwave therapy may also have improved function, by 11.42 points on a scale from 0 to 100, but this evidence was of low certainty, and the direct and indirect comparisons did not agree well.
For movement in the short term, manipulation under anaesthesia, in which the stiff shoulder is moved firmly while the person is anaesthetised, probably improved how far the arm could be raised forward. The gain was 33.29 degrees when the procedure was combined with a steroid injection into the joint, and 26.16 degrees when it was used alone. This evidence, too, was of moderate certainty. Shockwave therapy may have improved raising the arm forward, by 22.42 degrees, and to the side, by 18.62 degrees, but with low certainty and comparisons that did not agree well.
Over the medium term, the evidence for any gain was of low or very low certainty, and so was the evidence behind most of the other comparisons.
Side effects and the long term remain unclear
The trials reported side effects incompletely and in different ways. How the treatments compare in safety therefore remains highly uncertain, the researchers write, and when a trial reported no side effects, this cannot be taken as proof that there were none.
The researchers conclude that evidence of moderate certainty points to less pain in the short term with shockwave therapy, and to better forward movement of the arm in the short term with manipulation under anaesthesia, with or without a steroid injection. Most other effects remain uncertain. They add that the evidence is mainly short term and tied to the particular outcome measured. Long-term effects, overall benefits and harms, and whether the results hold beyond the form of frozen shoulder that arises without another known cause, which the trials mainly studied, are still unknown.
