A recent trial looking at the use of steroid injections and exercises for shoulder pain called the GRASP trial has recently been published in The Lancet this month with some interesting results. But yet again, just like there was with the recent SExSI trial there have been some comments and claims thrown around on social media misrepresenting what this trial shows. So this short blog is my attempt to clarify what the GRASP trial does and doesn’t show.

First things first I would like to congratulate all the authors and contributors to this trial as I do recognise doing research like this, as well as this, is both difficult, challenging, and time-consuming. So to get to the conclusion first the GRASP trial which stands for Getting it Right Addressing Shoulder Pain states that progressive exercise is NOT superior to best practice advice and that a steroid injection only offers a small, short-lasting clinically and statically insignificant effect on shoulder pain.

From this conclusion, I have been disappointed to see yet again some physios getting a bit too excited to say how exercise doesn’t help people with shoulder pain, and weaponising this study just as they did with the SExSI trial a few weeks ago to have a dig at my active treatment approach and ethos of ‘you cant go wrong getting strong’. Well, to put this simply as I can these individuals are wrong and misinterpreting and/or misunderstanding what this trial actually shows, usually because they haven’t read it in full just to have a pathetic pop at me instead. So let’s try and help them out a bit.

What is the GRASP?

The GRASP was a multi-centre, pragmatic superiority non-blinded randomised controlled trial, in which 708 patients (average age of around 55) with unilateral non-traumatic sub-acromial shoulder pain, of approx 4 months in duration were randomly allocated to one of 4 treatment conditions. Steroid injection + Best Practice Advice, or No Steroid Injection + Best Practice Advice, or Steroid Injection + Progressive Exercise, or No Steroid + Progressive Exercise.

The first thing which I think has confused and possibly misled a few people about this study is its title and the naming of the treatment groups which I do think at first glance give a false impression that this trial is comparing stand-alone advice with progressive exercises, with or without steroid injections. This is not so.

The ‘Best Practice’ group did have only one session with a physio lasting 1 hour in which they were given advice and education about shoulder pain, reassurance that the pain was not dangerous or harmful or anything serious or sinister. But they were also advised to do progressive exercise 5 x a week following a simple program of exercises with both leaflet and online video support. This group had no formal follow up sessions planned although they could get in contact with a physio if they had any issues.

The ‘Progressive Exercise’ group only differed from the ‘best practice’ group in that they had all of the above but also had up to an additional 5 follow up sessions with a physio over a 4 month period. The ‘progressive exercise’ group just like the ‘best practice’ group still had to do very similar exercises independently themselves 5 x week.

So basically both the ‘best practice’ and ‘progressive exercise’ groups had exactly the same advice, education, and reassurance and both groups were told to exercise 5 x week independently. The only difference is the ‘progressive exercise’ group had the option to attend 5 follow up sessions, which many did not with only around 25% attending the full 6 sessions either due to the physios deeming them fit for discharge or subjects not attending.


Personally, I think these groups would have been better named ‘single session’ and ‘multi session’ not ‘best practice’ and ‘progressive exercise’ as this gives a false impression that only one group had advice and the other group only exercised. In fact, the title of this study also appears to say it compares the effects of advice v progressive exercise on shoulder pain when it clearly doesn’t just do this. This study really compares the effect of single v’s multiple physio appointments in those with shoulder pain, and of course the effect of a single steroid injection.

We also don’t really know if ANY of the groups exercised that much as one of the biggest limitations of this study is that exercise adherence was not strictly monitored. Although the completed exercise rates in this trial were reported in the main paper as being around 90% for the best practice / single session group, and around 75% for the progressive exercise / multi session group they relied on subject self-reporting which often over-estimates adherence.

What I found interesting was in the supplementary files, the individual breakdown of how many subjects reported doing the exercises 5 x week for the first 8 weeks was around 43% for the single session only group, and about 60% for the multi-session group. This dropped to 15% for the single session group, and 18% for the multi session group by 6 months, and at 1 year it was less than 1% in both groups. This clearly shows that a lot of subjects were doing much less or nothing in both groups. We also don’t know how long, how intense, and which exercises the subjects were doing, which again I think is a huge limitation to this study.

The supplementary chart on self-reported exercise adherence in GRASP trial

So with this in mind, we have to consider that it could be that once educated, reassured, and given some attention and resources, those with shoulder pain improved with time and perhaps some occasional exercise of various intensity, duration, and type. Without a comparative group who were truly only given just advice alone with reassurance to carry on with their usual tasks and activities, we won’t know how much is from the advice or the exercise, but I am will to bet that there is a strong natural history effect here once people with shoulder pain are seen and reassured well.


The other interesting point with the GRASP study was the lack of significant effect of the steroid injection. Steroid shots are given a lot for people with shoulder pain and many are told that it will reduce pain to allow them to move more and do exercises better. And indeed the GRASP trial did show this, in some, for a short period of time.

However, it also shows that steroid injections are not as effective as many think and perhaps not the first thing to reach for in those with shoulder pain. Remember that the average duration of pain for subjects in this trial was 4 months and in terms of sub-acromial shoulder pain that can be classed as still ‘early days’. Personally, I would not often consider recommending a steroid injection for ‘simple’ non-traumatic sub-acromial shoulder pain in the first 6 months, and preferably 12 months.


So there you go a quick review of the GRASP trial which despite some dubious group naming is an excellent trial showing how steroids are not as effective or necessary as many think, and how advice and reassurance and maybe a little bit of exercise for a short period of time can help those with sub-acromial shoulder pain.

Once again to all my special fans out there, this unfortunately for you doesn’t give you any evidence or support that exercise doesn’t help people with shoulder pain, at best you can say it amuses them whilst natural history kicks in, just like most other treatments. However, if you are going to use a treatment to amuse a patient whilst natural history kicks in, why not use one that has a host of positive physiological and psychological benefits rather than all your silly poking, prodding, stabbing, scrapping or sticking them with stuff.

As always thanks for reading and remember, you can’t go wrong getting strong…

Also if you are interested in a fun, engaging and enlightening weekend taking a look at the evidence and how best to assess and manage people with shoulder pain then check out my LIVE+ONLINE and Face-2-Face Shoulder Complex courses here

Until next time, happy lifting




Related news