Internal impingement of the rotator cuff: the new subacromial impingement?

I am hearing more and more talk and reading more and more research about internal impingement of the rotator cuff, so it was about time I wrote a piece on it. Let me declare immediately and forcefully – I think it’s absolutely worthy of research and I support all efforts to advance knowledge on this concept. However, I do have concerns that we’re making the same mistakes we made for 50 years with subacromial impingement syndrome. Let me explain.

Firstly, a quick primer on internal impingement. Despite the catchy title of this blog, it is not new and was first described in the 1980s [1]. Crudely, it is explained as entrapment of the rotator cuff between the glenoid/labrum during arm elevation, and arm elevation + external rotation. It was historically believed to primarily afflict young throwing athletes, for example baseball players, however accumulating evidence suggests it might be common in non-athletic populations too [2].

Internal impingement is proposed to explain the articular surface side rotator cuff tear that is commonly observed during radiological and arthroscopic evaluation. It differs from subacromial impingement in its mechanism, which should be self-explanatory. Subacromial impingement occurs when the superior rotator cuff/bursa abrades the overlying coracoacromial arch. Internal impingement occurs when the articular surface of the rotator cuff gets ‘pinched’ or ‘kissed’ by the glenoid/labrum. Thus, it has been referred to as a ‘kissing’ lesion.

Subacromial impingement has been reported to occur in 50% of tested humans [2, 3], irrespective of whether they have shoulder pain or not. What about internal impingement? In a recent study by Lawrence et al, it was reported that 100% of asymptomatic participants displayed internal impingement during simple arm elevation [2]. There was no difference in the rate of internal impingement between those with and without a rotator cuff tear. The range of motion (ROM) where impingement occurred was between 90-150°, with a mean of 115°. As elevation ROM increased, the proximity of the rotator cuff footprint to the glenoid/labrum reduced in an inverse relationship (figure 1).

                                                                        Figure 1.

The authors mention that the location of rotator cuff/glenoid ‘kissing’ approximated the area where most degenerative rotator cuff tears are thought to originate (10-15mm posterior to bicipital groove). They conjecture this can’t be a coincidence and state that internal impingement may be a common mechanism of rotator cuff pathology.

A number of things stand out to me when looking at the literature on internal impingement:

1. It’s far more common than subacromial impingement – in fact it appears to be a normal phenomenon as it happens in everyone (that has been tested). If something occurs in everyone, is it pathologic or physiologic? It seems to me it is a physiological norm. However, we can still use this arthrokinematic knowledge to inform our treatment. If a patient has pain around 115° elevation ROM, we may wish to minimise time spent in this position for a period of time, before testing the waters at some stage again in the future, or experiment with movement modifications and see if it changes symptoms. Very clearly, this does not mean we should encourage avoiding this position ad infinitum.
2. It’s fairly accepted that rotator cuff health is at the mercy of many biological factors, including age, smoking, diabetes, occupation, and even psychological distress [4]. Yes, I think compression, both internal and external, may contribute to rotator cuff pathology too [5]. Is compression a primary driver of rotator cuff tears in the majority of individuals? I doubt it but it’s a plausible contributor we might consider.
3. It's worth considering the average ROM where internal impingement is reported to occur, approx. 115° elevation. This is at the upper end of the infamous 70-120° painful arc zone where it is suggested most individuals with RCRSP experience pain. It is plausible, therefore, that internal impingement could manifest as pain in this painful arc zone but it must be acknowledged that it is occurring at quite an advanced ROM. In fact, in the Lawrence study, it occurred on average at 76% of the participants maximum elevation ROM. I would posit the average individual going about their daily life would infrequently enter these ranges, and if they did it would be sporadic and mostly with minimal load. Of course there will be exceptions, for example those who play certain sports or do certain exercises and those who work in occupations where this ROM is required. However, even if an individual frequently goes into this ROM, can we say it's just internal impingement that is causing pathology and pain? I don't think so. Frequent loading cycles of the shoulder through ROM will subject the rotator cuff tendons to varied forces, including tensile, compressive, and shear. All of these forces cause the tendon to react, because tendons are mechanosensitive. The tendon reaction can be positive and negative, depending on context. We shouldn't simply demonise compression, although it has become fashionable to do so. And, moreover, tendons like load, crave load, and the solution to a load problem is loading!
4. Predicting in what context a rotator cuff tear will become related to symptoms is a further challenge. We all know the evidence of the prevalence of asymptomatic rotator cuff tears (I think?) [6]. Empirical predictors of symptomatic rotator cuff tears include older age, greater BMI, smoking, dominant arm, greater height, greater weight, greater critical shoulder angle, and hypertension [7]. It should be clear, most of these factors are unrelated to structure, morphometry, morphology, and kinematics. When thinking about causation in an epidemiological sense it is common to ask, “could something else explain this phenomenon?”. With rotator cuff pathology and shoulder pain there are many plausible alternative explanations, which should downgrade our credence that impingement is the primary contributor.
5. If internal impingement is a real and causally consequential mechanical process, how should it inform our treatment? Will it change anything we do? I propose, no. In my recent paper, we use the concept of exercise experiments in order to explore pain with movement and exercise [8]. Following this rationale, what we would do if we identified pain in the ROM where internal impingement may be occurring is simply modify the movement. We could modify the load, the plane of movement, the absolute range (i.e. don’t go beyond painful zone), the context, the lever arm, it goes on and on and on, and see what it does to their symptom profile. We emphasise what is painful now, might not be tomorrow or in 6 weeks. It’s momentarily sensitive for various reasons but there are still many positions and loads you can tolerate, let’s lean into that. There is no reason this philosophy can’t be applied to internal impingement and thus it’s business as usual for learned clinicians.

In summary, I fully advocate the continued investigation of internal impingement and I will follow the research closely. I am, however, concerned that internal impingement will be overused to lazily explain rotator cuff pathology and shoulder pain, just like its infamous predecessor subacromial impingement. If we have learnt anything over the last 20 years of researching shoulder pain and rotator cuff pathology, it’s that our simple intuitions of causation are often profoundly inadequate to explain complex pathologies and experiences. That’s not to say simple explanations shouldn’t be pursued, they absolutely should, but when they don’t stand up to experiment or basic logic, they should be viewed with caution.

References

1. Perry, J., Anatomy and biomechanics of the shoulder in throwing, swimming, gymnastics, and tennis. Clin Sports Med, 1983. 2(2): p. 247-70.
2. Lawrence, R.L., et al., In vivo evaluation of rotator cuff internal impingement during scapular plane abduction in asymptomatic individuals. J Orthop Res, 2023. 41(4): p. 718-726.
3. Lawrence, R.L., et al., Effect of glenohumeral elevation on subacromial supraspinatus compression risk during simulated reaching. J Orthop Res, 2017. 35(10): p. 2329-2337.
4. Leong, H.T., et al., Risk factors for rotator cuff tendinopathy: A systematic review and meta-analysis. J Rehabil Med, 2019. 51(9): p. 627-637.
5. Andrade, R., et al., Is Bony Morphology and Morphometry Associated With Degenerative Full-Thickness Rotator Cuff Tears? A Systematic Review and Meta-analysis. Arthroscopy, 2019. 35(12): p. 3304-3315 e2.
6. Barreto, R.P.G., et al., Bilateral magnetic resonance imaging findings in individuals with unilateral shoulder pain. J Shoulder Elbow Surg, 2019. 28(9): p. 1699-1706.
7. Zhao, J., et al., What Factors Are Associated with Symptomatic Rotator Cuff Tears: A Meta-analysis. Clin Orthop Relat Res, 2022. 480(1): p. 96-105.
8. Powell, J.K., et al., Is exercise therapy the right treatment for rotator cuff‐related shoulder pain? Uncertainties, theory, and practice. Musculoskeletal Care, 2024. 22(2).