Exercise is the constant in rotator cuff pain — what you add on top varies

Let's be upfront: the evidence on which exercise type, load and duration works best in rotator cuff-related shoulder pain is still thin. What exists is reasonably consistent about exercise as the base, but thin on everything you add on top of it. Here is what the recent literature actually supports — and where it stops.

Start with the strongest signal available. In a systematic review of 1,152 middle-aged and older adults with rotator cuff-related shoulder pain, Vico-Rodríguez et al. (2026) pooled trials comparing exercise-based interventions against usual care, other exercise approaches or adjunct therapies, and found significant improvements in pain (g = -1.454, p < 0.001), function (g = 0.457, p < 0.001) and ROM (g = -2.477, p < 0.001). Strength was the exception: improvements were smaller and did not reach statistical significance (g = 0.229, p = 0.085). The authors also flag high heterogeneity (I² = 95.96%), likely due to differences in protocols, participants and outcome assessment. Read that number before quoting any effect size from this literature to a colleague.

What the adjunct trial actually shows

The one trial here that adds something to exercise reports improvement in every arm — which is precisely why the added value of each adjunct stays unresolved. In an RCT of 29 patients with MRI-confirmed supraspinatus tendinopathy and sleep disturbance, Pawłuś et al. (2026) compared physiotherapy alone against PT plus subacromial steroid injection, PT plus suprascapular nerve block and PT plus PRP, and all groups demonstrated clinical improvement, most consistently in pain, insomnia severity, shoulder function and cognitive performance. With PT alone also improving significantly in pain, insomnia severity, shoulder function and processing speed, the incremental benefit of each injection arm is hard to read from these numbers.

Outcomes were assessed only up to 3 months, so no long-term follow-up is reported.

Where the interesting work is happening

One adjacent line of work is worth watching, without letting it carry your prescription. In the FORTIS pilot, only the PT-alone group showed significant correlations between BDNF changes and reductions in pain and insomnia severity — a correlation, not a proven mechanism.

What this means for your clinic

Expect pain, function and ROM to move; do not promise strength gains with the same confidence, since the pooled strength effect was not significant. Treat adjuncts — injections, PRP, nerve blocks — as additions whose specific value is still an open question, and say so to patients. And in a secondary analysis of an RCT of 84 patients with chronic rotator cuff tendinopathy, Åxman et al. (2026) found that logbook adherence declined over the 26-week programme, with pain logbook compliance falling from 76% to 68% and exercise logbook compliance from 75% to 61%.

Since logbook completion fell over the long programme in that trial, PhysioSync's Home Programs put the exercises in the patient's calendar while Smart Alerts flag low adherence, so you can step in and adjust the plan when it slips.