“Is there any research?” is a question we hear often from people considering an injection for shoulder arthritis or a rotator cuff tear. Our clinic has contributed published clinical research to that discussion.
In 2019, my colleagues and I published Treatment of shoulder osteoarthritis and rotator cuff tears with bone marrow concentrate and whole bone marrow injections in Cogent Medicine. The study examined changes in symptoms and function after treatment. It is important to understand both what we observed and what the design could not establish. (1)
Who was included?
The study included 47 patients and 50 shoulders; three patients had both shoulders treated. The diagnoses were shoulder osteoarthritis or rotator cuff tears. Patients received one or two treatments using bone marrow concentrate or whole bone marrow. In those who received a second treatment, the average interval was approximately 22 days. These are study details, not a recommended schedule for every patient. (1)
What is the difference between whole marrow and concentrate?
Both preparations came from the patient’s own iliac crest, the pelvic bone. Whole bone marrow was used without concentration; BMC underwent processing. Neither was a purified stem-cell product.
Our analysis grouped outcomes by the number of treatments. It was not designed to establish that BMC was superior to whole marrow, or the reverse. That distinction matters when a reader tries to use the paper to choose a preparation. (1)
Why we measured daily function
Resting pain and pain during activity do not tell the whole story of a painful shoulder. We also wanted to know whether patients could manage ordinary activities more easily.
The questionnaire used ten activities drawn from the Upper Extremity Functional Index. Examples included opening a door, washing dishes, combing hair, reaching overhead, carrying something heavy, rolling over in bed, doing household work, dressing, and participating in hobbies or sports. (1)
These are the kinds of changes patients ask about in the office. Someone may care more about putting on a shirt independently or getting comfortable in bed than about a small change in a scan.
Using selected items is also a methodological limitation: this was a modified functional measure, not the complete standard index. Its percentage changes should not be treated as interchangeable with every other shoulder questionnaire.
What did patients report?
Both treatment groups reported statistically significant improvements from baseline in pain and function. The study reported the following percentage changes: (1)
| Reported outcome | One-treatment group | Two-treatment group |
|---|---|---|
| Reduction in resting pain | 33.33% | 47.31% |
| Reduction in active pain | 28.13% | 42.92% |
| Mean patient-reported overall improvement | 42.25% | 50.17% |
| Increase in functional score | 15.81% | 36.89% |
The two-treatment group had a significantly greater improvement in active pain than the one-treatment group. The paper did not find statistically significant outcome differences between the osteoarthritis and rotator cuff tear groups. (1)
An average overall improvement of 50.17% does not mean that half the patients were cured, or that half the cartilage was restored. It was a patient-reported improvement measure. Group averages also do not predict the outcome of a particular patient.
The groups represented 20 shoulders after one treatment and 30 after two. Reported follow-up averaged 8.66 and 5.36 months, respectively. The unequal follow-up is another reason to avoid treating the percentages as a direct test of treatment frequency. Disease severity was not accounted for in the analysis. (1)
Why the comparison cannot prove that two treatments are better
Patients were not randomly assigned to one or two injections. Groups may differ in symptoms, severity, timing, expectations, and other characteristics. Without randomization and a suitable control group, a difference between groups cannot establish that an additional injection caused the difference.
For the same reason, improvement after treatment cannot be attributed entirely to the marrow preparation. Rehabilitation, changes in activity, natural variation in symptoms, and the contextual effects of treatment may contribute.
What the study does not demonstrate
- It does not establish superiority to PRP, cortisone, exercise, or shoulder replacement.
- It does not demonstrate that BMC outperforms whole marrow.
- It does not demonstrate cartilage regrowth or reliable reconnection of a torn tendon.
- It does not establish the response of an 89-year-old with end-stage arthritis.
- It does not establish a standard number of injections or an ideal treatment interval.
- Its small size cannot reliably characterize uncommon complications.
The treating clinic also conducted the research. That relationship should be transparent. Independent controlled studies are important for determining how much of an observed benefit is caused by the treatment and which patients are most likely to benefit.
How does this fit with other shoulder research?
A later pilot randomized study of nonconcentrated marrow for glenohumeral arthritis reported advantages over cortisone in some secondary outcomes, but not a statistically significant difference in its primary shoulder outcome. The trial was small and ended early. (2)
Separately, a 2024 randomized crossover trial studied combined marrow concentrate and platelet products for supraspinatus tears, rather than arthritis. The combined intervention and crossover design limit conclusions about marrow alone. (3)
These studies ask different questions. I do not combine them into a single claim that stem cells work for every painful shoulder.
What I want patients to take from our findings
We observed improvements that mattered to patients: less pain and easier use of the arm. Those observations deserve to be reported. They also deserve to be presented with enough detail that readers understand the uncertainty.
When considering treatment, the next step is to relate the evidence to the actual diagnosis, the amount of structural damage, and the patient’s goals. An encouraging group result should begin an informed discussion, not end it.
For the broader treatment decision, read our guide to bone-on-bone shoulder arthritis and replacement alternatives. For tendon injuries and the differences between PRP and marrow preparations, see our shoulder pain and rotator cuff overview.
References
1 Darrow M, Shaw B, Schmidt N, Boeger G, Budgett S. Treatment of shoulder osteoarthritis and rotator cuff tears with bone marrow concentrate and whole bone marrow injections. Cogent Medicine. 2019;6(1):1628883. doi:10.1080/2331205X.2019.1628883.
2 Dwyer T, Hoit G, Lee A, et al. Injection of Bone Marrow Aspirate for Glenohumeral Joint Osteoarthritis: A Pilot Randomized Control Trial. Arthroscopy, Sports Medicine, and Rehabilitation. 2021;3(5):e1431–e1440. doi:10.1016/j.asmr.2021.07.005.
3 Centeno CJ, Fausel Z, Dodson E, Berger DR, Steinmetz NJ. Percutaneous bone marrow concentrate and platelet products versus exercise therapy for the treatment of rotator cuff tears: a randomized controlled, crossover trial with 2-year follow-up. BMC Musculoskeletal Disorders. 2024;25(1):392. doi:10.1186/s12891-024-07519-6.
