Can You Return to Golf With a Rotator Cuff Tear Without Surgery?

Marc Darrow, MD, JD

Some golfers return to play with a rotator cuff tear without having surgery. Whether that is a reasonable goal depends on the tear, shoulder strength and stability, symptoms, and response to rehabilitation. PRP or bone marrow–based injections may be considered in selected cases, but they cannot be assumed to repair every tear or replace a needed operation.

During the July 18, 2026 episode of Living Pain-Free, Steve called from London and described returning to golf after shoulder dislocations and tendon injuries. He reported choosing injection treatment after several surgical consultations and said he could again hit long drives.

His experience is encouraging. The broadcast does not provide the imaging, examination findings, exact injection preparation, or follow-up required to establish structural healing. Returning to golf is a meaningful functional outcome, but it is not proof that every damaged tendon has returned to normal.

The first question is what is causing the shoulder pain

We emphasizes finding the source of symptoms rather than treating the word “tear” as a complete diagnosis. A shoulder can have more than one relevant problem: rotator cuff tendinopathy, a partial tear, bursitis, biceps tendon pain, labral injury, arthritis, or instability.

That does not mean every painful shoulder needs every structure injected. The examination and appropriate imaging should identify which findings explain the symptoms and which may be incidental. More injections are not automatically better treatment.

The existing discussion of shoulder labrum tears addresses another possible contributor after a dislocation. A labral injury and a rotator cuff tear are different conditions, even when they occur together.

Research also documents an association between cuff injury and instability. A surgical series of 842 shoulders with recurrent anterior instability in patients aged 40 or younger found cuff lesions in approximately 7%. Most were partial articular-sided supraspinatus lesions. This selected population shows that the problems can coexist; it does not establish that all cuff tears involve widespread joint degeneration. (1)

Partial tears, full-thickness tears, and retracted tendons are different

Tendinopathy describes a tendon disorder and does not necessarily mean a tear is present. A partial-thickness tear affects part of the tendon’s depth. A full-thickness tear extends through it. A larger tear may retract away from its attachment and be accompanied by muscle atrophy or fatty change.

These distinctions affect treatment. A study of chronic tendinopathy cannot establish whether an injection will reconnect a detached tendon after a traumatic dislocation. An improvement in pain also does not necessarily mean that the tendon can transmit force normally.

The 2025 AAOS guideline recognizes that both physical therapy and surgery can improve outcomes in symptomatic small-to-medium full-thickness tears. It also cautions that tear size and muscle changes may progress during long-term nonsurgical care. This supports an individualized decision that considers function and structural risk. (2)

Sudden weakness after trauma, inability to actively raise the arm, repeated dislocations, or new numbness deserves timely assessment. A suspected acute repairable tear should not be delayed solely to try a series of injections.

What is PRP intended to do?

Platelet-rich plasma is prepared from the patient’s blood. Processing concentrates platelets and produces a preparation containing signaling proteins involved in the tissue response to injury. It is not a stem cell preparation.

Platelet-derived signaling components include platelet-derived growth factor, transforming growth factor beta, insulin-like growth factors, vascular endothelial growth factor, and epidermal growth factor. These participate in processes such as cell signaling, matrix regulation, and blood-vessel responses. Their biological roles provide a rationale for investigation; their presence does not demonstrate that an injected tendon will regrow or heal completely. (3)

PRP formulations differ in platelet concentration, white blood cells, activation, volume, and injection site. Studies also differ in rehabilitation and number of injections. Those differences complicate comparisons, but they should not be used to dismiss every negative result or assume that an untested protocol must work better.

A placebo-controlled trial that did not show an advantage

A 2019 Swiss trial of interstitial supraspinatus tears enrolled 80 adults and compared two ultrasound-guided injections of PRP with saline, one month apart. The injections specifically targeted the tear. (4)

PRP did not produce superior tendon-healing or clinical outcomes. The PRP group also had more reported adverse events, including prolonged pain, frozen shoulder, or extension of the lesion, during follow-up. This is important evidence for that diagnosis and approach.

It would be inaccurate to describe these injections as placed at random or to explain the result simply as a failure to use ultrasound. A hypothesis that treating additional painful structures might produce different results would need its own comparative study.

PRP compared with cortisone: results depend on the study

A study involving 129 patients compared PRP, corticosteroid, prolotherapy, and lidocaine injections for rotator cuff lesions. Corticosteroid had advantages for early symptom relief, while the authors reported longer-term advantages with PRP within their 24-week follow-up. These findings do not establish that PRP is always superior or that dextrose and steroid injections have no role. (5)

A later retrospective study reviewed 98 charts and analyzed 75 patients. The PRP group received three injections two weeks apart, while the corticosteroid group received one. Both groups improved, with no significantly better outcome in either group. Because the study was not randomized and used different injection schedules, it cannot resolve the ideal treatment or dose. (6)

A double-blind randomized trial of 99 patients with partial tears or tendinopathy found greater improvement with PRP at three months than with corticosteroid. There was no sustained advantage at 12 months and no significant difference in conversion to surgery. A short-term difference is useful information, but it is not proof that PRP repairs the tendon or prevents an operation. (7)

Corticosteroid can provide symptom relief in selected patients. Repeated injections and possible future surgery introduce additional considerations. The choice should account for the diagnosis, risks, previous response, and functional goals rather than treating either injection as the universal answer.

PRP compared with placebo, exercise, or no injection

A 2020 meta-analysis of five randomized trials found no clear short- or medium-term advantage of PRP over comparison care. Longer-term pain outcomes favored PRP, but longer-term functional improvement was not significantly better. Pain reduction and restoration of shoulder capacity should therefore be described separately. (8)

A broader meta-analysis of 17 randomized trials reported favorable pain and functional outcomes for PRP across common shoulder diseases. Its pooled results support further interest, but a review covering several diagnoses and preparation methods does not establish the same benefit for each tear pattern or for traumatic shoulder instability. (9)

Differences between reviews can reflect which studies, diagnoses, follow-up periods, and outcomes were included. Neither a favorable pooled result nor an individual negative trial should be presented as the entire evidence base.

What did the two-injection study actually show?

A 2021 prospective study followed patients treated with two PRP injections after activity modification and physical therapy had not provided adequate relief. Improvements were reported through two years. (10)

The study had no randomized comparison group. It therefore supports an observation of sustained improvement, not a conclusion that two injections work better than one, outperform rehabilitation, or prevent every partial tear from progressing.

This distinction matters for a golfer who is considering another injection after incomplete improvement. A treatment decision needs a reassessment of the diagnosis and response. It should not rest on an unsupported rule that more sessions necessarily produce a better result.

A small tendinopathy study should not be described as a tear-repair trial

A 2021 prospective study included 30 patients with persistent rotator cuff tendinopathy. Patients with a rotator cuff tear or another shoulder disorder were excluded. Participants received a landmark-guided PRP injection followed by a six-week exercise program, and symptoms and shoulder scores improved over 12 weeks. (11)

The lack of a comparison group means the separate contributions of PRP, exercise, and time cannot be determined. More importantly, the study did not test whether PRP heals a torn tendon. Retaining that distinction makes the finding useful without overstating it.

Comparing PRP with hyaluronic acid

A 2022 prospective, nonrandomized comparison evaluated a single PRP injection against three hyaluronic acid injections for partial-thickness cuff tears. Both groups also performed rehabilitation exercises and improved over three months. PRP showed an advantage in passive shoulder abduction and the Constant-Murley score at the final assessment. (12)

Because allocation was not randomized, follow-up was short, and both groups received exercise, the study does not establish that PRP is universally better or that either approach produces complete structural healing.

New 2026 research: a specific partial supraspinatus tear

A randomized trial of 63 patients studied articular-sided partial supraspinatus tears. Two ultrasound-guided PRP injections plus physical therapy produced a greater reduction in tear volume and greater improvement in the Constant-Murley shoulder score than physical therapy alone. Both groups improved. (13)

This is useful structural and clinical evidence for a defined diagnosis. It does not establish complete healing in every patient, an optimal return-to-golf date, or effectiveness for a retracted full-thickness tear following dislocation.

The site’s partial supraspinatus tear article provides a focused place for this finding. It should supplement the broader evidence rather than be generalized to every painful shoulder.

How current guidelines interpret the mixed evidence

The 2025 AAOS guideline does not support routine PRP for rotator cuff tendinopathy or partial tears and also questions routine nonsurgical PRP for full-thickness tears. These options reflect inconsistent findings and do not prohibit an individualized discussion. The 2026 partial-tear trial adds evidence published after that guideline, but one study does not resolve the remaining uncertainty. (2)

The 2025 JOSPT guideline addresses diagnosis, nonsurgical care, rehabilitation, and return to function for tendinopathy and partial-thickness tears. Its scope should not be extended automatically to a large traumatic full-thickness tear. (14)

PRP during surgical repair is a separate treatment question

Some research concerns PRP applied during a surgical repair. That setting is different from treating an unrepaired tear with an office injection. A reduction in postoperative retear rates cannot establish that PRP alone will reconnect a detached tendon.

A 2022 narrative review of PRP and an editorial by E. Hohmann discussed reduced retear rates without consistent improvement in clinical outcomes. These contributions illustrate the distinction between structural healing and how the shoulder feels or functions; they are not independent trials establishing an injection-only alternative to surgery. (3) (15)

For a golfer weighing options, the useful question is whether a proposed treatment addresses an unrepaired tendon, supports a surgical repair, or aims primarily to improve symptoms. Those are different goals.

What about bone marrow concentrate?

Bone marrow concentrate, or BMC, contains a mixture of marrow cells and other components. It is different from PRP and from a purified or culture-expanded stem cell product. Results should not be transferred automatically between these preparations or donor-derived tissue products.

In a randomized crossover trial, 51 patients with supraspinatus tears received either BMC combined with platelet products or a home exercise program. The combination group improved more at the initial three-month comparison. The small sample, crossover, combined treatment, and disclosed commercial interests limit interpretation, particularly for longer-term comparisons. The study cannot isolate the effect of BMC or establish equivalence to surgical repair. (16)

Marc Darrow and colleagues also reported outcomes from 47 patients and 50 shoulders treated with whole marrow or BMC for cuff tears or shoulder arthritis. Patients reported improvement, but there was no randomized control group and the diagnoses were mixed. The study did not establish complete tendon regeneration or a standard golf-return timetable. (17)

These findings support further investigation of selected marrow-based procedures. They do not establish that a cell-containing product is necessarily more effective than PRP for every shoulder.

When can a golfer begin swinging again?

There is no single evidence-based deadline that applies to every tear and injection protocol. Steve reported getting back to the course in roughly three months. That is his account, not a schedule another patient should copy.

A review of 84 tendinopathy studies found substantial variation in post-PRP care and no direct comparisons establishing the best protocol. Reported activity restrictions cannot be assumed to validate a particular return-to-golf timetable. (18)

A 2026 international expert consensus favors rehabilitation and gradual, individualized return to activity after PRP for chronic tendinopathy. It is consensus guidance, not a trial proving an optimal loading schedule, and it does not replace precautions appropriate to a traumatic tear or dislocation. (19)

A practical progression from daily activity to the course

Early recovery should establish comfortable daily movement and an appropriate range of motion without repeatedly provoking the shoulder. Strength and control need rebuilding as tolerated. A shoulder that is comfortable at rest may still lack endurance for repeated swings.

When sport-specific loading is appropriate, putting and short chips may provide a lower-demand starting point. Partial swings, a limited number of practice shots, fuller swings, and longer rounds introduce additional demands. This is an illustrative progression, not a validated protocol or permission to start on a particular day.

Both the immediate response and symptoms later that day or the next morning matter. A flare, new weakness, loss of motion, or a sense that the shoulder is slipping should prompt reassessment rather than an attempt to push through. Ground impact, awkward lies, carrying a bag, and repeated practice can add load beyond the swing itself.

The existing golf shoulder article provides further discussion of golf-related symptoms. The objective is sustained capacity, not simply completing one pain-free shot.

When surgery still deserves consideration

A large traumatic tear, substantial weakness, tendon retraction, recurrent instability, or failure to regain useful function can make surgical assessment important. Improvement in pain should not obscure a meaningful loss of strength or a potentially repairable injury.

Nonsurgical care can be worthwhile for many patients. Surgery can also be worthwhile when it addresses a defined problem with a reasonable prospect of improving function. The decision should account for the examination, imaging, time since injury, activity goals, and the consequences of waiting.

Steve’s return to golf illustrates a successful personal outcome. The broader lesson is that function can improve without a normal-looking scan, and that treatment choices deserve an informed discussion. Neither a testimonial nor the word “tear” should determine the decision by itself.

References