A torn rotator cuff is one of the most common reasons adults end up in a shoulder surgeon's office. The pain at night, the weakness lifting your arm overhead, the MRI report with the word "tear" on it: sooner or later the conversation turns to arthroscopic repair. Many patients then ask a fair question: could a stem cell injection heal the tendon instead, and spare me the operation and the long rehab?
The honest answer depends heavily on what kind of tear you have. This article walks through what the research shows for surgery, for physiotherapy, and for stem cells, both as an injection on its own and as an add-on during surgery, so you can see where each option is strong and where the evidence runs out.
The honest headline: for small-to-medium full-thickness tears, the longest randomized trial found that surgical repair gave better results than physiotherapy at 10 and 15 years. For partial-thickness tears, small early studies of stem cell injections are encouraging but are not yet randomized against surgery. The best-studied use of stem cells in the rotator cuff is actually alongside surgery, where they appear to lower re-tear rates without clearly improving pain or function.
The quick answer
If you have a full-thickness tear, especially after an injury, in a younger or active person, or with clear weakness, a surgical opinion should come first. If you have a partial-thickness tear or tendinopathy and want to try something between physiotherapy and surgery, a stem cell injection is a reasonable option to discuss, as long as you understand the evidence is early.
Stem cells and surgery are less "either/or" than they sound. Several of the most cited stem cell studies in the shoulder tested cells during a surgical repair, to help the tendon heal back onto the bone, not as a replacement for the operation.[7][8]
First, what kind of tear do you have?
The rotator cuff is a group of four tendons that hold the ball of the shoulder in its socket and help lift and rotate the arm. "Rotator cuff tear" covers a wide range of problems:
- Tendinopathy: the tendon is thickened and painful but not torn.
- Partial-thickness tear: part of the tendon is frayed or split, but it is still attached to the bone.
- Full-thickness tear: a hole goes all the way through the tendon. Small and medium tears may still work reasonably well; large tears can retract and lose function.
- Massive or retracted tear: the tendon has pulled back and the muscle may have started to turn to fat, which makes any repair harder.
Age matters too. Degenerative tears, which develop slowly with age, are very common in people over 60 and do not always cause symptoms. Traumatic tears, from a fall or a heavy lift, behave differently and are more often repaired quickly. Our general guide to stem cells for rotator cuff injuries explains in more detail what the imaging needs to show.
Side by side
| Arthroscopic rotator cuff repair | Stem cell injection (no surgery) | |
|---|---|---|
| What it does | Stitches the torn tendon back onto the bone with anchors | Aims to support healing in the existing tendon and reduce pain |
| Strength of evidence | Moderate: several randomized trials against physiotherapy, with long follow-up in one | Limited: small early-phase studies and one industry-run randomized trial against exercise |
| Best suited to | Full-thickness tears, traumatic tears, younger or active patients, clear weakness | Partial-thickness tears and tendinopathy; possibly small, non-retracted tears |
| Procedure | Keyhole surgery under anaesthetic, usually day case | Outpatient cell collection and ultrasound-guided injection |
| Recovery | Sling for around 4–6 weeks; strengthening over months; full recovery often 6–12 months | Usually no sling; a few days of soreness; any benefit judged over 3–6 months |
| Main limitations | Re-tears are common, especially in large tears and older patients | Few controlled trials; durability beyond two years largely unknown |
| Main risks | Stiffness, infection, nerve injury, anaesthetic risks, failed repair | Soreness at the harvest and injection sites; serious events uncommon in published studies |
What the surgery evidence shows
Rotator cuff repair is a very common operation, but its evidence base is more mixed than many patients expect.
Surgery vs physiotherapy in the short term
A 2019 Cochrane review pooled randomized trials comparing repair with exercise-based treatment, mostly in people with small, degenerative supraspinatus tears. At one year, surgery probably made little or no difference to pain, and the improvement in function was small. The reviewers stressed that the trials had methodological problems and that their conclusions may not apply to traumatic tears, large tears or younger people.[1]
Surgery vs physiotherapy in the long term
The picture changes with time. In a Norwegian trial, 103 patients with full-thickness tears no larger than 3 cm were randomly assigned to tendon repair or to physiotherapy (with surgery later if needed). At 10 years, the repair group had better shoulder scores, less pain and better pain-free movement, and the gap between the groups had grown since the 1- and 5-year results.[2] At 15 years the advantage for repair persisted, and in shoulders treated with physiotherapy alone the average tear had roughly doubled in size, from about 16 mm to 32 mm.[3]
That trial is one of the strongest reasons to take a full-thickness tear seriously. Tears tend to grow, and a tear that is repairable now may not be repairable in five or ten years.
The re-tear problem
Surgery does not always hold. A large systematic review of 108 studies covering more than 8,000 repaired shoulders found an average re-tear rate of about 27% at around two years, with higher rates in older patients, larger tears and tendons with more fatty change in the muscle. Interestingly, patient-reported results generally improved whether or not the repair stayed intact.[4] This re-tear problem is exactly what much of the stem cell research in the shoulder has tried to address.
What the stem cell evidence shows
Stem cell research in the rotator cuff falls into two quite different groups, and it is important not to blur them.
1. Stem cells added during surgery
Here, cells (usually concentrated from bone marrow or fat) are applied at the repair site during an arthroscopic repair, with the aim of helping the tendon heal back to bone.
- Hernigou et al., 2014. 45 patients received bone marrow concentrate during their repair and were compared with 45 matched patients who did not. All 45 treated repairs had healed by six months versus 30 of 45 controls, and at ten years 87% of treated cuffs were intact versus 44% of controls. This was a case-controlled study, not a randomized trial.[5]
- Kim et al., 2017. 35 patients given fat-derived stem cells in fibrin glue during repair were matched with 35 who had repair alone. MRI re-tear rates were 14.3% versus 28.5%, but there were no differences in pain, range of motion or shoulder scores at around 28 months.[6]
- Two 2025 reviews pooled the comparative studies. Both found similar pain and function with or without stem cells, and fewer re-tears in the stem cell groups. One review noted that the literature is still preliminary and the certainty of evidence is limited.[7][8]
The fair reading: as an add-on to surgery, stem cells may help the repair stay intact, but so far that has not translated into clearly better pain or function scores.
2. Stem cell injections instead of surgery
This is what most patients mean when they ask about "stem cells vs surgery", and it is where the evidence is thinnest.
| Study | Design | What it found |
|---|---|---|
| Jo et al., Stem Cells, 2018[9] | First-in-human dose-escalation trial; 18 analysed patients with partial-thickness tears; fat-derived cells injected into the tendon; no control group | No treatment-related adverse events; pain and function scores improved substantially in the mid- and high-dose groups; MRI showed shrinking tendon defects at the high dose |
| Jo et al., Arthroscopy, 2020[10] | Two-year follow-up of the same programme (19 patients) | Improvements in pain and function maintained at two years; no patient needed surgery during follow-up |
| Centeno et al., BMC Musculoskelet Disord, 2024[11] | Randomized crossover trial; 51 patients with high-grade partial or full-thickness supraspinatus tears; bone marrow concentrate plus platelet products vs home exercise | Greater improvement in pain and function than exercise at three months; gains continued to two years. Authors are employed by or hold shares in Regenexx, which sells this treatment |
These results are encouraging, and the safety signal is reassuring. But the numbers are small, the Jo studies had no control group, the only randomized trial compared injections with home exercise rather than with surgery, and it was run by a company with a commercial interest in the outcome. No randomized trial has yet compared a stem cell injection directly with rotator cuff repair. Anyone who tells you injections are "as good as surgery" is going beyond the published evidence.
Cell injections are also used for tendon problems elsewhere in the body; our article on stem cells for tendinopathy covers that wider evidence.
Who may be a reasonable candidate for an injection
Based on the published studies, the patients most likely to be reasonable candidates for a stem cell injection rather than surgery tend to:
- Have tendinopathy or a partial-thickness tear confirmed on MRI or ultrasound, or a small full-thickness tear that has not retracted
- Still have good strength, with pain being the main problem rather than weakness
- Have tried at least three months of structured physiotherapy without enough improvement
- Have a degenerative rather than a sudden traumatic tear
- Have a clear reason to avoid or delay surgery, such as work, caregiving or medical risk, and accept that surgery may still be needed later
- Be in reasonable general health, with no active infection or cancer
When this may not be right for you
We would recommend seeing a shoulder surgeon rather than travelling for an injection if:
- Your tear followed an injury and you suddenly cannot lift your arm. Acute traumatic tears often do best with early repair, and waiting can make a repair harder.
- The tear is full-thickness and you are younger or physically active. The long-term trial evidence favours repair for small-to-medium full-thickness tears.[3]
- The tendon has retracted or the muscle shows significant fatty change on MRI. An injection cannot pull a retracted tendon back to the bone.
- You have marked weakness, or the shoulder has started to ride up in the socket on X-ray.
- You have shoulder arthritis, a frozen shoulder or nerve-related pain from the neck as the main problem; these need a different assessment.
- You have an active infection, active cancer, are pregnant or breastfeeding, or have a bleeding disorder that makes collection or injection unsafe.
- You would only consider it worthwhile if it guaranteed you would never need surgery. No honest clinic can promise that.
If you are not sure which group you fall into, a surgeon's opinion is valuable even if you lean towards a non-surgical route. We are happy to review your imaging and tell you if we think surgery is the better option.
How to think about the decision
It can help to ask three questions in order. First: is this a tear that is likely to grow or that needs mechanical repair? For full-thickness tears in active people, the honest answer is often yes, and that points to surgery. Second: if it is a partial tear or tendinopathy, have I given physiotherapy a fair chance? Many people improve with a well-designed exercise programme. Third: if not, what am I hoping an injection will achieve? A realistic hope is less pain and better function, with a chance of avoiding or delaying surgery. An unrealistic one is a guaranteed rebuilt tendon.
Before any injection, it is also worth checking that the cells are processed to proper standards. Our article on whether stem cell injections are safe summarises the side-effect research.
How our medical team assesses shoulders, knees, hips and other joints, and what treatment involves in Medellín, is set out on our page on stem cell therapy for shoulder, knee and joint injuries.
The bottom line
Rotator cuff repair and stem cell injections are not really competing for the same patient. For full-thickness tears, especially in younger or active people, surgery has the stronger long-term evidence. For partial tears and tendinopathy that have not settled with physiotherapy, stem cell injections are a promising option with encouraging early results and real uncertainty about how well, and how long, they work. And when surgery is needed, adding cells to the repair may help it hold, though that has not yet been shown to make patients feel better.
If you are unsure where your shoulder falls, send us your MRI or ultrasound report. We will tell you honestly whether an injection is reasonable, what it might and might not achieve, and when we think you would be better served by a shoulder surgeon.
