An injury that never finishes healing has usually stalled in the remodeling phase, the last stage of repair. The temporary collagen is never replaced and lined up. That is because the tissue lacked one of two things. Either it lacked the metabolic budget to build new tissue, or it lacked the load that tells the fibers which way to point.
Most injuries heal. That is worth saying first. A page like this draws people for whom it did not happen, and the odds matter. In pooled data on first-time outer ankle sprains, lasting pain fell from 48.6% at three months to 6.7% at twelve. The curve keeps improving for a year.
This article is about the tail end of that curve. What is different in the biology of injuries that are still there when everyone else has forgotten theirs?
What are the stages of injury healing, and where do they stall?
Inflammatory, proliferative, remodeling. The first two are strong. They run on autopilot in almost everyone. The third is slow, has conditions, and depends on inputs that are easy to get wrong.
In remodeling, the quick, temporary type III collagen laid down earlier is slowly replaced by type I. The new fibers line up in the direction of load. Two inputs control it. One is a metabolic budget big enough to build and link the new tissue. The other is a load signal telling the fibers which way to point. Take away either one, and the temporary tissue simply stays.
Why does an injury that looks fine on a scan still hurt?
The tissue that results is not inflamed. That is why the ending in tendonitis is misleading, and why anti-inflammatory approaches aimed at it disappoint. What you find instead is messy collagen and extra filler. You also find neovascularization — small vessels growing into tissue that should have few of them — and, next to those vessels, new nerve fibers.
That last detail explains a lot of confusion. The structure now has nerves it did not have before. So it causes pain out of proportion to how it looks. A patient describes severe pain. A scan shows a tendon that looks okay. Someone concludes the problem is exaggerated. The problem is real. It is not mainly a strength problem or an inflammation problem. It is a failed-remodeling problem.
Why is my injury not healing?
1. The mechanical signal was withdrawn
Remodeling collagen lines up with load. Take away load, and nothing tells the fibers where to point. Load can go away through avoiding pain, keeping the injured part still, or a rehab program that stopped when walking stopped hurting. This is the most common cause you can reverse, and the least exciting.
It is also the one with the best treatment evidence. A 2026 network meta-analysis compared exercise plans for patellar tendinopathy (kneecap tendon pain). No plan clearly beat heavy slow resistance training. The authors warned that the rankings are only exploratory. Read closely, that is a useful finding, not a letdown. The exact plan matters less than actually adding load step by step.
2. The metabolic budget was not there
Building and linking collagen is costly. It needs raw materials and the right hormone and inflammation setting. Pooled risk factors for rotator cuff tendon problems turn out to be more about body chemistry than mechanics. Insulin resistance, poor sleep, low vitamin D and ongoing nicotine use are the four that keep showing up. Each has its own article — vitamin D, sleep and nicotine.
3. The problem was never the structure everyone was treating
A repair cannot stall in a tissue that was never injured. Sometimes a diagnosis was assumed from a scan report. It was not proven by an exam and, when needed, by response to a test block. Then months of correct treatment can go to the wrong target.
4. There was too much load, applied wrongly
This is the mirror image of the first cause, and less common. Steady heavy load on a tissue in the middle of remodeling keeps it stuck in the earlier building stage. Telling progressive loading apart from repeated re-injury is a medical judgment, not a formula.
Where a catalyst fits, and where the evidence says it does not
A stalled repair is the state a biologic is designed for. It is the state the positive trials recruited. It is also a field where results vary more than they should. The most common reason is not the biology. It is that the treatments being compared were never the same treatment. And the patients getting them were never sorted by whether their terrain could carry out the order.
Part of the spread is that trials in this field are poorly described. One review looked at randomized trials of PRP for knee osteoarthritis. It found they follow the MIBO reporting standard badly. So pooled analyses compare treatments that were never the same treatment. Why the preparation is a medical decision here is covered in Why PRP Results Vary.
How do you restart an injury that stopped healing?
Confirm the target. Fix the terrain. Bring back progressive load and keep it going past the point where it stops hurting. Then, when a target and a window justify it, give a well-described trigger shot under imaging guidance. Doing this in reverse order is the most common way regenerative treatment is wasted.
Frequently Asked Questions
What does it mean when a doctor says my injury has healed but it still hurts?
Usually that the first injury settled down, but the remodeling phase did not finish. That leaves messy tissue with ingrown vessels and nerve fibers. It acts nothing like the original problem. Why that tissue is not inflamed the way the word tendonitis suggests is covered in Tendon PRP versus intra-articular PRP.
Why did my injection not work?
There are three common reasons, in about this order. The target was never confirmed. Nobody measured the shot. Or the body could not act on the signal. Each is covered in Why PRP Injections Fail.
Does smoking really matter for a tendon injury?
Yes, and nicotine itself, not just smoke. That is because nicotine narrows blood vessels and weakens collagen linking. Both act directly on the process that needs to finish. The details, including vaping and pouches, are in Nicotine and smoking effects on tendon and bone healing.
Is a stalled injury still treatable years later?
Yes. Every positive trial here took in people whose problem had lasted months or years. So a long history qualifies you for a look. It does not rule you out. What changes with time is the realistic target. That difference is explained in Realistic recovery timelines after orthobiologic treatment.
How do you tell a stalled repair from something that was misdiagnosed?
By re-checking how the injury works, not rereading the report. That is because a repair that never started on the right structure is a different problem from one that started and stopped. The reasoning is the same as in Your Scan Came Back Clean and You Are Still in Pain.
Is it possible for an injury to never heal?
Most injuries do heal. After a first ankle sprain, for example, lasting pain drops from 48.6% at three months to 6.7% at twelve. A small group is still hurting when everyone else has moved on. In those injuries, the last stage of repair, remodeling, stalled. The temporary collagen was never replaced and lined up. So the tissue stays painful until the cause is fixed.
What are the signs an injury is not healing properly?
The clearest sign is pain that is still there long after the normal recovery curve should have flattened out. Another is pain that feels far worse than the scan suggests. That mismatch happens because a stalled repair grows new small blood vessels and nerve fibers into the tissue. So it hurts out of proportion to how it looks. The pain is real, and it is not a sign of exaggeration.
What should I do if my injury is not healing?
Follow the order that restarts repair. First, confirm which structure is really injured. Use an exam and, when needed, a test block, not just a scan report. Second, fix the terrain: insulin resistance, poor sleep, low vitamin D and nicotine. Third, bring back progressive load and keep it going after the pain stops. Only then, when a target justifies it, is a trigger shot considered.
Find out what your recovery is actually limited by
Regen.MD begins with a physician-led Clinical Evaluation — a review of your history, your imaging and your metabolic data, and a written terrain roadmap before any procedure is discussed. Evaluation is contingent upon review of your data.
Questions? Call or text us.
Sources
- Liu Y, Li C, Yang F. Comparative effectiveness of exercise interventions for patellar tendinopathy: a systematic review and network meta-analysis of randomized controlled trials. BMC Sports Sci Med Rehabil, 2026. PubMed 42192475 doi:10.1186/s13102-026-01743-4
- Leong HT, Fu SC, He X, et al. Risk factors for rotator cuff tendinopathy: a systematic review and meta-analysis. J Rehabil Med, 2019. PubMed 31489438 doi:10.2340/16501977-2598
- D'Souza RS, Her YF, Hussain N, et al. Evidence-based clinical practice guidelines on regenerative medicine treatment for chronic pain: a consensus report from a multispecialty working group. J Pain Res, 2024. PubMed 39282657 doi:10.2147/JPR.S480559
- Manchikanti L, Navani R, Navani A, et al. Comprehensive evidence-based guidelines for regenerative therapies in the management of chronic low back pain: 2025 update from the American Society of Interventional Pain Physicians (ASIPP). Pain Physician, 2025. PubMed 41481869
- Stone AV, Abed V, Owens M, et al. Randomized controlled trials on platelet-rich plasma for knee osteoarthritis poorly adhere to the Minimum Information for Studies Evaluating Biologics in Orthopaedics (MIBO) guidelines: a systematic review. Am J Sports Med, 2024. PubMed 38282598 doi:10.1177/03635465231185289
- Michels F, Wastyn H, Pottel H, Stockmans F. The presence of persistent symptoms 12 months following a first lateral ankle sprain: a systematic review and meta-analysis. Foot Ankle Surg, 2022. PubMed 34961654 doi:10.1016/j.fas.2021.12.002
- Bensa A, Sangiorgio A, Deabate L, et al. PRP injections for the treatment of knee osteoarthritis: the improvement is clinically significant and influenced by platelet concentration — a meta-analysis of randomized controlled trials. Am J Sports Med, 2025. PubMed 39751394 doi:10.1177/03635465241246524