The Runner's Tendon: What the Trials Show

The road at the edge of town is empty at six in the morning, and the runner who turns onto it moves, for the first few hundred meters, like a man testing a ladder he does not trust. He is in his mid-forties. Above his right heel, the cord that connects calf to bone has been stiff on waking for a year now; it loosens somewhere in the second mile, and he has learned to call that a good day. He is a composite, this runner, a type assembled from the epidemiology rather than a person from any study, but the literature knows him in detail: the sport, the decade of life, the tendon.

The cord in question is the Achilles, the largest tendon in the body, and in a runner it works less like a rope than like a spring, stretching and recoiling with each stride, storing energy at the landing and returning it at the push-off. When it goes wrong slowly, without a single injuring moment, medicine long called the result tendinitis. The name turned out to be mistaken, and the mistake mattered, because a name is also a treatment plan.

A word, too, about the kind of exercise this story turns on. Eccentric loading means working a muscle while it lengthens rather than shortens: the braking half of movement, the controlled lowering, the descent rather than the climb. For the calf and its tendon, that means bearing weight down slowly against gravity instead of springing up against it. It is an unglamorous kind of effort, felt more as control than as exertion, and for most of the twentieth century nobody thought of it as medicine.

The word that had to change

The suffix -itis promises inflammation, and for decades the standard response followed from it: rest the tendon, calm it with anti-inflammatory drugs, wait. But when researchers examined the tissue of chronically painful Achilles tendons, the inflammation was not there. A 2000 review in Sports Medicine laid out the histology: no inflammatory cells, increased amounts of interfibrillar glycosaminoglycans, and collagen fibres whose structure and arrangement had gone disorderly; in situ microdialysis, sampling the tendon's own chemistry, confirmed the absence of inflammation. The condition was renamed tendinosis, and more broadly tendinopathy, a word that describes disrepair rather than fire.

The same review sketched who carries it and how the field was faring against it. The condition is most often seen among recreational male runners between 35 and 45, though it appears in sedentary patients as well. It was, the authors wrote, considered a troublesome injury to treat: the usual nonsurgical program of rest, anti-inflammatories, corrected alignment, stretching and strengthening rested on sparse scientific evidence, and it had been stated that nonsurgical treatment in general was not successful, with surgery required in about a quarter of patients. Surgery itself carried an asterisk: short-term results frequently very good, but the few long-term studies showed signs of possible deterioration with time, slow recovery of calf strength, and a progressive loss of heel-bone density on the operated side for up to a year.

What replaced the inflammation story was a subtler map. In a review in the British Journal of Sports Medicine, J. L. Cook and C. R. Purdam proposed that tendon pathology moves along a continuum, which would explain why people of different ages, with tendons under diverse loads, arrive with such different pain, irritability and capacity to function, and why some tendons recover with simple interventions while others remain resistant to all treatments. The model was offered explicitly for evaluation rather than as settled fact, and it gave clinicians a rational way to place treatments along the continuum. Its central kindness to patients is the permission it grants to stop comparing notes: the neighbor whose tendon settled in a month and the training partner who has fought his for two years may simply be standing at different points on the same road, needing different things from it.

The scale of the territory is easy to underestimate. The JAMA trial that appears later in this story opens with the field's own accounting: tendon disorders comprise 30 to 50 percent of all activity-related injuries. Among long-distance runners specifically, a systematic review found the incidence of lower-extremity injuries ranging from 19.4 to 79.3 percent, with the knee, not the Achilles, as the predominant site; the strong evidence on risk pointed to high weekly training distance in men and a history of previous injury, alongside the curious finding that an increase in weekly distance appeared protective against knee injuries. Running's ledger of harms, in other words, is real but tangled, and the tendon holds one well-studied page of it.

That four-fold spread in incidence, from roughly one runner in five to four in five, is itself instructive about this literature. It reflects studies that counted differently: different definitions of an injury, different populations of runners, different lengths of watching. A reader meets the same spread everywhere in sports medicine, and the sensible response is neither alarm at the high number nor comfort in the low one but attention to what, in each study, was actually counted. The same discipline serves for every result that follows.

What happened in Umeå

To feel the weight of what came next, it helps to hold the era's standard advice in mind. A painful tendon, on the inflammation theory, was an angry one, and the program that followed was pacification: rest from the provoking sport, anti-inflammatory drugs, perhaps new shoes or an orthotic, stretching, patience. The 2000 review's accounting of that program, cited above, is what makes the next study read the way it does: sparse evidence behind the usual measures, and the oft-stated claim that a quarter of patients would end up in an operating room.

The modern treatment era begins at a university hospital in Umeå, in northern Sweden, with a study of striking simplicity. In 1998, H. Alfredson and colleagues took 15 recreational athletes, twelve men and three women with a mean age of 44.3, all carrying chronic Achilles tendinosis that had outlasted conventional care, and all, at the start, in too much pain to run at all. The prescription was not rest. It was twelve weeks of heavy-load eccentric calf training: loading the tendon specifically as the muscle lengthens, the braking phase of the movement, deliberately and progressively heavy.

After twelve weeks, all fifteen were back at their pre-injury levels, running fully. Pain during activity had fallen significantly, and the injured leg's strength had risen until it no longer differed from the healthy side. The comparison the authors offered was blunt: fifteen other recreational athletes with the same diagnosis and similarly long symptoms, treated conventionally with rest, anti-inflammatories, shoe changes, orthoses, physical therapy and ordinary training programs, in whom conventional treatment succeeded not once, and all of whom ultimately went to surgery. The design deserves its caveats, a prospective study of fifteen with a comparison group rather than randomization, and the authors confined their own claim to “a very good short-term effect on athletes in their early forties.” Fifteen-for-fifteen against none-for-fifteen is the kind of result that demands randomized trials rather than replaces them. The remarkable thing is that the randomized trials, when they came, largely obliged.

Two roads, one destination

The most instructive of them ran in Copenhagen. In a 2015 randomized trial in the American Journal of Sports Medicine, 58 patients with chronic midportion Achilles tendinopathy were assigned to twelve weeks of either the classic eccentric program or heavy slow resistance training, a gym-based alternative that loads the tendon in both directions, slowly, under substantial weight. The two programs express different philosophies of the same idea: the eccentric protocol isolates the lowering phase in a spartan homework of repetitions, while heavy slow resistance moves the work into a gym, fewer sessions built around genuinely heavy loads moved at a deliberate tempo through the full motion, both of them arguments made to the tendon in its own mechanical language. Both groups improved significantly in validated function and pain scores, and the improvements held at one year. Alongside the clinical gains, the tendons themselves changed: thickness and the ingrowth of new blood vessels, both marks of the disorder, decreased. On none of these measures did the two programs differ.

Where they did differ was in what patients would actually do. Compliance with training sessions was 92 percent under heavy slow resistance and 78 percent under the eccentric program, a significant difference; patient satisfaction leaned the same way at twelve weeks, 100 percent versus 80, though that lean did not reach statistical significance then and faded further by one year. Compliance, in the end, was the one significant difference between two programs whose results did not differ.

A companion trial from the same Copenhagen group had already run the harder comparison at the knee. Thirty-nine men with patellar tendinopathy were randomized to peritendinous corticosteroid injections, eccentric decline-squat training, or heavy slow resistance. At twelve weeks everyone had improved, the cortisone group included. By the half-year follow-up, the two loading groups had held their gains and the cortisone group had deteriorated. Heavy slow resistance, alone among the three, was accompanied by an elevated turnover of the collagen network, the tendon's structural protein being taken apart and rebuilt, and satisfaction at half a year was highest in that group. The authors' summary was measured and damning at once: corticosteroid injection has good short-term but poor long-term clinical effects in patellar tendinopathy. The treatment with the good short-term effect, in this trial, was not the one with the good long-term effect. Two of the trial's quieter findings fill in the picture. The mechanical properties of the tendons, tested directly, were similar in healthy and injured tissue and unchanged by any treatment, a reminder of how loosely symptoms and structure can travel together in this condition. And the elevated collagen turnover under heavy slow resistance is the closest thing the literature offers to a mechanism seen in the act: the tissue's structural protein being dismantled and relaid while the patient's pain scores fell.

The permission slip

For a runner, the harshest sentence in the standard advice was always the first one: stop running. A Swedish randomized trial published in 2007 tested whether that sentence was necessary. Thirty-eight patients with Achilles tendinopathy all received the same graded loading rehabilitation; half were additionally required to stop running and jumping for the first six weeks, while the other half were allowed to continue those activities under a pain-monitoring model, a set of rules that permits training so long as pain stays within defined, tolerable bounds and settles afterward. The groups improved at statistically indistinguishable rates. On the hundred-point function score, the group that kept running went from 57 to 85 over a year, the resting group from 57 to 91, with no significant difference between them, and the authors' conclusion was the one runners had waited a generation to read: no negative effects could be demonstrated from continuing tendon-loading activity under the model. The finding is a permission, not a prescription; the trial found no demonstrable harm in continuing, not that continuing was superior.

The pain-monitoring model itself deserves a sentence of appreciation, because it reframes what pain is for. Under the old advice, pain was an alarm, and the only correct response to an alarm is to stop. Under the model, pain within tolerable, defined bounds that settles afterward becomes information, a gauge to train by rather than a siren to flee. That reframing turns out to have measurable stakes.

The same Gothenburg group later reported what became of patients treated with exercise alone, following 34 of them for five years in a case series, a design that observes rather than compares. Eighty percent had fully recovered from the initial injury, 65 percent with no symptoms at all; only two had sought any additional treatment. One correlation from that follow-up lingers: the more a patient feared movement, on a standard kinesiophobia scale, the less completely their heel-rise capacity recovered, a negative correlation of 0.590. The authors drew the practical inference in favor of the pain-monitoring model itself, a structure that gives an anxious tendon-owner something better than fear to steer by.

The needle that could not beat salt water

Against the patient work of loading programs, the market has always offered something faster, and in the late 2000s the leading candidate was platelet-rich plasma, described in the trial report itself as an increasingly used treatment: the patient’s own blood, spun down to concentrate the platelets and their growth factors, injected into the degenerated tendon in the hope of provoking regeneration. In 2010, JAMA published the trial the field needed. Fifty-four patients with chronic midportion Achilles tendinopathy, all performing eccentric exercises as usual care, were randomized double-blind to a PRP injection or to a saline one. Over 24 weeks the PRP group improved by 21.7 points on the hundred-point function score. The saline group improved by 20.5. The adjusted difference between them was 0.9 of a point (95 percent confidence interval, −12.4 to 10.6), an interval that excluded the 12-point benefit the investigators had defined in advance as clinically relevant. Both groups got substantially better while performing the exercises both arms shared; the platelets added nothing that salt water did not.

Who gets it, as far as anyone knows

The question of why one runner’s tendon fails while a training partner’s holds has an unsatisfying answer, delivered honestly by a 2019 systematic review in the British Journal of Sports Medicine. From 5,111 publications, the reviewers found ten cohort studies worth including, every one at high risk of bias, and extracted limited evidence for nine clinical risk factors, among them a prior lower-limb tendinopathy or fracture, use of the quinolone antibiotic ofloxacin, moderate alcohol consumption, training in cold weather, and decreased plantar-flexor strength, along with certain gait features and factors specific to heart-transplant patients. Just as telling was the list of twenty-six suspected factors that showed no association, including being overweight, static foot posture and overall physical activity level, three of the explanations most often offered across a shop counter. The reviewers singled out ofloxacin use, alcohol and plantar-flexor weakness as the potentially modifiable entries. The honest summary of the risk literature is that it is thin, and that several of its firmest folk beliefs have no evidentiary floor under them. A runner who has been told that his weight, his arches or simply his running are the reason his tendon aches is being offered explanations the cohort evidence, such as it is, declined to support; the factor with some of the strangest support, a common antibiotic, is one almost nobody would name unprompted.

One demographic note runs through all of this and lands close to home for the readers of a longevity site. The evidence base for tendon loading is not built on college athletes. Alfredson’s fifteen averaged 44.3 years old; the five-year exercise cohort averaged 51. These are studies of middle-aged tendons, which is precisely the population in which the condition concentrates, and their results sit alongside the broader questions of training in midlife taken up in a companion review of joint and tendon recovery for masters athletes on this site.

The slow spring

Set side by side, the studies describe a tissue with a particular temperament. The tendon did not respond to the conventional care of the 1990s; in the nonrandomized Umeå comparison, that program failed in all fifteen patients, every one of whom ultimately had surgery. Injections have not held up either: cortisone’s benefit had deteriorated by six months in the patellar trial, and platelets could not distinguish themselves from saline. What improved, in trial after trial, was the loaded tendon: heavy, slow, patient work in programs measured in twelve-week blocks, under which the Copenhagen investigators found the collagen network itself turning over, the structure being rebuilt rather than merely soothed. Which brings the story back to the road at the edge of town, and to the runner moving carefully through his first quarter mile. In these studies, the loaded tendons did as well as or better than the rested ones, and the injected tendons did no better than the loaded ones. That finding keeps a slow calendar, and the studies are candid about its terms: twelve weeks to the first verdict, a year to the durable one, five years before a cohort could report that four of five had fully recovered from the initial injury on exercise alone, 65 percent of them symptom-free. The tendon keeps slow time; the evidence, patiently gathered over more than two decades from Umeå to Copenhagen to The Hague, keeps it too.

Educational, not medical advice.