Usain Bolt and the Wrong Leg

What is actually documented

Start with what nobody disputes, because it is more interesting than the theory built on top of it. Bolt wrote in his autobiography that he has scoliosis, and that the curve leaves his right leg about 13 millimetres shorter than his left. In a 2011 interview he was plainer than most athletes are about anything: his spine is really curved, it got worse as he grew, and if he keeps his core and back strong it does not bother him.

He also said something that gets quoted less often. Early in his career, before anyone around him understood the condition, he got injured every year. That is a real pattern, and it matters later, though not for the reason the popular story gives. Read it again and notice what he is describing: not a spine that tears hamstrings, but a young athlete training through problems nobody had characterised yet, accumulating exactly the injury history that the risk literature treats as the strongest predictor there is. He is describing the boring explanation, in his own words, years before the interesting one was attached to him.

The injury itself is documented in his own words, posted the day after the race. A tear of the proximal myotendinous junction of biceps femoris in his left hamstring, with partial retraction. Three months of rehabilitation. That is the whole medical record available to the public, and it is worth noticing what it does not contain: no grade, no mention of scoliosis, no causal claim of any kind.

Then there is the asymmetry, which is real and was measured properly. A biomechanics group at Southern Methodist University studied his stride and found his left leg stays in contact with the ground about 14 percent longer than his right, while his right foot strikes with more peak force, roughly 1,080 pounds against 955. The researchers described this as a difference in how he generates his impulses. They did not connect it to injury. Asked whether the asymmetry helped or hurt him, the honest summary of that work is that it reached no firm conclusion.

The chain, link by link

The popular explanation has four links. Scoliosis tilts the pelvis. The tilt shortens a leg. The short leg overloads a hamstring. The overloaded hamstring eventually tears. It sounds like anatomy. Take the links one at a time.

One: the direction is probably backwards

Reviews of leg-length discrepancy describe discrepancies above roughly 20 millimetres producing altered loading and, among the results, functional scoliosis. That is the leg causing the curve, not the curve causing the leg. Telling the two apart in a given person is not a matter of opinion, it requires standing radiographs and block correction tests, which is why a 2026 systematic review exists on exactly that diagnostic problem. Nobody can determine the direction in Bolt from public information, and our article asserted it flatly.

Two: 13 millimetres is below the line

A review of leg-length inequality concluded that anatomic discrepancy does not appear to be clinically significant until it reaches roughly 20 millimetres, and that small discrepancies are close to universal, averaging around 5 millimetres in the population. Bolt's reported 13 millimetres is real and is larger than most. It is also below the threshold the review sets. A separate systematic review of leg-length discrepancy and stability found the published data inconsistent.

Three: asymmetry has been tested, and it did not predict injury

This is the link that has actually been examined properly, and it is where the chain breaks in the open. A 2024 study instrumented 836 recreational runners and followed them for six months, recording 107 injuries. Leg-length discrepancy was among the leading determinants of the gait asymmetry they measured. Greater asymmetry was not associated with higher injury risk. On two variables, flight time and peak braking force, greater asymmetry was associated with slightly lower risk.

It is not one study. An earlier cohort of 87 runners found no evidence that static biomechanical alignment of the lower limbs related to lower-limb injury, with patellofemoral pain the one exception. A prospective study of 263 footballers measured nine postural variables, scoliosis among them, and concluded that static postural deviations are not typically associated with non-contact lower-limb injury risk and are unlikely to add value to pre-season screening. A cohort of 207 rugby league players reached the same place: given the lack of association with injury, these deviations may not warrant intervention.

Four: nobody has connected it to hamstrings at all

The largest synthesis of hamstring injury risk factors pooled 78 studies, 8,319 hamstring injuries and 71,324 athletes. Its conclusion names older age and a history of hamstring injury as the strongest risk factors. Scoliosis is not in it. Spinal curvature is not in it. Pelvic obliquity and leg-length discrepancy are not in it. The 2023 international consensus on hamstring injuries does not mention them either.

What actually predicts a torn hamstring

Here is the part that makes the exotic explanation unnecessary. The established risk factors were sitting in plain view the whole time, and Bolt had them in abundance.

Older age is one, and he was thirty, which for a sprinter is late. A history of hamstring injury is the other, and by his own account he was getting injured every year early on. A recent hamstring injury raises the risk further still, and he had been managing hamstring trouble through that final season. Previous calf strain and previous ACL injury also appear in the pooled data. None of this requires a curve in the spine. It requires a thirty-year-old sprinter with a long injury history running the anchor leg of a relay.

1. Established

   History of hamstring injury

   Relative risk 2.7 across the pooled studies. A recent one raises it to 4.8.

2. Established

   Older age

   The other consistent prospective signal, and the one no training programme reverses.

3. Contested

   Eccentric strength and fascicle length

   A 2016 cohort found both predictive. A 2021 pooled analysis of 1,100 athletes found no difference between injured and uninjured, and a 2024 study found fascicle length unrelated to risk.

4. Absent

   Scoliosis and leg length

   Not named in the risk-factor meta-analysis, not in the international consensus, and not supported by any prospective study located.

Notice that the honest list includes a genuine disagreement. Eccentric hamstring weakness and short biceps femoris fascicles were reported as strong predictors in a 2016 cohort of elite footballers. A later pooled analysis of six studies and 1,100 athletes found no difference in eccentric strength between the injured and the uninjured, and a 2024 study of college American football players found fascicle length unrelated to injury risk. A four-year cohort of 614 professional players concluded that strength deficits are weak risk factors that do not let you identify an individual player at risk. That is what a live scientific question looks like, and it is more useful to a reader than a tidy story.

What the evidence does support

The advice in our earlier version was mostly reasonable. It was simply justified by the wrong thing. Strip out the mechanism and the guidance that remains is well evidenced, with one caveat worth stating.

The Nordic hamstring exercise is the most studied preventive intervention there is. A 2019 review of 15 studies and 8,459 athletes found it roughly halved hamstring injury rates, a relative risk of 0.49, holding at 0.55 after removing the studies at high risk of bias. That is a large effect for a single exercise.

The caveat: a 2021 methodological reappraisal argued that only five of those studies were true randomised trials, and reanalysing those gave a relative risk of 0.56 with a confidence interval running from 0.20 to 1.52. That interval crosses one, meaning the protective effect is not established at the standard the reappraisal demands. Both papers are real and both are worth knowing. And separately, a survey of elite teams found that only one club in seventeen actually ran the full original programme, which is its own kind of finding.

On coming back, the consensus is that return to sport should be criteria-based rather than calendar-based: medical clearance, no pain on testing, restored flexibility, performance on field testing, and psychological readiness. Functional performance ranks highest among those in a survey of practitioners. Protocols guided by a specific active-flexibility test have reported the lowest reinjury rates, in the range of one to four percent, against a background where recurrence across prospective studies has ranged from 14 to 63 percent.

Which brings the point home. In elite football, reinjuries are around 16 percent of hamstring injuries, and every recorded reinjury in one large study affected the same muscle, biceps femoris. The muscle Bolt tore. The thing that predicts a hamstring tear best is a hamstring that has torn before, and the intervention with the best evidence is loading it, patiently, for longer than feels necessary.

So the scoliosis is real, the shorter leg is real, the asymmetry in his stride is real and measured. What is not real is the line drawn between them and the muscle that failed. He was a thirty-year-old sprinter with years of hamstring trouble, running one last race in the cold. That explanation is duller. It is also the one the evidence supports, and it is the one that generalises to you.

None of this is medical advice, and none of it substitutes for assessment by a clinician who can examine you. If you have scoliosis, the honest position is that the prospective evidence does not show it predicts lower-limb muscle injury, and that no screening value has been demonstrated for static postural deviations. Train for the risk factors that are established, not for the ones that make a better story.