What did the study find?

Higher self-reported physical activity was associated with a lower risk of ALS (hazard ratio for high versus low activity 0.77, 95% CI 0.61 to 0.96). Device-measured activity pointed the same way (HR per 1 SD 0.75, 95% CI 0.58 to 0.97, in 96,570 people with 98 ALS events).

Does exercise cause motor neuron disease?

This study found the reverse association: more activity went with less disease. The suspicion came largely from observations of professional athletes, soldiers and very high performers, which this analysis cannot fully speak to, because a general population cohort contains few people training at that level.

Why was the middle of the range best?

The relationship for self-reported activity was non-linear, with lowest risk in those in the mid-range. That shape leaves room for the older worry at the very top end while still showing that being sedentary is not protective.

What about people at high genetic risk?

The association was similar in people carrying two copies of a risk variant in the UNC13A gene. In carriers of the C9ORF72 repeat expansion, the largest single genetic cause, the direction was the same but the result was not statistically significant, and only 535 carriers with 56 events were available.

Could early disease be causing the low activity, rather than the other way round?

That is the central objection and the study cannot rule it out. Motor neuron disease damages nerves for years before diagnosis, so people destined to develop it may already be moving less when their activity is measured. A 14-year follow-up narrows that window without closing it.

How common is motor neuron disease?

Rare. There were 541 diagnoses among 384,836 people over 14 years. That rarity is why a cohort this large was needed, and why the estimates still carry wide margins.

Should anyone change what they do?

Nothing here is a reason to start or stop exercising for the sake of your motor neurons. The reasons to be active are already well established and sit elsewhere. This is general information rather than medical advice.

Physical activity tracked 23% less motor neuron disease across 384,836 UK adults

Exercise has long been suspected of raising motor neuron disease risk, mostly from studies of elite athletes. An Oxford team followed 384,836 people for 14 years and found the opposite, with the lowest risk mid-range.

For twenty years the uncomfortable question about motor neuron disease has been whether exercise brings it on. Footballers get it. Soldiers get it. The pattern was noticed often enough that the role of physical activity in the risk of amyotrophic lateral sclerosis is debated in the literature rather than settled.

Writing in the Journal of Neurology, Neurosurgery & Psychiatry, an Oxford-led team put the question to 384,836 people. Higher self-reported physical activity was associated with a lower risk of ALS, by roughly a quarter.

What motor neuron disease does

Amyotrophic lateral sclerosis, usually shortened to ALS and known in Britain as motor neuron disease, attacks nerve cells called neurons in your brain and spinal cord.

Those particular cells have one job. They transmit messages from your brain and spinal cord to your voluntary muscles, the ones you choose to move. As they fail, the muscles they command stop receiving instructions, and weakness spreads from wherever it started.

It usually strikes between age 40 and 60, and it is fatal, because the muscles that fail eventually include the ones that let you breathe. Most people with ALS die from respiratory failure. There is no cure, and treatment adds months rather than years.

That is the context for the exercise question. If a common, recommended, otherwise healthy behavior were contributing to a disease like this, people would want to know.

Where the suspicion came from

Not from nothing. Clusters of cases among professional athletes have been documented for decades, most famously in Italian football and in the American baseball player whose name the disease still carries in the United States.

The trouble with those observations is that elite sport is not a random sample of exercise. It selects for particular bodies, particular injuries, particular chemistries and particular careers, and it is small. Extracting a general rule about walking to work from a cohort of professional footballers has never been sound, but in the absence of anything larger, the worry persisted.

What this study adds is scale, and a second way of measuring. Previously the evidence came from case series and occupational cohorts, none of them large enough to draw the shape of the relationship rather than argue about its direction.

What 384,836 people showed

The median follow-up was for 14.0 years, during which there were 541 incident diagnoses of ALS.

People in the highest activity group had about 23% less disease than the least active. The interval around that runs from roughly 4% to 39%, which is a real association and not a precise one.

Then the shape. The relationship was non-linear, with lowest risk in those in the mid-range self-reported activity. Being sedentary was the worst place to sit, but the very top of the range was not the best. The curve turns up again.

That detail matters more than the headline figure, because it is where the old suspicion could still live. A population cohort contains very few people training like professional athletes, so the far end of this curve is thinly populated and cannot settle what happens out there.

The wristbands agreed

Self-reported exercise is notoriously flattering. People round up, forget the sitting, and answer the question they wish had been asked.

So a subgroup wore wrist monitors. Higher overall device-measured activity was also associated with a lower risk of ALS, in 96,570 people among whom 98 cases occurred.

Two different measurement methods, with different biases, pointing the same way is worth more than either alone. The device data behaved slightly differently though: it showed a linear dose-response relationship, with no upturn at the top. Whether that reflects a real difference or simply too few very active people wearing monitors is not something the analysis can separate.

What genetics did and did not change

The team also asked whether the picture differs for people already at high inherited risk.

For carriers of a common risk variant in a gene called UNC13A, the association looked much the same. For carriers of the C9ORF72 repeat expansion, the single largest genetic cause of the disease, the result was directionally consistent but not conclusive, in a group of 535 people with 56 cases between them.

That is a small number to ask a question of, and the honest reading is that the study did not have enough carriers to answer it either way.

What a cohort like this cannot show

One objection sits above all the others, and it is not a technicality.

Motor neuron disease damages nerves for years before anyone notices a symptom. Someone who will be diagnosed at 62 may already be losing strength at 55, and someone losing strength moves less. Measure activity in that window and the arrow points backwards: the disease reduced the exercise, not the other way around.

Fourteen years of follow-up pushes that window back further than most studies manage, which is the strongest thing this design can do. It does not eliminate the problem, and the authors do not claim it does.

There is also the matter of who joins UK Biobank. Volunteers are healthier, better off and more active than the population they are drawn from, which tends to compress the range at the sedentary end, where much of this signal sits.

What it changes

Not what anyone should do on Monday. Regular exercise is one of the best things you can do for your health, and reducing your risk for many chronic diseases has never depended on what it does to motor neurons.

What shifts is the shape of a worry. A person who runs, lifts or cycles regularly has been carrying a low-grade question about whether they are doing themselves harm, seeded by headlines about athletes. On the largest population test yet run, higher activity tracked lower risk, and the sedentary end of the range looked worse rather than safer.

The authors are careful about how far that goes: higher self-reported and device-measured overall physical activity were associated with a lower risk of ALS overall, but with a potentially non-linear dose-response relationship. An association, in other words, and a curve with a bend in it that nobody has yet explained.

For the small number of people whose training genuinely resembles a professional athlete’s, the question stays open. For everyone else, it looks less like a risk than it did.