Does running a marathon make you healthier?

The one-sentence answer

Why this came up here

I asked this question before signing up, not after — it's the reason the site exists. I've been running for over a decade: four half marathons since 2018, the later ones finishing somewhere between 2:20 and 2:40. The first three ended the same way — blisters, a climbing heart rate, and a trot-walk-drag negotiation with the last four kilometres. The 2025 race was different. After a year of strength training and generally smarter preparation, I sped up in the final kilometre and crossed the line feeling like I had something left. For the first time, the people who turn around and run the course again to make it 42.2 km didn't seem completely unhinged.

So I signed up for the full distance — and then went looking for evidence about whether this is actually good for me. The longer version of how I got here is in why I'm doing this; a version of this piece first appeared on my personal blog.

The intuitive model

Think of exercise as a dose-response curve. Going from sedentary to running a few times a week buys a large share of the measurable benefit; each additional hour buys less. Marathon training sits far out on the flat part of that curve — still fitness-positive, but in small increments, while at least one risk (injury) scales with volume rather than flattening. The honest question isn't whether running is healthy; it's whether the far end of the curve is still worth the trip.

What the evidence says

The benefit is real, and mostly cheap. Cardiorespiratory fitness is one of the strongest measurable predictors of how long people live. A 2024 overview of meta-analyses, pooling 199 cohort studies, found that people with high fitness had roughly half the mortality risk of those with low fitness. The association is observational — more on that below — but it is unusually large and consistent. The catch for this project: most of that fitness is attainable at training volumes well below a marathon build, so the marginal benefit of the last hours is small.

Hearts kept under sustained load push back a little. A 2021 systematic review and meta-analysis found athletes had about 2.5 times the odds of atrial fibrillation compared with non-athlete controls, with the excess larger in athletes under 55. Two details temper it: mixed-sport athletes (soccer, rugby) carried more risk than endurance athletes, and these are career athletes, not people training up for one race — nothing here says a recreational runner inherits their odds. But a months-long block of sustained load is a push in that direction, and it would be tidier than honest to pretend the risk is zero.

Injury is the risk that actually scales. Toresdahl et al. (2021) followed first-time New York City marathoners through training: roughly half picked up an injury during the build. Most were minor, but about 1 in 10 runners had one serious enough to keep them from starting or finishing. This is the number my plan spends the most effort on, and the one the injuries & interruptions page exists to report honestly.

The scare stories are the smallest risk. Cardiac arrest during a race happens, and it is exceedingly rare — about 1 in 100,000 participants in a large NEJM registry of long-distance races. The risk is elevated while racing compared with sitting on a couch that hour, and still tiny in absolute terms.

One more cost, easy to miss: it is genuinely hard to hold on to muscle mass through a high-volume running block, and muscle is itself a longevity asset. That one is a measured outcome in this experiment, not a footnote — body composition is tracked on the dashboard.

What remains uncertain or contested

The fitness-mortality link is observational: fit people differ from unfit people in many ways studies can't fully adjust for, so "half the mortality risk" is an association, not a guaranteed payoff for getting fitter. The atrial fibrillation data comes from career athletes, and the injury numbers from a single cohort of first-timers in one city — I differ from both groups, in opposite directions: less lifetime load than a career athlete, more running base than a typical first-timer. And no study measures the thing I actually care about — the net health effect of one well-planned marathon build on one specific runner. That gap is what the experiment is for.

What's being observed in this project

When I put the question to Dr. Varun Reddy, the sports-medicine collaborator on this project, his view was the sentence the whole plan is built on: planned properly, most of the downside risk can be substantially reduced — leaving a small net upside (a little weight lost, slightly better cardiovascular markers) and a temporary downside that can be reclaimed. The plan is the "planned properly" part; his review of it is on record. Whether the small net upside actually shows up is what the dashboard tracks: VO₂ max (what that number means), resting heart rate, body composition, and everything that goes wrong along the way. This page will link forward to the results, whichever way they point.