The evidence on exercise is extremely poor. In cultures that promote athleticism the sheer willingness to exercise correlates so strongly with healthiness that it's extremely hard to control for. And even then meta studies give you at best only a few extra years.
Exercise itself is a risky activity. Over a lifetime the risk of injury is non-trivial.
Studies that go against this pro-exercise bias are usually ignored[0].
> (seems like there are open questions here, like how much is actually important)
Yes, see this part of my post. It does seem that lots of types of exercise are not at all necessary for health.
This trial, for example, shows that HIIT for 3 intervals of 2-3 minutes had better impact than HIIT for 4 intervals of 4 minutes. So obviously there's still a lot to learn.
There's also research showing impact on blood glucose levels when you exercise before/ after eating, etc. "Exercise" is a big lump of a term, but there's a massive amount of research into different kinds - far more research than with almost any of the pills you can take.
Oh, there's absolutely a huge body of evidence showing exercise has "beneficial" effects on this or that measure. But it's hard to go from there and claim a huge life expectancy increase.
> There's also research showing impact on blood glucose levels when you exercise before/ after eating, etc. "
And we had studies showing that fructose has undesirable effects in that area. But so far most studies tracking food consumption in real life situations don't show anything bad happening.
Most of the world already believes exercise must be good for you, how many researchers are designing studies trying to uncover negative effects? Probably very few.
When most studies struggle badly to show anything above 5 extra years it's in the placebo category. Optimism might as well get you more [0].
Longevity studies in humans are pretty hard. You have to wait until the person dies. So instead signals are used, like "had non fatal heart attacks", if you're lucky. Most signals are going to instead be closer to "lowered or raised some metric that, from other research, we believe will lead to longevity".
I don't know how many papers there are on negative vs positive effects of exercise, but HIIT has had plenty of research done explicitly to determine safety, or to see if there are more effective exercises with lower rates of injury.
No question that there are very, very few studies that are directly "X leads to longer or better life", it's just "X leads to Y, we separately think Y leads to a longer or better life". Of course that's not as good... but it's a symptom of those studies being expensive and difficult.
They are hard, but we do have some. And even the most naive physical activity level vs all-cause-mortality studies show a mild effect at best. No controls.
Common sense would tell you that physical activity has to spuriously correlate with healthiness. Many conditions will simply render you unable (or less able) to exercise.
So what are the odds a properly designed longitudinal study would actually yield a bigger life expectancy increase? Seems very unlikely to me. The ones that we do have set an upper bound as far as I'm concerned. Almost certainly has to be lower.
I understand that you're talking about a specific kind of exercise and most of what's out there doesn't speak to that. But nonetheless I'd be very skeptical given how little effect has been shown for general physical activity. And we do have plenty of similar studies showing random benefits of various exercises in a clinical setting.
The evidence on exercise is extremely poor. In cultures that promote athleticism the sheer willingness to exercise correlates so strongly with healthiness that it's extremely hard to control for. And even then meta studies give you at best only a few extra years.
Exercise itself is a risky activity. Over a lifetime the risk of injury is non-trivial.
Studies that go against this pro-exercise bias are usually ignored[0].
[0] "This result suggests that daily strenuous exercise as an occupation shortens rather than prolongs the lifespan." https://www.nature.com/articles/s41599-020-0476-6