30-year cohort study finds: multiple types of exercise associated with 19% lower mortality, even at same total activity level

The background: we know exercise helps, but does the type matter?

That regular physical activity goes along with a longer life is one of the best-documented findings in medicine. A large analysis published in 2023 pooled 94 groups of people with more than 30 million participants. It found that people who did the equivalent of 150 minutes of brisk walking a week had a roughly 31% lower risk of dying during follow-up than inactive people, and that even half that amount brought appreciable benefits^[2].

The World Health Organization therefore recommends that adults do 150–300 minutes of moderate activity or 75–150 minutes of vigorous activity per week, plus muscle-strengthening exercise on at least two days a week. For people aged 65 and over, it additionally recommends varied “multicomponent” activity that emphasises balance and strength training on at least three days a week^[3].

What has been much less clear is whether the kind of exercise matters, and whether doing several different activities brings anything beyond the total amount of exercise. There are reasons to suspect it might. In a small six-month trial in 160 older adults with obesity, endurance training mainly improved fitness, strength training mainly improved muscle strength, and combining the two improved physical performance the most^[4]. Long-term data on whether variety is linked to survival, however, have been missing. That is the gap the new study set out to fill^[1].

What the researchers did

The team at the Harvard T.H. Chan School of Public Health used two long-running US health studies: the Nurses' Health Study (women) and the Health Professionals Follow-Up Study (men). From 1986 onwards, participants filled in a questionnaire every two years about how much time they spent on activities such as walking, jogging, running, cycling, lap swimming, tennis or squash, rowing or callisthenics, weight training and climbing stairs. Most people answered it many times (median: 13 times per person)^[1].

The analysis included 111,467 people who, at the start, had no diabetes, heart disease, cancer, lung disease or neurological disease. To reduce the risk that people who were already ill simply exercised less, the researchers left a four-year gap between reported activity and the time at which deaths were counted. They also took into account other factors that differ between active and inactive people, including smoking, alcohol, diet quality, body weight, high blood pressure, high cholesterol and social contacts^[1].

To measure variety, they built a simple score: one point for each activity a person did consistently (at least 20 minutes a week, or at least five flights of stairs a day), averaged over the years. In the most varied fifth of participants, women typically did about three activities and men about four; in the least varied fifth, it was about one or less^[1].

Exercise amounts were measured in “MET-hours”, a unit that combines time and effort. As a rough guide, the authors note that 7.5 MET-hours of cycling a week corresponds to about an hour^[1].

What they found

Over more than three decades, 38,847 participants (about 34.9% of all participants) died: 9,901 from cardiovascular disease, 10,719 from cancer and 3,159 from lung disease^[1].

Almost every recorded activity was associated with reduced mortality. Comparing people who did the most of a given activity with those who did the least or none, the risk of death was 17% lower for walking, 15% lower for racquet sports, 14% lower for rowing or callisthenics, 13% lower for both running and weight training, 11% lower for jogging, 10% lower for stair climbing and 4% lower for cycling. These comparisons already took the person's other activities into account. Lap swimming was the exception: its risk ratio was 1.01, which means no measurable difference. The authors suspect this may be partly because people who report the same swimming time can swim very differently – leisurely or hard – which blurs the picture^[1].

More is better – up to a point. For total exercise, the lower risk of death from heart disease, cancer and lung disease largely levelled off at around 20 MET-hours a week. Individual activities showed similar plateaus. For walking and weight training, for example, the curve flattened at about 7.5 MET-hours a week^[1].

Variety made a difference of its own. People in the top fifth for variety had a 19% lower risk of death during follow-up than those in the bottom fifth. This figure already accounts for how much they exercised overall. Because of the large number of participants, a comparatively precise estimate can be made: the true reduction is likely between 15% and 22%. The link held for each major cause of death, with a 13% lower risk for cancer, 17% for cardiovascular disease and 41% for lung disease. Within every level of total activity (low, medium or high) those with more variety did better, and people who were in the top group for both amount and variety had a 21% lower risk than those in the bottom group for both^[1].

In raw numbers, the least varied group had about 23 deaths per 1,000 people per year of follow-up, compared with about 11 in the most varied group. That gap is much larger than the 19% figure because the groups also differed in other ways – above all, people who did more kinds of exercise also tended to exercise more overall. The adjusted figure strips out these differences as far as possible^[1].

What this means

The study adds a long-term, large-scale piece of evidence to an idea that is intuitive due to physiological considerations: different activities affect the body in different ways (for example fitness of the heart and lungs, muscle strength, body composition or bone strength^[1]), so it seems plausible that a mix covers more ground than any single activity. The finding that benefits from each individual activity levelled off at moderate amounts fits this picture. The authors suggest that people who spread their exercise across several activities may get the most out of each one^[1].

The results also fit well with existing advice. The WHO already recommends combining aerobic and muscle-strengthening activity, and varied multicomponent activity for older adults^[3]. The new data do not change these recommendations; they offer some support for them from a different angle.

For everyday life, the most robust message is still the oldest one: being active is linked to a longer life, and even modest amounts appear to help^[2]. Simple walking showed one of the strongest associations in this study^[1]. The new finding is that, if you already exercise, adding a different kind of activity may be worth considering.

What the study does not show is a “best” combination of activities. The authors deliberately did not define fixed combinations, because people change what they do over time. They call for future studies on whether certain activities work especially well together^[1].

Limitations

• Association does not imply causality. People who perform many kinds of exercise may differ from others in ways the researchers could not fully measure, for example in fitness, health or free time. The analysis adjusted for many factors, but some confounding factors may remain^[1].
• Reverse causation cannot be definitively excluded. People who are becoming ill may reduce their activities even before a disease is diagnosed. The four-year gap reduces this problem but cannot rule it out, especially for lung disease, which can cause symptoms long before it is diagnosed^[1].
• Self-reported exercise was used. Participants estimated their own activity, and the questionnaire did not capture how hard they exercised. Earlier checks showed that the questionnaire tracks real activity only moderately well^[1].
• A specific group of people was examined. Participants were mostly white health professionals in the USA, and women and men came from separate studies. The results may not apply equally to other populations^[1].
• Not every activity was included in the questionnaire from the beginning. Weight training, for example, was added to the questionnaires later, which limits the data for some activities^[1].