https://bjsm.bmj.com/content/bjsports/early/2026/05/28/bjsports-2025-110503.full.pdf
chatGPT:
Summary
This paper examines whether long-term resistance training is associated with lower mortality, and how its effects interact with aerobic exercise. It uses three large US prospective cohorts: Health Professionals Follow-up Study, Nurses’ Health Study, and Nurses’ Health Study II, giving 147,374 participants, 35,798 deaths, and up to 30 years of follow-up. Resistance training and aerobic activity were repeatedly assessed by questionnaire, rather than only once at baseline.
The main finding is that moderate resistance training is associated with lower all-cause mortality, even after adjustment for aerobic activity. The apparent optimum for all-cause mortality was around 90–119 minutes/week, associated with 13% lower all-cause mortality compared with no resistance training. Higher levels, ≥120 min/week, did not add further benefit.
Cause-specific results differed. The 90–119 min/week group had 19% lower cardiovascular mortality and 27% lower neurological disease mortality. Cancer mortality was different: lower risk appeared only at 1–59 min/week, with no clear benefit at higher resistance-training volumes.
The joint analysis is important. Aerobic activity had the stronger association with lower mortality. Resistance training alone gave modest benefit, while the lowest mortality was seen in those doing both substantial aerobic activity and resistance training. For example, those doing 30 to <45 MET-hours/week aerobic activity plus 60–119 min/week resistance training had an HR of 0.55 versus people with inadequate aerobic activity and no resistance training. But at very high aerobic activity, ≥45 MET-hours/week, resistance training added little extra apparent mortality benefit.
What is novel?
The novelty is not simply “exercise lowers mortality.” The useful additions are:
First, it looks at resistance training duration, not merely frequency such as “twice per week.” The authors note that prior work often used dichotomised frequency measures and single baseline exposure, limiting dose-response inference.
Second, it uses repeated measures over time, creating cumulative averages of resistance training and aerobic activity. This is stronger than a one-off baseline questionnaire because it better approximates long-term behaviour and reduces measurement error.
Third, it explicitly separates all-cause, cardiovascular, cancer, respiratory, neurological and other mortality, showing that the dose-response curves may not be the same for different outcomes. The neurological mortality result is particularly interesting because this outcome has been less well studied for resistance training.
Fourth, it provides a detailed joint resistance-training/aerobic-activity matrix, rather than asking whether resistance training is “independent” in a simplistic way. This shows that aerobic activity appears dominant for mortality reduction, while resistance training adds benefit mainly below very high aerobic activity.
Critique
This is a strong observational study, but it should not be read as proving that exactly 90–119 minutes/week of resistance training is biologically optimal. People who do resistance training differ systematically from those who do not. In Table 1, higher resistance-training groups were leaner, smoked less, had healthier diets and did more aerobic activity. The authors adjust for these factors, but residual “healthy user” confounding is still likely.
The exposure measure is also crude. “Resistance training minutes/week” does not distinguish load, intensity, proximity to failure, muscle groups, progression, frailty-adapted training, supervised versus unsupervised training, or whether someone is doing heavy compound lifting versus light machine circuits. A minute of resistance training is not physiologically equivalent across people.
The apparent lack of added benefit above 120 min/week may be real, but it could also reflect confounding or exposure misclassification. People doing very high resistance-training volumes may include mixed groups: very healthy exercisers, people training for sport, people compensating for illness or weight concerns, or people with higher injury burden. The study cannot fully separate these.
The cancer finding is the most puzzling. The lower cancer mortality only at 1–59 min/week, not at higher levels, could suggest a genuine J-shaped relationship, but it may also be chance, residual confounding, cancer-site heterogeneity, or survivorship effects. The authors mention possible mechanisms such as IGF-1, but the paper does not directly measure mediators, so mechanistic interpretation should be cautious.
Generalisation is limited. The cohorts are mostly health professionals, predominantly white, and generally health-literate. The authors themselves acknowledge that the cohorts were established before diversity was a major recruitment consideration.
Bottom line
The paper supports the practical view that doing some resistance training is associated with lower mortality, with a plausible useful range around 1–2 hours/week, and that combining it with aerobic activity is better than doing neither. However, the strongest mortality signal remains aerobic activity, and the paper should not be interpreted as showing that more than 120 minutes/week of resistance training is harmful or useless. It shows a plateau in observational mortality associations, not a definitive biological ceiling.