06/17/2026
Quality content!
Strength training lowers death risk. The benefit caps at about two hours a week.
A new analysis pooled three Harvard cohorts. 147,374 adults. Up to 30 years of follow-up. 35,798 deaths. Resistance training time was self-reported repeatedly across the follow-up window. The analysis is observational, which means it cannot prove causation. It can map the dose-response curve, and the shape of that curve is the entire point.
The shape is consistent across every major cause of death. Risk drops fast as weekly training rises from zero. By 60 minutes a week, most of the benefit has already happened. By 120 minutes a week, the curve goes flat. Past two hours, more lifting did not produce further reductions in any outcome the study measured. At the plateau, compared to no resistance training, all-cause death was 13% lower. Cardiovascular death was 19% lower. Neurological death was 27% lower. Cancer was the only outcome that did not follow the same curve. The cancer benefit appeared only at low volumes (1 to 59 minutes per week) at 9 to 12% lower risk, and by the time training crossed two hours, that signal was gone.
This is where the broader literature matters, because the Zhang paper is not an outlier. It is the largest cohort to date to confirm a curve that pooled data was already showing four years ago. Shailendra et al. (2022, Am J Prev Med) ran a meta-analysis of 10 prior studies looking at this exact question. They reported a 15% reduction in all-cause mortality from any resistance training, with a peak benefit of 27% reduction at around 60 minutes per week, and explicitly noted that reductions diminished at higher volumes. Same curve. Same plateau. Same peak benefit window. The Zhang 2026 analysis confirms the shape with a sample size more than ten times larger and a follow-up window long enough to capture neurological mortality, which prior cohorts were underpowered to detect.
This consistency matters because the field has had a quiet running disagreement about whether resistance training's mortality benefit is real or whether it is mostly a marker of people who are generally more health-conscious. The shape of the curve is the strongest argument against the marker hypothesis. If the relationship were purely confounded by lifestyle, you would not expect a clean, replicable dose-response with a plateau at the same time window across independent cohorts. You would expect the benefit to track total exercise volume and not flatten. It does flatten. And it flattens in the same place every time.
Adults who layered meaningful aerobic exercise on top of their resistance training saw the largest reductions in this analysis, up to 47% lower all-cause death at the high end of combined volume. This is also consistent with the broader literature. Every major cohort that has looked at combined activity has found that aerobic plus resistance outperforms either alone, often by a large margin. The lifting plateau is a real ceiling on what more strength training can buy you. It is not a ceiling on what more total exercise can buy you.
The mechanistic reading is that resistance training and aerobic training drive partly overlapping and partly distinct biology. Resistance training is doing most of its work on muscle mass, insulin signaling, glycemic control, and skeletal strength. Those adaptations saturate. The marginal lifter going from zero to two hours a week gains them. The marginal lifter going from two to four hours a week has already captured them. Aerobic training drives cardiac output, mitochondrial density, endothelial function, and VO2 max. Those pathways do not saturate at the same training volume. The combined effect on mortality is larger than either alone because the two stimuli are not doing the same job.
The minimum effective dose for the survival benefit of lifting is closer to 60 minutes a week than to the 150 minutes the guidelines push. The optimal dose is about two hours. Past two hours a week, hours four through ten are doing something for your physique, your strength, your enjoyment, your performance. They are not doing anything additional for your death risk in this dataset. If your goal is longevity rather than performance, two hours a week of lifting plus 150 to 300 minutes a week of moderate aerobic activity is the combination the data supports.
This is observational data. Self-reported exercise time is noisy. The cohorts skew toward US health professionals, which means the population is healthier and more educated than the general population. The study measured time, not intensity, exercise selection, or program design. The 47% figure for the combined group is the upper bound and applies to high aerobic volumes. The 13/19/27% reductions for lifting alone at two hours a week are the more defensible numbers.
Lifting cuts death risk. The benefit plateaus at around two hours a week. The way to move the number further is to stack cardio on top, not more sets.
Zhang et al., Br J Sports Med, 2026
Shailendra et al., Am J Prev Med, 2022
