REF-1570 · Lee et al., 2022 · Circulation · Paper · PubMed

When is the death-risk curve nearly at its lowest?

Near 150-300 min/wk vigorous (HR 0.79-0.77) or 300-600 moderate (0.74-0.69, still falling); no clear gain or harm above. Repeated self-report, health professionals: cohort, not proof.

Main caveat: Self-report, leisure only; random error remains despite validated questionnaires.
moderate evidenceMeasured on non-athletes to trained athletes

Participants without cardiovascular disease or cancer at baseline entered in 1988, with activity updated every 2 years and a median follow-up of 26 years. Against the guideline minimum, 2-4 times the vigorous dose went with 2-4% lower mortality, and 2-4 times the moderate dose with 3-13% lower.

Above about 300 min vigorous or 600 min moderate, there was no clear further association, and no sign of harm for death. In joint analyses, vigorous work went with lower mortality when moderate activity was under 300 min/week, but not clearly above that. The authors' statement that high vigorous activity showed no harm concerns these mortality estimates.

Protocol

Exposure: biennial report of activity types, cumulative average; categories from 0 to 600+ min/week. References: 0 min vigorous and 0-19 min moderate.
Adjustment: race, family history, hormone use, alcohol, energy intake, smoking, sleep, diet quality, BMI, and mutual adjustment of moderate and vigorous activity (r=0.2).
Bias steps: baseline disease excluded, 2-year lag, repeated measures.
Funding: NIH; no conflicts declared.

Key Results

Minutes/weekVigorous, all-cause HR (95% CI)Vigorous, CVD HRModerate, all-cause HR (95% CI)Moderate, CVD HR
1-74 / 20-740.87 (0.82-0.93)0.76 (0.67-0.86)0.91 (0.88-0.94)0.87 (0.81-0.94)
75-1490.81 (0.76-0.87)0.69 (0.60-0.78)0.84 (0.81-0.88)0.79 (0.73-0.85)
150-2240.79 (0.74-0.85)0.73 (0.64-0.85)0.80 (0.77-0.83)0.78 (0.72-0.84)
225-2990.77 (0.72-0.84)0.67 (0.57-0.79)0.79 (0.76-0.82)0.75 (0.68-0.82)
300-3740.78 (0.72-0.85)0.68 (0.56-0.82)0.74 (0.70-0.77)0.72 (0.65-0.79)
450-5990.76 (0.68-0.85)0.67 (0.53-0.84)0.69 (0.65-0.73)0.62 (0.55-0.69)
≥6000.74 (0.65-0.85)0.71 (0.54-0.93)0.68 (0.64-0.73)0.63 (0.56-0.71)
Row not shown, 375-449 min: vigorous all-cause 0.73 (0.65-0.81), CVD 0.65 (0.52-0.82), its lowest estimates; moderate 0.74 (0.70-0.78), CVD 0.71 (0.63-0.79). Confounding visible: vigorous 75-149 min was HR 0.50 (0.47-0.53) adjusted for age alone, against 0.81 after covariates.
Thin ends: 0 min vigorous is 91,573 of 2,984,545 person-years, about 3.1%, and only 3.1% men. At ≥600 min there are 26,638 person-years, about 0.9% (our arithmetic), and 59.8% men. Age and sex are handled in the models.
Joint analysis: combinations of 75-300 min vigorous and 150-600 min moderate went with about 35-42% lower mortality; for the insufficiently active adding either dose, the authors estimate about 22-31% lower mortality (their Discussion).
Single baseline measure: weaker associations and a tendency to a U-shape.
Incident cardiovascular disease: 75-149 min vigorous HR 0.78 (0.70-0.87), coronary 0.75 (0.64-0.87), stroke 0.86 (0.73-1.01).
Joint results: the 35-42% and 22-31% figures are in the full text only; the abstract gives the split at 300 min/week of moderate activity.

Mechanisms

Measured: none.
Authors' interpretation: repeated measures reduce measurement error and reverse causation, which may explain why others found U-shapes.
Inference: cumulative averages smooth over change, and participants who decline when ill still shift their average, so reverse causation is smaller rather than gone.

Practical Application

When is the death-risk curve nearly at its lowest? Near 150-300 min/wk vigorous (HR 0.79-0.77) or 300-600 moderate (0.74-0.69, still falling); no clear gain or harm above. Repeated self-report, health professionals: cohort, not proof.
Good for: someone training for health with no race. Either intensity went with lower mortality, and vigorous work went with the largest added difference when moderate activity was low.
Good for: answering whether high volume looked harmful for death in a long cohort: not shown, up to about 13.7 h/week vigorous on average in the top group.
Not good for: deciding an athlete's training hours, which rest on performance; see SYNTH-26.
Not good for: non-fatal rhythm or coronary questions; see REF-1577 and REF-1578.

Limitations

Self-report, leisure only; random error remains despite validated questionnaires.
Selected sample: white health professionals in the main, with possible healthy-user bias and limited generalisability.
Residual confounding, which the authors acknowledge.
Few at the extremes: reference and top categories are small and differ in sex mix.
Death as the outcome says nothing on injury, rhythm disorder or coronary plaque.
How this entry was checked and changed
Why this level

large observational cohort; repeated measures make it among the better designs of its type, but it stays at the observational cap