REF-1564 · Joensuu et al., 2024 · Medicine & Science in Sports & Exercise · Paper · PubMed

Does being active beat your genes?

In 180 identical twin pairs differing in activity, death HR 0.74 (99% CI 0.37 to 1.48): lower risk not shown, yet the CI allows a large benefit. Self-report; low power, cohort.

Main caveat: Low power in the key test: 87 deaths in 360 twins; the authors name small sample size as the main critique.
moderate evidenceMeasured on non-athletes to trained athletes

Activity was asked in 1975, 1981 and 1990, and vital status followed to the end of 2020, with follow-up starting at DNA sampling. In the cohort, genetic risk predicted cardiovascular death, but meeting guidelines had no significant independent association after adjustment, and no interaction with genetic risk was detected.

The authors state they could not confirm that adherence mitigates genetic risk or is causally linked to mortality. They also say this does not dispute trial evidence on risk factors and that activity may add quality of life even if years are unconfirmed.

Protocol

Sample: 4897 complete cases from 8838 genotyped participants, 54.3% women, aged 33-60. For the twin analysis, 1348 monozygotic pairs; 412 lost to missing data and 756 not discordant left 180.
Exposure: mean MET-h/week across three surveys; meeting the guideline was at least 7.5. Strength work and sitting were not measured.
Analysis: Cox models with Bonferroni correction, so all CIs are 99%, and P < 0.01 is the bar. Full adjustment added BMI, smoking, diet and alcohol; the twin-pair models were also adjusted for BMI and smoking.
Funding: Academy of Finland and Finnish foundations; no competing interests.

Key Results

ComparisonDeathsHR (99% CI)
Whole cohort, meets guideline vs not, all-cause, before lifestyle adjustment11950.86 (0.72 to 1.02)
Same, fully adjusted11950.93 (0.78 to 1.11)
Whole cohort, cardiovascular, before lifestyle adjustment3890.80 (0.59 to 1.09)
Same, fully adjusted3890.92 (0.67 to 1.25)
Twin pairs, more vs less active, all-cause870.74 (0.37 to 1.48)
Twin pairs, cardiovascular360.39 (0.10 to 1.46)
Twin pairs, high genetic coronary risk, all-cause360.69 (0.23 to 2.11)
Twin pairs, low genetic coronary risk, all-cause291.00 (0.30 to 3.37)
Genetic risk: per 1 SD of the coronary score, cardiovascular death HR 1.17 (1.03 to 1.33); across the three scores, 17-24% higher in the abstract.
Vigorous activity, whole cohort: 0.75 (0.59 to 0.97) before lifestyle adjustment, 0.94 (0.72 to 1.23) after.
Reporting: Table 1 prints 3879 meeting guidelines, while the men and women listed add to 3876 (our arithmetic). A small slip with no effect on results.

Mechanisms

Measured: polygenic scores and mortality only.
Authors' interpretation: activity may partly proxy a healthier lifestyle and a genotype favouring both activity and cardiovascular health (pleiotropy). Healthy-exerciser bias and long follow-up could also weaken or inflate associations.
Inference: BMI may lie on the path from activity to death, so adjusting for it could remove part of a real effect; the fully adjusted estimates may err low.

Practical Application

Does being active beat your genes? In 180 identical twin pairs differing in activity, death HR 0.74 (99% CI 0.37 to 1.48): lower risk not shown, yet the CI allows a large benefit. Self-report; low power, cohort.
Good for: the honest reply that genetics and healthy habits explain part of the usual activity-longevity gap, and that the causal size is unsettled.
Not good for: "exercise does not extend life". A point estimate of 0.74 with this CI is uninformative, not null.
Not good for: athletes' training doses. Nothing here reaches them, and the outcome is death rather than fitness or function.

Limitations

Low power in the key test: 87 deaths in 360 twins; the authors name small sample size as the main critique.
Self-report: three leisure-time surveys over 15 years, none after 1990, with occupation not covered.
Survivor sample: follow-up begins at DNA sampling, and complete cases were healthier than the full genotyped sample (84.6% vs 72.1% healthy).
Discordance reasons unknown: illness could make one twin less active, and that is reverse causation inside the pair (inference).
Weak contrast: most participants met the guideline, so the division may not separate exposure enough, as the authors note.
How this entry was checked and changed
Why this level

large observational cohort; the twin design controls genes and early environment by design but is still observational, and its key comparison has low power