SUPPLEMENT: Secular trends in risk of stroke according to body mass index and blood pressure, 1976-2017

Supplement for for submitted article.

Objective
To test the hypothesis that the associations of body mass index (BMI) and BMI-related risk
factors with risk of stroke have attenuated over time using cohorts recruited from the general
population over 4 decades.

Methods
We undertook prospective studies of 2 cohorts enrolled in 1976 to 1978 (13,567 participants
from the Copenhagen City Heart Study) and 2003 to 2015 (107,040 participants from the
Copenhagen General Population Study). Each cohort was recruited randomly from the Danish
general population 20 to 100 years of age. Participants were followed up from the date of
examination to date of emigration, death, or stroke event, whichever occurred first. Follow-up
ended in March 2017. We did not lose track of any individual. BMI and blood pressure were
modeled with splines and in categories. Main outcome was incident stroke, including ischemic
stroke, intracerebral hemorrhage, and subarachnoid hemorrhage.

Results
The crude incidence of stroke declined in extreme categories of BMI and blood pressure from
1977 to 2017. The multivariable-adjusted hazard ratios for stroke in participants with BMI ≥30
vs 18.5 to 24.9 kg/m2 were 1.4 (95% confidence interval 1.2–1.6) in the 1976–1978 cohort and
1.1 (1.0–1.2) in the 2003–2015 cohort (p = 0.008 for 1976–1978 vs 2003–2015). The corresponding
hazard ratios (confidence intervals) in participants with blood pressure ≥160/100
vs <140/90 mm Hg were 2.1 (1.9–2.3) and 1.5 (1.4–1.7), respectively (p < 0.001). Similar
secular trends were observed for diabetes mellitus but were not obvious for other risk factors.

Conclusion
The associations of high BMI and high blood pressure with higher risk of stroke were attenuated
across 2 Danish cohorts enrolled from 1976 through 2015.