Comparison between recent and long-term physical activity levels as predictors of cardiometabolic risk: a cohort study.

Objective: To compare recent and long-term physical activity levels as predictors of cardiometabolic risk in a risk factor-free adult population.

Design: A 12-year prospective cohort study.

Setting: The Finnish Public Sector study with surveys conducted in four waves at 4-year intervals.

Participants: 19 230 participants (mean age 50.2 (SD 9.1) years, 84% women) with no prevalent cardiometabolic risk factors at wave 3 were included. Physical activity was assessed at waves 1, 2 and 3. The long-term physical activity level was determined as the mean of activity from wave 1 to 3 (over 8 years).

Outcome Measure: 4-year incidence of cardiometabolic risk factors (obesity, hypertension, dyslipidaemia and diabetes) after wave 3, measured individually and as a sum (accumulation of two or more risk factors vs none). Logistic and multinomial logistic regression analyses were used for the analysis.

Results: Graded associations between higher physical activity levels and lower odds of all risk factors were observed (p for trend <0.01). In comparison with the persistently vigorously active participants (≥60 metabolic equivalent (MET)-hours/week), those who were persistently inactive (<7 MET-hours/week) were about four times more likely to develop obesity (OR=4.24, 95% CI=2.83 to 6.36), two times more likely to develop hypertension (OR=2.02, 95% CI=1.45 to 2.82) and dyslipidaemia (OR=1.82, 95% CI=1.03 to 3.22) and eight times more likely to develop diabetes (OR=7.84, 95% CI=1.78 to 34.6). The corresponding OR for accumulating two or more risk factors was 5.24-fold (95% CI=2.39 to 11.47). For recently inactive versus recently vigorously active, the estimates were weaker (OR=2.36, 95% CI=1.71 to 3.25 for obesity; 1.78, 95% CI=1.35 to 2.35 for hypertension; 1.71, 95% CI=1.04 to 2.82 for dyslipidaemia; 3.56, 95% CI=1.06 to 11.96 for diabetes and 2.66, 95% CI=1.48 to 4.78 for ≥2 risk factors).

Conclusion: Cardiometabolic risk associated with physical inactivity is better captured by repeated measurements of physical activity than by a single measurement of the most recent activity level.