Abstract MP94: Association of Physical Activity and Sleep Profiles with Framingham 10 year Cardiovascular Disease Risk in Finnish Men and Women

Abstract

Introduction: Health behaviors including sleep, physical activity (PA) and sedentariness directly and indirectly affect CVD risk. Risk of CVD mortality and morbidity decreases with increasing leisure time PA. Sleep length deviant from population mean relates to increased CVD risk. Sleep quality and circadian preference are suggested to associate with CVD risk and metabolic risk factors. It is unclear how PA modifies health risks of poor sleep and vice versa. Higher PA may improve sleep and poor or short sleep may reduce PA. Our aim is to recognize people with the same profile of PA, sedentariness and sleep and to study if these subpopulations differ in their CVD risk. Material and methods: The National FINRISK 2012 Study comprised a stratified random sample of 10 000 Finns, aged 25 to 74 years, who participated in a health examination and filled in questionnaires. Participation rate was 65% (3041 men, 3383 women). We calculated the Framingham 10 year risk score for participants initially free from CVD and used latent class analysis (LCA) that identifies unobserved underlying subpopulations based on item response probabilities to determine how PA, sedentary and sleep behaviors interrelate. We fitted models with 1-5 classes and chose a 4 class model based on statistical criteria and interpretational judgment. Analyses were stratified by sex as measurement invariance was not fulfilled. Furthermore, we performed analysis of covariance, weighted by the posterior probabilities of the LCA, in order to compare the CVD risk in the 4 profiles. Results: We found 4 latent classes or PA and sleep profiles in men and women. The most important indicators were employment status, leisure time PA, sleep opinion and sleep duration. Likelihoods for employment, high leisure time PA and sufficient, longer sleep were combined in the most prevalent profiles and likelihoods for unemployment, physical inactivity and short, insufficient sleep occurred in the least prevalent profiles. When comparing the Framingham 10 year CVD risk score between the profiles, statistically significant differences, adjusted for age and education, were found. Lower CVD risk was associated with profiles that included likelihood for high leisure time or occupational PA and evening preference. In men, the profile with highest risk also included high screentime and long naps (risk score 20.7, p<.001). In women the profile with highest CVD risk also included high screentime (risk score 9.5, p<.001). Conclusions: PA, sedentary and sleep behavior patterns in the Finnish population are represented by 4 different profiles differ in the likelihood for leisure time PA, occupational PA, sleep length and satisfaction with sleep. The risk of CVD seems to be higher in the most sedentary and worst sleeping profile and in morning types, suggesting that lifestyle interventions should take into account not only one behavior but the clustering of health behaviors.