Notable paradoxical phenomena in associations between cardiovascular health score, subclinical and clinical cardiovascular disease in the community: The Framingham Heart Study

IMPORTANCE: Cardiovascular Health (CVH) scores are inversely associated with prevalent subclinical (SubDz) and incident cardiovascular disease (CVD). However, the majority of people who develop CVD have intermediate or ideal CVH scores, while many with poor CVH profiles escape CVD development. OBJECTIVE: To describe the prevalence of paradoxical relations among CVH, SubDz, and CVD. DESIGN: Cohort study, Framingham Study data collected prospectively (1995-2016). SETTING: Population-based. PARTICIPANTS: 7,627 participants (mean age 49 years, 53% women) attending Offspring examinations 6/7 and Third Generation examinations 1/2. EXPOSURES: CVH score (range 0-14) constructed from poor, intermediate, or ideal status for each metric (smoking, diet, physical activity, blood pressure, body mass index, fasting glucose, total cholesterol); and prevalent SubDz (≥1 of: increased carotid intimal media thickness, CIMT; left ventricular hypertrophy, LVH; microalbuminuria, MA; elevated ankle brachial index, ABI; coronary artery calcium score ≥100,CAC). MAIN OUTCOME(S) AND MEASURE(S): Ideal CVH (scores 12-14), intermediate CVH (scores 8-11), and poor CVH (0-7). We described three distinct paradoxical phenomena, involving combinations of CVH, SubDz, and CVD, and generated CVD incidence rates and predicted CVD probabilities for all combinations. RESULTS: We observed 842 CVD events (median follow-up 13.7 years); 1,663 participants had SubDz. Most individuals with poor CVH (78%) or SubDz (57% for CIMT to 77% for LVH) did not develop CVD on follow-up. Among participants with incident CVD, the majority had intermediate or ideal CVH (68%) or absent SubDz (46% for CAC to 96% for ABI) at baseline. We observed similar paradoxical results in relations between CVH and prevalent SubDz. Poor CVH and prevalent SubDz were each associated with higher CVD incidence rates compared to intermediate or ideal CVH and absent SubDz, respectively. The predicted CVD probability was nearly three-times greater among participants with poor (22%) versus intermediate or ideal CVH (8%). Mean CVD predicted probabilities were nearly three (26% vs. 10% for MA) to six-times (29% vs. 5% for CAC) greater among participants with SubDz versus without SubDz. Findings were consistent within age and sex strata. CONCLUSIONS AND RELEVANCE: Although poor CVH and SubDz presence are associated with CVD incidence, paradoxical phenomena involving CVH, SubDz, and CVD are frequently prevalent in the community. Further studies to elucidate biological mechanisms underlying these phenomena are warranted.