Gordon · The American journal of medicine 1977 · prospective cohort study · n=2815

High density lipoprotein as a protective factor against coronary heart disease. The Framingham Study.

Cited 5082 times in the scientific literature.

Level 3 - non-randomized controlled study

Prospective observational cohort follow-up study

PubMed 193398 · doi:10.1016/0002-9343(77)90874-9 · record verified 2026-08-30

What was done

Fasting lipid and lipoprotein levels (HDL, LDL, total cholesterol, and triglycerides) were measured between 1969 and 1971 in 2,815 men and women aged 49 to 82 years within the Framingham Study. A subset free of baseline coronary heart disease (1,025 men and 1,445 women) was followed prospectively for approximately four years to assess incident coronary heart disease.

What was found

Coronary heart disease developed in 79 men and 63 women over the follow-up period. HDL cholesterol was strongly and inversely associated with incident coronary heart disease in both men and women (p < 0.001), remaining significant after adjustment for other lipids and standard risk factors. LDL cholesterol showed a weaker association with coronary heart disease incidence (p < 0.05). Triglycerides were associated with coronary heart disease incidence only in women and only before adjustment for other lipids. Total cholesterol was not associated with coronary heart disease risk in this older cohort. Quantitative effect sizes (hazard ratios or relative risks) were not reported in the abstract.

Why it matters

This study provided early epidemiological evidence that HDL cholesterol is an independent protective marker against coronary heart disease in older adults, showing that total cholesterol alone was insufficient for risk stratification.

Limits

The follow-up period was relatively short (approximately four years), and the abstract provides p-values without effect sizes or confidence intervals. The cohort was restricted to older adults (ages 49 to 82) from Framingham, limiting generalizability to younger or more diverse populations, and the observational design cannot establish causality.

Cited by