Introduction: Elevated blood cholesterol is a modifiable risk factor for cardiovascular disease (CVD). Surveillance of cholesterol levels is needed to study disease burden, consider priorities for CVD prevention and intervention, and understand the effect of diet, lifestyle and medication use. Previous studies show a cholesterol level decline in recent decades, but lack data to follow individuals from different birth cohorts from early and middle adulthood to older age. Hypothesis: To assess secular and longitudinal trends in total cholesterol levels. Methods: We investigated changes in age-specific cholesterol levels by repeated measurements in 37,968 women and men born 1905-1977 examined up to five times between 1979 and 2008 in the population-based Tromsø Study. Results: Mean total cholesterol decreased from 1979 to 2008 in both genders in all age groups. The decrease in age group 40-49 years was 1.2 mmol/l in women and 1.0 mmol/l in men. Both the 80th percentile and the 20th percentile of the population cholesterol distribution decreased. In birth cohorts born 1940-77, mean total cholesterol levels increased with age up to 2001 and decreased thereafter to 2007. Among those born 1910-39, the cholesterol decrease was evident after 1994-95. The longitudinal trends in those born 1940-77 were not influenced by use of lipid-lowering drugs, while the decrease with age in birth cohorts born 1910-39 was slightly weaker among never users of lipid-lowering drugs compared to the total sample. Conclusions: A longitudinal decrease in mean cholesterol levels was observed from 1994-95 or from 2001 depending on birth cohort. A secular decrease was observed from 1979-80 in all age groups. The findings points to a mass population effect rather than a treatment effect of high-risk individuals alone.