Cholesterol, Triglyceride, and Glucose Levels Across Birth Cohorts in the US

Abstract

Importance: Population mean levels of total cholesterol and fasting triglycerides have decreased substantially in the US in recent decades, but improvements in cardiometabolic health may have slowed among younger cohorts. Conversely, prevalence of diabetes (types 1 and 2) and obesity has increased, especially among adults younger than 65 years. It is unclear how trends in cholesterol, triglyceride, and glucose levels have changed across different birth cohorts and whether adverse trends in obesity are associated with these patterns.

Objective: To quantify national trends in total cholesterol, fasting triglyceride, and fasting glucose levels among cohorts born between 1920 and 1999 and examine the potential association of these patterns with body mass index (BMI).

Design, setting, and participants: This serial cross-sectional study used National Health and Nutrition Examination Survey (NHANES) data from the 1999-2000 to 2017-2020 cycles. Data were analyzed between November 1, 2023, and July 31, 2024. Participants included nonpregnant and noninstitutionalized US adults 18 years or older, born between 1920 and 1999, who had data collected from 1999 to 2020.

Exposure: Eight 10-year birth cohorts (from 1920 to 1999).

Main outcomes and measures: Total cholesterol, fasting triglyceride, and fasting glucose levels and BMI. Quantile regression models reported average marginal effects to quantify mean change in cardiometabolic outcome measures per decade of birth years. Parametric regression models estimated the association of birth cohort with outcomes, assessing BMI as the mediator.

Results: Of 52 006 participants weighted to represent 264 664 915 US adults, weighted median age was 46 (IQR, 33-60) years and 50.6% were women. For the 50th percentile of measures, mean difference per 1-decade younger birth cohort was -7.1 (95% CI, -8.2 to -6.1) mg/dL for total cholesterol level, -13.1 (95% CI, -15.1 to -11.1) mg/dL for fasting triglyceride level, and 2.7 (95% CI, 2.3-3.1) mg/dL for fasting glucose level. BMI appeared to attenuate the associations between birth cohort and lipid levels and enhanced the associations between birth cohort and fasting glucose levels. However, up to 80% of the associations between birth cohorts and cardiometabolic outcomes were not mediated through BMI.

Conclusions and relevance: In this cross-sectional study of 52 006 participants representing 264 664 915 US adults, population-level improvements in total cholesterol and triglyceride levels decelerated and adverse trends in glucose levels accelerated in more recent birth cohorts, which was partially mediated by concurrent increases in BMI. Public health initiatives that target antecedent health behaviors are needed to improve cardiometabolic health across generations.

Figure 1.. Sample Inclusion and Exclusion Flowchart

BMI indicates body mass index; NHANES, National Health and Nutrition Examination Survey. ^aParticipants could be excluded for more than 1 reason.

Figure 2.. Adjusted Trends of Total Cholesterol, Fasting Triglyceride, and Fasting Glucose Levels and Body Mass Index (BMI) by Birth Cohort

Data are from National Health and Nutrition Examination Survey (NHANES), 1999 to 2020. The curves are based on quantile regression models adjusted for sex, race (with non-Hispanic White race as the reference group), age, square of the age, square of the cohort, and age × cohort interaction. The models for total cholesterol level and BMI used NHANES mobile examination center subsample weights. The models for triglyceride and glucose levels used fasting subsample weights.