Meal Frequency and Timing Are Associated with Changes in Body Mass Index in Adventist Health Study 2

Abstract

Background: Scientific evidence for the optimal number, timing, and size of meals is lacking.Objective: We investigated the relation between meal frequency and timing and changes in body mass index (BMI) in the Adventist Health Study 2 (AHS-2), a relatively healthy North American cohort.Methods: The analysis used data from 50,660 adult members aged ≥30 y of Seventh-day Adventist churches in the United States and Canada (mean ± SD follow-up: 7.42 ± 1.23 y). The number of meals per day, length of overnight fast, consumption of breakfast, and timing of the largest meal were exposure variables. The primary outcome was change in BMI per year. Linear regression analyses (stratified on baseline BMI) were adjusted for important demographic and lifestyle factors.Results: Subjects who ate 1 or 2 meals/d had a reduction in BMI per year (in kg · m^-2 · y^-1) (-0.035; 95% CI: -0.065, -0.004 and -0.029; 95% CI: -0.041, -0.017, respectively) compared with those who ate 3 meals/d. On the other hand, eating >3 meals/d (snacking) was associated with a relative increase in BMI (P < 0.001). Correspondingly, the BMI of subjects who had a long overnight fast (≥18 h) decreased compared with those who had a medium overnight fast (12-17 h) (P < 0.001). Breakfast eaters (-0.029; 95% CI: -0.047, -0.012; P < 0.001) experienced a decreased BMI compared with breakfast skippers. Relative to subjects who ate their largest meal at dinner, those who consumed breakfast as the largest meal experienced a significant decrease in BMI (-0.038; 95% CI: -0.048, -0.028), and those who consumed a big lunch experienced a smaller but still significant decrease in BMI than did those who ate their largest meal at dinner.Conclusions: Our results suggest that in relatively healthy adults, eating less frequently, no snacking, consuming breakfast, and eating the largest meal in the morning may be effective methods for preventing long-term weight gain. Eating breakfast and lunch 5-6 h apart and making the overnight fast last 18-19 h may be a useful practical strategy.

Keywords: Adventist Health Study 2; BMI; meal frequency; meal timing; weight control.

FIGURE 1

Changes in BMI per year in the AHS-2 population by quartile of baseline BMI (interrupted lines) and over all subjects (solid line). The regression model was change in BMI per year = α + Σ(β_j ⋅ age_j) + Σ(β_k ⋅ covariate_k), where j represents the J-1 age range indicator variables, and k the K covariates. Data are the predicted values of change in BMI per year at a particular age, conditional on covariates at mean values (continuous variables) or reference values (categorical variables), with 95% confidence bands. AHS-2, Adventist Health Study 2.

FIGURE 2

The relation between meal frequency and timing and the change in BMI per year relative to the reference exposure value in the AHS-2 population. (A) Number of meals and snacks consumed per day (reference: 3 meals/d), (B) length of overnight fast (reference: 12–17 h), (C) consumption of breakfast (reference: no), and (D) timing of the largest meal (reference: dinner). P values are given for trends or differences (as appropriate). Data are predicted values ± 95% CIs conditional on covariates at mean or reference values. AHS-2, Adventist Health Study 2.

FIGURE 3

Modification by race of the association between the relative change in BMI according to whether lunch was the largest meal. The P value is for interaction by race. Data are predicted values of change in BMI per year conditional on the chosen values of race and lunch as the largest meal, with other covariates at mean or reference values.