The timing of food intake, body weight, and chronic disease risk
Abstract
ABSTRACT OF THE DISSERTATIONThe Timing of Food Intake, Body Weight, and Chronic Disease Risk By Valeria Elahy Doctor of Philosophy in Epidemiology University of California, Irvine, 2022 Associate Professor Andrew Odegaard, Chair Several studies have demonstrated an association between timing of eating, circadian rhythms, metabolism, and chronic disease risk. However, a long-term relationship between the timing of eating and disease remains unclear due to the length of time required to study this association, among other reasons. This dissertation intends to explore further the relationship between the time, type 2 diabetes risk, breast cancer risk, and weight loss maintenance over time by using breakfast and after-dinner snacks as proxies of eating timing. This dissertation consisted of three separate research projects aiming to 1) examine the association between the consumption of breakfast and after-dinner snack patterns and breast cancer risk among post-menopausal; 2) estimate the causal effect of long-term breakfast consumption and night snacking on type 2 diabetes risk via causal inference modeling among young adults; 3) investigate if consuming breakfast and evening snacks have a differential effect on weight loss maintenance among individuals with obesity undergoing a standard weight loss intervention. The first study conducted a prospective analysis of 70501 post-menopausal women aged 49 to 81 years from the Women's Health Initiative Observational cohort study. Each participant's breakfast and after-dinner snack intake were assessed at the study Year 1 exam. Multivariable Cox proportional hazards regression models examined breakfast and after-dinner snack consumption frequencies and the risk of invasive and in situ breast cancer diagnosed between 1996 and 2020. In the second study, we emulated a target trial using observational data from 3737 subjects from the Coronary Artery Risk Development in Young Adults (CARDIA) Study. We applied the parametric g-formula to estimate the 20-y diabetes risk under several hypothetical intervention strategies (breakfast consumption 0-1, 2-4, and 5-7 times/week and post-dinner snack consumption 0-1, 2-4, and 5-7 times/week). Finally, in the third study, we emulated a target trial using observational data from 372 subjects in the Innovative Approaches to Diet, Exercise and Activity (IDEA) study. We followed the subjects for 18 months after the initial behavioral weight loss program and estimated body weight and composition change under several hypothetical interventions (breakfast consumption 0-4 times/week vs. 5-7 times/week and evening snack consumption 0-2 times/week vs. 3-7 times/week). The estimates were adjusted for pre-baseline and time-varying confounders and censoring using inverse probability weights with marginal structural models. The analyses showed no association between breakfast meals or after-dinner snack habits and the risk of breast cancer in post-menopausal women. In addition, the estimates from causal inference analysis supported that avoiding post-dinner snacks might be beneficial in reducing the long-term risk of diabetes; however, the role of starting regular breakfast consumption in midlife may have no major impact on the 20-y risk of diabetes. Finally, regular breakfast consumption and minimizing evening snacking may have a modest impact on lessening weight and body fat regain over 18 months after initial weight loss. In conclusion, the frequency of breakfast and after-dinner snacks is associated with metabolic disease risk and body weight maintenance.