Polygenic risk score for obesity and the quality, quantity, and timing of workplace food purchases: A secondary analysis from the ChooseWell 365 randomized trial

Autor: Anne N. Thorndike, Douglas E. Levy, Richa Saxena, Jessica L. McCurley, Hassan S. Dashti, Marie-France Hivert
Rok vydání: 2020
Předmět:
Male
Central Nervous System
Physiology
Economics
Social Sciences
030204 cardiovascular system & hematology
Logistic regression
Weight Gain
Nervous System
Body Mass Index
0302 clinical medicine
Cognition
Risk Factors
Food choice
Medicine and Health Sciences
Medicine
Psychology
030212 general & internal medicine
Workplace
biology
General Medicine
Middle Aged
Quartile
Physiological Parameters
Female
Anatomy
Research Article
Adult
Employment
Decision Making
Cafeteria
Health Promotion
03 medical and health sciences
Food Preferences
Statistical significance
Food Quality
Genetics
Humans
Genetic Predisposition to Disease
Obesity
Nutrition
business.industry
Body Weight
Cognitive Psychology
Biology and Life Sciences
Odds ratio
Consumer Behavior
biology.organism_classification
medicine.disease
Diet
Food
Genetic Loci
Labor Economics
Cognitive Science
business
Body mass index
Demography
Boston
Neuroscience
Zdroj: PLoS Medicine
PLoS Medicine, Vol 17, Iss 7, p e1003219 (2020)
ISSN: 1549-1676
Popis: Background The influence of genetic risk for obesity on food choice behaviors is unknown and may be in the causal pathway between genetic risk and weight gain. The aim of this study was to examine associations between genetic risk for obesity and food choice behaviors using objectively assessed workplace food purchases. Methods and findings This study is a secondary analysis of baseline data collected prior to the start of the “ChooseWell 365” health-promotion intervention randomized control trial. Participants were employees of a large hospital in Boston, MA, who enrolled in the study between September 2016 and February 2018. Cafeteria sales data, collected retrospectively for 3 months prior to enrollment, were used to track the quantity (number of items per 3 months) and timing (median time of day) of purchases, and participant surveys provided self-reported behaviors, including skipping meals and preparing meals at home. A previously validated Healthy Purchasing Score was calculated using the cafeteria traffic-light labeling system (i.e., green = healthy, yellow = less healthy, red = unhealthy) to estimate the healthfulness (quality) of employees’ purchases (range, 0%–100% healthy). DNA was extracted and genotyped from blood samples. A body mass index (BMI) genome-wide polygenic score (BMIGPS) was generated by summing BMI-increasing risk alleles across the genome. Additionally, 3 polygenic risk scores (PRSs) were generated with 97 BMI variants previously identified at the genome-wide significance level (P < 5 × 10−8): (1) BMI97 (97 loci), (2) BMICNS (54 loci near genes related to central nervous system [CNS]), and (3) BMInon-CNS (43 loci not related to CNS). Multivariable linear and logistic regression tested associations of genetic risk score quartiles with workplace purchases, adjusted for age, sex, seasonality, and population structure. Associations were considered significant at P < 0.05. In 397 participants, mean age was 44.9 years, and 80.9% were female. Higher genetic risk scores were associated with higher BMI. The highest quartile of BMIGPS was associated with lower Healthy Purchasing Score (−4.8 percentage points [95% CI −8.6 to −1.0]; P = 0.02), higher quantity of food purchases (14.4 more items [95% CI −0.1 to 29.0]; P = 0.03), later time of breakfast purchases (15.0 minutes later [95% CI 1.5–28.5]; P = 0.03), and lower likelihood of preparing dinner at home (Q4 odds ratio [OR] = 0.3 [95% CI 0.1–0.9]; P = 0.03) relative to the lowest BMIGPS quartile. Compared with the lowest quartile, the highest BMICNS quartile was associated with fewer items purchased (P = 0.04), and the highest BMInon-CNS quartile was associated with purchasing breakfast at a later time (P = 0.01), skipping breakfast (P = 0.03), and not preparing breakfast (P = 0.04) or lunch (P = 0.01) at home. A limitation of this study is our data come from a relatively small sample of healthy working adults of European ancestry who volunteered to enroll in a health-promotion study, which may limit generalizability. Conclusions In this study, genetic risk for obesity was associated with the quality, quantity, and timing of objectively measured workplace food purchases. These findings suggest that genetic risk for obesity may influence eating behaviors that contribute to weight and could be targeted in personalized workplace wellness programs in the future. Trial registration Clinicaltrials.gov NCT02660086.
Hassan Dashti and colleagues investigate associations between genetic risk for obesity and food choice behaviors using workplace food purchases.
Author summary Why was this study done? Genetics play a role in the development of obesity, yet the influence of genetic risk for obesity on food choice behaviors is not well understood. Workplace cafeteria purchasing data provided an opportunity for objective, real-time assessment of employees’ food choices. What did the researchers do and find? We conducted a secondary analysis of baseline data of cafeteria food purchases prior to the start of a workplace intervention in a health-promotion randomized control trial using data from 397 employees who accepted to provide DNA for genetic analyses. We observed that the highest quartile of a genome-wide polygenic score for body mass index (highest genetic risk for obesity) was associated with lower dietary quality of all purchases, higher quantity of food purchases, later time of breakfast purchases, and lower likelihood of preparing dinner at home relative to the lowest quartile. What do these findings mean? Genetic risk for obesity was associated with the quality, quantity, and timing of objectively measured workplace food purchases, suggesting that genetic risk for obesity may influence eating behaviors that contribute to weight.
Databáze: OpenAIRE
Externí odkaz:
Nepřihlášeným uživatelům se plný text nezobrazuje