The purpose of this research is to investigate the association between consumption of chocolate and measures of adiposity in a large, representative sample of US adults.

Cross-sectional data from 13,626 nondiabetic adults (≥20 years) participating in the National Health and Nutrition Examination Survey study were aggregated using 5 study cycles from 2005-2006 through 2013-2014. Chocolate consumption was determined based on 2 24-hour dietary recalls. Body mass index (BMI) and waist circumference were objectively measured. We used multivariable linear regression to test associations of 1) any chocolate consumption (yes/no), and 2) the total amount of chocolate consumption (grams/day, in quartiles) with BMI and waist circumference. Models controlled for sociodemographic, lifestyle, health-related, and dietary covariates.

Overall, 11.1% of the population self-reported any chocolate consumption in either of their 2 24-hour dietary recalls. Adjusted linear regression models showed that individuals who reported any chocolate consumption had 0.92 kg/m² (95% confidence interval, 0.53-1.32) lower BMI, and 2.07 cm (95% confidence intervals, 1.22-2.92) lower waist circumference than those who reported no chocolate consumption. In models examining the association of amount of chocolate consumption and weight status, compared with those who did not consume chocolate, lower BMI (P for trend = .003) and waist circumference (P for trend = .001) were observed in the first, second, and third quartiles of total chocolate consumption.

In this representative sample of US adults, chocolate consumption was associated with lower markers of adiposity. Further research using a longitudinal or experimental design is needed to establish the direction of causation.