Cross-sectional Association between Chocolate Intake^a and Body Mass Index (BMI)^b in the ARIC^c Cohort.

a<p>Frequency of chocolate intake was assessed by means of a semi-quantitative food frequency question.</p>b<p>BMI, Body Mass Index - weight in Kg divided by height in m squared.</p>c<p>ARIC, Atherosclerosis Risk in Communities Cohort.</p>d<p>N is the number of participants who provided data values at visit 1 and 3. Missing values of the outcome, exposure and confounder variables were replaced by multiple imputation.</p>e<p>Model A - adjusted for age, age squared, race (non-black, black), sex (male, female);</p>f<p>Model B - adjusted for model A variables plus alcohol intake (g/week), smoking status (never, former, current <20/day, current ≥20/day), educational level (grade school or less; some high school; high school graduate; vocational school; some college; graduate or professional school), prevalent illness (preexisting, physician diagnosed heart attack, stroke, diabetes, or cancer), and dietary caloric intake.</p>g<p>Model C - adjusted for model B variables plus energy-adjusted dietary vegetable, fruit and fat levels.</p>h<p>Prevalent obesity-related illness was based on yes/no responses to questions about the existence of physician diagnosed heart attack, stroke, diabetes, and cancer.</p>j<p>the smoking confounder was a continuous variable in these two models.</p>k<p>Data are presented as mean and 95% confidence interval, based on a random intercept model using combined data from ARIC visit 1 and 3 (see Statistical Methods).</p>l<p>P for linear trend was assessed with the linear version of the exposure variable in the model. P for quadratic trend was assessed with both linear and quadratic versions of the exposure variable in the model.