Choline/betaine habitual intake and incident coronary heart disease

Increased whole blood levels of total choline have previously been found to be associated with cardiovascular disease. We estimated the association between dietary choline and betaine intake and risk of incident CHD in the Women's Health Initiative (WHI) Observational Study (n=93,676) and the WHI Dietary Modification trial data (n=48,835) collected over 8-12 years of follow-up. Food consumption data, quantified from the WHI Food Frequency Questionnaire (FFQ) were converted into estimates of daily intake of choline and betaine. These FFQ estimates were compared to four-day food records (FDFR) estimates, and reliability was quantified through the comparison of FFQs administered at baseline and one year follow-up. Based on the intra-individual variability for all interrelated nutrients models, we estimated the associations between choline and betaine and incident CHD corrected for reliability. Median daily choline and betaine intakes were 264 (IQR: 201−341) and 164 (IQR: 114−230). Only 10% of the population (n=14,313) met the adequate intake of 425 mg/day. Median daily intakes of choline and betaine respectively were 268 mg and 166 mg in Whites, 242 mg and 164 mg in Blacks, 250 mg and 123 mg in Hispanic/Latinos; 233 mg and 150 mg in Asian or Pacific Islanders. A mixed model assessing the joint intra-individual variability for all interrelated nutrients yielded reliability coefficients of (betaine) 0.50; (choline) 0.59; (total energy intake) 0.55; (total folate) 0.30; (B6) 0.60, (B12) 0.55; (methionine) 0.56. Correlations between FFQ and FDFR estimates were highest for natural folate (0.32); choline (0.30); methionine (0.27). Correlations were lowest for betaine (0.22); total folate (0.12); B6 (0.08); B12 (0.10). Hazard ratios (HRs) associated with a one standard deviation (SD) nutrient density adjusted choline intake (35.30 mg/dy per kilocarlorie) were 1.26 [95% confidence interval (CI): 1.09,1.46] for Blacks and 1.06 (95% CI: 1.01,1.11) for Whites. After reliability correction, HRs were further from the null in Blacks [1.42 (95% CI: 1.03,1.86)] and Whites [HR: 1.13 (95% CI: 1.02,1.23)].;In conclusion, choline intake was postively associated with incident CHD in Whites and Blacks with a substantially greater effect estimate within Blacks. Correction for reliability demonstrated that measurement error results in substantial attenuation of effect estimates.