The EPIC-Norfolk Study
Fibre intake in relation to serum total cholesterol levels and CHD risk: a comparison of dietary assessment methods.
European Journal of Clinical Nutrition 2011 ; 66: 296-304.
Ward HA, Keogh R, Lentjes M, Luben RN, Wareham NJ, and Khaw KT
DOI : 10.1038/ejcn.2011.184
PubMed ID : 22045223
PMCID : 0
Abstract
Prospective diet diaries may be more accurate than retrospective food frequency questionnaires (FFQ). The objective of this study was to compare FFQ and 7-day diet diary (7DD) measurements of fibre intake with the incidence of coronary heart disease (CHD).
We compared 7DD and FFQ fibre intake in a nested case-control study in a population of 25639 men and women aged 40-79 years, surveyed in 1993-97 and followed up until 2007. Among 2151 CHD cases and 5354 controls, FFQ and 7DD fibre intake (6 g/day) was examined in relation to serum total cholesterol and CHD using linear and logistic regression adjusted for age and additionally for body mass index, physical activity, smoking status, family history of CHD, social class, diabetes, alcohol, energy, saturated fat and use of lipid lowering medication, antihypertensive medication or aspirin.
Age-adjusted serum total cholesterol was inversely associated with 7DD fibre among men and women, but with FFQ fibre among men only. In the multivariate analysis, associations with 7DD fibre were attenuated among men (regression coefficient -0.036 mmol/l, s.e. 0.021, P-value 0.087) and women (regression coefficient -0.069 mmol/l, s.e. 0.036, P-value 0.053), and were non-significant for FFQ fibre. Among men, age-adjusted CHD risk was inversely associated with 7DD fibre (odds ratio (OR) 0.84, 95% confidence interval (CI) 0.79-0.90), but not with FFQ fibre (OR 0.96, 95% CI 0.90-1.12). Among women, age-adjusted CHD risk was inversely associated with 7DD fibre (OR 0.83, 95% CI 0.75-0.93), and had a weaker inverse borderline-significant association with FFQ fibre (OR 0.93, 95% CI 0.87-1.01). Multivariate models yielded similar results.
Inconsistencies in diet-CHD relationships in population studies may be associated with the use of different dietary assessment methods.