Abstract
This prospective study explores high sensitivity C-reactive protein (hs-CRP) levels in relation to dietary patterns at two time points in European children.
Out of the baseline sample of the IDEFICS study (n = 16,228), 4020 children, aged 2-9 years at baseline, with available hs-CRP levels and valid data from a food frequency questionnaire (FFQ) at baseline (T0) and 2 years later (T1) were included. K-means clustering algorithm based on the similarities between relative food consumption frequencies of the FFQ was applied. hs-CRP was dichotomized according to sex-specific cutoff points. Multilevel logistic regression was performed to assess the relationship between dietary patterns and hs-CRP adjusting for covariates.
Three consistent dietary patterns were found at T0 and T1: ‘animal protein and refined carbohydrate’, ‘sweet and processed’ and ‘healthy’. Children allocated to the ‘protein’ and ‘sweet and processed’ clusters at both time points had significantly higher odds of being in the highest category of hs-CRP (OR 1.47; 95% CI 1.03-2.09 for ‘animal protein and refined carbohydrate’ and OR 1.44; 95% CI 1.08-1.92 for ‘sweet and processed’) compared to the ‘healthy’ cluster. The odds remained significantly higher for the ‘sweet and processed’ pattern (OR 1.39; 95% CI 1.05-1.84) when covariates were included.