Studies have reported an inverse association between a diet rich in fruits and vegetables and type 2 diabetes (T2D), but data on high-risk ethnic minority groups is limited. We investigated whether serum carotenoids, as biomarkers for fruit and vegetable intake, mediate ethnic differences in the prevalence of T2D.

Age-adjusted serum carotenoid concentrations were compared using ANCOVA. A cross-sectional analysis was performed using Cox regression to estimate prevalence ratios (PR) and their 95 % CI of the association between serum carotenoid concentrations and T2D. To study whether serum carotenoids potentially mediate the differences in the prevalence of T2D across ethnic groups, we compared PR of the model including known risk factors and the model additionally adjusted for serum carotenoid concentrations using the Dutch group as reference.

A study among six ethnic groups living in Amsterdam, the Netherlands.

Data on 204 Dutch, 203 South Asian Surinamese, 204 African Surinamese, 203 Turkish and 200 Moroccan-origin participants from the Healthy Life in an Urban Setting (HELIUS) study were used.

Serum carotenoid concentrations differed across ethnic groups. After adjusting for confounders, the serum concentrations of total carotenoids (PR 0·67, 95 % CI 0·54, 0·84), α-carotene (PR 0·57, 95 % CI 0·42, 0·77), β-carotene (PR 0·45, 95 % CI 0·32, 0·63) and β-cryptoxanthin (PR 0·73, 95 % CI 0·58, 0·92) were inversely associated with T2D. Despite the associations, serum carotenoids did not mediate the ethnic differences in the prevalence of T2D.

The limited contribution of serum carotenoids to ethnic differences in T2D suggests that a focus on increasing fruit and vegetable intake alone will not likely eliminate ethnic differences in T2D prevalence.