Timing of initial cereal exposure in infancy and risk of islet autoimmunity

Dietary exposures in infancy have been implicated, albeit inconsistently, in the etiology of type 1 diabetes mellitus (DM).

To examine the association between cereal exposures in the infant diet and appearance of islet autoimmunity (IA).

Birth cohort study conducted from 1994 to 2002 with a mean follow-up of 4 years.

Newborn screening for HLA was done at St Joseph's Hospital in Denver, Colo. First-degree relatives of type 1 DM individuals were recruited from the Denver metropolitan area.

We enrolled 1183 children at increased type 1 DM risk, defined as either HLA genotype or having a first-degree relative with type 1 DM, at birth and followed them prospectively. We obtained exposure and outcome measures for 76% of enrolled children. Participants had variable lengths of follow-up (9 months to 9 years).

Blood draws for the detection of insulin autoantibody, glutamic acid decarboxylase autoantibody, or IA-2 autoantibody were performed at 9, 15, and 24 months and annually thereafter. Children with IA (n = 34) were defined as those testing positive for at least 1 of the autoantibodies on 2 or more consecutive visits and who tested positive or had diabetes on their most recent visit.

Children initially exposed to cereals between ages 0 and 3 months (hazard ratio [HR], 4.32; 95% confidence interval [CI], 2.0-9.35) and those who were exposed at 7 months or older (HR, 5.36; 95% CI, 2.08-13.8) had increased hazard of IA compared with those who were exposed during the fourth through sixth month, after adjustment for HLA genotype, family history of type 1 DM, ethnicity, and maternal age. In children who were positive for the HLA-DRB1*03/04,DQB8 genotype, adjusted HRs were 5.55 (95% CI, 1.92-16.03) and 12.53 (95% CI, 3.19-49.23) for initial cereal exposure between ages 0 to 3 months and at 7 months or older, respectively.

There may be a window of exposure to cereals in infancy outside which initial exposure increases IA risk in susceptible children.