Aims/hypothesis: This study aimed to establish the relation between early infant nutrition and signs of beta-cell autoimmunity in young children.
Methods: We identified and observed from birth 2949 infants with increased genetic risk of Type I (insulin-dependent) diabetes mellitus (HLA DQB1*02/ *0302 or DQB1*0302/x, x = other than *02, *0301 or *0602) and monitored them for islet cell antibodies at 3 to 6 month intervals. If an infant seroconverted to islet cell antibody positivity, all of his or her samples were also analysed for autoantibodies to insulin, GAD65 (GADA) and to the protein tyrosine phosphatase related IA-2 molecule (IA-2A). Our case-control study comprises the first 65 children who seroconverted to islet cell antibody positivity before the age of 4 years and 390 control children who were islet cell antibody-negative (six control children/ case). We monitored the duration of exclusive and total breastfeeding and the age at which cows' milk was introduced.
Results: Infants who had been breastfed exclusively for at least 4 months had lower risk of seroconversion to positivity for IA-2A or all four autoantibodies [odds ratio (OR) 0.24; 95 % CI 0.06-0.94 and OR 0.17; 95 % CI 0.03-0.86, respectively] than those infants who had been breastfed exclusively for less than 2 months. The risk of seroconversion to positivity for IA-2A or all four autoantibodies was higher in those younger than 2 months (OR 4.37; 95 % CI 1.33-14.42 and OR 5.02; 95 % CI 1.27-19.89) or aged 2 to 3.9 months (OR 5.50; 95 % CI 1.21-25.04 and 6.19; 95% CI 1.10-34.84) when they first received cows' milk than in those aged 4 months or older.
Conclusions/interpretation: These observations suggest that short-term breastfeeding and the early introduction of cows' milk-based infant formula predispose young children who are genetically susceptible to Type I diabetes to progressive signs of beta-cell autoimmunity.