Infant Feeding and Risk of Type 1 Diabetes in Two Large Scandinavian Birth Cohorts

Abstract

Objective: Our aim was to study the relation between the duration of full and any breastfeeding and risk of type 1 diabetes.

Research design and methods: We included two population-based cohorts of children followed from birth (1996-2009) to 2014 (Denmark) or 2015 (Norway). We analyzed data from a total of 155,392 children participating in the Norwegian Mother and Child Cohort Study (MoBa) and the Danish National Birth Cohort (DNBC). Parents reported infant dietary practices when their child was 6 and 18 months old. The outcome was clinical type 1 diabetes, ascertained from nationwide childhood diabetes registries. Hazard ratios (HRs) were estimated using Cox regression.

Results: Type 1 diabetes was identified in 504 children during follow-up, and the incidence of type 1 diabetes per 100,000 person-years was 30.5 in the Norwegian cohort and 23.5 in the Danish cohort. Children who were never breastfed had a twofold increased risk of type 1 diabetes compared with those who were breastfed (HR 2.29 [95% CI 1.14-4.61] for no breastfeeding vs. any breastfeeding for ≥12 months). Among those who were breastfed, however, the incidence of type 1 diabetes was independent of duration of both full breastfeeding (HR per month 0.99 [95% CI 0.97-1.01]) and any breastfeeding (0.97 [0.92-1.03]).

Conclusions: Suggestive evidence supports the contention that breastfeeding reduces the risk of type 1 diabetes. Among those who were breastfed, however, no evidence indicated that prolonging full or any breastfeeding was associated with a reduced risk of type 1 diabetes.

Figure 1

Flowchart showing participants from the MoBa (A) and the DNBC (B) included in the analysis. *In MoBa, 95,200 mothers participated during one or more pregnancies (40.6% of eligible pregnancies). In DNBC, 91,326 mothers participated during one or more pregnancies (about 30% of eligible mothers). †The DNBC interview when the infants were 6 months old did not include detailed questions about age at introduction of formula until April 2000, resulting in missing data for age at introduction of formula and thus also for duration of full breastfeeding for 12,402 of the 66,676 children whose mothers completed the 6-month interview (including 48 children with type 1 diabetes [T1D]).

Figure 2

Association of duration of any breastfeeding (A) and full breastfeeding (B) with risk of type 1 diabetes (T1D) in the MoBa and DNBC. The vertical lines show the 95% CIs. The upper confidence limit was 10.4 (arrow in panel B). HRs were adjusted for the following covariates: parental type 1 diabetes, infant sex, mode of delivery, infant birth weight, and gestational age at delivery, as well as the mother’s parity, age at delivery, education, smoking during pregnancy, and BMI before pregnancy. Exact numbers and incidence rates are shown in Supplementary Table 3. A total of 14 children had type 1 diabetes among those with no breastfeeding, reflecting the relatively wide CIs for this group. The pooled trend-adjusted HR per month difference in any breastfeeding was 0.99 (95% CI 0.97–1.01) overall and 0.99 (95% CI 0.97–1.01; P = 0.89) after excluding those not breastfed at all. The pooled trend-adjusted HR per month difference in full breastfeeding was 0.97 (95% CI 0.92–1.03) overall and 0.99 (95% CI 0.94–1.05; P = 0.73) after excluding those not breastfed at all.