Perinatal risk factors for type 1 diabetes mellitus

Higher birth weight (over 4000 g) slightly increases the risk of developing type 1 diabetes in childhood. Newborns weighing over 4 kg have an approximately 10% higher risk, with a linear increase of about 3% for every additional 500 g [1]. In type 2 diabetes, both low and high birth weight increase the risk. In type 1 diabetes, by contrast, low birth weight (under 2500 g) even appears to be protective.

The proposed mechanism is related to the accelerator hypothesis. A heavier newborn has a higher insulin demand, which overloads the pancreatic beta cells. They therefore become more visible to the immune system. Although the effect is modest, it is one of the best replicated findings in the research on perinatal factors for type 1 diabetes.

Cesarean delivery has been associated with an approximately 20% higher risk for the child to develop type 1 diabetes, compared with vaginal delivery [2]. The main hypothesis is that newborns delivered by cesarean section are not exposed to the mother's vaginal and intestinal microbiota. This alters early intestinal colonization, with reduced levels of Bacteroides and Bifidobacterium.

However, recent studies have considerably weakened the causal argument [3]. For example, the prospective TEDDY study has followed children at high genetic risk since birth. It found no association between cesarean section and the subsequent development of type 1 diabetes-specific autoimmunity in the child. Currently, this association is considered possible, but various confounding factors could explain part of the additional risk.

Yes, breastfeeding appears to offer a moderate protective effect against type 1 diabetes. Prolonged breastfeeding (over 6 months) reduces the risk of type 1 diabetes by 61%, and exclusive breastfeeding for at least 2–3 months reduces the risk by 32%. Compared with prolonged breastfeeding, the absence of breastfeeding is associated with an approximately 2.5-fold higher risk of type 1 diabetes [4]. The biggest difference appears to be between breastfeeding, however briefly, and not breastfeeding at all.

The protective mechanisms of breastfeeding include the transfer of immunoglobulins (type A, which protect the intestinal mucosa) and the promotion of beneficial bacteria (Bifidobacterium). They also include the strengthening of the intestinal barrier and the delay of exposure to complex foods. However, avoiding cow's milk proteins through hydrolyzed formulas does not reduce the risk of type 1 diabetes.

The timing of solid food introduction appears to matter, and the optimal window is probably between 4 and 6 months of age. Introducing gluten at 3–6 months (compared with before 3 months) is associated with a significant reduction in type 1 diabetes risk [4]. Very late introduction of gluten (after 9 months) could increase the risk, but this has not been confirmed.

In children at increased risk, delaying gluten introduction from 6 to 12 months does not increase the risk of type 1 diabetes [5]. The amount of gluten-containing cereals consumed may matter, not just the timing of introduction. The biological explanation centers on the maturation of the intestinal immune system. In the 4–6 month window, the gut-associated lymphoid tissue learns to tolerate foods. Introducing them too early (immature barrier) or too late (missed learning window) may favor immune activation rather than tolerance.

The hypothesis that early introduction of cow's milk in the diet could trigger type 1 diabetes dominated research for decades. Early observational studies reported significant associations. The molecular mimicry hypothesis proposed that bovine serum albumin (BSA) causes cross-reactions with a beta-cell autoantigen. In fact, later introduction of cow's milk (after 2–3 months) is associated with a lower risk of type 1 diabetes [4].

Extensively hydrolyzed casein formula was compared with conventional cow's milk formula over a median follow-up of 11.5 years. The incidence of type 1 diabetes was similar, meaning the hydrolyzed formula had no effect [6]. If intact cow's milk proteins had been a major trigger, the hydrolyzed formula should have reduced the incidence, but it did not. Cross-reactivity between anti-BSA antibodies and the pancreatic beta-cell autoantigen has never been proven at the molecular level. Therefore, cow's milk exposure remains at most a minor modulating factor, and its avoidance cannot be recommended as a prevention strategy for type 1 diabetes.

Overall, preterm birth (before 37 weeks) is associated with a slightly increased risk of type 1 diabetes. Very preterm birth (before 32 weeks) is associated with a lower risk than birth at term, while birth at 32–36 weeks slightly increases the risk [3]. However, moderate and late preterm births predominate numerically, which explains the slightly increased overall risk.

Children born very preterm have a very immature immune system at birth. They are also exposed to glucocorticoids, administered to mothers for lung maturation. This exposure and differences in pancreatic beta-cell development may explain the protective (apparently paradoxical) effect.

Yes, advanced maternal age is associated with a slightly increased risk of type 1 diabetes in the child. The relationship appears to be linear, with a 5–10% increase in risk for every five additional years of maternal age, with no plateau at any particular age. The association persists even after adjustment for birth weight, gestational age, birth order, maternal diabetes and breastfeeding [7]. Paternal age has shown no clear association with the risk of type 1 diabetes.

Proposed mechanisms include the accumulation of environmental exposure over the mother's lifetime, age-related epigenetic changes, increasing maternal body mass index with advancing age, and possible changes in immune function. Your age alone should not cause you additional concern about having a child with type 1 diabetes, because the effect is modest.

Certain maternal infections during pregnancy can increase the child's risk of type 1 diabetes, but the effect varies considerably depending on the type of infection. Overall, maternal infections during pregnancy increase the risk by 31%, with a more pronounced effect for enterovirus infections (54% increased risk) [8]. The strongest established association is with congenital rubella. Approximately 12–20% of children with congenital rubella syndrome develop type 1 diabetes, and over 40% have impaired glucose tolerance.

Maternal cytomegalovirus infection has been associated with an approximately fourfold higher risk, but the data are limited. Respiratory infections in the first trimester might be associated with subsequent development of type 1 diabetes in the child. This association has not, however, been confirmed. Regarding COVID-19, there is not yet sufficient evidence concerning maternal infection during pregnancy.

Yes, maternal gestational diabetes is associated with a nearly twofold risk of type 1 diabetes in the child. Gestational diabetes increases the child's risk of type 1 diabetes, with an odds ratio (OR) of 1.94 [9]. Pre-existing maternal type 1 diabetes increases the risk even more (estimates range between 3 and 11 times, depending on the study).

The accelerator hypothesis states that intrauterine hyperglycemia programs insulin resistance in the child, increasing the metabolic stress on beta cells and subsequently accelerating autoimmune destruction. The second proposed mechanism is epigenetic programming, in which gestational diabetes induces DNA methylation changes in cord blood, so that exposed children show accelerated epigenetic aging. The third implicated mechanism is the mother's genetic susceptibility to type 1 diabetes, which is passed on to the child. The mother may even subsequently develop autoimmune diabetes herself, as the figure below shows.

First-born children appear to have a slightly higher risk of type 1 diabetes than younger siblings. The effect is small, though, and does not appear in all studies. It is more pronounced and more consistent during the first five years of the child's life and then decreases with age [10].

As a possible explanation, the hygiene hypothesis assumes that first-born children have lower early microbial exposure than children with older siblings, which could affect the development of immune tolerance. The evidence currently available is not sufficient to support the hygiene hypothesis in the development of type 1 diabetes. Therefore, you have no reason for additional concern based on the order in which your children were born.

There is suggestive but not definitive evidence that antibiotics administered repeatedly in the first year of life are associated with a slightly increased risk of type 1 diabetes. The risk is higher in children born by cesarean section. It appears that only broad-spectrum antibiotics administered in the first year of life have been associated with a higher risk of type 1 diabetes. Narrow-spectrum antibiotics had no effect in this regard [11].

It appears that the antibiotics themselves are not to blame for this association, but rather the reason for which they were prescribed. The most recent studies, which attempt to account for this, have not found significant causal associations. The proposed mechanism is disruption of the gut microbiome [12].

Probably not, according to recent data. Neonatal jaundice was historically associated with a slightly increased risk of type 1 diabetes, but this association has weakened considerably. Today, phototherapy is no longer associated with type 1 diabetes [13].

The association appears present only in older birth cohorts (1973–1982) and absent in recent decades, likely reflecting changes in Rh incompatibility prophylaxis and phototherapy practices [3]. Bilirubin at moderate concentrations actually has antioxidant properties and could theoretically protect beta cells.

Blood group itself is not a risk factor for type 1 diabetes. Blood group incompatibility between mother and child was historically associated with an increased risk of type 1 diabetes. In modern cohorts followed from birth, this association no longer appears [3] [14].

It is interesting that the HLA-DR3 variant (a risk factor for type 1 diabetes) is over-represented in children with type 1 diabetes. It also appears more frequently in children with blood group incompatibility with their mother. This shared genetic susceptibility may explain the epidemiological association without implying a direct causal link.