Carl Johan Wingren1, Daniel Agardh, Juan Merlo. 1. Department of Clinical Sciences, Faculty of Medicine, Lund University, Unit for Social Epidemiology, Malmö, Sweden. carl_johan.wingren@med.lu.se
Abstract
OBJECTIVE: An association between small for gestational age (SGA) and risk for celiac disease (CD) in childhood has previously been reported. However, this association may reflect residual confounding by genetic or environmental factors. For example, presence of subclinical CD in the mother might be a common cause of both SGA and CD in the offspring. We investigate whether SGA is causally associated with CD before age six years by applying both conventional population-based regression models and sibling analysis that investigates the association in siblings discordant for SGA. MATERIAL AND METHODS: Using the Swedish Medical Birth Registry, we identified all singleton children born in Sweden during 1987-1993 (792,401). Of these we included 681,954 children in the study and identified 2641 cases of CD using the Swedish National In-Hospital Registry. We applied both conventional Cox regression analysis and a quasi-experimental sibling design that to some extent simulates a counterfactual situation of exposure, reducing possible confounding effects of genetic and shared environmental factors. RESULTS: We identified an increased risk of CD in both boys (hazard ratio (HR) 1.70, 95% confidence interval (CI) 1.25-2.32) and girls (HR 1.30, 95% CI 0.99-1.70) using conventional Cox regression models. Using sibling analysis, the association between SGA and CD was confirmed in boys (HR 4.23, 95% CI 1.19-15.04) but not in girls (HR 1.00, 95% CI 0.45-2.20). CONCLUSIONS: Our results support a causal association between SGA and CD risk in boys but not in girls, although the mechanisms underlying this difference are still unclear.
OBJECTIVE: An association between small for gestational age (SGA) and risk for celiac disease (CD) in childhood has previously been reported. However, this association may reflect residual confounding by genetic or environmental factors. For example, presence of subclinical CD in the mother might be a common cause of both SGA and CD in the offspring. We investigate whether SGA is causally associated with CD before age six years by applying both conventional population-based regression models and sibling analysis that investigates the association in siblings discordant for SGA. MATERIAL AND METHODS: Using the Swedish Medical Birth Registry, we identified all singleton children born in Sweden during 1987-1993 (792,401). Of these we included 681,954 children in the study and identified 2641 cases of CD using the Swedish National In-Hospital Registry. We applied both conventional Cox regression analysis and a quasi-experimental sibling design that to some extent simulates a counterfactual situation of exposure, reducing possible confounding effects of genetic and shared environmental factors. RESULTS: We identified an increased risk of CD in both boys (hazard ratio (HR) 1.70, 95% confidence interval (CI) 1.25-2.32) and girls (HR 1.30, 95% CI 0.99-1.70) using conventional Cox regression models. Using sibling analysis, the association between SGA and CD was confirmed in boys (HR 4.23, 95% CI 1.19-15.04) but not in girls (HR 1.00, 95% CI 0.45-2.20). CONCLUSIONS: Our results support a causal association between SGA and CD risk in boys but not in girls, although the mechanisms underlying this difference are still unclear.