Introduction Celiac disease (CD) is a chronic small bowel disease with a strong heritability. Several perinatal conditions are today considered as risk factors for CD in children. One of these conditions is being born small for gestational age (SGA). However, if the association between SGA and CD risk in children is causal is unknown. Therefore, we aimed to apply a quasi-experimental family based (QEFB) design that is a powerful strategy for studying causal relationships.

Methods Using the Swedish Medical Birth Registry linked to a number of other national databases we identified all singleton children born in Sweden between 1987 and 1993 (n =781 624). We applied a QEFB design, and compared sibling with discrepant exposure (ie, being or not SGA) in relation to their risk of CD from birth until they were 2-year old. We also performed classical logistic regression analyses adjusting for known risk factors for CD but without the QEFB design.

Results In the classical adjusted logistic regression analysis, we found an association between being SGA and CD risk: OR 1.34 95% CI 1.03 to 1.74. However when applying the QEFB design and conditional logistic regression this association disappeared: OR 1.05 95% CI 0.53 to 2.06.

Conclusion Our results suggest that previous finding indicating an association between being SGA and CD risk were confounded. These previous results might be explained by the fact that the offspring from mothers with CD are frequently SGA and, because the strong heritability, they have also a higher risk for CD.