Our study has several limitations. First, data on gestational age and birth weight were unvalidated self-reports.
However, validation studies demonstrated good self-report of related pregnancy factors (16,19). Also, several
validation studies show moderate to high reliability of self-report of preterm birth and infant birth weight when
compared with medical records (29–31). Second, we used categories of gestational age and birth weight instead of
continuous measures, which may make it difficult to see subtle changes in disease risk. Furthermore, these
categorizations prevented us from examining small-for-gestational age or large-for-gestational-age infants; as such we
had to restrict our analysis of birth weight to term births. Thus, our finding of an early elevation in T2DM risk for
women who delivered infants that were term low birth weight may not be generalizable to women who delivered
preterm infants, who were small for gestational age. We also had limited power to examine recurrent complications in
later pregnancies. Finally, we had limited ability to evaluate this association among minorities, who have both a higher
prevalence of these pregnancy complications and T2DM.