The risk of developing autism is increased for siblings from closely spaced pregnancies and is highest for those conceived within a year of their sibling’s birth. In a study of sibling pairs born in California, children conceived a year or less after their firstborn sibling had 3.4 times as high odds of having autism as those conceived at least three years later.1 Autism odds for children conceived 12–23 and 24–35 months after their firstborn sibling were 1.9 and 1.3 times as high, respectively.

Using California’s Birth Master Files, investigators identified firstborn and second-born full sibling pairs born between 1992 and 2002. They collected paternal and maternal demographic information, as well as information on gestation and source of medical payment. They calculated the period between the two pregnancies (the interpregnancy interval) by subtracting the second-born infant’s gestational age from the total number of days between the infants’ birth dates. Sibling pairs were excluded if the second sibling’s gestational age was missing or if the interpregnancy interval was less than 20 weeks, greater than 44 weeks or a negative number. The investigators identified children who had satisfied diagnostic criteria for autism by researching case files from the state Department of Developmental Services (DDS).

Preliminary analyses revealed an association between shorter interpregnancy inter-vals and characteristics of lower autism risk among second-born siblings without autism—younger parental ages, lower maternal education, enrollment in Medi-Cal (California’s publicly financed health insurance program) and earlier birth year. Logistic regression analyses of 662,730 sibling pairs, which controlled for these characteristics, and in which the firstborn did not have autism, found more than threefold elevated odds of developing autism among second siblings conceived less than 12 months later (odds ratio, 3.4), compared with those conceived at least 36 months later. The odds for second siblings conceived 12–23 months and 24–35 months after the first were also significantly elevated (1.9 and 1.3). Although preterm birth and low birth weight were associated with interpregnancy intervals shorter than one year, the odds of autism remained the same when either of these was added to the model. The association held across social and demographic characteristics, although it weakened in the oldest and youngest parents.

Investigators also conducted a conditional logistic regression analysis of 5,327 sibling pairs in which either child had received an autism diagnosis. Children conceived less than 12 months after a firstborn sibling were at increased risk for autism, compared with that sibling (odds ratios, 2.0 for an interpregnancy interval of 3–5 months, 2.3 for an interval of 6–8 months and 2.1 for a 9–11-month interval).

The investigators note some bias-related limitations in their study. For example, children with autism who were not diagnosed or did not receive services through DDS were not identified by the study. Also, the data were subject to administrative errors, such as potential erroneous reporting of gestational ages. Nevertheless, one of the study’s strengths was its use of California DDS records, which allowed analyses of birth records that spanned 11 years and autism diagnoses in a heavily populated and diverse state.

Noting the absence of a definitive link between close birthspacing and autism, the investigators call for further exploration into the underlying mechanisms of each. They add, however, that short interpregnancy intervals, which may be planned, also may be associated with poor prenatal health and, in some cases, unintended pregnancy.—A. Kott