ADVANCING PATERNAL AND MATERNAL AGE ARE BOTH IMPORTANT FOR AUTISM RISK

King et al.¹ addressed 2 questions using data from the California Department of Developmental Services. The first concerned possible bias in estimating the association between parental age and autism as a result of pooling data across birth years. In their analysis, pooling caused a “reversal paradox” for paternal age—that is, the association between paternal age and autism was substantially inflated in pooled data. On this basis, they argued that the relation between paternal age and autism has been overestimated in previous research.

We, however, do not agree with this conclusion. Previous studies have considered birth years and the effects of pooling data. Birth year is often included as a confounder in the statistical models, and data by birth year are often inspected by researchers to check for such possible pooling artifacts.^2–6 In fact, a recent study by Grether et al.⁵ used the same California data source as King et al. and found no evidence for an inflated paternal-age effect when pooling data. For every 10-year increase in paternal age the pooled odds ratio (OR) for autism was 1.22 (95% confidence interval [CI]=1.18, 1.26). The equivalent OR for individual years was lower than was the pooled OR in 7 of the 14 individual years and higher in 6 (Table 2 in Grether et al.⁵). These results suggested that bias, if it exists, is only minimal. Furthermore, the research clearly indicated that an association between paternal age and autism in offspring can be detected in most populations and is not an artifact of confounding by maternal age or of pooling across birth cohorts.

The Maneki Neko (literally “Beckoning Cat”) is a common Asian sculpture, often made of ceramic, which is believed to bring good luck to the owner. The sculpture depicts a cat (traditionally a Japenese Bobtail) beckoning with an upright paw, and is usually displayed—many times at the entrance—in shops, restaurants, pachinko parlors, and other businesses. In the design of the sculptures, a raised right paw supposedly attracts money, while a raised left paw attracts customers. Image from Punchstock.com. Printed with permission.

The second question that King et al. addressed is the relative importance of paternal or maternal age. The authors used a statistical method of decomposition—about which we have some qualms—to examine this. However, is it really important whether the association with autism is stronger for paternal or for maternal age? Is there a “competition” between the two? At this point, an abundance of large and well-designed studies have demonstrated that advancing maternal and paternal age are associated with autism. One can draw different conclusions about their relative importance depending upon the perspective adopted. For example, men can and often do conceive at much later ages than can women, broadening the paternal age range. So, even if 10 years of maternal age confers greater risk than does 10 years of paternal age, the oldest men carry greater risk than do the oldest women (Table 1 in Grether et al.⁵).

Regardless of the relative importance of paternal and maternal age, it is noteworthy that so many credible epidemiologic studies (including that conducted by King et al.) have now confirmed that paternal and maternal age at birth are related to autism risk. The evidence is substantial enough to justify a search for the underlying mechanisms. Genomic alterations could be involved in either the case of men or of women,⁷ as could other factors. All possible mechanisms can and should be studied in both human and animal models.⁸