Data from the Baby Siblings Research Consortium confirm and specify the nature of the female protective effect in autism: A commentary on Messinger et al.

The manuscript describes the BSRC observation that across a number of developmental competencies, modest sex differences were observed, and the discrepancy between boys and girls with ASD was not different in magnitude from that between typically developing boys and girls. The authors conclude that therefore sex differences were “not autism-specific.” Examining males and females within the category of the disease-of-interest is an important exercise in understanding sex differences, but there can be profound ASD-specific sex effects on a disease that render sex differences among affected subjects relatively minor and indistinguishable from those differentiating typical boys and girls. Conduct disorder provides an excellent example—the differences between males and females who meet diagnostic criteria are minimal (by definition) and yet males are six times more likely to be affected than females.

In the case of ASD, the first major clue generated by these data was that the magnitude of sex difference for any given trait within any subject group was dwarfed by the magnitude of difference across groups (ASD versus non-ASD). Since the “quantum leap” from typicality to clinical abnormality (the magnitude of which, in relation to within-group differences, is illustrated in figure) occurred over three times more commonly for male siblings than for female siblings (this was the second major clue generated by the study data), the data actually represent a remarkable manifestation of sex-specific modulation of inherited liability, which is the fundamental nature of the so-called female protective effect. The refocusing of the problem on contrasts across groups rather than within groups is represented by comparing the arrows and relative sample sizes in Fig. 1, which is a schematic reconstruction-to-scale of Figure three of the original manuscript.

Bild vergrößern

In prior published work, we have presented evidence for the manner in which a female protective effect might operate in the absence of the Carter effect, essentially via a categorical mechanism of protection by which most females are protected from most forms of ASD [4, 7‐9]. The BSRC data reported in the manuscript is entirely consistent with such modulation of risk by sex, and we have shown in quantitative trait analyses among school-aged children in both the AGRE and IAN registries how such protection manifests itself among females in multiplex autism families [4, 10].