In your August issue, Robinson et al.1 reported that children prenatally exposed to low to moderate levels of alcohol within the first 3 months of pregnancy had significantly less behavioural problems on the Child Behaviour Checklist (CBCL) when compared with the offspring of mothers who abstained. This paper adds to the relatively meagre body of literature suggesting that alcohol consumption at low levels during pregnancy may not be associated with adverse behavioural outcomes; however, certain methodological limitations must be carefully considered when interpreting these results.
A serious limitation is the selective attrition of participants; the one-third of mothers lost to follow-up were more socially disadvantaged than those who participated. Although the sociodemographic characteristics of the remaining participants were similar to the Western Australian population, this does not negate the fact that the study included a select subset of participants from its original sample. Children born to more socially disadvantaged mothers may be particularly vulnerable to the teratogenic effects of alcohol for numerous reasons, such as poorer prenatal nutrition and stress. There may be important interactions among socioeconomic determinants and toxic exposures. Therefore, the conclusion that moderate drinking may not adversely affect the fetus could be erroneous, considering the sample is over-represented by healthy, resilient mothers.
Another concern is the large number of confounding variables that were not sufficiently accounted for, such as maternal psychopathology. As the authors noted, in general, individuals who consume moderate levels of alcohol tend to have better mental health than those who abstain.1 Recently, a prospective population-based study found a ‘U’-shaped association between alcohol use and the risk of depression and anxiety.2 It is possible that the group of abstainers in this study had higher rates of psychopathology than the group of light to moderate drinkers, which may have contributed to their child’s poorer CBCL scores. A retrospective review found that children with a history of parental psychopathology had significantly poorer scores on three subscales of the CBCL, when compared with those without parental psychopathology.3
Moreover, ratings on the CBCL could also be affected by the responders’ level of emotional impairment. Najman et al.4 found that mothers exhibiting clinically significant symptoms of anxiety and depression reported more behavioural problems on CBCL, when compared with their children’s own responses on the Youth Self-Report. It is unfortunate that the authors did not try to rectify this shortcoming by utilising other informants, including teachers and the children themselves, so as to corroborate the parent reports.
The attempt by Robinson et al. to conceive a biological mechanism to explain the better behavioural outcomes in offspring exposed to alcohol is problematic. The suggestion that moderate alcohol consumption ‘fixes’ maternal anxiety and hence renders a favourable stress-axis environment for the unborn seems like an alcohol advertisement from the 1970s, when the industry fought tooth and nail against the existence of fetal alcohol syndrome.
It is extremely concerning that, in the cohort of Robinson et al., most women continued to consume alcohol at similar levels throughout their entire pregnancy. This represents an unacceptable climate that should alarm medical professionals and public health authorities. With millions of women exhibiting problematic drinking, and an estimated 1% of babies affected by fetal alcohol spectrum disorder, this blasé and permissive approach is potentially dangerous, and may even encourage alcohol consumption during pregnancy in women at risk.