Levels and patterns of prenatal alcohol exposure associated with increased risk of a clinically meaningful adverse...

Recent interest in the screening and diagnosis of fetal alcohol spectrum disorders (FASD) by the American Psychiatric Association and the American Association of Pediatrics has highlighted the lack of empirical evidence regarding the levels and patterns of prenatal alcohol exposure (PAE) associated with clinically significant adverse effects on cognitive and behavioral function. We used benchmark dose analysis to address this issue.

Benchmark dose analyses were performed on composite cognitive function scores derived from data obtained at school age, adolescence, and early adulthood from 2227 participants from six prospective, longitudinal cohorts, whose mothers were interviewed about their alcohol and drug use during pregnancy or shortly thereafter. We compared single-predictor models based on average alcohol consumed per day during pregnancy with a two-predictor, semiparametric model based on (1) average alcohol dose per drinking occasion and (2) drinking frequency.

Our two-predictor model showed that at lower levels of drinking frequency, a relatively high dose/occasion is required for an increased risk of poor intellectual function, whereas at higher levels of drinking frequency, a lower dose/occasion is sufficient to increase that risk. In addition, lower doses/occasion were associated with clinically meaningful adverse effects in participants born to older than to younger mothers. The single-predictor models proved less adequate because the dose/occasion for a substantial proportion of participants exposed at the benchmark doses generated by those models was too low to increase the risk of clinically significant effects.

Given that the cognitive and behavioral deficits seen in FASD resemble those also seen in a range of other disorders, whether the patient's PAE is sufficient to increase the risk of clinically meaningful impairment is a critical element in an FASD diagnosis. Accurate diagnosis is needed to determine which patients warrant interventions that have been shown to be effective in remediating PAE-related neurobehavioral impairment.