Alcohol use during pregnancy is common and remains a significant threat to the development and health of exposed offspring. A recent survey by the National Birth Defects Prevention Study 1
found that nearly one-third of women drank alcohol at some time during their pregnancy; 22.5% drank during the first month of pregnancy and 25% drank during the first trimester. Longitudinal analyses of the 1991–2005 Behavioral Risk Factor Surveillance System surveys indicate that the rate of past-month alcohol use among pregnant women was unchanged across this 14-year period 2
. Thus, drinking during pregnancy remains prevalent despite an advisory from the Surgeon General 3
to limit alcohol use if pregnant or contemplating pregnancy.
Alcohol exposure during gestation is associated with varied effects on development, depending on the amount and timing of exposure. Heavy alcohol use during pregnancy is the cause of Fetal Alcohol Syndrome (FAS), a well-defined syndrome that includes growth deficits, facial dysmorphology, and CNS abnormalities 4
. However, most women who consume alcohol during pregnancy are light-to-moderate drinkers in early pregnancy and quit or decrease their alcohol use by mid-pregnancy. 5–7
Moderate or light alcohol consumption during pregnancy is associated with fewer and less severe effects.
Children with FAS and those with high levels of prenatal alcohol exposure (PAE) have higher scores on the aggression subscales of the Child Behavior Checklist (CBCL) 8–11
. Sood et al., 12
found a dose-response association between PAE and behavior problems. At low levels of PAE, offspring had more aggressive and externalizing behaviors. Higher levels of exposure predicted more delinquent behaviors and a higher total problem score on the CBCL. By contrast, a recent longitudinal study, 13
found no effects of low levels of PAE on problem behaviors; moderate first trimester PAE predicted scores in the clinical range of aggression on the CBCL, and both moderate and heavy exposure in late pregnancy predicted total externalizing behaviors. Using a large representative U.S. sample, D'Onofrio et al. 14
examined the association of PAE to maternal ratings of behavior problems among siblings with differing levels of exposure, a study design that made it possible to control for unmeasured environmental and genetic confounders. They reported a dose-response relation between PAE and externalizing behaviors and concluded the association was causal.
Conduct Disorder (CD) in DSM-IV is a diagnosis 15
that describes a pattern of severe behavior problems persisting for more than 12 months with at least one behavior present in the past six months. The diagnostic criteria include aggression toward people and animals, destruction of property, deceitfulness or theft, and serious rule violations. The cumulative prevalence of DSM-IV CD by age 16 in a population sample was 9% 16
, with a higher rate among boys than girls (14.1% versus 3.8%). In the National Comorbidity Survey-Replication, the lifetime prevalence of CD among the subsample of adults aged 18–44 years was 9.5%, with a higher rate among males (12%) than females (7.1%). 17
CD was significantly more prevalent among a small sample of heavily alcohol-exposed children when compared to a matched control group. 18
Using a retrospective measure of prenatal exposure to alcohol and tobacco, Disney et al. 19
assessed the relation of PAE to symptoms of DSM-IV CD in a population sample of adolescents from the Minnesota Twin Family Study. PAE was associated with a greater number of CD symptoms when prenatal tobacco exposure (PTE) and parental psychopathology (Substance Use Disorders and Antisocial Personality Disorder) were controlled.
Maternal smoking during pregnancy also has been associated with externalizing behaviors and with CD. Wakschlag et al. 20
reported that smoking more than a half pack/day during pregnancy increased the risk of CD 4.4 times in boys compared to boys with no prenatal tobacco exposure (PTE). Using a dimensional rating of DSM-IV psychiatric symptoms, Fergusson et al. 21
reported a significant association between PTE and CD symptoms after adjusting for confounders such as socioeconomic factors, harsh discipline, child abuse, and parental criminality. Other research indicates that when confounders such as parental psychopathology and PAE are controlled, there is no significant effect of PTE on externalizing behaviors. D’Onofrio et al. 22
compared the rate of conduct problems among tobacco-exposed offspring and their non-exposed siblings from the National Longitudinal Survey of Youth and concluded that the effect of PTE on conduct problems was not significant when genetic and environmental influences on externalizing problems were controlled. In a study of a population-based birth cohort, the Generation R Study, Roza et al. 23
similarly found that after adjustment for confounds such as socioeconomic factors and parental psychopathology, the association between prenatal exposure to maternal and paternal smoking and offspring externalizing behaviors was not significant. Another study examined risk for externalizing disorders among youth at high or low risk for developing alcoholism; prenatal alcohol and tobacco exposure were not significantly related to externalizing behaviors when family history of alcohol dependence was considered. 24
Twin studies provide support for genetic transmission of CD. 25,26
Parental history of alcohol or drug dependence has also been associated with externalizing behavior, possibly through a shared genetic liability for deviant behaviors. 24, 27
Using a twin-family study design, Hicks et al. 28
explored the familial resemblance of four externalizing disorders (CD, adult antisocial behavior, alcohol and drug dependence) among parents and twin-offspring and concluded that a general vulnerability to the four externalizing disorders exists that is highly heritable. Other common risk factors identified include male gender, IQ, race, poverty, home environment, parenting, life events, single parenthood, and parental psychopathology. 29–31
This report will focus on the relation of PAE to a diagnosis of CD in a birth cohort. We hypothesized that PAE would significantly predict CD after adjusting for the effects of other significant risk factors including other prenatal exposures. In contrast to other reports, trimester-specific exposures were ascertained and the outcome was CD diagnosis, rather than a count of CD symptoms, offering a more rigorous test of the hypothesis. Covariates included prenatal exposure to tobacco, marijuana, cocaine, and other illicit drugs, income, race, gender, parenting style, life events, home environment, family history of alcohol problems, and maternal lifetime psychopathology.