ARCHIVED - Early primary school outcomes associated with maternal use of alcohol and tobacco during pregnancy and with exposure to parent alcohol and tobacco use postnatally

2. Literature Review

Exposure to Prenatal and Postnatal Alcohol and Tobacco

2.1 Methodological Considerations

In discussing the associations between maternal substance use and children’s health, it is important to note at the outset several important caveats, summarized eloquently by Richter and Richter (2001), as well as Huizink and Mulder (2006). In this area of research, “gold standard” experimental designs (randomized controlled trials) are precluded because of obvious ethical concerns. Thus, the ability to draw causal conclusions is limited. Most research studies are based on cross-sectional designs with clinical samples – highly exposed children (sometimes termed “cases”) who are selected and compared to non-exposed children. In these types of studies, information on alcohol and smoking during pregnancy is collected retrospectively, up to 12 months or more after birth of the child. Use of a retrospective design increases the possibility of recall bias, wherein the mother is hesitant to admit substance use, or forgets the amount she consumed. A prospective longitudinal design is preferable, since the women are typically recruited during pregnancy and their child is followed for a long period (e.g. from 3 months to 8 years of age in the current BBBF longitudinal study).

A number of confounding factors can mediate the demonstrated associations between prenatal substance exposure and effects on children. These confounds may include, for example, socio-economic status (SES) and other demographic variables such as maternal education; prenatal nutrition, caffeine, drug use and psychological stress; prenatal medical care; and the postnatal environment such as exposure to second-hand smoke, quality of the parent–child interactions, and other familial risk factors. Two studies indicate that the association between prenatal tobacco exposure and effects on children may be influenced more by confounders than the association between prenatal alcohol exposure and effects on children (D’Onofrio et al., 2008; Sen & Swaminathan, 2007). Finally, it should be noted that the methods of assessment, specific outcome measures and level of substance use vary greatly among studies. Results from different studies, therefore, are not always directly comparable (Huizink & Mulder, 2006).

Nevertheless, the preponderance of evidence clearly points to strong associations between prenatal alcohol and tobacco exposure and adverse consequences on children’s physical, cognitive, social/emotional and behavioural development (Richter & Richter, 2001).

Although a relatively large body of literature has examined effects of prenatal exposure to alcohol and tobacco in newborns and infants, there are fewer studies on older children. Thus, in the sections below, we focus on effects on preschool and primary school-aged children.

It is important to assess developmental outcomes in childhood since these outcomes predict health and well-being into adolescence and adulthood (Pihlakoski et al., 2006).

2.2     Alcohol Exposure in Prenatal and Postnatal Periods

In Canada, the rate of mothers who report drinking any alcohol during pregnancy is approximately 10.5%, according to the 2005 Canadian Community Health Survey (CCHS). In this nationally representative survey, women who had given birth in the previous five years were asked about their alcohol use during pregnancy. In the 2005 survey, the rate of alcohol use was slightly lower than in the previous two CCHS surveys: 12.4% of women in 2003 and 12.2% of women in 2000–01 reported drinking alcohol during pregnancy (Public Health Agency of Canada, 2008). In the CCHS, women over the age of 35 or between 15 and 19 years were generally more likely to report alcohol consumption than mothers between the ages of 20 and 34. Regionally, Quebec had the highest rate of maternal drinking during pregnancy in 2005 at 17.7%, and Newfoundland and Labrador had the lowest rate at 4.1%. 

Slightly higher rates of maternal alcohol use during pregnancy were reported in the National Longitudinal Survey of Children and Youth (NLSCY) – in 2002–03, 15.6% of mothers reported consuming alcohol during their pregnancy and in 2000–01, 13.9% did so (Government of Canada, 2007).

A more in-depth analysis of the 2000–01 CCHS data from Alberta indicates that younger mothers (under 20 years) were more likely to binge drink (i.e. drink 5 or more drinks on one occasion) than mothers over 26 years of age. These analyses also indicate that higher-income pregnant women in Alberta were more likely to be drinkers; however, when they did drink, lower-income pregnant women were more likely to binge drink once per month or more.

These data are supported by the 2006–07 Maternity Experiences Survey results, which indicated that approximately 10.5% of women reported drinking during pregnancy. The Maternity Experiences Survey is a project of the Public Health Agency of Canada’s Canadian Perinatal Surveillance System and was conducted by Statistics Canada. The study surveyed women 15 years of age and older who had had a singleton birth in the three months prior to the 2006 Census (Public Health Agency of Canada, 2009).

These figures indicate that a substantial number of Canadian children will continue to be exposed to alcohol in the prenatal stage unless there are dramatic changes in maternal behaviour.

2.2.1 Heavy Drinking During Pregnancy and Fetal Alcohol Spectrum Disorder

Alcohol is established as a significant teratogen, and results in a host of cognitive, social and behavioural deficits such as impairments in general intellectual functioning, language and academic achievement; developmental delays; and problems with learning, memory, adaptive functioning, attention, inhibition, and state regulation (Bailey et al., 2004; Mattson, Riley, Gramling, Delis & Jones, 1998; Streissguth & O’Malley, 2000). The consequences of alcohol use in pregnancy range from subtle problems to the unique cluster of abnormalities that constitutes Fetal Alcohol Syndrome (FAS) – the most severe of the four conditions that comprise Fetal Alcohol Spectrum Disorder (FASD) (Chudley et al., 2005; Jacobson & Jacobson, 2002).

FAS was first described in 1973 by Jones and Smith, and is caused by heavy drinking during pregnancy. A diagnosis of FAS requires evidence of four main features (growth deficiency, facial malformation, central nervous system damage and confirmed (or unconfirmed) prenatal alcohol exposure), although substantial developmental disabilities are also evident in children without facial malformation (Chudley et al., 2005). The term “FASD” refers collectively to a number of disabilities associated with prenatal exposure to alcohol. The three conditions in the spectrum, all permanent and preventable, include FAS, partial-FAS (pFAS), and alcohol-related neurodevelopmental disorder (ARND). The latter two terms are applied to children who have confirmed prenatal alcohol exposure and who exhibit some, but not all, of the FAS features (Chudley et al., 2005). In Canada, it is estimated that 9 out of 1,000 babies each year are born with FASD (Public Health Agency of Canada, 2007).

The majority of studies to date have focused on binge drinking during pregnancy and the associations with FAS among exposed children (Huizink & Mulder, 2006). Substantial evidence indicates that binge-like drinking patterns, in which the fetus is exposed to high blood alcohol concentrations over relatively short periods of time, are particularly harmful for offspring, and place the fetus at the highest risk of FASD (Maier & West, 2001). Binge drinking is often defined as 5 or more drinks on one occasion; one standard drink is equivalent to 0.5 oz. of absolute alcohol (AA) (Streissguth, Barr & Sampson, 1990). For example, Streissguth, Barr and Sampson (1990) reported that children whose mothers reported any binge drinking in the period prior to pregnancy recognition demonstrated poorer academic performance on reading and arithmetic at age 7 than children whose mothers abstained or did not binge drink during pregnancy.

Notwithstanding the profound negative effects of heavy drinking during pregnancy, a growing body of literature has documented adverse effects on children’s functioning at low to moderate levels of prenatal alcohol exposure (e.g. Jacobson & Jacobson, 1994; Sayal, Heron, Golding & Emond, 2007; Sood et al., 2001).

2.2.2 Moderate and Low Levels of Drinking During Pregnancy: Cognitive and Behavioural Outcomes

Recently, the effects of low levels of prenatal alcohol consumption have come under particular scrutiny. Controversial results from the U.K. Millennium Cohort study published online late in 2008 indicated that children born to mothers who drank up to 1 to 2 drinks per week or per occasion during pregnancy were not at increased risk of clinically relevant behavioural difficulties or cognitive deficits compared with children of abstinent mothers. The odds ratios in this study actually indicated lower risks of these problems at age 3 among children of light drinkers, even after controlling for possible confounds including socio-economic factors, current drinking, mother’s mental health, and child–parent relationship (Kelly et al., 2009). This study has received much media attention and prompted several commentaries and debates among researchers and clinicians, some of whom listed numerous methodological limitations of Kelly and colleagues’ study (e.g. Gijsen, Fulga, Garcia-Bourmissen & Koren, 2008; Nathanson, Jayesinghe & Roycroft, 2007; Sayal, 2009).

Indeed, Kelly and colleagues’ (2009) results were surprising in light of a growing body of literature that has documented the adverse effects of low and moderate levels of prenatal alcohol on behaviour, IQ, learning, and other educational outcomes among early school-aged children (Jacobson & Jacobson, 1994; Jacobson, Chiodo, Sokol & Jacobson, 2002; Sayal, Heron, Golding & Emond, 2007; Sood et al., 2001). For example, Sayal and colleagues (2007) reported an increased risk of behavioural and emotional problems (composite score of these problems) among girls whose mothers drank less than 1 drink per week during pregnancy. These effects were observed for parent ratings at age 47 months and 81 months, and were confirmed by later teacher ratings between 7 and 9 years of age. Similarly, in a prospective study of 501 mother-child pairs, Sood et al. (2001) reported that children with any prenatal exposure to alcohol were 3.2 times more likely to have delinquent behaviour scores in the clinical range. Other behavioural outcomes related to prenatal alcohol exposure include psychosocial deficits and problem behaviours, which have been found in FAS children and in children who were prenatally exposed to moderate levels of alcohol. These children were at increased risk of psychiatric disorders (Streissguth, Barr, Kogan & Bookstein, 1996) and were more likely to be rated as hyperactive, disruptive, impulsive or delinquent (Roebuck, Mattson & Riley, 1999).

The level of cognitive deficits among children with low–moderate prenatal alcohol exposure has not been studied extensively. Several studies, however, indicate that moderate alcohol exposure is associated with cognitive deficits in primary school-aged children, including IQ decrements, learning and memory problems and deficits in information-processing speed (Carmichael-Olson et al., 1997; Streissguth, 2007; Streissguth, Barr & Sampson, 1990; Wilford, Leech & Day, 2006). Streissguth, Barr and Sampson (1990), for example, found that moderate alcohol exposure (defined in this study as 2 or more drinks/day) was related to a 6-point decrease in IQ and lower reading and arithmetic achievement test scores at age 7, after adjustment for 15 covariates including prenatal tobacco exposure. In a prospective study of 636 mother-child pairs, Wilford, Leech and Day (2006) reported that moderate alcohol exposure (approximately 1 drink per day) during the first and second trimesters was related to decreases in composite IQ score as well as verbal, abstract/visual, and quantitative subscales at age 10 among African American children.

It appears that there may be dose–response effect of alcohol on child outcomes, wherein the heavier the level of maternal drinking during pregnancy, the greater the magnitude of negative effects on the exposed child (Goldschmidt, Richardson, Stoffer, Geva & Day, 1996; Jacobson and Jacobson, 2002; Jacobson, Jacobson, Sokol, Chiodo & Corobana, 2004; Sood et al., 2001). For example, Jacobson, Jacobson, Sokol, Chiodo and Corobana (2004) reported that each additional ounce of absolute alcohol (AA) per day (approximately 2 standard drinks) during pregnancy was related to a 2.9 point decrease in overall IQ at age 7.

In the section below, we examine the effects on externalizing and internalizing behaviour problems in particular.

2.2.3 Externalizing and Internalizing Behaviour Problems

An emerging literature has begun to document the associations between prenatal alcohol exposure and externalizing behaviour problems in school-aged children. Specifically, researchers have documented higher rates of inattentive, hyperactive, aggressive and antisocial behaviour in alcohol-exposed children compared with children with no exposure to alcohol (Brown et al., 1991; Mattson & Riley, 2000; Nanson & Hiscock, 1990; Sood et al., 2001).

In the aforementioned study by Sood et al. (2001), low levels of prenatal alcohol exposure (i.e. 1 alcoholic drink per week) were significantly associated with higher externalizing (aggressive and delinquent), internalizing (anxious/depressed and withdrawn), and other behaviour problems at 6 to 7 years of age. These results persisted even after careful control for confounding factors, including prenatal tobacco exposure, maternal age, education, marital status, SES and the home environment. Similarly, in a smaller sample of 88 Caucasian children 6 to 13 years old, heavy prenatal alcohol exposure (mothers were known to abuse alcohol, but children did not have diagnosis of FAS) was related to greater externalizing (attention, aggression, delinquency), internalizing (total score) and total behaviour problem scores (Mattson & Riley, 2000). In one retrospective study of children with FAS or fetal alcohol effects (FAE) (now part of the FASD) or children with attention deficit hyperactivity disorder (ADHD) and controls (Nanson and Hiscock, 1990), parents rated both groups of children as being more hyperactive and more inattentive than the children with no FAS and no ADHD.

Few studies have compared parent and teacher rating of externalizing problems. However, Brown et al. (1991), did compare these two groups of informants. Although teacher reports reflected more social competence problems, depression and externalizing behaviours in 5-year-old children whose mothers continued to drink during pregnancy compared with those whose mothers stopped drinking or who never drank, parent reports revealed no such differences (Brown et al., 1991).

Recently, researchers have also turned their attention to the associations between prenatal alcohol exposure and children’s internalizing problems such as depression and anxiety. O’Connor and colleagues have published a series of reports indicating associations between prenatal alcohol and childhood-onset depression (O’Connor & Kasari, 2000; O’Connor & Paley, 2006). For example, O’Connor and Paley (2006) used SEM to investigate the pathways from prenatal alcohol exposure to child depressive symptoms and the mediating effects of maternal and child characteristics, in a small sample of children aged 4 to 5 years. Results indicated that prenatal alcohol exposure was associated with more negative child affect. In turn, mothers of more negative children were less emotionally connected to their children, and those children had higher levels of depressive symptomatology. Interestingly, these results could not be explained by current maternal drinking patterns (O’Connor & Paley, 2006). Similarly, analyses from a large prospective sample of children prenatally exposed to moderate levels of alcohol indicated an association between higher rates of internalizing problems at age 10 and greater prenatal alcohol exposure, after controlling for significant covariates that also predicted problem behaviours (Day & Richardson, 2000).

2.2.4 Postnatal Alcohol Exposure

Both prenatal and postnatal alcohol exposure appear to shape children’s developmental trajectories (O’Connor & Paley, 2006). It is of importance, then, that prenatal alcohol use is strongly correlated with postnatal use (Carmichael-Olson, O’Connor & Fitzgerald, 2001). Such an association raises the possibility that there is some aspect of postnatal drinking that could account for effects attributed to prenatal drinking, and deserves to be addressed.

A large body of literature has examined the adverse effects of children of alcoholics (termed COAs). These studies suggest that children of alcoholics are at higher risk for a variety of emotional, behavioural and other developmental problems (Fitzgerald, Davies & Zucker, 2002).

Despite the well-documented adverse effects of postnatal exposure to alcohol, several researchers have noted that the effects of maternal current drinking do not have much of an effect on the strong association between prenatal alcohol and internalizing behaviour problems (O’Connor & Paley, 2006), aggressive behaviour, or social competence of school-aged children (Brown et al., 1991). It is possible that prenatal alcohol exposure may have effects on behaviour problems and socio-emotional functioning that are independent of current maternal drinking, or that postnatal alcohol use must be at a relatively high level to significantly contribute to adverse effects on children (O’Connor & Paley, 2006).

2.3 Tobacco Exposure in the Prenatal and Postnatal Periods

In the 2005 Canadian Community Health Survey (CCHS), 13.4% of women reported smoking cigarettes during pregnancy, and 14.1% of women reported being exposed to second-hand (environmental) tobacco smoke during their pregnancy (Public Health Agency of Canada, 2008). These rates have decreased since the 2000–01 CCHS, when 17.7% of women reported smoking during pregnancy and 22.4% reported being exposed to second-hand smoke. Similar rates of smoking during pregnancy were observed in the National Longitudinal Survey of Children and Youth (NLSCY); in 2002/03, 15.9% of women reported smoking during pregnancy, and earlier in 2000/01, 18.5% reported this behaviour (Government of Canada, 2007). In the 2005 CCHS survey, younger mothers, and mothers with less than a high school education were more likely to report this behaviour. Regionally, in the 2005 CCHS, British Columbia and Ontario had the lowest rates of maternal smoking during pregnancy (9.7% and 10.3%, respectively); Nunavut and Northwest Territories had the highest rates (59.5% and 32.8%, respectively). Data from the NLSCY indicate that 35% of women who reported smoking during pregnancy smoked 10 or more cigarettes a day.

Unfortunately, it appears that the majority of smokers will continue this behaviour throughout their pregnancy. In the U.S. National Pregnancy and Health Study, approximately two-thirds of women who smoked prior to their pregnancy continued smoking into the last trimester (National Institute on Drug Abuse, 1996).

2.3.1 Level of Prenatal Tobacco Exposure Associated with Adverse Outcomes

Research indicates a dose–response gradient, wherein the adverse effects on children exposed prenatally to tobacco (and its numerous by-products) is dependent on the frequency and quantity of maternal smoking during the gestation period (Richter & Richter, 2001).

The greater the exposure, the more likely the child is to suffer. For example, birth weight decreases in direct proportion to the number of cigarettes smoked (Cornelius & Day, 2007).

The timing of exposure also affects the outcomes in the exposed child, with the most pronounced effects of smoking on birth weight, for example, occurring during the third trimester (Richter & Richter, 2001).

Dose–response relationships have also been documented with other childhood outcomes, including cognitive and behavioural functioning (Huizink & Mulder, 2006; Martin, Dombrowski, Mullis, Wisenbaker & Huttunen, 2006). In a prospective longitudinal study of 676 Finnish children, Martin and colleagues (2006) classified maternal tobacco use during pregnancy as none, light (1–5 cigarettes/day) and heavy (6 or more cigarettes per day). At 12 years of age, children of light smokers exhibited levels of behaviour problems and academic achievement that were intermediate between those reported for non-smokers and for heavy smokers.

2.3.2 Effects on Growth, Cognitive and Behavioural Outcomes

A substantial body of literature has documented the adverse effects of maternal smoking during pregnancy on birth weight and infant growth (Cornelius & Day, 2007; Richter & Richter, 2001). Children born to mothers who smoke are also at risk of health conditions such as cleft palate, decreased lung function and middle ear disease; these effects are independent of the adverse health effects of environmental tobacco smoke (Richter & Richter, 2001).

A smaller literature base is available for effects of prenatal tobacco exposure beyond the neonatal and infant period. The available research does indicate relationships between prenatal tobacco exposure and childhood cognitive and behavioural developmental deficits, such as lower scores in general intellectual functioning, reduced verbal ability, increased activity, inattention and impulsivity, and higher rates of conduct disorder and other behaviour problems (Cornelius & Day, 2007; Huizink & Mulder, 2006; Richter & Richter, 2001).

In terms of cognitive outcomes, in the Ottawa Prenatal Prospective Study for example, tobacco exposure was significantly related to lower cognitive functioning and poorer language development at 2, 3 and 4 years of age (Fried & Watkinson, 1990; Fried, O’Connell & Watkinson, 1992). When those children were 9 to 12 years old, prenatal tobacco exposure was negatively associated with language and reading abilities. Similar results on cognitive functioning were reported by Milberger, Biederman, Faraone, Chen & Jones (1996) and Olds, Henderson and Tatelbaum (1994). In these three studies, associations between prenatal tobacco exposure and cognitive deficits remained significant after adjustment for confounds such as SES, education, marital status and parental IQ. However, none of the studies controlled for ongoing exposure to environmental tobacco smoke. Some researchers assert that associations between prenatal tobacco exposure and cognitive development can be explained by differences in genetics or the home environment, such as postnatal exposure to second-hand smoke; this area is discussed in section 2.3.4 (D’Onofrio et al., 2008; Eskenazi & Trupin, 1995).

Although prenatal exposure to tobacco appears to influence cognitive functioning, a stronger association is apparent with children’s behaviour problems (D’Onofrio et al., 2008). For example, one 10-year longitudinal study reported that mothers who smoked frequently while pregnant were more than four times as likely as less frequent smokers or non-smokers to have sons who developed a conduct disorder, and were more than five times as likely to have daughters who became dependent on drugs (Weissman, Warner, Wickramaratne & Kandel, 1999). It appears that there are clear long-term adverse effects of prenatal tobacco exposure on behaviour, according to results from a New Zealand birth cohort study (Fergusson, Woodward & Horwood 1998). Fergusson and colleagues reported that children exposed, compared with those not exposed to maternal smoking during pregnancy, had higher symptom rates of chronic disease, substance abuse, and depression at 16 to 18 years of age. The effects remained after the authors controlled for socio-economic disadvantage, impaired child-rearing behaviour, and parental and family problems. The bulk of the literature on behavioural outcomes has focused on Attention Deficit Hyperactivity Disorder (ADHD) and other externalizing behaviours. We review some of these studies in the next section. 

2.3.3  Externalizing and Internalizing Behaviour Problems

Smoking during pregnancy has been consistently linked with externalizing problems in childhood, especially in boys (e.g. Ashford, van Lier, Timmermans, Cuijpers & Koot, 2008; Martin, Dombrowski, Mullis, Wisenbaker & Huttenen, 2006; Wakschlag, Pickett, Cook, Benowitz & Leventhal, 2002; Williams et al., 1998). For example, prenatal exposure to tobacco (mother smoked 1 or more cigarette(s)/day) was related to significantly higher parent-rated activity levels at age 5 in a sample of 676 children from the Helsinki Longitudinal Project in Finland (Martin, Dombrowski, Mullis, Wisenbaker & Huttenen, 2006). Among the same sample at age 12, children who were prenatally exposed to tobacco were rated by their teachers as being more distractible and less mature than children who had no prenatal exposure to tobacco. Martin and colleagues controlled for a range of possible confounds, including SES, maternal age and maternal psychological distress but did not control for postnatal or environmental tobacco smoke exposure. In a population-based cohort of 1,452 twin pairs aged between 5 and 16 years from the Greater Manchester Twin Register, maternal prenatal smoking was found to have a statistically significant relationship with both parent and teacher ratings of ADHD, even after control for two sets of potential confounds – genetic factors and familial/environmental influences (Thapar et al., 2003). Linnet et al. (2003) found consistent evidence of independent effects of smoking on a variety of symptoms related to ADHD in 4- to 7-year-old children, after statistical control for factors known to confound the relationships with ADHD (e.g. familial psychopathology). In a sample of 4,879 children from an Australian longitudinal study, Williams et al. (1998) found a dose–response relationship between externalizing behaviour problems and maternal smoking during pregnancy at 5 years of age. Weaker relationships were evident for internalizing behaviour problems. The associations appeared to be independent of a wide range of possible confounds, such as SES, education, social class, marital status and mental health. Williams and colleagues concluded that these results are strongly suggestive of a causal relationship.

Unfortunately, none of the four studies described above appeared to statistically adjust for the effects of postnatal or environmental exposure to tobacco smoke. There remains disagreement in the literature about the importance of confounding factors on the relationship between prenatal tobacco exposure and child behaviour problems. For example, although Williams et al. (1998) concluded that the relationship is causal, Maughan, Taylor, Caspi and Moffitt (2004) asserted that the association between prenatal tobacco and conduct disorder may be better accounted for by confounds, including antisocial behaviour in both parents, depression in mother and family environment. D’Onofrio et al. (2008) agreed, suggesting that environmental and genetic factors account for the associations between prenatal tobacco exposure and externalizing problems.

Compared with the knowledge base for externalizing outcomes, the relationship between prenatal tobacco exposure and internalizing behaviours is less well documented. Results for these studies have been mixed. Weitzman, Gormaker and Sobol (1992) employed a sample of 2,256 children aged 4 to 11 years from the U.S. National Longitudinal Survey of Youth (NLSY). Three groups of children were compared: those whose mothers smoked both during and after pregnancy; those whose mothers smoked only during pregnancy; and those whose mothers smoked only after pregnancy. Weitzman et al. did not include a direct comparison with mothers who did not smoke at all, so results are less clear with respect to the unique influence of prenatal smoking. However, it was clear that children whose mothers smoked both during and after pregnancy had significantly increased levels of depression and anxiety compared with children whose mothers smoked only after or only during pregnancy. This association remained after adjusting for the child’s sex, birth weight, and various demographic and maternal characteristics. More recently, Ashford, van Lier, Timmermans, Cuijpers and Koot (2008) also used a longitudinal sample, and reported that prenatal tobacco exposure was a predictor of internalizing (and externalizing) behaviour problems in 396 children over the period of ages 5 to 18 years.

However, two studies have found that the effect of prenatal smoking on internalizing problems diminished after controlling for potentially confounding variables. For instance, the increased risk of internalizing problems among exposed children was found to disappear after controlling for variables such as socio-demographic factors, maternal anxiety and depression, birth weight, and pre- and perinatal complications (Williams et al., 1998) or after adjusting for socio-economic disadvantage, impaired child-rearing behaviours, and parental and family problems (Fergusson, Woodward & Horwood, 1998).

2.3.4 Postnatal Tobacco Exposure

Many women who smoke cigarettes during pregnancy continue to smoke after the pregnancy (Cornelius & Day, 2007). Children born to mothers who smoked during pregnancy are thus likely to continue to be exposed after the pregnancy. The most often cited consequence of postnatal exposure to environmental tobacco smoke (ETS) is an increased risk of sudden infant death syndrome (SIDS) (Cornelius & Day, 2007). Behavioural and cognitive outcomes, however, are also affected by postnatal exposure to ETS; however, results are mixed. For example, Cornelius, Goldschmidt, DeGenna and Day (2007) reported that environmental tobacco smoke was not a significant predictor of child behaviour at age 6 when prenatal tobacco exposure was considered. Weitzman, Gormaker and Sobol (1992), in contrast, reported a significant relationship between both prenatal and postnatal exposure and children’s behaviour problems, even after controlling for confounds such as child’s age, sex, family structure and income.

2.4. Effects of the Combination of Prenatal alcohol and Tobacco

It is widely acknowledged that alcohol and tobacco use during pregnancy typically occur in combination (Cornelius & Day, 2007; Sen & Swaminathan, 2007). Specifically, research indicates that between 40% and 76% of women who report smoking during the first trimester of their pregnancy report concurrent alcohol use (Cornelius, Taylor, Geva & Day, 1995; Day, Cornelius & Goldschmidt, 1992; Streissguth, Barr & Sampson, 1990). Despite these statistics, few studies have assessed the effects of the interaction of both substances on the exposed child.

In the Ottawa Prenatal Prospective Study, among children aged 3 and 4 years, heavier maternal use of both alcohol and tobacco during pregnancy was related to statistically lower average scores for child comprehension and motor skills compared with groups reporting lighter use of the two substances (Fried, O’Connell & Watkinson, 1992; Fried & Watkinson, 1990). This effect was not evident at ages 5 and 6. Other reports have indicated that prenatal alcohol exposure has greater effects than prenatal tobacco. In analyses from the U.S. NLSY, Sen and Swaminathan (2007) examined the effects of both substances on children’s behaviour problems between 4 and 10 years of age. Results indicated that whereas prenatal alcohol exposure continued to have effects on behaviour problems after controlling for confounds, prenatal smoking largely ceased to have any significant effects after controlling for maternal mental health and background, and postnatal smoking and drinking.

Based on this review of the literature, the present study was designed to examine whether prenatal and postnatal exposure to alcohol and tobacco, separately or in combination via mothers’ reports of drinking and smoking during pregnancy, had any lasting association with a wide range of children’s developmental outcomes over the first four years of primary school. Although other studies have documented the effects of maternal substance use on child health, there are several strengths of the present study:

• First, the prospective longitudinal study design (the BBBF Project) allowed us to determine the effects of both prenatal and postnatal substance exposure, and to assess developmental outcomes in the same children at several time-points.
• Second, the considerable sample size and diversity of participating families may increase the generalizability of results to other disadvantaged children in Canada.
• Third, we assessed effects on a wide range of child outcomes in the cognitive/academic, social/emotional, behavioural and health domains.
• Finally, the separate and combined effects of prenatal exposure to alcohol and tobacco were assessed.