The Association between Prenatal Exposure to Cigarettes and Cortisol Reactivity in 7-Month-Old Infants

Abstract

We examined the association between prenatal exposure to cigarettes and adrenocortical responses to stress in 7-month old infants. Cortisol levels were assessed twice prior to and twice following affect-eliciting procedures in 111 (59 exposed and 52 nonexposed) infants. Cortisol reactivity was defined as the difference between the peak poststressor cortical level and the pretask cortisol level. Higher values indicated higher cortisol reactivity. Exposed infants had higher peak cortisol reactivity than non exposed infants. There were no differences in pretask cortisol levels. Maternal hostility mediated the association between cigarette exposure and peak cortisol reactivity. Furthermore, infant gender moderated this association such that exposed boys had significantly higher peak cortisol reactivity than nonexposed infants or exposed girls. These findings provide additional evidence that prenatal cigarette exposure is associated with dysregulation during infancy and that early adverse, non-social experiences may have relatively long-lasting effects on cortisol reactivity in infants.

Prenatal exposure to cigarettes has been associated with disrupted regulatory processes among exposed offspring. Early cigarette exposure may increase the concentrations of norepinephrine and dopamine in the central catecholaminergic systems of the developing brain (Lichtensteiger et al., 1988). These regions of the brain are involved in regulatory activities and reactivity to stress (Robbins, 1997). During infancy, studies have found evidence of regulatory difficulties among cigarette-exposed infants such as altered auditory responsiveness and habituation, increased physiological arousal and negative affect, and differential physiological responding to sensory challenges as measured by heart rate, heart rate variability, cortisol levels and EEG beginning prenatally and persisting throughout the first year of life (Franco, Grosswasser, Hassid, Lanquart, Scaillet, & Kahn, 1999; Franco, Chabanski, Szliwowski, Dramaix, & Kahn, 2000; Fried, Watkinson, Dillon, & Dulberg, 1987; Schuetze & Zeskind, 2001; Streissguth, Barr, & Martin, 1983; Willoughby et al., 2007; Zeskind & Gingras, 2006). These findings are particularly significant when considering that one of the primary developmental tasks for infants is to cope with sensory challenges from the external environment as well as to adequately perform homeostatic processes in the presence of both endogenous and exogenous sources of stimulation (Porges, 1996). During the first weeks of life, regulation tends to be endogenously based (Porges, 1996). Beyond the neonatal period, the regulatory system consists of the ability of the infant to regulate the latency and intensity of reactions to environmental stimulation (reactivity) as well as use strategies that would modulate these reactions (regulation; Stifter & Braungart, 1995).

Several studies have suggested that the hypothalamic-pituitary adrenal (HPA) axis may be sensitive to the impact of early substance exposure (e.g., Jacobson, Bihun, & Chiodo, 1999; Ramsay, Bendersky, & Lewis, 1996). Cortisol is the primary hormone produced by the adrenal gland in response to activation of the HPA axis in humans and can be measured non-invasively in saliva (Kirschbaum & Hellhammer, 1992). Salivary cortisol levels in infants have been extensively studied as an indicator of individual differences in the psychobiology of the stress response (see Stansbury & Gunnar, 1994).

Cortisol plays an essential role in the individual’s ability to cope with the stressors of daily life such that elevations in cortisol reflect failure to cope with stressful and challenging events. Cortical adaptation to stress can be considered in the context of allostasis and allostatic load (McEwen, 2001). Allostasis is defined as the extent of activity necessary for an infant to maintain a stable state in the presence of stressors - i.e., to adapt (McEwen, 2000). Allostatic load, or “wear and tear” results with repeated adaptations to stressors. Allostatic load can be characterized as repeated or chronic stress, failure to habituate to repeated challenge, failure to respond appropriately to environmental challenges, or failure to terminate a response once the stressor has subsided (McEwen, 2001). Consequently, both hyper and hyporesponsivity of the HPA axis can lead to allostatic load which may have nonoptimal developmental consequences for infants.

An increasing number of studies with infants have used cortisol to measure responses to stressors such as inoculations, laboratory challenges or maternal separations (Donzella, Gunnar, Krueger, & Alwin, 2000;Hertstgaard, Gunnar, Larson, Brodersen & Lehman, 1992; Larson, Gunnar, & Hertsgaard, 1991; Lewis & Ramsay, 1995; Ramsay & Lewis, 2003; Schmidt et al., 1999; Stansbury & Gunnar, 1994). To date, the majority of studies examining adrenocortical stress responses in infants have compared a prestressor cortisol level to a single post-stressor cortisol level in an attempt to measure the cortisol reaction to a given stressor. However, reactivity and regulation are widely believed to be independent dimensions of infant responses to stressors (Eisenberg & Fabes, 1992a, 1992b; Fox, 1994; Kopp, 1989; Rothbart & Bates, 1998; Thompson, 1998). In fact, studies have demonstrated the independent nature of reactivity and regulation in both behavioral responses by infants to challenge (Lewis & Thomas, 1990; Worobey & Lewis, 1989) as well as in cortisol measures of reactivity and regulation (Blair et al., 2006; Ramsay & Lewis, 2003). Thus, one goal of this study is to obtain both reactivity and regulation measures of cortisol in a sample of cigarette exposed infants to examine their adrenocortical responses to environmental challenges.

Recently, several studies have examined the impact of early substance exposure on the HPA axis. Infants who were prenatally exposed to both alcohol and cigarettes had higher basal cortisol levels and a lower cortisol response to a stressor at 2 months of age (Ramsay, Bendersky, & Lewis, 1996). At 13 months of age, prenatal exposure to alcohol was related to increased basal levels while prenatal exposure to cocaine was related to lower basal levels (Jacobson, Bihun, & Chiodo, 1999). However, to date we have little information about the association of prenatal cigarette exposure to cortisol indices of both reactivity and regulation.

Furthermore, the extent to which these effects are due to the direct, teratological impact of prenatal exposure to cigarettes or are mediated by other maternal factors that are associated with maternal smoking is not clear. Previous studies highlight the importance of examining both direct and indirect pathways to the development of regulatory processes among cigarette-exposed infants. In fact, investigators have recently suggested that one trajectory to later developmental problems among cigarette-exposed infants may be through problematic maternal characteristics (Wakschlag & Hans, 2002). Maternal negative affect may be particularly relevant when examining affect among infants prenatally exposed to cigarettes. For example, cigarette smoking has consistently been associated with higher levels of depressive symptomatology and hostility, characterized by negative attitudes and beliefs towards other people and frequent and intense bouts of anger (Miller, Smith, Turner, Guijarro & Hallet, 1996), among the general population (Breslau, Kilbey & Andreski, 1993; Fergusson, Goodwin, & Horwood, 2003; Whiteman, Fowkes, Deary, & Lee, 1997) as well as among pregnant women (Pritchard, 1994; Rodriquez, Bohlin & Lindmark, 2000; Schuetze & Eiden, 2006; Zhu, & Valbo, 2002).

The importance of considering negative affect of pregnant smokers is underscored by the rapidly increasing body of literature that shows a range of nonoptimal developmental outcomes among the infants and children of women with increased levels of depression and hostility. Independent of cigarette smoking, parents who exhibit higher levels of depression are less sensitive caregivers, provide less stimulation to their infants, and show more withdrawn and intrusive interaction styles (Cohn & Campbell, 1992; Field, Hernandez-Reif, & Diego, 2006; Jameson, Gelfand, Kuczar, & Teti, 1997; Pelaez, Field, Pickens, & Hart, 2008). Similarly, maternal anger/hostility has been found to increase negative maternal behavior among substance-using women (Eiden, Chavez, & Leonard, 1999; Hans Bernstein, & Henson, 1999). Thus, one pathway to altered infant affect may be through maternal psychopathology.

Other substance use during pregnancy may also mediate the association between prenatal exposure to cigarettes and infant reactivity and regulation. Maternal cigarette use during pregnancy is likely to be associated with the use of other substances such as alcohol which is known to impact regulatory processes (Gingras & O’Donnell, 1998). Thus, the impact of maternal cigarette use during pregnancy can only be studied by measuring use of other substances in addition to cigarettes. Further, postnatal use of cigarettes and alcohol, in particular, are likely to have significant influences on maternal characteristics and, consequently, on infant regulation.

Evidence is also accumulating that suggests gender may moderate the association between prenatal exposure to cigarettes and developmental outcome. Studies have consistently found that boys are more vulnerable to a range of developmental problems (Gualitieri & Hicks, 1985) including externalizing behavior problems (Campbell, Shaw & Gilliom, 2000). Physiologically, some, but not all (e.g., Harper, Leake, Hodgman, & Hoppenbrowers, 1982), studies show gender-related differences in HR among infants. Among the studies that have found gender-related differences in HR, male infants have lower baseline HRs than female infants during the neonatal period (Nagy, Orvos, Barados, & Molnar, 2000) and at 6 months of age (Richards et al., 1984) but the reverse is found at approximately 1 month of age (Richards et al., 1984). Furthermore, gender differences seem particularly pronounced among at-risk infants (e.g., Hay, 1997; Weinberg, Tronick, Cohn, & Olsen, 1999) including substance-exposed infants (Moe & Slining, 2001). Specifically, there is some consensus in the literature that the male fetus may be more susceptible to teratogenic influences affecting the central nervous system than the female fetus (Flannery & Leiderman, 1994; Hans, 1994; Weinberg, Zimmerburg, & Sonderegger, 1992). For instance, cigarette exposed boys, but not girls, display lower sociability and negative emotionality as infants and were at an increased risk for conduct problems later in childhood (Wakschlag & Hans, 2002). Thus, biological risk may exacerbate gender differences for regulatory processes. Consequently, another goal of the present study was to determine if gender moderated the association between cigarette exposure and physiological regulation.

The purpose of this study was to examine both direct and indirect pathways from prenatal cigarette exposure to cortisol reactivity and regulation. Previous studies have indicated that a developmental shift in adrenocortical functioning occurs between 2 and 6 months of age (Lewis & Ramsay, 1995; Ramsay & Lewis, 1994). Consequently, we assessed cortisol reactivity and regulation among our cigarette-exposed infants at our 7 month assessment. Because of existing evidence suggesting that reactivity and regulation are independent indices of infant responses to stress (e.g., Lewis, 1989), we assessed cortisol levels at four time points in an attempt to obtain separate cortisol measures of reactivity and regulation. Cortisol was assessed twice prior to and twice following affect-eliciting procedures. Two poststressor samples of infant cortisol were obtained to allow us to examine the dampening of the cortisol response following the peak cortisol response (Ramsay & Lewis, 2003). Since the peak cortisol response typically occurs at approximately 20 minutes after a stressor (Dickerson & Kemeny, 2004), the first poststressor cortisol sample was obtained 20 minutes after a procedure designed to elicit negative affect. The second poststressor sample was obtained 20 minutes after the first poststressor sample.

Consistent with McEwen’s (2001) allostatic load concept, we hypothesized that infants prenatally exposed to cigarettes would have higher pretask cortisol, higher cortisol reactivity and slower recovery in response to affect arousing procedures compared to nonexposed infants. We also examined the possibility that infant gender would moderate the association between prenatal cigarette exposure and cortisol measures and that maternal psychopathology or other prenatal substance exposure would mediate the association between prenatal cigarette exposure and measures of cortisol regulation and reactivity.

Method

Participants

Participants consisted of mother-infant dyads recruited into a longitudinal study of cigarette exposure and infant development who had completed a laboratory visit at 7 months infant age. A total of 111 dyads were recruited into either a maternal smoking group (prenatal exposure [PE]; n = 59) or nonsmoking (control; n = 52) group. Of the 111 mother-infant dyads who completed the 2–4 week laboratory visit, 7 declined to participate in the 7-month assessment or did not show up after repeated reschedules (4 exposed, 3 nonexposed), and 13 were unable to be located (4 exposed, 9 nonexposed. Thus, the final sample used in these analyses consisted of 51 mothers who smoked during pregnancy and 40 mothers who reported not smoking or being exposed to secondhand smoke during pregnancy. There were no significant differences between families with complete versus missing data at 7 months on any demographic or substance use variable.

An outreach worker on the project staff recruited all participants after delivery from a local area hospital. Mothers ranged in age from 18 to 42 (M = 28.12, SD = 6. 49). The majority of mothers were African-American (71%), had a high school or below education (78%) and were single (88%). The Hollingshead 2-factor index was used to calculate socioeconomic status (SES: Hollingshead, 1975). The average SES was 2.80 (SD = 1.43). Thus, the sample consisted of predominantly low-income families with single mothers.

Once a family was recruited into the PE group, the closest matching non-smoking family was recruited. The two groups were matched on maternal age, maternal education and infant sex. Infants ranged from 2358 to 4541 grams at birth (M = 3284.69, SD = 544.59). All infants were fullterm (>36 weeks gestational age) with no major medical problems at birth. 54 percent of the infants were male.

Only mothers who reported no illicit drugs (other than occasional marijuana) and no more than moderate amounts of alcohol (Average Daily Alcohol consumption of less than .50 ounces of ethanol or one drink a day) during pregnancy were recruited for the study. Maternal report of no illicit substance use during pregnancy was confirmed via urine toxicologies. Urine screens are routinely conducted on all mothers who receive prenatal care through the hospital’s prenatal clinic and are obtained at birth for the remaining mothers. Urine toxicologies consisted of standard urine screening for drug level or metabolites of cocaine, opiates, benzodiazepines, and tetrahydrocannabinal (THC). Urine was rated positive if the quantity of drug or metabolite was >300 ng/ml. All mothers recruited into this study were negative for cocaine, opiates, and benzodiazepines.

The study received approval from the institutional review boards of the hospital as well as the primary institutions with which the authors are affiliated. In addition, informed written consent was obtained from all recruited participants. To meet Health Insurance Portability and Accountability Act (HIPPA) guidelines, additional written consent to view the medical records of both the mother and the infant was obtained at the time of recruitment. Participants received $30.00 and a $5.00 infant toy at the 2–4 week laboratory visit and $40.00 and a $10.00 infant toy at the 7 month laboratory visit.

Procedure

All mothers were screened after delivery for initial eligibility and matching criteria. Interested and eligible mothers were given detailed information about the study, asked to sign consent forms and were scheduled for their first visit which took place when the infant was 2–4 weeks of age. A second visit was scheduled when the infant was 7 months old. Both visits consisted of a combination of maternal interviews, observations of mother-infant interactions, and physiological and observational assessments of infant arousal and arousal regulation. For the current analyses, the data from the caregiver interview and the physiological assessment of infants at the 7-month assessment were used.

Assessment of Fetal Growth and Risk Status

Three measures of growth at birth were used in this study: birth weight (in grams), birth length (in centimeters) and head circumference (in centimeters). All measurements were taken by obstetrical nurses in the delivery room and recorded in the infant’s medical chart. Research staff recorded this information from the charts after recruiting the mother-infant dyad. Medical chart review at the time of recruitment was also used to complete the Obstetrical Complication Scale (OCS; Littman & Parmelee, 1978), a scale designed to assess the number of perinatal risk factors experienced by these infants. Higher numbers on this scale indicate lower risk status.

Identification of Cigarette Exposure

The Timeline Follow-back Interview (TLFB; Sobell & Sobell, 1995), administered during the first laboratory visit, was used to assess maternal smoking during pregnancy. Participants were provided a calendar and asked to identify events of personal interest (i.e., holidays, birthdays, vacations, etc.) as anchor points to aid recall. This method has been established as a reliable and valid method of obtaining longitudinal data on substance use patterns and has good test-retest reliability and is highly correlated with other intensive self-report measures and with salivary cotinine (Brown, Burges, Sales, Whiteley, Evans & Miller, 1998). The use of the TLFB for these substances resulted in the following variables for each of the substances used for the month prior to pregnancy, each trimester of pregnancy and for the 7 months following pregnancy: number of days cigarettes was smoked and average number of cigarettes smoked per week. In addition, the TLFB was used to obtain data on other substance use prior to pregnancy recognition including total number of joints used (for marijuana), total number of standard drinks consumed and mean standard drinks per drinking day (for alcohol).

Group differences in level of substance use during pregnancy were initially examined separately for each trimester. However, since the pattern of group differences did not vary by trimester, composite variables for total pregnancy use were created for each substance use variable and used in all subsequent analyses. The number of cigarettes smoked per week during pregnancy for smokers ranged from .92 to 152.78 (M = 39.01, SD = 36.67).

The TLFB was also used to obtain data on how many cigarettes were smoked (either by mother or another household member) in the presence of the infant after birth. The average reported number of cigarettes smoked per week in the presence of the infant ranged from 0 to 57.5 (M = 22.06, SD = 81.0). In order to confirm maternal self-reports of the infant’s postnatal exposure to cigarette smoke, salivary cotinine levels were assessed for infants during their postnatal visit. Cotinine, which is a metabolite of nicotine, is a biochemical marker which effectively quantifies exposure to cigarettes (Jarvis, 1989). It can be detected at low concentrations and is specific to tobacco (Feyerabend & Russell, 1990). Salivary concentrations of cotinine have been shown to be equal to those in the blood (Jarvis et al., 1984) and, thus, are an accurate, yet noninvasive, way of measuring ETS exposure. However, since cotinine can only be used as a biochemical marker for recent passive smoke exposure of up to three days prior to the assessment point (Greenberg et al., 1991), it can not be used as a biomarker for fetal exposure to cigarettes or for exposure to ETS over a longer duration. One ml of saliva was collected from each infant by placing one end of a 6″ cotton dental roll in the infant’s mouth while the other end was held by the research technician. The infant was allowed to suck (mouth) the cotton roll for 1–3 minutes. Once the end was saturated, it was removed from the infant’s mouth. That end was then cut and placed into the barrel of a 10cc syringe. The syringe plunger was used to express the saliva from the cotton roll into a small 0.5ml vial. Samples were frozen immediately after collection until assayed using Enzyme-linked immunosorbent assay. Cotinine levels ranged from 0 to 214 ng/ml (M = 34.0, SD = 64.15) for prenatally-exposed infants. Out of 40 infants whose mothers reported no smoking during pregnancy and no prenatal exposure to ETS, three had infants who tested positive for cotinine at 7 months of age indicating some postnatal exposure to secondhand smoke. Because the cotinine levels for these infants were low (<10 ng/ml) and there were no other differences between these infants and the other infants in the control group, these infants were retained in control group for group analyses.

Cortisol

At 7 months of age, infant reactivity and regulation was assessed using behavioral paradigms from the Laboratory Temperament Assessment Battery (LabTAB; Goldsmith & Rothbart, 1999) designed to elicit affect. The paradigms used consisted of an arm restraint paradigm designed to elicit anger/frustration (Goldsmith & Rothbart, 1999; Stifter & Braungart, 1995), a puppet show designed to elicit positive affect and the presentation of four frightening masks designed to elicit fear. Four salivary cortisol samples were collected surrounding the administration of these three paradigms. An initial pretask cortisol sample (pretask 1) was collected when infants first arrived for their 7-month assessment. At this point caregivers were interviewed on the child’s health that day, the child’s emotional state that day, as well as the times the child had eaten that day. Children were then placed in a high chair where a resting state was induced by having the infant watch a 5-minute segment of a neutral “Baby Einstein” video (see Calkins, 1997 for similar procedures for inducing rest). Following the 5-minute video segment, assessments continued with a 3-minute puppet game designed to elicit positive affect (Goldsmith & Rothbart, 1999), an additional three minutes of video, and a second salivary cortisol sample (pretask 2) collection. Immediately after the second saliva collection, an arm restraint paradigm designed to elicit negative affect (Goldsmith & Rothbart, 1999) was conducted, an additional three minutes of video, followed by three minutes of playing time with blocks. Assessment continued with the presentation of four masks each for ten seconds. Children were than taken out of the highchair and placed on a blanket with a basket of toys. Mothers were instructed to allow children to play alone for three minutes, before asking the caregiver to join the child for an additional five minutes of free play. A third saliva sample (poststressor 1) was then obtained which was approximately 20 minutes following the arm restraint procedure. Infants were then measured for growth and participated in a 10-minute mother-infant feeding interaction. The final sample (poststressor 2) was then obtained 20 minutes later (40 minutes following the arm restraint procedure.

The saliva samples were obtained by placing in the infant’s mouth one end of a 6″ cotton dental roll while the other end was held by the research technician. The infant was allowed to suck (mouth) the cotton roll for 1–3 minutes. Once the end was saturated, it was removed from the infant’s mouth, that end was cut and placed into a barrel of a 10cc syringe. The syringe plunger was used to express the saliva from the cotton roll into a small 0.5ml vial. Samples were stored in a freezer provided for that purpose at −80 degrees C until shipped to Salimetrics laboratories for analysis. Samples were packaged in accordance with Centers for Disease Control guidelines for the transport of biological specimens.

All samples were assayed for salivary cortisol using a commercially available high-sensitive enzyme immunoassay (Salimetrics, State College, PA) without modification to the manufacturers recommended protocol. The test uses only 25 ul of saliva, has a lower limit of sensitivity of .007ug/dl, range of sensitivity from .007 to 3.0 ug/dl, and average intra-and inter-assay coefficients of variation (CV) 5% and 10% respectively. In each assay batch, analytical controls representing lower and higher cortisol levels were included. All samples and controls were assayed in duplicate. Duplicates with CV% greater than 7% were repeated on a sample-by-sample basis. Sample values that represented extreme scores (in comparison to reference ranges) were retested using x2 and x4 dilutions. The mean of the duplicate tests was used in the analyses.

Time of Day

Understanding of the circadian rhythm in cortisol production is important when collecting salivary cortisol in infants (Larson, Gunnar, and Hertsgaard, 1991). Beginning at 3 months of age, the adrenocortical system produces most of its daily cortisol between 4 a.m. and 6 a.m., and then declines throughout the rest of the day (Larson et al., 1991). Therefore, an attempt was made to schedule appointments around similar times. 87% (n = 79) of all visits were scheduled at 10 a.m. with the remaining 13% scheduled between 12:30 and 2 p.m. There was no significant correlation of timing of first salivary cortisol sample collection with cortisol levels at any of the four assessment points. Consequently, time of day was not considered further in data analyses.

Behavioral Data

A series of behavioral measures were used to assess reactivity during the puppet show and arm restraint paradigms using the guidelines of the LabTAB Manual (Goldsmith & Rothbart, 1999). These included peak intensity of positive affect, anger, sadness and fear as well as latency to affect expressions for each episode. The intensity of affect expression was scored using Affex (Izard, Dougherty, & Hembree, 1983) adapted from Izard’s Maximally Discriminative Facial Movement Coding System (1979). Higher scores indicated higher arousal or reactivity. Latency scores ranged from 0 to 180, with higher scores indicating longer latencies and lower reactivity. Maternal positive vocalizations during arm restraint were also coded every 5 seconds (a 0 versus 1 variable for the absence or presence of positive vocalizations) and total scores per episode were computed by taking the overall frequency of the presence of positive vocalizations. Two coders blind to all information about the families coded reactivity. Inter-rater reliability was calculated for 12% of the observations and ranged from 89 to 95% across the different episodes (K = .68 and .76). Similarly, inter-rater reliability for latency to affect expressions ranged from 95% to 98% (K = .93 and .94).

Assessment of Maternal Negative Affect

The Brief Symptom Inventory (BSI, Derogatis, 1993) was used as a measure of general psychosocial functioning. The BSI is a brief form of Symptom Checklist 90-R and is a widely used mental health screening measure in a variety of clinical and research settings. It consists of 53 items rated on a 5-point scale ranging from 0 ‘not at all’ to 4 ‘extremely’. The items are grouped into nine scales of Anxiety, Hostility, Somatization, Obsessive-Compulsive, Interpersonal Sensitivity, Depression, Phobic Anxiety, Paranoid Ideation, and Psychoticism. These sub-scales have been reported to have high internal consistency and have been used in a large number of studies, including studies of maternal cocaine use (e.g., Eiden et al., 2002; Singer et al., 2002).

The Buss-Perry Aggression Scale (BPA, Buss & Perry, 1992) was used to measure components of hostility and aggression. This scale consists of 29 items rated on a 5-point scale ranging from 0 “extremely uncharacteristic of me’ to 4 ‘extremely characteristic of me’. The items are divided among four scales: Physical Aggression, Verbal Aggression, Anger and Hostility.

Results

Sample Characteristics

Group differences for the demographic, substance use and obstetric risk status variables for the two exposure groups are presented in Tables 1 and 2. Results of Multivariate Analyses of Variance (MANOVA) indicated a marginal effect of group status on demographics, F(4, 86) = 2.32, p = .06. Univariate analyses indicated that mothers who smoked during pregnancy had completed fewer years of education. Other substance use during pregnancy also significantly differentiated the two exposure groups, F(4, 86) = 10.61, p < .001. Univariate analyses indicated that mothers who smoked cigarettes during pregnancy consumed significantly more alcohol on average per week during pregnancy. There were, however, no significant differences in marijuana use between the two groups. Finally, results of MANOVA indicated no significant effect of group status on obstetrical risk, F(4, 86) = 0.47, p> .10.

Table 1.

Group Differences for Maternal Characteristics and Substance Use

	Non-Smokers		Smokers
	n = 52		n = 59
Variables	M	SD	M	SD
Demographic Characteristics
Age (years)	28.0	5.51	28.21	7.16
Education (years completed)**	12.13	1.28	11.32	1.45
SES (Hollingshead two-factor)	2.92	1.36	2.72	1.34
Parity	3.24	1.53	3.51	1.90
% Single	72%		64%
Substance Use
Prenatal: Average # of cigarettes/week	0	0	39.01	36.67
Prenatal: Average # of standard drinks/week**	.03	.08	3.16	7.7
Prenatal: Average # of joints/week	3.53	15.03	2.42	12.12
Postnatal: Average # of cigarettes/week+	.08	.28	22.08	81.0

⁺p < .10

^*p < .05

^**p < .01

Table 2.

Group Differences for Infant Characteristics

	Nonexposed		Exposed
	n = 52		n = 59
Variables	M	SD	M	SD
Sex (% male)
Fetal Growth
Gestational age (weeks)	39.31	1.28	39.0	1.49
Birth Weight (grams)+	3386.97	545.28	3233.1	557.82
Birth Length (cm)	49.15	4.85	48.33	5.43
Birth Head Circumference (cm)	33.92	1.26	34.02	1.77
Obstetrical Complication Scale	92.88	16.99	88.92	14.90
Postnatal Variables
Age at 7 Month Visit	32.37	3.46	32.32	3.14
Weight at 7 Month Visit (grams)+	9278.81	2642.02	8763.00	1146.85
Length at 7 Month Visit (cm)	68.8	6.35	67.14	7.42
Head circumference-7 Month Visit (cm)	44.47	2.28	44.43	1.81
Cotinine Levels**	4.42	26.27	71.43	227.86

⁺p < .10

^*p < .05

^**p < .01

Data Preparation

Consistent with previous studies (e.g., Gunnar, Mangelsdorf, Larson, & Herstgaard, 1989; Magnano, Gardner & Karmel, 1992; Lewis & Ramsay, 1995; Ramsay & Lewis, 2003), cortisol data were screened for outliers which were defined as values greater than 3 SDs above the mean for each assessment point. There was one infant with pretask 1 values, one with pretask 2, six with poststressor 1, and four with poststressor 2 values that were outliers. These outliers were winsorized following Tukey (1977), replacing values that were 3 SDs above the mean with the value of 3 SD above the mean. There were no values 3 SD or more below the mean. The cortisol data were quite skewed and kurtotic and were transformed (log 10) before further analyses were conducted. There were no differences in any of the results when participants with winsorized values were included or not included in the data set.

As with other physiological data, cortisol values are subject to the Law of Initial Value (Lewis & Ramsay, 1995). Thus, the association between the initial physiological values (i.e., pretask) and subsequent cortisol responses was examined using Pearson. Cortisol values at all time points were positively correlated with each other. There were no negative associations. The correlation between two adjacent samples were quite high ranging from .64 (for the correlation between pretask 1 and poststressor 2 samples) to .82 (for the correlation between pretask 1 and pretask 2), p < .001. The correlation between pretask 1 to poststressor 1 was r = .66, and between pretask 2 and poststressor 1 was r = .79. The correlation between pretask 2 and poststressor 2 was also high, r = .76. Thus, consistent with the Law of Initial Value, peak cortisol reactivity was adjusted for pretask measures. In order to obtain a cortisol response score that was statistically independent from the pretask levels, a residualized peak cortisol reactivity variable was created by regressing the peak cortisol reactivity variable on the pretask values. This residualized reactivity variable was used in all subsequent analyses.

Preliminary Analyses

The initial cortisol sample was used to examine pretask group differences. Although this value may not be a true baseline and may reflect some reaction to stressors associated with preparing to leave the home and the car ride to the laboratory, it does reflect the initial cortisol level before major affective stress. Consistent with previous work (e.g., Ramsay & Lewis, 2003), cortisol reactivity was defined as the difference between the peak poststressor cortisol level (Time 3 or 4) and the pretask cortisol level. Higher values for this measure indicated a greater peak response and higher cortisol reactivity. Previous studies examining cortisol regulation in infants have assessed cortisol regulation by calculating the slope in the cortisol levels from the peak cortisol level through the final post stressor cortisol level. More negative slopes indicate a more rapid dampening of the cortisol response. However, 63% (n = 57) of infants in this sample had their peak response during the last cortisol assessment point. This slope measure could, therefore, not be calculated for most of the participants and was not considered further in these analyses.

Individual Differences in Reactivity and Recovery

Consistent with Granger and Fortunata (under review), infants who showed at least a 10% increase in cortisol levels from pretask 1 to poststressor 1 were identified as reactors. A difference of 10% is two times larger than the rate of error in the assay and two times larger than the lower limit of the assays sensitivity and is, therefore, large enough to be considered a meaningful and reliable change. 25% (n = 28) of the infants in this study (18 exposed and 10 nonexposed) met the criteria to be considered reactors. There was no significant difference in likelihood of being a reactor on the basis of exposure status, X² (1) = .52, p > .10. A multivariate analysis of variance (MANOVA) with behavioral measures of reactivity and regulation as the dependent variables and reactor group status (reactors versus nonreactors) yielded a significant multivariate effect of group status, F(8, 102) = 2.37, p < .05. Univariate analyses followed by simple contrasts indicated that reactors had significantly shorter latency to sadness during both trials of the arm restraint task. Reactors also had a significantly more intense expression of sadness during the first trial and of anger during the second trial of the arm restraint task. A MANOVA with measures of maternal negative affect as the dependent variables also yielded a significant multivariate effect of group status, F(13, 97) = 2.34, p < .05. Univariate analyses followed by simple contrasts indicated that reactors had mothers with significantly more symptoms of depression, as measured by the BSI, and higher levels of anger, as assessed by the BPA anger subscale.

Among the reactors, infants who declined in cortisol levels by more than 10% from poststressor 1 to poststressor 2 were classified as recovering from reactors (n= 10, 35.7%). The remaining infants were classified as being delayed in recovery. Infants in the reactor recovery group were significantly more likely to be exposed infants than to be nonexposed infants, X² = 6.54, p < .05. MANOVA with behavioral measures of reactivity and regulation as the dependent variables yielded a marginal effect of group status (recoverers versus delayed recoverers), F(8, 19) = 2.7,2 p > .10. MANOVA with maternal negative affect (anxiety, depression, anger) as the dependent variables yielded a significant effect of group status, F(3, 24) = 5.12, p < .01. Univariate analyses followed by simple contrasts found that recoverers had mothers with significantly more symptoms of depression and anxiety and with higher levels of anger.

Assessment of Covariates

Correlational analyses were used to determine covariates for all subsequent analyses. Variables that were associated (p < .10) with either cigarette exposure or the outcome (cortisol or behavioral reactivity/regulation) variables were used as covariates. We first examined the association between the substance use and demographic and infant risk variables (see Table 3). Prenatal cigarette and alcohol exposure were both negatively associated with maternal education. Prenatal cigarette exposure was also negatively associated with gestational age and marginally with birthweight. Maternal education, birthweight and gestational age were, therefore, included as covariates in all subsequent analyses.

Table 3.

Associations among Study Variables

	Maternal Education	Maternal Age	SES	Parity	Gestational Age	Birth Weight	BPA	BSI – Depression	BSI - Anxiety	Postnatal Cigarette Exposure
Prenatal Cigarette Exposure	−.31**	.10	−.05	.11	−.23*	−.22+	.23*	.09	.21+	.13
Prenatal Alcohol Exposure	−.36**	.16	−.03	.27*	.01	.17	.19+	−.06	.03	.39***
Prenatal Marijuana Exposure	.03	.11	.29**	−.04	.06	−.21+	−.16	.02	.05	−.08
Pretask Cortisol	−.11	.10	.02	.19	.15	.08	−.01	−.09	−.05	.16
Peak Cortisol Reactivity	−.21+	.18	.19+	.22*	.13	.18	.09	−.04	−.09	−.11
Intensity of Anger – Trial 1	.047	−.11	−.20+	−.057	.23*	−.041	−.20+	−.11	−.16	−.067
Intensity of Sadness – Trial 1	.02	−.083	−.20+	.012	.24*	.017	−.19	−.13	−.123	.028
Latency to Anger – Trial 1	−.07	−.20+	.114	−.16	.18	−.01	.20+	.026	−.16	.055
Latency to Sadness – Trial 1	−.09	−.09	.16	.001	.28*	−.13	.047	−.12	−.04	.20+

⁺p < .10

^*p < .05

^**p < .01

^***p < .001

We next examined the associations between demographic, infant risk and outcome variables (cortisol and behavioral regulation variables). Parity was positively associated with peak cortisol reactivity variable. Maternal education and SES were marginally associated with peak cortisol reactivity. Consequently, maternal education, parity and SES were included as covariates in the corresponding analyses. Gestational age and SES were significantly and marginally, respectively, associated with intensity of anger and with intensity of sadness during the first arm restraint episode. GA was and postnatal cigarette exposure was significantly and marginally, respectively, associated with latency to sadness during the first arm restraint episode. Finally, there were no significant associations between cortisol variables and time of last meal, number of hours of sleep during the previous night, frequency of wakings during the night or administration of medication during the previous 24 hours (Hibel et al., 2006).

Associations between Cigarette Exposure and Cortisol Measures of Infant Regulation and Reactivity

Hierarchical linear regression analyses were conducted to examine the association between prenatal cigarette exposure and measures of infant regulation and reactivity. All covariates were entered as predictors in the first step and the average number of cigarettes smoked per week during pregnancy was entered as a predictor in the second step. Separate regression analyses were first conducted for each criterion cortisol variable. After controlling for all covariates, prenatal cigarette exposure was significantly associated with peak cortisol reactivity, t = 2.87, p < .01, but not with pretask 1, t = 1.54, p = .13. Separate regression analyses were then conducted for each behavioral measure of infant reactivity and regulation. After controlling for the covariates as described above, prenatal exposure to cigarettes was not associated with any of the behavioral indices of infant regulation and reactivity.

Associations between Cortisol and Behavioral Indices of Regulation

There were no associations between cortisol levels and any of the behavioral indices of infant reactivity or regulation during the puppet show. There was a significant association between intensity of anger (r = .25), intensity of sadness (r = −.24) and vocalizations (r = −.24) during the first arm restraint trial and peak cortisol reactivity. There was a marginal association between latency to sadness during the first arm restraint trial and peak reactivity (r = −.22). Results of a linear regression analysis with peak reactivity as the criterion variable and intensity of anger, intensity of sadness, latency to sadness and vocalizations during the first arm restraint trial as the predictors indicated that only intensity of anger was a significant predictor of peak cortisol reactivity, t = 2.09, p < 0.05. There were no associations between behavioral indices of infant reactivity or regulation during the second arm restraint trial and cortisol levels.

Mediational Analyses

The next step was to examine if there was an indirect association between cigarette exposure and levels of cortisol via maternal negative affect or other substance use during pregnancy. Two approaches to examining indirect or Mediational pathways have been discussed in recent years (e.g., MacKinnon et al., 2002). The first is the widely used causal steps approach to mediation that clearly specify that in order to test mediation, the independent (IV), dependent (DV), and mediator variables must all be associated with each other (Baron & Kenny, 1986; Judd & Kenny, 1981). The shortcomings of this method have been discussed, with three primary ones being that they do not provide a statistical test of the indirect effect of an IV on a DV via a third variable; that large sample sizes (n = 500 or more) are required to have adequate power to test Mediational effects with small to medium effect sizes; and that the condition that IV and DV have to be significantly associated with each other excludes many “inconsistent” intervening variable models in which the direct and indirect effects have opposite signs and may cancel each other out (Mackinnon, Krull, & Lockwood, 2000). Given our more moderate sample size, we chose to analyze the role of maternal psychopathology and other prenatal substance exposure using an intervening variable approach discussed by MacKinnon and colleagues (2000; 2002).

For maternal negative affect, only the BPA total score was associated with both amount of cigarette exposure during pregnancy (r = .25, p < .05) and peak cortisol reactivity, (r = .26, p < .05). Although both the Anxiety and Depression subscales of the BSI were positively associated with prenatal cigarette exposure, they were not associated with either of the cortisol variables. Thus, only the BPA score was explored further as a possible mediator of the association between cigarette exposure and cortisol reactivity. Cigarette group status was dummy coded (no smoking versus smoking during pregnancy). The first step in this process was to estimate the association between maternal smoking and scores on the BPA using Linear Regression with the score on the BPA as the criterion variable and maternal smoking as the predictor (see Table 3). These analyses indicated that after including all covariates, maternal smoking was significantly associated with BPA. In the next step, the association between BPA and peak cortisol reactivity was estimated. Hierarchical Linear Regression was used with the peak cortisol reactivity as the criterion variable. The Product of Coefficients Test for the intervening variable effect was used to calculate the significance of the indirect effect (see MacKinnon et al., 2002). The standard error of the intervening variable effect was calculated with the formula suggested by MacKinnon (1994). The significance of the intervening variable effect was tested by dividing the estimate of the intervening variable effect by its standard error, which was then compared to the values of the normal distribution (MacKinnon et al., 2002). The intervening variable effect for BPA when comparing cigarette-exposed infants to nonexposed infants was significant, z = 1.97, p < .05. Thus, infants who were prenatally exposed to cigarettes had mothers with higher levels of hostility, and mothers with higher levels of hostility had infants with higher peak cortisol levels of reactivity.

The same procedure was followed to examine prenatal exposure to other substances (alcohol or marijuana) as potential mediators of the association between prenatal exposure to cigarettes and cortisol indices of infant regulation or reactivity. Prenatal exposure to cigarettes was significantly associated with prenatal alcohol exposure (see Table 4) but not with prenatal marijuana exposure. In the next step, the association between prenatal alcohol exposure and the cortisol variables were estimated. Prenatal exposure to alcohol was significantly associated with both pretask cortisol and with peak cortisol reactivity. The intervening variable effects for prenatal alcohol exposure on the association between prenatal cigarette exposure and pretask cortisol, z = 1.51, and peak cortisol reactivity, z = 1.83, were not significant. Thus, prenatal alcohol exposure did not mediate the associations between prenatal cigarette exposure and either cortisol measure.

Table 4.

Hierarchical Linear Regression Models – Mediational Analyses

Predictor Variables	Unstandardized Coefficients Beta	S.E.	Standardized Beta	R2	F
Maternal Hostility
Regression 1: Outcome – Hostility
Smokers vs. Nonsmokers	.46	.149	.348***	.27	4.35***
Regression 2: Outcome – Peak Reactivity
Smokers vs. Nonsmokers	.115	.045	.26*	.07	6.68*
Prenatal Alcohol Exposure
Regression 3: Outcome – Prenatal Alcohol Exposure
Smokers vs. Nonsmokers	.39	.195	.24*	.20	2.91*
Regression 4: Outcome – Pretask Cortisol
Smokers vs. Nonsmokers	.074	.032	.244*	.06	5.52*
Regression 5: Outcome – Peak Cortisol Reactivity
Smokers vs. Nonsmokers	.19	.04	.461***	.21	22.88***

^*p < 0.05

^**p < 0.01

^***p < 0.001

Moderational Analyses

Next, using hierarchical linear regression, we conducted exploratory analyses to examine if any associations between maternal smoking during pregnancy and cortisol levels were moderated by infant gender. The covariates were entered in the first step, followed by maternal cigarette smoking during pregnancy and infant gender in the second step. The interaction term was entered into the regression equation in the third step. This term was the product of the dummy coded cigarette use variable with the z-transformed gender variable as recommended by Baron and Kenny (1986). Evidence for the moderating effect of gender could be established at this step if the interaction term explained a significant proportion of the variance over and above that accounted for by the main effects of its two contributing variables (Baron & Kenny, 1986). The interaction term was a significant predictor of peak cortisol reactivity indicating that infant gender moderated the association between prenatal exposure to cigarettes and cortisol reactivity. Among exposed infants, boys had significantly higher peak cortisol reactivity (see Figure 1).

Figure 1.

Interaction between infant gender and exposure status for peak cortisol reactivity.

Discussion

A growing body of research has found that infants who were prenatally exposed to cigarettes have a higher level of problems with behavioral regulation. Less is known, however, about underlying physiological regulation processes in this population. Therefore, the primary purpose of this study was to examine the effects of prenatal exposure to cigarettes on cortisol measures of reactivity and regulation at 7 months of life. The results generally support previous findings of an association between prenatal substance exposure and cortisol levels during the first year of life and of higher vulnerability among exposed boys for regulatory processes.

Contrary to our hypothesis, there was no association between prenatal exposure to cigarettes and pretask cortisol levels. This finding is not consistent with other studies that have found higher pretask cortisol levels at 2 (Ramsay et al., 1996) and 13 months of age (Jacobson et al., 1999) and a trend to higher prenatal cortisol levels at 6 months of age (Ramsay et al., 1996) among substance exposed infants. One possible explanation for this discrepancy may be the differences in the substances to which the infants were exposed prenatally. Infants in the Jacobson et al. (1999) study were exposed to alcohol, cocaine and other illicit drugs and findings indicated a different cortisol pattern for alcohol versus cocaine. Similarly, infants in the Ramsay et al. (1996) study were exposed to both alcohol and cigarettes. While the small sample size in the study did not allow analyses of separate effects of the two substances, the authors speculated that the prenatal exposure to alcohol but not cigarettes was responsible for the higher prestressor cortisol levels. As such, future studies are necessary to examine the effects of prenatal exposure to specific substances on adrenocortical functioning.

Results of the present study did find a positive association between prenatal exposure to cigarettes and peak cortisol reactivity indicating elevated stress in response to environmental challenge relative to nonexposed infants. When considered in the context of no group differences in pretask cortisol levels, the group differences for peak cortisol reactivity suggest that cigarette exposed infants may respond differently to stressors than nonexposed infants even when resting cortisol levels are similar. Thus, these results provide additional evidence that exposed infants are more reactive than nonexposed infants and may explain, in part, why cigarette exposed offspring are more likely to display externalizing behavior problems later in childhood (e.g., Wakschlag & Hans, 2002). It is important to note that our finding of a group difference for peak cortisol reactivity is not consistent with previous findings that infants exposed to alcohol and cigarettes did not differ from nonexposed infants in their cortisol response to a stressor at 6 months of age (Ramsay et al., 1996). However, there were several significant differences between the Ramsay et al (1996) study and the present study that may explain this discrepancy. First, Ramsay and colleagues used an inoculation as their environmental stressor. It is possible that the pain response elicited by an inoculation may affect adrenocortical functioning differently than a stressor such as gentle arm restraint which primarily elicits frustration. Second, only one poststressor cortisol sample was collected at approximately 20 minutes after the stressor. Given the considerable individual variation in the timing of the peak cortisol response (Ramsay & Lewis, 2003), the collection of a single poststressor sample obtained by Ramsay and colleagues (1996) may not have captured the peak cortisol reactivity of the infants in their sample. Third, failure to find differences in cortisol reactivity among the exposed infants in the Ramsay et al. (1996) study may reflect a ceiling effect of the higher pretask (baseline) cortisol levels among the exposed infants. Finally, as described above, it was not possible to examine the independent effects of alcohol and cigarettes on cortisol levels in the infants. Given previous findings that alcohol exposure is associated with adrenocortical hyperresponsivity in both the animal and human literature (Haley et al., 2006; Jacobson et al., 1999; Weinberg, Nelson, & Tayler, 1986), alcohol exposure in these infants may have masked any effects of prenatal exposure to cigarettes.

Several behavioral measures or reactivity were associated with peak cortisol reactivity indicating that infants who had increased physiological responses to environmental challenge were more likely to who behavioral reactions as well. In particular, infants with higher physiological reactivity displayed more intense expressions of anger during the arm restraint paradigm. The findings of an association between cortisol and behavioral measures of reactivity have been mixed in previous studies with some studies showing no association (Ramsay & Lewis, 2003) and others finding moderate associations (e.g., Gunnar, 1986; Lewis & Thomas, 1990). Future studies should further explore the relations between behavioral and cortisol indices of reactivity and regulation to determine if the association between intensity of anger and peak cortisol reactivity found in this study is a consistent pattern of reactivity in this high-risk substance exposed population.

Consistent with previous studies (Haley et al., 2006; Jacobson et al., 1999; Weinberg, Nelson, & Taylor, 1986), higher prenatal exposure to alcohol was significantly associated with higher pretask cortisol levels providing additional evidence for an association between alcohol exposure and adrenocortical hyperresponsivity for baseline/pretask cortisol levels. We also found a significant association between prenatal exposure to alcohol and higher peak cortisol reactivity. However, prenatal exposure to alcohol did not mediate the association between prenatal cigarette exposure and peak cortisol reactivity indicating that prenatal exposure to cigarettes had a unique association with peak cortisol reactivity.

The findings of the present study also indicated that the pathway linking prenatal cigarette exposure to peak cortisol reactivity was mediated by maternal hostility. Specifically, exposed infants who had mothers with higher levels of hostility experienced the highest cortisol responses. This provides empirical evidence supporting one specific mechanism by which cigarette exposure may impact physiological regulation during infancy. These findings can be considered within the context of a vulnerability model which indicates that experiencing aspects of nonoptimal caregiving such as adverse maternal psychosocial functioning in the presence of substance exposure increases developmental vulnerability (Luther & Zelazo, 2003).

Finally, the present findings provide additional evidence that boys are more vulnerable than girls to teratogenic effects as early as infancy. Although prenatal exposure to cigarettes was associated with peak cortisol reactivity among both boys and girls, the effects were exacerbated for exposed boys suggesting an enhanced vulnerability to the effects of smoking for boys. This finding is consistent with the general developmental literature which indicates that boys tend to have less optimal behavioral regulation such as higher levels of negative affect and more difficulty regulating their emotions than girls (Carter, Mayes, & Pajer, 1990; Weinberg et al., 1999). Furthermore, recent studies have suggested that regulatory behaviors are particularly vulnerable to prenatal cigarette exposure. For example, smoking during pregnancy has been associated with low sociability and negative emotionality and higher reactivity only among boys during infancy (Wakschlag & Hans, 2002; Willoughby et al., 2007). Others have found that boys, but not girls, were more likely to develop conduct problems and to display antisocial behaviors later in childhood (Fergusson et al., 1998; Weissman, Warner, Wickramaratne & Kandel, 1999). The present findings contribute to this body of knowledge by indicating that gender differences also exist in physiological responsivity to environmental challenge. Longitudinal research is needed, however, to examine whether this gender difference for higher cortisol reactivity persists throughout infancy and into the childhood years. In addition, future studies should explore whether increased adrenocortical reactivity during infancy is one possible pathway to the tendency for boys to display increased externalizing behaviors during childhood.

It is important to note some of the limitations of this study. First, because of the small number of infants who showed adrenocortical reactivity and then consequently showed signs of recovery to the environmental challenges in this study, we were unable to examine regulation/recovery from stress. Future studies are needed to obtain more poststressor samples to get a meaningful cortisol measure of regulation in this population of infants. However, it is important to note that exploratory analyses on the subsample of infants that did recover found some interesting differences in their caregiving environment relative to infants who had not yet recovered by the second poststressor sample. Infants who were classified as recovers were more likely to be exposed infants and had mothers with a more reactive psychosocial pattern as indicated by more symptoms of depression, anxiety and anger. Thus, these infants experienced a combination of both social and biological risk factors enhancing their developmental vulnerability. It is possible that these infants may have learned that their caregiving environment was not a reliable moderator of their stressful experiences and consequently, out of necessity, became more skilled independent regulators to environmental challenges without the need for social intervention. Conversely, infants of mothers with less risky psychosocial profiles may have been delayed in their recovery because they have come to depend on social moderation of their stress responses which was unavailable in this laboratory paradigm. Because of the small number of infants involved and the preliminary nature of these analyses, future studies should more systematically explore the differences in infant characteristics and social context between infants who recover quickly and infants with more delayed recovery patterns to an environmental stressor.

A second limitation concerns the assessment of cigarette exposure in these participants. The accurate assessment of prenatal substance use is always difficulty. Pregnant women are often hesitant to divulge information regarding the use of substances during pregnancy and accurate recall may be difficult. Although we did assess salivary cotinine levels in the infants to confirm postnatal exposure to cigarettes for the infants, no biomarker was collected to confirm maternal reports of cigarette smoking during pregnancy. However, in spite of the possibility that some women in the control group may have misreported their smoking status, it is important to note that there were group differences in cortisol for the infants. Similarly, both measures of psychosocial functioning are based on maternal self-report. Consequently, women may have misrepresented their levels of symptomatology. However, it is noteworthy that there are significant group differences in several aspects of psychosocial functioning and that these domains of psychosocial functioning are associated with observed behavioral and physiological indices of regulation among the infants.

A third limitation is that we did not assess cortisol levels in these infants at home in addition to the cortisol levels obtained during the laboratory visits. Studies have indicated that pretask cortisol levels obtained in laboratory settings are lower than cortisol levels obtained at similar times in the infant’s home (Gunnar et al., 1989). Thus, the increase in cortisol from pretask to the poststressor assessments may reflect a return to baseline rather than a stress response. Future studies should assess cortisol levels obtained from home and laboratory settings at similar times to further explore this possibility.

Despite these limitations, these findings provide additional evidence that early adverse experiences may have relatively long-lasting effects on cortisol reactivity in infants and lend support to the theory of reactivity posited by Boyce and Ellis (2005) which states that increased reactivity to stressors reflects an increased biological sensitivity to context. This heightened biological sensitivity has the potential for negative effects when the child also experiences adverse environmental contexts and positive effects when the child is exposed to highly responsive and supportive environments. The combination of prenatal exposure to cigarettes, increased levels of negative maternal psychosocial functioning and infant gender in the present study creates an adverse context that increases the likelihood of negative health and behavioral outcomes among highly reactive infants.

Recently, the developmental literature has focused on role that the child’s social context has on their adrenocortical responses to stress (e.g., Gunnar & Donzella, 2002). Generally, responsive and sensitive caregiving appears to buffer adrenocortical responses to stressors among infants and young children. Our findings suggest that other, non-social experiences also play an important role in influencing the psychobiology of the stress response in infants. Since many of these non-social experiences such as prenatal exposure to both licit and illicit drugs and other environmental contaminants (e.g., lead) are more likely in nonoptimal social environments such as increased parental psychopathology, lower socioeconomic status backgrounds, lower parental education and employment, and child maltreatment, future studies should explore the role that environmental contaminants play in the association between social influences and adrenocortical functioning.