Clin Invest Med 1996; 19 (4): 231-242
This research was supported by a grant from the Medical Research Council of Canada. Dr. Koren is a Career Scientist of the Ontario Ministry of Health.
(Original manuscript submitted Aug. 11, 1995; received in revised form Dec. 20, 1995; accepted Jan. 10, 1996)
Copyright 1996, Canadian Medical Association
Design: Prospective study of 94 mother-infant pairs.
Setting: Newborn nurseries of two hospitals in Toronto.
Participants: 93 mothers, including active smokers, passive smokers and nonsmokers, and their newborns (including one set of twins).
Interventions: Hair collected from mothers and neonates shortly after birth was analysed by radioimmunoassay.
Main outcome measures: Maternal data on demographic variables, obstetric history, diseases, drugs taken and smoking; infant data on demographic variables and birth indicators (gestational age, method of delivery, weight, head circumference, length, presence of meconium, need for resuscitation, need for special care and congenital malformations); hair concentrations of nicotine and cotinine.
Results: Neonates of active smokers had more adverse outcomes than other infants, including lower birth weight, smaller head circumference, shorter length and more perinatal complications. The difference was statistically significant. Amounts of cotinine extracted from mothers' and infants' hair showed significant differences among active smokers, passive smokers and nonsmokers. When other people in the household and the mother smoked, neonatal hair concentrations of nicotine and cotinine in the infants were threefold higher than those in the infants of mothers who were the only smokers in their households.
Conclusions: Hair concentrations of nicotine and cotinine are a powerful biological marker of the extent of intrauterine exposure to tobacco smoke.
Objectif : Établir le degré d'exposition foetale à la fumée de cigarettes, qui ne peut être extrapolé avec certitude d'après les seules données maternelles. Les auteurs ont donc mesuré la concentration de nicotine et de cotinine dans les cheveux de mères et de leurs nouveau-nés et étudié les corrélations entre ces mesures, les antécédents maternels et les mesures d'évaluation de la grossesse.
Conception : Étude prospective de 94 paires de mères et de nouveau-nés.
Contexte : Pouponnière de deux hôpitaux de Toronto.
Sujets : Quatre-vingt-treize mères avec tabagisme actif ou passif ou sans tabagisme, et leurs nouveau-nés (y compris un couple de jumeaux).
Interventions : Les spécimens de cheveux obtenus des mères et des nouveau-nés peu après la naissance ont été analysés par essai radio-immunologique.
Principale mesure des résultats : Chez les mères, des données démographiques, l'anamnèse obstétricale, les maladies, l'anamnèse médicamenteuse et le tabagisme ont été étudiés. Chez les nouveau-nés, des données démographiques, l'âge de grossesse, le mode d'accouchement, le poids, la circonférence de la tête, la longueur, la présence de méconium, les mesures de réanimation, les interventions spéciales requises et les malformations congénitales ont été analysés. Chez les mères et les nouveau-nés, la concentration en nicotine et cotinine dans les cheveux a été mesurée.
Résultats : Les nouveau-nés de mères avec tabagisme actif avaient plus de résultats défavorables que les autres nouveau-nés, y compris le faible poids à la naissance, une plus petite circonférence de la tête, une longueur corporelle plus courte et un taux de complications périnatales plus élevé. Ces différences étaient significatives sur le plan statistique. Des différences significatives ont aussi été observées quant à la concentration de cotinine extraite des cheveux des mères et des nouveau-nés selon qu'il s'agissait de sujets avec tabagisme actif ou passif ou de sujets non fumeurs. En présence de tabagisme maternel et chez d'autres membres de la famille, la concentration de nicotine et de cotinine dans les cheveux des nouveau-nés était trois fois plus élevée que chez les nouveau-nés dont la mère était la seule à fumer à la maison.
Conclusions : La concentration de nicotine et de cotinine dans les cheveux est un fort marqueur biologique du degré d'exposition intra-utérine à la fumée de cigarettes.
In Canada and the United States, approximately 30% of all women 18 years of age and older are regular cigarette smokers[1,2] and about 60% of women who smoke before pregnancy continue to smoke while pregnant.[3-5] The fetus of an active-smoking mother is involuntarily exposed to tobacco smoke and may be considered a passive smoker.
Cigarette smoking in pregnancy has long been associated with risks to the fetus, including premature delivery,[6] intrauterine growth retardation[7-9] and sudden infant death syndrome.[10-12]
Taylor and Wadsworth[10] found that rates of infant admission to hospital due to respiratory illness and of the incidence of bronchitis among children less than 5 years of age were significantly associated with smoking in pregnancy. Postnatal smoking alone showed no significant effect. Furthermore, assessments of the infants of smokers during the early postpartum period have shown that these infants have poorer pulmonary function than infants born to nonsmokers.[11]
There is some indication that children exposed to cigarette smoke prenatally exhibit poorer performance on cognitive and language-development tests.[12] After controlling for confounding variables, intelligence quotient scores for children born to women who smoked during pregnancy remain four points lower than those for infants born to nonsmoking women.[13] Furthermore, children exposed to tobacco smoke prenatally are consistently reported to display symptoms of attention deficit disorder with hyperactivity and extreme behavioural problems.[12,14,15]
It is estimated that 25% to 30% of nonsmoking women are actually passive smokers; that is, they are exposed to environmental tobacco smoke generated by the active smokers with whom they live or work.[16,17] Environmental tobacco smoke is a combination of smoke emitted by smouldering cigarettes (i.e., sidestream smoke) and exhaled by smokers. In this case, the fetus is considered to be exposed to tertiary tobacco smoke.
Many of the adverse effects observed among infants of active-smoking women have also been described (but to a lesser degree) among infants of passive smokers. Numerous studies have associated passive smoking with a higher incidence of low birth weight and a decrease in mean birth weight, ranging from 24 to 125 g below normal.[17-21]
Cigarette smoke contains more than 3000 toxic or carcinogenic compounds, including hydrogen cyanide, nicotine, carbon monoxide, benzo[a]pyrene, ammonia, catechol and N-nitrosamines.[22] The concentrations of these toxins are actually higher in undiluted sidestream smoke than in smoke inhaled through filtered cigarettes. With respect to fetal pathophysiology, carbon monoxide and nicotine have received the most attention.
Carbon monoxide results from the incomplete combustion of tobacco. Carbon monoxide is readily absorbed from the lungs, producing chronically elevated carboxyhemoglobin levels in smokers, from 5% to 6% in people who smoke a package of cigarettes a day. Carboxyhemoglobin levels are less than 2% in nonsmokers.[23] In neonatal cord blood, carboxyhemoglobin levels are approximately two times greater than in maternal blood,[24] indicating that carbon monoxide has a greater affinity for fetal hemoglobin. Because carbon monoxide decreases the affinity of hemoglobin for oxygen, it impairs oxygen transport and use in fetal tissues, resulting in chronic fetal hypoxia.[25]
The possibility of nicotine-mediated hypoxia in fetal development has also been suggested. Studies of humans and nonhuman primates have shown that nicotine causes a decrease in uteroplacental blood flow and accumulates in fetal circulation.[26,27] There is also evidence that nicotine interferes with the placental cholinergic system, inhibiting active amino-acid uptake and transport by placental villi.[28] It has been postulated that fetal nutritional requirements for these amino acids are compromised, which contributes to intrauterine growth retardation.
Although cigarette smoke contains many toxins, most infants born to active and passive smokers are healthy. To identify which infants are likely to be susceptible to the adverse effects of cigarette smoke, it is essential to first determine the extent of fetal exposure. Today, investigators rely mainly on maternal self-reports to assess gestational tobacco exposure. However, this method is not very reliable or accurate. The metabolic processes that govern the fate of these toxic constituents in the body vary greatly among smokers, as do individual smoking habits. With respect to passive smoking, estimating exposure to environmental tobacco smoke is much more complicated. Additional factors that need to be considered are the number of active smokers around the pregnant woman, the number of hours of exposure, room size, ventilation and the distance from the source of smoke. Therefore, the extent of fetal exposure to tobacco byproducts and subsequent fetal injury depend largely on these multiple factors.
Furthermore, smoking behaviour may vary during pregnancy. Several studies have examined the smoking behaviour of pregnant women. Approximately 25% to 30% of women who smoke quit smoking when planning a pregnancy or when a pregnancy is confirmed; 60% to 65% reduce the number of cigarettes smoked; 10% to 15% do not change their smoking habits; and 2% to 10% increase their smoking during pregnancy.[3-5,29] Hence, a single interview with a pregnant woman may not adequately reflect the variation in smoke exposure during the course of the pregnancy. Among women who continue to smoke during pregnancy there is evidence of deception. Some women may not disclose their smoking habits, and others may underestimate the amount they smoke.[30]
An objective measure of cigarette smoking that accounts for variables other than the claimed number of cigarettes smoked clearly needs to be found. To date, the presence of nicotine in the body has been found to be the most specific measure of exposure to cigarette smoke.[31,32] More specifically, nicotine and cotinine have been found to accumulate in adult hair. Measuring concentrations in hair allows for assessment of long-term cigarette smoke exposure. However, investigations involving this technique to identify intrauterine exposure to cigarette smoke have been sparse.
Another issue is that, in the past, the main focus of studies of the effects of cigarette smoking during pregnancy has been to establish differences between active smokers and nonsmokers. Much less attention has been paid to pregnant women who involuntarily inhale other people's smoke. Recently, however, measurements of cotinine in urine, plasma and saliva have been used to verify passive exposure to tobacco smoke and the consequences to human health.
Based on the foregoing, we identified the following objectives for this study.
[Table of Contents]
Patients and methods
Between September and December 1992, one investigator (C.E.) attended the newborn nursery at the Toronto Hospital-General Division three times per week to enroll the study participants. All subjects were clearly informed about the nature of the study, both verbally and through documentation. After obtaining verbal consent, all mothers were asked to complete a brief questionnaire detailing their smoking habits. At this time, maternal and neonatal hair samples were collected for each enrolled mother-
infant pair. The participants also granted the investigators their consent to review neonatal hospital charts for clinical information.
In addition to the mothers at the Toronto Hospital-General Division, seven mothers from York-
Finch General Hospital (Downsview, Ont.) were included in this study. These mothers had previously been identified as active smokers in another study of hair analysis to assess the prevalence of cocaine use in Toronto.
To accurately measure nicotine exposure from tobacco smoke, we took into account nicotine derived from other sources. Women who reported using nicotine-replacement therapies (for example, nicotine policrilex gum or nicotine patches), smoking cigars or chewing tobacco were excluded from the study. Women who did not speak English were also excluded.
To ensure confidentiality, all records and samples collected for a given subject were assigned a common identification number. The only information on the record forms that could be directly linked to an enrolled subject was the hospital chart number of the infant.
The study protocol was approved by the Research Ethics Committee at the Hospital for Sick Children in Toronto.
As mentioned previously, all mothers participating in the study were administered a brief questionnaire to assess their exposure to tobacco smoke. Data collected included the average number of cigarettes smoked during each trimester of pregnancy and exposure to sidestream smoke.
Subjects were identified as active smokers, passive smokers or nonsmokers based on self-reported exposure to cigarette smoke. Active smokers were defined as women who admitted to smoking cigarettes at any time in their pregnancy. Active smokers were then divided into two groups: women who lived with other smokers and those who were the sole smokers in their household. Passive smokers were defined as nonsmoking women who were exposed, for at least 2 hours per day throughout the pregnancy, to smoke from other people's cigarettes, either at home or in the workplace. Women classified as nonsmokers reported no regular exposure to cigarette smoke during their pregnancy.
Maternal data collected included date of birth, race, marital status, obstetric history, maternal illnesses and drugs being taken.
Data retrieved from the neonatal charts included sex, date of birth, method of delivery, gestational age, birth weight, head circumference, birth length, presence of meconium at delivery, method of resuscitation, special care requirements and congenital malformations. Resuscitation was categorized as spontaneous or supportive. Supportive resuscitation was defined as cases in which newborns required oxygen stimulation, oxygen bagging, oxygen suction, oxygen by mask, endotracheal tube and suction, oropharyngeal suction, bulb suction or tracheal aspiration at delivery. Infants admitted to the neonatal intensive care unit (NICU) were classified as requiring special care.
All of the information collected from the study participants was recorded on a form designed for this study and labelled with the subject's identification number.
A small amount of hair (20 to 25 hair shafts from mothers, 5 to 7 from infants) was cut from the posterior vertex, close to the scalp, of each participant with the use of fine scissors. The posterior vertex is generally chosen for hair sampling because most (approximately 85%) of the hair in this region is in the anagen phase of growth.[33] For the neonates, we obtained the longest hair from the occipital region of the scalp, which may represent the earliest scalp hair.[34] All hair samples were collected 1 to 3 days after delivery. These sampling procedures caused no pain or discomfort to any of the participants. Collected hair samples were put into individual envelopes, labelled with the corresponding identification numbers and stored at room temperature.
The most appropriate hair-washing conditions to remove sources of external contamination without removing systemically deposited drugs have not been established. Various washing solvents, including hexane,[35] ethanol, phosphate buffer,[36] sodium dodecyl sulfate,[37,38] dilute hydrochloric acid[39] and dichloromethane, have been used by different investigators.[40] There has been some debate as to whether some of these solvents are too stringent and may, in fact, remove systemically deposited drugs.[41] Some investigators have chosen not to wash hair at all. Extensive work in this field has shown that most external contamination is removed by washing with shampoo.[36] Therefore, in this study, hair samples were washed with a mild detergent, rinsed with distilled water and dried in a warm (37°C) oven over night.
The following day each hair sample was finely cut into small segments. For the purposes of this study, the hair was not differentially analysed according to length. Instead, the segments were thoroughly mixed together to determine the average nicotine or cotinine content in the sample. Two to 5 mg of each hair sample were weighed on an analytical balance (Mettler AE 100) and placed in a glass container with 1 mL of 0.6 N sodium hydroxide. Parafilm was used to seal the glass vials to prevent evaporation. The samples were agitated over night at 50°C to digest the hair. The following day the samples were neutralized with 50 to 70 µL of concentrated hydrochloric acid, and 100-µL aliquots of the neutral solutions were used to measure nicotine and cotinine concentrations by radioimmunoassay, as described by Langone, Gjika and Van Vunakis.[41]
The radioimmunoassay materials required for determination of nicotine and cotinine concentrations were purchased from the Department of Biochemistry, Brandeis University, Waltham, Mass. The procedures in this study involved adding fixed amounts of tritiated nicotine or tritiated cotinine to each sample, followed by incubation with the respective antiserum, raised in rabbits, for 1 hour at 37°C. A sufficient amount of antibody was used to bind 40% to 50% of the total radioactive ligand. For nicotine and cotinine assays we used the same isogel buffer, consisting of 0.01 mmol/L of trimethamine hydrochloride, 0.14 mmol/L of sodium chloride and 0.1% gelatin with a pH of 7.4. After allowing the reaction to reach equilibrium, a goat antirabbit (gamma) globulin was added to each sample to separate the antibody-bound nicotine or cotinine from the free analyte. Following overnight storage at 4°C, the antibody-bound fractions were extracted from solution by centrifugation at 1000 g for 45 minutes. The temperature of the centrifuge (Beckman Model TJ-6) was also set at 4°C.
The amount of radioactivity in the precipitate was expressed in average counts per minute, with a counting time of 2 minutes per sample, measured with a Beckman LS 5000 CE scintillation counter.
The concentration of nicotine or cotinine in each sample was determined by comparison with a standard curve.
Recovery of analytes was established by adding known amounts of nicotine and cotinine to a hair sample that had previously shown a negative test result for both nicotine and cotinine. After the hair was completely digested, 100-µL aliquots of the solution were analysed, and the recovery of each analyte was calculated. Recovery values of 87% for nicotine and 92% for cotinine were calculated on the basis of six experiments.
Cotinine has a much longer elimination half-life than nicotine. Self-reported smoking habits have been validated by measurements of this metabolite in plasma, urine and saliva.[31,32,35,42,43] On the basis of these facts, we hypothesized that cotinine levels in hair are more sensitive than nicotine levels in distinguishing between different levels of smoke exposure in infants.
To date, no study has assessed cotinine levels in neonatal hair. Previous studies of adult hair have demonstrated that cotinine concentrations in the hair of nonsmokers range from 0 to 0.4 ng/mg of hair, with a mean of 0.1 ng/mg of hair.[44,45] Kintz, Ludes and Mangin[45] have reported that the range of cotinine concentrations in the hair of passive smokers and nonsmokers is the same. We calculated the sample size in this study so that a twofold difference in the hair levels of cotinine between the infants of passive smokers and those of nonsmokers could be detected. On the assumption that the newborns of nonsmokers would have hair levels of cotinine similar to those reported in adults and that the standard deviation would be 25% of the range, a sample size of at least 16 was determined to be sufficient to detect the desired difference, with a power of 80% and a Type I error of 0.05.
A chi-squared analysis was used to compare proportional differences between active, passive and nonsmoking groups. When the expected values in a contingency table were less than five per cell, a Fisher's exact test for 2 x 2 tables was used to analyse the data.
Mean birth measurements and mean hair concentrations of nicotine and cotinine in active, pas-
sive and nonsmoking women and their newborns were compared through an analysis of variance (ANOVA), in the case of normally distributed data, and through the Mann-Whitney U test with corrections for multiple comparisons using the Bonferoni method, in the case of data that were not normally distributed.[46]
Correlations between maternal and neonatal hair concentrations of nicotine and cotinine, and between the number of cigarettes smoked according to self-reports and hair measurements, were analysed with a regression analysis for normally distributed data. Logarithmic transformation of the data or Spearman's rank correlation coefficient was used to compare nonparametric data.
A total of 93 mothers and their newborns were included in this study. Of these subjects, 36 were classified as active smokers, 22 were passive smokers and 35 were nonsmokers. One passive smoker delivered twins, leading to a total of 94 mother-infant pairs. The mean age of these women was 31 years, with a standard deviation (SD) of 6 years and a range from 17 to 43 years. Most (80%) of the study participants were white, 11% were black and 5% were of Asian descent. There was no difference in age or ethnic background among active, passive and nonsmoking women.
In terms of marital status, 18% of the women were single, 76% were married, 4% were in a common-law relationship and 2% were divorced. Most single women (73%) were active smokers.
Of all of the women, 29% were primiparous. The mean gravidity was 3 (SD 2, range of 1 to 10); the mean parity was 2 (SD 1, range 1 to 5); the mean number of spontaneous abortions was 0.5 (SD 1, range 0 to 6); and the mean number of therapeutic abortions was 0.2 (SD 1, range 0 to 3).
Among the 36 active smokers, four (11%) women quit smoking during the first trimester of pregnancy. Twenty women reported that they were the only smokers in their household, and 16 stated that they lived with another smoker, in most cases their spouse. The mean number of cigarettes smoked per day by this group was 11.8 (SD 9, range of 1 to 40), according to self-report. Most active-smoking women (58%) were light smokers, smoking 1 to 10 cigarettes per day. Few women in this group were heavy smokers, and only one woman smoked more than 25 cigarettes per day. The smoking patterns of these women are summarized in Table 1.
Among the 22 women who reported passive exposure to cigarette smoke, 21 said that their primary source of exposure was someone in their home. Only one woman stated that she had significant exposure to cigarette smoke in her workplace.
A total of 55 male and 39 female infants were enrolled in this study. Gestational age ranged from 31 to 43 weeks and birth weights from 2040 to 4460 g. The three groups of infants (by smoking status of their mothers) did not differ with respect to sex, gestational age, head circumference or Apgar scores at 1 and 5 minutes of life. A significantly lower mean birth weight was observed among the infants born to active smokers, even after controlling for gestational age (Table 2). The infants of active smokers also had a significantly shorter mean length when compared with the infants of passive and nonsmoking mothers.
Three low-birth-weight infants (weighing less than 2500 g) were identified in this study population; all of these infants were born to active smokers. Five out of the six infants born prematurely (before 37 weeks of gestation) were born to active smokers and one was born to a passive smoker. A Fisher's exact test did not reveal a signficant difference in the proportion of low-birth-weight infants among active, passive and nonsmoking groups; however, a greater proportion of premature infants was born to active smokers than to nonsmokers (p = 0.03). Table 3 summarizes complications or events noted during delivery. Data regarding the method of delivery, presence of meconium, spontaneous or supportive resuscitation or special care requirements were not explicitly recorded on all source documents for the infants enrolled in this study. Therefore, the "total" column in Table 3 represents only the data available from the hospital patient medical records.
Thirteen of the infants in this study were admitted to the NICU for special care. There were significantly more infants admitted to the NICU from the active-smoking group (10) than from the passive-smoking group (2) and the nonsmoking group (1)
(p <0.01). Among infants born to active smokers, there were two cases of opioid withdrawal and six of respiratory distress.
Five congenital abnormalities were found in this study sample. Among the active-smoking group, the following cases were noted: one sacrococcygeal teratoma, one right-kidney agenesis and one left-ear anomaly. One infant with a genital hydrocele was born to a passive smoker and one infant with hypospadias was born to a nonsmoker.
Preliminary results of our analysis have been published elsewhere.[47] The important final results presented here are those concerning the concentrations of cotinine and nicotine in the hair of mothers who were the sole smokers in their household versus those who lived with other smokers, and their infants. Among the active-smoking mothers, there was no significant difference in hair levels of nicotine or cotinine between those who were the only smokers in their household and those who lived with other household members who smoked. However, the infants born to women who smoked and lived with other smokers had significantly greater levels of cotinine in their hair than the infants of women who were the sole smokers in their household (p = 0.05, Tables 4, 5, 6 and 7). Hair concentrations of cotinine in the infants of women who were the sole smokers in their households did not differ from the levels in the infants of passive smokers.
[Table of Contents]
Discussion
In a preliminary report on this study, we showed that infants of passive smokers had significantly higher concentrations of cotinine in their hair at birth than infants of nonsmokers.[47] This analysis reveals that infants exposed prenatally to maternal and paternal smoking (in most cases, the other smoker in the household was the mother's spouse) had higher levels of cotinine in their hair than infants exposed to maternal smoking alone (Fig. 1). The higher levels of cotinine in the former group are not a result of the mothers in this group smoking more cigarettes than mothers in other groups. Women who were the sole smokers in their households smoked on average the same number of cigarettes as women whose partners also smoked.
Although there was a distinction in the concentrations of cotinine in the hair of the infants of passive smokers versus that of the infants of all active smokers, we did not find any significant difference when infants of the passive-smoking women were compared with the infants of the women who were the sole smokers in their household. Two important factors need to be examined in this situation. First, more than 50% of the smokers in this study were light smokers. Wald and colleagues[48] previously reported that urinary cotinine levels in passive smokers and in light, active smokers overlap. Therefore, consistent exposure to second-hand smoke may not differ much from light smoking. Second, our sample size of women who were the sole smokers in their household may not be large enough to allow the observed difference to reach statistical significance.
Small amounts of nicotine and cotinine were detected in the hair of nonsmokers and their infants. This is not surprising, since there are small amounts of nicotine in common foods such as potatoes, tomatoes, eggplants and tea.[49] Also, cigarette smoke is ubiquitous in the environment because more than 30% of Canadians are regular smokers. Therefore, complete avoidance of cigarette smoke is almost impossible. Despite these uncontrolled sources of exposure, the relatively different hair concentrations of cotinine allowed for a distinction between different groups of infants on the basis of their exposure to cigarette smoke.
Misclassification of active smokers as passive smokers or nonsmokers due to deception by the participants cannot be ruled out. Another possibility is that passive smokers may have been misclassified as nonsmokers, since significant exposure to environmental tobacco smoke was not accounted for in the questionnaire. If this was the case, then the true difference in neonatal hair concentrations may have been diluted and the results of our study are conservative. This implies that hair accumulation of cotinine in passive smokers may be even higher.
Our findings have recently been supported by those of Ostrea and colleagues,50 who analysed 55 meconium samples for cotinine content. Meconium concentrations of cotinine were found to be significantly higher in infants of passive smokers than in those of nonsmokers. However, the investigators observed no difference between cotinine concentrations in the meconium of infants of light, active smokers and of passive smokers. A moderate correlation was found between nicotine metabolites in meconium and smoking exposure by category (i.e., none, passive, light and heavy smoking). These results are similar to our observations concerning hair concentrations.
Among active smokers in our study, there was no correlation between the self-reported number of cigarettes smoked and maternal or neonatal hair concentrations of nicotine or cotinine. However, it is well known that self-reports by pregnant smokers are subject to recall and reporting bias.
Feldman and colleagues[51] showed that women who sought medical advice about drug exposure during pregnancy tended to make a more concerted effort to accurately report how much they smoked. However, 49% of the same women reported that they smoked less when they were subsequently interviewed after delivery.
In a study by Pley and associates,[29] women attending a prenatal clinic recorded the number of cigarettes smoked per day in a diary. The same pregnant women were interviewed about smoking at the first clinic visit. A subsequent review of the diary entries indicated that these women smoked significantly more cigarettes than they reported during the interview. These studies underscore the importance of using objective measures to determine exposure to cigarette smoke.
The nature of our study was clearly explained to the participants, and the women were approached in a nonthreatening manner; however, the women were interviewed only once by a person to whom they had not been previously introduced. The degree to which this situation may have played a role in the disclosure of smoking is unknown. In future, it may be better for the attending physician or nurse to recruit participants for studies of this nature. On the other hand, it is recognized that smoking is commonly and systematically underreported to groups that perceive affirmation of smoking as undesirable (such as physicians and nurses).[52] Since most pregnant women are advised of the potential harmful effects of cigarette smoking by their physician, an interview with an unbiased person who is not involved with the patient's primary care may be more likely to elicit an accurate response.
Most of the women in our study who were active smokers were light to moderate smokers, yet there was a great degree of variation in the amount of nicotine and cotinine extracted from the hair of these women and their infants. This variation may be attributed to the fact that the number of cigarettes smoked is only one determinant of systemic exposure to these compounds in the body. Individual differences in tobacco-smoke uptake (i.e., the way a cigarette is smoked) and metabolism are additional factors to be considered. Indeed, the sum of these individual factors determines total systemic maternal and fetal exposure to cigarette smoke, as evinced by the excellent correlation between maternal and neonatal hair levels of cotinine in our study.
We did not section the hair in these analyses to examine differences in nicotine and cotinine distribution over time, nor did we take into account the length of the hair. For some women with long hair, nicotine and cotinine in the hair shaft may have represented tobacco exposure before as well as during pregnancy. This possibility would introduce some error in interpreting the relation of maternal self-reports to hair analysis.
The contribution of nicotine deposited onto hair in the environment may be substantial. The total nicotine content in hair has been shown to be directly proportional to both the air concentration of nicotine and the duration of exposure.[53] This fact may account for the large nicotine-to-cotinine ratio (12:1) observed in the maternal hair samples. External contamination is less of a problem in analysis of hair for cotinine content because this compound is derived from hepatic metabolism. Hence, its route of entry into hair is through the systemic circulation. Furthermore, hair obtained from neonates shortly after birth is very unlikely to be externally contaminated by environmental tobacco smoke because smoking is prohibited in hospitals. This hypothesis is supported by the smaller nicotine-to-cotinine ratio observed in the neonatal hair samples than in the maternal samples in our study.
There is no consensus about which washing procedures should be employed to remove externally deposited nicotine. We chose to wash the hair with detergent rather than solvents, in order to prevent extraction of systemically deposited drugs.
With respect to pregnancy outcome, a dose-response relation between neonatal hair levels of nicotine or cotinine and birth measurements was not found. But this is not surprising, since no correlation between the reported exposure to tobacco smoke and birth measurements was found either. By contrast, our results are concordant with previous studies showing a higher incidence of premature delivery, decreased birth weight and greater special care requirements for infants of smokers than for those of nonsmokers.
Approximately 15% of the women who were invited to participate in this study refused because they wanted to avoid the inconvenience of the interview or objected to the cutting of their newborn's hair. It is possible that there may be some selection bias as a result. Perhaps women who were passive smokers and who delivered small infants were less likely to participate owing to feelings of guilt about or concern for their newborn. However, this explanation is highly suspect, given the fact that there was adequate representation of active smokers with small infants in this sample.
In summary, our study documents the usefulness of measuring nicotine and cotinine concentrations in maternal and neonatal hair to understand variation in adverse neonatal outcomes. Larger series of active smokers, passive smokers and nonsmokers need to be evaluated to establish the dose- and time-
response of various perinatal end points.
[Table of Contents]
References
The full text may also be ordered from the Canada Institute for Scientific and Technical Information (CISTI) or Institute for Scientific Information (ISI).