Children's Health Articles

Fetal Growth and Length of Gestation in Relation to Prenatal Exposure to

Environmental Tobacco Smoke Assessed by Hair Nicotine Concentration
Jouni J.K. Jaakkola/-^ Niina Jaakkola,^ and Kolbj0rn Zahlsen^
^Environmental Health Program, The Nordic School of Public Health, Goteborg, Sweden; ^Environmental Epidemiology Unit,
Department of Public Health, University of Helsinki, Helsinki, Finland; ^Department of Clinical Pharmacology, Regional and
University Hospital of Trondheim, Trondheim, Norway

source population that included all the 2,751

We assessed the efFects of prenatal exposure to environmental tobacco smoke on fetal growth and children born from 1 May 1996 to 30 April
length of gestation. The study population consisted of 389 nonsmoking women who were 1997 in two geographically defined hospital
selected from a population-based study in southeast Finland on the basis of questionnaire infor- districts in southeast Finland (Kymi and
mation after delivery (response rate 94%). The final exposure assessment was based on nicotine Porvoo Hospital Districts). All mothers were
concentration of maternal hair sampled after the delivery, which measures exposure during the asked after delivery to complete a self-admin-
past 2 months (i.e., the third trimester). The exposure categories were defined a priori as high istered questionnaire (response rate 94.2%).
(nicotine concentration > 4.00 (lg/g; n = 52), medium (0.75 to < 4.00 pg/g; n = 186), and low as The respondents, 2,591 mothers, had 2,568
the reference category (< 0.75 Jig/g; n = 151). In logistic regression analysis, controlling for con- singletons and 23 twin pairs. A total of 1,621
founding, the risk of preterm delivery (< 37 weeks) was higher in the high [adjusted odds ratio (62.3%) women were nonsmokers. They
(OR) = 6.12; 95% confidence interval (Cl), 1.31-28.7] and medium exposure categories were used as a framework for selecting 472
(adjusted OR = 1.30; 95% Cl, 0.30-5.58) compared with the reference category, and there was a nonsmoking women for the present study:
1.22 (95% Cl, 1.07-1.39) increase in adjusted OR with a 1 jig/g increase in hair nicotine concen- 189 women reported exposure to ETS either
tration. The corresponding adjusted OR was 1.06 (95% Cl, 0.96-1.17) for low birth weight and at home or outside the home, and 283
1.04 (95% Cl, 0.92-1.19) for small-for-gestational-age. Key a/ordi; biomarkers, birth weight, hair women reported no exposure to environmen-
nicotine, preterm delivery, small for gestational age, tobacco smoke pollution. Environ Health tal tobacco smoke. The latter group was
Perspect 109:557-561 (2001). [Online 21 May 2001] selected randomly from the unexposed with a
http://ehpnetl.niehs.mh.gov/docs/2001/W9p557-561jaakkola/abstract.html one-to-one balance according to the spouse's
current smoking, so that in l4l women the
spouse was a current smoker, and in 142
The adverse efFects of maternal smoking dur- environmental conditions such as air change, women the spouse was a never smoker or quit
ing pregnancy on fetal growth (1,2), as well as area, and volume of the space. Few studies smoking more than 12 months ago. Out of
on the risk of preterm delivery (3) are well have used biomarkers of exposure such as the 472 hair samples, we excluded 43 samples
established. People exposed to environmental serum cotinine (14,28,30) and saliva cotinine with weight less than 2 mg, 38 samples with a
tobacco smoke (ETS) encounter mainly the (29) measured during pregnancy or after deliv- concentration below detection limit (0.05
same compounds as in the mainstream smoke ery. Both of these biomarkers share a common Hg/g), and 2 samples from a twin pregnancy.
inhaled directly by the smoker, although the feature of having a short half-life (4), and The present study focused on 389 singleton
concentrations and time patterns differ (2,4). therefore they measure exposure only during a newborns of women whose hair sample pro-
Although individual studies have often been couple of days before sampling. Hair nicotine vided a hair nicotine measurement above
inconclusive (5-31), a recent meta-analysis by is a new promising biomarker; the method is detection limit. The characteristics of the
Windham and colleagues (32) suggests that noninvasive, and a sample of 0—2 cm from the study population are presented in Table 1.
exposure to ETS during pregnancy has a proximal hair gives a good estimate of the Data collection. The Environment and
small effect on birth weight and the risk of exposure during the past 2 months (34,35). Pregnancy questionnaire administered at the
term low birth weight. The effects on the Taking into account the time frame, a hair birth clinic inquired about maternal health in
length of gestation and risk of preterm deliv- sample at birth would describe exposure dur- general and during pregnancy; parents' educa-
ery have been studied less (10,19,23,25), and ing the last trimester, which is considered the tion, profession, and behavioral factors such
the evidence of any effect is weak. most important period for fetal growth. In a as smoking and exercise; and details of the
Assessment of exposure during a relevant recent case-control study by Nafstad and col- home and other environments. We requested
time period and control of confounding are leagues (36), the risk of small-for-gestational- records for the women in this study from the
the most critical issues of validity in studies of age births was related to the hair nicotine maternity health clinics in the two areas and
the effects of ETS on pregnancy outcomes concentration of newborns and of their non- received 97% of the records. Additional infor-
(32,33). In the first reports, exposure assess- smoking mothers after the delivery. We fur- mation on the child's birth weight, gestational
ment was based only on information on the ther studied the relation between exposure to age, and maternal smoking habits during
spouse's smoking (5-7,9,11-13,15-17,21, ETS measured as maternal hair nicotine con-
22,24,26). Later studies collected informa- centration after delivery and the risk of being
tion on multiple sources, such as other family small for gestational age. With the present Address correspondence to J.J.K. Jaakkola,
study design, we were also able to study the Environmental Health Program, The Nordic School
members and work exposure or on quantity of Public Health, P.O. Box 12133, SE-402 42
measured as daily duration of exposure or effects on birth weight and the risks of low
birth weight and preterm delivery. We also Goteborg, Sweden. Telephone: 4-46-31-693980. Fax:
number of cigarettes smoked indoors .f46-31-691777. E-mail: jouni.jaakkola@nlw.se.
(10,18-20,25,31). Questionnaire or inter- assessed the relations between hair nicotine We thank the physicians and nurses in the Kymi
view information on sources of exposure is concentration and reported exposure to ETS. and Porvoo Hospital Districts for their help and
sensitive to information bias, especially if car- support that made this study possible. This study
Methods was supported by the Ministry of Social Aflfairs and
ried out after the delivery and if the accuracy
Health, Finland.
of the smoking information is compromised Study population. The Finnish Prenatal Received 15 August 2000; accepted 7 December
due to inaccurate recall and variation of Environment and Health Study focused on a 2000.

Environmental Health Perspectives • VOLUME 1091 NUMBER 61 June 2001 557

Children's Health • Jaakkola et al.

pregnancy were obtained from the Finnish trimester, quit during the first trimester, no different categories of exposure defined by
Medical Birth Registry established in 1987 information). the hair nicotine concentration and ques-
and run by the Ministry of Social Affairs and Covariates. We used the following poten- tionnaire information. The relations
Health. The study protocol was approved by tial confounders as covariates in the analyses: between birth weight and the determinants
the Ethical Committees of the Department of sex, birth order, maternal age, body mass of interest were estimated in linear regres-
Public Health, University of Helsinki, three index before pregnancy, marital status, a sion analysis controlling for potential con-
participating hospitals, and the Ministry of combined index of maternal and paternal founding. Eor the three dichotomous
Social AfJairs and Health. education as an indicator of socioeconomic outcomes, odds ratio (OR) was the measure
Health outcomes. The primary health status (low: neither parent with vocational of effect. We used logistic regression analysis
outcomes were fetal growth and preterm education; medium: either one or both par- to estimate adjusted ORs. We adjusted for
delivery. We used three different measures of ents with vocational school as highest educa- all the potential determinants of the out-
fetal growth: birth weight in grams, low birth tion; high: either one or both parents with comes listed above.
weight (< 3,000 g), and small for gestational college or university education as a reference
age (SGA). A higher than traditional cut category), alcohol consumption during preg- Results
point for low birth weight, 3000 g, was cho- nancy, and employment during pregnancy. Characteristics of the study population. The
sen to increase the number of cases from 7 to Statistical methods. We estimated the characteristics of the study population were
AG. Small for gestational age was defined as prevalences (percent) of the reproductive compared according to the three exposure
birth weight in the lowest 10th percentile outcomes with 95% confidence intervals categories, high, medium and low (refer-
according to the week of gestation, which (GIs) based on the binomial distribution ence), that were defined by the hair nicotine
was calculated from the present source popu- and the mean of birth weight with confi- concentration (Table 1). The three groups
lation of nonsmoking women. Preterm deliv- dence intervals based on the t distribution. were similar in most of the factors that were
ery was defined as length of gestation less Eirst, we compared the mean birth weight possible determinants of the pregnancy out-
than 37 weeks. We retrieved information on and the risks of low birth weight, small for comes. Alcohol consumption during preg-
gestational age from maternity health clinic gestational age, and preterm delivery in the nancy was slightly more common among the
records. Gestational age was practically
always verified by ultrasound examination Table 1. Characteristics of the study population, 389 mothers and newborns, the Finnish Prenatal
during the 18th week of gestation. Environment and Health Study, 1996-1997.
Exposure assessment. The determinant of Exposure category (pg/g)
interest was prenatal exposure to ETS. Low Medium High
Exposure assessment was based on the nico- <0.75 0.75 to < 4.00 >4.00 Total
tine concentration of hair taken after delivery Characteristic No % No. % No. % No. %
(34). We also used maternity clinic medical Total 151 186 52 389
records, the questionnaire at the birth clinic, Cpv
OCA
and birth registry data as additional sources Male 86 57.0 96 51.6 30 57.7 212 54.5
of information. Female 65 43.0 90 48.4 22 42.3 177 45.5
Maternal age (years)
Hair nicotine concentration. The proxi- <25 28 18.5 47 25.3 9 17.3 84 21.6
mal 0-2 cm segment of maternal hair was 26-30 56 37.1 72 38.7 23 44.2 151 38.8
analyzed for nicotine using gas chromatogra- 31-35 50 33.1 48 25.8 14 26.9 112 28.8
phy/mass spectrometry (34). The detection >36 17 11.3 19 10.2 6 11.6 42 10.8
Maternal body mass index
limit was 0.05 fig nicotine/g hair when the <20 21 22
14.1 12.0 7 13.5 50 13.0
sample was at least 15 mg. Nicotine concen- 20 to < 25 90 60.4 97 52.7 32 61.5 219 56.9
tration in hair samples < 1.5 mg was not 25 to < 30 28 18.8 49 26.6 5 9.6 .82 21.3
reported due to uncertainty in the estimated >30 10 6.7 16 8.7 8 15.4 34 8.8
concentrations. Parity
Nulliparous 54 35.8 78 41.9 19 36.5 151 38.8
Maternity clinic medical records. 1 56 37.1 65 35.0 19 36.5 140 36.0
According to nationwide instructions, during >2 41 27.1 43 23.1 14 27.0 98 25.2
the first prenatal visit health care personnel Marital status
are advised to record information about the Married or cohabiting 146 96.7 181 97.3 50 98.0 377 97.2
smoking habits of the mother and her spouse Single 5 3.3 5 2.7 1 2.0 11 2.8
Combined parental education^
on a standard form. Low 22 15.0 36 19.8 12 24.0 70 18.5
Questionnaire information. We requested Medium 75 51.0 105 57.7 26 52.0 206 54.3
information about smoking habits of the par- High 50 34.0 41 22.5 12 24.0 103 27.2
ents and other family members in general and Alcohol consumption during pregnancy

during pregnancy. We also asked about the No 108 72.0 127 68.7 33 64.7 268 69.4
Yes 42 28.0 58 31.3 18 35.3 118 30.6
number of cigarettes smoked daily indoors by Worked during pregnancy
the spouse and other family members and the No 55 36.9 63 34.2 17 3i3.3 234 64.8
amount of time exposed to tobacco smoke Yes 94 63.1 121 65.8 34 66.7 155 35.2
outside home. The quantitative information Maternity leave (weeks)

was missing for a large proportion of subjects <34 97 64.2 107 57.5 30 57.7 234 60.2
34to<36 45 29.8 64 34.4 19 36.5 138 32.9
and was not used in the main analyses. >36 9 6.0 15 8.1 3 5.8 27 6.9
Birth registry information. The birth
Number of subjects with missing information: maternal body mass index, 4; marital status, 1; parental education, 10; alco-
registry information provided categorical
hol consumption, 3: and work during pregnancy, 5. Low exposure is the reference category.
information on the mother's smoking dur- •Low: neither parent with vocational education; medium: either one or both parents with vocational school as highest
ing pregnancy (no, yes, smoked after the first education; high: either one or both parents with college or university education.

558 VOLUME 1091 NUMBER 6 1 June 2001 • Environmental H e a l t h Perspectives

 Children's Health « Fetal growth, preterm delivery, and passive smoking

high exposure group (35.3%) than in the a higher cut point, 3,000 g, which maintains The risk of low birth weight was related
reference group (28.0%). the original idea of indicating a possible fetal to maternal hair nicotine with an exposure-
Exposure to environmental tobacco growth disturbance. The cut point of 2,500 g response pattern, although the lower 95%
smoke. The mean and median hair nicotine was defined during a time when the average confidence limits were < 1.00 (Table 4). The
concentrations were related to the reported birth weight was substantially lower due to adjusted OR contrasting the medium expo-
exposure to environmental tobacco smoke maternal nutrition and other factors. There sure to the reference category was 1.28 (95%
both at home and at work, as shown in were 46 children (11.8%) who were consid- Cl, 0.59-2.60), the high exposure 1.55
Table 2. In women who reported no expo- ered low birth weight children by this crite- (95% Cl, 0.55-4.43), and 1.06 (95% Cl,
sure either at home or work, the hair nico- rion. A total of 35 newborns were considered 0.96-1.17) for a unit increase in hair nico-
tine concentration was substantially higher if small for gestational age (9.1%), and there tine. A corresponding pattern was found
the spouse was a current smoker (medians: were 16 preterm deliveries (4.1%). when using questionnaire information as a
1.32 vs. 0.61 jig/g). We also asked about the Table 3 shows that the mean birth measure of exposure. The risk was related to
number of cigarettes per day smoked indoors weight was lower and the prevalences of work exposure and combined home and
during each trimester. This seemed to be a adverse pregnancy outcomes were in general work exposure, which corresponded well
difficult question, because only 29% of higher the higher the exposure, defined with the hair nicotine concentration distrib-
women reporting exposure estimated quan- either by the hair nicotine concentration or ution in the questionnaire information-
tity. Among those reporting quantity, the questionnaire information on exposure at based exposure categories.
mean concentration (± SD) was higher the home and work. The relations between exposure and small
higher the number of cigarettes reported In linear regression analysis adjusting for for gestational age were in general weaker,
(1-9 cigarettes/day, 2.68 ± 1.99 pg/g; 10-19 confounding, the mean birth weight was although indicating an exposure-response
cigarettes/day, 3.14 ± 2.40 pg/g; > 20 ciga- slightly but not significantly (17g, 95% Cl, pattern.
rettes/day, 5.17 ± 7.24 pg/g). ~178-145) lower in the high exposure cate- The risk of preterm delivery was related
Exposure to environmental tobacco smoke gory than in the reference category. We also to exposure assessed both by hair nicotine
and pregnancy outcomes. The mean birth fitted continuous nicotine concentration in concentration and questionnaire informa-
weight (± SD) in the study population was the model, but there was no clear association tion. We estimated an adjusted OR of 1.22
3,555 ± 512 g, which corresponds well with ("0.9 g birth weight per microgram per gram (95% Cl, 1.07-1.39) for a unit increase in
the average birth weight of Finnish children nicotine; 95% Cl, "20-18). Birth weight hair nicotine concentration. The risk of
(3,550 ± 582 g) (37). Only seven of the new- was not significantly related to either preterm delivery was increased in the pres-
borns (1.8%) were low birth weight children, reported exposure at home (yes vs. no expo- ence of work exposure alone (2.35; 95% Cl,
according to the traditional definition of sure: -99 g, 95% Cl, "273-75) or work (yes 0.50-11.1) and especially with both work
below 2,500 g (Table 3). We decided to apply vs. no exposure: "101 g, 95% Cl, "258-56). and home exposure (8.89; 9 5 % Cl,
1.05-75.3), but not with home exposure
Table 2. Maternal hair nicotine concentration after delivery by reported exposure to environmental only (0.65; 95% Cl, 0.06-6.81).
tobacco smoke during pregnancy, the Finnish Prenatal Environment and Health Study, 1996-1997.
Hair nicotine concentration (ua/q) Discussion
No. Mean±SD 25% Median 75% The present results are consistent with
Total 389 100.0 1.96 ±2.54 0.46 1.07 2.56 adverse effects of exposure to ETS on fetal
Home exposure^ growth. The results also suggest an increase
No 319 82.0 1.67 ±2.01 0.45 0.86 2.22 in the risk of preterm delivery. The strength
Yes 70 18.0 3.30 ±4.27 1.27 2.03 4.22
Work exposure^
of the present study lies in the use of mater-
No 299 78.1 1.79 ±2.56 0.45 0.92 2.26 nal hair nicotine concentration as a measure
Yes 84 21.9 2.57 ±2.84 0.61 1.85 3.31 of exposure during the third trimester and
Combined reported exposure use of objectively measured health outcomes.
No. spouse nonsmoker 128 32.9 0.87 ±1.02 0.25 0.61 1.00 Validity of results. The use of hair nico-
No. spouse smoker 126 32.4 2.27 ±2.46 0.63 1.32 3.14
Home yes, work no 51 13.1 2.93±4.23 0.90 1.89 3.83 tine concentration provided an objective
Home no, work yes 65 16.7 2.07 ±2.01 0.57 1.51 2.71 exposure assessment with comparable infor-
Home yes, work yes 19 4.9 4.29 ±4.34 ^^58 2^89 5.10 mation for newborns with and without nor-
"Any exposure during pregnancy. mal pregnancy. This approach minimized

Table 3. Mean birth weight and the prevalence of low birth weight (LBW), small for gestational age (SGA), and preterm delivery (< 37 weeks) according to mater-
nal hair nicotine concentration (pg/g) and reported exposure to ETS during pregnancy, the Finnish Prenatal Environment and Health Study, 1996-1997.
op
Total Birth weight (g) LBW (< 2,500 g) LBW (< 3,000 g) SG Preterm"
Exposure category No. % (mean ± SD) No. % No. % No. % No. %
Total 389 100.0 3,555 ±512 7 1.8 46 11.8 35 9.1 16 4.1
Hair nicotine concentration Ipg/g)
<0.75 151 38.8 3.559 ± 472 1 0.7 15 9.9 12 8.0 4 2.7
0.75 to < 4.00 186 47.8 3,554 ± 534 5 2.7 23 12.4 18 9.7 7 3.8
>4.00 52 13.4 3.547 ± 547 1 1.9 8 15.4 5 9.8 5 9.6
Reported exposure
No, spouse nonsmoker 128 32.9 3,590 ± 464 1 0.8 12 9.4 10 7.8 5 3.9
No. spouse smoker 126 32.4 3,587 ± 502 2 1.6 16 12.7 11 0.7 3 2.4
Home yes. work no 51 13.1 3,488 ± 458 1 2.0 5 9.8 5 9.8 2 3.9
Home no. work yes 65 16.7 3,459 ± 538 2 3.1 9 13.9 6 9.4 4 6.2
Home yes, work yes 19 4.9 3.616 ±833 1 5.3 4 21.1 3 17.7 2 10.5

Three newborns with gestational age < 35 weeks were excluded.

Environmental Health Perspectives • VOLUME 1091 NUMBER 61 June 2001 559

Children's Health • Jaakkola et al.

the problems of random and systematic error The proportion of women employed during meta-analysis of their own and 22 previous
in the exposure assessment, which could be pregnancy was similar among exposed and studies from 1966-1995 and reported a
introduced in a retrospective questionnaire. unexposed subjects, and we adjusted for pooled effect estimate of 25 g (95% Cl,
However, the nicotine levels corresponded employment in all the models, including 16-41). Combining all eight studies that
well to the reported exposure information those assessing the relations between hair examined low birth weight gave a pooled
given that several uncontrolled intervening nicotine concentration and pregnancy out- OR of 1.00 (95% Cl, 0.90-1.10). In the
factors, such as air change and volume in comes. present study, the adjusted OR for low birth
environmental spaces, could influence the The total eligible study population con- weight was 1.53 when contrasting high
relation between source strength and expo- sisted of 1,621 nonsmoking women, but the exposure to the reference, but the 95% Cl
sure. We used a priori criteria for defining cost of hair nicotine analysis was the limiting was wide (0.55-4.43). The corresponding
exposure categories on the basis ofa previous economic factor. To optimize the power and estimate from 11 studies of small-for-gesta-
Norwegian case-control study (36). All of the cost of the study, we selected the study tional-age births or term low birth weight
the outcomes were based on objective mea- subjects on the basis of questionnaire infor- was 1.19 (1.08-1.32), which is similar to the
surements that were made independently mation on exposure to ETS and spouse's results of the present study.
from the exposure assessment. current smoking. Only a few previous studies have used a
The high participation rate in the study, Synthesis with previous knowledge. There biomarker in the assessment of exposure (14,
94%, minimized the influence of any are several suggested mechanisms for the 28,30,36). In a study of 1,231 women,
hypothesized selection bias. The hair nicotine efFects of tobacco smoke on fetal growth (38). Haddow et al. (14) found an effect of 104 g
concentration was not available to some Most of these studies have been carried out (95% Cl, 35-173) in exposed nonsmoking
mothers due to a too-small hair sample, but on active smokers. Because sidestream smoke women defined by serum cotinine from 1 to
there is no reason to believe that the availabil- contains the same compounds as mainstream 10 ng/mL compared with unexposed (< 0.5
ity would be related to either exposure or smoke, it is likely that the mechanisms are ng/mL). Their effect estimate was not
studied outcomes. similar both in active and passive smoking. adjusted for gestational age, and therefore it
Use of medical records and questionnaire Inhaled carbon monoxide and nicotine incorporates the effects of reducing gestation
information allowed us to take into account increase fetal carboxyhemoglobin and reduce length, as well as reducing intrauterine
an extensive number of potential con- placental blood flow, resulting in low fetal tis- growth rate. Eskenazi et al. (28) reported a
founders. The exposed and unexposed were sue oxygenation. Structural changes in the small reduction of 45 g (95% Cl, "36-126)
relatively similar with respect to measured placenta of smoking mothers have been related to exposure assessed by serum coti-
determinants of the outcomes. The effect of observed as additional evidence of a plausible nine (2-10 ng/mL vs. <2 ng/mL) around 27
work exposure on preterm delivery was biologic mechanism (39). Smoking during weeks of gestation. There was no effect on
stronger compared with the effect of home pregnancy increases metabolites of tobacco in the risk of preterm delivery, and the eflFect on
exposure, although the concentrations of hair the fetus that may have toxic effects (40). the risk of low birth weight was weak (1.35;
nicotine due to reported home and work Maternal smoking has also been attributed to 95% Cl, 0.60-3.30). Peacock et al. (30)
exposure were similar. This could be fetal zinc deficiency (41). Zinc is considered reported a 0.2% (95% Cl, "2.4-2.8) reduc-
explained by confounding by unknown an essential trace element for many aspects of tion in birth weight in newborns of women
working conditions. Another explanation is growth and development (42). whose serum cotinine during pregnancy was
that pregnant women can influence the The present results of a small effect of in the upper quintile compared with new-
home exposure more than work exposure, prenatal exposure to ETS on fetal growth are borns of women in the lowest quintile. In
and thus they could better avoid exposures consistent with the previous studies. the case-control study by Nafstad et al. (36),
perceived disturbing at home than at work. Windham and colleagues (32) carried out a the risk of small-for-gestational-age births

Table 4. Crude and adjusted odds ratios for low birth weight (LBW), small for gestational age (SGA), and preterm delivery according to maternal hair nicotine con-
centration (pg/g) and reported exposure to environmental tobacco smoke during pregnancy, the Finnish Prenatal Environment and Health Study, 1996-1997.
LBW (< 3000 g) S(3A» Preterm (< 37 weeks)*
Exposure category Crude OR Adjusted OR* Crude OR Adjusted OR* Crude OR Adjusted OR*
Nicotine concentration (pg/g)
< 0.75 (reference category) 1.00 1.00 1.00 1.00 1.00 1.00
0.75 to < 4.00 1.28 1.28 1.24 1.05 1.47 1.30
(0.64-2.55) (0.59-2.60) (0.58-2.66) (0.44-2.49)' (0.41-5.01) (0.30-5.58)
' > 4.00 1.65 1.55 1.25 1.18 3.91 6.12
(0.66-4.15) (0.55-4.43) (0.42-3.74) (0.34-4.19) (1.01-15.2) (1.31-28.7)
per pg/g 1.06 1.06 1.04 1.04 1.12 1.22
(0.97-1.17) (0.96-1.17) (0.92-1.17) (0.92-1.19) (1.01-1.25) (1.07-1.39)
Reported exposure
No exposure, spouse nonsmoker (reference category) 1.00 1.00 1.00 1.00 1.00 1.00
No exposure, spouse smoker 1.52 1.92 1.11 1.41 0.59 1.00
(0.68-3.42) (0.79-4.70) (0.45-2.72) (0.52-3.82) (0.14-2.52) (0.20-5.07)
Home yes, work no 1.18 1.13 1.31 1.06 1.02 0.65
(0.39-3.58) (0.34-3.78) (0.42-4.04) (0.30-3.73) (0.19-5.45) (0.06-6.81)
Home no, work yes 1.67 1.43 1.19 1.02 1.58 2.35
(0.65-4.25) (0.50-4.12) (0.41-3.43) (0.31-3.31) (0.41-6.07) (0.50-11.1)
Home yes, work yes 2.76 2.08 2.46 1.47 2.82 8.89
(0.78-9.79) (0.44-9.73) (0.61-10.0) (0.23-9.32) (0.51-15.7) (1.05-75.3)
Values shown in parentheses are 95% Cl.
Three newborns with gestational age < 35 weeks were excluded. 'Logistic regression analysis: adjusted for sex, birth order, maternal age, body mass index before pregnancy, marital
status, index of socioeconomic status, alcohol consumption during pregnancy, and employment during pregnancy.

560 VOLUME 1091 NUMBER 61 June 2001 • Environmental Health Perspectives

 Children's Health « Fetal growth, preterm delivery, and passive smoking

was related to maternal hair nicotine concen- onset of delivery. Am J Obstet Gynecol 179:1051-1055 in Shanghai. Am J Public Health 83:207-210 (1993).
tration both in nonsmoking and smoking (19981. 25. Fortier I, Marcoux S, Brisson J. Passive smoking during
4. Guerin MR, Jenkins RA, Tomkins BA. Chemistry of pregnancy and the risk of delivering a small-for-gesta-
mothers. The adjusted OR for the medium Environmental Tobacco Smoke: Composition and tional-ago infant. Am J Epidemiol 139:294-301 (1994).
exposure category (0.75-4.00 [ig/g ) was 3.4 Measurement. Boca Raton, FLLewis Publishers, 1992. 26. Martinez FD, Wright AL, Taussig LM, and the Group
(95% CI, 1.3-8.6) and 2.1 (95% CI, 5. MacMahon B, Alpert M, Salber EJ. Infant weight and Health Medical Associates. The effect of paternal smok-
parental smoking habits. Am J Epidemiol 82:247-261 (1966). ing on the birthweight of newborns whose mothers did
0.4-10.1) for the high exposure category
6. Comstock GW, Lundin FE. Paternal smoking and perina- not smoke. Am J Public Health 84:1489-1491 (1994).
(> 4.00 |ig/g) compared with the reference tal mortality. Am J Obstet Gynecol 98:706-718 (1967). 27. Roquer JM, Figueras J, Botet F, Jimenez R. Influence on
category (<0.75 ng/g). 7. Underwood PB, Kesler KF, O'Lane J M , Callagen DA. fetal growth of exposure to tobacco smoke during preg-
Parental smoking empirically related to pregnancy out- nancy. Acta Paediatr Scand 84:118-121 (1995).
The possible effect of prenatal exposure to come. Dbstet Gynecol 29:1-6 (19671. 26. Eskenazi B, Prehn AW, Christiansson RE. Pasive and active
ETS on the length of gestation and the risk of 8. Borlee I, Bouckaert A, Lechat MF, Misson CB. Smoking maternal smoking as measured by serum cotinine: the
preterm delivery has received less attention, parents during and before pregnancy: weight, length and effect on birthweight. Am J Public Health 85:396-398 (1995).
and the results of the previous studies are head circumference of progeny. Eur J Obstet Gynecol 29. Rebaglatio M, Florey C du V, Bolumar F. Exposure to
ReprodBiol 6:171-177 (1978). environmental tobacco smoke in nonsmoking pregnant
inconsistent. Martin and Bracken (10) con- 9. Magnus P, Berg K, Bjerkedal T, Nance WE. Parental women in relation to birth weight. Am J Epidemiol
ducted a prospective cohort study of 3,891 determinants of birthweight. Clin Gen 26:397-405 (1984). 142:531-537 (1995).
women in New Haven, Connecticut, and 10. Martin TR, Bracken MB. Association of low birth weight 30. Peacock JL, Cook GC, Carey IM, Jarvis MJ, Bryant AE,
with passive smoke exposure in pregnancy. Am J Anderson HR, Bland JM. Maternal cotinine level during
found no effect of passive smoking during Epidemiol 124:633-642(1986). pregnancy and birthweight for gestational age. Int J
pregnancy on gestational age. In a cohort 11. Rubin HD, Krasilnikoff PA, Leventhal JM, Berget A. Effect Epidemiol 27:647-656 (1998).
study of 4,687 Swedish pregnant women by of passive smoking on birth-weight. Lancet 2:415-417 31. Sadler L, Belanger K, Saftlas A, Leaderer B, Hellenbrand
Ahlborg and Bodin (19), the adjusted risk (1986). K, McSharry J-E, Bracken MB. Environmental tobacco
12. MacArthur C, Knox EG. Passive smoking and birthweight. smoke exposure and small-for-gestational-age birth. Am
ratio for preterm delivery was 0.49 (95% CI, Lancet 2:37-38 (1987). JEpidemiol 150:695-705 (1999).
0.23-1.06) in women exposed in the home 13. Schwartz-Bickenbach D, Schulte-Hobein B, Abt S, Plum 32. Windham GC, Eaton A, Hopkins B. Evidence for an asso-
only and 1.27 (95% CI, 0.70-2.31) in the C, Nau H. Smoking and passive smoking during preg- ciation between environmental tobacco smoke exposure
nancy and early infancy: effects on birthweight, lactation and birthweight: a meta-analysis and new data. Pediatr
workplace. In a study of 4,644 Canadian period, and cotinine concentrations in mother's milk and Perinat Epidemiol 13:35-57 (1999).
women by Fortier and colleagues (25), the infant's urine. Toxicol Lett 35:73-81 (1987). 33. Jaakkola MS, Jaakkola JJK. Assessment of exposure to
adjusted OR in women exposed in the home 14. Haddow JE, Knight GJ, Palomaki GE, McCarthy JE. environmental tobacco smoke. Eur Respir J 10:2384-2397
only was 0.93 (95% CI, 0.64-1.31) and 0.92 Second-trimester serum cotinine levels in nonsmokers in (1997).
relation to birth weight. Am J Obstet Gynecol 159:481-484 34. Zahlsen K, Nilsen OG. Nicotine in hair of smokers and
(95% CI, 0.58-1.51) in those exposed in the (1988). non-smokers: sampling procedure and gas chromato-
workplace only. None of the previous studies 15. Cambell MJ, Lewry J, Wailoo M. Further evidence for the graphic/mass spectrometric analysis. Pharmacol Toxicol
using biomarkers reported the relation effects of passive smoking on neonates. Postgrad Med J 75:143-149(1994).
64:663-665(1988). 35. Nafstad P, Jaakkola JJK, Hagen JA, Zahlsen K, Magnus
between exposure and risk of preterm deliv- 16. Brooke OG, Anderson HR, Bland J M , Peacock JL, P. Hair nicotine concentrations in mothers and children
ery. The present study shows for the first time Stewart CM. Effects on birthweight of smoking, alcohol, in relation to parental smoking. J Expo Anal Environ
an exposure-related increase in the risk of caffeine, socioeconomic factors, and psychosocial Epidemiol 7:235-239 (1997).
stress. Br Med J 298:795-801 (1989). 36. Nafstad P, Fugelseth D, Qvigstad E, Zahlsen K, Magnus
preterm delivery related to exposure to ETS.
17. Chen Y, Pederson LL, Lefcoe NM. Passive smoking and P, Lindemann R. Nicotine concentration in the hair of
Concluding remarks. The present results low birthweight. Lancet 2:54-55 (1989). nonsmoking mothers and size of offspring. Am J Public
18. Lazzaroni F, Bonassi S, Manniello E, Morcaldi L, Repetto Health 88:120-124 (1998).
based on objective assessment of exposure E, Ruocco A, Calvi A, Cotellessa G. Effect of passive 37. Merilainen J, Kojo-Austinen H, Tuominon K, Hemminki E,
during a period most relevant for fetal smoking during pregnancy on selected perinatal para- Teperi J. Finnish Perinatal Statistics 1987 |in Finnishl.
growth provide further evidence of the meters. Int J Epidemiol 19:960-966(1990). Laakintohallituksen julkaisuja 158. Helsinki:Finnish National
19. Ahlborg G Jr, Bodin L. Tobacco smoke exposure and Board of Health, 1990.
adverse effects of tobacco smoke exposure. pregnancy outcome among working women. A prospec- 38. Misra DP, Nguyen RHN. Environmental tobacco smoke
The results suggest that prenatal exposure tive study at prenatal care centers in Orebro County, and low birth weight: a hazard in the workplace? Environ
may also increase the risk of preterm delivery. Sweden. Am J Epidemiol 133:336-347 (t991). Health Perspect 107(suppl 6):897-904 (1999).
20. Ogawa H, Tominaga S, Hori K, Noguchi K, Kanou I, 39. Asmussen I. Ultrastructure of the human placenta at
Matsubara M. Passive smoking by pregnant women and term. Observations of placentas from newborn children
REFERENCES AND NOTES fetal growth. J Epidemiol Community Health 45:164-168 of smoking and non-smoking mothers. Acta Obstet
(1991). Gynecol Scand 56:119-126 (1977).
1. Kramer MS. Determinants of low birth weight: method- 21. Saito R. The smoking habits of pregnant women and 40. Meberg A, Sande H, Foss OP, Stenwig JT. Smoking during
ological assessment and meta-analysis. Bull WHO their husbands, and the effect on their infants. Nippon pregnancy—effects on the fetus and on thiocyanate lev-
65.663-737 0987). Koshu EiseiZasshi 38:124-131 (t99t). els in mother and baby. Acta Paediatr Scand 68:547-552
2. California Environmental Protection Agency, Office of 22. Mathai M, Skinner A, Lawton K, Weindling AM. Maternal (1979).
Environmental Health Hazard Assessment. Health Effects smoking, urinary cotinine levels and birth-weight. Aust 41. Kuhnert B, Kuhnert PM, Debanne S, Williams TG. The
of Exposure to Environmental Tobacco Smoke: Final NZ J Obstet Gynaecol 30:33-36 (1990). relationship between cadmium, zinc, and birth weight in
Report. Sacramento. CA:California Environmental 23. Mathai M, Vijayasri R, Babu S, Jeyaseelan L. Passive pregnant women who smoke. Am J Obstet Gynecol
Protection Agency. 1997. maternal smoking and birthweight in a South Indian pop- 157:1247-1251 (1987).
3. Kyrklund-Blomberg NB. Cnattingius S. Preterm birth and ulation. Br J Obstet Gynaecol 99:342-343 (1992). 42. Apgar J. Zinc and reproduction. Ann Rev Nutr 5:43-48
maternal smoking: risks related to gestational age and 24. Zhang J, Ratcliffe JM. Paternal smoking and birthweight (1985).

Visit EHP CaiDTcidir (Q)pp(Q)[r..v^uuu. Online

• updated weekly
• most ads include e-mail links for instant response

Just point and click lournalxom

Environmental Health Perspectives • VOLUME 1091 NUMBER 61 June 2001 561