Exposure to Tobacco Smoke in Utero and the Risk of Stillbirth and Death in the First Year of Life

Abstract

The authors examined the association between exposure to tobacco smoke in utero and the risk of stillbirth and infant death in a cohort of 25,102 singleton children of pregnant women scheduled to deliver at Aarhus University Hospital, Aarhus, Denmark, from September 1989 to August 1996. Exposure to tobacco smoke in utero was associated with an increased risk of stillbirth (odds ratio = 2.0, 95% confidence interval: 1.4, 2.9), and infant mortality was almost doubled in children born to women who had smoked during pregnancy compared with children of nonsmokers (odds ratio = 1.8, 95% confidence interval: 1.3, 2.6). Among children of women who stopped smoking during the first trimester, stillbirth and infant mortality was comparable with that in children of women who had been nonsmokers from the beginning of pregnancy. Conclusions were not changed after adjustment in a logistic regression model for the sex of the child; parity; or maternal age, height, weight, marital status, years of education, occupational status, and alcohol and caffeine intake during pregnancy. Approximately 25% of all stillbirths and 20% of all infant deaths in a population with 30% pregnant smokers could be avoided if all pregnant smokers stopped smoking by the sixteenth week of gestation.

fetal death, infant mortality, pregnancy, smoking, tobacco

Although a smaller proportion of children die today than at any other time within the last 100 years, stillbirth and infant death are still a major concern, and the search for preventable causes is of the utmost importance. Exposure to tobacco smoke in utero is one of the possible preventable factors with potential influence on stillbirth and infant death. Smoking reduces fetal oxygenation through increased blood levels of carboxyhemoglobin and through impairment of oxygen unloading (1). Nicotine also causes vasoconstriction (2), which, along with the reduced prostacyclin synthesis (3), increases vascular resistance and decreases fetal blood flow (4, 5). Exposure to smoking in utero may thus cause damage to the developing organs, making them vulnerable to disease, or it may interfere with the immune system, predisposing the child to infectious or other diseases.

Results of previous studies on smoking in pregnancy and infant death have been inconsistent (6–26), and it is possible that an apparent association between smoking and mortality can be explained by differences between smokers and nonsmokers in social class and in other lifestyle factors. Only if smoking is causally related to mortality is it possible to reduce the number of infant deaths by interventions directed at reducing the number of pregnant smokers.

The aim of this study was to evaluate the association between exposure to tobacco smoke in utero and the risk of stillbirth and increased mortality in the first year of life while taking into account a number of potential confounders and effect modifiers.

MATERIALS AND METHODS

All pregnant women scheduled for delivery at the Department of Obstetrics and Gynaecology, Aarhus University Hospital, from September 1989 to August 1996 were invited to participate in the study. The women were asked to answer three questionnaires--the first two before the routine antenatal visit at 16 weeks of gestation and the third before the visit at 30 weeks of gestation.

Information from the first questionnaire was used to establish the women's medical record. It provided information on medical and obstetric history, maternal age, smoking habits before pregnancy and during the first trimester, and alcohol intake during pregnancy. The second questionnaire provided information on marital status, education, occupational status, and caffeine intake during pregnancy. Information from the third questionnaire was used only to record smoking habits during pregnancy. Information about delivery was obtained from birth registration forms filled out by the attending midwife immediately after delivery. Before data entry, all birth registration forms were manually checked and compared with the medical records by a research midwife.

Information about stillbirths was obtained from the data we collected at our department and from the Danish Medical Birth Register through record linkage by using the mother's personal identification number. Information about deaths during the first year of life, including causes of death, was obtained from the Registry of Causes of Death, administered by the Danish National Board of Health, and from the Civil Registration System. Four children who were deceased according to the data from the Civil Registration System were not listed in the Registry of Causes of Death. The information was checked in the children's medical records, according to which the four children were deceased. Stillbirth was defined as delivery of a dead fetus occurring at 28 completed weeks of gestation or later, and infant death was defined as death of a liveborn infant occurring before age 1 year.

The study population was restricted to Danish-speaking women who answered the first questionnaire and who carried their pregnancy to a livebirth or a delivery after 28 weeks of gestation (n = 26,326). Twins were excluded from the study (n = 882). The study population was further restricted to those with valid information about the mother's smoking habits during pregnancy (n = 25,102). The second questionnaire was answered by 18,598 (74 percent) women and the third by 16,500 (66 percent).

Women were categorized as smokers if they smoked one or more cigarettes per day when 16 or 30 weeks pregnant. Only minimal changes in smoking habits appear after 16 weeks of gestation (27). Smoking was analyzed in two categories (smokers vs. nonsmokers), and also in the ordered categories 0, 1–9, and 10 or more cigarettes per day. Furthermore, nonsmokers were analyzed in two categories; those who stopped smoking before 16 weeks of pregnancy were compared with those who were nonsmokers from the beginning of pregnancy. Caffeine intake was calculated from daily intake of coffee (1 cup = 100 mg), tea (1 cup = 50 mg), chocolate (1 cup = 50 mg), and cola (250 ml = 50 mg). Caffeine intake was analyzed in two categories (<400 vs. ≥400 mg/day).

Statistical analyses

The association between smoking during pregnancy and mortality from the third trimester to age 1 year is presented as odds ratios with 95 percent confidence intervals. Other variables accounted for in the analyses were categorized as seen in table 1. These variables were evaluated in logistic regression analyses in two ways. They were included one at a time, and if they changed the measure of association between smoking and mortality by more than 10 percent, they remained in the final model (28). Subsequently, all of the covariates in table 1 were included at the same time based on a priori information that they might potentially confound the results. All covariates were entered as a number of dummy variables equal to the number of categories minus one. Effect modification was evaluated by stratified analyses and by logistic regression with interaction terms. The association between smoking and infant mortality was also evaluated by Cox regression analyses. However, because the results were similar to those from logistic regression analyses, they are not presented. The attributable fraction was calculated from the formula: attributable fraction = proportion of exposed in the population × (odds ratio − 1)/proportion of exposed in the population × (odds ratio − 1) + 1 (29).

TABLE 1.

Stillbirth and infant mortality according to smoking habits during pregnancy and other sociodemographic and lifestyle factors, Denmark, September 1989 to August 1996

	No.*	Smokers (%)	Stillbirth		Infant mortality
			No.	%	No.	%
Cigarettes/day
0	17,599		63	3.6	69	3.9
1–9	3,268		19	5.8	27	8.3
≥10	4,235		34	8.0	27	6.4
Parity
Primiparous	12,813	28.7	63	5.0	71	5.6
Multiparous	12,289	31.1	53	4.3	52	4.2
Sex of the child
Male	12,903	29.5	69	5.3	70	5.5
Female	12,198	30.3	46	3.8	53	4.4
Maternal age (years)
15–24	4,651	41.6	18	3.9	28	6.0
25–29	11,432	27.0	48	4.6	52	5.0
30–50	10,019	27.4	50	5.0	43	4.3
Marital status
Married or living together	20,270	27.4	94	4.6	95	4.7
Single	919	56.0	3	3.3	6	6.6
Missing	3,913	36.6	19	4.9	22	5.6
Education (years)
<10	7,558	42.5	32	4.2	35	4.7
≥10	10,546	19.1	44	4.2	47	4.5
Missing	6,998	32.5	40	5.7	41	5.9
Working status during pregnancy
Working	12,014	26.4	53	4.4	53	4.4
Unemployed	3,872	42.3	19	5.0	21	5.5
Student	1,902	15.8	7	3.7	5	2.6
Missing	7,314	32.7	37	5.1	44	6.0
Caffeine intake during pregnancy (mg/day)
<400 mg	12,148	19.3	41	3.4	55	4.5
≥400 mg	6,447	46.9	40	6.2	29	4.5
Missing	6,507	32.8	35	5.4	39	6.0
Alcohol intake during pregnancy (drinks/week)
0	16,606	28.8	67	4.0	83	5.0
1–2	5,609	27.4	28	5.0	23	4.1
3–4	1,758	38.7	10	5.7	8	4.6
≥5	557	52.1	7	12.6	4	7.3
Missing	572	36.5	4	7.0	5	8.8
Maternal height (cm)
<160	1,551	35.7	12	7.7	7	4.5
160–164	5,186	31.8	28	5.4	31	6.0
165–169	7,597	29.9	37	4.9	37	4.9
≥170	10,333	27.7	39	3.8	44	4.3
Missing	435	37.5	0	0	4	9.2
Maternal weight (kg)
<50	1,158	41.8	5	4.3	9	7.8
50–59	9,484	31.2	41	4.3	39	4.1
60–69	9,144	27.2	40	4.4	49	5.4
≥70	4,817	28.8	25	5.2	23	4.8
Missing	499	37.7	5	10.0	3	6.1

^*

Includes livebirths and stillbirths. Calculation of infant mortality is based on livebirths only.

The study was approved by the regional ethics committee, the Danish Data Protection Agency, and the Danish National Board of Health.

RESULTS

Among the 25,102 pregnant women, 17,599 (70 percent) were nonsmokers, and 7,503 (30 percent) were smokers. Among the nonsmokers, 2,651 (15 percent) had been smokers at the beginning of pregnancy but had stopped in the first trimester. Among the smokers, 3,268 (44 percent) smoked 1–9 cigarettes per day, and 4,235 (56 percent) smoked 10 or more cigarettes per day. The mean number of cigarettes smoked per day during pregnancy was 9.3 (standard deviation, 4.8).

The rate of stillbirth was 4.6/1,000. The infant mortality rate was 4.9/1,000. The rates of stillbirth and infant mortality according to smoking habits during pregnancy and a number of sociodemographic and lifestyle factors are presented in table 1.

Compared with nonsmokers, smokers had about twice the risk of stillbirth (table 2) and of infant death (table 3). The risk of stillbirth and infant death among children born to women who stopped smoking in the first trimester was comparable with that in children of women who were nonsmokers from the beginning of pregnancy (tables 2 and 3). The mean age at death among children of smokers was 55 days, and among children of nonsmokers, it was 42 days (mean difference, 14 days, 95 percent confidence interval: −40, 14).

TABLE 2.

Stillbirths among children of women with various smoking habits during pregnancy, Denmark, September 1989 to August 1996

	No.	Stillbirths		OR*	95% CI*	OR†	95% CI
		No.	%
Nonsmokers from 16 weeks gestation	17,599	63	3.6	Reference		Reference
Smokers, irrespective of amount (cigarettes/day)	7,503	53	7.1	2.0	1.4, 2.9	1.9	1.3, 2.9
1–9	3,268	19	5.8	1.6	1.0, 2.7	1.5	0.9, 2.4
≥10	4,235	34	8.0	2.1	1.2, 3.9	1.8	1.2, 2.8
Nonsmokers during entire pregnancy	14,948	54	3.6	Reference		Reference
Stopped smoking before 16 weeks gestation	2,651	9	3.4	0.9	0.5, 1.9	0.9	0.5, 1.9

^*

OR, odds ratio; CI, confidence interval.

^†

Adjusted for parity, sex of the child, maternal age, marital status, years of education, working status during pregnancy, caffeine and alcohol intake during pregnancy, and maternal weight and height.

TABLE 3.

Deaths among children of women with various smoking habits during pregnancy, Denmark, September 1989 to August 1996

	No.*	Infant deaths		OR†	95% CI†	OR‡	95% CI
		No.	%
Nonsmokers from 16 weeks gestation	17,536	69	3.9	Reference		Reference
Smokers, irrespective of amount (cigarettes/day)	7,450	54	7.3	1.8	1.3, 2.6	1.9	1.3, 2.8
1–9	3,249	77	8.3	2.1	1.4, 3.3	1.7	1.1, 2.7
≥10	4,201	27	6.4	1.6	1.0, 2.5	1.5	0.9, 2.3
Nonsmokers during entire pregnancy	14,894	59	4.0	Reference		Reference
Stopped smoking before 16 weeks gestation	2,642	10	3.8	1.0	0.5, 1.9	0.9	0.5, 1.9

^*

Calculation of infant mortality is based on livebirths only.

†

OR, odds ratio; CI, confidence interval.

‡

Adjusted for parity, sex of the child, maternal age, marital status, years of education, working status during pregnancy, caffeine and alcohol intake during pregnancy, and maternal weight and height.

Adjustment for sex of the child; parity; and maternal age, marital status, years of education, occupational status, alcohol intake during pregnancy, and height and weight changed the estimates of association only nonsubstantially (table 4). We obtained results similar to those in tables 2 and 3 when, independent of parity, only a woman's first pregnancy in the study period was included in the analyses (data not shown).

TABLE 4.

Distribution of causes of death among infants born to smokers and nonsmokers, Denmark, September 1989 to August 1996

	Children of smokers		Children of nonsmokers
	No.	%	No.	%
Intrapartum events	5	9.3	5	7.2
Congenital abnormalities	20	37.0	33	47.8
Conditions due to preterm delivery	2	3.7	4	5.8
Infections	3	5.6	1	1.4
Respiratory disorders	1	1.9	6	8.7
Endocrine disorders	2	3.7	3	4.3
Sudden infant death syndrome	12	22.2	8	11.6
Miscellaneous	9	16.7	9	13.0
Total	54		69

Distribution of causes of infant deaths among children of smokers and nonsmokers is shown in table 4. Exclusion from the infant mortality rate of those children who died from sudden infant death syndrome gave a result comparable with that presented in table 3 (odds ratio = 1.6, 95 percent confidence interval: 1.1, 2.4).

The association between smoking and stillbirth was independent of caffeine intake. In women whose intake of caffeine was less than 400 mg per day, smoking increased the risk of infant death by 40 percent (odds ratio = 1.4, 95 percent confidence interval: 0.8, 2.6), and when caffeine intake was 400 mg or more per day, smoking increased the risk of infant death more than fourfold (odds ratio = 4.4, 95 percent confidence interval: 1.8, 10.8). This difference was mainly due to low infant mortality among nonsmokers with a high intake of caffeine.

If our results reflect causal relations between smoking, stillbirth, and infant mortality, approximately 25 percent of all stillbirths and 20 percent of all infant deaths could be avoided if all pregnant smokers stopped smoking by the sixteenth week in a population with 30 percent pregnant smokers.

DISCUSSION

The association between smoking during pregnancy and mortality during the last part of the pregnancy and during the first year of life has often been studied (6–9, 11–21, 24), but the results are ambiguous. Several early studies found no significant differences in the risk of death before and during labor or of perinatal mortality when pregnant smokers were compared with nonsmokers (15, 21, 23). These studies were all small, and relevant confounders were not included in the analyses. Other, larger studies have found increased risk attached to smoking during pregnancy (12, 13, 17). Cnattingius et al. (13) included almost 300,000 births and found that smoking increased the risk of intrauterine fetal death by 40 percent during the third trimester, while the risk of death within the first 6 days of life was nonsignificantly increased by 20 percent. Kleinman et al. (12) found that, among primiparae, the risk of death before and during delivery, of neonatal mortality, and of death during the second to twelfth months of life increased with the number of cigarettes the mother smoked during the pregnancy (12). These results had been adjusted for differences in the mothers' age, race, parity, marital status, and level of education. However, differences in alcohol intake between smokers and nonsmokers had not been taken into consideration, a fact that may explain part of the differences found in the mortality rates of children born to smokers and those born to nonsmokers.

In this study of 25,102 deliveries, smoking increased the risk of stillbirth and infant death. Because of prospective collection of information about maternal lifestyle factors and sociodemographic factors, we could adjust for a number of potential confounders, including alcohol intake during pregnancy. Apparently, these variables could not explain our findings. However, information about socioeconomic variables was missing in about one third of all women, which may constrain the possibility to adjust fully for these factors. We found a higher risk associated with smoking than in previous studies that have found significantly increased mortality in infants of smokers (7, 10, 12, 13). This may be due to the fact that our study was conducted in a homogeneous population with low overall mortality, reflecting lower prevalence of competing risks.

To our knowledge, no previous study has been able to examine the effect of smoking cessation during the first trimester. We found that among children of women who stopped smoking before 16 weeks of gestation the risk of stillbirth and infant death was comparable with that in children of women who had been nonsmokers from the beginning of pregnancy. This result indicates that smoking exerts its influence on mortality at some point after the first trimester and supports the causal role of fetal exposure to tobacco smoke in late pregnancy. The result also emphasizes the importance of providing smoking cessation programs for pregnant women who indicate at the first antenatal visit that they smoke. In our population, very few women gave up smoking after the first trimester, and therefore, we could not study the effect of smoking cessation after that period.

We did not have information on maternal smoking habits after delivery, but these are likely to be closely associated with smoking habits during pregnancy. Discrimination between the effects of maternal smoking during pregnancy and after birth on infant death therefore remains speculative. If passive exposure to cigarette smoking after delivery explained our findings, the risk of respiratory deaths would most likely be increased among children of smokers compared with children of nonsmokers. However, children of smokers were not at increased risk of respiratory deaths compared with children of nonsmokers. Furthermore, the distribution of causes of infant deaths showed that exposure to tobacco smoke in utero apparently increased the risk of death due to a variety of causes, which may indicate that the harmful effect of exposure to tobacco smoke is not restricted to a single organ.

Information about smoking habits was obtained from self-administered questionnaires at 16 and 30 weeks of gestation. Information about smoking habits at 30 weeks of gestation was missing in about one third of the population. However, previous studies in the same population have shown that only minimal changes in smoking habits occur after the first trimester (27, 30). Furthermore, studies of pregnant women indicate that they underreport their smoking (31, 32). The underreporting, either by completely denying the habit or by reporting a smaller number of cigarettes than the amount they smoke, would cause an underestimation of the true effect of smoking and make it difficult to find the correct dose-response association. Underreporting may thus explain why we found the same risk of infant death among women who smoked fewer than 10 cigarettes per day compared with those who smoked 10 or more cigarettes per day.

Information about death was obtained from Danish registers. Because information about death is held in more than one register and because we had access to the medical records of the mothers and their children, we could manually check the information from the registers when there was disagreement between different sources of information. With this procedure, the number of misclassified deaths was reduced.

In countries where smoking is prevalent, exposure to tobacco smoke in utero is probably the most important preventable risk factor for stillbirth and infant death. Assuming that the associations found in this study reflect causal associations, in Denmark, one fourth of all stillbirths and one fifth of all deaths occurring during the first year of life could be avoided if all pregnant women stopped smoking by the sixteenth week of their pregnancy.

Correspondence to Dr. Kirsten Wisborg, Perinatal Epidemiological Research Unit, Department of Gynaecology and Obstetrics, Aarhus University Hospital, DK-8200 Aarhus N, Denmark (e-mail: ).