Background: Maternal cigarette smoking has been causally associated with an increased risk for stillbirth. Preliminary reports suggest an increased risk for stillbirth with smokeless tobacco use during pregnancy.
Methods: We conducted a population-based prospective cohort study to investigate this association by using a house-to-house approach to recruit 1217 women who were between 3 and 7 months' gestation. Of these, 96% were contacted after delivery to determine the pregnancy outcome. Demographic and maternal variables which were apparently associated either with stillbirth or with smokeless tobacco use (OR ≥ 1.5) were included as potential confounders. Stillbirth was defined as any delivery of a dead fetus after 20 completed weeks of gestation. We used time-to-event methods to analyze the risk of stillbirth.
Results: Overall occurrence of stillbirth among singleton deliveries in this population was 4.1%. Smokeless tobacco use was reported by 17% of women; 8.9% of smokeless tobacco users had a stillbirth compared with 3.1% among nonusers (life-table adjusted hazard ratio = 3.1; 95% confidence interval = 1.7–5.6). After adjustment by the Cox proportional hazards procedure for age, educational and socioeconomic background, working status of mother, parity, prenatal care variables, and place of delivery, the risk for stillbirth in users was 2.6 (95% confidence interval-1.4–4.8). Most women used mishri (a pyrolyzed tobacco product often used as dentifrice), and there was a dose–response relationship between the daily frequency of use and stillbirth risk. The risk of stillbirth associated with smokeless tobacco use was greater in earlier gestational periods.
Conclusions: Smokeless tobacco use during pregnancy increases stillbirth risk, with a risk at least as great as that associated with maternal cigarette smoking.
From the Epidemiology Unit, Tata Institute of Fundamental Research, Mumbai, India.
Submitted 28 January 2005; accepted 26 July 2005.
Supported by WHO India.
Correspondence: Sreevidya Subramoney, 601, Great Eastern Chambers, Plot No 28, Sector 11, CBD Belapur, Navi Mumbai 400614. E-mail: firstname.lastname@example.org.
Perinatal mortality is one of the most sensitive indices of maternal and child health. Each year, approximately 8 million perinatal deaths occur globally, 98% in the developing countries; of these, 60% are stillbirths. Rates of perinatal and neonatal mortality in developing countries have decreased more slowly than rates of infant mortality.1
In western countries, maternal cigarette smoking has been shown to be causally associated with an increased risk for stillbirth; the relative risk ranges from 1.2 to 1.6.2 While cigarette smoking is prevalent among women in developed countries, the use of smokeless tobacco is prevalent among women in many developing countries.3 Smokeless tobacco use also is becoming increasingly common among young girls globally.4
Particularly high prevalence of smokeless tobacco use among women (between 15% and 60%) has been reported in several states in the northeastern, eastern, and western regions of India.5,6 In the Mumbai Cohort study,7 of 59,527 lower-middle and lower-class women age 35 years and older, 58% reported current tobacco use, virtually all of which was smokeless.
Evidence on stillbirth risk with smokeless tobacco use is sparse.8,9 A hospital-based cross-sectional study8 on 1388 singleton births in Pune, India, reported that the rate of stillbirth in tobacco chewers was 3 times higher than that in nonusers. A multicenter hospital-based case–control study in India9 reported a 1.5 times higher risk of perinatal death (95% CI = 1.3–1.7) in tobacco users. We report on the effect of smokeless tobacco use during pregnancy on stillbirth in a cohort of nonsmoking pregnant women in the city of Mumbai.
A cohort of 1217 pregnant women was interviewed on a house-to-house basis from 8 selected geographic areas of Mumbai City. These geographic areas corresponded to Corporation health post areas. There are about 180 health posts in the city serving the health needs of the population, including maternal and child health needs. We selected 8 health posts on the basis of the following criteria: those registering at least 100 new pregnancies per month, accessibility, and enthusiasm of the health post authorities about participating in the study.
Community Health Volunteers attached to these Health posts (n = 177) regularly monitor and list pregnant women in their respective areas during their house-to-house visits. All pregnant women willing to register for antenatal care with the Corporation Health posts are listed.
A project social worker, accompanied by the Community Health Volunteer, visited all the listed women and screened them for eligibility before interviewing them with a pilot-tested, structured questionnaire. Eligible women were defined as those who were between 12 and 28 weeks of gestation and who planned to remain in the study area for delivery. Every woman provided oral informed consent to study participation before enrollment in the study. Community Health Volunteers, working in their areas for several years, were well known in the community and had an established rapport with the community. Thus, there were no refusals for study participation. A second interview at recruitment (10%) by a different social worker was used as a quality control check.
Demographic information and medical and obstetric history (including the date of the last normal menstrual period, and the gestational age in months) was elicited at recruitment. Weight and height of the women were measured using a bathroom weighing scale and an inch tape, respectively. Women's weight before pregnancy generally was not known. Detailed information on all types of tobacco consumption was obtained during this phase. Ultrasound estimates of gestational age were abstracted from antenatal records whenever available. After interview, the women were provided with health education regarding optimal antenatal care and were advised to keep regular contact with the health care system. There was no specific focus on smokeless tobacco use.
The Community Health Volunteers and the social workers regularly monitored the recruited women. A follow-up interview was conducted for 1167 women (96%) at their residence after the delivery. Women were asked their tobacco habits just before delivery and gestational age at birth. Medical records of delivery were abstracted. We obtained the date of delivery from records and verified this date with the woman. Women were advised on immunization schedules for their infant and the importance of breast-feeding. If the infant was reported dead, the time of death was obtained (born dead or died minutes or hours after birth).
Smokeless tobacco use was categorized as light (1–4 times per day) or heavy (5 or more times per day). Women in the state of Maharashtra (which includes Mumbai City) commonly use mishri. Mishri is pyrolyzed and powdered tobacco, the use of which starts as a dentifrice. Gutka and betel quid are other commonly-used forms of smokeless tobacco. These include other ingredients such as the areca nut and are generally retained in the mouth for longer periods of time and their juices sucked and swallowed, or spit out. Very few women smoke, and those who do, smoke bidi (a small amount of tobacco flakes, approximately 0.2 g, hand rolled in a dry tendu leaf tied with a thread).
Any fetus that did not breathe or show any other evidence of life at birth after a minimum 20 weeks of gestation was defined as a stillbirth. Early neonatal deaths and stillbirths are very clearly differentiated in the Maternal and Child Health program in India, which has been operational for several decades; doctors are trained to specifically differentiate the 2 outcomes. In addition, mothers were specifically asked whether the child cried and was well, and breast-fed immediately after birth or was resuscitated or transferred to a special care nursery. Our analyses are restricted to stillbirths, as our perinatal mortality statistics were not complete.
Gestational age at delivery was determined on the basis of ultrasound reports before the completion of 30 weeks of pregnancy whenever available (n = 453). For the rest, we use the number of days from the date of start of the last normal menstrual period to the date of delivery. The agreement of ultrasound with self-reports among women for whom both were available was 70% for preterm deliveries and 95% for term deliveries.
Of the 1167 women who were followed-up, we excluded 2 bidi smokers, 8 women who gave birth to twins, 21 who had aborted before 20 weeks, and 26 with no date of birth and with secondhand information only. Thus, 1110 women were available for analysis of stillbirth rates. Of these 1110 women, 45 had delivered at home; all these women were interviewed after delivery and included in the study.
Socioeconomic status classification involved calculation of an index for each woman. This index, which is considered standard for urban Indian families,10 uses the sum of standard scores assigned to categories of the educational and occupational status of the father and the per-capita income of the family. Only 2 women had upper-class socioeconomic status score; these women were included along with the middle class women for all analyses.
Demographic and other variables were individually examined for their association with stillbirth; hazard ratios (HRs) and 95% confidence intervals for the risk of stillbirth were calculated using the Cox proportional hazards procedure, with days of gestation as the timescale. Their association with smokeless tobacco use was also individually examined. Potential confounding variables were those that were apparently associated with either the exposure or the outcome (OR ≥ 1.5). HIV status was available for 500 women, all of whom were negative. Only 2 women reported alcohol consumption. Therefore, HIV and alcohol use were not included in the analysis. Additional variables considered but not included in the multivariate model were religion, language, type of accommodation, height of mother, weight of mother, interpregnancy interval, and previous preterm delivery, as none of these variables showed association with stillbirth or with smokeless tobacco use.
Stillbirth rates in smokeless tobacco users and nonusers were compared. We then performed multiple covariate analysis using time-to-event methods as described by Smith,11 with days of gestation as the timescale. Frequency and type of smokeless tobacco use were similarly analyzed. Stillbirth risk at particular gestational periods was calculated similarly using life table methods (as the ratio of the number of stillbirths during the period to all ongoing pregnancies at the start of that period) and adjusted for potential confounders using the Cox proportional hazards procedure. Survival curves were derived using the life table procedure.
The response rate among pregnant women in the community was 100%. In 10% rechecks, the intersocial-worker agreement was 100% for tobacco use and over 90% for most other variables. Of the 1217 recruited women, regular tobacco use during pregnancy was reported by 208 (17.1%). Smokeless tobacco use predominated (99% of tobacco users; n = 206), with mishri being the most common (81%). Smokeless tobacco use per day was 1–2 times for 46%, 3–4 times for 21%, and higher for the rest. Reduction or cessation of tobacco use late in the pregnancy was reported by only 5 women; hence they were not analyzed separately.
The variables that qualified as potential confounders (Table 1) include age of mother, working status of mother, number of antenatal care visits, number of Tetanus toxoid doses taken, place of delivery, previous stillbirth and previous cesarean section (all associated with an increased risk of stillbirth). Additionally, educational status, socioeconomic score, and parity, were included as potential confounders, as they were associated with smokeless tobacco use.
Of 1110 women, 46 (4.1%) had stillbirths; 28 among 908 nonusers (3.1%) and 18 among 202 smokeless tobacco users (8.9%). The unadjusted HR of stillbirth in users (calculated by life table methods) was 3.1 (95% CI = 1.7–5.6); after adjustment for potential confounders (Table 1, excluding previous stillbirth and previous cesarean delivery) by the Cox proportional hazards procedure, the risk was 2.6 (1.4–4.8).
The unadjusted hazard of stillbirth with smokeless tobacco use among women who had experienced a previous pregnancy (n = 848) was 2.6 (1.3–5). After adjustment for all potential confounders listed in Table 1 (including previous stillbirth and previous cesarean delivery) the HR among these women was 2.1 (1.03–4.4).
Table 2 shows the crude and adjusted risk of stillbirth for the 2 common types of smokeless tobacco, mishri and gutka, and the increase in stillbirth risk with increasing use of mishri. This increased risk remained substantial after adjustment for important covariates by the Cox proportional hazards procedure.
The excess risk of stillbirth associated with smokeless tobacco use was apparent throughout gestation (Figure 1). Table 3 shows the proportion of surviving fetuses, stillbirth rates, and HRs by gestational age. HRs were higher at lower gestational periods.
We previously reported an effect of smokeless tobacco use on birth weight,12 reducing babies' weights by approximately 100 g. In the present analysis we find that smokeless tobacco use during pregnancy confers an increased risk for stillbirth with a strong dose-related influence and independent of confounding factors. This risk is at least as large as the effect of maternal cigarette smoking. The risk associated with smokeless tobacco use was higher in earlier gestational periods.
Calculation of gestational age for approximately 60% of the women was based on the last menstrual period, which is likely to have resulted in some misclassification error. This imprecision was unavoidable in a population-based study in India, as in many other countries. However, because this error has occurred in both smokeless tobacco user and nonuser groups, it is unlikely to have had a major influence on the calculated relationship between smokeless tobacco use and stillbirth.
Other limitations are the lack of data on infant sex and birth weight for some infants, so that these variables could not be included in the analysis. The possibility of bias owing to differential enrollment of women does not arise in our study, as all women in the selected areas were listed by a house-to-house approach and all listed women were enrolled.
Tobacco in smokeless form contains several carcinogenic and toxic substances.13 Heavy metals such as lead and cadmium have been found in smokeless tobacco at levels that may present potential risks to the fetus. Exposure to cotinine in fetuses of mishri users has been demonstrated,14 indicating that nicotine and perhaps other toxic substances can cross the placental barrier.
Smokeless tobacco use during pregnancy has been associated with growth restriction,12 preterm delivery,12 and placental morphologic changes.15 Mediation of increased stillbirth risk through these mechanisms is a possibility. Also, nicotine may have a direct toxic effect by inducing a change in central respiratory control mechanism and eliciting fetal hypoxia-ischemia.16 Although the reproductive health effects of cigarette smoking have been studied extensively (and smoking during pregnancy is established as a cause of stillbirth), the biologic mechanisms remain to be fully delineated.2
It is interesting that use of gutka (containing areca nut in addition to tobacco) conferred a somewhat greater risk of stillbirth than the use of mishri (adjusted HR = 5.5 for gutka compared with 2.5 for mishri). In a case–control study among aborigines in Taiwan17 (62 subjects with adverse pregnancy outcomes and 124 age-matched controls), chewing of betel quid (including areca nut but not tobacco) was associated with a 2.8-fold higher risk of adverse pregnancy outcomes. However, a prospective study by De Costa and Griew18 (400 betel chewers and 400 nonchewers) did not suggest an increased risk for perinatal mortality.
The perinatal mortality rate in India and Bangladesh stands at 65–80/1000 total births, in comparison to 20–25/1000 total births in Srilanka and Thailand, and 3–5/1000 total births in the developed countries.1 With 17% prevalence of smokeless tobacco use and 3 times higher risk of stillbirth in users, about 25% of stillbirths in the study population may be attributable to smokeless tobacco use (calculated as [i − io]/i, where i is the overall incidence of stillbirth in the population, and io the incidence in the nonexposed).
High rates of smokeless tobacco use have been reported among youth (13–15 years) globally including in India,4 especially so among girls.19 This may have serious consequences for reproductive health.
We thank all the women who participated in the study, the Community Health Volunteers, and the doctors at the health posts selected for the study.
1. World Health Organisation. Perinatal Mortality
: A Listing of Available Information.
Maternal Health and Safe motherhood Program (WHO/FRH/MSM/96. 7), Geneva: WHO; 1996.
2. Office on Smoking and Health. USDHHS. Women and Smoking.
A report of the Surgeon General 2001. Atlanta: Center for Disease Control and Prevention, National Center for Chronic disease Prevention and Health Promotion.
3. Gupta PC, Ray CS. Smokeless tobacco and health in India and South Asia. Respirology
4. The Global Youth Tobacco Survey Collaborative Group.Tobacco use among youth: a cross-country comparison. Tob Control
5. Rani M, Bonu S, Jha P, et al. Tobacco use in India: prevalence and predictors of smoking and chewing in a national cross sectional household survey. Tob Control
6. Sinha DN, Gupta PC, Pednekar MS, et al. Tobacco use among school personnel in Bihar, India. Tob Control
7. Gupta PC. Socio-demographic characteristics of tobacco use among 99,598 individuals in Bombay, India using hand-held computers. Tob Control
8. Krishna K. Tobacco chewing in pregnancy. Br J Obstet Gynaecol
9. Shah D, Shroff S, Ganla K. Factors affecting perinatal mortality in India (perinatal audit). Prenat Neonat Med
10. Gupta MC, Mahajan BK. Modified Kuppuswamy's scale for socioeconomic status of urban Indian families, (converted to All India Consumer Price Index 1997–98). In: Textbook of Preventive and Social Medicine.
1st ed. New Delhi: Jaypee Brothers; 1992:82–86.
11. Smith GCS. Estimating risks of perinatal death. Am J Obstet Gynecol
12. Gupta PC, Sreevidya S. Smokeless tobacco use, birthweight and gestational age: a population based prospective cohort study of 1217 women in Mumbai, India. BMJ
13. International Agency for Research on Cancer: IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans. Vol. 37, Tobacco Habits Other Than Smoking; Betel-quid and Areca-nut Chewing; and Some Related Nitrosomines
. Lyon: International Agency for Research on Cancer; 1984.
14. Sarkar S, Bhide MB, Bhide AM, et al. Transplacental exposure of human fetuses to nicotine/cotinine in masheri addict mothers and consequent loss of weight in babies at birth. In: O'Neil IK, Bartsch H, eds. Nitrosocompounds: Biological Mechanisms, Exposures and Cancer Etiology,
Kailua-Kona, Hawai: IARC Technical report No 11; 1991.
15. Agrawal P, Chansoriya M, Kaul KK. Effect of tobacco chewing by mothers on placental morphology. Indian Pediatr
16. Slotkin TA. Fetal nicotine or cocaine exposure: which one is worse? J Pharmacol Exp Ther
17. Yang MS, Chang FT, Chen SS, et al. Betel quid chewing and risk of adverse pregnancy outcomes among aborigines in southern Taiwan. Public Health
18. De Costa C, Griew AR. Effects of betel chewing on pregnancy outcome. NZJ Obstet Gynaecol
© 2006 Lippincott Williams & Wilkins, Inc.
19. Global Youth Tobacco Survey Collaborating Group. Differences in worldwide tobacco use by gender: findings from the Global Youth Tobacco Survey. J Sch Health