Chapter 1 Smoking-Related Health Problems Unique to Women INTRODUCTION Smoking habits and attitudes among women and teenage girls have differed in the past from the habits and attitudes among men and teenage boys. Women tended to smoke fewer cigarettes, were less likely to inhale, and were more likely to smoke low "tar" and nico- tine and filter-tipped brands. Surveys have indicated, however, that the smoking habits of women are becoming more like men's, Women are taking up the habit at an earlier age and have become heavier smokers. This has made them more vulnerable not only to lung cancer and other smoking-related diseases, but also to specific health problems that are unique to their sex. For example, research on the relationship between cigarette smoking and the outcome of pregnancy has established that there are definite risks to both the fetus and the mother associated with cigarette smoking during pregnancy. Moreover, women who use oral contraceptives are at greater risk of cardiovascular disease if they smoke cigarettes. There is also evidence that nicotine is present in the breast milk of lactat- ing mothers who smoke. The following is a review of the current information on these and other health consequences of smoking unique to women. EFFECTS OF SMOKING ON THE OUTCOME OF PREGNANCY There are definite health risks associated with smoking and preg- nancy, including effects on birth weight, perinatal mortality, and long-term physical and intellectual development of the child. This section reviews each of these subjects and also includes information about the likely mechanism of action of smoke and its contents on the mother and the products of conception. Smoking and Birth Weight In 1957, Simpson published her original finding that babies born to women who smoke during their pregnancy weigh on the average 200 grams (g) less than the babies born to women who do not smoke 1 (34). Since then, more than 100 articles on this relationship have led to the general acceptance that smokers' babies generally weigh 150 to 250 g less than nonsmokers' babies, and twice as many of the former weigh less than 2500 g (13). The 1973 report of The Health Consequences of Smoking presented evidence to support a causal association between cigarette smoking and fetal growth re- tardation (39). A strong dose-response relationship was also estab- lished in that report, with differences in weight being in direct pro- portion to the number of cigarettes smoked. The following additional points were summarized in the 1973 report to further support the causal association between cigarette smoking during pregnancy and lower birth weight: 1. Results are consistent in all studies, retrospective and pro- spective, from many different countries, races, cultures, and geographic settings. 2. The relationship between smoking and reduced birth weight is independent of other factors that influence birth weight, such as race, parity, maternal size, socioeconomic status, sex of child, and all others that have been studied. 3. If a woman gives up smoking by the fourth month of preg- nancy, her risks of delivering a low-birth-weight baby is similar to that of a nonsmoker. Subsequent to the 1973 report, additional reports have further discussed and corroborated the association between smoking in pregnancy and low birth weight (19, 25,33, 35). Smoking and Perinatal Mortality A strong, probably causal, association between cigarette smoking and higher late fetal and infant mortality rates among smokers' infants is now well established (38). Retrospective and prospective studies have revealed a statistically significant relationship between cigarette smoking and an elevated mortality risk among the infants of smokers. In three of these studies of sufficient size to permit ad- justment for other risk factors, a highly significant independent as- sociation between smoking and mortality was established. Part of the discrepancy in results between these studies and those in which a significant association between smoking and infant mortality was not demonstrated may be explained by a lack of adjustment for risk factors other than smoking. The 1973 report also presented evidence indicating that the higher relative risks occurred among populations with risk factors other than smoking being present, such as socioeconomic status, age, parity, race, and previous pregnancy history. 2 Since 1973, a series of articles by Meyer, et al. analyzed data from the Ontario Perinatal Mortality Study of all single births in ten Ontario teaching hospitals in 1960-61 (26, 27, 28). The study in- volved 5 1,490 births, including 701 fetal deaths and 655 early neo- natal deaths, and was supplemented by clinical records with inter- views of mothers in the hospital, interviews with anesthetists and attending physicians, and autopsy records (29). Perinatal mortality increased significantly with smoking and was also affected by such factors as maternal age, parity, socioeconomic status, previous pregnancy history, hemoglobin level, and other risk factors (29). Smoking frequencies also varied by many of these characteristics. Smoking and other risk factors were cross-tabulated among 52 data subgroups. In all subgroups, the mortality increase with smoking was dose related, but not in a simple, linear way. The increased risk of perinatal mortality associated with light smoking among young, low-parity, nonanemic mothers was less than 10 percent. At the other extreme, mothers with other risk factors of high parity, public hospital status, with previous low-birth-weight infants, or with hemoglobin less than 11 g had further increased perinatal mortality risks of 70- 100 percent when they were smokers. The most significant risk factor (mortality rate of 78 per 1,000 total births) was anemia, defined as a hemoglobin of less than 8.0 g. The failure of some earlier studies to find a significant increase in perinatal mortality with maternal smoking may be due to selection of study populations from the end of the spectrum, where light smoking is associated with only a slight increase in perinatal risk. This evidence points up how population selection could influence study findings and shows that exposure to the effects of smoking during pregnancy is much more dangerous for the babies of some women than for others. These findings are corroborated by a num- ber of studies in which fetal, neonatal, or perinatal mortality rates are compared for smoking and nonsmoking women, controlling for the effects of various risk factors previously mentioned (1, 12, 22, 36). Additional data were published in 1976- 1977 (26, 27) and re- vealed that frequencies of low birth weight (under 2500 g), preterm delivery (< 38 weeks), perinatal mortality, abruptio placentae, pla- centa previa, bleeding during pregnancy, and prolonged and pre- mature rupture of the membranes increased directly and signifi- cantly (p< 0.00001) as the level of maternal smoking increased (Tables 1 and 2). The 1976 paper used multiple regression analysis to measure the independent effect of smoking on the various risk factors. The probabilities of these complications were also com- pared (Figure 1). Risks of placenta previa and abruptio placentae were higher for smokers than for nonsmokers at all gestations, with 3 TABLE 1. Adjusted rates and F ratios for maternal smoking and other impor- tant factors affecting birth ,weight, gestation, placental complica. tions, and perinatal mortality Factor *Adjusted Rates of Outcome tF Ratio Maternal Smoking Level NOIIS < 1 Pack Per Day > 1 Pack Per Day Previous Pregnancy History No Prwious Pregnancy Previous Pregnancy, 0 Loss Previous Pregnancy, Loss Hospital Pay Status Private Public - Birth Weight < 2500 Grams Per 1000 Births 49.4 182.8 75.7 113.7 70.0 123.5 57.8 134.0 60.0 87.4 84.0 Gestation < 38 Weeks Per 1000 Births Maternal Smoking Level None 77.1 50.6 < 1 Pack Per Day 92.2 > ' Pack Per Day 115.9 Previous Pregnancy History No Previous Pregnancy 69.1 182.6 Previous Pregnancy, 0 Loss 85.7 Previous Pregnancy, Loss 193.9 Hospital Pay Status Private 78.9 120.3 Public 116.2 Placenta Previa Per 1000 Births Maternal Smoking None 6.5 11.7 < 1 Pack Per Day a.1 > 1 Pack Per Day 12.5 Previous Pregnancy History No Previous Pregnancy 8.8 14.4 Previous Pregnancy. 0 Loss 6.6 Previous Pregnancy, Loss 15.8 (Hospital pay status not a significant factor) 4 TABLE 1. Adjusted rates and F ratios for maternal smoking and other impor- tant factors affecting birth weight, gestation, placental complica- tions, and perinatal mortality (con/inu~d) ,",84d Smoking < 1 Pack Per Day > I Pack Per Day Previous Pregnancy History NO Previous Pregnancy Previous Pregnancy, 0 Loss Previous Pregnancy, Loss Hospital Pay Status Prwate Public Abruptio Placentae Per 1000 Births 16.4 17.1 20.3 27.6 18.8 25.6 17.6 37.4 17.5 20.7 25.0 Perinatal Mortality Per 1000 Births Maternal Smoking Rone < 1 Pack Per Day > 1 Pack Per Day Previous Pregnancy HistOW No Previous Pregnancy Previous Pregnancy, 0 Loss Previous Pregnancy, LOSS 23.5 8.4 28.2 31.8 23.1 97.4 23.6 68.7 Hospital Pay Status Private 23.3 44.2 Public 36.1 *Adjusted rates show independent effect of the factor given, adjusted for all other factors in regression. They are: maternal smoking, hospital pay status, mothers' bhthpla~, height, prepregnant weight, sex of child, previous pregnancy history, and age-parity. tF ratio degrees of freedom: numerator = number of subgroups -1, denominator = infhity. (All differencesshown are highly significant. F ratios indicate the relative importance of the factor.) SOURCE: Personal correspondence, based on data in Meyer, M.B., et al. (26). 0.02 - - 0 A 0.01 - - 0.006 0.008 0.004 - - 6I .s P 0.002 - - 2 9 z 0.001 = = 4 0.0006 ti 0.0006 t 0.0004 FIGURE l.-Risks of selected pregnancy complica- tions for smoking and nonsmoking mothers, by period of gestational age at delivery for A, abruptio pla- centae, B, placenta previa, C, premature rupture of membranes (PROM) SOURCE: Meyer M B et al. (27). --L -`--r-- ~. .-- 4 0.04-e t : I I I I I II I II I I I I I II I I I 20 24 28 32 36 40 44 Gestation: Weeks 0.004 - - .s B 0.001 - - & 0.0008 f 0.0006 8 $ 0.0004-- 6 g 0.0002-- % % t 0.0001 -L 0 El I Ill I I t I I I I I I I l Ill1 I I I I 20 24 28 32 36 40 44 Gestation: Weeks 6 0.1 0.06 0.06 0 C 0.001 0.0008 0.0008 g 0.0004 L z .g 0.0002 D 2 e 0. 0.0001 I I I I I I I I I I II I 1 I I I I I I I I I 20 24 28 32 36 40 44 Gestation: Waaks TABLE 2. Perinatal mort&ty and selected pregnancy complications, by maternal smoking levels Smoking level (packs per day) (rates per 1.000 total births) Outcome Perinatal Mortality Abruptio Placentae Placenta Previa Bleeding During Pregnancy Rupture of Membranes > 48 Hours Rupture of Membranes Only at Admission 0 23,358 Births) 23.3 28.0 16.1 10.6 6.4 8.2 116.5 141.6 15.8 23.3 30.3 39.3 &`s* Births) 2* X 33.4 27.8* 28.9 47.32 13.1 28.6* 180.1 201.9f 35.8 109.9f 45.0 45.7i o Cochran's chi square for trends. *p < 0.00001. SOURCE: Meyer, M.B.. et al. (27). 7 relatively larger differences in the earlier weeks of pregnancy. The risk of premature rupture of membranes was more than three times greater for smokers than for nonsmokers among deliveries that oc- curred before 34 weeks gestation and remained higher than the risk for nonsmokers through term (Figure 1C). A prospective investigation of 9,169 pregnant women was con- ducted by Goujard, et a.l. (15), and results showed a substantial increase in stillbirths among smokers. A large proportion of this increase was due to abruptio placentae. There were 100 stillbirths, classified into tive categories of causes: vascular, abruptio placen- tae, mechanical, miscellaneous (syphilis, Rh, malformations, etc.), and unknown (Table 3). The abruptio placentae category exclu- sively represented cases without toxemia, the one toxemic case being classified with the vascular causes. The higher proportion of smokers is significant for only two of the categories: abruptio placentae (p = 0.005) and unknown causes (p = 0.0005). Although the numbers were small, the risk of stillbirths by abruptio placentae is six times higher among smokers. TABLE 3. Stillbirths according to cause in relation to maternal smoking during pregnancy Comparison Number of Percent With Live Stillbirths Deliveries Smokers Births* Cause of Death: Vaaculer Abruptio Placentae Mechanical Miscellaneous (Syphilis, Rh, Malformations, . . .) Unknown Detailed Records Not 0~ 13 13 24 31 2s 46 15 :: p = 0.005 p = 0.0005 Available 5 TOTAL 100 2ti p = 0.0001 Livebirths 9069 i When p is not given, the difference is not significant. SOURCE: Goujard, J., et al. (15). 12 Long-Term Effects on Physical and Intellectual Development Three studies (6, 16, 40) report on long-term effects of smoking in pregnancy. Data from two of the studies presented below demon- strate an association between smoking during pregnancy and im- paired physical and intellectual development in the offspring. Additional reports further substantiate this association (10, 11). Butler and Goldstein (6) analyzed the National Child Develop- ment Study, a longitudinal study of 17,000 children born in Britain from March 3 to 9, 1958. The test procedures included a reading 8 test at the age of 7 years, and a mathematics test, a reading test, and a general ability test at the age of 11. At both ages the height of the child was also measured. Analyses at both ages were based on smoking habits of the mother after the fourth month of pregnancy Statistically significant differences in height and reading ability between smoking categories (0, l-9, or lO+ cigarettes daily) were found at both 7 and 11 years of age. When account was taken for such factors as mother's height, age, social class as determined by father's occupation, number of older and younger children in the household, and the sex of the child, there was a deficit of height and reading ability in the off- spring of mothers who smoked, the extent of which increased with the amount smoked. These results establish an association of smoking in pregnancy with later intellectual development, although the gap between child- ren of smokers (at all levels of smoking) and nonsmokers does not appear to change between the ages of 7 and 11 years. Smoking in pregnancy is associated with an impairment of both mental and physical growth, although compared with other social and biologi- cal factors, the effects are small. In the study by Wingerd and Schoen (40), the net effects of various factors on length at birth and height at 5 years were deter- mined in 3,707 single-born, white California children. Children of smoking mothers were found to be shorter (p< 0.001) at birth and at 5 years than children of nonsmoking mothers. (Intellectual de- velopment was not measured in this study.) In contrast to these results, Hardy and Mellits (16) found very few significant differences in a number of body measurements and intellectual functions up to the age of 7 years between children of smokers and nonsmokers. A possible explanation for this discre- pancy is that their sample was too small, and a weight-matched con- trol group could add a bias. Whereas the British study by Butler and Goldstein involved a sample size of over 5,000 children, Hardy and Mellits based their t'indings on only 88 matched pairs of children. Calculations by the authors of the British study show that with the small sample used by Hardy and Mellits there was only about a 20 percent probability of detecting statistically significant differences in the heights of children born to smoking and nonsmoking mothers. CARBON MONOXIDE AND CARBOXYHEMOGLOBIN LEVELS IN MATERNAL AND FETAL CIRCULATION AND THE POS- SIBLE MECHANISMS OF SMOKING EFFECTS ON PREGNANCY There is evidence to show that carboxyhemoglobin (COHb) levels are substantially elevated in pregnant women who smoke and may result in damage to placental and fetal blood vessels. Higher levels 9 of COHb in both fetal and maternal blood may also be a factor in the increased incidence of low birth weight of infants born to women who smoke. Cole, Hawkins, and Roberts (7) studies the smoking habits of a group of pregnant women and related these to the level of COHb in the circulating blood. A group of 222 patients attending antenatal clinics at a London hospital were questioned about their smoking habits. Ninety-three (42 percent) were smokers, and 129 (58 per- cent) were nonsmokers. Simultaneous maternal and cord blood samples were taken at normal delivery and at Caesarean section from 28 patients, and the COHb and fetal hemoglobin levels of the samples were measured. Results showed that women who smoke during pregnancy have a significantly higher level of COHb in their blood than women who do not smoke (p< 0.01). The mean COHb levels were 1.2 percent (range 0 to 2.4 percent) for the non- smokers and 4.1 percent (range 0.5 to 14 percent) for the smokers. There was a positive correlation betwen the number of cigarettes smoked on the day of sampling and the COHb level (correlation coefficient 0.82) (Figure 2). With the exception of two patients, FIGURE 2.-Number of cigarettes normally smoked per day compared with COHb level at time of sam- pling in 93 pregnant women. $ = Mean range of COHb levels for 129 nonsmokers 15 - o 10 20 30 40 Cigarettes Smoked Per Day SOURCE: Cole, P.V., et al. (7) 10 all the fetal COHb 1 evels were demonstrably higher than the re- spective maternal ones. The mean fetal/maternal COHb ratio was 1.84 to 1 (standard deviation ~0.85). Hemoglobin has a 210 times greater affmity for carbon monoxide (CO) than for oxygen. It is obvious, therefore, that cigarette smoking during pregnancy dimin- ishes the oxygen carrying capacity of both fetal and maternal blood. This affects maternal oxygenation by increased pulmonary venous admixture and diminishes the oxygen available to the fetus at the tissue level by its effect on fetal oxyhemoglobin dissociation. In a 1975 report by Dow, Rooney, and Spence (1 1), a signifi- cantly greater rise in COHb concentration in response to smoking a single cigarette was shown in pregnant women (3.9 percent increase) as opposed to nonpregnant women (2.1 percent increase). This was more pronounced when anemia was present (5.0 percent increase) and appeared to be inversely related to the hemoglobin concentra- tion. Three groups of women, all smokers, were selected for this study. The first group consisted of 10 normal, pregnant women late in the second trimester of pregnancy, with hemoglobin levels of over 11 g per 100 milliliters (ml). The second group consisted of 10 women also late in the second trimester but whose hemoglobin levels were less than 10 g/ 100 ml. Apart from anemia at the time of admission to the study, these patients were normal, The third group consisted of 10 normal, nonpregnant women with normal hemoglo- bin levels (over 11 g/l00 ml). The change in COHb was estimated spectrophotometrically in response to smoking the first cigarette of the morning, the women having rested for at least 30 minutes. A sample of venous blood was withdrawn before and 2 minutes after smoking the cigarette. The cigarettes were of a standard size and of a "non-mild" (i.e., not low "tar" and nicotine) variety. The women were instructed to take a puff every 40 seconds, inhaling as deeply as possible, to a total of 10 puffs. In the nonpregnant group, the mean rise in COHb concentration (&standard error of mean) was 2.1kO.2 percent. A significantly greater increase was found in the normal pregnant group (mean rise 3.950.4 percent; t=3.91; p