- Email: [email protected]
Smoking in pregnancy, exhaled carbon monoxide, and birth weight
Smoking in Pregnancy, Exhaled Carbon Monoxide, and Birth Weight ROGER H. SECKER-WALKER, MB, PAMELA M. VACEK, PhD, BRIAN S. FLYNN, SeD, AND PHILIP B. MEAD, MD Objective: To examine the relation of cigarette consumption and exhaled carbon monoxide levels during pregnancy and to assess the effect of these smoking measures on birth weight. Methods: Cigarette consumption and exhaled carbon mon• oxide levels were recorded at the first prenatal visit and the 36-week visit from women who smoked early in pregnancy. Analysis of variance was used to compare birth weights for differing levels of cigarette consumption and exhaled carbon monoxide. Correlation and regression analyses were used to estimate the effects of the smoking measures at both prena• tal visits on birth weight. Results: Cigarette consumption and exhaled carbon mon• oxide levels at both visits were associated significantly with birth weight. After the first prenatal visit, a reduction in cigarette consumption of at least nine cigarettes per day or in exhaled carbon monoxide of 8 parts per million (ppm) was associated with gains in birth weight of 100 g or more. The proportion of low birth weight (LBW) infants increased significantly with increasing levels of cigarette consumption and with increasing concentrations of exhaled carbon mon• oxide. Conclusion: Substantial reductions in cigarette consump• tion or in exhaled carbon monoxide levels after the first prenatal visit are needed to achieve gains in birth weight. Not smoking, or having an exhaled carbon monoxide level less than 5 ppm minimizes the likelihood of having an LBW infant. (Obstet Gynecol 1997;89:648-53. © 1997 by The American College of Obstetricians and Gynecologists.)
Cigarette smoking has a small but important effect on birth weight, recognized for more than 3 decades. 1,2 Among factors known to influence birth weight, smok• ing is one of the few that can be changed.3-S Carbon monoxide, by binding to fetal hemoglobin and reducing the availability of oxygen, is thought to be an important From the Office of Health Promotion Research, the Biometry Facility, the Departments ofFamily Practice and Obstetrics and Gynecology, and the Vermont Cancer Center, Burlington, Vermont. Support for this study was provided by the National Institutes of Health, grant numbers HL29957 and CA22435.
648 0029-7844/97/$17.00
PH 50029-7844(97)00103-8
cause of the fetal growth restriction associated with cigarette smoking.6 Carbon monoxide, which has a half-life of 1-4 hours, can be measured readily in exhaled air,7,8 and this measure has been used in several smoking-eessation interventions in pregnancy.9-14 Quitting smoking has been a major goal of efforts to change pregnant women's smoking behavior, but the majority of pregnant smokers do not quit. 15 Reducing cigarette consumption is a more achievable goal. Al• though this is an important behavioral change in the process of quitting smoking,16 it may not be accompa• nied by any advantages in fetal growth because smok• ers who reduce their cigarette consumption tend to maintain carbon monoxide concentrations close to their original levelsP MacArthur and Knox18 reported no impact of reduc• tions in maternal smoking on birth weight, whereas Hebel et al19 reported that only reductions to fewer than five cigarettes per day were accompanied by an increase in infant birth weight. However, in the Birmingham Trial 11,20,21 infant birth weights among women who did not quit smoking were higher for those who had at least a 50% reduction in salivary cotinine concentration dur• ing pregnancy than for those whose salivary cotinine was not reduced this much. We report our observations of exhaled carbon mon• oxide concentrations measured during pregnancy in relation to women's reported cigarette consumption and to their infants' birth weights. We were particularly interested in examining the extent to which reductions in cigarette consumption or in exhaled carbon monox• ide were accompanied by increases in birth weight.
Materials and Methods To provide a personalized rationale to quit smoking and to monitor changes in self-reported smoking be• havior, we measured exhaled carbon monoxide levels at selected prenatal visits during a randomized con-
Obstetrics & Gynecology
trolled trial of smoking cessation and relapse preven• tion advice during pregnancy. The trial took place between October 1988 and October 1992 at University Associates in Obstetrics and Gynecology's Maternal Infant Care clinic, a state-supported clinic for under• served women, and in the adolescent clinic. Five hun• dred twenty-one women who were smoking early in pregnancy participated in this trial. For this study of the relation between cigarette con• sumption, exhaled carbon monoxide, and birth weight, we focused on the 392 women in that trial who had singleton births for which birth weight information was available. At the first prenatal visit, details of smoking behavior, measurement of exhaled carbon monoxide concentration, and demographic and obstetric informa• tion were obtained. Smoking behavior and exhaled carbon monoxide were assessed again at the 36-week visit. Birth weight and gestational age at the time of delivery, estimated from the date of the last menstrual period or, when this was uncertain, from fetal ultra• sound measurements, were abstracted from the labor and delivery log or from the hospital discharge sum• mary. Exhaled carbon monoxide concentration in parts per million (ppm) was measured by trained personnel using Vitalograph Carbon Monoxide Monitors (Vitalo• graph Inc., Lenexa, KS). These monitors are small, hand-held, battery-operated devices with disposable mouthpieces. Women were instructed to take a deep breath and then to exhale through the mouthpiece as completely as possible. The carbon monoxide concen• tration was read from the monitor after 30 seconds. Table 1 presents the demographic characteristics, smoking behavior, and exhaled carbon monoxide levels of the 392 women in this study. Their average age was about 22 years; only 15% had more than a high school education; 47% were primigravid. Less than 2% were non-white. The demographic characteristics, smoking behavior, and exhaled carbon monoxide levels of these women did not differ significantly from those of the other women enrolled in the controlled trial, except that they started smoking about 1 year later and entered prenatal care about 1.5 weeks later. Three hundred one women reported smoking one or more cigarettes per day at their first prenatal visit, whereas 91 (23%) reported having quit since knowing they were pregnant. Among the women who had quit, the reported time since their last cigarette ranged from 14 hours to 247 days, with a median of 30 days. Data were available on 354 women at their 36-week visit, 244 of whom were smokers and 110 who reported not smoking. Four of these women did not have carbon monoxide measurements at this visit. Fifteen women delivered before the 36-week visit, and the remaining 23
VOL. 89, NO.5, PART 1, MAY 1997
Table 1. Demographic and Other Characteristics at the First Prenatal Visit Characteristic
n = 392
Age (y) Education < High school High school > High school Type of care Maternal Infant Care clinic Adolescent clinic Primigravid Weeks pregnant at first visit Age started smoking (y) Smokers (n = 301) Cigarettes per day Exhaled carbon monoxide (ppm) Quitters (n = 91) Exhaled carbon monoxide (ppm)
22.3 :t 5.0
164 (42%) 170 (43%) 58(15%) 343 (87%) 49 (13%) 179 (46%) 15.2 :t 7.5 14.8:t 3.0 12.7:t 6.9 16.4:t 9.6 4.2:t 4.3
ppm = parts per million. Data are presented as mean :t standard deviation or n (%).
women who lacked 36-week data had transferred to other obstetric practices or voluntarily dropped out of the study. These 38 women differed neither demo• graphically nor in their cigarette consumption or ex• haled carbon monoxide levels at the first visit from the women who were assessed at the 36-week visit. Analysis of variance was used to compare birth weights in groups based on cigarette consumption or exhaled carbon monoxide.· Relation between each smoking measure and birth weight were estimated by Pearson correlations. Regression analysis was used to estimate the effects on birth weight of smoking levels at both the first prenatal and 36-week visits, as measured by either self-reported consumption or exhaled carbon monoxide concentration. Trends in the proportion of low birth weight (LBW) infants for differing levels of smoking and exhaled carbon monoxide were assessed 'using the Mantel-Haenszel test for linear associations. We used P :::; .05 as the level of significance.
Results Exhaled carbon monoxide levels increased with increas• ing cigarette consumption: The correlation between these two measures was 0.57 at the first visit (P < .001) and 0.66 at the 36-week visit (P < .001). Between the first and 36-week visits, 35% of the women reported reducing their cigarette consumption, 23% reported increasing theirs, and 42% reported no change. There were significant correlations between the amounts smoked at the two time points: r = 0.66, P < .001 for cigarettes per day; and r = 0.64, P < .001 for exhaled carbon monoxide.
Seeker-Walker et al
Carbon Monoxide and Birth Weight
649
Table 2. Cigarette Consumption and Exhaled Carbon
Monoxide at First and 36-Week Visits, and Birth Weight Birth weight (g)
First visit Smoking measure
36-wk visit
n
Mean
110
13
3500 ~ 523" 3311 ~ 537 3123 ~ 61Q§ 3151 ~ 554§ 3051 ~ 511§
3511 ~ 3265 ~ 3230 ~ 3239 ~ 2800 ~
93 72 125 102
3470 ~ 5271 3239 ~ 617"" 3187 ~ 537"" 3107 ~ 591""
111 65 112 62
n
Mean
91
~
SO
Cigarettes per day 0 1-5 6-10 11-20 >20
53 110 125
58 86 87 13
Exhaled carbon monoxide (ppm) <5 5-9 10-19 ~20
~
SO
441 t 474* 515* 387* 645"
3486 ~ 469" 3362 ~ 493 3241 ~ 478 tt 3058 ~ 376tt
SO = standard deviation; ppm = parts per million. For each set of means, groups with different symbols are signifi• cantly different (P < .05) using Student-Newman-Keuls procedure to adjust for multiple comparisons. Groups with the same symbol are not significantly different from each other or from the groups not denoted by a symbol.
There were no significant relations between cigarette consumption at the first or 36-week visit and gestational age at delivery, but exhaled carbon monoxide and gestational age were correlated weakly: r = -0.14, P < .01 for the first visit; and r = -0.14, P = .01 for the 36-week visit. On average, women with exhaled carbon monoxide levels of 20 ppm or more at both visits delivered about 5 days earlier than those who reported not smoking. The mean (± standard deviation) birth weight of singleton infants was 3243 ± 578 g for women with first-visit data and 3310 :!:: 486 g for women with 36-week data. Table 2 shows the mean birth weights for different levels of cigarette consumption and for differ• ent levels of exhaled carbon monoxide at the first and 36-week visits. Women who reported not smoking at the first visit had infants that were significantly heavier than the infants of mothers who were smoking more than five cigarettes per day at this visit. At the 36-week visit, women who reported not smoking had infants that were significantly heavier than the infants whose mothers were still smoking at this visit, regardless of the level of smoking. Women who reported smoking more than 20 cigarettes per day had infants with the lowest birth weights, although the power to detect statistically significant differences was limited because of the small number of women in this group. Only differences of 142 g or more between this group and women who smoked 11-20 cigarettes per day at the first visit had an 80% chance of detection.
650 Seeker-Walker et al
Carbon Monoxide and Birth Weight
Women with exhaled carbon monoxide concentra• tions less than 5 ppm at the first visit had infants that were significantly heavier than the infants of women with higher levels. The mean birth weights for other concentrations of exhaled carbon monoxide did not differ significantly from one another. At the 36-week visit, women with exhaled carbon monoxide concentra• tions less than 5 ppm had infants that were significantly heavier than the infants of women whose levels were 10 ppm or higher, but were not different from the infants of women with levels of 5-9 ppm. Both of the smoking measures showed a significant linear relation with birth weight. The correlations be• tween birth weight and cigarette consumption were r = -0.23 (n = 392) and r = -0.28 (n = 354) for the first and 36-week visits, respectively. The correlation between birth weight and exhaled carbon monoxide concentra• tion was r = -0.22 (n = 392) and r = -0.32 (n = 350) for the first and 36-week visits, respectively. All of these correlations were highly significant (P < .001), but these smoking measures explained only 5-10% of the vari• ability in birth weight. Adjustment for gestational age reduced the correlation coefficient for carbon monoxide slightly at the first visit (r = -0.18) but had little effect on the other coefficients. The two smoking measures did not differ significantly in their ability to predict birth weight at either the first or the 36-week visit. Using the smoking measures at both visits to predict birth weight yielded the following regression equations. For cigarette (cig) consumption: Birth weight = 3481 - 7.8 (cigs/ day, visit 1) - 11.4 (cigs/ day, 36 weeks), r
= 0.09, P < .001;
and for exhaled carbon monoxide (CO) level: Birth weight
= 3528 -
4.5 (CO, visit 1)
- 13.5 (CO, 36 weeks), r=O.l1, P < .001. The coefficients for the smoking measures at 36 weeks were significant in both equations (P < .01), but those at visit 1 were not significant in either. However, these first-visit coefficients were also not significantly differ• ent from the 36-week coefficients. When derived sepa• rately for primigravid and multigravid women, the coefficients for cigarettes per day at visit 1 and 36 weeks and for exhaled carbon monoxide at visit 1 and 36 weeks did not differ significantly from those in either equations. The equation for cigarette consumption shows that reducing consumption by nine cigarettes per day would increase birth weight by 11.4 x 9 = 103 g; the equation for carbon monoxide levels shows that reducing exhaled carbon monoxide by 8 ppm would increase birth weight by 13.5 X 8 = 108 g.
Obstetrics & Gynecology
Gestational age accounted for 21.5% of the variability in birth weight in these women, but including gesta• tional age in these equations scarcely changed the regression coefficients for cigarettes per day for either visit. When gestational age (GA) was included, the equation for cigarette (cig) consumption was: Birth weight = (137 X GA) - 1954 - 7.4 (cigs/ day, visit 1) - 9.1 (cigs/ day, 36 weeks), r = 0.29, P < .001. For exhaled carbon monoxide, the coefficient for the first visit was reduced, but that for 36 weeks changed very little. With gestational age included, the equation for exhaled carbon monoxide (CO) was: Birth weight = (130
X
GA) - 1664
- 1.7 (CO, visit 1) - 12.4 (CO, 36 weeks), r
= 0.28, P < .001.
These adjustments for gestational age only slightly modified the effects of changing cigarette consumption. For example, the equation for cigarette consumption shows that reducing consumption by nine cigarettes per day would increase birth weight by 9.1 X 9 = 82 g, 21 g less than without the adjustment. The equation for carbon monoxide level shows that reducing exhaled carbon monoxide by 8 ppm would increase birth weight by 12.4 X 8 = 99 g, 9 g less than without the adjustment. Interestingly, in the equation for carbon monoxide level and gestational age, the coefficient for exhaled carbon monoxide at the first visit is significantly different from that at the 36-week visit (P = .05), which implies that for a fixed gestational age, smoking later in pregnancy has a greater effect on birth weight. The percentage of infants with LBW among the 392 women having data on both smoking measures for their first prenatal visit was 8.4%, and for the 354 women with both measures at their 36-week visit was 5.4%. Table 3 shows the percentage of mothers having LBW infants for different levels of cigarette consumption and for different levels of exhaled carbon monoxide at these visits. There were highly significant associations be• tween increasing proportions of LBW infants and in• creasing cigarette consumption, and with increasing levels of exhaled carbon monoxide at both visits. Simi• lar trends were seen for the proportions of LBW infants of primigravid women and of multigravid women in relation to increasing levels of cigarette consumption and exhaled carbon monoxide. However, the trends for the 36-week smoking measures among the primigrav• ids, for which there were only eight LBW infants, were not statistically significant.
VOL. 89, NO.5, PART 1, MAY 1997
Table 3. Cigarette Consumption and Exhaled Carbon Monoxide at First and 36-Week Visits, and Low Birth Weight Mothers having L8W infants 36-wk visit
First visit Smoking measure Cigarettes per day 0 1-5 6-10 ~11
Exhaled carbon monoxide (ppm) <5 5--9 10-19 ~20
Total births
No. <2500 g
%
1.1 7.5 10.9 11.6"
110 86 100
0 3 7 9
0.0 5.2 8.1 9.0t
2.2 5.6 8.8 15.7*
111 65 112 62
2 3 8 6
1.8 4.6 7.1 9.7S
Total births
No. <2500 g
%
91 53 110 138
1 4 12 16
93
2 4 11 16
72
125 102
58
LBW = low birth weight; ppm = parts per million. Mantel-Haenszel test for linear associations: "P == .004. t P == .002. P < .001. S P == .017.
*
Discussion We found that levels of smoking at the first prenatal and 36-week visits, whether measured in terms of self-reported cigarettes per day or exhaled carbon mon• oxide concentration, were related significantly to birth weight and to increases in the chance of having an LBW infant. Gestational age at delivery was weakly related to exhaled carbon monoxide level, but not to cigarette consumption. The reduction in birth weight and the increase in the proportion of LBW infants in relation to increasing levels of cigarette consumption are compa• rable to those reported by others. 6,22 Our results indicate that for pregnant women who are current smokers at their first prenatal visit, their infants' potential birth weights are already compro• mised compared with those who were smokers at the time they learned they were pregnant but then quit. Birth weights are further compromised by subsequent levels of smoking. However, the extent of this further compromise is lessened for women who reduce their cigarette consumption, compared with continued smoking at the same level. The reductions in birth weight per cigarette smoked per day-about 8 g for the first visit and 11 g for the 36-week visit-are compara• ble to the estimate of Mathai et al 23 of 12 g. Haddow et al,22 in a study of 4211 pregnant women, reported similar correlations between cigarette consumption at 15-21 weeks' gestation and birth weight (-0.19) and between serum cotinine, a metabolite of nicotine, and birth weight (-0.23). They found in this large sample
Seeker-Walker et al
Carbon Monoxide and Birth Weight
651
that serum cotinine was a significantly better predictor of birth weight than was cigarette consumption. We are not aware of other reports of the relation between exhaled carbon monoxide and birth weight, but the reduction in birth weight we observed is com• parable to that reported in relation to increasing con• centrations of cord blood carboxyhemoglobin.24 As seen in the equation using carbon monoxide level and ges• tational age, the influence of smoking late in pregnancy, as judged by exhaled carbon monoxide levels at the 36-week visit, appears to have a greater effect on birth weight than carbon monoxide levels at the first prenatal visit. This is consistent with the observation that women who quit smoking before 20 weeks' gestation have infants whose birth weights are comparable to those of non-smokers' infants. u8 Many factors .other than smoking are known to influence the birth weight of singleton infants, includ• ing gestational age at delivery, gravidity, maternal age, height and weight, maternal weight gain during preg• nancy, substance use, ethnicity, social and economic factors, and infant gender. 2,3,25 Of these, only smoking, use of other substances, and maternal weight gain during pregnancy can be altered by interventions. Our estimates for the proportion of the variance due to cigarette consumption, unadjusted for these other pre• dictors, were 5% for the first visit and 8% at the 36-week visit. These estimates are larger than the adjusted esti• mates of Hebel et al19 for cigarette consumption mea• sured early in pregnancy, at 2%, or late in pregnancy, at 4%. The differences may reflect the correlation between cigarette consumption and the factors for which Hebel et al19 adjusted, particularly education and maternal weight gain. Our study clearly confirms the detrimental effect of smoking on birth weight and provides estimates of the size of that effect, whether determined from reported cigarette consumption or exhaled carbon monoxide. Reductions in cigarette consumption or in exhaled carbon monoxide levels must be substantial if gains in birth weight of 100 g or more are to be achieved. The higher average infant birth weights of women who stopped smoking either before or after their first prena• tal visit, or of those who had exhaled carbon monoxide levels less than 5 ppm, argue strongly that quitting smoking is the optimal behavior change to maximize birth weight and minimi~e the likelihood of having an LBW infant.
22.
References
23.
1. Simpson WJ. A preliminary report on cigarette smoking and the incidence of prematurity. Am J Obstet GynecoI1957;73:808-15. 2. Butler NR, Goldstein H, Ross EM. Cigarette smoking in pregnancy:
652 Seeker-Walker et al
Carbon Monoxide and Birth Weight
3.
4. 5.
6. 7. 8. 9.
10.
11.
12.
13.
14.
15. 16.
17.
18. 19.
20.
21.
24.
Its influence on birth weight and perinatal mortality. BMJ 1972;2: 127-30. Brooke OG, Anderson HR, Bland JM, Peacock JL, Stewart CM. Effects on birth weight of smoking, alcohol, caffeine, socioeconomic factors, and psychosocial stress. BMJ 1989;298:795-801. Sexton M, Hebel R. A clinical trial of change in maternal smoking and its effect on birth weight. JAMA 1984;251:911-5. MacArthur Co Newton JR, Knox EG. Effect of anti-smoking health education on infant size at birth: A randomized trial. Br J Obstet GynaecoI1987;94:295-300. Abel EL. Smoking during pregnancy: A review of effects on growth and development of offspring. Hum Bioi 1980;52:593-625. Frederiksen L, Martin JE. Carbon monoxide and smoking behav• ior. Addict Behav 1979;4:21-30. Jarvis MJ, Russell MAH, Saloojee Y. Expired air carbon monoxide: A simple breath test of tobacco smoke intake. BMJ 1980;281:484-5. Hughes JR, Epstein LH, Andrasik F, Neff DF, Thompson DS. Smoking and carbon monoxide levels during pregnancy. Addict Behav 1982;7:271-6. Bauman KE, Bryan ES, Dent CW, Koch GG. The influence of observing carbon monoxide level on cigarette smoking by public prenatal patients. Am J Public Health 1983;73:1089-91. Burling TA, Bigelow GE, Robinson JC, Mead AM. Smoking dUring pregnancy: Reduction via objective assessment and directive ad• vice. Behav Ther 1991;22:31-40. Price JH, Krol RA, Desmond SM, Losh DP, Roberts SM, Snyder FF. Comparison of three antismoking interventions among pregnant women in an urban setting: A randomized trial. Psychol Rep 1991;68:595-604. Rush D, Orme J, King J, Eiser JR, Butler NR. A trial of health education aimed to reduce cigarette smoking among pregnant women. Paediatr Perinat Epidemiol 1992;6:285-97. Hartmann KE, Thorp 1M, Pahel-Short L, Koch MA. A randomized controlled trial of smoking cessation intervention in pregnancy in an academic clinic. Obstet Gynecol 1996;87:621-6. Fingerhut LA, Kleinman JC, Kendrick JS. Smoking before, during and after pregnancy. Am J Public Health 1990;80:541-4. Prochaska JO, DiClemente Cc. Stages and processes of self-change of smoking: Toward an integrative model of change. J Consult C1in Psychol 1983;51:390-5. Benowitz NL, Jacob P, Kozlowski LT, Yu L. 1nfluence of smoking fewer cigarettes on exposure to tar, nicotine and carbon monoxide. N Engl J Med 1986;315:1310-3. MacArthur C, Knox EG. Smoking in pregnancy: Effects of stopping at different stages. Br J Obstet GynaecoI1988;95:551-5. Hebel JR, Fox NL, Sexton M. Dose-response of birth weight to various measures of maternal smoking during pregnancy. J Clin Epidemiol 1988;41:483-9. Windsor RA, Lowe ]B, Perkins LL. Behavioral impact and cost benefit of health education methods for pregnant smokers in public health maternity clinics: The Birmingham Trial II. Am J Public Health 1993;83:201-6. Li CQ Windsor RA, Perkins RL, Goldenburg RL, Lowe JB. The impact on infant birthweight and gestational age of cotinine• validated smoking reduction during pregnancy. JAMA 1993;269: 1519-24. Haddow JE, Knight GJ, Palomaki GE, Kloza EM, Wald NJ. Ciga• rette consumption and serum cotinine in relation to birthweight. Br J Obstet Gynaecol 1987;94:678-81. Mathai M, Skinner A, Lawton K, Weindling AM. Maternal smok• ing, urinary cotinine levels and birthweight. Aust N Z J Obstet Gynaecol 1990;30:33-6. Wouters EJ, de Jong PA, Cornelissen PJH, Kurver PHJ, van Gel WC, van Woensel CLM. Smoking and low birth weight: Absence
Obstetrics & Gynecology
of Influence by carbon monoxide. Eur J Obstet Gynecol Reprod Bioi 1987;25:3~41. 25. Showstack JA, Budetti PP, Minkler D. Factors associated with birthweight: An exploration of the roles of prenatal care and length of gestation. Am J Public Health 1984;74:1003-8.
Received August 21, 1996. Received in revised form November 26, 1996. Accepted January 15, 1997.
Address reprint requests to:
Roger H. Seeker-Walker, MB 1 South Prospect Street Burlington, VT 05401
Copyright Cl 1997 by The American College of Obstetricians and Gynecologists. Published by Elsevier Science Inc.
STANDARDS FOR REPORTING TRIALS The CONSORT listing of standards for reporting randomized trials has been adopted as policy by Obstetrics & Gynecology. Investigators who are planning, conducting, or reporting randomized trials should be thoroughly familiar with these standards. A copy can be obtained by contacting: Obstetrics & Gynecology, 1100 Glendon Avenue, Suite 1655, Los Angeles, CA 90024-3520; FAX (310) 208·2838.
VOL. 89, NO.5, PART 1, MAY 1997
Secker-Walker et al
Carbon Monoxide and Birth Weight
653
Cigarette smoking has a small but important effect on birth weight, recognized for more than 3 decades. 1,2 Among factors known to influence birth weight, smok• ing is one of the few that can be changed.3-S Carbon monoxide, by binding to fetal hemoglobin and reducing the availability of oxygen, is thought to be an important From the Office of Health Promotion Research, the Biometry Facility, the Departments ofFamily Practice and Obstetrics and Gynecology, and the Vermont Cancer Center, Burlington, Vermont. Support for this study was provided by the National Institutes of Health, grant numbers HL29957 and CA22435.
648 0029-7844/97/$17.00
PH 50029-7844(97)00103-8
cause of the fetal growth restriction associated with cigarette smoking.6 Carbon monoxide, which has a half-life of 1-4 hours, can be measured readily in exhaled air,7,8 and this measure has been used in several smoking-eessation interventions in pregnancy.9-14 Quitting smoking has been a major goal of efforts to change pregnant women's smoking behavior, but the majority of pregnant smokers do not quit. 15 Reducing cigarette consumption is a more achievable goal. Al• though this is an important behavioral change in the process of quitting smoking,16 it may not be accompa• nied by any advantages in fetal growth because smok• ers who reduce their cigarette consumption tend to maintain carbon monoxide concentrations close to their original levelsP MacArthur and Knox18 reported no impact of reduc• tions in maternal smoking on birth weight, whereas Hebel et al19 reported that only reductions to fewer than five cigarettes per day were accompanied by an increase in infant birth weight. However, in the Birmingham Trial 11,20,21 infant birth weights among women who did not quit smoking were higher for those who had at least a 50% reduction in salivary cotinine concentration dur• ing pregnancy than for those whose salivary cotinine was not reduced this much. We report our observations of exhaled carbon mon• oxide concentrations measured during pregnancy in relation to women's reported cigarette consumption and to their infants' birth weights. We were particularly interested in examining the extent to which reductions in cigarette consumption or in exhaled carbon monox• ide were accompanied by increases in birth weight.
Materials and Methods To provide a personalized rationale to quit smoking and to monitor changes in self-reported smoking be• havior, we measured exhaled carbon monoxide levels at selected prenatal visits during a randomized con-
Obstetrics & Gynecology
trolled trial of smoking cessation and relapse preven• tion advice during pregnancy. The trial took place between October 1988 and October 1992 at University Associates in Obstetrics and Gynecology's Maternal Infant Care clinic, a state-supported clinic for under• served women, and in the adolescent clinic. Five hun• dred twenty-one women who were smoking early in pregnancy participated in this trial. For this study of the relation between cigarette con• sumption, exhaled carbon monoxide, and birth weight, we focused on the 392 women in that trial who had singleton births for which birth weight information was available. At the first prenatal visit, details of smoking behavior, measurement of exhaled carbon monoxide concentration, and demographic and obstetric informa• tion were obtained. Smoking behavior and exhaled carbon monoxide were assessed again at the 36-week visit. Birth weight and gestational age at the time of delivery, estimated from the date of the last menstrual period or, when this was uncertain, from fetal ultra• sound measurements, were abstracted from the labor and delivery log or from the hospital discharge sum• mary. Exhaled carbon monoxide concentration in parts per million (ppm) was measured by trained personnel using Vitalograph Carbon Monoxide Monitors (Vitalo• graph Inc., Lenexa, KS). These monitors are small, hand-held, battery-operated devices with disposable mouthpieces. Women were instructed to take a deep breath and then to exhale through the mouthpiece as completely as possible. The carbon monoxide concen• tration was read from the monitor after 30 seconds. Table 1 presents the demographic characteristics, smoking behavior, and exhaled carbon monoxide levels of the 392 women in this study. Their average age was about 22 years; only 15% had more than a high school education; 47% were primigravid. Less than 2% were non-white. The demographic characteristics, smoking behavior, and exhaled carbon monoxide levels of these women did not differ significantly from those of the other women enrolled in the controlled trial, except that they started smoking about 1 year later and entered prenatal care about 1.5 weeks later. Three hundred one women reported smoking one or more cigarettes per day at their first prenatal visit, whereas 91 (23%) reported having quit since knowing they were pregnant. Among the women who had quit, the reported time since their last cigarette ranged from 14 hours to 247 days, with a median of 30 days. Data were available on 354 women at their 36-week visit, 244 of whom were smokers and 110 who reported not smoking. Four of these women did not have carbon monoxide measurements at this visit. Fifteen women delivered before the 36-week visit, and the remaining 23
VOL. 89, NO.5, PART 1, MAY 1997
Table 1. Demographic and Other Characteristics at the First Prenatal Visit Characteristic
n = 392
Age (y) Education < High school High school > High school Type of care Maternal Infant Care clinic Adolescent clinic Primigravid Weeks pregnant at first visit Age started smoking (y) Smokers (n = 301) Cigarettes per day Exhaled carbon monoxide (ppm) Quitters (n = 91) Exhaled carbon monoxide (ppm)
22.3 :t 5.0
164 (42%) 170 (43%) 58(15%) 343 (87%) 49 (13%) 179 (46%) 15.2 :t 7.5 14.8:t 3.0 12.7:t 6.9 16.4:t 9.6 4.2:t 4.3
ppm = parts per million. Data are presented as mean :t standard deviation or n (%).
women who lacked 36-week data had transferred to other obstetric practices or voluntarily dropped out of the study. These 38 women differed neither demo• graphically nor in their cigarette consumption or ex• haled carbon monoxide levels at the first visit from the women who were assessed at the 36-week visit. Analysis of variance was used to compare birth weights in groups based on cigarette consumption or exhaled carbon monoxide.· Relation between each smoking measure and birth weight were estimated by Pearson correlations. Regression analysis was used to estimate the effects on birth weight of smoking levels at both the first prenatal and 36-week visits, as measured by either self-reported consumption or exhaled carbon monoxide concentration. Trends in the proportion of low birth weight (LBW) infants for differing levels of smoking and exhaled carbon monoxide were assessed 'using the Mantel-Haenszel test for linear associations. We used P :::; .05 as the level of significance.
Results Exhaled carbon monoxide levels increased with increas• ing cigarette consumption: The correlation between these two measures was 0.57 at the first visit (P < .001) and 0.66 at the 36-week visit (P < .001). Between the first and 36-week visits, 35% of the women reported reducing their cigarette consumption, 23% reported increasing theirs, and 42% reported no change. There were significant correlations between the amounts smoked at the two time points: r = 0.66, P < .001 for cigarettes per day; and r = 0.64, P < .001 for exhaled carbon monoxide.
Seeker-Walker et al
Carbon Monoxide and Birth Weight
649
Table 2. Cigarette Consumption and Exhaled Carbon
Monoxide at First and 36-Week Visits, and Birth Weight Birth weight (g)
First visit Smoking measure
36-wk visit
n
Mean
110
13
3500 ~ 523" 3311 ~ 537 3123 ~ 61Q§ 3151 ~ 554§ 3051 ~ 511§
3511 ~ 3265 ~ 3230 ~ 3239 ~ 2800 ~
93 72 125 102
3470 ~ 5271 3239 ~ 617"" 3187 ~ 537"" 3107 ~ 591""
111 65 112 62
n
Mean
91
~
SO
Cigarettes per day 0 1-5 6-10 11-20 >20
53 110 125
58 86 87 13
Exhaled carbon monoxide (ppm) <5 5-9 10-19 ~20
~
SO
441 t 474* 515* 387* 645"
3486 ~ 469" 3362 ~ 493 3241 ~ 478 tt 3058 ~ 376tt
SO = standard deviation; ppm = parts per million. For each set of means, groups with different symbols are signifi• cantly different (P < .05) using Student-Newman-Keuls procedure to adjust for multiple comparisons. Groups with the same symbol are not significantly different from each other or from the groups not denoted by a symbol.
There were no significant relations between cigarette consumption at the first or 36-week visit and gestational age at delivery, but exhaled carbon monoxide and gestational age were correlated weakly: r = -0.14, P < .01 for the first visit; and r = -0.14, P = .01 for the 36-week visit. On average, women with exhaled carbon monoxide levels of 20 ppm or more at both visits delivered about 5 days earlier than those who reported not smoking. The mean (± standard deviation) birth weight of singleton infants was 3243 ± 578 g for women with first-visit data and 3310 :!:: 486 g for women with 36-week data. Table 2 shows the mean birth weights for different levels of cigarette consumption and for differ• ent levels of exhaled carbon monoxide at the first and 36-week visits. Women who reported not smoking at the first visit had infants that were significantly heavier than the infants of mothers who were smoking more than five cigarettes per day at this visit. At the 36-week visit, women who reported not smoking had infants that were significantly heavier than the infants whose mothers were still smoking at this visit, regardless of the level of smoking. Women who reported smoking more than 20 cigarettes per day had infants with the lowest birth weights, although the power to detect statistically significant differences was limited because of the small number of women in this group. Only differences of 142 g or more between this group and women who smoked 11-20 cigarettes per day at the first visit had an 80% chance of detection.
650 Seeker-Walker et al
Carbon Monoxide and Birth Weight
Women with exhaled carbon monoxide concentra• tions less than 5 ppm at the first visit had infants that were significantly heavier than the infants of women with higher levels. The mean birth weights for other concentrations of exhaled carbon monoxide did not differ significantly from one another. At the 36-week visit, women with exhaled carbon monoxide concentra• tions less than 5 ppm had infants that were significantly heavier than the infants of women whose levels were 10 ppm or higher, but were not different from the infants of women with levels of 5-9 ppm. Both of the smoking measures showed a significant linear relation with birth weight. The correlations be• tween birth weight and cigarette consumption were r = -0.23 (n = 392) and r = -0.28 (n = 354) for the first and 36-week visits, respectively. The correlation between birth weight and exhaled carbon monoxide concentra• tion was r = -0.22 (n = 392) and r = -0.32 (n = 350) for the first and 36-week visits, respectively. All of these correlations were highly significant (P < .001), but these smoking measures explained only 5-10% of the vari• ability in birth weight. Adjustment for gestational age reduced the correlation coefficient for carbon monoxide slightly at the first visit (r = -0.18) but had little effect on the other coefficients. The two smoking measures did not differ significantly in their ability to predict birth weight at either the first or the 36-week visit. Using the smoking measures at both visits to predict birth weight yielded the following regression equations. For cigarette (cig) consumption: Birth weight = 3481 - 7.8 (cigs/ day, visit 1) - 11.4 (cigs/ day, 36 weeks), r
= 0.09, P < .001;
and for exhaled carbon monoxide (CO) level: Birth weight
= 3528 -
4.5 (CO, visit 1)
- 13.5 (CO, 36 weeks), r=O.l1, P < .001. The coefficients for the smoking measures at 36 weeks were significant in both equations (P < .01), but those at visit 1 were not significant in either. However, these first-visit coefficients were also not significantly differ• ent from the 36-week coefficients. When derived sepa• rately for primigravid and multigravid women, the coefficients for cigarettes per day at visit 1 and 36 weeks and for exhaled carbon monoxide at visit 1 and 36 weeks did not differ significantly from those in either equations. The equation for cigarette consumption shows that reducing consumption by nine cigarettes per day would increase birth weight by 11.4 x 9 = 103 g; the equation for carbon monoxide levels shows that reducing exhaled carbon monoxide by 8 ppm would increase birth weight by 13.5 X 8 = 108 g.
Obstetrics & Gynecology
Gestational age accounted for 21.5% of the variability in birth weight in these women, but including gesta• tional age in these equations scarcely changed the regression coefficients for cigarettes per day for either visit. When gestational age (GA) was included, the equation for cigarette (cig) consumption was: Birth weight = (137 X GA) - 1954 - 7.4 (cigs/ day, visit 1) - 9.1 (cigs/ day, 36 weeks), r = 0.29, P < .001. For exhaled carbon monoxide, the coefficient for the first visit was reduced, but that for 36 weeks changed very little. With gestational age included, the equation for exhaled carbon monoxide (CO) was: Birth weight = (130
X
GA) - 1664
- 1.7 (CO, visit 1) - 12.4 (CO, 36 weeks), r
= 0.28, P < .001.
These adjustments for gestational age only slightly modified the effects of changing cigarette consumption. For example, the equation for cigarette consumption shows that reducing consumption by nine cigarettes per day would increase birth weight by 9.1 X 9 = 82 g, 21 g less than without the adjustment. The equation for carbon monoxide level shows that reducing exhaled carbon monoxide by 8 ppm would increase birth weight by 12.4 X 8 = 99 g, 9 g less than without the adjustment. Interestingly, in the equation for carbon monoxide level and gestational age, the coefficient for exhaled carbon monoxide at the first visit is significantly different from that at the 36-week visit (P = .05), which implies that for a fixed gestational age, smoking later in pregnancy has a greater effect on birth weight. The percentage of infants with LBW among the 392 women having data on both smoking measures for their first prenatal visit was 8.4%, and for the 354 women with both measures at their 36-week visit was 5.4%. Table 3 shows the percentage of mothers having LBW infants for different levels of cigarette consumption and for different levels of exhaled carbon monoxide at these visits. There were highly significant associations be• tween increasing proportions of LBW infants and in• creasing cigarette consumption, and with increasing levels of exhaled carbon monoxide at both visits. Simi• lar trends were seen for the proportions of LBW infants of primigravid women and of multigravid women in relation to increasing levels of cigarette consumption and exhaled carbon monoxide. However, the trends for the 36-week smoking measures among the primigrav• ids, for which there were only eight LBW infants, were not statistically significant.
VOL. 89, NO.5, PART 1, MAY 1997
Table 3. Cigarette Consumption and Exhaled Carbon Monoxide at First and 36-Week Visits, and Low Birth Weight Mothers having L8W infants 36-wk visit
First visit Smoking measure Cigarettes per day 0 1-5 6-10 ~11
Exhaled carbon monoxide (ppm) <5 5--9 10-19 ~20
Total births
No. <2500 g
%
1.1 7.5 10.9 11.6"
110 86 100
0 3 7 9
0.0 5.2 8.1 9.0t
2.2 5.6 8.8 15.7*
111 65 112 62
2 3 8 6
1.8 4.6 7.1 9.7S
Total births
No. <2500 g
%
91 53 110 138
1 4 12 16
93
2 4 11 16
72
125 102
58
LBW = low birth weight; ppm = parts per million. Mantel-Haenszel test for linear associations: "P == .004. t P == .002. P < .001. S P == .017.
*
Discussion We found that levels of smoking at the first prenatal and 36-week visits, whether measured in terms of self-reported cigarettes per day or exhaled carbon mon• oxide concentration, were related significantly to birth weight and to increases in the chance of having an LBW infant. Gestational age at delivery was weakly related to exhaled carbon monoxide level, but not to cigarette consumption. The reduction in birth weight and the increase in the proportion of LBW infants in relation to increasing levels of cigarette consumption are compa• rable to those reported by others. 6,22 Our results indicate that for pregnant women who are current smokers at their first prenatal visit, their infants' potential birth weights are already compro• mised compared with those who were smokers at the time they learned they were pregnant but then quit. Birth weights are further compromised by subsequent levels of smoking. However, the extent of this further compromise is lessened for women who reduce their cigarette consumption, compared with continued smoking at the same level. The reductions in birth weight per cigarette smoked per day-about 8 g for the first visit and 11 g for the 36-week visit-are compara• ble to the estimate of Mathai et al 23 of 12 g. Haddow et al,22 in a study of 4211 pregnant women, reported similar correlations between cigarette consumption at 15-21 weeks' gestation and birth weight (-0.19) and between serum cotinine, a metabolite of nicotine, and birth weight (-0.23). They found in this large sample
Seeker-Walker et al
Carbon Monoxide and Birth Weight
651
that serum cotinine was a significantly better predictor of birth weight than was cigarette consumption. We are not aware of other reports of the relation between exhaled carbon monoxide and birth weight, but the reduction in birth weight we observed is com• parable to that reported in relation to increasing con• centrations of cord blood carboxyhemoglobin.24 As seen in the equation using carbon monoxide level and ges• tational age, the influence of smoking late in pregnancy, as judged by exhaled carbon monoxide levels at the 36-week visit, appears to have a greater effect on birth weight than carbon monoxide levels at the first prenatal visit. This is consistent with the observation that women who quit smoking before 20 weeks' gestation have infants whose birth weights are comparable to those of non-smokers' infants. u8 Many factors .other than smoking are known to influence the birth weight of singleton infants, includ• ing gestational age at delivery, gravidity, maternal age, height and weight, maternal weight gain during preg• nancy, substance use, ethnicity, social and economic factors, and infant gender. 2,3,25 Of these, only smoking, use of other substances, and maternal weight gain during pregnancy can be altered by interventions. Our estimates for the proportion of the variance due to cigarette consumption, unadjusted for these other pre• dictors, were 5% for the first visit and 8% at the 36-week visit. These estimates are larger than the adjusted esti• mates of Hebel et al19 for cigarette consumption mea• sured early in pregnancy, at 2%, or late in pregnancy, at 4%. The differences may reflect the correlation between cigarette consumption and the factors for which Hebel et al19 adjusted, particularly education and maternal weight gain. Our study clearly confirms the detrimental effect of smoking on birth weight and provides estimates of the size of that effect, whether determined from reported cigarette consumption or exhaled carbon monoxide. Reductions in cigarette consumption or in exhaled carbon monoxide levels must be substantial if gains in birth weight of 100 g or more are to be achieved. The higher average infant birth weights of women who stopped smoking either before or after their first prena• tal visit, or of those who had exhaled carbon monoxide levels less than 5 ppm, argue strongly that quitting smoking is the optimal behavior change to maximize birth weight and minimi~e the likelihood of having an LBW infant.
22.
References
23.
1. Simpson WJ. A preliminary report on cigarette smoking and the incidence of prematurity. Am J Obstet GynecoI1957;73:808-15. 2. Butler NR, Goldstein H, Ross EM. Cigarette smoking in pregnancy:
652 Seeker-Walker et al
Carbon Monoxide and Birth Weight
3.
4. 5.
6. 7. 8. 9.
10.
11.
12.
13.
14.
15. 16.
17.
18. 19.
20.
21.
24.
Its influence on birth weight and perinatal mortality. BMJ 1972;2: 127-30. Brooke OG, Anderson HR, Bland JM, Peacock JL, Stewart CM. Effects on birth weight of smoking, alcohol, caffeine, socioeconomic factors, and psychosocial stress. BMJ 1989;298:795-801. Sexton M, Hebel R. A clinical trial of change in maternal smoking and its effect on birth weight. JAMA 1984;251:911-5. MacArthur Co Newton JR, Knox EG. Effect of anti-smoking health education on infant size at birth: A randomized trial. Br J Obstet GynaecoI1987;94:295-300. Abel EL. Smoking during pregnancy: A review of effects on growth and development of offspring. Hum Bioi 1980;52:593-625. Frederiksen L, Martin JE. Carbon monoxide and smoking behav• ior. Addict Behav 1979;4:21-30. Jarvis MJ, Russell MAH, Saloojee Y. Expired air carbon monoxide: A simple breath test of tobacco smoke intake. BMJ 1980;281:484-5. Hughes JR, Epstein LH, Andrasik F, Neff DF, Thompson DS. Smoking and carbon monoxide levels during pregnancy. Addict Behav 1982;7:271-6. Bauman KE, Bryan ES, Dent CW, Koch GG. The influence of observing carbon monoxide level on cigarette smoking by public prenatal patients. Am J Public Health 1983;73:1089-91. Burling TA, Bigelow GE, Robinson JC, Mead AM. Smoking dUring pregnancy: Reduction via objective assessment and directive ad• vice. Behav Ther 1991;22:31-40. Price JH, Krol RA, Desmond SM, Losh DP, Roberts SM, Snyder FF. Comparison of three antismoking interventions among pregnant women in an urban setting: A randomized trial. Psychol Rep 1991;68:595-604. Rush D, Orme J, King J, Eiser JR, Butler NR. A trial of health education aimed to reduce cigarette smoking among pregnant women. Paediatr Perinat Epidemiol 1992;6:285-97. Hartmann KE, Thorp 1M, Pahel-Short L, Koch MA. A randomized controlled trial of smoking cessation intervention in pregnancy in an academic clinic. Obstet Gynecol 1996;87:621-6. Fingerhut LA, Kleinman JC, Kendrick JS. Smoking before, during and after pregnancy. Am J Public Health 1990;80:541-4. Prochaska JO, DiClemente Cc. Stages and processes of self-change of smoking: Toward an integrative model of change. J Consult C1in Psychol 1983;51:390-5. Benowitz NL, Jacob P, Kozlowski LT, Yu L. 1nfluence of smoking fewer cigarettes on exposure to tar, nicotine and carbon monoxide. N Engl J Med 1986;315:1310-3. MacArthur C, Knox EG. Smoking in pregnancy: Effects of stopping at different stages. Br J Obstet GynaecoI1988;95:551-5. Hebel JR, Fox NL, Sexton M. Dose-response of birth weight to various measures of maternal smoking during pregnancy. J Clin Epidemiol 1988;41:483-9. Windsor RA, Lowe ]B, Perkins LL. Behavioral impact and cost benefit of health education methods for pregnant smokers in public health maternity clinics: The Birmingham Trial II. Am J Public Health 1993;83:201-6. Li CQ Windsor RA, Perkins RL, Goldenburg RL, Lowe JB. The impact on infant birthweight and gestational age of cotinine• validated smoking reduction during pregnancy. JAMA 1993;269: 1519-24. Haddow JE, Knight GJ, Palomaki GE, Kloza EM, Wald NJ. Ciga• rette consumption and serum cotinine in relation to birthweight. Br J Obstet Gynaecol 1987;94:678-81. Mathai M, Skinner A, Lawton K, Weindling AM. Maternal smok• ing, urinary cotinine levels and birthweight. Aust N Z J Obstet Gynaecol 1990;30:33-6. Wouters EJ, de Jong PA, Cornelissen PJH, Kurver PHJ, van Gel WC, van Woensel CLM. Smoking and low birth weight: Absence
Obstetrics & Gynecology
of Influence by carbon monoxide. Eur J Obstet Gynecol Reprod Bioi 1987;25:3~41. 25. Showstack JA, Budetti PP, Minkler D. Factors associated with birthweight: An exploration of the roles of prenatal care and length of gestation. Am J Public Health 1984;74:1003-8.
Received August 21, 1996. Received in revised form November 26, 1996. Accepted January 15, 1997.
Address reprint requests to:
Roger H. Seeker-Walker, MB 1 South Prospect Street Burlington, VT 05401
Copyright Cl 1997 by The American College of Obstetricians and Gynecologists. Published by Elsevier Science Inc.
STANDARDS FOR REPORTING TRIALS The CONSORT listing of standards for reporting randomized trials has been adopted as policy by Obstetrics & Gynecology. Investigators who are planning, conducting, or reporting randomized trials should be thoroughly familiar with these standards. A copy can be obtained by contacting: Obstetrics & Gynecology, 1100 Glendon Avenue, Suite 1655, Los Angeles, CA 90024-3520; FAX (310) 208·2838.
VOL. 89, NO.5, PART 1, MAY 1997
Secker-Walker et al
Carbon Monoxide and Birth Weight
653