Smoking, maternal age, fetal growth, and gestational age at delivery Shi Wu Wen, MS: Robert L. Goldenberg, MD: Gary R. Cutter, PhD," Howard J. Hoffman, MA,c Suzanne P. Cliver, BA: Richard O. Davis, MD: and Mary B. DuBard, MAa
Birmingham, Alabama, and Bethesda, Maryland The relationship between smoking and maternal age and their combined effects on birth weight, Intrauterine growth retardation, and preterm delivery were studied. Smoking lowers birth weight both by decreasing fetal growth and by lowering gestational age at delivery. However, the effect of smoking on both fetal growth and gestational age is significantly greater as maternal age advances. In a multiple logistic regression model adjusting for race, parity, marital status, maternal weight, weight gain, and alcohol use, smoking was associated with a fivefold increased risk of growth retardation in women older than 35 but less than a twofold increased risk in women younger than 17. Smoking reduced birth weight by 134 gm in young women but 301 gm in women older than 35. Smoking in older women also was associated with more instances of preterm delivery and a lower mean gestational age when compared to women 25 or younger. (AM J OBSTET GVNECOL 1990;162:53-8.)
Key words: Smoking, growth retardation, preterm delivery, low birth weight In developed countries, maternal smoking is a major risk factor for low birth weight (LB W). 1. 2 Less well studied is the effect of maternal smoking on the two components of LBW, intrauterine growth retardation (rUGR)' and preterm delivery.' Since rarely, if ever, has the effect of smoking on these two important obstetric outcome measures been studied in the same population, we initiated this study. However, in performing the preliminary analyses, it became apparent to us that the effect of smoking on LBW varied with maternal age. It is this relationship, the changing effect of smoking on the LBW rate and its components, the preterm delivery and IUGR rates, as maternal age advances that will be the subject of this report.
Material and methods Women seen for prenatal care and delivered at the University of Alabama at Birmingham or Cooper Green Hospital in Birmingham, Alabama, from January 1983 through January 1988 were included in this
From the Perinatal EpIdemiology Umt, Department of ObstetrlC5 and Gynecology: and the School of Publzc Health,' The Universit.~ of Alabama at BIrmingham, and the Bwmetry Branch, Preventwn Research Program, Natzonal InstItute of ChIld Health and Human Development, National InstItutes of Health.' Supported zn part by National Instltutes of Health contract No. NO 1-HD-4-2811. Recelvedfor publzcatzon December 28,1988; revzsedJune 1,1989; accepted June 16, 1989. Reprint requests: Robert L. Goldenberg, MD, Department of Obstetncs and Gynecology, The University of Alabama at Birmzngham. Unzversity Statwn, Blrmzngham, AL 35294. 611114647
study. During this period of time there were approximately 20,000 births in the two hospitals. Deliveries in women with no prenatal care and therefore no study data, those transferred to our institution for delivery only, and those with multiple births, fetal deaths, or congenital malformations were excluded from the study. In addition, because maternal diabetes has such a major independent effect on birth weight and because physicians tend toward early delivery of women with this condition, infants of women with both gestational and insulin-requiring diabetes also were excluded. As a result, 17,149 pregnancies were analyzed. Of these, 1610 cases were coded as hypertension at delivery. After an initial analysis showing that the effect of smoking and maternal age on birth weight was similar when hypertensive smokers were compared with nonhypertensive smokers and because of a tendency of physicians toward early delivery of women with hypertensive disease, these women also were excluded so that there were 15,539 births in the final analysis. The prenatal data used in this study came from computerized medical records" and were based primarily on history collected at the first prenatal visit. Smoking was defined as the use of cigarettes during the pregnancy before the first visit. Alcohol use was also defined as use during pregnancy before the first visit. Maternal age was the age of the mother when she registered for prenatal care. Race was coded as white or nonwhite with virtually all nonwhites being black. Women were divided into those without a previous birth reaching 20 weeks' gestation (parity = 0) and those with at least one previous pregnancy of ~20 weeks (parity ~ 1). In the
53
54 Wen et al.
January 1990 Am J Obstet Gynecol
Table I. Population studied CharacterIStic
Race Black White Marital status Married Single Parity 0 2:1 Alcohol use No Yes Maternal age (yr) 2:16 17-19 20-25 26-30 31-35 2:36 TOTAL
% with characterIStic
Smokers %
Nonsmokers (%)
70.2 29.8
20.2 53.7
79.8 46.3
37.9 62.1
35.9 26.5
64.1 73.5
43.4 56.6
24.3 34.4
75.7 65.9
86.6 13.4
25.5 60.1
74.5 39.9
7.5 23.4 43.9 17.2 6.3 1.7 100
15.7 21.6 32.5 39.4 37.4 34.5 29.9
84.3 78.4 67.5 60.6 62.6 65.5 71.1
The percent of the study population with certain characteristics and, of the women with these characteristics, the percent who are smokers and nonsmokers.
Table II. Percents and 95% confidence intervals of IUCR and preterm delivery in smokers and nonsmokers in various maternal age groups /UGR %
Preterm delzvery %
Nonsmokers Maternal age (yr)
%
s16 17-19 20-25 26-30 31-35 2:36
7.3 5.5 5.0 4.7 5.3 3.7
I
Smokers
95% CI
%
5.6-9.0 4.6-6.4 4.4-5.7 3.6-5.8 3.5-7.2 0.8-6.6
8.6 7.9* 10.2* 11.1 * 11.5* 13.1*
I
Nonsmokers
95% CI
%
4.4-12.9 5.9-10.0 8.8-11.5 9.1-13.1 8.0-14.9 5.9-20.3
15.9 12.7 10.4 10.2 10.3 8.4
I
Smokers
95% CI
%
13.5-18.3 11.4-14.0 9.5-11.4 8.6-11.7 7.8-12.8 4.2-12.7
13.3 14.3 1l.5 13.9* 14.9* 10.6
I
95% CI
8.1-18.5 11.7-16.9 10.1-12.9 11.7-16.0 11.0-18.7 4.1-17.1
*Significant differences (p < 0.05) between smokers ~nd nonsmokers.
multiple regression analyses, the maternal weight used was that recorded at the first prenatal visit. Maternal weight gain per week after 20 weeks' gestation was derived by subtracting the weight at the first visit after 20 weeks' gestation from the weight at the last prenatal visit and dividing by the elapsed time in weeks. Gestational age at delivery was calculated from the best estimate of the expected date of confinement in the prenatal record. The best-estimate expected date of confinement was generally derived at the first or second prenatal visit and was based on all data available including last menstrual period, ultrasonography, physical examination, quickening, and auscultation of the fetal heart. A preterm delivery was defined as one that occurred before 37 completed weeks. IUCR was defined as a birth weight below the 10th percentile for
each gestational age according to the standards of Brenner et al. 6 To isolate the effect of smoking from other p
Birmingham, Alabama, and Bethesda, Maryland The relationship between smoking and maternal age and their combined effects on birth weight, Intrauterine growth retardation, and preterm delivery were studied. Smoking lowers birth weight both by decreasing fetal growth and by lowering gestational age at delivery. However, the effect of smoking on both fetal growth and gestational age is significantly greater as maternal age advances. In a multiple logistic regression model adjusting for race, parity, marital status, maternal weight, weight gain, and alcohol use, smoking was associated with a fivefold increased risk of growth retardation in women older than 35 but less than a twofold increased risk in women younger than 17. Smoking reduced birth weight by 134 gm in young women but 301 gm in women older than 35. Smoking in older women also was associated with more instances of preterm delivery and a lower mean gestational age when compared to women 25 or younger. (AM J OBSTET GVNECOL 1990;162:53-8.)
Key words: Smoking, growth retardation, preterm delivery, low birth weight In developed countries, maternal smoking is a major risk factor for low birth weight (LB W). 1. 2 Less well studied is the effect of maternal smoking on the two components of LBW, intrauterine growth retardation (rUGR)' and preterm delivery.' Since rarely, if ever, has the effect of smoking on these two important obstetric outcome measures been studied in the same population, we initiated this study. However, in performing the preliminary analyses, it became apparent to us that the effect of smoking on LBW varied with maternal age. It is this relationship, the changing effect of smoking on the LBW rate and its components, the preterm delivery and IUGR rates, as maternal age advances that will be the subject of this report.
Material and methods Women seen for prenatal care and delivered at the University of Alabama at Birmingham or Cooper Green Hospital in Birmingham, Alabama, from January 1983 through January 1988 were included in this
From the Perinatal EpIdemiology Umt, Department of ObstetrlC5 and Gynecology: and the School of Publzc Health,' The Universit.~ of Alabama at BIrmingham, and the Bwmetry Branch, Preventwn Research Program, Natzonal InstItute of ChIld Health and Human Development, National InstItutes of Health.' Supported zn part by National Instltutes of Health contract No. NO 1-HD-4-2811. Recelvedfor publzcatzon December 28,1988; revzsedJune 1,1989; accepted June 16, 1989. Reprint requests: Robert L. Goldenberg, MD, Department of Obstetncs and Gynecology, The University of Alabama at Birmzngham. Unzversity Statwn, Blrmzngham, AL 35294. 611114647
study. During this period of time there were approximately 20,000 births in the two hospitals. Deliveries in women with no prenatal care and therefore no study data, those transferred to our institution for delivery only, and those with multiple births, fetal deaths, or congenital malformations were excluded from the study. In addition, because maternal diabetes has such a major independent effect on birth weight and because physicians tend toward early delivery of women with this condition, infants of women with both gestational and insulin-requiring diabetes also were excluded. As a result, 17,149 pregnancies were analyzed. Of these, 1610 cases were coded as hypertension at delivery. After an initial analysis showing that the effect of smoking and maternal age on birth weight was similar when hypertensive smokers were compared with nonhypertensive smokers and because of a tendency of physicians toward early delivery of women with hypertensive disease, these women also were excluded so that there were 15,539 births in the final analysis. The prenatal data used in this study came from computerized medical records" and were based primarily on history collected at the first prenatal visit. Smoking was defined as the use of cigarettes during the pregnancy before the first visit. Alcohol use was also defined as use during pregnancy before the first visit. Maternal age was the age of the mother when she registered for prenatal care. Race was coded as white or nonwhite with virtually all nonwhites being black. Women were divided into those without a previous birth reaching 20 weeks' gestation (parity = 0) and those with at least one previous pregnancy of ~20 weeks (parity ~ 1). In the
53
54 Wen et al.
January 1990 Am J Obstet Gynecol
Table I. Population studied CharacterIStic
Race Black White Marital status Married Single Parity 0 2:1 Alcohol use No Yes Maternal age (yr) 2:16 17-19 20-25 26-30 31-35 2:36 TOTAL
% with characterIStic
Smokers %
Nonsmokers (%)
70.2 29.8
20.2 53.7
79.8 46.3
37.9 62.1
35.9 26.5
64.1 73.5
43.4 56.6
24.3 34.4
75.7 65.9
86.6 13.4
25.5 60.1
74.5 39.9
7.5 23.4 43.9 17.2 6.3 1.7 100
15.7 21.6 32.5 39.4 37.4 34.5 29.9
84.3 78.4 67.5 60.6 62.6 65.5 71.1
The percent of the study population with certain characteristics and, of the women with these characteristics, the percent who are smokers and nonsmokers.
Table II. Percents and 95% confidence intervals of IUCR and preterm delivery in smokers and nonsmokers in various maternal age groups /UGR %
Preterm delzvery %
Nonsmokers Maternal age (yr)
%
s16 17-19 20-25 26-30 31-35 2:36
7.3 5.5 5.0 4.7 5.3 3.7
I
Smokers
95% CI
%
5.6-9.0 4.6-6.4 4.4-5.7 3.6-5.8 3.5-7.2 0.8-6.6
8.6 7.9* 10.2* 11.1 * 11.5* 13.1*
I
Nonsmokers
95% CI
%
4.4-12.9 5.9-10.0 8.8-11.5 9.1-13.1 8.0-14.9 5.9-20.3
15.9 12.7 10.4 10.2 10.3 8.4
I
Smokers
95% CI
%
13.5-18.3 11.4-14.0 9.5-11.4 8.6-11.7 7.8-12.8 4.2-12.7
13.3 14.3 1l.5 13.9* 14.9* 10.6
I
95% CI
8.1-18.5 11.7-16.9 10.1-12.9 11.7-16.0 11.0-18.7 4.1-17.1
*Significant differences (p < 0.05) between smokers ~nd nonsmokers.
multiple regression analyses, the maternal weight used was that recorded at the first prenatal visit. Maternal weight gain per week after 20 weeks' gestation was derived by subtracting the weight at the first visit after 20 weeks' gestation from the weight at the last prenatal visit and dividing by the elapsed time in weeks. Gestational age at delivery was calculated from the best estimate of the expected date of confinement in the prenatal record. The best-estimate expected date of confinement was generally derived at the first or second prenatal visit and was based on all data available including last menstrual period, ultrasonography, physical examination, quickening, and auscultation of the fetal heart. A preterm delivery was defined as one that occurred before 37 completed weeks. IUCR was defined as a birth weight below the 10th percentile for
each gestational age according to the standards of Brenner et al. 6 To isolate the effect of smoking from other p
