Tobacco Documents Online

ABS, TRA T Objectives. The associations of infant birth outcomes with maternal pregravid obesity, gesmtional weight gain, and prenatal cigarett~ smoking Methods. A retrospective ahaly- sis of 1343 obose and normal-weight gravidas evaluated the associations of cigarette smoking, gestational weight change, and pregravid body mass index with birthweight, low birthweight, and small- and large-for- gestational-age births. Results. Smoking was associ- ated with the delivery of lower- birthweight infants for both obese and normal-weight women, and ges- rational weight gain did not eliminate the birthweight-lowering effects of smoking. Women at highest risk of delivering lower-birthweight infants were obese smokers whose gesta- tional gains were less than 7 kg and normal-weight smokers whose gesta- tional gains were less than 11.5 kg. Conclusions. To balance the risks of small and large-size infants, gains of 7 to 11.5 kg for obese women and 11.5 to 16 kg for normal- weight women appear appropriate. (Am J Public Health. 1997;87:591- ~96) The Effects of Cigarette Smoking anc Gestational Weight Change on Birth Outcomes in Obese and Normal-Weight Women o 0~ 0 Wendy L. Hellerstedt, PhD, MPH, John H. Himes, PhD, MPH, Mary. Story, PhD, Irene R. Alton, MS, RD, and Laura E. Edwards, MD Introduction Prenatal cigarette smoking is one of the most important preventable causes of low birthweight~ and fetal growth retarda- tion.: It is estimated that 20% to 30% of the low-birthweight bh-ths in the United States are attributable to smoking) De- spite many public health efforts and high public awareness of the dangers of smok- ing, the overall prevalence of smoking during pregnancy is 21% in the United States.4 The effects of tobacco on birth- weight are probably due to several mecha- nisms. While most reports emphasize placental pathology, cigarette smoking could also affect maternal nutrition and, consequently, fetal nutrition,s If pregnant smokers are nutritionally compromised, it is plausible that higher pregravid weights (e.g., obesity) or increased gestational weight gain may counteract some of the smoking-related effects. In 1990, the Institute of Medicine developed pregravid weight-specific guide- lines for gestational weight gain.5 These recommendations were presented accord- ing to pregravid body mass index (weight [kg]/height [m]2). Among the recommen- dations were gestational gains between 11.5 and 16 kg for normal-weight women (i.e., body mass index = 19.8 to 26) and at least 7 kg for obese women (i.e., body mass index greater than 29). Since this report, two studies including obese women have suggested that an upper limit of 11.5 kg may optimize birthweight.9.1° Cur- rendy, there is no agreement about the appropriate upper bound of gestational gain for the obese~ or whether weight loss, no weight gain, or gains less than 7 kg for these women compromise birthweightY- Clarification of recommendations is impor- tant because the prevalence of obesity among women of childbearing age is increasing in the United S~ates.14 The Institute of Medicine report also stated that cigarette-smoking .g.g.g.g.ravidas may re- quire specific nutritional counseling, but did not comment on whether different gestational weight-gain recommenda- tions should be made for smokers and nonsmokers. This study examined the relationship of the Institute of Medicine's gestational weight-gain recommendations and ciga- rette smoking to birthweight in a hospital- based retrospective study of 1343 obese and normal-weight gravidas. Of particular interest were whether the birthweight- lowering effect of smoking was similar for infants of obese and normal-weight women, and whether higher-than-recom- mended gestadonal weight gains could compensate for the effect of cigarette smoking on birthweight without also increasing the risk for delivery of high- birthweight infants. Methods The pregnancy and delivery records of all 21 185 deliveries at St. Paul- Ramsey Medical Center from January 1977 through August 1993 were reviewed Wendy L. Hellerstedt. John H. I-limes, and Mary Story are with the School of Public Health, University of Minnesota, Minneapolis. Irene R. Alton is with Health Start Inc. St. Paul, Minn. Laura E. Edwards is with St. Paul-Ramsey Medical Center, St. Paul, Minn. Requests for reprints should be sent to Wendy L. Hellerstedt. PhD, MPH, University of Minnesota, Division of Health Policy and Man- agement, Maternal-and Child Health, D355 Mayo Bldg, Box 97. 420 Delaware St SE, Minneapolis, MN 55454. This paper was accepted June 28, 1996. Editor's Note. See related annotation by Rasmussen and Adams tp 543) in this issue. A/~rJJ 1997. VoL B7, No. ~ American Journal of Public Health 591

Hel~erstedt et TABLE 1---Characteristlc~ of Obese and Normal.Weight Gravidas Who Delivered at SL Paul--Ramsey Medical Center, January 1977 through August 1993 Obese Normal Weight (n = 683) (n = 660) Demographic characteristics Age group, • '~15 yrs 16-34 yrs >35 yrs Previous births, %a.b 0 1--3 ~4 Race/ethnicity, White Black Hispanic Native Amedcan 0.2 0.8 91.0 93.5 8.8 5.8 31.8 33.3 64.7 64.9 3.5 1.8 68.8 69.1 20.4 20.6 6.6 6.5 3.8 3.2 Pregravid weight and gestational weight change Mean pregravid body mass index (SO)* 38.3 (4.6) 22.8 (1.6) Range 29.1-66.4 19.6-26.0 Mean pregravid weight, kg'(SD)* 103.5 (13.7) 61.1 (5.9) Range 63.5-176.9 47.6-83.9 Mean gestational weight change, kg (SD)c,* 9.6 (8.3) 14.6 (5,5) Range - 18.1-43.5 - 2.3--32.2 Prenatal characteristics Adequate utilization of prenatal care, %~ 73.9 Cigarette smoking, %b 26.4 Alcohol drinking, %b 10.0 Illicit drug use, %~ 4.7 69.1 26.2 9.1 5.6 Not~. Body mass index = weight (kg)/height (m)=. =Matching variables. bNo group differences in chi-square analy~es. CAdjusted for infant gestational age at birth. "P < .0001. retrospectively. The center, a major urban hospital in St. Paul, Mirm, is one of the major providers for the area's indigent population. The initial sample included almost every obese woman who delivered a singleton during the time period. The record of each obese woman identified was matched to that of one normal-weight woman by race/ethnicity, delivery date, and broad categories of age (i.e., <-15, 16 through 34, and ->35 years of age) and parity (i.e., nulliparous, parity I through 3, and parity ->4). The original sample consisted of 771 obese and 771 normal- weight women (n = 1542). After exclu- sions because of missing data, siblings, and fetal deaths, the final sample con- sisted of 683 obese and 660 normal- weight women, to The four outcomes examined in this study were the continuous variable birth- weight (in grams), and dichotomous vari- ables for low birthweight (<2500 g), small for gestational age (< sex-specific 10th percentilelS), and large for gesta- fional age (> sex-specific 90th percen- tile~S). Of interest were the independent, and potentially combined, effects on these birthweight variables of (I) prenatal ciga- rette smoking and pregravid obesity, and (2) smoking and gestational weight gain. Smokers were defined as women who had documentation of any cigarette smoking during pregnancy. Total gestational weight change was defined as the difference between self-reported pregravid weight and measured weight at the last prenatal care visit, which occurred, on average, 1 week wior to delivery. Most of the univariate chi-square, multiple linear (i.e., general linear mod- els), and logistic regression analyses were conducted for obese and normal-weight women separately.t6.All regression analy- ses were adjusted for the following potential confounders: continuous vari- ables for maternal age and pregravid body mass index and dichotomous variables for female infant sex, maternal Black race, maternal Native American race, nullipar- ity, prenatal alcohol use, prenatal illicit drug use, adequacy of prenatal care,17 presence of gestafional hypertension, and presence of gestational diabetes. All analy- ses of birthweight variables were also adjusted for gestational age at birth, which was based on the best estimate from reported menstrual data, early prenatal exam, and early ultrasound studies. Logistic regression analyses com- puted the adjusted odds ratios and 95% confidence intervals for birthweight out- comes for women in smoking-specific weight-change categories relative to the odds for women in a reference group reflecting hypothesized low risk. For obese women, the reference group was nonsmokers who gained 7 to l 1.5 kg; this reference group was compared with the following groups: smokers who gained less than 7 kg, smokers who gained 7 through l l.5 kg, smokers who gained more than 11.5 kg, nonsmokers who gained less than 7 kg, and nonsmokers who gained more than l l.5 kg. For normal-weight women, the following groups were compared with the reference group of nonsmokers who gained 11.5 through 16 kg: smokers who gained less than 11.5 kg, smokers who gained 11.5 through 16 kg, smokers who gained more than 16 kg, nonsmokers who gained less than l 1.5 kg, and nonsmokers who gained more than 16 kg. Results Table 1 compares the characteristics of the 683 obese and 660 normal-weight women. The age match used broad categories, thus explaining the difference (P <--.0001) in the mean age of obese (27.1 years; 5.7 SD) and normal-weight women (~.4 years;. 5.9 SD). Obese women, compared with normal-weight women, gained on average one third less weight during pregnancy and had a wider variation in gestational weight change. Overall, 38% of obese women lost weight or gained less than 7 kg, compared with 7% of normal-weight women. Obese women were also less likely than normal- weight women to gain more than 11.5 kg (40% vs 73%, respectively). There was no difference in prenatal smoking prevalence between obese and normal-weight women.

Obesity, Smoking, and Birthweight Pregravid Obesi~. and Birth Outcomes The mean gestationat age at birth did not differ between infants of obese wo- men (39.0 weeks; 2.6 SD) and normal- weight women (39.1 weeks; 2.2 SD), but the frequency of preterm births was higher (P = .04) among obese women (10.5%~ than among normal-weight women (7.3%). Preterm (i.e., less than 37 completed weeks of age at birth) was strongly associated with low birthweight: 76% of the low-birthweight infants of obese women and 64% of the low- birthweight infants of normal-weight women were preterm. The frequency of combined term and preterm low birthweight was similar for infants of obese and normal-weight women (8.2% vs 8.1%, respectively). However, the gestational-age-adjusted mean birth- weight for infants of obese women was higher than that for normal-weight women (P <-.001): 3420 g (760 SD) compared with 3285 g (620 SD). The frequency of small-for-gestational-age births was lower among infants of obese women than of normal-weight women (6.2% vs 9.1%; P = .042), and the frequency of large-for- gestational-age births was higher (14.1% vs 8.0%; P -< .001). Smoking and Birth Outcomes Prenatal smoking was not associated with mean gestational age or with the frequency of preterm births for either obese or normal-weight women. The distribu- tion of adverse birthweight outcomes generally varied significantly by prenatal smoking status in chi-square analyses, but significance did not always persist after adjustment for potential confounders in logistic regression analyses (Table 2). Of interest was the change in low-birth- weight risk for infants of obese smokers compared with infants of nonsmokers after adjustment. Crude analyses showed a twofold increase in low-birthweight risk, but the adjusted analyses showed a fivefold increase in low-birthweight risk for infants of obese smokers compared with nonsmokers, although the confidence interval for the estimate of the odds ratio included 2.0. Further examination of the logistic model showed that variation in gestational age was associated with this hicreased odds ratio for obese smokers. Gesta.~onal Weight Change and Birth Outcomes The distributional patterns of each of the birthweight outcomes showed signifi- cant linear trends across gestational • TABLE 2--Oistribution of Birth Outcomes for 683 Obese and 660 Normal-Weight Women, by Prenatal Smoking Status Prenatal Smoking Status Birth Outcomes Birthweight <10% Weight for >90% Weight for <2500 g Age (Small for Age (Large for (Low Birthweight) GestationalAge) Gestational Age) Obese women (body mass index >29) Smoker (n = 180), % 13.9 10.6 8.9 Nonsmoker (n = 503), % 6.8 4.6 15.9 P= .003 .OIM .020 Adjusted ORh 5.1 1.8 0.5 95% CI 1.8, 14.3 0.9, 3.7 0.3, 0.9 Normal-weight women (body mass index = 19.8-26.0) Smoker (n = 173), % 8.7 15.0 3.5 Nonsmoker (n = 487), % 7.2 7.0 9.7 P= .526 .002 .010 Adjusted ORb 1.0 1.9 0.3 95% CI 0.3, 2.8 1.0, 3.5 0.1, 0.9 Note. OR = odds ratio; CI = confidence interval. "Unadjusted P, chi-square analysis. bAdjusted odds ratio, smokers vs nonsmokers, from multivariate logistic regression analysis, controlled for gestational age, infant sex, pregravid body mass index, maternal race, age, padty, alcohol use, drug use, prenatal care adequacy, gestational diabetes, and gestational hypertension. TABLE 3---Distribution of Birth Outcomes for 683 Obese and 660 Normal- Weight Women, by Categories of Gestational Weight Gain Gestational Weight Gain Birth Outcomes Birthweight <10% Weight for >90% Weight for <2500 g Age (Small for Age (Large for (Low Birthweight) GestationalAge) Gestational Age) Obese women (body mass index >29) Lost/no gain (n = 75), % 16.0 0.5-6.5 kg (n = 181), % 11.1 7-11.5 kg (n = 168), % 8.3 12-16 kg (n = 126), % 4.0 >16 kg (n = 133), % 6.0 P" .003 Normal-weight women (body mass index = 19.8-26.0) <11.5 kg (n = 176), % 14.2 11.5-16 kg (n = 240), % 5.4 >16 kg (n = 244), % 4.9 p, .001 10.7 9.3 6.6 10.5 6.0 11.3 4.0 17.5 5.3 21.8 .115 .001 15.9 2.8 7.5 6.7 5.7 13.1 .001 <.001 "Unadjusted P, ManteI-Haenszel chi-square analysis for linear trend. weight--change categories for obese and normal-weight women, except for the frequency of small-for-gestational-age in- fants among obese women (Table 3). Accordingly, the frequencies of low- birthweight infants for obese and normal- weight women and of small-for-gesta- tional-age infants for normal-weight women decreased with increasing catego- ries of gain, while the frequen.ey of large-for-gestational-age infants increased with increasing categories of gain. April ! 997. Vol. 87. No. 4 American Journal of Public Health 593

TABLE 4---Frequencies of Low-Birthweight (LBW), Small-for-GestationaI-Age (SGA), and Large-for-GestationaI-Age (LGA) Births for 683 Obese and 660 Normal-Weight Women, by Prenatal Smoking and Compliance with the Institute of Medicine's Gestational Weight-Gain Recommendations Smokers Nonsmokers Gained Less Gained More Gained Less Gained More than IOM Gained Within than IOM than IOM Gained Within than IOM Recommended IOM Range Recommended Recommended IOM Range Recommended Obese women (body mass index >29) No. 75 40 LBW, % 17.3 10.0 SGA, % 13.3 10.0 LGA, % 5.3 10.0 Normal-weight women (body mass index = 19.8-26.0) No. 63 55 LBW, % 17.5 3.6 SGA, % 28.6 10.9 LGA, % 0 1.8 65 181 128 194 12.3 10.5 7.8 2.6 7.7 5.5 4.7 3.6 12.3 12.2 11.7 22.2 55 113 185 1~ 3.6 12.4 6.0 5.3 3.6 8.9 6.5 6.4 9.1 4.4 8.1 14.3 Note. IOM = Institute of Medicine. Recommendations are from IOM report,s Gestational Weight Change, Smoking, and Birth Outcomes Among obese women, gestational weight gain did not vary significandy by smoking status; smokers gained an aver- age of 9.3 kg (8.5 SD) compared with 9.7 kg (8.3 SD) for nonsmokers. However, normal-weight smokers gained less than nonsmokers (P = .0015): 13.4 kg (6.0 SD) versus 15.0 kg (5.3 SD), respectively. Stratified analyses of obese and normal- weight women showed that the frequen- cies of low-birthweight and smail-for- gestational-age births were highest among smokers who gained less weight than recommended by the Institute of Medi- cine and the frequency of large-for- gestational-age births was highest among nonsmokers who gained more than recom- mended (Table 4). These findings gener- ally persisted in multivariate analyses. For obese women, compared with nonsmok- ers who gained 7 through 11.5 kg, • smokers who gained less than 7 kg were at significantly higher risks for delivering low-birthweight infants (adjusted odds ratio [OR] = 7.7; 95% confidence inter- val [CI] = 1.5, 40.0; P = .016) and small- for-gestational-age infants (adjusted OR = 3.2; 95% CI = 1.1, 10.1; P = .04). For normal-weight women, smokers who gained less than 11.5 kg were at signifi- cantly higher risk for delivering small-for- gestational-age infants (adjusted OR = 4.3; 95% CI = 1.8, 10.3; P=.001), compared with nonsmokers who gained 11.5 through 16 kg. The combined effects of smoking and gestational weight change influenced the distribution of large-for- gestational-age births only among infants of obese women in multivariate analyses. Compared with infants of obese nonsmok- ers who gained 7 through 11.5 kg, the only group at significantly higher risk of large-for-gestadonal-age births was non- smokers who gained more than 11.5 kg (adjusted OR = 2.3; 95% CI = 1.2, 4.5; P = .014). To allow further understanding of the influence of smoking and gestational weight gain, as well as pregravid body mass index as a continuous variable, multiple linear regression was conducted on birthweight (in grams). All of the potentially confounding variables used in the main analyses were entered. Signifi- cant variables in the model for obese women (R2 = .56) that were negatively associated with birthweight were smoking (adjusted coefficient = - 189; SE = 43; P-< .0001), Black race, nulliparity, and female infant sex. Pregravid body mass index (adjusted coefficient = 8.1; SE = 4.0; P = .05), gestational age, gestadonal diabetes, and gestational weight change (adjusted coefficient = 5.8; SE = 1.0; P<-.0001) were positively associated with birdaweight. An interaction term for smoking and gestational weight change was tested in the final model but was not significant (P > .90). For normal-weight women, the model predicting birthweight (R2 = .53) was very similar to that for obese women: smoking (adjusted coefficient =-109; SE = 38; P = .005), Black race, nullipar- ity, and female infant sex were negatively associated with birthweight while gesta- tional age, gestational diabetes, and gesta- tional weight change (adjusted coeffi- cient = 7.8; SE = 1.4; P --- .0001) were positively associated. Contrary to the findings for obese women, smoking ap- peared to be less strongly associated with birthweight for normal-weight women, and pregravid body mass index was not associated. An interaction term for smok- ing and gestational weight change was tested in the final model and was not associated with birthweight (P > .30). The samples of obese and normal- weight women were combined to exam- ine whether gestational weight gain modi- fied the influence of smoking on the continuous variable, birthweight (in grams). Three regression models were computed. The first model, including covariates and pregravid body mass in- dex, examined the relationship of smok- ing with birthweight; the second model included all of the variables in the first model plus gestational weight change. In the first model, smoking was associated with a decrease in birthweight (adjusted coefficient = - 169; SE = 29; P -- .0001). The addition of gestational weight change in the second model produced a slight weakening of this effect (adjusted smok- ing coefficient =-155; SE = 29; P <- .0001). An interaction term for smoking and gestational weight change was added 594 Arflericall JotllTla.l of Public Health April 1997. Vol. 87. No. 4

Obesity, Smoking, and Birthweight to a third model, but was not significant (P = .12). A similar series of analyses were conducted for obese and normal- weight women separately and resulted in similar findings: the coefficient for smok- ing decreased by about 15 g for infants of both normal-weight and obese women when gestational weight change (P -< .001 for both groups) was added to the model. Discussion Because cigarette smoking is so strongly associated with birthweight,~ it is important to fully understand the nature of its relationship to birthweight and to specifically evaluate whether maternal characteristics or behaviors modify the effects of tobacco. We believe this is the first study to evaluate the Institute of Medicine's recommendations5 for gesta- tional weight gain for women who are at high risk for delivering lower-birthweight infants because of prenatal cigarette smok- ing. We found that the lower bound of these recommendations appears appropri- ate for obese and normal-weight nonsmok- ers and smokers. Although gestational weight gain did not significantly weaken the birthweight-lowedng effects of smok- ing, there was evidence of a combined effect of less-than-recommended gesta- tional weight gain and smoking on birthweight for obese and normal-weight women. These data indicate that to optimize birthweight, obese and normal- weight smokers should be encouraged to gain at least 7 kg and 11.5 kg, respec- tively, during pregnancy. The Institute of Medicine was unable to specify an upper bound of gestational weight gain for obese women because of insufficient data.~ For obese nonsmokers in this study, gains beyond 11.5 kg were associated with slightly reduced risks for low-birthweight and small-for-gestational- age births, but were also associated with significantly increased risk for large-for- gestational-age births, suggesting that 11.5 kg may be an appropriate upper bound for them. For infants of obese smokers, gains greater than 11,5 kg did not significantly influence the risk for large-for-gestational-age births; there was no decrease in risk for low birthweight, and the decrease in risk for small-for- gestational-age births was not appre- ciable, suggesting that an upper bound of 11.5 kg may be appropriate for obese smokers as well. These data, suggesting optimal out- comes for obese women who gain 7 through 11.5 kg, are in agreement with data fi'om a recent study by Cogswell et al.,9 who found in a nationally representa- tive database that an upper bound of 11.5 kg for obese women minimized risks for both low birthweight and birthweight greater than 4500 g, and with a report by Parker and Abramsts that showed an' increased risk for small-for-gestational- age births among infants of obese women who gained less than 7 kg. Neither of these reports examined outcomes by maternal smoking status. In this study, the mean birthweight of infants of obese women was as strongly affected by smoking as that of normal- weight women. While high pregravid body mass index may counteract some of the effects of cigarette smoking on small- for-gestational-age births and low birth- weight, it appears that whatever protec- tion is provided by obesity, it is by no means complete. In this study, the risk for low-birthweight infants was five times higher for obese smokers than for non- smokers, and differences in weight gain between obese smokers and nonsmokers did not explain this surprisingly large effect of cigarette smoking. Gravidas who are obese and who smoke may have other, unmeasured behaviors or characteristics that put them at high risk for the delivery of lower-birthweight infants. The finding of an increased risk for lower-birthweight infants among obese smokers does not agree with older reports by Gain et al.6 and Luke et al.,s who have reported no smoking effect on infant birthweight among heavy women. However, these studies had small samples of obese women (i.e., less than 75 women), which limits the inferences that might be drawn from them. Our finding that the risk for delivery of low-birthweight infants was not associated with smoking in normal- weight women was also surprising and cannot be easily explained, especially because these smokers had lower mean weight gains than nonsmokers. It should be noted, however, that the risk for small-for-gestational-age births among normal-weight smokers compared with nonsmokers was significantly elevated. There are several limitations to this study. Because the association of cigarette smoking and birthweight is related to dosage and timing of exposure,Lt9-22 the simple categorization of women as smok- ers or nonsmokers obscured the nature of the association between birthweight and the timing and intensity of smoking. Further, it is plausible that some women were misclassified, and we cannot address the degree or direction of error associated with self-reports of smoking. Self-report is the most prevalent and practical method of ascertaining prenatal smoking status, but because smoking is known to endan- ger not only the mother but the fetus. nondisclosure is likely.:3 The effects of broad categorization and misclassification could have weakened our ability to find significant relationships and leave open the question about whether the true association of smoking and birthweight is actually stronger than presented in this The data on gestational weight gain were derived from self-reported pregravid weight and measured predelivery weight. While self-reported pregravid weights are thought to be reasonably valid, within a standard deviation of about 1.5 kg,24 obese women may be more likely to underreport pregravid weight than normal- weight women."~ In this study, if obese women underreported their pregravid weights, our estimation of gestational gain would have been biased toward high gains. Because the estimated gestational weight change among obese women in this study tended toward surprisingly low gains and weight loss, rather than high gains, we suspect that obese women did not significantly underreport their pre- gravid weights. It is clear that no level of gestational weight gain will eliminate the effects of cigarette smoking on birthweight, and any effort that effectively decreases smoking during pregnancy is important. Also, good birth outcomes occur among high-risk women who experience a broad range of total gestatioaal weight gain. Thus, further scrutiny of the meaning of gestational weight gain is indicated. Gestational weight gain is com- monly defined as total or net pregnancy weight gain, largely because of data availability and difficulties in meas0ring and analyzing trends in weight gain. The pattern of weight gain in pregnancy, or abrupt changes in maternal weight, may also influence birthweight.2~26,27 Future studies should address patterns of gesta- tional weight gain and should specifically examine obese women so that the manage- ment of their prenatal care may be better understood. Also, because gestational weight gain is not a surrogate measure of nutritional adequacy2s and because women vary widely in their energy needs during pregnancy,29 further work should continue to examine whether nutritional modifica- tion can attenuate the strong effect of tobacco on birthweight. [] o o April 1997. Vol. 87. No. 4. American Journal of Public Health 595

~ ~ ~ H~eller~tedt et al. Acknowledgments These data were presented at the 123nt annual meeting of the American Public Health Associa- tion, San Diego, Calif, October 29 through November 2, 1995. The authors wish to thank Andrew Mosca, Zhiping Huang, and Andrea Wilcox for their assistance with this manuscript. References 1. Kramer MS. Intrauterine growth and gesta- tional duration determinants. Pediatrics. 1987:80:502-511. 2. Wen SW, Goldenberg RL, Cutter GIL et al. Smoking, maternal age, fetal growth, and gestational age at delivery. Am J Obstet Gynecol. 1990;162:53-58. 3. Healthy People 2000: National Health Promotion and Disease Prevention Objec- tives. Washington, DC: Public Health Service; 1991. DHHS publication PHS 91-50212. 4. Williamson DF, Seaxiula MK, Kenddck JS, Binkin NJ. Comparing the prevalence of smoking in pregnant and nonpregnant women, 1985 to 1986. JAMA. 1989;261: 70-74. 5. Institute of Medicine, Subcommitte~ on Nutritional Status and Weight Gain during Pregnancy. Nutrition during Pregnancy. Washington DC: National Academy Press; 1990. 6. Garn SM, Hoff K, McCab¢ KD. Is there nutritional mediation of the "smoking effect" on the fetus.'? Am J Clin Num. !979;32:118 I-1187. Editorial. 7. Metcoff J, Costilo~ R Crosby WM, et aL Effect of food supplementation (WIC) during pregnancy on birth weight. Am J Clin Nute 1985;41:933-947. 8. Luke B, Hawkins MM, Petrie RH. Influ- ence of smoking, weight gain, and pre- gravid weight for height on intrauterine growth. Am J Clin Nuo:. 198h34:1410- 1417. 9. Cogswell ME, Serdula MK, Hungerford DW, Yip R. Gestational weight gain among average-weight and overweight women-- what is excessive? Am J Obstet Gynecol. 1995; 172:705-7 ! 2. 10. Edwards L, Hellerstedt W, Alton I, Story M, Himes J. Pregnancy complications and birth outcomes in obese and normal-weight women: effects of gestational weight change. Obstet Gynecol. 1996;87:389-- 394. 11. Johnson JWC, Longmate JA, Frentzen B. Excessive maternal weight and pregnancy outcome. Am J Obstet Gynecol. 1992;167: 353-372. 12. Fr~ntzen BH, Dimperio DL, Cruz AC. Maternal weight gain: effect on infant birth weight among overweight and average- weight low-income women. Am J Obstet Gynecol. 1988; 159:1114-1117. 13. Ratner RE, Hamner LH III, lsada NB. Effects of gestational weight gain in morbidly obese women: I. maternal morbid- ity. Ara J P erinatoL 1991;8:21-24. 14. Kucamarski RJ, F'legal KM, Campbell SM, Johnson CL. Increasing prevalence of overweight among US adults: the National Health and Nutrition Examination Sur- veys, 1960 to 1991. JAMA. 1994;272:205- 211. 15. Williams RL Creasy RK, Cunninghara GC, Hawes WE, Non'is FD, Tashiro M. Fetal growth and perinatal viability in California. Obstet Gynecol. 1982;59:624- 632. 16. SAS Language and Procedures: Usage, Version 6. Cary, NC: SAS Institute lnc; 1989. ~ 17. Kessner DM, Singer J, Kalk CE, Schle- singer ER. Infant death: an analysis by maternal risk and health care. Washington, DC: National Academy of Sciences; 1973. 18. Parker JD, Abrams B. Prenatal weight gain advice: an examination of the recent prenatal weight gain recommendations of the Institute of Medicine. Obstet Gynecol. 1992;79:664-669. 19. Walsh RA. Effects of maternal smoking on adverse pregnancy outcomes: examination of the criteria of causation. Hum Biol. 1994:66:1059-1092. 20. Rush D, Cassano P. Relationship of ciga- rette smoking and social class to birth weight and perinatal mortality among all births in Britain, 5-11 April 1970. J Epidemiol Community Health. 1983;37: 249-255. 21. Hebel JR, Fox NL, Sexton M. Dose- response of birth weight to various mea- sures of maternal smoking during preg- nancy. J Clin Epidemiol. 1988;41:483- 489. 22. Lieberman E, Gremy I, Lang JM, Cohen AP. Low birthwcight at term and the timing of fetal exposure to maternal smoking. Am J Public Health. 1994;84:1127-1131. 23. Floyd RL, Rimer BK, Giovino GA, Mullen PD, Sullivan SE. A review of smoking in pregnancy: effects on pregnancy outcomes and cessation efforts. Annu Rev Public Health. 1993;14:379-411. 24. Susser M. Review of dietary behavior during pregnancy. In: Dobbing J, ed. Maternal Nutrition in Pregnancy: Eating for Two. London, Eng: Academic Press; 1981:64-65. 25. Yu SM, Nagey DA. Validity of serf- reported pregravid weight. Ann Epidemiol. 1992;2:715-721. 26. Abrams B, Newman V, Key T, Parker J. Maternal weight gain and prelerm delivery. Obstet GynecoL 1989;74:577-583. 27. Siega-Riz AM, Adair LS, Hobel CJ. Institute of Medicine maternal weight gain recommcndat/ons and pregnancy outcome in a predominantly Hispanic population. Obstet Gynecol. 1994;84:56.%-573. 28. Susser M. Maternal weight gain, infant birth weight, and diet: causal sequences. Am J Clin Nutr. 1991;53:1384-1396. 29. King JC, Butm NF, Bronstein MN, Kopp LE, Lindquist SA. Energy metabolism during pregnancy: influence of maternal energy status. Am J Clin Nutr. 1994; 59(suppl):4395-4455. 0 O~ o.~ O~ 0 596 Amer~c~m Journa~ of Public Haa/th April 1997. Vol. 87, No. 4