Background
Women with at least 1 prior occurrence of premature birth often have demographic and medical risk factors that are not modifiable. However, smoking cessation could be a targeted intervention in which a woman with a history of premature birth may be able to reduce her future risk of recurrence.
Objective
This study aims to assess how trimester-specific smoking patterns influence the risk of recurrent premature birth.
Study Design
This was a population-based retrospective cohort study of singleton nonanomalous live births in Ohio, 2006–2012 using vital statistics birth records. This analysis was limited to women with at least 1 prior premature birth. Rates of birth <37 weeks were compared among nonsmokers, women who smoked in the 3 months prior to pregnancy and quit in the first vs quit in the second vs quit in the third trimester. Multivariate logistic regression analyses assessed the association between smoking cessation at various time points in pregnancy and recurrent premature birth while adjusting for maternal race, education, Medicaid enrollment, and marital status.
Results
We analyzed the outcomes of 36,432 women with a prior premature birth who subsequently delivered at 20–42 weeks. One third of women with a prior premature birth smoked during pregnancy. Of smokers, 16% quit early in the first trimester, 7% quit in the second, 5% quit in the third trimester, and 72% smoked throughout pregnancy. The rate of recurrent premature birth in nonsmokers was high 28% in this cohort. Smoking in pregnancy with cessation in the first or second trimester was not significantly associated with an increase in recurrent premature birth rates (first trimester, 29% adjusted odds ratio, 0.97 [95% confidence interval, 0.9–1.1], and second trimester, 31% adjusted odds ratio, 1.10 [95% confidence interval, 0.9–1.3], respectively). However, quitting late in pregnancy (third trimester) was associated with a high rate (43%) of delivery <37 weeks, adjusted odds ratio, 1.81 (95% confidence interval, 1.48–2.21). Continued smoking throughout pregnancy was also associated with an increased recurrent premature birth (32%), adjusted odds ratio, 1.14 (95% confidence interval, 1.07–1.22), despite adjustment for concomitant premature birth risk factors.
Conclusion
Smoking cessation in pregnancy and its relationship to preterm birth has been studied extensively, and it is widely accepted that smoking in pregnancy increases preterm birth rates. However, this study provides novel information quantifying the risk of recurrent preterm birth and stratifies the increased risk of recurrent preterm birth by trimester-specific smoking behavior. Although women with even 1 prior premature birth are at an inherently high risk of recurrence, women who stopped smoking early in the first 2 trimesters experienced similar preterm birth rates compared with nonsmokers. However, delayed smoking cessation or smoking throughout pregnancy significantly increased recurrent premature birth risk. Smoking cessation is a potential modifiable risk factor for recurrent preterm birth in high-risk pregnancies. This study highlights the importance of early pregnancy smoking cessation in those at especially high risk, women with a prior preterm birth.
Preterm birth, defined as birth prior to 37 weeks of gestation, is the leading cause of perinatal morbidity and mortality in developed countries. In the United States, there are a variety of maternal and fetal characteristics associated with spontaneous preterm birth and recurrent preterm birth.
Differences in maternal demographic characteristics, pregnancy history, nutritional status, and adverse behaviors have all been identified as factors affecting the rate of spontaneous preterm birth and recurrent preterm birth in given populations. Whereas factors such as race and ethnicity are fixed, behavioral factors such as tobacco smoking, alcohol use, and drug abuse may be targeted areas for change.
Targeting tobacco abuse during pregnancy as one of the modifiable risk factors for the prevention of recurrent spontaneous preterm birth is of considerable importance because of the prevalence of its use in this high-risk population. Nationwide in 2010, as many as 23% of women smoked tobacco in the 3 months prior to pregnancy, and 54% of those women quit smoking during pregnancy. In the state of Ohio, only 24% of women who smoked tobacco during the three months prior to pregnancy quit before conceiving, and 16% of those smokers continued to smoke through the third trimester of pregnancy.
Pregnancy is a time in a woman’s life when she may more strongly consider behavioral changes, and indeed, smoking cessation interventions have been demonstrated to significantly reduce the number of mothers who continue to smoke into the third trimester of pregnancy. However, some mothers may decline intervention if they believe that their pregnancy is already too far along for smoking cessation to make a difference in birth outcome. It would therefore be advantageous to offer firm data supporting smoking cessation at any gestational age.
This study seeks to assess how trimester-specific smoking patterns may influence recurrent preterm birth risk. Although there have been studies that compare risk of preterm birth among never-smokers, smokers who quit during pregnancy, and smokers who smoke throughout pregnancy, the influence of smoking pattern changes on preterm birth in the highest-risk group, women with a prior preterm birth, has not previously been reported. We therefore sought to perform a large-scale, population-based study of recent US births to determine whether timing of smoking cessation during pregnancy influences the risk of recurrent preterm birth.
Materials and Methods
The protocol for this study was approved by the Human Subjects Institutional Review Board of the Ohio Department of Health, and a deidentified data set was provided for analysis. This study was exempt from review by the Institutional Review Board at the University of Cincinnati (Cincinnati, OH).
We performed a population-based, retrospective cohort study including birth certificate data of all live births including women with preterm birth that occurred in Ohio over a 7 year period (2006–2012). Analyses were limited to singleton nonanomalous live births between 20 and 42 weeks’ gestation to women with a prior preterm birth <37 weeks and with available data on smoking history.
We aimed to compare the exposure of trimester-specific, self-reported smoking behaviors during pregnancy with the outcome of recurrent preterm birth rates. The US birth certificate contains data on maternal smoking behaviors quantified by the number of cigarettes or packs of cigarettes smoked during 4 time periods: 3 month before pregnancy, the first 3 months of pregnancy, second 3 months of pregnancy, and the third trimester of pregnancy.
From these groups, subjects were categorized into 1 of 5 exposure groups based on their self-reported smoking behaviors: nonsmokers, quit smoking in the first trimester (smoked in the 3 months prior to conception but reported no smoking in the first trimester or after); quit smoking in the second trimester (smoked before conception and the first trimester but reported no smoking in the second trimester or after); quit smoking in the third trimester (smoked before conception and the first and second trimester but reported no smoking in the third trimester); and smoked throughout pregnancy (smoked preconception and in all trimesters).
Women who smoked more than 1 cigarette per day were considered smokers for the purpose of this study. Despite the subjective nature of the self-reported data, Pickett et al demonstrated the reliability of self-reported reproductive smoking behaviors.
The primary outcome was recurrent preterm birth defined as birth at <37 weeks’ gestation in women with a history of 1 or more preterm births <37 weeks. Data on prior preterm birth were recorded in the 2003 version of the US birth certificate, which was utilized for all births in the data set used for this study.
Gestational age was defined using the US Vital Statistics Birth Certificate best obstetric estimate variable that combines dating of the last menstrual period and the earliest ultrasound measurement. Nonsmokers were the referent group for exposure comparisons, and term birth at 37–42 weeks was the referent group for preterm birth outcome comparisons.
We compared demographic, socioeconomic, prenatal, and maternal health characteristics among all 5 exposure groups. Dichotomous variables were compared using χ 2 tests, and continuous variables were compared using an analysis of variance. We used a multinominal logistic regression analysis with estimated adjusted ORs to assess the dose-response relationship between cigarettes smoked per trimester and recurrent preterm birth rates after adjusting for the confounding influences of the mother’s race, Medicaid enrollment, and marital status. Covariates were selected using significant differences found during univariate analyses of the exposure groups, those with biological plausibility, and significance within the adjusted model.
Analyses were performed using STATA 12.1 software (StataCorp, College Station, TX). Results are reported as rates with associated P values and odds ratios (ORs) with 95% confidence intervals (CIs). Comparisons were considered statistically significant if the value was P < .05 or the 95% CI did not include the null value 1.0.
Results
During the study period 2006–2012, there were a total of 1,034,552 live births in the state of Ohio. After exclusions of births with fetal anomalies, multiple gestation births, and births to mothers with no prior preterm birth, our study population was limited to 36,432 live births to women with a prior preterm birth. There were a few cases (n = 897 [2.5%]) with missing smoking data or smoking data that did not fall into any of the 5 exposure groups; thus, 35,535 is the total number of births included in this analysis.
The reference group of nonsmokers comprised 68.4% of births in the study period with nearly one third of women having smoked cigarettes. Of women who reported some smoking during the study period, only a minority (16.1%) quit smoking in the first trimester. Seven and a half percent quit smoking in the second trimester, 4.7% quit smoking in the third, and the greatest majority (71.7%) smoked throughout the entire pregnancy.
The majority of births during this study period were to non-Hispanic white mothers (n = 24,947 [70.2%] P < .01) who were more frequently smokers than nonwhite mothers (n = 8012 [76.4%] compared with n = 2480 [23.6%] P < .01, Table 1 ). Non-Hispanic white mothers were also more likely to smoke throughout the entire pregnancy when compared with nonwhite mothers (n = 5902 [78.4%] compared with n = 1614 [21.4%] P < .01).
| Characteristics | Never-smokers (n = 25,043) | Quit smoking in the first trimester (n = 1689) | Quit smoking in the second trimester (n = 784) | Quit smoking in the third trimester (n = 492) | Smoked throughout pregnancy (n = 7527) |
|---|---|---|---|---|---|
| Demographic factors | |||||
| Maternal race | |||||
| Non- Hispanic white | 16,905 (66.6%) | 1203 (4.7%) | 563 (2.2%) | 344 (1.4%) | 5902 (23.6%) |
| Non- Hispanic black | 6450 (72.4%) | 410 (4.6%) | 182 (2.0%) | 122 (1.4%) | 1390 (15.6%) |
| Hispanic | 1170 (77.5%) | 56 (3.7%) | 31 (2.1%) | 21 (1.4%) | 188 (12.5%) |
| Other | 466 (86.5%) | 17 (3.1%) | 6 (1.1%) | 3 (0.6%) | 36 (6.7%) |
| Parity | 2 (1,3) | 2 (1,3) | 2 (1,3) | 2 (1,3) | 2 (1,3) |
| Maternal age, mean, y | 29.2 ± 5.7 | 27.4 ± 5.0 | 26.6 ± 5.2 | 26.7 ± 5.5 | 27.1 ± 5.2 |
| Age group, y | |||||
| Younger than 20 | 778 (62.6%) | 52 (4.2%) | 29 (2.3%) | 32 (2.6%) | 299 (24.1%) |
| 20–34 | 19,493 (66.7%) | 1464 (5.0%) | 678 (2.3%) | 404 (1.4%) | 6492 (22.2%) |
| 35 or older | 4772 (80.0%) | 173 (2.9%) | 77 (1.3%) | 56 (0.9%) | 736 (12.4%) |
| Socioeconomic factors | |||||
| Less than high school diploma | 3868 (15.5%) | 289 (17.1%) | 172 (21.9%) | 147 (29.9%) | 2762 (36.7%) |
| Married | 15,978 (63.8%) | 740 (43.8%) | 290 (37.0%) | 145 (29.5%) | 2376 (31.6%) |
| Women, Infants, and Children enrollment | 10,077 (40.2%) | 959 (56.8%) | 489 (62.4%) | 303 (61.6%) | 5003 (66.5%) |
| Medicaid | 9626 (38.4%) | 952 (56.4%) | 511 (65.2%) | 346 (70.3%) | 5630 (74.8%) |
| Prenatal care | |||||
| Limited (≤5 visits) | 2731 (10.9%) | 190 (11.3%) | 102 (13.0%) | 120 (24.4%) | 1530 (20.3%) |
| Early prenatal care (≤12 wks gestation) | 13,503 (53.9%) | 927 (54.9%) | 378 (48.2%) | 199 (40.5%) | 3092 (41.1%) |
| Late prenatal care (>20 wks gestation) | 1621 (6.5%) | 95 (5.6%) | 65 (8.3%) | 40 (8.1%) | 898 (11.9%) |
| Maternal health indicators | |||||
| Prior cesarean delivery | 7785 (31.1%) | 477 (28.2%) | 228 (29.1%) | 137 (27.9%) | 2061 (27.4%) |
| Obesity, prepregnancy BMI ≥30 kg/m 2 | 6780 (27.1%) | 490 (29.0%) | 203 (25.9%) | 112 (22.8%) | 1558 (20.7%) |
| Gestational weight gain mean | 28.9 ± 17.9 | 31.9 ± 19.1 | 32.3 ± 19.0 | 26.8 ± 17.7 | 27.1 ± 18.4 |
| Chronic hypertension | 1084 (4.3%) | 80 (4.7%) | 21 (2.7%) | 21 (4.3%) | 250 (3.3%) |
| Pregestational diabetes mellitus | 462 (1.8%) | 27 (1.6%) | 14 (1.8%) | 6 (1.2%) | 129 (1.7%) |
| Gestational hypertension/preeclampsia | 1777 (7.1%) | 101 (6.0%) | 48 (6.1%) | 20 (4.1%) | 312 (4.2%) |
| Gestational diabetes mellitus | 2110 (8.4%) | 143 (8.5%) | 71 (9.1%) | 41 (8.3%) | 440 (5.9%) |
Mothers younger than 20 years were more likely to be smokers compared with mothers older than 20 years (n = 620 [44.3%] compared with n = 10,080 [29.3%] P < .01). Women who received limited prenatal care (≤5 visits) or presented after 20 weeks’ gestation for prenatal care were more likely be smokers than women who received early prenatal care at or prior to 12 weeks’ gestation (n = 3040 [41.1%] compared with n = 4596 [25.4%] P < .01).
Similarly, women who had limited or late prenatal care were also nearly twice as likely to smoke for the duration of the pregnancy when compared with those who initiated prenatal care in the first trimester (32.8% compared with 17.1%, P < .01).
Table 1 displays additional maternal characteristics found to be associated with undesirable smoking behaviors of smoking throughout pregnancy or quitting late in pregnancy. These characteristics include low maternal education level; unmarried status; Women, Infants, and Children enrollment; and Medicaid-funded delivery.
Seventy-eight percent (n = 4325) of women who self-identified as heavy smokers (more than 20 cigarettes per day) smoked throughout the entire pregnancy ( Table 2 ). Women who smoked throughout the entire pregnancy on average smoked more cigarettes per day in the preconception period than the women who quit smoking in the first trimester (17.3 ± 10.7 compared with 11.7 ± 9.7). However, women who smoked throughout the duration of their pregnancy reported a reduction in their average number of cigarettes smoked per day from 17.3 ± 10.7 in the preconception period to 10.6 ± 8.2 during pregnancy.
| Variables | Never-smokers (n = 25,043) | Quit smoking in the first trimester (n = 1689) | Quit smoking in the second trimester (n = 784) | Quit smoking in the third trimester (n = 492) | Smoked throughout pregnancy (n = 7527) |
|---|---|---|---|---|---|
| Preconception smoking behaviors | |||||
| Average number of cigarettes smoked per day | 0 | 11.7 ± 9.7 | 14.8 ± 10.8 | 17.2 ± 11.1 | 17.3 ± 10.7 |
| Heavy smokers >20/day | 0 | 537 (31.8%) | 339 (43.2%) | 286 (58.1%) | 4325 (57.5%) |
| Smoking behaviors in final 3 months of smoking | |||||
| Average number of cigarettes smoked per day | 0 | 11.7 ± 9.7 | 9.4 ± 7.9 | 8.6 ± 8.9 | 10.6 ± 8.2 |
| <10 | 0 | 672 (39.8%) | 409 (52.2%) | 280 (56.9%) | 3008 (40.0%) |
| 19–20 | 0 | 480 (28.4%) | 212 (27.0%) | 131 (26.6%) | 2832 (37.6%) |
| ≥20 | 0 | 537 (31.8%) | 163 (20.8%) | 81 (16.5%) | 1687 (22.4%) |
Birth outcomes significantly associated with smoking throughout pregnancy compared with nonsmokers included intrauterine growth restriction, birthweight less than the 10th percentile for gestational age, low birthweight <2500 g, an increased rate of neonatal intensive care unit admission, and higher neonatal transfer rate to a tertiary care facility ( P < .01). The absolute rates of other maternal and fetal birth outcomes were similar among the smoking groups; however, the P values were statistically significant at P < .01, likely as a result of the large sample size ( Table 3 ).
| Variables | Never-smokers (n = 25,043) | Quit smoking in the first trimester (n = 1689) | Quit smoking in the second trimester (n = 784) | Quit smoking in the third trimester (n = 492) | Smoked throughout pregnancy (n = 7527) |
|---|---|---|---|---|---|
| Maternal | |||||
| Spontaneous vaginal delivery | 14,541 (58.0%) | 966 (57.2%) | 449 (57.3%) | 276 (56.1%) | 4510 (59.9%) |
| Operative vaginal delivery | 822 (3.3%) | 75 (4.4%) | 31 (4.0%) | 14 (2.9%) | 239 (3.2%) |
| Cesarean delivery | 9648 (38.5%) | 647 (38.3%) | 304 (38.8%) | 201 (40.9%) | 2771 (36.8%) |
| Premature rupture of membranes | 1542 (6.2%) | 118 (7.0%) | 59 (7.5%) | 71 (14.4%) | 548 (7.3%) |
| Fetal growth restriction, BW <10th percentile | 3225 (12.9%) | 230 (13.6%) | 119 (15.2%) | 106 (21.5%) | 1811 (24.1%) |
| Meconium | 979 (3.9%) | 66 (3.9%) | 22 (2.8%) | 22 (4.5%) | 372 (4.9%) |
| Induction of labor | 5148 (20.6%) | 394 (23.3%) | 178 (22.7%) | 76 (15.5%) | 1404 (18.7%) |
| Fetal intolerance to labor | 2061 (8.2%) | 141 (8.4%) | 59 (7.5%) | 58 (11.8%) | 487 (6.5%) |
| Fetal | |||||
| BW, mean ± SD | 2964 ± 764 | 2951 ± 745 | 2841 ± 819 | 2343 ± 1061 | 2743 ± 667 |
| LBW <2500 g | 5353 (21.4%) | 354 (21.0%) | 198 (25.3%) | 235 (47.8%) | 2314 (30.7%) |
| ELBW <1500 g | 1408 (5.6%) | 85 (5.0%) | 61 (7.8%) | 133 (27.0%) | 394 (5.2%) |
| NICU admission | 4464 (17.8%) | 309 (18.23%) | 153 (19.5%) | 172 (35.0%) | 1462 (19.4%) |
| Infant transfer | 1533 (6.1%) | 92 (5.5%) | 70 (8.9%) | 57 (11.6%) | 640 (8.5%) |
| 5 minute Apgar score <7 | 1110 (4.4%) | 64 (3.8%) | 44 (5.6%) | 83 (16.9%) | 330 (4.4%) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
