You are here

Article Text

Article menu

Download PDFPDF

Research paper
Benefits of smoking bans on preterm and early-term births: a natural experimental design in Switzerland
  1. Ana M Vicedo-Cabrera1,2,
  2. Christian Schindler1,2,
  3. Dragana Radovanovic3,
  4. Leticia Grize1,2,
  5. Fabienne Witassek3,
  6. Julia Dratva1,2,
  7. Martin Röösli1,2,
  8. Laura Perez1,2
  1. 1Department of Epidemiology and Public Health, Swiss Tropical and Public Health Institute, Basel, Switzerland
  2. 2University of Basel, Basel, Switzerland
  3. 3AMIS Plus Data Center. Epidemiology, Biostatistics and Prevention Institute, University of Zurich, Zurich, Switzerland
  1. Correspondence to Dr Ana M Vicedo-Cabrera, Department of Epidemiology and Public Health, Swiss Tropical and Public Health Institute, Socinstrasse 57, P.O. Box, Basel CH-4002, Switzerland; anamaria.vicedocabrera{at}unibas.ch

Abstract

Background Birth outcomes are relevant for future children's heath. Capitalising on a natural experimental design in Switzerland, we evaluated how regional smoking bans introduced at different time points affected birth outcomes, including preterm and early-term births.

Methodology We used birth registry data of all singleton neonates born in Switzerland (2007–2012). We developed canton-specific interrupted time-series followed by random meta-analysis to evaluate the benefits of smoking bans on preterm (<37 gestational weeks) and early-term (37–38 gestational weeks) births. Heterogeneity across type of ban and contextual characteristics was explored through metaregression. A time-to-event approach was used for evaluating duration of pregnancy under the smoking bans and effects, taking into account individual maternal factors.

Results We observed a decrease in the risk of preterm birth of 3.6% (95% CI, −9.3% to 2.5%), and early-term birth of 5.0% (95% CI −7.5% to −2.5%). Results showed a clear dose–response relationship. Greater risk reductions were obtained for preterm births in areas with more comprehensive bans (−6.8%; 95% CI −12.1% to 0.1%), and for pregnancies with the longest gestational time under smoking bans (HR, 0.991; 95% CI 0.984 to 0.997 per 10% increase in duration). Benefits were unequal across outcomes and characteristics of cantons and mothers.

Conclusion Smoking bans resulted in improved birth outcomes in Switzerland with cantons that adopted more comprehensive smoking bans achieving greater benefits. Early-term births constitute a previously ignored though important group.

  • Secondhand smoke
  • Prevention
  • Public policy

https://doi.org/10.1136/tobaccocontrol-2015-052739

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    In the last decade, the introduction of smoking bans in public and work places had lead to cardiorespiratory benefits in diverse populations worldwide.1–3 These effects are well documented. Fewer studies have explored the relationship between smoking bans and birth outcomes.4 ,5

    Prenatal exposure to tobacco smoke is one of the most important modifiable determinants of adverse birth outcomes, such as preterm births (<37 gestational weeks).6–13 Preterm births are associated with perinatal and infant mortality and morbidity,14–17 as well as long-term effects that persist into adulthood.18–21 Past studies from different settings have observed reductions in the risk of preterm birth of 3–25% following the introduction of smoking bans,4 ,5 ,22 whereas, others failed to find consistent associations.23 Nonetheless, important questions remain regarding the extent and magnitude of these benefits across populations and contexts. Maternal tobacco exposure has been recently associated with early-term births (37–38 gestational weeks),13 a new risk group of growing public health interest given increased mortality and morbidity risk compared to full-term pregnancies.24–27 The benefits of smoking ban for this risk group has never been evaluated. The benefits of smoking ban on birth outcome are likely mediated through a reduction of active maternal smoking, and the mother's exposure to secondhand smoke (SHS).28 The contributions of each pathway may depend on the context and characteristics of the population that need to be better understood.

    In Switzerland, smoking in public places and work places is regulated at both the federal and cantonal level (cantons are the administrative units of the Swiss Federal state). For example, on 1 May 2010, the Swiss Parliament approved a nationwide federal smoking ban covering most indoor public places and work places, with several exceptions in the hospitality sector, including authorised smoking establishments with <80 m2 and dedicated smoking rooms in larger establishments. When the federal ban was enacted, 12 of 26 cantons had previously introduced their own laws, most with a higher level of protection for hospitality workers. Consequently, there is a patchwork of laws with different degrees of protection, implemented at different times, within a small geographical area. This immediately suggests a natural experimental design, where the timing of the implementation of bans across cantons can be used as the ‘random’ intervention allocation time, and presents a unique opportunity to examine the health benefits of smoking bans.

    We capitalised on this opportunity, and evaluated how smoking bans affected birth outcomes, including preterm and early-term births at the cantonal and federal level. We explored patterns of benefits between different types of bans implemented, timing of the introduction of the smoking ban in pregnancies and other contextual and individual maternal characteristics to better understand the potential pathways contributing to change in tobacco smoke exposure among the studied population.

    Methods

    Study population and outcome definition

    Our study population included all singleton neonates born in Switzerland whose mothers resided in the country, and whose gestational age at birth was between 22 and 42 complete weeks. Birth data was obtained from birth certificates collected by the Swiss Federal Office of Statistics (Bundesamt für Statistik, BFS). We included only births from 1 January 2007 onwards, because information on gestational age at birth was available for most cantons only after this point; information was available in the cantons Vaud and Valais in 2008, and Geneva in 2009. Gestational age at birth, reported in weeks, was determined by ultrasound examinations. Additional information was collected, including the neonate's gender, and the mother's age, parity, year of conception, civil status, nationality and canton of residence. All analyses were conducted for two separate birth outcomes: preterm births, defined as infants born with a gestational age <37 weeks, and early-term births, defined as infants born with a gestational age between 37 and 38 weeks.29

    Before-and-after analysis

    We performed a canton-specific interrupted time-series analysis to estimate changes in risk of preterm and early-term birth following the introduction of the smoking bans. Quasi-Poisson regression models were applied including the weekly count of each outcome, and the smoking ban impact was modelled with a step function (ie, assigning a value of 1 to weeks of the ban period, and 0 to the preban weeks). With this design, each canton-specific preban series acts as its own ‘comparison or control group’. We controlled for potential long-term and seasonal trends with a linear time function (measured in weeks) and a season indicator (categorical variable with 4 strata, 1 for each season). To properly account for the temporal variation of the number and characteristics of pregnancies at risk, we applied the modification of the pregnancies-at-risk approach, previously described by Vicedo-Cabrera and colleagues.30 ,31 In brief, it consists of the inclusion of an outcome-specific offset with the weekly number of ongoing gestations at risk, along with a weekly correcting factor to correct for the baseline probability of giving birth according to the mean gestational age of the pregnancies at risk included in the offset (further details in online supplementary methods S1). Results were expressed as per cent change in risk derived from the relative risk (RR) estimates.

    In a second stage, we performed a random-effects meta-analysis using the canton-specific results to obtain a nationwide per cent change in risk estimates. We excluded the cantons of Neuchatel, Geneva and Ticino, because their preban period was <1 year (see online supplementary figure S1). We also performed several sensitivity analyses, to assess the robustness of our results: inclusion of all cantons, non-linear time trend functions, introduction of temporal variation of tobacco taxes, potential postban gradual changes in risk, and effects on full-term births, defined as infants born between 39 and 42 gestational weeks (detailed explanation in online supplementary methods S2).

    Metaregression

    We performed a metaregression analysis to explore potential modifications of impacts according to the type of smoking ban and cantonal characteristics. We hypothesised that more strict bans resulted in larger reduction of SHS exposure in public places, and that cantons with higher levels of SHS exposure, preban, also benefited most from the smoking bans. On the contrary, we hypothesised that cantons with higher prevalence of smoking, or with less healthy population, benefited less from the smoking bans because of competing risks with regard to tobacco exposure (eg, unchanged maternal smoking prevalence or tobacco smoke exposure at home).

    The indicators included in the metaregression were (1) indicator of extent of ban with cantons following only the federal smoking ban (‘Federal Law’), versus those with more prohibitive bans and additional occupational exposure restrictions (‘Federal Law + restrictions’); (2) proxy indicators for the preban level of SHS exposure in public places and work including mean number of third-sector establishments, previous antitobacco measures, and a cantonal socioeconomic status (SES) score (see paragraph below) and (3) proxy indicator for active maternal smoking and health awareness including smoking prevalence in the general public and among women of child-bearing age (CBA, 18–44 years old) before smoking bans, language (proxy of the disparities in lifestyle and cultural attitudes across regions in Switzerland), and health score for the canton (see paragraph below).

    The contextual SES and health score were extracted from a principal components analysis (PCA). The PCA was performed on a priori selected cantonal indicators as detailed in the online supplementary methods S3, table S1–S3. The PCA consists in evaluating the correlation between indicators to extract non-correlated composite factors. For our study, three components were extracted explaining 85% of the overall variance. One was highly correlated with indicators related to the economical position of the canton (foreign population, urban population, gross domestic product, educational level, population density, family size and status index). The other two components were correlated with health and unemployment variables and they were summarised in a composite index hypothesised as capturing the health profile of the cantonal population. Scores were included in the metaregression after categorising into tertiles. Other indicators were included as below or above federal median, or dichotomised as necessary.

    Time-to-event analysis

    We conducted a time-to-event analysis to assess whether the risk of preterm and early-term births decreased with increased duration of pregnancy under the smoking bans, whether the gestational trimester into which the smoking ban was introduced represented a window of susceptibility, and how these effects were modified by maternal characteristics.

    Gestational age was considered as the time axis, and the third trimester (beginning in week 27) as the risk period, so each pregnancy entered the risk set at the beginning of this trimester, and time to birth was analysed in consecutive periods of 1 week, with censoring up to the gestational week when birth occurred. Only births >26 gestational weeks were retained in this analysis due to left-truncation. To avoid fixed-cohort bias, we excluded births with a conception date earlier than 26 weeks before the start of the study, or later than 42 weeks before the end of the study (detailed explanation in online supplementary methods S4).32

    We considered duration of exposure to smoking bans separately for each of the three gestational trimesters. For the first and second trimester, exposure was defined as the proportion of days which occurred during smoking bans. To ensure that the exposure preceded the outcome, for the third trimester, this was defined as the proportion of days between the start of the third trimester and the end of the preceding week, and was updated every week. To explore windows of susceptibility to exposure, duration indicators were included separately in models. To represent cumulative exposures, we combined proportions for the second and third trimesters, and for all three trimesters.

    We applied Cox proportional hazard regression models to obtain HRs of preterm and early-term births for different proportion terms. All models were adjusted by gender and for individual maternal characteristics, such as age, parity, civil status, nationality and canton of residence. The proportional hazard assumption was tested through Schoenfeld residuals. In addition, we estimated the HR for the time under the ban for all trimesters combined, according to different categories of maternal characteristics, and to the contextual variables previously applied in the metaregression analyses.

    Results

    Description of the population

    Totally, 446 492 births were included in our study; 5.5% of total births were preterm, while 27.1% were early-term (table 1). According to the inclusion criteria defined for each substudy (see online supplementary figure S2), 400 784 births were included in the before-and-after analysis, and 413 516 births were included in the time-to-event analysis, with similar percentages for preterm and early-term births (5.4% and 26.8%, and 5.2% and 27.4%). Higher prevalence of both outcomes was observed in male neonates, and for mothers >35 years old. Higher prevalence of preterm births was observed in primiparous, unmarried women of Swiss nationality; early-term births showed no such relationship.

    View this table:
    Table 1

    Description of the study population

    Changes in the risk of birth outcomes following the introduction of smoking bans

    Following the introduction of smoking bans, we found a decrease in the risk of preterm births of 3.6% (95% CI −9.3% to 2.5%), and a decrease in the risk of early-term births of 5.0% (95% CI −7.5% to −2.5%) nationwide (figure 1). Observed decreases did not vary significantly across cantons (preterm births, I2=7.42%, p value=0.433; early-term births, I2=0.01%, p value=0.395). Thirteen cantons showed decreased risk for preterm births, and 17 cantons showed decreased risk for early-term births, with consistent beneficial effects in the most populated cantons.

    Figure 1
    Figure 1

    Forest plot with the meta-analytic results of the canton-specific estimates of the change in risk (95% CI) of preterm and early-term birth associated with the introduction of the smoking ban. Results are ordered according to the total population in 2010 in each canton (descending) (Cantons of Geneva, Neuchatel and Ticino were excluded). All models adjusted by: long-term and seasonal trends, correcting factor of the offset. Residual (restricted) maximum likelihood to estimate the between-study variance, and Knapp and Hartung adjustment was applied. Canton abbreviations: Aargau (AG); Appenzell Innerrhoden (AI); Appenzell Ausserrhoden (AR); Bern (BE); Basel-Landschaft (BL); Basel- Stadt (BS); Fribourg (FR); Glarus (GL); Graubunden (GR); Jura (JU); Luzern (LU); Nidwalden (NW); Obwalden (OW); St. Gallen (SG); Schaffhausen (SH); Solothurn (SO); Schwyz (SZ); Thurgau (TG); Uri (UR); Vaud (VD); Valais (VS); Zug (ZG); Zürich (ZH).

    The risk of preterm and early-term births varied with the type of smoking ban introduced, and other contextual factors, however, differences did not reach statistical significance (figure 2). For preterm births, effects were stronger in cantons with more restrictive bans (−6.2%; 95% CI −12.1% to 0.1%), with previous antitobacco measures (−4.2%; 95% CI −9.9% to 1.8%), with more third-sector establishments (−5.4%; 95% CI −11.3% to 0.9%), with intermediate SES scores (−10.4%; 95% CI −17.7% to −2.4%) and higher health scores (−6.6%; 95% CI −19.2% to 8.0%). For early-term births, effects were stronger in cantons with lower SES scores (−7.0%; 95% CI −13.0% to −0.6%) and lower smoking prevalence in CBA women (−6.4%; 95% CI −9.6% to −3.1%).

    Figure 2
    Figure 2

    Per cent change in risk estimates (95% CI) of preterm and early-term birth associated with the introduction of the smoking ban according to contextual and smoking-related indicators of each canton (Geneva, Neuchatel and Ticino were excluded). Type of smoking ban: Fed. Law: cantons which followed only the Federal Smoking Ban, Fed. Law+restr: cantons which applied additional restrictions to the Federal Smoking Ban. Categories of smoking prevalence (adult population (general) and child-bearing-aged women (CBA-women)), and number of establishments of the third sector (proxy of hospitality venues) defined as below or above the median value among all cantons. Categories of socioeconomic status (SES) and health status based on tertiles of the corresponding scores derived from the principal components analysis (SES/health status: lower/better (1st tertile), intermediate (2nd tertile), higher/worse (3rd tertile).

    Effects were slightly lower if all cantons were included in the meta-analysis, or when a non-linear time trend was introduced (see online supplementary table S4). The greatest changes in risk were observed immediately following the introduction of smoking bans; gradual changes in risk following smoking ban showed no consistent pattern. We found no changes in full-term birth rates following smoking bans (−0.8%; 95% CI −2.5% to 0.9%).

    Duration and window of susceptibility to smoking bans

    Most of the pregnancies ended before (42.0%) or started after (43.8%) the introduction of smoking bans. When the smoking bans were introduced, 4.9% of pregnancies were in the first trimester, 4.7% in the second trimester, and 4.6% in the third trimester (results not shown).

    We observed a 0.9% decrease in the risk of preterm birth for each 10% increase in the proportion of days of pregnancy that occurred during smoking bans (HR 0.991; 95% CI 0.984 to 0.997). Estimates were less consistent and smaller for early-term births (HR 0.997; 95% CI 0.994 to 1.000) (table 2). Similar effects were observed for preterm births if the first or second trimester was considered the relevant window of exposure (HR 0.992; 95% CI 0.986 to 0.997; HR 0.991; 95% CI 0.984 to 0.997). Instead, only postban exposure during the third trimester showed consistent results for early-term births (HR 0.996; 95% CI 0.994 to 0.998).

    View this table:
    Table 2

    HRs (95% CI) of preterm (26–36 gestational weeks) and early-term births (37–38 gestational weeks) for a 10% increase in the proportion of pregnancy time under smoking ban for the combined and single susceptibility periods

    For preterm births, reductions in risk were essentially independent of individual maternal characteristics (figure 3). For early-term births, reductions were larger and significant for mothers who were >35 years old (HR 0.993; 95% CI 0.987 to 0.998), multiparous (HR 0.996; 95% CI 0.992 to 1.000) or married (HR 0.996; 95% CI 0.992 to 1.000). Effects across contextual factors varied by outcome (see online supplementary figure S2). For early-term births, we found notable differences compared to those from the metaregression; there were stronger effects in cantons with high smoking prevalence, high SES score and worse health score. For preterm births, patterns were rather similar.

    Figure 3
    Figure 3

    HRs (95% CI) of preterm (26–36 gestational weeks) and early-term births (37–38 gestational weeks) for a 10% increase in the proportion of pregnancy time under smoking ban (for the whole gestation), according to different maternal characteristics.

    Discussion

    Our study shows that the introduction of smoking bans in Switzerland was associated with a reduction in the risk of 5% in early-term births and of 3.5%, but not statistically significant, in preterm births. Our study is the first to differentiate between preterm and early-term births, and the first to show a clear dose–response relationship; cantons that adopted more comprehensive smoking bans, and women for which the smoking ban was introduced the earliest in their pregnancy achieved greater improvements. Benefits, however, differed across outcomes and characteristics of cantons and mothers.

    Our findings are consistent with results from past studies that showed reductions in the risk of preterm births of between 3% and 25% following the introduction of smoking bans in different settings.4 ,5 ,22 However, the wide variability of the magnitude of the benefits observed in these studies, along with the inconsistent evidence obtained in other works,23 may be partly explained by the different study populations and the heterogeneous methodologies applied in each case.

    To the best of our knowledge, our study is the first to differentiate between preterm and early-term births. Early-term births may not present the same high risks as preterm births, however, early-term neonates are less physiologically mature, and face greater chances of developing health problems throughout their lives, compared with full-term neonates.24–27 Early-term births are more common than preterm births. Therefore, mothers at risk for early-term births, and early-term neonates, represent a relatively large group, with important implications for public health. Smoking bans or other policy interventions targeted at reducing early-term births may have considerable long-term health and cost benefits.

    Our study is the first to show a clear dose–response relationship between smoking bans and the risk of adverse birth outcomes.33 ,34 Cantons that adopted more comprehensive smoking bans showed greater reductions in the risk of preterm births. Furthermore, our time-to-event analysis suggests that the sooner in pregnancy smoking bans were introduced, the greater the benefits for these high-risk pregnancies, with potential windows of susceptibility in the first and second trimesters. We also found smaller but consistent reductions in early-term risk for exposure during the third trimester. These findings are supported by previous studies, which have shown that reducing smoke exposure during pregnancy improves birth outcomes, and that benefits are linked to when during pregnancy such reduction occurs.35 ,36

    To the best of our knowledge, this is the first investigation exploring the heterogeneity of the smoking ban impact on birth outcomes to try to explain pathways of tobacco exposure changes in the absence of complete individual information. Our results suggest that preterm and early-term births are associated with different at-risk groups that might have responded differently to smoking bans. Reductions in preterm births were observed primarily in cantons with supposedly higher SHS exposure in public places before the ban (lower SES score and more establishments of third sector), with potentially lower maternal smoking (higher health scores and lower prevalence of smoking among women), but with benefits showing across all women independently of individual characteristics. Preterm neonates show more complications during gestation, and may be more susceptible to hazardous environmental factors.13 ,37 ,38 These results are relevant for public health and public policy because they suggest that these high-risk groups may have more largely benefited from measures achieving even small reductions in SHS exposure at population level. There were less strong gradients in risk reductions for early-term births with indicators of SHS exposure in public places. In addition, benefits for this outcome mostly materialised for women who were potentially of higher SES (older, multipara and married), and in cantons with higher active smoking in women. This might suggest that among the risk group of early-term birth, those with greater health awareness (or SES) principally benefited. Past studies have shown that low preban exposure, and more affluent families, benefited the most from smoking bans, and that there is reduction of maternal smoking after smoking bans.39–42 We can speculate that among our subgroup of women at risk of giving early-term births, a pathway of change in exposure to tobacco smoke may have been through smoking cessation. Although we acknowledge that these are speculations that should be taken with caution because these hypotheses cannot be tested due to the lack of specific data on smoking cessation and SHS exposure.

    Our results are also consistent with our initial hypothesis that cantons with worse health status and higher prevalence of smoking in general population obtained fewer benefits from the smoking bans. As stated, this might be due to a lesser reduction in SHS exposure in private environments (homes, cars, etc) in these cantons which we were not able to explore in our study. However, variable changes in exposures from smoking ban at private places have been observed.43–45

    Studies that evaluate the health benefits of smoking bans have been questioned because of potential biases resulting from concurrent secular trends and other temporal factors.46 We conducted two types of analysis to address two independent research questions. The before-and-after study aims at exploring how the risk of birth outcomes changed after the introduction of the ban; the time-to-event analysis evaluates if and how the risk changes according to the pregnancy time spent under the smoking ban. Despite using two very different approaches, we were able to find consistent results across both hypotheses. In addition, we took several steps to control for temporal factors, and tested the robustness of our models through sensitivity analyses. We initially assumed a linear time trend in the time-series analysis; however, we also tested non-linear functions, with no major changes in effects. In addition, we tested our models on full-term births, an outcome unlikely to be associated with smoking bans, obtaining a null effect. The fact that results were homogeneous across regions, despite the fact that smoking bans were implemented at different times and in different ways across 26 cantons, is a strong argument in favour of the robustness of our results and against the presence of confounding. Finally, our time-to-event approach ruled out any potential exposure misclassification in ongoing pregnancies when smoking bans were introduced. We used routinely collected registry-based data, providing nearly complete coverage of the population in all cantons, thus preventing selection bias. Although the individual information contained is limited, the natural experimental design partially overcomes this shortcoming because time-invariant confounders are controlled.

    Important knowledge gaps remain. Smoking ban exposure was assigned according to the canton of residence, because information about workplace was not available. We cannot disregard the possibility of potential exposure misclassification due to differences between the canton of residence and canton of workplace, however, intercanton commuting rates are relatively low in women (only 8% in 2010) according to the data of the BFS. More detailed geographic information could improve our study, and highlight important differences in SHS exposure between residence and workplace in pregnant women.

    Many factors influence the benefits of smoking laws. We could only use indirect evidence to understand the contribution of contextual and individual factors in the benefits of smoking bans. We were unable to collect information about changes in maternal smoking exposure following the introduction of smoking bans, and could not evaluate its contribution in improving birth outcomes. We know that smoking prevalence among CBA women in Switzerland remained fairly stable during the years when the smoking bans were introduced with the number of smoking mothers generally low, and most quitting smoking early in pregnancy (http://www.swisstph.ch/fileadmin/user_upload/Pdfs/swifs/SWIFS_Schlussbericht.pdf, date accessed 28 Sep 2015). A considerable decrease in SHS exposure (including public and private spaces) was observed between 2007 and 2012 in Switzerland (from 16% to 6%) coinciding with the introduction of the bans (http://www.bfs.admin.ch/bfs/portal/en/index/themen/14/02/02/key/03.html). However, we have no information of potential reduction of SHS at home or during specific conditions such as pregnancy. More information about maternal smoking patterns and SHS exposure are needed to reduce this uncertainty. Other policies such as increase in tobacco taxes have been found to be more effective in improving perinatal health than smoking bans.23 However, in our study, smoking ban benefits remained or even increased when considering tobacco taxes, as shown in the sensitivity analysis.

    In conclusion, our study confirms that smoking bans lead to improved birth outcomes including the previously ignored group of early-term births. Benefits were the greatest in cantons with presumably higher exposure to SHS in public places preban, and that could reduce this exposure the most with stricter smoking bans. Patterns of benefits across outcomes, populations and contexts were difficult to interpret given lack of individual data on exposure changes.

    What this paper adds

    • Introduction of smoke-free laws in public places has been associated with population-level cardiorespiratory benefits worldwide. However, a limited number of studies have investigated whether these improvements extend to pregnant women and newborns.

    • The maternal and contextual factors influencing the extent and magnitude of the impact of smoking bans on perinatal health are still little understood.

    • This is the first study to show a clear dose–response relationship between smoking bans and the risk of preterm births with greater benefits in cantons that adopted more comprehensive smoking bans. Early-term births constitute a previously ignored, though important at-risk group, that also benefits from smoking bans.

    References

    1. ↵
      1. Lee PN,
      2. Fry JS,
      3. Forey BA
      . A review of the evidence on smoking bans and incidence of heart disease. Regul Toxicol Pharmacol 2014;70:7–23. doi:10.1016/j.yrtph.2014.06.014
      OpenUrlCrossRefPubMed
    2. ↵
      1. Callinan JE,
      2. Clarke A,
      3. Doherty K, et al
      . Legislative smoking bans for reducing secondhand smoke exposure, smoking prevalence and tobacco consumption. Cochrane Database Syst Rev 2010;(4):CD005992. doi:10.1002/14651858.CD005992.pub2
    3. ↵
      1. Rajkumar S,
      2. Schmidt-Trucksäss A,
      3. Wellenius GA, et al
      . The effect of workplace smoking bans on heart rate variability and pulse wave velocity of non-smoking hospitality workers. Int J Public Health 2014;59:577–85. doi:10.1007/s00038-014-0545-y
      OpenUrlCrossRefPubMed
    4. ↵
      1. Been JV,
      2. Nurmatov UB,
      3. Cox B, et al
      . Effect of smoke-free legislation on perinatal and child health: a systematic review and meta-analysis. Lancet 2014;383:1549–60. doi:10.1016/S0140-6736(14)60082-9
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. McKinnon B,
      2. Auger N,
      3. Kaufman JS
      . The impact of smoke-free legislation on educational differences in birth outcomes. J Epidemiol Community Health 2015;69:937–43. doi:10.1136/jech-2015-205779
      OpenUrlAbstract/FREE Full Text
    6. ↵
      1. Salihu HM,
      2. Wilson RE
      . Epidemiology of prenatal smoking and perinatal outcomes. Early Hum Dev 2007;83:713–20. doi:10.1016/j.earlhumdev.2007.08.002
      OpenUrlCrossRefPubMedWeb of Science
    7. ↵
      1. Qiu J,
      2. He X,
      3. Cui H, et al
      . Passive smoking and preterm birth in urban China. Am J Epidemiol 2014;180:94–102. doi:10.1093/aje/kwu092
      OpenUrlAbstract/FREE Full Text
    8. ↵
      1. Lødrup Carlsen KC,
      2. Jaakkola JJ,
      3. Nafstad P, et al
      . In utero exposure to cigarette smoking influences lung function at birth. Eur Respir J 1997;10:1774–9.
      OpenUrlAbstract
    9. ↵
      1. Leite M,
      2. Albieri V,
      3. Kjaer SK, et al
      . Maternal smoking in pregnancy and risk for congenital malformations: results of a Danish register-based cohort study. Acta Obstet Gynecol Scand 2014;93:825–34. doi:10.1111/aogs.12433
      OpenUrlCrossRefPubMed
    10. ↵
      1. Rehan VK,
      2. Asotra K,
      3. Torday JS
      . The effects of smoking on the developing lung: insights from a biologic model for lung development, homeostasis, and repair. Lung 2009;187:281–9. doi:10.1007/s00408-009-9158-2
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      1. Miyake Y,
      2. Tanaka K,
      3. Arakawa M
      . Active and passive maternal smoking during pregnancy and birth outcomes: the Kyushu Okinawa Maternal and Child Health Study. BMC Pregnancy Childbirth 2013;13:157. doi:10.1186/1471-2393-13-157
      OpenUrlCrossRefPubMed
    12. ↵
      1. Ko TJ,
      2. Tsai LY,
      3. Chu LC, et al
      . Parental smoking during pregnancy and its association with low birth weight, small for gestational age, and preterm birth offspring: a birth cohort study. Pediatr Neonatol 2014;55:20–7. doi:10.1016/j.pedneo.2013.05.005
      OpenUrlCrossRefPubMed
    13. ↵
      1. Brown H,
      2. Speechley K,
      3. Macnab J, et al
      . Biological determinants of spontaneous late preterm and early term birth: a retrospective cohort study. BJOG 2015;122:491–9.
      OpenUrlCrossRefPubMed
    14. ↵
      1. Anderson P,
      2. Doyle LW
      , Victorian Infant Collaborative Study Group. Neurobehavioral outcomes of school-age children born extremely low birth weight or very preterm in the 1990s. JAMA 2003;289:3264–72.
      OpenUrlCrossRefPubMedWeb of Science
    15. ↵
      1. Berkowitz GS,
      2. Papiernik E
      . Epidemiology of preterm birth. Epidemiol Rev 1993;15:414–43.
      OpenUrlFREE Full Text
    16. ↵
      1. Escobar GJ,
      2. Clark RH,
      3. Greene JD
      . Short-term outcomes of infants born at 35 and 36 weeks gestation: we need to ask more questions. Semin Perinatol 2006;30:28–33. doi:10.1053/j.semperi.2006.01.005
      OpenUrlCrossRefPubMedWeb of Science
    17. ↵
      1. Liu L,
      2. Johnson HL,
      3. Cousens S, et al
      . Global, regional, and national causes of child mortality: an updated systematic analysis for 2010 with time trends since 2000. Lancet 2012;379:2151–61. doi:10.1016/S0140-6736(12)60560-1
      OpenUrlCrossRefPubMedWeb of Science
    18. ↵
      1. Mattioli S,
      2. Farioli A,
      3. Legittimo P, et al
      . Tobacco smoke and risk of childhood acute non- lymphocytic leukemia: findings from the SETIL study. PLoS ONE 2014;9:e111028. doi:10.1371/journal.pone.0111028
      OpenUrlCrossRefPubMed
    19. ↵
      1. Thacher JD,
      2. Gruzieva O,
      3. Pershagen G, et al
      . Pre- and postnatal exposure to parental smoking and allergic disease through adolescence. Pediatrics 2014;134:428–34. doi:10.1542/peds.2014-0427
      OpenUrlAbstract/FREE Full Text
    20. ↵
      1. Kovess V,
      2. Keyes KM,
      3. Hamilton A, et al
      . Maternal smoking and offspring inattention and hyperactivity: results from a cross-national European survey. Eur Child Adolesc Psychiatry 2015;24:919–29. doi:10.1007/s00787-014-0641-9
      OpenUrlCrossRefPubMed
    21. ↵
      1. Dior UP,
      2. Lawrence GM,
      3. Sitlani C, et al
      . Parental smoking during pregnancy and offspring cardio-metabolic risk factors at ages 17 and 32. Atherosclerosis 2014;235:430–7. doi:10.1016/j.atherosclerosis.2014.05.937
      OpenUrlCrossRefPubMed
    22. ↵
      1. Levy D,
      2. Mohlman MK,
      3. Zhang Y
      . Estimating the Potential Impact of Tobacco Control Policies on Adverse Maternal and Child Health Outcomes in the United States Using the SimSmoke Tobacco Control Policy Simulation Model. Nicotine Tob Res 2015. Published Online First: 18 September 2015. doi:doi:10.1093/ntr/ntv178
    23. ↵
      1. Hawkins SS,
      2. Baum CF,
      3. Oken E, et al
      . Associations of tobacco control policies with birth outcomes. JAMA Pediatr 2014;168:e142365. doi:10.1001/jamapediatrics.2014.2365
      OpenUrl
    24. ↵
      1. Zhang X,
      2. Kramer MS
      . Variations in mortality and morbidity by gestational age among infants born at term. J Pediatr 2009;154:358–62, 362.e1. doi:10.1016/j.jpeds.2008.09.013
      OpenUrlCrossRefPubMedWeb of Science
    25. ↵
      1. Donovan EF,
      2. Besl J,
      3. Paulson J, et al
      . Infant death among Ohio resident infants born at 32 to 41 weeks of gestation. Am J Obstet Gynecol 2010;203:58.e1–5. doi:10.1016/j.ajog.2010.01.071
      OpenUrlCrossRefPubMed
    26. ↵
      1. Schonhaut L,
      2. Armijo I,
      3. Pérez M
      . Gestational age and developmental risk in moderately and late preterm and early term infants. Pediatrics 2015;135:e835–41. doi:10.1542/peds.2014-1957
      OpenUrlAbstract/FREE Full Text
    27. ↵
      1. Brown HK,
      2. Speechley KN,
      3. Macnab J, et al
      . Neonatal morbidity associated with late preterm and early term birth: the roles of gestational age and biological determinants of preterm birth. Int J Epidemiol 2014;43:802–14. doi:10.1093/ije/dyt251
      OpenUrlAbstract/FREE Full Text
    28. ↵
      1. Been JV,
      2. Mackay DF,
      3. Millett C, et al
      . Impact of smoke-free legislation on perinatal and infant mortality: a national quasi-experimental study. Sci Rep 2015;5:13020. doi:10.1038/srep13020
      OpenUrlCrossRefPubMed
    29. ↵
      WHO. WHO: recommended definitions, terminology and format for statistical tables related to the perinatal period and use of a new certificate for cause of perinatal deaths. Modifications recommended by FIGO as amended October 14, 1976. Acta Obstet Gynecol Scand 1977;56:247–53.
      OpenUrlPubMedWeb of Science
    30. ↵
      1. Darrow LA,
      2. Strickland MJ,
      3. Klein M, et al
      . Seasonality of birth and implications for temporal studies of preterm birth. Epidemiology 2009;20:699–706. doi:10.1097/EDE.0b013e3181a66e96
      OpenUrlCrossRefPubMedWeb of Science
    31. ↵
      1. Vicedo-Cabrera AM,
      2. Iñíguez C,
      3. Barona C, et al
      . Exposure to elevated temperatures and risk of preterm birth in Valencia, Spain. Environ Res 2014;134C:210–17. doi:10.1016/j.envres.2014.07.021
      OpenUrl
    32. ↵
      1. Strand LB,
      2. Barnett AG,
      3. Tong S
      . Methodological challenges when estimating the effects of season and seasonal exposures on birth outcomes. BMC Med Res Methodol 2011;11:49. doi:10.1186/1471-2288-11-49
      OpenUrlCrossRefPubMed
    33. ↵
      1. Nguyen KH,
      2. Wright RJ,
      3. Sorensen G, et al
      . Association between local indoor smoking ordinances in Massachusetts and cigarette smoking during pregnancy: a multilevel analysis. Tob Control 2013;22:184–9. doi:10.1136/tobaccocontrol-2011-050157
      OpenUrlAbstract/FREE Full Text
    34. ↵
      1. Cox B,
      2. Martens E,
      3. Nemery B, et al
      . Impact of a stepwise introduction of smoke-free legislation on the rate of preterm births: analysis of routinely collected birth data. BMJ 2013;346:f441.
      OpenUrlAbstract/FREE Full Text
    35. ↵
      1. Kawashima A,
      2. Koide K,
      3. Ventura W, et al
      . Effects of maternal smoking on the placental expression of genes related to angiogenesis and apoptosis during the first trimester. PLoS ONE 2014;9:e106140. doi:10.1371/journal.pone.0106140
      OpenUrlCrossRefPubMed
    36. ↵
      1. Jaddoe VWV,
      2. Troe E-JWM,
      3. Hofman A, et al
      . Active and passive maternal smoking during pregnancy and the risks of low birthweight and preterm birth: the Generation R Study. Paediatr Perinat Epidemiol 2008;22:162–71. doi:10.1111/j.1365-3016.2007.00916.x
      OpenUrlCrossRefPubMedWeb of Science
    37. ↵
      1. Spong CY,
      2. Mercer BM,
      3. D'Alton M, et al
      . Timing of indicated late-preterm and early- term birth. Obstet Gynecol 2011;118:323–33. doi:10.1097/AOG.0b013e3182255999
      OpenUrlCrossRefPubMedWeb of Science
    38. ↵
      1. Raju TNK
      . Developmental physiology of late and moderate prematurity. Semin Fetal Neonatal Med 2012;17:126–31. doi:10.1016/j.siny.2012.01.010
      OpenUrlCrossRefPubMedWeb of Science
    39. ↵
      1. Page RL,
      2. Slejko JF,
      3. Libby AM
      . A citywide smoking ban reduced maternal smoking and risk for preterm births: a Colorado natural experiment. J Womens Health (Larchmt) 2012;21:621–7. doi:10.1089/jwh.2011.3305
      OpenUrlCrossRefPubMedWeb of Science
    40. ↵
      1. Franchini M,
      2. Caruso C,
      3. Perico A, et al
      . Assessment of foetal exposure to cigarette smoke after recent implementations of smoke-free policy in Italy. Acta Paediatr 2008;97:546–50. doi:10.1111/j.1651-2227.2008.00762.x
      OpenUrlCrossRefPubMedWeb of Science
    41. ↵
      1. Moore GF,
      2. Currie D,
      3. Gilmore G, et al
      . Socioeconomic inequalities in childhood exposure to secondhand smoke before and after smoke-free legislation in three UK countries. J Public Health (Oxf) 2012;34:599–608. doi:10.1093/pubmed/fds025
      OpenUrlAbstract/FREE Full Text
    42. ↵
      1. Brown T,
      2. Platt S,
      3. Amos A
      . Equity impact of European individual-level smoking cessation interventions to reduce smoking in adults: a systematic review. Eur J Public Health 2014;24:551–6. doi:10.1093/eurpub/cku065
      OpenUrlAbstract/FREE Full Text
    43. ↵
      1. Akhtar PC,
      2. Haw SJ,
      3. Currie DB, et al
      . Smoking restrictions in the home and secondhand smoke exposure among primary schoolchildren before and after introduction of the Scottish smoke-free legislation. Tob Control 2009;18:409–15. doi:10.1136/tc.2009.030627
      OpenUrlAbstract/FREE Full Text
    44. ↵
      1. Kairouz S,
      2. Lasnier B,
      3. Mihaylova T, et al
      . Smoking restrictions in homes after implementation of a smoking ban in public places. Nicotine Tob Res 2015;17:41–7. doi:10.1093/ntr/ntu125
      OpenUrlAbstract/FREE Full Text
    45. ↵
      1. Ho SY,
      2. Wang MP,
      3. Lo WS, et al
      . Comprehensive smoke-free legislation and displacement of smoking into the homes of young children in Hong Kong. Tob Control 2010;19:129–33. doi:10.1136/tc.2009.032003
      OpenUrlAbstract/FREE Full Text
    46. ↵
      1. Ramsay CR,
      2. Matowe L,
      3. Grilli R, et al
      . Interrupted time series designs in health technology assessment: lessons from two systematic reviews of behavior change strategies. Int J Technol Assess Health Care 2003;19:613–23. doi:10.1017/S0266462303000576
      OpenUrlCrossRefPubMedWeb of Science

    Footnotes

    • Contributors AMV-C contributed to the study design, data collection, performance of the statistical analysis and interpretation of the results, drafting and revision of the article. CS contributed to the study design, advisory in the statistical analysis, interpretation of the results and revision of the manuscript. DR contributed to the study design, and revision of the manuscript. LG contributed to data collection, interpretation of the results and revision of the manuscript. FW, JD contributed to the study design, and revision of the manuscript. MR advisory in the statistical analysis, and interpretation of the results and revision of the manuscript. LP designed the overall study idea, conceptualisation of the study design, advisory in the statistical analysis, interpretation of the results, writing and revision of the manuscript.

    • Funding This study was funded by the Tobacco Prevention Fund of the Federal Office of Public Health in Switzerland (decision number 13.008336).

    • Competing interests None declared.

    • Provenance and peer review Not commissioned; externally peer reviewed.