- Email: [email protected]
National, regional, and global prevalence of smoking during pregnancy in the general population: a systematic review and meta-analysis
Articles
National, regional, and global prevalence of smoking during pregnancy in the general population: a systematic review and meta-analysis Shannon Lange, Charlotte Probst, Jürgen Rehm, Svetlana Popova
Summary
Background Smoking during pregnancy has been linked to numerous adverse health consequences for both the developing fetus and mother. We estimated the prevalence of smoking during pregnancy by country, WHO region, and globally and the proportion of pregnant women who smoked during pregnancy, by frequency and quantity, on a global level. Methods For this systematic review and meta-analysis, we did a comprehensive systematic literature search for studies reporting the prevalence of smoking during pregnancy in the general population, published between Jan 1, 1985 and Feb 1, 2016, using several electronic bibliographic databases (CINAHL, Embase, ERIC, Medline, Medline in process, PsychINFO, Scopus, and Web of Science), without language or geographical restrictions. We included original research studies published in a peer-reviewed journal and assessed study quality using a tool specifically developed for use in systematic reviews addressing questions of prevalence. Studies were excluded if they did not include lifetime non-smokers in their sample or estimate, used a sample not generalisable to the general population of the respective country, or did not provide primary data. To estimate the prevalence by country, we did country-specific random-effects meta-analyses for countries with two or more available empirical studies, and we predicted the prevalence using a multilevel fractional response regression model with country-specific indicators for countries with one or no study. We estimated the proportion of female daily smokers who do not quit once pregnant by calculating the regional and global averages of the prevalence of daily smoking during pregnancy and of the prevalence of daily smoking in women. To estimate the global prevalence, by frequency and quantity, we did random-effects meta-analyses using available data from all countries and applied the respective proportions to the global prevalence estimate. We did a time–trend analysis using a univariate multilevel fractional response model. The review protocol is available on PROSPERO, registration number CRD42017075837. Findings Of 21 329 studies identified, 295 were retained for data extraction. We calculated estimates via meta-analysis for 43 countries and via statistical modelling for 131 countries. The three countries with the highest estimated prevalence of smoking during pregnancy were Ireland (38·4%, 95% CI 25·4–52·4), Uruguay (29·7%, 16·6–44·8), and Bulgaria (29·4%, 26·6–32·2). The global prevalence of smoking during pregnancy was estimated to be 1·7% (95% CI 0·0–4·5). The prevalence of smoking during pregnancy was 8·1% (95% CI 4·0–12·2) in the European Region, 5·9% (3·2–8·6) in the Region of the Americas, 1·2% (0·7–1·7) in the Southeast Asian Region, 1·2% (0·0–3·7) in the Western Pacific Region, 0·9% (0·0–1·9) in the Eastern Mediterranean Region, and 0·8% (0·0–2·2) in the African Region. Globally, 72·5% (95% CI 70·4–75·0) of pregnant women who smoked were daily smokers, and 27·5% (25·4–29·6) of them were occasional smokers; 51·8% (95% CI 50·0–53·5) women who smoked were light smokers, 34·8% (33·1–36·4) were moderate smokers, and 13·5% (12·3–14·7) were heavy smokers. Furthermore, the proportion of women who smoked daily and continued to smoke daily during pregnancy was 52·9% (95% CI 45·6–60·3), ranging from 30·6% (95% CI 25·6–36·4) in the European Region to 79·6% (44·2–100·0) in the Western Pacific Region.
Lancet Glob Health 2018 Published Online May 30, 2018 http://dx.doi.org/10.1016/ S2214-109X(18)30223-7 See Online/Comment http://dx.doi.org/10.1016/ S2214-109X(18)30246-8 Institute for Mental Health Policy Research, Centre for Addiction and Mental Health, Toronto, ON, Canada (S Lange MPH, C Probst PhD, Prof J Rehm PhD, S Popova PhD); Institute of Medical Science, Faculty of Medicine, (S Lange, Prof J Rehm, S Popova), Dalla Lana School of Public Health, (Prof J Rehm, S Popova), Department of Psychiatry (Prof J Rehm), Factor-Inwentash Faculty of Social Work (S Popova), University of Toronto, Toronto, ON, Canada; and Institute of Clinical Psychology and Psychotherapy, Center of Clinical Epidemiology and Longitudinal Studies, Technische Universität Dresden, Dresden, Germany (C Probst, Prof J Rehm) Correspondence to: Dr Svetlana Popova, Institute for Mental Health Policy Research, Centre for Addiction and Mental Health, Toronto, ON M5S 2S1, Canada [email protected]
Interpretation Smoking during pregnancy is still a prevalent behaviour in many countries. These findings should inform smoking prevention programmes and health promotion strategies, as well as draw attention to the need for improved access to smoking cessation programmes for pregnant women. Funding Centre for Addiction and Mental Health. Copyright © The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY-NC-ND 4.0 license.
Introduction Smoking tobacco during pregnancy is associated with a number of pregnancy complications such as ec topic pregnancy, placental abruption, placenta praevia,
pre-eclampsia, and a range of poor fetal outcomes such as fetal mortality, stillbirth, and tobacco-induced abortions.1–6 Preterm birth, low birthweight, and sudden infant death syndrome are also linked to smoking during
www.thelancet.com/lancetgh Published online May 30, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30223-7
1
Articles
Research in context Evidence before this study Smoking during pregnancy is associated with a wide range of pregnancy complications and several adverse outcomes for the exposed child. Although the detrimental health effects of smoking during pregnancy are well established, most countries do not have up-to-date population-level prevalence data on smoking during pregnancy. To fill these knowledge gaps, we did a comprehensive epidemiological study to estimate the actual (based on existing empirical studies; for 43 countries) and predicted (for countries with one or no study; 131 countries) prevalence of smoking during pregnancy in the general population by country, WHO region, and globally. We also estimated the proportion of pregnant women globally who smoked during pregnancy, by frequency and quantity, and the proportion of women who smoked daily and continued to smoke daily during pregnancy, by WHO region and globally. Added value of this study Globally, 52·9% (95% CI 45·6–60·3) of women who smoked daily continued to smoke daily during pregnancy. The European Region had the highest estimated prevalence of
pregnancy, with established causality.1,2,7,8 Evidence also suggests an association between smoking during pregnancy and major birth malformations such as oral clefts, craniosynostosis, gastroschisis,1,9,10 and intellectual impairment later in life.1,11,12 Economic estimates have shown that the expenditures of medical care that are attributable to birth complications are much higher for women who smoke during pregnancy than for non-smokers.13 This association suggests that the birth complications of women who smoke during pregnancy are more severe and need more intense interventions than those of women who do not smoke during pregnancy. However, the magnitude of such smoking-attributable health-care expenditures has decreased over time.14,15 Alarmingly, surveillance data from the Global Youth Tobacco Surveys16 suggest that the difference in the prevalence of smoking between the two sexes is narrowing, especially in low-income and middle-income countries. In view of the effects of smoking on reproduction, these data underscore the importance of sex-specific prevalence estimates, especially for women of childbearing age. On the basis of findings from the Global Burden of Disease Study, WHO recently reported that the global agestandardised prevalence of daily smoking in women in 2015 was 5·4%.17 However, although the detrimental health effects of smoking during pregnancy are well established, to the best of our knowledge no estimates exist of the prevalence of smoking during pregnancy for all countries. In one notable study, Caleyachetty and colleagues18 estimated the prevalence of smoking during pregnancy for 54 low-income and middle-income 2
smoking during pregnancy (8·1%, 95% CI 4·0–12·2) but the lowest estimated proportion of women who smoked daily and continued to smoke daily during pregnancy (30·6%, 25·6–36·4). The African Region had the lowest estimated prevalence of smoking during pregnancy (0·8%, 0·0–2·2) but the second highest estimated proportion of women who smoked daily and continued to smoke daily during pregnancy (61·9%, 30·6–100·0). The Western Pacific Region had the highest proportion of women who smoked daily and continued to smoke daily during pregnancy (79·6%, 44·2–100·0). Implications of all the available evidence To the best of our knowledge, this is the first study to estimate the prevalence of smoking during pregnancy for all countries. Although the global prevalence of smoking during pregnancy appears to be relatively low (1·7%, 95% CI 0·0–4·5), smoking during pregnancy is still prevalent in many countries. Knowledge of the prevalence of smoking during pregnancy can inform strategies for developing targeted anti-smoking campaigns and intervention programmes.
countries using data from Demographic and Health Surveys, which were done between 2001 and 2012. Nevertheless, most countries do not have up-to-date population-level prevalence data on smoking during pregnancy, and the global prevalence is there fore unknown. Yet knowledge of the prevalence of ing during pregnancy can inform strategies for smok develop ing targeted advertising anti-smoking cam paigns and intervention programmes. Therefore, to fill these knowledge gaps, we did the first comprehensive epidemiological study to (1) estimate the prevalence of smoking (any frequency or quantity) during pregnancy in the general population, by country, WHO region (ie, African Region, Eastern Mediterranean Region, European Region, Region of the Americas, Southeast Asian Region, and Western Pacific Region), and globally; (2) estimate the proportion of pregnant women who smoke during pregnancy, by frequency and quantity, on a global level; (3) estimate the proportion of women who smoke daily and continue to smoke daily during pregnancy, by WHO region and globally; and (4) determine whether the prevalence of smoking during pregnancy in the general population has changed in the past 30 years. This study is intended to provide further insight into the global burden of smoking during pregnancy and provide clinicians and public health authorities with the best available data at this time. Smokeless tobacco was excluded from this study because data are scarce, and important potential predictors, such as the prevalence of smokeless tobacco use in women in all countries, are not available at this time.
www.thelancet.com/lancetgh Published online May 30, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30223-7
Articles
Methods We have adhered to the Guidelines for Accurate and Transparent Health Estimates Reporting guidelines19 and to the standards set out in Preferred Reporting Items for Systematic Reviews and Meta-Analyses.20
21 305 records identified by database search
24 records identified through other sources
8073 excluded (duplicates)
Search strategy and selection criteria A comprehensive systematic literature search was done to identify all studies that have reported the prevalence of smoking during pregnancy in the general population in any country. We searched CINAHL, Embase, ERIC, Medline, Medline in process, PsychINFO, Scopus, and Web of Science for studies published between Jan 1, 1985, and Feb 1, 2016 (ie, a 30-year period), using terms that are listed in the appendix (p 4) and without language or geographical restrictions. The content pages of the major epidemiological journals and citations in the relevant articles were also manually screened. The full review protocol is available in PROSPERO, registration number CRD42017075837. Articles were retained if they: consisted of original, quantitative research published in a peer-reviewed journal; reported the prevalence of smoking during pregnancy in the general population; and provided a measure of uncertainty (CI or SE) for the prevalence or at least two of the following: sample size, number of cases, or prevalence. Articles were excluded if they: did not include lifetime non-smokers in their sample or estimate; used a sample not generalisable to the general population of the respective country (ie, studies had to use a sample that was generalisable to the general population at either the national or subnational level; as such, we included nationally or regionally representative surveys or hospital-based studies with full coverage of births within a specific timeframe); or did not provide primary data (ie, estimates had to be based on individuallevel data). For the purposes of this study, the term smoking refers to smoking tobacco in the form of cigarettes. A lifetime non-smoker was defined as a woman who had never smoked a cigarette. Conversely, a smoker was defined as a woman who had smoked at least one cigarette during her pregnancy, either daily (at least one cigarette every day) or occasionally (at least one cigarette per occasion [less than daily]). We appraised the quality of each included study using a modified version of a critical appraisal tool for use in systematic reviews addressing questions of prevalence.21
Data analysis For countries with two or more available empirical studies reporting the prevalence of smoking during pregnancy, we did country-specific random-effects metaanalyses.22 Before the meta-analyses, prevalence estim ates were transformed using a double arcsine transformation23 so that the data followed a normal distribution (an assumption needed when statistically
13 256 identified for screening
12 235 excluded after full-text screening
1021 full-text articles assessed for eligibility
726 excluded 546 lacked relevant data 180 did not meet inclusion criteria
See Online for appendix
For the PROSPERO database see http://www.crd.york.ac.uk/ PROSPERO/
295 eligible studies
Figure 1: Study selection
combining estimates). The resulting combined point estimates and respective CI were then back-transformed to facilitate interpretation. For countries with either one or no available study, the prevalence (any frequency or quantity) was predicted using a multilevel fractional response regression model.24 A fractional response model does not rely on any distributional assumptions; however, the outcome variable must range between zero and one.24 A multilevel approach was chosen to account for the likelihood that estimates within countries would be more similar to one another than would estimates between countries. Additional methodological details on the estimation of the prevalence of smoking (any frequency or quantity) during pregnancy in the general population, by country, are presented in the appendix (pp 2, 3). To estimate the prevalence of smoking during pregnancy by WHO region and globally, we calculated a weighted average of the country-specific point estimates, weighted by the number of livebirths in each country for the latest available year (2001–15).25 To estimate the proportion of pregnant women who smoked during pregnancy by frequency and quantity on a global level, we first did random-effects meta-analyses of the frequency and quantity of smoking, using available data from all countries. Frequency was categorised as either occasional smoking (at least one cigarette per occasion, but less than daily) or as daily smoking (at least one cigarette every day). Quantity was categorised as either light smoking (up to ten cigarettes per smoking day), moderate smoking (11–19 cigarettes per smoking day), or heavy smoking (≥20 cigarettes per smoking day). We then
www.thelancet.com/lancetgh Published online May 30, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30223-7
3
Articles
Prevalence (%) 0–1·5 1·6–2·5 2·6–5·0 5·1–7·5 7·6–10·0 10·1–15·0 15·1–20·0 ≥20·1 No data
Figure 2: Global prevalence of smoking during pregnancy in 2015
applied the estimated proportions of occasional and daily smoking (frequency) and light, moderate, and heavy smoking (quantity) to the estimated global prevalence of smoking during pregnancy. To derive the CI for each respective point estimate, we did Monte Carlo simulations,26 generating 100 000 samples per category, and using the 2·5th and 97·5th percentiles of the resulting distribution as the CI. To estimate the proportion of daily female smokers in the general population who continued to smoke daily during pregnancy by WHO region and globally, we first calculated a weighted average of the country-specific point estimates (data obtained directly from the Institute of Health Metrics and Evaluation), weighted by the adult female population in each country. We then applied the estimated proportion of pregnant women who smoked daily to the regional and global estimates of the prevalence of smoking during pregnancy. We did Monte Carlo simulations to estimate the CI for each estimated proportion. We used a univariate mixed-effects fractional response model24 to determine whether the prevalence of smoking during pregnancy has changed in the past 30 years. Given that most studies came from the USA, we also did a sensitivity analysis by removing all studies from the USA to determine whether they were driving the observed trend. 4
Role of the funding source The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.
Results Of 21 329 studies identified in the literature search, 295 studies with relevant data were retained for data extraction (figure 1; appendix p 5). These 295 empirical studies contained 373 estimates from 104 countries on the prevalence of smoking during pregnancy, as several studies reported estimates of the prevalence of smoking during pregnancy for multiple geographies (ie, for more than one country or for more than one site within a country).18,27–33 Data were available for 32 of 47 countries in the African Region, 16 of 35 countries in the Region of the Americas, six of 21 countries in the Eastern Mediterranean Region, 35 of 53 countries in the European Region, six of 11 countries in the Southeast Asian Region, and nine of 27 countries in the Western Pacific Region. 46 estimates were available for light smoking, 34 estimates for moderate smoking, and 29 estimates for heavy smoking during pregnancy. The prevalence of smoking occasionally was reported in 15 studies, whereas the prevalence of smoking daily during pregnancy was
www.thelancet.com/lancetgh Published online May 30, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30223-7
Articles
Frequency
Quantity
Moderate (34·8%) Occasional (27·5%)
Daily (72·5%)
Heavy (13·5%)
Light (51·8%)
Figure 3: Proportion of pregnant women who smoked during pregnancy, by frequency and quantity, globally Frequency was categorised as occasional smoking (at least one cigarette per occasion, but less than daily) or daily smoking (at least one cigarette every day). Quantity was categorised as light smoking (1–10 cigarettes per smoking day), moderate smoking (11–19 cigarettes per smoking day), and heavy smoking (≥20 cigarettes per smoking day). 25·0
During pregnancy All women
Prevalence (%)
20·0 15·0 10·0 5·0
y all ob Gl
Re gi on Re gi o no Ea ste Am f th rn M e e ed rica ite s rra ne a Re n gi Eu on ro pe a So ut Reg n he io as n tA si R an W es egio te n rn Pa cifi Re c gi on
0
Af ric an
reported in 20 studies. The critical appraisal of included studies and a complete reference list is available in the appendix (p 11, 24). The prevalence of smoking during pregnancy was estimated for 147 countries (via meta-analysis for 43 countries [with two or more available empirical studies] and via statistical modelling [prediction] for 131 countries). The prevalence of smoking during pregnancy could not be estimated for 20 countries because of missing data for one or more predictor variables. The final model was based on 372 of the 373 available estimates (ie, those with complete covariate data) from 104 countries and included the following predictor variables: age-standardised prevalence of daily smoking in women of childbearing age (β=0·80 [referring to an increase of 10%], 95% CI 0·62–0·97), gross domestic product at purchasing price parity (β=–0·08 [referring to an increase of 10 000 international dollars], 95% CI –0·21 to 0·05), proportion of the population living in urban areas (β=0·35 [referring to a 10% increase], 95% CI 0·24–0·46), and study year (β=–0·14 [referring to a 10 year increase], 95% CI –0·30 to 0·01). The five countries with the highest estimated prevalence of smoking (any frequency or quantity) during pregnancy were Ireland (38·4%, 95% CI 25·4–52·4; I²=96·5%, based on meta-analysis of five studies), Uruguay (29·7%, 16·6–44·8; based on meta-analysis of three studies), Bulgaria (29·4%, 26·6–32·2; based on statistical modelling), Spain (26·0%, 22·4–29·7; I²=96·3%, based on meta-analysis of ten studies), and Denmark (25·2%, 18·6–32·4; I²=99·8%, based on meta-analysis of six studies). The five countries with the lowest prevalence of smoking during pregnancy were Tanzania (0·2%, 0·0–0·6; based on meta-analysis of two studies), Burundi (0·3%, 0·0–0·6; based on statistical modelling), St Lucia (0·3%, 0·0–0·6; based on statistical modelling), Sri Lanka (0·3%, 0·0–0·6; based on statistical modelling), and Malawi (0·3%, 0·0–0·7; based on statistical modelling). The prevalence of smoking during pregnancy, by country, is shown in figure 2 and in the appendix (p 18). The results of the tests of heterogeneity and publication bias for the meta-analyses of smoking during pregnancy in the general population, by country, are shown in the appendix (p 21). The global prevalence of smoking during pregnancy was estimated to be 1·7% (95% CI 0·0–4·5). The highest prevalence of smoking during pregnancy was in the European Region (8·1%, 4·0–12·2), and the lowest prevalence of smoking during pregnancy was in the African Region (0·8%, 0·0–2·2). Globally, 72·5% (95% CI 70·4–75·0) of women who smoked during pregnancy were daily smokers, and 27·5% (25·4–29·6) were occasional smokers (figure 3). In terms of quantity, more than half of all women who smoked during pregnancy (51·8%, 95% CI 50·0–53·5) were light smokers, 34·8% (33·1–36·4) were moderate smokers, and 13·5% (12·3–14·7) were heavy smokers.
Figure 4: Prevalence of daily smoking in women and pregnant women in the general population and estimated proportion of women who smoke daily who continue to smoke daily during pregnancy, by WHO region
Results of the tests of heterogeneity and publication bias for the meta-analyses of frequency and quantity of cigarettes smoked during pregnancy in the general population, by country, are shown in the appendix (p 22). The proportion of women who smoked daily and continued to smoke daily during pregnancy ranged from 30·6% (95% CI 25·6–36·4) in the European Region to 79·6% (44·2–100·0) in the Western Pacific Region (figure 4). Globally, this proportion was estimated to be 52·9% (95% CI 45·6–60·3). The estimated regional averages of the prevalence of daily smoking in women and daily smoking in pregnant women in the general population is shown in the appendix (p 23). The univariate multilevel fractional response model showed that the prevalence of smoking during pregnancy has significantly decreased in the past 30 years (β=–0·47 [referring to a 10 year increase], 95% CI –0·56 to –0·39). After removing all studies from the USA, the decrease in the prevalence of smoking during pregnancy in the past
www.thelancet.com/lancetgh Published online May 30, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30223-7
5
Articles
30 years did not change meaningfully (β=–0·38 [referring to a 10 year increase], –0·46 to –0·29).
Discussion This is the first study to estimate the prevalence of smoking during pregnancy nationally, regionally, and globally and to estimate the proportion of pregnant women globally who smoke during pregnancy by frequency and quantity. This is also the first study to estimate the proportion of women who smoke daily and continue to smoke daily during pregnancy, regionally and globally. Although social norms discourage women from smoking during pregnanc, our findings show that smoking during pregnancy is still a prevalent behaviour in many countries. Specifically, the estimated prevalence of smoking during pregnancy was more than 10% in 29 (17%) of 174 countries and more than 20% in 12 (7%) countries. Our findings also indicate that more than 70% of pregnant women who smoked during pregnancy did so daily rather than occasionally. This finding is in line with previous findings that women who quit smoking before pregnancy had smoked fewer cigarettes before pregnancy than women who did not quit,34 suggesting that daily smokers are less likely to quit smoking upon pregnancy recognition. We also found that, globally, more than half of women who smoked daily continued to do so while they were pregnant. The clinical implications of these findings are note worthy, particularly the fact that although the prevalence appears to be decreasing overall, smoking during pregnancy is still prevalent in several countries. Obstetricians, gynaecologists, midwives, and other pre natal health-care workers in these countries should be aware of the magnitude of the problem and promote smoking cessation interventions that target pregnancy, whether pharmacological, psychological, or both. The preconception period is the optimal time to intervene.35 Thus, preconception prevention strategies need to be implemented in countries where the prevalence of smoking during pregnancy remains high. The imple mentation of such strategies might be particularly challenging in low-income and middle-income countries because of obstacles such as economic factors, low perception of risks in the public, lack of policies that promote cessation, poor health-care systems, lack of infrastructure, and industry action.36 In low-income and middle-income countries where the prevalence of smoking during pregnancy is high, the focus should be on building capacity by, for example, integrating smoking cessation services with other health-care services, train ing health-care professionals, and developing guide lines.36 WHO-recommended tobacco control policies and legislations have increased globally, with the largest increases in low-income and middle-income countries. However, the negative relation between higher imple mentation and management of tobacco control and 6
smoking prevalence is stronger in men than in women.37 In view of the adverse effects of smoking on the developing fetus, this relation underscores the importance of sex-sensitive prevention initiatives and smoking cessation programmes that specifically target women of childbearing age. As such, the prevalence estimates presented here serve to inform stakeholders and policy makers of the magnitude of the public health problem in their respective countries. The predicted prevalence estimates of smoking during pregnancy are particularly useful for countries where existing tobacco use surveillance systems or data are absent. We have therefore provided countries with a working estimate of the prevalence of smoking during pregnancy, which can be refined when data become available. Women with intended conceptions are more likely than women with unintended pregnancies to quit smoking upon pregnancy recognition.38 However, a large percentage of pregnancies globally are unplanned.39 As such, WHO recommends that a five-step process be used in all health-care settings to the extent possible.40 These five steps are: (1) asking about smoking history and current smoking behaviour; (2) advising smoking women of the value of quitting and the risks of continuing; (3) assessing a smoking woman’s motivation and thoughts on quitting; (4) assisting the smoker in all attempts to quit with the use of materials or referrals to intervention programmes; and (5) arranging follow-up throughout pregnancy and the post-partum period.41 Importantly, the primary focus should be on smoking cessation. However, if a woman continues to smoke while pregnant, she should be supported in reducing the quantity of smoking. A few limitations of our study should be acknowledged. First, the pooled prevalence of smoking during pregnancy is based on data from the past 30 years. As shown in the time–trend analysis, the prevalence of smoking during pregnancy decreased with time, which could have resulted in a slight inflation of both the estimates obtained via meta-analysis and the regional and global estimates because they are weighted averages of both actual and predicted data. However, the time– trends of smoking during pregnancy might differ between countries, so the predicted estimates for some countries in which the prevalence of smoking in women increased in the past three decades might be slightly underestimated.42 Second, because of the data availability, we used age-standardised prevalence of daily smoking in women of childbearing age in the statistical model, as opposed to the prevalence of smoking at any frequency. However, the prevalence of daily smoking could be a stronger indicator of smoking during pregnancy. Third, it was not possible to obtain country-specific estimates of the prevalence of smoking during pregnancy by frequency and quantity because of data limitations. Instead, these indicators were pooled across countries
www.thelancet.com/lancetgh Published online May 30, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30223-7
Articles
and applied to the global prevalence estimate. We acknowledge that the estimates of the prevalence of smoking during pregnancy by frequency and quantity are likely to be biased towards the countries from which the data originated; however, these are the best available global estimates at this time. Fourth, the predicted prevalence estimates for the 131 countries with either one or no available study might diverge from the true prevalence because the data from which the values were predicted carry some measurement error, and other relevant explanatory variables might affect the prevalence of smoking during pregnancy and not be possible to account for. Other valid approaches to predicting country-specific prevalence of smoking during pregnancy exist (eg, a dynamic model) but rely on additional information (eg, data on smoking initiation and the prevalence of smokers and never-smokers). However, taking into consideration the study and that we were limited to the information reported in the included studies, we consider the present model to yield the best estimates possible. Finally, as would be expected when pooling estimates across time and locations, we observed high heterogeneity in the meta-analyses. However, heterogeneity can be overestimated when summarising studies with large sample sizes.43 The findings highlight a number of key areas for improvement in all parts of the world with respect to tobacco use during pregnancy, which includes both smoked and smokeless tobacco. First, women of child bearing age need to be educated about the potential detrimental effects of tobacco use on the developing fetus and infant. It is important that the messages about tobacco use during pregnancy be provided to women clearly and consistently.40 Prenatal care provides an opportune time for health-care professionals to ask about tobacco use and provide appropriate referrals to smoking cessation programmes whenever necessary. Second, evidence-based smoking cessation interventions should be integrated into routine prenatal care. Third, universal screening protocols need to be established for all pregnant women and women of childbearing age. Finally, prevention efforts and intervention programmes need to be expanded to reach the broader population of women of childbearing age.44 Contributors SL led the conception and design of the study, the development of the data collection instrument, data collection and quality assessment, did the statistical analysis, interpreted the data, and wrote and revised the manuscript. CP contributed to study design, did the statistical analysis, assisted in data interpretation, and wrote and revised the manuscript. JR and SP contributed to data interpretation and revised the intellectual content of the manuscript. Declaration of interests We declare no competing interests. References 1 Centers for Disease Control. Publications and reports of the Surgeon General. How tobacco smoke causes disease: the biology and behavioral basis for smoking-attributable disease: a report of the Surgeon General. Atlanta, GA: Centers for Disease Control and Prevention, 2010.
2 3 4 5 6 7
8 9 10 11 12
13 14 15 16 17
18
19 20
21 22 23 24 25 26
DiFranza JR, Lew RA. Effect of maternal cigarette smoking on pregnancy complications and sudden infant death syndrome. J Fam Pract 1995; 40: 385–94. Pineles BL, Park E, Samet JM. Systematic review and meta-analysis of miscarriage and maternal exposure to tobacco smoke during pregnancy. Am J Epidemiol 2014; 179: 807–23. Rogers JM. Tobacco and pregnancy. Reprod Toxicol 2009; 28: 152–60. Salihu HM, Wilson RE. Epidemiology of prenatal smoking and perinatal outcomes. Early Hum Dev 2007; 83: 713–20. Tikkanen M, Nuutila M, Hiilesmaa V, Paavonen J, Ylikorkala O. Prepregnancy risk factors for placental abruption. Acta Obstet Gynecol Scand 2006; 85: 40–44. Ko TJ, Tsai LY, 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: 20–27. Taylor JA, Sanderson M. A reexamination of the risk factors for the sudden infant death syndrome. J Pediatr 1995; 126: 887–91. Little J, Cardy A, Munger RG. Tobacco smoking and oral clefts: a meta-analysis. Bull World Health Organ 2004; 82: 213–18. Werler MM, Sheehan JE, Mitchell AA. Association of vasoconstrictive exposures with risks of gastroschisis and small intestinal atresia. Epidemiology 2003; 14: 349–54. Drews CD, Murphy CC, Yeargin-Allsopp M, Decoufle P. The relationship between idiopathic mental retardation and maternal smoking during pregnancy. Pediatrics 1996; 97: 547–53. Mortensen EL, Michaelsen KF, Sanders SA, Reinisch JM. A dose-response relationship between maternal smoking during late pregnancy and adult intelligence in male offspring. Paediatr Perinat Epidemiol 2005; 19: 4–11. Centers for Disease Control. Medical-care expenditures attributable to cigarette smoking during pregnancy—United States, 1995. MMWR Morbid Mortal Wkly Rep 1997; 46: 1048–50. Adams EK, Miller VP, Ernst C, Nishimura BK, Melvin C, Merritt R. Neonatal health care costs related to smoking during pregnancy. Health Econ 2002; 11: 193–206. Adams EK, Melvin CL, Raskind-Hood C, Joski PJ, Galactionova E. Infant delivery costs related to maternal smoking: an update. Nicotine Tob Res 2011; 13: 627–37. Warren CW, Jones NR, Peruga A, et al. Global youth tobacco surveillance, 2000–2007. MMWR Surveill Summ 2008; 57: 1–28. GBD 2015 Tobacco Collaborators. Smoking prevalence and attributable disease burden in 195 countries and territories, 1990–2015: a systematic analysis from the Global Burden of Disease Study 2015. Lancet 2017; 389: 1885–906. Caleyachetty R, Tait CA, Kengne AP, Corvalan C, Uauy R, Echouffo-Tcheugui JB. Tobacco use in pregnant women: analysis of data from Demographic and Health Surveys from 54 low-income and middle-income countries. Lancet Glob Health 2014; 2: e513–20. Stevens GA, Alkema L, Black RE, et al. Guidelines for accurate and transparent health estimates reporting: the GATHER statement. Lancet 2016; 388: e19–23. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. PLoS Med 2009; 6: e1000100. Munn Z, Moola S, Riitano D, Lisy K. The development of a critical appraisal tool for use in systematic reviews addressing questions of prevalence. Int J Health Policy Manag 2014; 3: 123–28. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986; 7: 177–88. Freeman MF, Tukey JW. Transformations related to the angular and the square root. Ann Math Stat 1950; 21: 607–1. Papke LE, Wooldridge J. Econometric methods for fractional response variables with an application to 401 (k) plan participation rates. J Appl Econ 1996; 11: 619–32. United Nations. Population and vital statistics report. New York, NY: United Nations, 2016. Graham C, Talay D. Stochastic Simulation and Monte Carlo Methods: Mathematical Foundations of Stochastic Simulation. Berlin: Springer, 2013.
www.thelancet.com/lancetgh Published online May 30, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30223-7
7
Articles
27 Bloch M, Althabe F, Onyamboko M, et al. Tobacco use and secondhand smoke exposure during pregnancy: an investigative survey of women in 9 developing nations. Am J Public Health 2008; 98: 1833–40. 28 Grant TM, Huggins JE, Sampson PD, Ernst CC, Barr HM, Streissguth AP. Alcohol use before and during pregnancy in western Washington, 1989–2004: implications for the prevention of fetal alcohol spectrum disorders. Am J Obstet Gynecol 2009; 200: 278.e1–8. 29 Ostrea EM Jr, Villanueva-Uy E, Ngerncham S, et al. An epidemiologic study comparing fetal exposure to tobacco smoke in three Southeast Asian countries. Int J Occup Environ Health 2008; 14: 257–62. 30 Pichini S, Morini L, Marchei E, et al. Ethylglucuronide and ethylsulfate in meconium to assess gestational ethanol exposure: preliminary results in two Mediterranean cohorts. Can J Clin Pharmacol 2009; 16: e370–75. 31 Primatesta P, Del Corno G, Bonazzi MC, Waters WE. Alcohol and pregnancy: an international comparison. J Public Health Med 1993; 15: 69–76. 32 Rodriguez A, Olsen J, Kotimaa AJ, et al. Is prenatal alcohol exposure related to inattention and hyperactivity symptoms in children? Disentangling the effects of social adversity. J Child Psychol Psychiatry 2009; 50: 1073–83. 33 Smedberg J, Lupattelli A, Mardby AC, Nordeng H. Characteristics of women who continue smoking during pregnancy: a cross-sectional study of pregnant women and new mothers in 15 European countries. BMC Pregnancy Childbirth 2014; 14: 213. 34 Curtin SC, Matthews TJ. Smoking prevalence and cessation before and during pregnancy: data from the birth certificate, 2014. Natl Vital Stat Rep 2016; 65: 1–14.
8
35 Dean SV, Imam AM, Lassi ZS, Bhutta ZA. Importance of intervening in the preconception period to impact pregnancy outcomes. Nestle Nutr Inst Workshop Ser 2013; 74: 63–73. 36 Abdullah AS, Husten CG. Promotion of smoking cessation in developing countries: a framework for urgent public health interventions. Thorax 2004; 59: 623–30. 37 Anderson CL, Becher H, Winkler V. Tobacco control progress in low and middle income countries in comparison to high income countries. Int J Environ Res Public Health 2016; 13: 1039. 38 Kost K, Landry DJ, Darroch JE. Predicting maternal behaviors during pregnancy: does intention status matter? Fam Plann Perspect 1998; 30: 79–88. 39 Sedgh G, Singh S, Hussain R. Intended and unintended pregnancies worldwide in 2012 and recent trends. Stud Fam Plann 2014; 45: 301–14. 40 WHO. WHO recommendations for the prevention and management of tobacco use and second-hand smoke exposure in pregnancy. Geneva, Switzerland: World Health Organization, 2013. 41 Fiore MJC, Baker TB, Bailey WC, et al. A clinical practice guideline for treating tobacco use and dependence: 2008 update. A US Public Health Service report. Am J Prev Med 2008; 35: 158–76. 42 Ng M, Freeman MK, Fleming TD, et al. Smoking prevalence and cigarette consumption in 187 countries, 1980–2012. JAMA 2014; 311: 183–92. 43 Rücker G, Schwarzer G, Carpenter JR, Schumacher M. Undue reliance on I(2) in assessing heterogeneity may mislead. BMC Med Res Methodol 2008; 8: 79. 44 Orleans CT, Barker DC, Kaufman NJ, Marx JF. Helping pregnant smokers quit: meeting the challenge in the next decade. Tob Control 2000; 9 (suppl 3): III6–11.
www.thelancet.com/lancetgh Published online May 30, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30223-7
National, regional, and global prevalence of smoking during pregnancy in the general population: a systematic review and meta-analysis Shannon Lange, Charlotte Probst, Jürgen Rehm, Svetlana Popova
Summary
Background Smoking during pregnancy has been linked to numerous adverse health consequences for both the developing fetus and mother. We estimated the prevalence of smoking during pregnancy by country, WHO region, and globally and the proportion of pregnant women who smoked during pregnancy, by frequency and quantity, on a global level. Methods For this systematic review and meta-analysis, we did a comprehensive systematic literature search for studies reporting the prevalence of smoking during pregnancy in the general population, published between Jan 1, 1985 and Feb 1, 2016, using several electronic bibliographic databases (CINAHL, Embase, ERIC, Medline, Medline in process, PsychINFO, Scopus, and Web of Science), without language or geographical restrictions. We included original research studies published in a peer-reviewed journal and assessed study quality using a tool specifically developed for use in systematic reviews addressing questions of prevalence. Studies were excluded if they did not include lifetime non-smokers in their sample or estimate, used a sample not generalisable to the general population of the respective country, or did not provide primary data. To estimate the prevalence by country, we did country-specific random-effects meta-analyses for countries with two or more available empirical studies, and we predicted the prevalence using a multilevel fractional response regression model with country-specific indicators for countries with one or no study. We estimated the proportion of female daily smokers who do not quit once pregnant by calculating the regional and global averages of the prevalence of daily smoking during pregnancy and of the prevalence of daily smoking in women. To estimate the global prevalence, by frequency and quantity, we did random-effects meta-analyses using available data from all countries and applied the respective proportions to the global prevalence estimate. We did a time–trend analysis using a univariate multilevel fractional response model. The review protocol is available on PROSPERO, registration number CRD42017075837. Findings Of 21 329 studies identified, 295 were retained for data extraction. We calculated estimates via meta-analysis for 43 countries and via statistical modelling for 131 countries. The three countries with the highest estimated prevalence of smoking during pregnancy were Ireland (38·4%, 95% CI 25·4–52·4), Uruguay (29·7%, 16·6–44·8), and Bulgaria (29·4%, 26·6–32·2). The global prevalence of smoking during pregnancy was estimated to be 1·7% (95% CI 0·0–4·5). The prevalence of smoking during pregnancy was 8·1% (95% CI 4·0–12·2) in the European Region, 5·9% (3·2–8·6) in the Region of the Americas, 1·2% (0·7–1·7) in the Southeast Asian Region, 1·2% (0·0–3·7) in the Western Pacific Region, 0·9% (0·0–1·9) in the Eastern Mediterranean Region, and 0·8% (0·0–2·2) in the African Region. Globally, 72·5% (95% CI 70·4–75·0) of pregnant women who smoked were daily smokers, and 27·5% (25·4–29·6) of them were occasional smokers; 51·8% (95% CI 50·0–53·5) women who smoked were light smokers, 34·8% (33·1–36·4) were moderate smokers, and 13·5% (12·3–14·7) were heavy smokers. Furthermore, the proportion of women who smoked daily and continued to smoke daily during pregnancy was 52·9% (95% CI 45·6–60·3), ranging from 30·6% (95% CI 25·6–36·4) in the European Region to 79·6% (44·2–100·0) in the Western Pacific Region.
Lancet Glob Health 2018 Published Online May 30, 2018 http://dx.doi.org/10.1016/ S2214-109X(18)30223-7 See Online/Comment http://dx.doi.org/10.1016/ S2214-109X(18)30246-8 Institute for Mental Health Policy Research, Centre for Addiction and Mental Health, Toronto, ON, Canada (S Lange MPH, C Probst PhD, Prof J Rehm PhD, S Popova PhD); Institute of Medical Science, Faculty of Medicine, (S Lange, Prof J Rehm, S Popova), Dalla Lana School of Public Health, (Prof J Rehm, S Popova), Department of Psychiatry (Prof J Rehm), Factor-Inwentash Faculty of Social Work (S Popova), University of Toronto, Toronto, ON, Canada; and Institute of Clinical Psychology and Psychotherapy, Center of Clinical Epidemiology and Longitudinal Studies, Technische Universität Dresden, Dresden, Germany (C Probst, Prof J Rehm) Correspondence to: Dr Svetlana Popova, Institute for Mental Health Policy Research, Centre for Addiction and Mental Health, Toronto, ON M5S 2S1, Canada [email protected]
Interpretation Smoking during pregnancy is still a prevalent behaviour in many countries. These findings should inform smoking prevention programmes and health promotion strategies, as well as draw attention to the need for improved access to smoking cessation programmes for pregnant women. Funding Centre for Addiction and Mental Health. Copyright © The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY-NC-ND 4.0 license.
Introduction Smoking tobacco during pregnancy is associated with a number of pregnancy complications such as ec topic pregnancy, placental abruption, placenta praevia,
pre-eclampsia, and a range of poor fetal outcomes such as fetal mortality, stillbirth, and tobacco-induced abortions.1–6 Preterm birth, low birthweight, and sudden infant death syndrome are also linked to smoking during
www.thelancet.com/lancetgh Published online May 30, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30223-7
1
Articles
Research in context Evidence before this study Smoking during pregnancy is associated with a wide range of pregnancy complications and several adverse outcomes for the exposed child. Although the detrimental health effects of smoking during pregnancy are well established, most countries do not have up-to-date population-level prevalence data on smoking during pregnancy. To fill these knowledge gaps, we did a comprehensive epidemiological study to estimate the actual (based on existing empirical studies; for 43 countries) and predicted (for countries with one or no study; 131 countries) prevalence of smoking during pregnancy in the general population by country, WHO region, and globally. We also estimated the proportion of pregnant women globally who smoked during pregnancy, by frequency and quantity, and the proportion of women who smoked daily and continued to smoke daily during pregnancy, by WHO region and globally. Added value of this study Globally, 52·9% (95% CI 45·6–60·3) of women who smoked daily continued to smoke daily during pregnancy. The European Region had the highest estimated prevalence of
pregnancy, with established causality.1,2,7,8 Evidence also suggests an association between smoking during pregnancy and major birth malformations such as oral clefts, craniosynostosis, gastroschisis,1,9,10 and intellectual impairment later in life.1,11,12 Economic estimates have shown that the expenditures of medical care that are attributable to birth complications are much higher for women who smoke during pregnancy than for non-smokers.13 This association suggests that the birth complications of women who smoke during pregnancy are more severe and need more intense interventions than those of women who do not smoke during pregnancy. However, the magnitude of such smoking-attributable health-care expenditures has decreased over time.14,15 Alarmingly, surveillance data from the Global Youth Tobacco Surveys16 suggest that the difference in the prevalence of smoking between the two sexes is narrowing, especially in low-income and middle-income countries. In view of the effects of smoking on reproduction, these data underscore the importance of sex-specific prevalence estimates, especially for women of childbearing age. On the basis of findings from the Global Burden of Disease Study, WHO recently reported that the global agestandardised prevalence of daily smoking in women in 2015 was 5·4%.17 However, although the detrimental health effects of smoking during pregnancy are well established, to the best of our knowledge no estimates exist of the prevalence of smoking during pregnancy for all countries. In one notable study, Caleyachetty and colleagues18 estimated the prevalence of smoking during pregnancy for 54 low-income and middle-income 2
smoking during pregnancy (8·1%, 95% CI 4·0–12·2) but the lowest estimated proportion of women who smoked daily and continued to smoke daily during pregnancy (30·6%, 25·6–36·4). The African Region had the lowest estimated prevalence of smoking during pregnancy (0·8%, 0·0–2·2) but the second highest estimated proportion of women who smoked daily and continued to smoke daily during pregnancy (61·9%, 30·6–100·0). The Western Pacific Region had the highest proportion of women who smoked daily and continued to smoke daily during pregnancy (79·6%, 44·2–100·0). Implications of all the available evidence To the best of our knowledge, this is the first study to estimate the prevalence of smoking during pregnancy for all countries. Although the global prevalence of smoking during pregnancy appears to be relatively low (1·7%, 95% CI 0·0–4·5), smoking during pregnancy is still prevalent in many countries. Knowledge of the prevalence of smoking during pregnancy can inform strategies for developing targeted anti-smoking campaigns and intervention programmes.
countries using data from Demographic and Health Surveys, which were done between 2001 and 2012. Nevertheless, most countries do not have up-to-date population-level prevalence data on smoking during pregnancy, and the global prevalence is there fore unknown. Yet knowledge of the prevalence of ing during pregnancy can inform strategies for smok develop ing targeted advertising anti-smoking cam paigns and intervention programmes. Therefore, to fill these knowledge gaps, we did the first comprehensive epidemiological study to (1) estimate the prevalence of smoking (any frequency or quantity) during pregnancy in the general population, by country, WHO region (ie, African Region, Eastern Mediterranean Region, European Region, Region of the Americas, Southeast Asian Region, and Western Pacific Region), and globally; (2) estimate the proportion of pregnant women who smoke during pregnancy, by frequency and quantity, on a global level; (3) estimate the proportion of women who smoke daily and continue to smoke daily during pregnancy, by WHO region and globally; and (4) determine whether the prevalence of smoking during pregnancy in the general population has changed in the past 30 years. This study is intended to provide further insight into the global burden of smoking during pregnancy and provide clinicians and public health authorities with the best available data at this time. Smokeless tobacco was excluded from this study because data are scarce, and important potential predictors, such as the prevalence of smokeless tobacco use in women in all countries, are not available at this time.
www.thelancet.com/lancetgh Published online May 30, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30223-7
Articles
Methods We have adhered to the Guidelines for Accurate and Transparent Health Estimates Reporting guidelines19 and to the standards set out in Preferred Reporting Items for Systematic Reviews and Meta-Analyses.20
21 305 records identified by database search
24 records identified through other sources
8073 excluded (duplicates)
Search strategy and selection criteria A comprehensive systematic literature search was done to identify all studies that have reported the prevalence of smoking during pregnancy in the general population in any country. We searched CINAHL, Embase, ERIC, Medline, Medline in process, PsychINFO, Scopus, and Web of Science for studies published between Jan 1, 1985, and Feb 1, 2016 (ie, a 30-year period), using terms that are listed in the appendix (p 4) and without language or geographical restrictions. The content pages of the major epidemiological journals and citations in the relevant articles were also manually screened. The full review protocol is available in PROSPERO, registration number CRD42017075837. Articles were retained if they: consisted of original, quantitative research published in a peer-reviewed journal; reported the prevalence of smoking during pregnancy in the general population; and provided a measure of uncertainty (CI or SE) for the prevalence or at least two of the following: sample size, number of cases, or prevalence. Articles were excluded if they: did not include lifetime non-smokers in their sample or estimate; used a sample not generalisable to the general population of the respective country (ie, studies had to use a sample that was generalisable to the general population at either the national or subnational level; as such, we included nationally or regionally representative surveys or hospital-based studies with full coverage of births within a specific timeframe); or did not provide primary data (ie, estimates had to be based on individuallevel data). For the purposes of this study, the term smoking refers to smoking tobacco in the form of cigarettes. A lifetime non-smoker was defined as a woman who had never smoked a cigarette. Conversely, a smoker was defined as a woman who had smoked at least one cigarette during her pregnancy, either daily (at least one cigarette every day) or occasionally (at least one cigarette per occasion [less than daily]). We appraised the quality of each included study using a modified version of a critical appraisal tool for use in systematic reviews addressing questions of prevalence.21
Data analysis For countries with two or more available empirical studies reporting the prevalence of smoking during pregnancy, we did country-specific random-effects metaanalyses.22 Before the meta-analyses, prevalence estim ates were transformed using a double arcsine transformation23 so that the data followed a normal distribution (an assumption needed when statistically
13 256 identified for screening
12 235 excluded after full-text screening
1021 full-text articles assessed for eligibility
726 excluded 546 lacked relevant data 180 did not meet inclusion criteria
See Online for appendix
For the PROSPERO database see http://www.crd.york.ac.uk/ PROSPERO/
295 eligible studies
Figure 1: Study selection
combining estimates). The resulting combined point estimates and respective CI were then back-transformed to facilitate interpretation. For countries with either one or no available study, the prevalence (any frequency or quantity) was predicted using a multilevel fractional response regression model.24 A fractional response model does not rely on any distributional assumptions; however, the outcome variable must range between zero and one.24 A multilevel approach was chosen to account for the likelihood that estimates within countries would be more similar to one another than would estimates between countries. Additional methodological details on the estimation of the prevalence of smoking (any frequency or quantity) during pregnancy in the general population, by country, are presented in the appendix (pp 2, 3). To estimate the prevalence of smoking during pregnancy by WHO region and globally, we calculated a weighted average of the country-specific point estimates, weighted by the number of livebirths in each country for the latest available year (2001–15).25 To estimate the proportion of pregnant women who smoked during pregnancy by frequency and quantity on a global level, we first did random-effects meta-analyses of the frequency and quantity of smoking, using available data from all countries. Frequency was categorised as either occasional smoking (at least one cigarette per occasion, but less than daily) or as daily smoking (at least one cigarette every day). Quantity was categorised as either light smoking (up to ten cigarettes per smoking day), moderate smoking (11–19 cigarettes per smoking day), or heavy smoking (≥20 cigarettes per smoking day). We then
www.thelancet.com/lancetgh Published online May 30, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30223-7
3
Articles
Prevalence (%) 0–1·5 1·6–2·5 2·6–5·0 5·1–7·5 7·6–10·0 10·1–15·0 15·1–20·0 ≥20·1 No data
Figure 2: Global prevalence of smoking during pregnancy in 2015
applied the estimated proportions of occasional and daily smoking (frequency) and light, moderate, and heavy smoking (quantity) to the estimated global prevalence of smoking during pregnancy. To derive the CI for each respective point estimate, we did Monte Carlo simulations,26 generating 100 000 samples per category, and using the 2·5th and 97·5th percentiles of the resulting distribution as the CI. To estimate the proportion of daily female smokers in the general population who continued to smoke daily during pregnancy by WHO region and globally, we first calculated a weighted average of the country-specific point estimates (data obtained directly from the Institute of Health Metrics and Evaluation), weighted by the adult female population in each country. We then applied the estimated proportion of pregnant women who smoked daily to the regional and global estimates of the prevalence of smoking during pregnancy. We did Monte Carlo simulations to estimate the CI for each estimated proportion. We used a univariate mixed-effects fractional response model24 to determine whether the prevalence of smoking during pregnancy has changed in the past 30 years. Given that most studies came from the USA, we also did a sensitivity analysis by removing all studies from the USA to determine whether they were driving the observed trend. 4
Role of the funding source The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.
Results Of 21 329 studies identified in the literature search, 295 studies with relevant data were retained for data extraction (figure 1; appendix p 5). These 295 empirical studies contained 373 estimates from 104 countries on the prevalence of smoking during pregnancy, as several studies reported estimates of the prevalence of smoking during pregnancy for multiple geographies (ie, for more than one country or for more than one site within a country).18,27–33 Data were available for 32 of 47 countries in the African Region, 16 of 35 countries in the Region of the Americas, six of 21 countries in the Eastern Mediterranean Region, 35 of 53 countries in the European Region, six of 11 countries in the Southeast Asian Region, and nine of 27 countries in the Western Pacific Region. 46 estimates were available for light smoking, 34 estimates for moderate smoking, and 29 estimates for heavy smoking during pregnancy. The prevalence of smoking occasionally was reported in 15 studies, whereas the prevalence of smoking daily during pregnancy was
www.thelancet.com/lancetgh Published online May 30, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30223-7
Articles
Frequency
Quantity
Moderate (34·8%) Occasional (27·5%)
Daily (72·5%)
Heavy (13·5%)
Light (51·8%)
Figure 3: Proportion of pregnant women who smoked during pregnancy, by frequency and quantity, globally Frequency was categorised as occasional smoking (at least one cigarette per occasion, but less than daily) or daily smoking (at least one cigarette every day). Quantity was categorised as light smoking (1–10 cigarettes per smoking day), moderate smoking (11–19 cigarettes per smoking day), and heavy smoking (≥20 cigarettes per smoking day). 25·0
During pregnancy All women
Prevalence (%)
20·0 15·0 10·0 5·0
y all ob Gl
Re gi on Re gi o no Ea ste Am f th rn M e e ed rica ite s rra ne a Re n gi Eu on ro pe a So ut Reg n he io as n tA si R an W es egio te n rn Pa cifi Re c gi on
0
Af ric an
reported in 20 studies. The critical appraisal of included studies and a complete reference list is available in the appendix (p 11, 24). The prevalence of smoking during pregnancy was estimated for 147 countries (via meta-analysis for 43 countries [with two or more available empirical studies] and via statistical modelling [prediction] for 131 countries). The prevalence of smoking during pregnancy could not be estimated for 20 countries because of missing data for one or more predictor variables. The final model was based on 372 of the 373 available estimates (ie, those with complete covariate data) from 104 countries and included the following predictor variables: age-standardised prevalence of daily smoking in women of childbearing age (β=0·80 [referring to an increase of 10%], 95% CI 0·62–0·97), gross domestic product at purchasing price parity (β=–0·08 [referring to an increase of 10 000 international dollars], 95% CI –0·21 to 0·05), proportion of the population living in urban areas (β=0·35 [referring to a 10% increase], 95% CI 0·24–0·46), and study year (β=–0·14 [referring to a 10 year increase], 95% CI –0·30 to 0·01). The five countries with the highest estimated prevalence of smoking (any frequency or quantity) during pregnancy were Ireland (38·4%, 95% CI 25·4–52·4; I²=96·5%, based on meta-analysis of five studies), Uruguay (29·7%, 16·6–44·8; based on meta-analysis of three studies), Bulgaria (29·4%, 26·6–32·2; based on statistical modelling), Spain (26·0%, 22·4–29·7; I²=96·3%, based on meta-analysis of ten studies), and Denmark (25·2%, 18·6–32·4; I²=99·8%, based on meta-analysis of six studies). The five countries with the lowest prevalence of smoking during pregnancy were Tanzania (0·2%, 0·0–0·6; based on meta-analysis of two studies), Burundi (0·3%, 0·0–0·6; based on statistical modelling), St Lucia (0·3%, 0·0–0·6; based on statistical modelling), Sri Lanka (0·3%, 0·0–0·6; based on statistical modelling), and Malawi (0·3%, 0·0–0·7; based on statistical modelling). The prevalence of smoking during pregnancy, by country, is shown in figure 2 and in the appendix (p 18). The results of the tests of heterogeneity and publication bias for the meta-analyses of smoking during pregnancy in the general population, by country, are shown in the appendix (p 21). The global prevalence of smoking during pregnancy was estimated to be 1·7% (95% CI 0·0–4·5). The highest prevalence of smoking during pregnancy was in the European Region (8·1%, 4·0–12·2), and the lowest prevalence of smoking during pregnancy was in the African Region (0·8%, 0·0–2·2). Globally, 72·5% (95% CI 70·4–75·0) of women who smoked during pregnancy were daily smokers, and 27·5% (25·4–29·6) were occasional smokers (figure 3). In terms of quantity, more than half of all women who smoked during pregnancy (51·8%, 95% CI 50·0–53·5) were light smokers, 34·8% (33·1–36·4) were moderate smokers, and 13·5% (12·3–14·7) were heavy smokers.
Figure 4: Prevalence of daily smoking in women and pregnant women in the general population and estimated proportion of women who smoke daily who continue to smoke daily during pregnancy, by WHO region
Results of the tests of heterogeneity and publication bias for the meta-analyses of frequency and quantity of cigarettes smoked during pregnancy in the general population, by country, are shown in the appendix (p 22). The proportion of women who smoked daily and continued to smoke daily during pregnancy ranged from 30·6% (95% CI 25·6–36·4) in the European Region to 79·6% (44·2–100·0) in the Western Pacific Region (figure 4). Globally, this proportion was estimated to be 52·9% (95% CI 45·6–60·3). The estimated regional averages of the prevalence of daily smoking in women and daily smoking in pregnant women in the general population is shown in the appendix (p 23). The univariate multilevel fractional response model showed that the prevalence of smoking during pregnancy has significantly decreased in the past 30 years (β=–0·47 [referring to a 10 year increase], 95% CI –0·56 to –0·39). After removing all studies from the USA, the decrease in the prevalence of smoking during pregnancy in the past
www.thelancet.com/lancetgh Published online May 30, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30223-7
5
Articles
30 years did not change meaningfully (β=–0·38 [referring to a 10 year increase], –0·46 to –0·29).
Discussion This is the first study to estimate the prevalence of smoking during pregnancy nationally, regionally, and globally and to estimate the proportion of pregnant women globally who smoke during pregnancy by frequency and quantity. This is also the first study to estimate the proportion of women who smoke daily and continue to smoke daily during pregnancy, regionally and globally. Although social norms discourage women from smoking during pregnanc, our findings show that smoking during pregnancy is still a prevalent behaviour in many countries. Specifically, the estimated prevalence of smoking during pregnancy was more than 10% in 29 (17%) of 174 countries and more than 20% in 12 (7%) countries. Our findings also indicate that more than 70% of pregnant women who smoked during pregnancy did so daily rather than occasionally. This finding is in line with previous findings that women who quit smoking before pregnancy had smoked fewer cigarettes before pregnancy than women who did not quit,34 suggesting that daily smokers are less likely to quit smoking upon pregnancy recognition. We also found that, globally, more than half of women who smoked daily continued to do so while they were pregnant. The clinical implications of these findings are note worthy, particularly the fact that although the prevalence appears to be decreasing overall, smoking during pregnancy is still prevalent in several countries. Obstetricians, gynaecologists, midwives, and other pre natal health-care workers in these countries should be aware of the magnitude of the problem and promote smoking cessation interventions that target pregnancy, whether pharmacological, psychological, or both. The preconception period is the optimal time to intervene.35 Thus, preconception prevention strategies need to be implemented in countries where the prevalence of smoking during pregnancy remains high. The imple mentation of such strategies might be particularly challenging in low-income and middle-income countries because of obstacles such as economic factors, low perception of risks in the public, lack of policies that promote cessation, poor health-care systems, lack of infrastructure, and industry action.36 In low-income and middle-income countries where the prevalence of smoking during pregnancy is high, the focus should be on building capacity by, for example, integrating smoking cessation services with other health-care services, train ing health-care professionals, and developing guide lines.36 WHO-recommended tobacco control policies and legislations have increased globally, with the largest increases in low-income and middle-income countries. However, the negative relation between higher imple mentation and management of tobacco control and 6
smoking prevalence is stronger in men than in women.37 In view of the adverse effects of smoking on the developing fetus, this relation underscores the importance of sex-sensitive prevention initiatives and smoking cessation programmes that specifically target women of childbearing age. As such, the prevalence estimates presented here serve to inform stakeholders and policy makers of the magnitude of the public health problem in their respective countries. The predicted prevalence estimates of smoking during pregnancy are particularly useful for countries where existing tobacco use surveillance systems or data are absent. We have therefore provided countries with a working estimate of the prevalence of smoking during pregnancy, which can be refined when data become available. Women with intended conceptions are more likely than women with unintended pregnancies to quit smoking upon pregnancy recognition.38 However, a large percentage of pregnancies globally are unplanned.39 As such, WHO recommends that a five-step process be used in all health-care settings to the extent possible.40 These five steps are: (1) asking about smoking history and current smoking behaviour; (2) advising smoking women of the value of quitting and the risks of continuing; (3) assessing a smoking woman’s motivation and thoughts on quitting; (4) assisting the smoker in all attempts to quit with the use of materials or referrals to intervention programmes; and (5) arranging follow-up throughout pregnancy and the post-partum period.41 Importantly, the primary focus should be on smoking cessation. However, if a woman continues to smoke while pregnant, she should be supported in reducing the quantity of smoking. A few limitations of our study should be acknowledged. First, the pooled prevalence of smoking during pregnancy is based on data from the past 30 years. As shown in the time–trend analysis, the prevalence of smoking during pregnancy decreased with time, which could have resulted in a slight inflation of both the estimates obtained via meta-analysis and the regional and global estimates because they are weighted averages of both actual and predicted data. However, the time– trends of smoking during pregnancy might differ between countries, so the predicted estimates for some countries in which the prevalence of smoking in women increased in the past three decades might be slightly underestimated.42 Second, because of the data availability, we used age-standardised prevalence of daily smoking in women of childbearing age in the statistical model, as opposed to the prevalence of smoking at any frequency. However, the prevalence of daily smoking could be a stronger indicator of smoking during pregnancy. Third, it was not possible to obtain country-specific estimates of the prevalence of smoking during pregnancy by frequency and quantity because of data limitations. Instead, these indicators were pooled across countries
www.thelancet.com/lancetgh Published online May 30, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30223-7
Articles
and applied to the global prevalence estimate. We acknowledge that the estimates of the prevalence of smoking during pregnancy by frequency and quantity are likely to be biased towards the countries from which the data originated; however, these are the best available global estimates at this time. Fourth, the predicted prevalence estimates for the 131 countries with either one or no available study might diverge from the true prevalence because the data from which the values were predicted carry some measurement error, and other relevant explanatory variables might affect the prevalence of smoking during pregnancy and not be possible to account for. Other valid approaches to predicting country-specific prevalence of smoking during pregnancy exist (eg, a dynamic model) but rely on additional information (eg, data on smoking initiation and the prevalence of smokers and never-smokers). However, taking into consideration the study and that we were limited to the information reported in the included studies, we consider the present model to yield the best estimates possible. Finally, as would be expected when pooling estimates across time and locations, we observed high heterogeneity in the meta-analyses. However, heterogeneity can be overestimated when summarising studies with large sample sizes.43 The findings highlight a number of key areas for improvement in all parts of the world with respect to tobacco use during pregnancy, which includes both smoked and smokeless tobacco. First, women of child bearing age need to be educated about the potential detrimental effects of tobacco use on the developing fetus and infant. It is important that the messages about tobacco use during pregnancy be provided to women clearly and consistently.40 Prenatal care provides an opportune time for health-care professionals to ask about tobacco use and provide appropriate referrals to smoking cessation programmes whenever necessary. Second, evidence-based smoking cessation interventions should be integrated into routine prenatal care. Third, universal screening protocols need to be established for all pregnant women and women of childbearing age. Finally, prevention efforts and intervention programmes need to be expanded to reach the broader population of women of childbearing age.44 Contributors SL led the conception and design of the study, the development of the data collection instrument, data collection and quality assessment, did the statistical analysis, interpreted the data, and wrote and revised the manuscript. CP contributed to study design, did the statistical analysis, assisted in data interpretation, and wrote and revised the manuscript. JR and SP contributed to data interpretation and revised the intellectual content of the manuscript. Declaration of interests We declare no competing interests. References 1 Centers for Disease Control. Publications and reports of the Surgeon General. How tobacco smoke causes disease: the biology and behavioral basis for smoking-attributable disease: a report of the Surgeon General. Atlanta, GA: Centers for Disease Control and Prevention, 2010.
2 3 4 5 6 7
8 9 10 11 12
13 14 15 16 17
18
19 20
21 22 23 24 25 26
DiFranza JR, Lew RA. Effect of maternal cigarette smoking on pregnancy complications and sudden infant death syndrome. J Fam Pract 1995; 40: 385–94. Pineles BL, Park E, Samet JM. Systematic review and meta-analysis of miscarriage and maternal exposure to tobacco smoke during pregnancy. Am J Epidemiol 2014; 179: 807–23. Rogers JM. Tobacco and pregnancy. Reprod Toxicol 2009; 28: 152–60. Salihu HM, Wilson RE. Epidemiology of prenatal smoking and perinatal outcomes. Early Hum Dev 2007; 83: 713–20. Tikkanen M, Nuutila M, Hiilesmaa V, Paavonen J, Ylikorkala O. Prepregnancy risk factors for placental abruption. Acta Obstet Gynecol Scand 2006; 85: 40–44. Ko TJ, Tsai LY, 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: 20–27. Taylor JA, Sanderson M. A reexamination of the risk factors for the sudden infant death syndrome. J Pediatr 1995; 126: 887–91. Little J, Cardy A, Munger RG. Tobacco smoking and oral clefts: a meta-analysis. Bull World Health Organ 2004; 82: 213–18. Werler MM, Sheehan JE, Mitchell AA. Association of vasoconstrictive exposures with risks of gastroschisis and small intestinal atresia. Epidemiology 2003; 14: 349–54. Drews CD, Murphy CC, Yeargin-Allsopp M, Decoufle P. The relationship between idiopathic mental retardation and maternal smoking during pregnancy. Pediatrics 1996; 97: 547–53. Mortensen EL, Michaelsen KF, Sanders SA, Reinisch JM. A dose-response relationship between maternal smoking during late pregnancy and adult intelligence in male offspring. Paediatr Perinat Epidemiol 2005; 19: 4–11. Centers for Disease Control. Medical-care expenditures attributable to cigarette smoking during pregnancy—United States, 1995. MMWR Morbid Mortal Wkly Rep 1997; 46: 1048–50. Adams EK, Miller VP, Ernst C, Nishimura BK, Melvin C, Merritt R. Neonatal health care costs related to smoking during pregnancy. Health Econ 2002; 11: 193–206. Adams EK, Melvin CL, Raskind-Hood C, Joski PJ, Galactionova E. Infant delivery costs related to maternal smoking: an update. Nicotine Tob Res 2011; 13: 627–37. Warren CW, Jones NR, Peruga A, et al. Global youth tobacco surveillance, 2000–2007. MMWR Surveill Summ 2008; 57: 1–28. GBD 2015 Tobacco Collaborators. Smoking prevalence and attributable disease burden in 195 countries and territories, 1990–2015: a systematic analysis from the Global Burden of Disease Study 2015. Lancet 2017; 389: 1885–906. Caleyachetty R, Tait CA, Kengne AP, Corvalan C, Uauy R, Echouffo-Tcheugui JB. Tobacco use in pregnant women: analysis of data from Demographic and Health Surveys from 54 low-income and middle-income countries. Lancet Glob Health 2014; 2: e513–20. Stevens GA, Alkema L, Black RE, et al. Guidelines for accurate and transparent health estimates reporting: the GATHER statement. Lancet 2016; 388: e19–23. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. PLoS Med 2009; 6: e1000100. Munn Z, Moola S, Riitano D, Lisy K. The development of a critical appraisal tool for use in systematic reviews addressing questions of prevalence. Int J Health Policy Manag 2014; 3: 123–28. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986; 7: 177–88. Freeman MF, Tukey JW. Transformations related to the angular and the square root. Ann Math Stat 1950; 21: 607–1. Papke LE, Wooldridge J. Econometric methods for fractional response variables with an application to 401 (k) plan participation rates. J Appl Econ 1996; 11: 619–32. United Nations. Population and vital statistics report. New York, NY: United Nations, 2016. Graham C, Talay D. Stochastic Simulation and Monte Carlo Methods: Mathematical Foundations of Stochastic Simulation. Berlin: Springer, 2013.
www.thelancet.com/lancetgh Published online May 30, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30223-7
7
Articles
27 Bloch M, Althabe F, Onyamboko M, et al. Tobacco use and secondhand smoke exposure during pregnancy: an investigative survey of women in 9 developing nations. Am J Public Health 2008; 98: 1833–40. 28 Grant TM, Huggins JE, Sampson PD, Ernst CC, Barr HM, Streissguth AP. Alcohol use before and during pregnancy in western Washington, 1989–2004: implications for the prevention of fetal alcohol spectrum disorders. Am J Obstet Gynecol 2009; 200: 278.e1–8. 29 Ostrea EM Jr, Villanueva-Uy E, Ngerncham S, et al. An epidemiologic study comparing fetal exposure to tobacco smoke in three Southeast Asian countries. Int J Occup Environ Health 2008; 14: 257–62. 30 Pichini S, Morini L, Marchei E, et al. Ethylglucuronide and ethylsulfate in meconium to assess gestational ethanol exposure: preliminary results in two Mediterranean cohorts. Can J Clin Pharmacol 2009; 16: e370–75. 31 Primatesta P, Del Corno G, Bonazzi MC, Waters WE. Alcohol and pregnancy: an international comparison. J Public Health Med 1993; 15: 69–76. 32 Rodriguez A, Olsen J, Kotimaa AJ, et al. Is prenatal alcohol exposure related to inattention and hyperactivity symptoms in children? Disentangling the effects of social adversity. J Child Psychol Psychiatry 2009; 50: 1073–83. 33 Smedberg J, Lupattelli A, Mardby AC, Nordeng H. Characteristics of women who continue smoking during pregnancy: a cross-sectional study of pregnant women and new mothers in 15 European countries. BMC Pregnancy Childbirth 2014; 14: 213. 34 Curtin SC, Matthews TJ. Smoking prevalence and cessation before and during pregnancy: data from the birth certificate, 2014. Natl Vital Stat Rep 2016; 65: 1–14.
8
35 Dean SV, Imam AM, Lassi ZS, Bhutta ZA. Importance of intervening in the preconception period to impact pregnancy outcomes. Nestle Nutr Inst Workshop Ser 2013; 74: 63–73. 36 Abdullah AS, Husten CG. Promotion of smoking cessation in developing countries: a framework for urgent public health interventions. Thorax 2004; 59: 623–30. 37 Anderson CL, Becher H, Winkler V. Tobacco control progress in low and middle income countries in comparison to high income countries. Int J Environ Res Public Health 2016; 13: 1039. 38 Kost K, Landry DJ, Darroch JE. Predicting maternal behaviors during pregnancy: does intention status matter? Fam Plann Perspect 1998; 30: 79–88. 39 Sedgh G, Singh S, Hussain R. Intended and unintended pregnancies worldwide in 2012 and recent trends. Stud Fam Plann 2014; 45: 301–14. 40 WHO. WHO recommendations for the prevention and management of tobacco use and second-hand smoke exposure in pregnancy. Geneva, Switzerland: World Health Organization, 2013. 41 Fiore MJC, Baker TB, Bailey WC, et al. A clinical practice guideline for treating tobacco use and dependence: 2008 update. A US Public Health Service report. Am J Prev Med 2008; 35: 158–76. 42 Ng M, Freeman MK, Fleming TD, et al. Smoking prevalence and cigarette consumption in 187 countries, 1980–2012. JAMA 2014; 311: 183–92. 43 Rücker G, Schwarzer G, Carpenter JR, Schumacher M. Undue reliance on I(2) in assessing heterogeneity may mislead. BMC Med Res Methodol 2008; 8: 79. 44 Orleans CT, Barker DC, Kaufman NJ, Marx JF. Helping pregnant smokers quit: meeting the challenge in the next decade. Tob Control 2000; 9 (suppl 3): III6–11.
www.thelancet.com/lancetgh Published online May 30, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30223-7