Maternal smoking and the risk of still birth: systematic review and meta-analysis
© Marufu et al.; licensee BioMed Central. 2015
Received: 24 October 2014
Accepted: 16 February 2015
Published: 13 March 2015
Smoking in pregnancy is known to be associated with a range of adverse pregnancy outcomes, yet there is a high prevalence of smoking among pregnant women in many countries, and it remains a major public health concern. We have conducted a systematic review and meta-analysis to provide contemporary estimates of the association between maternal smoking in pregnancy and the risk of stillbirth.
We searched four databases namely MEDLINE, EMBASE, Psych Info and Web of Science for all relevant original studies published until 31st December 2012. We included observational studies that measured the association between maternal smoking during pregnancy and the risk of stillbirth.
1766 studies were screened for title analysis, of which 34 papers (21 cohorts, 8 case controls and 5 cross sectional studies) met the inclusion criteria. In meta-analysis smoking during pregnancy was significantly associated with a 47% increase in the odds of stillbirth (OR 1.47, 95% CI 1.37, 1.57, p < 0.0001). In subgroup analysis, smoking 1-9 cig/day and ≥10 cig/day was associated with an 9% and 52% increase in the odds of stillbirth respectively. Subsequently, studies defining stillbirth at ≥ 20 weeks demonstrated a 43% increase in odds for smoking mothers compared to mothers who do not smoke, (OR 1.43, 95% CI 1.32, 1.54, p < 0.0001), whereas studies with stillbirth defined at ≥ 24 weeks and ≥ 28 weeks showed 58% and 33% increase in the odds of stillbirth respectively.
Our review confirms a dose-response effect of maternal smoking in pregnancy on risk of stillbirth. To minimise the risk of stillbirth, reducing current smoking prevalence in pregnancy should continue to be a key public health high priority.
Smoking in pregnancy is a major public health problem in the developed countries . The World Health Organisation (WHO) predicts that this will reach ‘epidemic’ proportions in developing countries in the near future . Within the current challenging economic climate in many countries, smoking in pregnancy imposes a significant burden on population health and resources, and is associated with a range of poor outcomes for both mother and child, such as ectopic pregnancy, miscarriage, placental abruption, preterm birth and low birth weight [2,3]. The harmful effects of tobacco smoke exposure in pregnancy can be avoided  and it is one of the most prevalent modifiable risk factors for adverse pregnancy outcomes [5,6].
In many developed countries, the rates of smoking in pregnancy have been declining over recent decades  with current prevalence estimates between 10% – 19% , and data from the UK suggesting that one out of eight women smoke throughout the pregnancy . Moreover, this decline has not been consistent across all social classes; lower rates of decline have been noted across less advantaged socioeconomic groups . Smoking in pregnancy in developed countries tends to be higher among women who have low income and poor education .
Stillbirth rates widely vary across low, middle and high-income countries [12-14]. The lowest rates have been observed in Finland with 2.0 per 1000 live births and in the developing world rates are much higher at more than 40 per 1000 live births in countries like Nigeria and Pakistan . Two previous systematic reviews and meta-analyses have estimated the size of effect of smoking during pregnancy and the risk of stillbirth. A report by the United States Surgeon General  showed a relative risk of stillbirth of 1.2-1.8 within smokers versus non-smokers. The analysis was based on three studies [16-18] conducted in two countries, the United States and Sweden. Recent stillbirth statistics by Flenady et al  was based on a systematic review from high-income countries; and four studies were included in the meta-analysis of the association between any smoking in pregnancy and the risk of stillbirth yielding an odds ratio of 1.36 (95% CI 1.27, 1.46). Results from these published reviews were limited to literature from developed countries; yet stillbirth rates are much higher in developing countries, and it is therefore imperative to conduct a comprehensive review, which reflects the impact of smoking in pregnancy on stillbirth. We have conducted a systematic review and meta-analysis to provide contemporary estimates of the association between maternal smoking in pregnancy and the risk of stillbirth.
A detailed electronic search was performed through four databases namely MEDLINE, EMBASE, Psych Info and Web of Science. All relevant published studies with sufficient data on maternal cigarette or tobacco smoking during pregnancy with the outcome of stillbirth were included. Due to the nature of the research objective, there were no randomised control trials (RCTs) identified, and observational studies (such as cohort design, case control studies and cross sectional surveys) were considered using the standard guidelines of Meta-analysis of Observational Studies in Epidemiology (MOOSE) . All relevant studies published in the English language, up to 31st December 2012 were included in the review. Case reports, non-English publications and those, which only involved passive or environmental smoke information, were excluded. We carefully considered studies where the study populations were similar to avoid duplication of cases; where studies were based on exactly the same population during the same time period, but there were differences in sample sizes, we selected the results from the publication with the larger sample size for inclusion in the meta-analysis, though those excluded from quantitative analysis were included in a narrative synthesis.
We developed search terms based on medical subject headings (MeSH), free text words and words in the title or abstract. We combined search terms of exposure (smoking during pregnancy) with the outcome of the study (stillbirth). The MeSH terms used included combinations of the terms: stillbirth, maternal smoking, and pregnancy, also applied with special characters ($, *), wherever required. The study protocol was agreed within the team, however it was not published prior to commencement of the review. We also checked the reference list of all identified papers and the most recent similar reviews for any additional studies. A complete online search strategy along with search terms is attached (Additional file 1).
Study selection and outcome definition
The electronic search was performed with a three-stage approach; first titles were screened by TM and AA, then abstracts were screened by two authors independently (TM and SL). For studies, which appeared to be eligible, full texts were obtained and reviewed independently by two authors who further performed data extraction independently on pre-piloted forms, and discrepancies were resolved by consensus. We also obtained full texts for those studies where a decision could not be made based on title and abstract. Data extraction forms included followings; study design, exposure validation, definition of exposure (smoking) and outcome (stillbirth), confounders adjustments, sample size and study location.
Stillbirth at ≥ 20 weeks of gestation
Stillbirth at ≥ 24 weeks of gestation
Stillbirth at ≥ 28 weeks of gestation
Where both unadjusted and adjusted measures were reported, the latter were extracted. Quality assessment was based on a Newcastle Ottawa Scale (NOS). Cohort and case control studies were awarded up to a maximum quality score of nine and cross sectional studies were given up to a maximum score of seven . A definition of ‘high quality’, was given to cohort and case control studies with a score of 7 or above or cross-sectional studies with a score of 5 or above; the rest were deemed ‘low quality’ . Each quality assessment (NOS) was conducted by at least by two authors independently and then findings were verified.
Meta-analysis was performed using Rev-Man 5.3  with a random effects model. The I2 statistic was applied to calculate heterogeneity (I2 more than 75% was considered high heterogeneity, more than 50% moderate and 25% was considered as low heterogeneity) . The odds ratio for the overall effect of smoking in pregnancy from each study was used, which was the main outcome reported in included literature. The ratio was calculated from the available data, wherever feasible. Studies included in the meta-analysis were further considered for subgroup analysis based on study quality, definition of stillbirth and cigarette consumption (1-9 cig/day and ≥10 cig/day). The funnel plot method was used to assess publication bias. Studies without statistical data presentation, and those where cases were potentially overlapping between publications, were considered in a narrative synthesis.
Table of included studies
Exposure: smoking during pregnancy and validity
Definition of stillbirth (outcome) and gestational period
Factors adjusted for in individual studies if available
Studies included in meta-analysis
Wood et al (2009)
Alberta Perinatal Health Program, 1991-2004 Canada
Fetal death after at least 24 completed weeks of gestation.
Adjusted OR 1.04 (0.74-1.47)
Maternal age, diabetes, hypertension, previous caesarean section.
Smith et al (2007)
Scotland morbidity records 1991-2001, UK
Fetal death after at least 24 completed weeks of gestation.
OR for smoking was calculated 2.15 (95% CI 1.74-2.65)
Age, deprivation, height, BMI, Marital Status.
Hogberg and Cnattingius (2007)
Sweden – Swedish National study 1983-2001- Participants with two consecutive pregnancies.
Fetal death after at least 28 completed weeks of gestation.
OR for smoking in both Pregnancies 1.35 (1.15-1.58). 1-9cig/day OR 1.16 (0.92-1.46) ≥10 cig/day 0R 1.55 (1.17-2.04). OR for smoking in the 1st pregnancy only OR 1.02 (0.79-1.30), 1-9 cig/day OR 1.13 (0.9-1.39), ≥10 cig/day OR 1.14 (0.87-1.51). OR for smoking in 2nd pregnancy only 0.84 (0.55-1.26), 1-9cig/day OR 1.10 (0.89-1.35), ≥ 10 cig/day OR 1.45 (1.16-1.81).
Maternal age, education, cohabiting with infant’s father, mother’s country of birth, year of second delivery, inter-pregnancy interval and stillbirth in first pregnancy.
1-9 cig/day, ≥10 cig/day.
Salihu et al (2008)
USA – Missouri 1978-1997.
In-utero fetal death at least 20 week into gestation.
a) Bidirectional case cross-over OR 1.20 (1.03-1.39) 1-9 cig/day OR 1.16 (0.88-1.53) 10-19 cig/day OR 1.10 (0.90-1.24) ≥ 20 cig/day OR 1.21 (1.01-1.50),
Maternal race and age.
b) Case Control OR 1.34 (1.26-1.43) 1-9 cig/day OR 0.94 (0.76-1.70) 10-19 cig/day OR 1.31(1.22-1.41) ≥20 cig/day OR 1.43 (1.31-1.54).
Ahlenius and Thomassen (1999)
94,270 in 1984 and 124,201 in 1991.
Sweden - Using the Swedish medical birth register in 1984 and 1991
Death of a fetus before or during delivery with a gestational duration of at least 28 weeks.
Combined (1984 and 1991) OR 1.37 (1.17 to 1.61)
Maternal age and parity.
1-9 cig/day, ≥10 cig/day.
1-9 Cig/day OR 1.22
≥10 Cig/day OR 1.62 (1.31-2.00)
1.47 (1.19 to 1.82)
>10 Cig/day OR 1.97 (1.51-2.59)
Aliyu et al (2007)
USA Missouri – Using Missouri maternally linked cohort data 1978 – 1997
In utero fetal death at < = 20 weeks gestation
OR(originally presented as HRs) 1.43 (1.36-1.51)
1-9 cig/day, 10-19 cig/day, ≥ 20 cig/day.
Gray et al (2009)
UK- Scotland, National study 1994-2003.
24-44 weeks’ gestation
OR 1.75 (1.61-1.91)
McCowan et al (2007)
New Zealand - National Women’s
Birth of a baby with no signs of life delivered at ≥
OR 1.33 (0.99-1.79)
Infant sex, maternal ethnicity, parity, marital status, maternal age at delivery, history of previous miscarriage, history of previous abortion, previous low-birth weight infant, previous C-section.
Hospital, Auckland, New Zealand from 1993 to 2000
20 weeks or with a birth weight of > 400 g if gestation was not known.
Miller et al (2010)
Canada – Ontario state 2004-2006
Death of fetus at 20 weeks of gestation
Overall OR 1.58 (1.38-1.81)
Age, parity, multiple gestation
Moshin et al (2005)
Australia – New South Wales 1998-2002
Fetal death at ≥ 20 weeks gestation or of 400 g or more birth weight.
Overall adjusted OR 1.17 (1.05–1.31) Unadjusted OR 1.05 (0.94–1.17).
Maternal age, maternal country of birth, socioeconomic status, maternal diabetes mellitus, and first antenatal care visit.
Raymond et al (1994)
Sweden National study 1983-1989
Fetus death at 28 weeks of gestation or more.
OR 1.4 (1.2-1.4)
Age, parity, pregnancy complications, hypertensive diseases, diabetes, placental complications and intrauterine growth retardation (IUGR).
Reddy et al (2006)
USA – 12 clinic centres and 19 hospital sites. 2002-2008.
Fetal death at 23 weeks gestation and beyond.
OR (originally presented as HRs) 1.57 (1.21-2.02)
Sutan et al (2010)
UK- Scotland, National study (NHS Hospital)
Late fetal death from 20 weeks gestation.
OR 1.64 (1.46-1.84)
Tuthill et al (1999)
Wales – All Wales Perinatal Survey 1993-1995
Mortality from 20 completed weeks of gestation
Social class, infection, placental abruption, sudden infant death syndrome.
Wisborg et al (2001)
Denmark – Aarhust University Hospital 1989-1996.
Delivery of a dead fetus occurring at 28 weeks or more gestation.
OR 1.9 (1.3-2.9),
Sex of child, parity, maternal age, education, alcohol, caffeine intake.
1-9 cig/day, ≥10 cig/day.
1-9 cig/day OR 1.5 (0.9-2.4) ≥10 cig/day 1.8 (1.2-2.8)
Winbo et al (2001)
Sweden – National study 1983 - 1995
Death of fetus after competing 28 weeks of gestation.
OR stillbirths ≥37 weeks compared with live births 1.08 (0.96-1.22)
Ferraz and Gray (1991)
Brazil – Natal city and state of Rio Grande do Norte 1984-1986
Stillborn cases weighing 500 g or more
OR 1.4 (1.0-2.0)
Dodds et al (2006)
Canada – Nova Scotia and Eastern Ontario 1999-2001
Not specified - Death before delivery of a fetus weighing 500 g or more
OR (originally presented in HRs) Smoking during 1st trimester 2.13 (1.23-3.64)
Efkarpidis et al (2004)
UK – Nottingham City Hospital 1991-1997.
Definition of stillbirth not stated- pregnancies considered from ≥ 24 weeks
OR 1.2 (0.64-1.6)
Froen et al (2002)
Norway – Oslo 1986-1995.
Intrauterine death before onset of labour of a fetus > =22 weeks gestation or > = 500 g body mass
OR calculated Using Stats 1.81 (1.02-3.15) 1-9 cig/day OR 1.0 (0.43-2.12), ≥10 cig/day 0R 3.04 (1.13-8.19)
1-9 cig/day, ≥10 cig/day.
Goy et al (2008)
Canada – Nova Scotia and Eastern Ontario 1999-2001.
Not specified - Death before delivery of a fetus weighing 500 g or more
OR 2.06 (1.22-3.47)
Province of origin, age, inactivity during pregnancy, stillbirth in previous pregnancy, used fertility treatment
Little and Weiberg (1993)
Fetal death at 28 weeks or more gestation
OR calculated using Stata 1.29 (1.09-1.53)
OR 1-19 cig/day 1.3 (1.09-1.57), 20-29 cig/day OR 1.39 (1.06-1.82), ≥30 cig/day OR 0.67 (0.41-1.12).
≥ 30 cig/day.
Mishra et al (2005)
India - Survey conducted 98-99
Delivery of a dead baby at > 28 weeks of pregnancy
OR 1.23 (0.92-1.64)
Cooking smoke, anaemia, BMI, education, standard of living, house type.
Unadjusted OR 1.60 (1.23-2.08).
Robson et al (2006)
Australia – New South Wales 1998-2003.
OR 1.65 (1.15-2.38)
Age, parity, medical/obstetric complications
1-10 cig/day >10 cig/day.
1- 10 cig/day OR 1.16 (0.80-1.69) >10 cig/day OR 1.68 (1.17-2.41).
Studies included in narrative synthesis
Cnattingius et al (1988)
Sweden National Study – Using Swedish medical birth registry covering more than 99% of all births in Sweden. 1983-1985
Late fetal death – still birth occurring at 28 weeks of gestation or later
OR (originally presented as RR) 1.37 (1.20-1.57)
1-9 cig/day 1.32 (1.12-1.55)
≥10 cig/day 1.45 (1.21-1.75)
Sweden – 5 centres (Boras, Helsingborg, Karlskrona, Kristianstastad and Lund) 1983-1989.
Definition of stillbirth not stated, pregnancies considered
OR calculated using Stata. OR 1.18 (0.62-2.26)
1-10 cig/day OR 1.33 (0.58-3.09) >10 cig/day 1. 04 (0.43-2.50)
≥ 28 weeks
Aliyu et al (2008)
USA Missouri – Using Missouri maternally linked cohort data 1978 – 1997.
In utero fetal death at < = 20 weeks gestation.
OR (originally presented as HRs) 1.48 (1.40-1.56)
< 35 years old, ≥ 35yers old. Not validated
< 35 years smokers1.43 (1.35-1.51)
≥35 years Smokers OR 2.4 (2.06-2.80)
Kleinman et al (1988)
USA Missouri 1978-1983
(Fetal mortality) Fetal death at 20 weeks or more.
< 1 pack/day ≥1 pack/day.
<1 pack/day 1.36(1.16-1.59) ≥1 pack/day 1.62 (1.34-1.97)
<1 pack/day 1.21 (1.06-1.38) ≥1pack/day 1.15(0.99-1.34)
Raatikainen et al (2006)
Finland-Kuopio University Hospital 1989-2001.
Fetal death was defined as intrauterine death before
22 weeks of pregnancy or 500 g fetal weight.
Aliyu et al (2010)
US Missouri – Using Missouri maternally-linked cohort data files 1978 – 1997.
In utero fetal death at > = 20 weeks gestation.
<15 years 3.3 (1.4-7.8), 15-19 years 1.7 (1.5-1.9), 20-24 years 1.5 (1.4-1.6)
No sufficient data to calculate Odds ratios.
Bai et al (2000)
Australia- Liverpool hospital Sydney March 1996 to June 1998
Not Clear (births occurring at less than 20 weeks gestation age exclusion criteria).
Results presented in Percentages. No odds ratios and no sufficient data to calculate OR.
Huang et al (2000)
Canada – The Royal Victoria Hospital in Montreal 1961-1974 and 1978-1996.
Fetal deaths occurring before labour without evidence of significant fetal, maternal or placental pathology weighing 500 g or more.
USA – Missouri 1978-1990
Smokers: Yes Not validated
Fetal death greater or equal to 20 weeks gestation.
Results in RR no sufficient data to calculate OR
Smoking in 1st and not in 2nd pregnancy RR 1.11, Not smoking in 1st but in 2nd pregnancy RR 1.6.smoking in both pregnancies RR 1.19.
Salihu et al (2004)
USA – 1995-1997.
Intrauterine fetal demise after 20 weeks of gestation.
Estimates in Hazard ratios not enough data to calculate OR.
≥ 40 years 2.71 (1.88-3.91), 30-39 years 1.58 (0.96-2.60), 20-29 years 1.41 (1.16-1.71)
Out of the 34 eligible studies (Table 1), eight were case control studies [23-28,54,55], five cross sectional [16,18,40,41,45] and twenty-one were cohort studies [12-14,17,29-39,42-44,56-58]. Fourteen were conducted in Europe, [13,14,16,17,25,27,29,32,36-39,43,54], four in Australia [34,41,56,58], fourteen in North America [12,18,23,26,28,30,31,33,35,42,44,45,55,57], one in Asia  and one in South America . The largest sample size observed was 7,792,990  with the smallest being 202 . Ten studies [13,16,25,26,30-32,35,54,55] did not present their results with odds ratios but had sufficient data to calculate the ratio using STATA 12, which was further used in the meta-analysis. Seven studies [18,42-45,57,58] did not present estimates or sufficient data to calculate an odds ratio; therefore they were included in the narrative synthesis. Based on NOS scale, overall study quality was moderately satisfactory. Ten cohort studies [12,14,17,32,35-39,42] and four case control studies [23,24,54,55] were considered of high quality with seven or more points, and only two cross sectional studies [16,45] were of high quality having five or more points. Scores ranged from 2 to 9 and the median score was 6.
The type of exposure
Eleven studies clearly defined the level of exposure (smoking) [14,16,17,25,29,31,33,34,38,54,56]. Most of the studies used categories of 1-9cig/day and ≥10 cig/day to categorise participants according to the number of cigarettes or packs consumed daily, but the Robson study  used slightly different categories. Seventeen studies did not define smoking according to level of cigarette consumption [12,13,24,26-28,30,32,35,36,40,41,43,45,55,57,58] but according to the smoking status (yes/no). Fifteen studies collected exposure information during pregnancy [12-14,16,17,27,29,30,32,33,38,42,43,54,56] and the rest of the studies [18,23-26,31,34,35,39-41,44,45,55,57,58] obtained the information after delivery. None of the studies reported biochemical validation of smoking status such as salivary cotinine assessments.
Data collection on outcome
The definition of stillbirth varied across studies. Twenty-one out of thirty-four studies included stillbirth from early gestational age ≥ 20 weeks [18,23,30,31,33,34,36,37,41-43,45,56-58], ≥22 weeks , ≥23 weeks , and four studies used birth weight to estimate the gestational week of stillbirth (Birth weight ≥ 500 g, [24,26,28,44]. The rest of the studies used late gestational ages as the cut off values for stillbirth, ≥24 weeks in four studies [12,13,27,32], ≥28 weeks in nine studies [14,16,17,25,29,38-40,55]. In all studies, outcome data was either obtained from medical or clinical records, medical data sets or birth registry records.
Fourteen studies included in the quantitative synthesis adjusted for one or more confounders (Table 1). Maternal age was the most common adjusted factor. Five studies adjusted for socioeconomic status (SES) including education [13,14,37,38,40]. Other factors that were adjusted for include parity, BMI, ethnicity/race, infant sex, perinatal care and marital status, alcohol, caffeine intake and cohabitation. Four studies [24,27,36,39] did not adjust for any confounders and six studies [26,31,32,35,54,55] had OR calculated therefore considered as not adjusted for any confounders. Factors adjusted in cohort studies varied widely. Several studies [12,17,28,34,37,41,56] considered one or more medical conditions such as pregnancy complications, diabetes mellitus and hypertension as possible confounders.
Meta-analysis of maternal smoking and risk of stillbirth
Subgroup analysis by study quality
Results from high quality score studies showed that women who smoke during pregnancy, are 41% more likely to have a stillbirth compared to women who do not smoke during pregnancy (OR 1.41, 95% CI 1.28, 1.55, p <0.0001, 14 studies). Low quality score studies showed that women who smoke during pregnancy were at 49% increased odds of stillbirth compared to women who do not smoke during pregnancy (OR 1.49, 95% CI 1.33, 1.67, p <0.0001, 11 studies). There was no significant difference between the two subgroups (p = 0.44).
Subgroup analysis by categories of cigarette consumption
Subgroup analysis by gestation of stillbirth
Subgroup analysis by study location
Only two studies were conducted in developing countries [24,40]. Both studies support evidence from research conducted in developed countries concluding that women who smoke during pregnancy are at a risk of stillbirth. A study  demonstrated a very marginal difference showing those women who smoke during pregnancy are at least 40% at risk of experiencing a stillbirth (OR 1.40, 95% CI 1.0, 1.96), similar to the pooled odds ratio of studies conducted in North America (OR 1.39, 95% CI 1.29, 1.50) P value < 0.0001, 9 studies) and Europe (OR 1.55, 95% CI 1.36, 1.78, P value < 0.0001,11 studies).
Seven studies [18,42-45,57,58] presented their results in other than odds ratio and there was no sufficient data to calculate the odds ratio, and three studies [16,25,30] were derived from same study population as [17,31] respectively. Six of the studies [16,18,30,42,45,57] observed an association between maternal smoking in pregnancy and the risk of stillbirth but the other four [25,43,44,58] observed no significant differences. Two studies [16,42] observed a dose related response with high risk ratio of stillbirth in mothers who smoke ≥10 cig/day compared to 1-9 cig/day and those who smoke ≥1 pack/day in contrast to <1 pack/day respectively. Three studies (51, 35 54) showed the risk of stillbirth to be higher in young mothers <15 years old (HR 3.3, 95% CI 1.4, 7.8) and also in older mothers (≥35 years, HR 3.2, 95% CI 2.2, 4.5) (≥40 years, HR 2.71, 95% CI 1.88, 3.91) respectively. A study from Sweden suggested that placental abruption is likely to be common in smokers .
Result summary: maternal smoking in pregnancy and the risk of stillbirth
i) Overall pooled Results from main meta-analysis (25 studies)
ii) Subgroup analysis by daily amount of cigarettes consumption
1-9 cig/day (6 studies)
≥ 10 cig/day (7 studies)
iii) Subgroup analysis by gestation period at which stillbirth occurred
Stillbirth ≥ 20 weeks (11 studies)
Stillbirth ≥ 24 week (4 studies)
Stillbirth ≥ 28 weeks (7 studies)
iv) Subgroup analysis by geographic location
North America (9 studies)
Europe (11 studies)
Asia (1 study)
South America (1 study)
Australia (3 studies)
v) Subgroup analysis by quality
High score (14 studies)
Low score (11 studies)
Strengths and limitations
We implemented a comprehensive search strategy with strict adherence to the protocol, and results were presented in accordance with MOOSE guidelines . Search strategy, data extraction, analysis and quality assessment was performed independently by authors and findings were confirmed within the team. There was no evidence of publication bias. However, the review has some limitations. We could not obtain one study within available timelines , and due to limited data presentation, seven studies [18,42-45,57,58] were considered in the narrative synthesis. Secondly, only English language studies were considered mainly for practical purposes and the available timeframe.
Heterogeneity was explored using a variety of subgroup analyses; a high level of heterogeneity was found amongst cohort studies and the pooled estimate derived from these should be treated with some caution. This could be partially attributable to the different factors controlled for in the cohort studies as explained previously. The overall heterogeneity (I2 = 77%, Figure 2) was higher mainly due to cohort studies, however in case of case control studies (I2 = 20%) and cross-sectional studies (I2 = 0%), very low levels of heterogeneity was observed. However, this estimate was consistent with synthesised estimates derived from studies with other designs and this consistency suggests it is likely to be valid. It appears that the effect could differ by age [24,30,38,45,57], but data on maternal age was not available in included studies. Study quality was found to be satisfactory.
The most recent systematic review by Flenady et al  on maternal smoking and the risk of stillbirth, focusing on high-income countries, reported a 36% increase in the odds of stillbirth (OR 1.36). The review considered cohort and case control studies published between 1998 and 2009 and excluded those which did not control for confounding factors, hospital based studies and took a slightly different approach to dealing with duplicate studies. Acknowledging the difference in the approach, our study is in agreement with the results from Flenady’s et al  review, and showed a 46% increased risk of stillbirth (OR 1.46 95% CI 1.36, 1.55). We included 25 studies in our main meta-analysis, conducted in both developing and developed countries making our findings generalisable. Our review has also analysed the available data according to the definition of stillbirth in gestational weeks, number of cigarettes and study quality, which were not reported in the previous published literature. The review also provides strong evidence that the risk is higher at higher levels of cigarette consumption, indicating a dose related response relationship, outlined in some of the included studies [15,38,42].
The consistency in the size of the effect observed in this study throughout different subgroup analyses and across different study designs, each with their own biases and strengths and limitations, as well as the consistency across studies of differing quality, suggests that this is a true estimate of the size of effect. Further support is obtained from studies where mothers change in smoking behaviour from one pregnancy to another. Hogberg and Cnattingius  suggested that mothers who quit smoking from first to second pregnancy reduced their risk of stillbirth to the same level as non-smokers in both pregnancies, while those who smoked in both pregnancies had a 35% increased risk of stillbirth (OR 1.35) compared to non-smokers.
Our results also provide evidence that the effect of maternal smoking in pregnancy on risk of stillbirth is not dependent on the definition of stillbirth or on the stage of pregnancy at which it occurs. Moreover we did not ascertain, at what point during pregnancy the effect of the exposure is occurring as most studies evaluated the exposure at one point in time only. Previous studies [26,38] suggested that the timing of exposure may influence the risk of stillbirth and in particular  suggested that those who quite early in pregnancy may have similar risk to non-smokers. Future studies of this association should measure exposure at different stages of gestation. There are important potential confounders for the effect of maternal smoking on risk of stillbirth, the most important ones are SES, maternal age, maternal weight and medical comorbidities, majority of studies have adjusted for these. Only two studies did not adjust for any confounders [36,43] with no further information.
The exact pathophysiology of fetal exposure to tobacco smoke is not entirely understood, however based on available evidence, possible mechanisms have been conceptualised . Nicotine has been known to cause narrowing of the placental vessels [60,61], which coupled with reduced prostacycline synthesis , resulting in increased vascular resistance and consequently impairing blood supply to the fetus . Carbon monoxide in tobacco smoke also reduces fetal oxygenation by forming carboxy-haemoglobin in turn interfering with oxygenation transfer . The correlation between smoking and stillbirth is likely to be explained through increased prevalence of placental complications and fetal growth restriction . The resultant physiological effect of these changes increases the risk of fetal morbidity (small for gestation age and preterm birth) and subsequently may lead to fetal death . Therefore, there is strong biological plausibility for smoking in pregnancy causing stillbirth [46,65].
Recent Cochrane reviews have concluded that behavioural interventions can be effective in helping women to stop smoking in pregnancy  even though there remains no evidence for the effectiveness of pharmacotherapy such as NRT in pregnancy . The result of our study suggests that such interventions are important to reduce the risk of stillbirth.
Our findings support that smoking greatly increases the risk of stillbirth. Smoking in pregnancy is an established cause of a range of pregnancy complications and poor pregnancy outcomes. Every opportunity must be utilised to record smoking status during pregnancy, to give advice and support, including necessary facilities to help women stop smoking. Although we have conducted a comprehensive review, most of the available evidence on the risk from smoking is from developed countries. Smoking is rapidly increasing in the developing world where stillbirth is a major problem; therefore along with focusing on research initiatives, it is important to invest efforts on educating women on the risks of smoking to their unborn child and providing smoking cessation support for pregnant women in these parts of world.
New castle Ottawa scale
We acknowledge the support of the Greenfield Medical Library staff of University of Nottingham, UK.
- World Health Organisation (WHO). Recommendations for the prevention and management of tobacco use and second- hand smoke exposure in pregnancy. Geneva: World Health Organization; 2013.Google Scholar
- Royal College of Physicians. Passive smoking and children: a report of the Tobacco Advisory Group of the Royal College of Physicians. London: Royal College of Physicians; 2013.Google Scholar
- Royal College of Physicians. Smoking and the young. London: Royal college of Physicians; 1992.Google Scholar
- Cnattingius S, Bergstrom R, Lipworth L, Kramer MS. Pre-pregnancy weight and the risk of adverse pregnancy outcomes. N Eng J Med. 1998;338:147–52.View ArticleGoogle Scholar
- Fretts RC. Etiology and prevention of stillbirth. Am J Obstet Gynecol. 2005;193:1923–35.View ArticlePubMedGoogle Scholar
- Lawn JE, Blencowe H, Pattinson R, Cousens S, Kumar R, Ibiebele I, et al. Stillbirths: Where? When? Why? How to make the data count? Lancet. 2011;377:1448–63.View ArticlePubMedGoogle Scholar
- Cnattingius S. The epidemiology of smoking during pregnancy: smoking prevalence, maternal characteristics, and pregnancy outcomes. Nicotine Tob Res. 2004;6:S125–40.View ArticlePubMedGoogle Scholar
- Flenady V, Koopmans L, Middleton P, Froen JF, Smith GC, Gibbons K, et al. Major risk factors for stillbirth in high-income countries: a systematic review and meta-analysis. Lancet. 2011;377:1331–40.View ArticlePubMedGoogle Scholar
- NHS Information centre. The Infant Feeding Survey 2010. Available at http://www.hscic.gov.uk [Accessed January 2013].
- Scollo MM, Winstanley MH. Tobacco in Australia: Facts and issues. 4th ed. Melbourne: Cancer Council Victoria; 2012. Available from http://www.TobaccoInAustralia.org.au.Google Scholar
- Ebert LM, Fahy K. Why do women continue to smoke in pregnancy? Women Birth. 2007;20:161–8.View ArticlePubMedGoogle Scholar
- Wood SL, Chen S, Ross S, Sauve R. The risk of unexplained antepartum stillbirth in second pregnancies following caesarean section in the first pregnancy. BJOG. 2008;115:726–31.View ArticlePubMedGoogle Scholar
- Smith GCS, Shah I, White IR, Pell JP, Crossley JA, Dobbie R. Maternal and biochemical predictors of antepartum stillbirth among nulliparous women in relation to gestational age of fetal death. BJOG. 2007;114:705–14.View ArticlePubMedGoogle Scholar
- Hogberg L, Cnattingius S. The influence of maternal smoking habits on the risk of subsequent stillbirth: Is there a causal relation? BJOG. 2007;114:699–704.View ArticlePubMedPubMed CentralGoogle Scholar
- United States Department of Health and Human Services (USDHHS). Health Consequences of Tobacco use among women. In: Women and smoking: a report of the surgeon general. Rockville MD: U.S: Department of Health and Human Services, Centre for Disease Control and Prevention, National Centre for Chronic Disease Prevention and Health Promotion, Office of Smoking and Health; 2001. p. 177–450.Google Scholar
- Cnattingius S, Haglund B, Meirik O. Cigarette smoking as risk factor for late fetal and early neonatal death. BMJ. 1988;297:258–61.View ArticlePubMedPubMed CentralGoogle Scholar
- Raymond EG, Cnattingius S, Kiely JL. Effects of maternal age, parity and smoking on the risk of stillbirth. BJOG. 1994;101:301–6.View ArticleGoogle Scholar
- Schramm WF. Smoking during pregnancy: Missouri longitudinal study. Paediatr Perinat Epidemiol. 1997;11:73–83.View ArticlePubMedGoogle Scholar
- Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting. JAMA. 2000;283:1–5.View ArticleGoogle Scholar
- Zegers-Hochschild F, Adamson GD, de Mouzon J, Ishihara O, Mansour R, Nygren K, et al. The International Committee for Monitoring Assisted Reproductive Technology (ICMART) and the World Health Organization (WHO) Revised Glossary on ART Terminology. Hum Reprod. 2009;24:2683–7.View ArticlePubMedGoogle Scholar
- Wells G, Shea B, O’Connell D, Peterson J, Welch V, Losos M, et al. Newcastle-Ottawa scale (NOS) for assessing the quality of non-randomised studies in meta-analysis. Available from: http://www.ohri.ca/programs/clinical_epidemiology/oxford.htm (accessed March 10, 2012).
- Higgins JPT, Green S (editors). Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [updated March 2011]. The Cochrane Collaboration 2011. Available from http://www.cochrane-handbook.org.
- Salihu HM, Sharma PP, Getahun D, Hedaytzadeh M, Peters S, Kirby RS, et al. Prenatal tobacco use and risk of stillbirth: a case-control and bidirectional case-crossover study. Nicotine Tob Res. 2008;10:159–66.View ArticlePubMedGoogle Scholar
- Ferraz EM, Gray RH. A case-control study of stillbirths in Northeast Brazil. Int J Gynaecol Obstet. 1991;34:3–19.View ArticleGoogle Scholar
- Walles B, Tyden T, Herbst A, Ljungblad U, Rydhstrom H. Maternal health-care program and makers for late fetal death. Acta Obstet Gynecol Scand. 1994;73:773–8.View ArticlePubMedGoogle Scholar
- Dodds L, King WD, Fell DB, Armson BA, Allen A, Nimrod C. Stillbirth risk factors according to timing of exposure. Ann Epidemiol. 2006;16:607–13.View ArticlePubMedGoogle Scholar
- Efkarpidis S, Alexopoulos E, Kean L, Liu D, Fay T. Case-control study of factors associated with intrauterine fetal deaths. MedGenMed. 2004;6:53.PubMedPubMed CentralGoogle Scholar
- Goy J, Dodds L, Rosenberg MW, King W. Health-risk behaviours: examining social disparities in the occurrence of stillbirth. Paediatr Perinat Epidemiol. 2008;22:314–20.View ArticlePubMedGoogle Scholar
- Ahlenius I, Thomassen P. The changing panorama of late fetal death in Sweden between 1984 and 1991. Acta Obstet Gynecol Scand. 1999;78:408–14.View ArticlePubMedGoogle Scholar
- Aliyu MH, Salihu HM, Wilson RE, Alio AP, Kirby RS. The risk of intrapartum stillbirth among smokers of advanced maternal age. Arch Gynecol Obstet. 2008;278:39–45.View ArticlePubMedGoogle Scholar
- Aliyu MH, Salihu HM, Wilson RE, Kirby RS. Prenatal smoking and risk of intrapartum stillbirth. Arch Environ Occup Health. 2007;62:87–92.View ArticlePubMedGoogle Scholar
- Gray R, Bonellie SR, Chalmers J, Greer I, Jarvis S, Kurinczuk JJ, et al. Contribution of smoking during pregnancy to inequalities in stillbirth and infant death in Scotland 1994-2003: retrospective population based study using hospital maternity records. BMJ. 2009;339:b3754.View ArticlePubMedPubMed CentralGoogle Scholar
- Miller EC, Cao HL, Wen SW, Yang Q, Lafleche J, Walker M. The risk of adverse pregnancy outcomes is increased in preeclamptic women who smoke compared with nonpreeclamptic women who do not smoke. Am J Obstet Gynecol. 2010;203:1–8.View ArticleGoogle Scholar
- Mohsin M, Bauman AE, Jalaludin B. The influence of antenatal and maternal factors on stillbirths and neonatal deaths in New South Wales. J Biosoc Sci. 2006;38:643–57.View ArticlePubMedGoogle Scholar
- Reddy UM, Laughon SK, Sun L, Troendle J, Willinger M, Zhang J. Prepregnancy risk factors for antepartum stillbirth in the United States. Obstet Gynecol. 2010;116:1119–26.View ArticlePubMedPubMed CentralGoogle Scholar
- Sutan R, Campbell D, Prescott GJ, Smith WCS. The risk factors for unexplained antepartum stillbirths in Scotland, 1994 to 2003. J Perinatol. 2010;30:311–8.View ArticlePubMedGoogle Scholar
- Tuthill DP, Stewart AH, Coles EC, Andrews J, Cartidge PH. Maternal cigarette smoking and pregnancy outcome. Paediatr Perinat Epidemiol. 1999;13:245–53.View ArticlePubMedGoogle Scholar
- Wisborg K, Kesmodel U, Henriksen TB, Olsen SF, Secher NJ. Exposure to tobacco smoke in utero and the risk of stillbirth and death in the first year of life. Am J Epidemiol. 2001;154:322–7.View ArticlePubMedGoogle Scholar
- Winbo I, Serenius F, Dahlquist G, Kallen B. Maternal risk factors for cause-specific stillbirth and neonatal death. Acta Obstet Gynecol Scand. 2001;80:235–44.View ArticlePubMedGoogle Scholar
- Mishra V, Retherford RD, Smith KR. Cooking smoke and tobacco smoke as risk factors for stillbirth. Int J Environ Health Res. 2005;15:397–410.View ArticlePubMedGoogle Scholar
- Robson S, Cameron CA, Roberts CL. Birth outcomes for teenage women in New South Wales, 1998-2003. Aust N Z J Obstet Gynaecol. 2006;46:305–10.View ArticlePubMedGoogle Scholar
- Kleinman JC, Pierre MB, Madans Jr JH, Land GH, Schramm WF. The effects of maternal smoking on fetal and infant mortality. Am J Epidemiol. 1988;127:274–82.PubMedGoogle Scholar
- Raatikainen K, Huurinainen P, Heinonen S. Smoking in early gestation or through pregnancy: a decision crucial to pregnancy outcome. Prev Med. 2007;44:59–63.View ArticlePubMedGoogle Scholar
- Huang DY, Usher RH, Kramer MS, Yang H, Morin L, Fretts RC. Determinants of unexplained antepartum fetal deaths. Obstet Gynecol. 2000;95:215–21.PubMedGoogle Scholar
- Salihu HM, Shumpert MN, Aliyu MH, Alexander MR, Kirby RS, Alexander GR. Stillbirths and infant deaths associated with maternal smoking among mothers aged not greater than or equal to 40 years: a population study. Am J Perinatol. 2004;21:121–9.View ArticlePubMedGoogle Scholar
- Moga M, Preda GH. Smoking and pregnancy. J Environ Prot Ecol. 2008;9:566–73.Google Scholar
- Bukowski R, Carpenter M, Conway D, Coustan D, Dudley DJ, Goldenberg RL, et al. Association between stillbirth and risk factors known at pregnancy confirmation. JAMA. 2011;306:2469–79.View ArticleGoogle Scholar
- Kunzel W, Misselwitz B. Unexpected fetal death during pregnancy – a problem of unrecognised fetal disorders during antenatal care? Eur J Obstet Gynecol Reprod Biol. 2003;110:S86–92.View ArticlePubMedGoogle Scholar
- Xu B, Jarvelin MR, Rantakallio P. Maternal smoking in pregnancy and sex differences in perinatal death between boys and girls. Soc Biol. 1998;45:273–7.PubMedGoogle Scholar
- Kallen K. The impact of maternal smoking during pregnancy on delivery outcome. Eur J Public Health. 2001;11:329–33.View ArticlePubMedGoogle Scholar
- Verkerk PH, Buitendijk SE, Verloove-Vanhorick SP. Differential misclassification of alcohol and cigarette consumption by pregnancy outcome. Int J Epidemiol. 1994;23:1218–25.View ArticlePubMedGoogle Scholar
- Maleckiene L, Nadisauskiene R, Bergstrom S. Scio-economic demographic and obstetric risk factors for late death of unknown etiology in Luthuania: a case—referent study. Acta Obstet Gynecol Scand. 2001;80:321–5.PubMedGoogle Scholar
- Beyerleinl A, Von Kriesl R, Hummel M, Lack N, Schiessl B, Giani G. Improvement in pregnancy-related outcomes in the offspring of diabetic mothers in Bavaria, Germany, during 1987–2007. Diabet Med. 2010;27:1379–84.View ArticleGoogle Scholar
- Froen JF, Arnestad M, Vege A, Irgens, Rognum T, Saugstad O. Comparative epidemiology of sudden infant death syndrome and sudden intrauterine unexplained death. Arch Dis Child. 2002;87:118–21.View ArticleGoogle Scholar
- Little RE, Weinberg CR. Risk-factors for antepartum and intrapartum still-birth. Am J Epidemiol. 1993;137:1177–89.PubMedGoogle Scholar
- McCowan LM, George-Haddad M, Stacey T, Thompson JM. Fetal growth restriction and other risk factors for stillbirth in a New Zealand setting. Aust N Z J Obstet Gynaecol. 2007;47:450–6.View ArticlePubMedGoogle Scholar
- Aliyu MH, Salihu HM, Alio AP, Wilson RE, Charkrabarty S, Clayton HB. Prenatal smoking among adolescents and risk of fetal demise before and during labor. J Pediatr Adolesc Gynecol. 2010;23:129–35.View ArticlePubMedGoogle Scholar
- Bai J, Wong FW, Gyneshwar R, Stewart HC. Profile of maternal smokers and their pregnancy outcomes in south western Sydney. J Obstet Gynaecol Res. 2000;26:127–32.View ArticlePubMedGoogle Scholar
- Flenady V. Causes and risk factors of stillbirth in Australia. Perinatal Society of Austarlia and New Zealand Congress 2008. Gold Coast. Paediatr Child Health. 2008:2-10.
- Marrow RJ, Ritchie JW, Bull SB. Maternal cigarette smoking: the effects on umbilical and uterine blood flow velocity. Am J Obstet Gynecol. 1988;159:1069–71.View ArticleGoogle Scholar
- Lambers DS, Clark KE. The maternal and fetal physiologic effects of nicotine. Semin Perinatol. 1996;20:115–26.View ArticlePubMedGoogle Scholar
- Ylikorkala O, Viinikka L, Lehtovirta P. Effect of nicotine on fetal prostacyclin and thromboxane in humans. Obstet Gynecol. 1985;66:102–5.PubMedGoogle Scholar
- Bruner JP, Forouzan I. Smoking and Bucally administered nicotine. Acute effect on uterine and umbilical artery Doppler flow velocity wave forms. J Reprod Med. 1991;36:435–40.PubMedGoogle Scholar
- Salihu HM, Aliyu HM, Kirby RS. In utero nicotine exposure and fetal growth inhibition among twins. Am J Perinatol. 2005;22:421–7.View ArticlePubMedGoogle Scholar
- Hammoud AO, Bujold E, Sorokin Y, Schild C, Krapp M, Baumann P. Smoking in pregnancy revisited: findings from a large population-based study. Am J Obstet Gynecol. 2005;192:1856–63.View ArticlePubMedGoogle Scholar
- Hajek P, Stead LF, West R, Jarvis M, Hartmann-Boyce J, Lancaster T. Relapse prevention interventions for smoking cessation. Cochrane Database Syst Rev 2013, Issue 8. Art. No.: CD003999. doi:10.1002/14651858.CD003999.pub4.
- Coleman T, Chamberlain C, Davey MA, Cooper SE, Leonardi-Bee J. Pharmacological interventions for promoting smoking cessation during pregnancy. Cochrane Database Syst Rev 2012, Issue 9. Art. No.: CD010078. doi:10.1002/14651858.CD010078
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.