Chronic respiratory diseases (CRDs) accounted for the leading contributor to global mortality in the past decades, seriously endangering human health worldwide. Some of the most common are chronic obstructive pulmonary disease (COPD), asthma, occupational lung diseases, and pulmonary hypertension. Moreover, epidemiological evidence indicated that the incidences of lung cancer (LC) and idiopathic pulmonary fibrosis (IPF) have risen in the past decades, reducing patients’ quality of life.

Smoking is a well-established risk factor for these aforementioned respiratory diseases.

Despite various smoking cessation measures, about 12% of women smoke during pregnancy1, resulting in their fetus being exposed to smoke, thus leading to various long-term health problems in the offspring2,3. Growing evidence indicates that smoking during pregnancy disturbs fetal lung development4,5, causing a negative effect on the pulmonary health of the offspring in childhood with an increased risk for wheezing, hospitalization for respiratory infections, and childhood asthma6-9. Whether maternal smoking during pregnancy has a continued impact on the offspring’s lung health during adulthood, remains uncertain. A few previous studies have indicated an association between maternal smoking and adult lung function10,11, besides, intrauterine exposure to maternal tobacco smoking was related to more adult respiratory symptoms, but there was no strong evidence that maternal smoking influences adult lung health after multivariable adjustment as these were performed using observational studies, which are vulnerable to confounding bias12. A recent study stemming from the UK Biobank cohort reported that maternal smoking might bring about an excess reduction in forced expiratory volume in one second (FEV1)/forced vital capacity (FVC), and risk of COPD, but that the results are heterogeneous due to the individual smoking, and the findings showed that there was no strong evidence that maternal smoking influenced adult lung health among never smokers13. Thus, whether maternal smoking around birth represents a strong determinant of CRDs in offspring remains uncertain because the available evidence is scarce.

Undoubtedly, well-designed randomized controlled trials (RCTs) are the gold standard for deducing causality, but their use is frequently limited because of practical and ethical considerations. Mendelian randomization (MR) is a desirable approach that can overcome these challenges by nature, as genetic variants are assorted randomly at conception and fixed at birth; they can be applicable to assess the relationships between maternal smoking and CRDs in their offspring by exploiting genetic variants as instruments for the exposure. Based on data from the largest available genome-wide association study (GWAS), we performed a comprehensive MR study to ascertain the relationships between maternal smoking around birth and a wide range of possible CRDs in their offspring during adulthood.


Study design

This is a two-sample MR study design based on summary-level data. An MR analysis depends on the assumptions (Figure 1) that: the genetic variants are strongly associated with the exposure (the relevance assumption); are not associated with confounders of the exposure-outcome relationship (the independence assumption); and have an effect on the outcome through the exposure only and not through any other causal pathway (the exclusion restriction assumption)14.

Figure 1

Overall design of the two-sample Mendelian randomization analysis in this study

Data sources and instrumental variable selection

Exposure events were maternal smoking in the time period around birth (as defined in each database), and GWAS data for exposure were obtained from GWAS Catalog: GCST90041844, covering 494132 participants. Outcome events were the CRDs in the offspring during adulthood, including respiratory insufficiency, COPD-related respiratory insufficiency, emphysema, COPD, COPD hospital admissions, early onset COPD, later onset COPD, asthma, idiopathic pulmonary fibrosis (IPF), lung cancer (LC), small cell lung carcinoma (SCLC), and lung squamous cell carcinoma (LUSC). The GWAS data sources for outcomes are described in detail in Table 1.

Table 1

The detailed information of GWAS data in outcomes

OutcomesGWAS IDSample sizeCasesControlsp (with SNPs associated with exposure)
Diseases of the respiratory systemfinn-b-J10_RESPIRATORY2187921072611115315.0×10-7
Respiratory insufficiencyfinn-b-RESPIRATORYINSUFF1376458781367675.0×10-7
COPD-related respiratory insufficiencyfinn-b-COPD_INSUFFICIENCY18775410311867235.0×10-7
COPD, hospital admissionsfinn-b-COPD_HOSPITAL21879265002122925.0×10-7
Early onset COPDfinn-b-COPD_EARLY21570535082121975.0×10-6
Later onset COPDfinn-b-COPD_LATER21528430872121975.0×10-7
Idiopathic pulmonary fibrosisebi-a-GCST9001812043723513694358665.0×10-7
Lung cancerieu-a-9662720911348158615.0×10-6
Small cell lung carcinomaieu-a-988233712791205805.0×10-6
Squamous cell lung cancerieu-a-989624677704547635.0×10-6

[i] GWAS: genome-wide association study. COPD: chronic obstructive pulmonary disease. SNPs: single-nucleotide polymorphisms.

As at least 10 instrumental variables (IVs) are required for a MR study15, we selected instrumental variables of p<5×10-7 or p<5×10-6 for MR analysis. The parameters used to eliminate linkage disequilibrium among variables were kb=10000 and r2=0.01. The F statistic is used to estimate sample overlap effects and weak instrumental bias, and an F>10 is sufficient to limit bias from weak instrumental variables16.

As the smoking status of offspring may affect their risk of developing respiratory diseases, we needed to take this into account in any association, and hence, as the genes rs10226228 were associated with nicotine-dependent smoking of cigarettes per day, and the rs10883802, rs11783093, rs1563245, rs414763, rs414763, rs6011779, rs62477310, and rs7938812 were all related to current tobacco smoking, the rs12042107 and rs876793 were related to past tobacco smoking, while the rs2183947 was related to pack-years of adult smoking as proportion of life span exposed to smoking. Therefore, these single-nucleotide polymorphisms (SNPs) were regarded as an unreliable instrumental variable for maternal smoking around birth (Table 2). Besides, the details of the per allele associations with exposure plotted against per allele associations with outcome are provided in the Supplementary file.

Table 2

Detailed information on confounding SNPs that were removed during our GWAS analysis

OutcomesGWAS IDRemoved SNPs related to smoking by offspring
Diseases of the respiratory systemfinn-b-J10_RESPIRATORYrs10226228, rs10883802, rs11783093, rs12042107, rs2183947, rs576982, rs6011779, rs62477310, rs709400
Respiratory insufficiencyfinn-b-RESPIRATORYINSUFFrs10226228, rs10883802, rs11783093, rs12042107, rs2183947, rs576982, rs6011779, rs62477310, rs709400
COPD-related respiratory insufficiencyfinn-b-COPD_INSUFFICIENCYrs10226228, rs10883802, rs11783093, rs12042107, rs2183947, rs6011779, rs62477310, rs709400
Emphysemafinn-b-J10_EMPHYSEMArs10226228, rs10883802, rs11783093, rs12042107, rs2183947, rs576982, rs6011779, rs62477310, rs709400
COPDfinn-b-J10_COPDrs10226228, rs10883802, rs12042107, rs2183947, rs62477310, rs709400
COPD, hospital admissionsfinn-b-COPD_HOSPITALrs10226228, rs10883802, rs12042107, rs2183947, rs62477310, rs709400
Early onset COPDfinn-b-COPD_EARLYrs10226228, rs10883802, rs11783093, rs12042107, rs1563245, rs2183947, rs414763, rs6011779, rs62477310, rs7938812, rs876793
Later onset COPDfinn-b-COPD_LATERrs10226228, rs10883802, rs11783093, rs12042107, rs2183947, rs576982, rs6011779, rs62477310, rs709400
Asthmaebi-a-GCST90014325rs10226228, rs10883802, rs11783093, rs12042107, rs218394, rs6011779, rs62477310, rs709400
Idiopathic pulmonary fibrosisebi-a-GCST90018120rs10226228, rs10883802, rs11783093, rs12042107, rs2183947, rs576982, rs6011779, rs62477310, rs709400
Lung cancerieu-a-966rs10226228, rs10883802, rs2624839, rs414763, rs62477310, rs709400, rs7938812, rs876793
Small cell lung carcinomaieu-a-988rs10226228, rs10883802, rs2624839, rs414763, rs62477310
Squamous cell lung cancerieu-a-989rs10883802, rs26248397, rs414763, rs62477310, rs876793

[i] GWAS: genome-wide association study. COPD: chronic obstructive pulmonary disease. SNPs: single-nucleotide polymorphisms.

Statistical analysis

We used a two-sample MR analysis to estimate the direct effect of maternal smoking around birth on the risk of offspring CRDs during adulthood . All MR analysis, except for asthma, used fixed-effects models with the inverse-variance-weighted (IVW) model, MR-Egger regression, weighted-median estimator (WME), and weighted mode (VM), while the MR analysis for asthma outcome was conducted using the random effects models. Among these methods, the IVW model is used as the primary method of MR analysis to assess the causal effects, which summarizes effect sizes from multiple independent studies by calculating the weighted mean of the effect sizes using the inverse variance of the individual studies as weights. However, in the presence of horizontal pleiotropy, IVW may not be consistent and may result in the deviation for causal inference. The MR-Egger regression can be used to assess the horizontal pleiotropy of selected IVs, is applied under a weaker assumption that the direct or pleiotropic effects of the genetic variants on the outcome are independent of the genetic associations with the exposure, the so-called ‘instrument strength independent of direct effect’ (InSIDE) assumption17. The WME method offers a consistent estimate of causal effects by utilizing the weighted median of Wald under the condition that at least 50% of variants adhere to the criteria of a valid IV for the exclusion restrictions. Utilizing the estimation of individual proportions, the WM method categorizes SNPs based on their similarity and computes the counter-variance weighted count of SNPs in each group18.

After removing twelve SNPs (rs10226228, rs10883802, rs11783093, rs1563245, rs414763, rs414763, rs6011779, rs62477310, rs7938812, rs12042107, rs876793, and rs2183947), a leave-one-out sensitivity analysis was performed to examine the effect of individual SNPs on causal estimates. The examination of heterogeneity involved the utilization of Cochran’s Q statistic and the related p-values to ascertain the consistency of causal relationships across all SNPs. The horizontal pleiotropy was calculated based on the MR-Egger intercept and p-values. Besides, MR-Pleiotropy Residual Sum and Outlier (MR-PRESSO) analysis, employed to assess the pleiotropy effects of outlier SNPs and correct abnormal findings attributable to such outliers, involves regressing SNP outcomes on SNP exposure and utilizing the square of residuals to identify outliers.

The sensitivity analyses were conducted by three tests: 1) the leave-one-out sensitivity test was used to determine the stability of individual SNPs in this MR study by excluding IVs in sequence; 2) the robustness of various IVs was tested by Cochrane’s Q-statistic, in which p>0.05 represents non-significant heterogeneity; and 3) the horizontal pleiotropy was calculated based on the MR-Egger intercept and p>0.05 indicates no horizontal pleiotropy.

The results are presented as odds ratios (ORs) with 95% confidence intervals (CIs) for convenience of interpretation. All analyses were performed using R software, version 4.2.0.


Results of the Mendelian randomization study testing causal association

The results of MR analysis showed that before removing SNPs related to smoking by the offspring, maternal smoking around birth increased the appearance of respiratory diseases in the offspring by 17% (OR=1.17; 95% CI: 1.05–1.30), increased the risk of respiratory dysfunction by 2.29-fold (OR=3.29; 95% CI: 1.72–6.30), and increased the risk of respiratory dysfunction related to COPD by 3.84-fold (OR=4.84; 95% CI: 2.15–10.91). The risk of developing emphysema increased by 1.85-fold (OR=2.85; 95% CI: 1.12–7.24), the risk of developing COPD increased by 81.6% (OR=1.82; 95% CI: 1.36–2.43), the risk of COPD and hospital admissions increased by 80.3% (OR=1.80; 95% CI: 1.32–2.47), and the risk of early onset COPD increased by 54% (OR=1.54; 95% CI: 1.20–1.972). The risk of developing late onset COPD increased by 66% (OR=1.66; 95% CI: 1.66–4.269), and the risk of developing asthma increased by 24.4% (OR=1.244; 95% CI: 1.11–1.395). The risk of developing IPF increased by 0.2% (OR=1.002; 95% CI: 1.0–1.003), the risk of developing lung cancer increased by 20.4% (OR=1.204; 95% CI: 0.9–1.47), the risk of developing small cell lung cancer increased by 20% (OR=1.20; 95% CI: 0.99–1.47), and the risk of squamous cell lung cancer increased by 24.3% (OR=1.24; 95% CI: 1.02–1.53).

After removing SNPs associated with smoking by the offspring, maternal smoking still led to a 14% increase in the risk of respiratory diseases in the offspring (OR=1.14; 95% CI: 1.01–1.28), a 1.41-fold increase in the risk of respiratory insufficiency (OR=2.41; 95% CI: 1.04–5.60), and a 14% increase in the risk of respiratory insufficiency related to COPD (OR=1.14; 95% CI: 1.01–1.28). The risk of COPD increased by 74.2% (OR=1.74; 95% CI: 1.21–2.52), the risk of COPD and hospital admissions increased by 65.9% (OR=1.66; 95% CI: 1.12–2.46), the risk of early onset COPD increased by 29.6% (OR=1.30; 95% CI: 1.01–1.67), the risk of late onset COPD increased by 94.4% (OR=1.95; 95% CI: 1.18–3.21), and the risk of asthma increased by 33.6% (OR=1.336; 95% CI: 1.161–1.538). However, after removing the SNP of smoking by the offspring, the causal relationship between maternal smoking and IPF (OR=1.00; 95% CI: 1.00–1.00), the causal relationship between maternal smoking and lung cancer (OR=1.20; 95% CI: 0.96–1.50), and the causal relationship between maternal smoking and small cell lung cancer (OR=1.11; 95% CI: 0.77–1.60) were no longer statistically significant, while the causal relationship with squamous cell lung cancer (OR=1.23; 95% CI: 0.99–1.52) still existed (Table 3, Figure 2).

Table 3

The relationship between maternal smoking and respiratory diseases in the offspring

OutcomeMR analysis before removing SNPs related to smoking by offspringMR analysis after removing SNPs associated with smoking by offspring
SNPsMethodsOR (95% CI)pSNPsMethodsOR (95% CI)p
Diseases of the respiratory system25Inverse variance weighted1.171 (1.052–1.303)0.00416Inverse variance weighted1.14 (1.013–1.284)0.030
MR Egger1.303 (0.806–2.106)0.292MR Egger1.119 (0.654–1.915)0.687
Weighted median1.253 (1.087–1.444)0.002Weighted median1.21 (1.033–1.417)0.018
Weighted mode1.336 (1.007–1.773)0.056Weighted mode1.319 (0.918–1.894)0.155
Respiratory insufficiency26Inverse variance weighted3.292 (1.72–6.298)<0.00117Inverse variance weighted2.413 (1.039–5.603)0.040
MR Egger68.06 (3.724–1243)0.009MR Egger4.371 (0.088–216.6)0.470
Weighted median2.22 (0.857–5.746)0.100Weighted median1.585 (0.532–4.718)0.408
Weighted mode1.509 (0.201–11.34)0.693Weighted mode1.16 (0.176–7.658)0.880
COPD-related respiratory insufficiency26Inverse variance weighted4.84 (2.148–10.906)<0.00118Inverse variance weighted3.119 (1.323–7.352)0.009
MR Egger58.99 (1.688–2062)0.034MR Egger4.862 (0.088–268.0)0.451
Weighted median3.233 (1.219–8.58)0.018Weighted median2.187 (0.736–6.501)0.159
Weighted mode2.447 (0.377–15.89)0.357Weighted mode2.168 (0.396–11.86)0.385
Emphysema26Inverse variance weighted2.849 (1.121–7.242)0.02817Inverse variance weighted2.174 (0.801–5.904)0.127
MR Egger380.8 (8.673–16719)0.005MR Egger63.49 (0.696–5789)0.092
Weighted median2.758 (0.905–8.411)0.074Weighted median1.863 (0.486–7.141)0.364
Weighted mode2.246 (0.269–18.77)0.462Weighted mode1.465 (0.158–13.549)0.741
COPD23Inverse variance weighted1.816 (1.357–2.43)<0.00117Inverse variance weighted1.742 (1.205–2.519)0.003
MR Egger2.115 (0.514–8.706)0.311MR Egger1.269 (0.217–7.403)0.795
Weighted median1.489 (0.986–2.249)0.058Weighted median1.198 (0.737–1.948)0.466
Weighted mode1.08 (0.469–2.489)0.858Weighted mode1.009 (0.466–2.184)0.982
COPD, hospital admissions23Inverse variance weighted1.803 (1.318–2.467)<0.00117Inverse variance weighted1.659 (1.119–2.461)0.012
MR Egger1.965 (0.428–9.03)0.395MR Egger1.216 (0.184–8.039)0.842
Weighted median1.512 (1.008–2.267)0.046Weighted median1.316 (0.818–2.115)0.258
Weighted mode1.4 (0.624–3.142)0.423Weighted mode0.992 (0.423–2.325)0.985
Early onset COPD66Inverse variance weighted1.54 (1.204–1.972)0.00155Inverse variance weighted1.296 (1.009–1.665)0.043
MR Egger0.996 (0.482–2.057)0.991MR Egger0.727 (0.365–1.447)0.368
Weighted median1.428 (1.02–1.998)0.038Weighted median1.295 (0.891–1.88)0.175
Weighted mode1.498 (0.693–3.239)0.308Weighted mode1.333 (0.677–2.623)0.409
Later onset COPD26Inverse variance weighted2.663 (1.661–4.269)<0.00117Inverse variance weighted1.944 (1.178–3.207)0.009
MR Egger13.821 (1.779–107.374)0.019MR Egger1.658 (0.15–18.266)0.686
Weighted median2.347 (1.314–4.192)0.004Weighted median1.958 (0.996–3.849)0.051
Weighted mode1.828 (0.421–7.934)0.428Weighted mode2.505 (0.642–9.769)0.205
Asthma26Inverse variance weighted1.244 (1.11–1.395)<0.00118Inverse variance weighted1.336 (1.161–1.538)<0.001
MR Egger0.996 (0.63–1.576)0.987MR Egger0.955 (0.553–1.648)0.870
Weighted median1.176 (1.025–1.349)0.020Weighted median1.187 (1.009–1.397)0.039
Weighted mode1.101 (0.815–1.486)0.536Weighted mode1.139 (0.838–1.55)0.417
Idiopathic pulmonary fibrosis27Inverse variance weighted1.002 (1–1.003)0.01518Inverse variance weighted1.001 (0.999–1.003)0.224
MR Egger1.003 (0.997–1.009)0.320MR Egger1.002 (0.995–1.01)0.525
Weighted median1.002 (1–1.004)0.132Weighted median1.001 (0.999–1.004)0.316
Weighted mode1.002 (0.998–1.006)0.271Weighted mode1.002 (0.997–1.006)0.488
Lung cancer45Inverse variance weighted1.204 (0.985–1.47)0.06937Inverse variance weighted1.203 (0.964–1.501)0.103
MR Egger1.168 (0.595–2.296)0.654MR Egger1.288 (0.641–2.591)0.482
Weighted median1.276 (0.98–1.662)0.071Weighted median1.276 (0.948–1.718)0.108
Weighted mode1.339 (0.72–2.489)0.361Weighted mode1.317 (0.671–2.586)0.429
Small cell lung carcinoma41Inverse variance weighted1.179 (0.838–1.657)0.34436Inverse variance weighted1.11 (0.77–1.601)0.577
MR Egger1.758 (0.391–7.917)0.467MR Egger1.599 (0.334–7.644)0.560
Weighted median1.205 (0.776–1.872)0.407Weighted median1.151 (0.718–1.844)0.559
Weighted mode1.253 (0.494–3.177)0.637Weighted mode1.145 (0.443–2.956)0.781
Squamous cell lung cancer48Inverse variance weighted1.243 (1.015–1.523)43Inverse variance weighted1.229 (0.992–1.523)0.059
MR Egger1.101 (0.557–2.175)MR Egger1.144 (0.57–2.295)0.707
Weighted median1.206 (0.904–1.609)Weighted median1.205 (0.891–1.631)0.227
Weighted mode1.214 (0.624–2.362)Weighted mode1.24 (0.652–2.359)0.516

[i] COPD: chronic obstructive pulmonary disease. SNPs: single-nucleotide polymorphisms.

Figure 2

Forest plot of the relationship between maternal smoking before and after birth and chronic respiratory diseases in offspring

Sensitivity analysis

Our sensitivity analyses included heterogeneity analysis and tests for horizontal pleiotropy (Table 4). After removing confounders associated with offspring smoking, there was no horizontal pleiotropy (p>0.05) in all MR results. Besides, the findings of heterogeneity analysis indicated the absence of statistically significant heterogeneity (p>0.05) in all MR results except for the MR analyses with asthma (Q=30.913, p=0.020) as the outcome event. Moreover, there were no outliers in all MR-PRESSO results.

Table 4

Heterogeneity and horizontal pleiotropy in the present Medelian Randomization study

OutcomesHeterogeneityHorizontal pleiotropy
QpIntercept in MR-Egger regressionp (MR-Egger intercept analysis)
Diseases of the respiratory system12.1080.6710.0010.946
Respiratory insufficiency10.2750.852-0.0210.764
COPD-related respiratory insufficiency23.1530.144-0.0160.827
COPD, hospital admissions22.7590.1200.0110.746
Early onset COPD59.9610.2680.0200.084
Later onset COPD18.2280.3110.0060.896
Idiopathic pulmonary fibrosis11.7270.8160.0000.728
Lung cancer17.5990.996-0.0020.840
Small cell lung carcinoma10.3541.000-0.0120.641
Squamous cell lung cancer27.8130.9550.0020.833

[i] COPD: chronic obstructive pulmonary disease. SNPs: single-nucleotide polymorphisms.


This study utilized GWAS data to investigate whether the exposure to maternal smoking around birth is associated with CRDs of the offspring during adulthood, as proposed by epidemiologic studies. The results found were: 1) maternal smoking around birth may be defined as a dangerous exposure for lung development in their offspring, inducing respiratory insufficiency, emphysema, and COPD-related respiratory insufficiency; 2) the intrauterine exposure to tobacco smoke may increase the risk of diseases of the respiratory system, especially the chronic airway inflammatory diseases including COPD and asthma; and 3) smoking by pregnant women may result in their offspring being more prone to suffer IPF, and increase the incidence of lung cancer in the offspring, despite that this was not statistically significant.

Tobacco smoke contains thousands of chemical compounds. Nicotine, as one of the leading chemical components in smoke, can enter fetal circulation through the placental barrier and spread throughout the body, which can lead to the development of diseases19. In this process, nicotine can interact with nicotinic acetylcholine receptors (nAChRs) in the fetal lung, leading to change in the structure and function of the lung of the offspring2,20,21. Smoking in pregnant women has a negative effect on the pulmonary health of their offspring4. A prospective study found that FEV1 and forced expiratory flow (FEF) between 25 and 75% of FVC of offspring who had been exposed to maternal smoking in utero, and continued to decrease in early adulthood8. Meta-analyses have demonstrated a significant association between exposure to maternal smoking during pregnancy and the risk of developing bronchopulmonary dysplasia (BPD)22, which might increase the risk of COPD23. An animal study reported that maternal exposure to cigarette smoke increased receptors for advanced glycation end-products (RAGE) and in its signaling elements associated with increased oxidative stress and inflammatory cytokines in the offspring’s lungs, inducing the proliferation of lung cells and changing the structure and function of the lung of the offspring, resulting in poor lung function and causing respiratory insufficiency4. The limitation of observational studies is that they are susceptible to confounding by unmeasured differences between the exposed and unexposed populations, and our findings provide additional evidence for a potential effect of maternal smoking around birth on their offspring’ poor lung function (including respiratory insufficiency and COPD-related respiratory insufficiency) and pulmonary structural change (such as emphysema).

Cigarette smoking is a key environmental risk factor for chronic airway inflammatory diseases such as asthma and COPD. Previous studies illustrated that maternal smoking poses a risk for their fetus, by altering lung growth and development in utero, and possibly priming the immune system by inducing specific epigenetic changes, increasing the morbidity of bronchopulmonary dysplasia (BPD) and leading to COPD in the offspring24-26. Our study used SNPs as instrumental variables to elucidate the role of maternal smoking around the time of delivery as a cause of elevated risk of COPD in their offspring. Recently, an MR study has reported that maternal smoking around birth increases the risk of childhood asthma based on childhood asthma of 1993 cases from ukb-d-ASTHMA_CHILD27. In contrast, our two-sample MR analysis had a much larger outcome cohort (ebi-a-GCST90014325, including 56167 cases and 352255 controls) and strengthened the evidence for an effect of maternal smoking around birth on their offspring’s asthma during adulthood, providing more convincing evidence by removing SNPs associated with smoking by the offspring in the MR analysis.

Cohort studies have evaluated the longitudinal association of smoking with IPF28,29, and an MR study investigated the causal association between smoking and the risk of IPF30. Cohort studies have found that smoking could increase the risk of IPF in a dose-response manner, and a two-sample MR study30,31 confirmed that a potential causal effect of smoking on IPF, while a one-sample MR study reported that smoking is unlikely to be a causal factor for IPF31. Our study found that the offspring might be prone to suffer IPF if they had been exposed to smoking in utero, but might be more vulnerable to exposure to tobacco smoking after birth. This study finding strengthens the evidence for an effect of smoking on IPF in people, acquired by exposure30.

Smoking has been widely recognized as a risk factor for numerous types of cancer, and studies have confirmed the causal effect of smoking on the risk of various tumors, including lung cancer32,33. A clinical study established smoking cessation could decrease the risk of death from lung cancer34. In our study, the results indicate that maternal smoking around birth might promote the incidence of lung cancer but could not be defined as a factor for lung cancer owing to the MR analysis after removing SNPs associated with smoking in offspring, providing additional evidence for a causal effect of exposure to smoking after birth on lung cancer.


Although a two-sample MR study is a powerful approach to investigate the relationship between exposures and outcomes, we should be careful with our findings because of several limitations. First, the participants in our study were from the European Pedigree GWAS database. Hence, definitions of exposure to cigarette smoking and its exact timing are defined as categorized in this database. The results, hence, need to be verified in other populations. Second, there may be developmental compensations during offspring growth, which may influence the effects due to instrumental variables. Third, the potential confounding factors, such as the exact timing of maternal smoking around birth and the effects of secondhand smoke on chronic diseases, including CRDs, have not been investigated in this study. Thus, passive smoking may introduce variability in the MR analysis and should be noted to elucidate the effect of maternal smoking around birth on the offspring’s adult lung health and CRDs. Fourth, horizontal pleiotropy is a significant concern for the reliability of MR results. Nevertheless, the MR-Egger regression test showed no clear directional pleiotropy, and the likelihood of this bias is reduced because consistent estimates were observed across multiple MR methods, which have different assumptions.


Our study compressively investigated the effect of maternal smoking around birth on the offspring’s adult lung health and CRDs, and the results indicated that smoking during pregnancy may lead to offspring respiratory insufficiency and increase the incidence of chronic airway inflammatory diseases (e.g. asthma and COPD), during adulthood. Thus, it is critical to enhance policies for smoking cessation during pregnancy.