Abstract
Objective
This study aimed to investigate the association between smoking behaviours during early pregnancy and the risk and severity of gestational diabetes mellitus (GDM), with a particular focus on smoking status, smoking intensity and secondhand smoke exposure.
Design
Secondary analysis of prospectively collected cohort data.
Setting
Multi-centre study conducted in South Korea (Korean Pregnancy Outcome Study) between March 2013 and January 2017.
Participants
From 4537 pregnant women initially enrolled, 3457 singleton pregnancies were included after excluding cases with transfer, loss to follow-up, twin pregnancies, miscarriages and pre-existing diabetes mellitus. All participants were women of Korean ethnicity.
Secondary outcome measures
Primary outcome was GDM and its subtypes (A1GDM: diet-controlled; A2GDM: insulin-requiring). Secondary outcomes were associations with active smoking (before pregnancy and during early pregnancy), smoking intensity dose–response relationships (pack-years) and secondhand smoke exposure among never-smokers.
Results
Among 3457 participants, 231 women (6.7%) were diagnosed with GDM (198 A1GDM, 33 A2GDM). Active smoking before pregnancy (adjusted OR (aOR) 3.98, 95% CI 1.58 to 9.30) and during early pregnancy (aOR 9.90, 95% CI 2.97 to 29.45) were significantly associated with A2GDM, while no significant association was observed with A1GDM. A clear dose-response relationship was observed, with smoking intensity >4 pack-years markedly increasing A2GDM risk (aOR 20.68, 95% CI 6.75 to 59.39). Detailed pack-year analysis showed 4–6 pack-years (aOR 20.57, 95% CI 5.80 to 65.46) and >6 pack-years (aOR 25.98, 95% CI 3.21 to 146.45). Among never-smokers, secondhand smoke exposure showed a borderline association with overall GDM risk (aOR 1.33, 95% CI 0.98 to 1.81).
Conclusions
Maternal active smoking before and during early pregnancy, as well as higher smoking intensity, was associated with an increased risk of pharmacologically treated GDM (A2GDM). Although secondhand smoke exposure did not reach statistical significance, the trend suggested a potential association with GDM risk among never-smokers. These findings provide important evidence for public health strategies for prenatal care, as smoking cessation and environmental smoke avoidance during prenatal and early antenatal care in women reduce the risk of gestational diabetes.
Keywords: Risk Factors, Diabetes in pregnancy, Smoking Reduction
STRENGTHS AND LIMITATIONS OF THIS STUDY.
Large prospective cohort with standardised data collection protocols, representative of the Korean reproductive-age population.
Firth’s penalised logistic regression was employed to address small-sample bias in severity-based analyses.
Smoking assessment relied on self-reported data, potentially subject to recall and social desirability bias.
Despite using Firth’s method, the small insulin-requiring gestational diabetes mellitus sample size (n=33) still limited statistical power.
Data collected between 2013 and 2017, though smoking prevalence has remained stable in this population according to national statistics.
Introduction
Gestational diabetes mellitus (GDM) is characterised by glucose intolerance during pregnancy.1 It is divided into two types based on severity: A1GDM, which can be managed with dietary control alone, and A2GDM, which requires pharmacological intervention for blood glucose regulation.2 The prevalence of GDM ranges from 1% to 28% worldwide,3 4 with a notable increase in affected pregnant women observed in recent international studies.5,7 GDM is associated with various adverse outcomes, including macrosomia, preterm birth, neonatal hypoglycaemia, stillbirth and a higher incidence of caesarean section. Furthermore, it heightens the risk of type 2 diabetes in mothers as well as predisposes offspring to obesity and glucose intolerance in childhood and adulthood.8 Particularly, the risk of some of these complications is higher in cases of A2GDM than A1GDM.9 These findings emphasise the critical importance of implementing intervention strategies during pregnancy.
Several risk factors for GDM have been identified, including maternal obesity, advanced maternal age, physical inactivity, multiparity, a family history of type 2 diabetes, a history of GDM or macrosomic delivery and polycystic ovary syndrome.10 Among these, smoking has been demonstrated to adversely affect both mother and fetus, including increasing the risk of GDM.11 While the precise mechanisms are not fully understood, smoking has been reported to contribute to impaired glucose homeostasis by increasing oxidative stress, inflammation, hyperglycaemia and insulin resistance.12,14
Previous studies have consistently shown that high smoking intensity and exposure are linked to an elevated risk of perinatal complications. Numerous studies have established active smoking as a significant risk factor for GDM.14,16 More recently, one research study has revealed a dose–response relationship between smoking intensity and the development or risk of A2GDM only.17 Furthermore, second-hand smoke exposure has been shown to increase the risk of GDM even among non-smoking pregnant women.18 These findings underscore the importance of addressing smoking exposure in both active smokers and non-smokers during pregnancy to mitigate the risk of GDM and its associated complications.
Despite ongoing research efforts, evidence regarding the differential impact of smoking on the development of A1GDM versus A2GDM remains limited. Therefore, further prospective studies are warranted to clarify the effects of smoking on GDM. This study aims to comprehensively investigate the association between smoking behaviours (including smoking status, intensity and secondhand smoke exposure) before and during early pregnancy, and the severity of GDM, using data from a large-scale prospective pregnancy cohort.
Materials and methods
Study population
This study is a secondary analysis that utilised data from the Korean Pregnancy Outcomes Study (KPOS),19 a longitudinal prospective cohort study aimed at investigating pregnancy-related complications and risk factors in South Korea. Initially, 4537 pregnant women were enrolled in the KPOS between 15 March 2013 and 31 January 2017, and all participants provided written informed consent. Cases involving transfer, loss to follow-up, twin pregnancies, miscarriages and pre-existing diabetes mellitus (DM) were excluded, resulting in 3457 cases included in this secondary analysis (online supplemental figure 1). At their initial visit, specifically around 12 weeks of gestation, participants provided comprehensive information including sociodemographic details, medical history, reproductive information, psychological health status and health-related behaviours.
GDM diagnosis followed a two-step screening protocol. Initial screening was performed using a 50 g glucose challenge test (GCT) between 24 and 28 weeks of gestation. Women with GCT values ≥140 mg/dL proceeded to a diagnostic oral glucose tolerance test. Detailed diagnostic criteria and thresholds are provided in our previous report.19 Women diagnosed with GDM were classified as A2GDM if insulin or other pharmacological treatment was required and documented in their medical records during pregnancy, and as A1GDM if managed with diet and lifestyle modifications alone. Based on pregnancy outcomes, participants were categorised into three groups: non-GDM (n=3226), A1GDM (n=198) and A2GDM (n=33).
Patient and public involvement
Patients and the public were not involved in the design, conduct or reporting of this secondary data analysis.
Data collections on smoking behaviour
Smoking behaviours were assessed through a self-administered questionnaire administered at four different time points: around 12 weeks of gestation (visit 1), 24 weeks (visit 2), 36 weeks (visit 3) of pregnancy and 6–8 weeks post partum (visit 4). Participants were categorised into two main groups: active smokers and never smokers. The distribution of active smokers across these pregnancy visits is presented in online supplemental table 3. Active smoking status was further divided into two subgroups. Women who had quit smoking before becoming pregnant were classified as ‘smoking before pregnancy’, while those who had smoked at any time after pregnancy recognition until visit 1 (approximately 12 weeks of gestation) were classified as ‘smoking until early pregnancy’, regardless of whether they had quit before visit 1. The classification was determined based on self-reported smoking status and timing of cessation collected at visit 1.
For active smokers, two additional questions were asked regarding daily cigarette consumption and duration of smoking. Tobacco intensity was quantified using the ‘pack-year’ measurement unit, defined as smoking one pack of cigarettes (20 cigarettes) per day for 1 year. For never smokers, exposure to secondhand smoke (yes/no) was assessed based on exposure to smoke from their husbands, other family members or colleagues at visit 1. The distribution of secondhand smoke exposure among never smokers across pregnancy visits is presented in online supplemental table 4.
Statistical analysis
Baseline characteristics were compared among the non-GDM, A1GDM and A2GDM groups. For continuous variables, differences were evaluated using analysis of variance. For categorical variables, the χ2 test was used.
Covariates for multivariable logistic regression models were selected based on two criteria: (1) variables showing statistical significance (p<0.05) in univariable analyses and (2) sociodemographic factors and cardiovascular risk factors considered as potential confounders. The final adjusted models included maternal age, pre-pregnancy body mass index, systolic blood pressure, diastolic blood pressure, sociodemographic factors (education, household income, occupation), parity and previous GDM history, hypertension, family history of DM, family history of hypertension and secondhand smoke exposure (for analyses of active smoking status and pack-years). Due to multicollinearity between parity and previous GDM history (as previous GDM can only occur in multiparous women), these two variables were combined into a single categorical variable with three levels: nulliparous women, multiparous women without previous GDM and multiparous women with previous GDM.20
Multivariable logistic regression analysis was conducted to estimate adjusted ORs (aORs) and 95% CIs. For analyses involving A2GDM (which had a relatively small sample size of n=33), Firth’s penalised logistic regression was used to reduce small-sample bias in the estimates. For analyses of overall GDM and A1GDM, standard logistic regression was applied.
Results
Analysis of general characteristics
Table 1 presents the baseline characteristics of participants across the three groups. Maternal age, pre-pregnancy body mass index and blood pressure (both systolic and diastolic) were higher in both GDM groups compared with the non-GDM group, with the A2GDM group showing the highest values. The A2GDM group also showed lower educational attainment and household income, and higher proportions of multiparity and previous GDM history. A marked disparity was observed in the prevalence of a first-degree family history of DM, with the highest occurrence in the A2GDM group (66.7%). In contrast, family history of hypertension, physical activity, mental health parameters and use of nutritional supplements showed no notable differences across the three groups (online supplemental table 1). Additionally, baseline characteristics stratified by smoking status (never smoking, smoking before pregnancy and smoking until early pregnancy) are presented in online supplemental table 2.
Table 1. Characteristics of the study participants.
| Variables | Non-GDM (n=3226) | A1GDM (n=198) | A2GDM (n=33) |
|---|---|---|---|
| Maternal age (years) | 33.2±3.7 | 34.2±3.5 | 37.2±3.5 |
| Mean difference (95% CI) | 1.1 0.6 to 1.6) |
4.0 (2.7 to 5.3) |
|
| Pre-pregnancy BMI (kg/m2) | 21.0±2.8 | 23.1±3.8 | 24.7±3.9 |
| Mean difference (95% CI) | 2.1 (1.6 to 2.7) |
3.6 (2.2 to 5.0) |
|
| SBP (mm Hg) | 113.3±12.9 | 119.7±14.6 | 122.9±16.8 |
| Mean difference (95% CI) | 6.4 (4.3 to 8.5) |
9.6 (3.7 to 15.6) |
|
| DBP (mm Hg) | 65.9±9.5 | 69.7±10.1 | 72.6±11.8 |
| Mean difference (95% CI) | 3.9 (2.4 to 5.3) |
6.7 (2.5 to 10.9) |
|
| Education | |||
| ≤High school | 255 (7.9) | 22 (11.1) | 7 (21.2) |
| >High school | 2971 (92.1) | 176 (88.9) | 26 (78.8) |
| Household income | |||
| ≤4 million KRW/month | 957 (29.7) | 58 (29.3) | 19 (57.6) |
| >4 million KRW/month | 2269 (70.3) | 140 (70.7) | 14 (42.4) |
| Occupation | |||
| Unemployed | 1180 (36.6) | 89 (44.9) | 12 (36.4) |
| Employed | 2046 (63.4) | 109 (55.1) | 21 (63.6) |
| History | |||
| Parity and GDM history | |||
| Nulliparous | 1940 (60.1) | 97 (49.0) | 19 (57.6) |
| Multiparous, no previous GDM | 1246 (38.6) | 87 (43.6) | 12 (36.4) |
| Multiparous, previous GDM | 40 (1.2) | 14 (7.1) | 2 (6.1) |
| HTN | |||
| No | 3191 (98.9) | 190 (96.0) | 31 (93.9) |
| Yes | 35 (1.1) | 8 (4.0) | 2 (6.1) |
| PCOS | |||
| No | 3177 (98.5) | 192 (97.0) | 32 (97.0) |
| Yes | 49 (1.5) | 6 (3.0) | 1 (3.0) |
| First-degree family history of HTN | |||
| No | 2021 (63.3) | 112 (56.6) | 19 (57.6) |
| Yes | 1174 (36.7) | 86 (43.4) | 14 (42.4) |
| First-degree family history of DM | |||
| No | 2568 (80.1) | 137 (69.2) | 11 (33.3) |
| Yes | 638 (19.9) | 61 (30.8) | 22 (66.7) |
Continuous variables are presented as mean±SD with mean differences (95% CIs) compared to the non-GDM group shown in Table 1. Categorical variables are presented as count (percentage).
A1GDM, diet-controlled GDM; A2GDM, insulin-requiring GDM; BMI, body mass index; DBP, diastolic blood pressure; DM, diabetes mellitus; GDM, gestational diabetes mellitus; HTN, hypertension; PCOS, polycystic ovary syndrome; SBP, systolic blood pressure.
Analysis of smoking status
The distribution of smoking status differed significantly among the three groups (table 2), with the A2GDM group showing a notably higher proportion of active smokers compared with both non-GDM and A1GDM groups (p<0.05).
Table 2. Distribution of participants by smoking status.
| Smoking status | Non-GDM (n=3226) | A1GDM (n=198) | A2GDM (n=33) | P value |
|---|---|---|---|---|
| Never smoking | 2896 (89.8) | 175 (88.4) | 20 (60.6) | <0.001 |
| Smoking before pregnancy | 256 (7.9) | 17 (8.6) | 8 (24.2) | |
| Smoking until early pregnancy | 74 (2.3) | 6 (3.0) | 5 (15.2) |
Data are presented as counts (percentages).
P values among the three subjects were obtained from the χ2 test.
A1GDM, diet-controlled GDM; A2GDM, insulin-requiring GDM; GDM, gestational diabetes mellitus.
Table 3 presents the associations between smoking status and GDM risk through multivariable logistic regression analysis. For overall GDM, smoking until early pregnancy showed an elevated risk in the unadjusted model (OR 2.21, 95% CI 1.09 to 4.06); however, this association was attenuated and no longer statistically significant after adjustment for cardiovascular risk factors and sociodemographic variables (aOR 1.88, 95% CI 0.89 to 3.65). Smoking before pregnancy was not significantly associated with overall GDM risk in either model.
Table 3. ORs for GDM according to the smoking status.
| Smoking status | Unadjusted OR (95% CI) | Adjusted OR (95% CI)† |
|---|---|---|
| Non-GDM vs overall GDM | ||
| Never smoking | 1 (ref.) | 1 (ref.) |
| Smoking before pregnancy | 1.45 (0.92 to 2.20) |
1.21 (0.74 to 1.89) |
| Smoking until early pregnancy | 2.21 (1.09 to 4.06) |
1.88 (0.89 to 3.65) |
| Non-GDM vs A1GDM | ||
| Never smoking | 1 (ref.) | 1 (ref.) |
| Smoking before pregnancy | 1.10 (0.64 to 1.79) |
0.94 (0.53 to 1.57) |
| Smoking until early pregnancy | 1.34 (0.52 to 2.88) |
1.17 (0.44 to 2.64) |
| Non-GDM vs A2GDM | ||
| Never smoking | 1 (ref.) | 1 (ref.) |
| Smoking before pregnancy | 4.68 (1.97 to 10.18) |
3.98 (1.58 to 9.30) |
| Smoking until early pregnancy | 10.43 (3.58 to 25.80) |
9.90 (2.97 to 29.45) |
| A1GDM vs A2GDM | ||
| Never smoking | 1 (ref.) | 1 (ref.) |
| Smoking before pregnancy | 4.16 (1.57 to 10.48) |
2.98 (0.91 to 9.52) |
| Smoking until early pregnancy | 7.24 (2.05 to 24.96) |
4.36 (1.01 to 18.27) |
Unadjusted OR and 95% CI estimated with logistic regression (crude) or Firth’s penalised logistic regression for analyses involving A2GDM.
Adjusted OR and 95% CI estimated with logistic regression (for overall GDM and A1GDM) or Firth’s penalised logistic regression (for analyses involving A2GDM) adjusting for maternal age, pre-pregnancy body mass index, systolic blood pressure, diastolic blood pressure, education, occupation, parity and previous GDM history, first-degree family history of diabetes mellitus, first-degree family history of hypertension and secondhand smoking.
A1GDM, diet-controlled GDM; A2GDM, insulin-requiring GDM; GDM, gestational diabetes mellitus; ref, reference group.
For A1GDM, neither smoking before pregnancy nor smoking until early pregnancy showed statistically significant associations in adjusted models. In contrast, A2GDM demonstrated substantial and persistent associations: aOR 3.98 (95% CI 1.58 to 9.30) for smoking before pregnancy and aOR 9.90 (95% CI 2.97 to 29.45) for smoking until early pregnancy.
In direct comparisons between A1GDM and A2GDM, both smoking exposures showed elevated risks for A2GDM in crude models. After adjustment, the association with smoking before pregnancy was attenuated and no longer statistically significant (aOR 2.98, 95% CI 0.91 to 9.52), while smoking until early pregnancy remained significant (aOR 4.36, 95% CI 1.01 to 18.27).
Analysis of smoking intensity
To examine the association between cumulative smoking exposure and GDM risk, we analysed pack-years using five categories (online supplemental table 5). The analysis revealed a progressive association between pack-year levels and A2GDM risk, with a notable threshold around 4 pack-years. Based on this pattern, we selected 4 pack-years as the cut-off for our primary analysis.
Table 4 presents the association between smoking intensity (>4 pack-years) and GDM risk. For A2GDM, a substantially elevated risk was observed compared with non-GDM, which remained robust after adjustment (aOR 20.68; 95% CI 6.75 to 59.39). In contrast, A1GDM showed no statistically significant association with smoking intensity (aOR 1.84; 95% CI 0.61 to 4.45). In the direct comparison between A1GDM and A2GDM, smoking over 4 pack-years was significantly associated with A2GDM risk in both crude and adjusted models (aOR 8.15; 95% CI 1.91 to 36.12).
Table 4. The association between smoking intensity and the risk of GDM.
| Smoking intensity by pack-year | Unadjusted OR (95% CI) | Adjusted OR (95% CI)† | |
|---|---|---|---|
| Non-GDM vs A1GDM | Never smoking | 1 (ref.) | 1 (ref.) |
| 0<pack year≤4 | 1.04 (0.61 to 1.67) | 0.87 (0.50 to 1.44) | |
| 4<pack year | 1.88 (0.64 to 4.38) | 1.84 (0.61 to 4.45) | |
| Non-GDM vs A2GDM | Never smoking | 1 (ref.) | 1 (ref.) |
| 0<pack year≤4 | 3.21 (1.21 to 7.42) | 2.81 (1.00 to 6.99) | |
| 4<pack year | 23.81 (9.24 to 55.81) | 20.68 (6.75 to 59.39) | |
| A1GDM vs A2GDM | Never smoking | 1 (ref.) | 1 (ref.) |
| 0<pack year≤4 | 3.01 (1.04 to 7.95) | 2.19 (0.65 to 6.81) | |
| 4<pack year | 11.67 (3.58 to 40.56) | 8.15 (1.91 to 36.12) |
Unadjusted OR and 95% CI estimated with logistic regression (crude) or Firth’s penalised logistic regression for analyses involving A2GDM.
Adjusted OR and 95% CI estimated with logistic regression (for overall GDM and A1GDM) or Firth’s penalised logistic regression (for analyses involving A2GDM) adjusting for maternal age, pre-pregnancy body mass index, systolic blood pressure, diastolic blood pressure, education, occupation, parity and previous GDM history, first-degree family history of diabetes mellitus, first-degree family history of hypertension and secondhand smoking.
A1GDM, diet-controlled GDM; A2GDM, insulin-requiring GDM; GDM, gestational diabetes mellitus; ref, reference group.
Analysis of secondhand smoking
Among never smokers, the prevalence of secondhand smoke exposure did not differ significantly among the three groups (online supplemental table 6). As shown in table 5, although secondhand smoke exposure did not reach statistical significance in the adjusted model (aOR 1.33; 95% CI 0.98 to 1.81), the observed trend suggested a potential association with overall GDM risk.
Table 5. OR for GDM by secondhand smoking in the never-smoker group.
| Exposure to secondhand smoking | Unadjusted OR (95% CI) | Adjusted OR (95% CI)† |
|---|---|---|
| Non-GDM vs overall GDM | ||
| No | 1 (ref.) | 1 (ref.) |
| Yes | 1.31 (0.97 to 1.76) | 1.33 (0.98 to 1.81) |
| Non-GDM vs A1GDM | ||
| No | 1 (ref.) | 1 (ref.) |
| Yes | 1.32 (0.96 to 1.80) | 1.33 (0.96 to 1.83) |
| Non-GDM vs A2GDM | ||
| No | 1 (ref.) | 1 (ref.) |
| Yes | 1.28 (0.51 to 3.04) | 1.31 (0.51 to 3.23) |
| A1GDM vs A2GDM | ||
| No | 1 (ref.) | 1 (ref.) |
| Yes | 0.97 (0.37 to 2.42) | 0.88 (0.31 to 2.39) |
Unadjusted OR and 95% CI estimated with logistic regression (crude) or Firth’s penalised logistic regression for analyses involving A2GDM.
Adjusted OR and 95% CI estimated with logistic regression (for Overall GDM and A1GDM) or Firth’s penalised logistic regression (for analyses involving A2GDM) adjusting for maternal age, pre-pregnancy body mass index, systolic blood pressure, diastolic blood pressure, education, occupation, parity and previous GDM history, first-degree family history of diabetes mellitus and first-degree family history of hypertension.
A1GDM, diet-controlled GDM; A2GDM, insulin-requiring GDM; GDM, gestational diabetes mellitus; ref, reference group.
Discussion
This study examined associations between smoking behaviours during early pregnancy and GDM severity using prospectively collected cohort data. We analysed smoking status, intensity and secondhand smoke exposure in relation to GDM subtypes classified by treatment modality.
Our findings revealed that active smoking was associated with A2GDM, the more severe form requiring pharmacological intervention. Both smoking before pregnancy and smoking continuing into early pregnancy were associated with elevated A2GDM risk. Women with a cumulative smoking history exceeding 4 pack-years showed substantially elevated risk for A2GDM. These findings align with previous studies reporting associations between prenatal smoking and GDM risk.21 22 However, the stronger association observed with A2GDM compared with A1GDM suggests that smoking may influence not only GDM development but also its progression to forms requiring pharmacological treatment. The association with higher cumulative smoking exposure supports the role of chronic or high-intensity smoking in glucose dysregulation during pregnancy.
Notably, smoking was not significantly associated with A1GDM in our study, suggesting that smoking effects may be more closely related to severe forms of GDM. These findings indicate that cumulative smoking history may serve as a clinically relevant marker for identifying women at higher risk of severe GDM. In our study population, higher cumulative exposure (particularly exceeding 4 pack-years) showed substantially elevated risks. Healthcare providers could consider incorporating detailed smoking assessments into risk stratification during early prenatal care.
The impact of secondhand smoke on GDM risk remains debated in the literature. While some studies reported no significant associations,23 24 others have demonstrated potential associations.25 In our study, secondhand smoke exposure among never-smokers showed a borderline association with overall GDM risk, suggesting a potential relationship that warrants further investigation. These findings underscore the importance of promoting smoke-free environments during pregnancy. Targeted education and public health efforts are needed to inform pregnant women, partners, family members and coworkers about potential risks associated with secondhand smoke exposure and to encourage smoke-free environments during the prenatal period.
This study has several strengths. We used a large, well-characterised prospective cohort with standardised data collection protocols. The analysis of GDM subtypes based on treatment modality provided insights into the differential effects of smoking on GDM severity. We employed Firth’s penalised logistic regression to address potential small-sample bias in analyses involving the A2GDM subgroup. The cohort’s demographic characteristics were representative of the general reproductive-age population in South Korea, supporting the generalisability of our findings within similar populations.
However, several limitations warrant consideration. First, smoking behaviour was assessed through self-reported questionnaires, which may be subject to recall and social desirability bias. Pregnant women may underreport smoking due to social stigma, and previous studies have shown that self-reported smoking often underestimates actual exposure compared with biochemical validation.26 Second, the relatively small sample size of the A2GDM group (n=33) posed challenges for subgroup analyses. Although we employed Firth’s penalised logistic regression to reduce small-sample bias, statistical power remained limited, potentially contributing to wide CIs and increased risk of type II errors. Third, our data was collected between 2013 and 2017. However, smoking prevalence and intensity among women of reproductive age have remained relatively stable over this period, according to Statistics Korea, suggesting that our findings remain applicable to contemporary populations.27 Finally, residual confounding from unmeasured variables cannot be entirely excluded despite comprehensive covariate adjustment. Future research should incorporate objective biomarkers of smoking exposure, larger sample sizes for rare outcomes and contemporary cohorts to validate these findings and explore underlying mechanisms linking smoking intensity to GDM severity.
Conclusions
This study examined associations between smoking behaviours during early pregnancy and GDM severity using prospective cohort data from Korean pregnant women. Our findings revealed that active smoking was associated with A2GDM, the more severe form requiring pharmacological intervention. Higher cumulative smoking exposure showed a substantially elevated risk. In contrast, smoking was not significantly associated with A1GDM. Secondhand smoke exposure among never-smokers showed a borderline association with overall GDM risk, warranting further investigation. These findings suggest that smoking history may serve as a clinically relevant marker for identifying women at risk of severe GDM. Healthcare providers could consider incorporating smoking assessments into prenatal risk stratification and offering targeted smoking cessation support during preconception and early pregnancy care. Such interventions may contribute to reducing the burden of severe GDM and improving maternal and fetal health outcomes.
Supplementary material
Acknowledgements
We express our sincere gratitude to the Division of Maternal-Fetal Medicine in the Cheil General Hospital and CHA Hospital, which were involved in the KPOS research (the Korea National Institute of Health research project: 2015-E6302-00).
Footnotes
Funding: This work was supported by the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI; Ministry of Health and Welfare, Republic of Korea) (grant numbers: RS-2022-KH129953 and RS-2025-02213591) and the Korea Disease Control and Prevention Agency (KDCA) National Institute of Health (NIH) research project (project No. 2024-ER1104-01).
Prepublication history and additional supplemental material for this paper are available online. To view these files, please visit the journal online (https://doi.org/10.1136/bmjopen-2025-112308).
Provenance and peer review: Not commissioned; externally peer reviewed.
Patient consent for publication: Not applicable.
Ethics approval: This study involves human participants. This study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Review Boards of Cheil General Hospital (IRB number: CGH-IRB-2013-10) and CHA Hospital (IRB number: 2013-14-KNC13-018, CHAMC 2021-04-052-001). This analysis using data from KPOS was also approved by the CHA Bundang Medical Center (CHAMC 2021-04-052). Participants gave informed consent to participate in the study before taking part.
Data availability free text: Data are available upon reasonable request. The data that support the findings of this study are not publicly available due to privacy and ethical restrictions but are available from the corresponding author (HMR) upon reasonable request and with permission of the Korean Pregnancy Outcome Study (KPOS) steering committee and institutional review board.
Patient and public involvement: Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Data availability statement
Data are available upon reasonable request.
References
- 1.Mack LR, Tomich PG. Gestational Diabetes: Diagnosis, Classification, and Clinical Care. Obstet Gynecol Clin North Am. 2017;44:207–17. doi: 10.1016/j.ogc.2017.02.002. [DOI] [PubMed] [Google Scholar]
- 2.Kunarathnam V, Vadakekut ES, Mahdy H. Gestational diabetes. 2025. https://www.ncbi.nlm.nih.gov/books/NBK545196/ Available. [PubMed]
- 3.ACOG Practice Bulletin No. 190: Gestational Diabetes Mellitus. Obstet Gynecol. 2018;131:e49–64. doi: 10.1097/AOG.0000000000002501. [DOI] [PubMed] [Google Scholar]
- 4.Saeedi M, Cao Y, Fadl H, et al. Increasing prevalence of gestational diabetes mellitus when implementing the IADPSG criteria: A systematic review and meta-analysis. Diabetes Res Clin Pract. 2021;172:108642.:S0168-8227(20)30899-8. doi: 10.1016/j.diabres.2020.108642. [DOI] [PubMed] [Google Scholar]
- 5.Roglic G. Global Report on Diabetes. Geneva: World Health Organization; 2016. [Google Scholar]
- 6.Wang Z, Kanguru L, Hussein J, et al. Incidence of adverse outcomes associated with gestational diabetes mellitus in low- and middle-income countries. Int J Gynaecol Obstet. 2013;121:14–9. doi: 10.1016/j.ijgo.2012.10.032. [DOI] [PubMed] [Google Scholar]
- 7.Zhu Y, Zhang C. Prevalence of Gestational Diabetes and Risk of Progression to Type 2 Diabetes: a Global Perspective. Curr Diab Rep. 2016;16:7. doi: 10.1007/s11892-015-0699-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Capobianco G, Gulotta A, Tupponi G, et al. Materno-Fetal and Neonatal Complications of Diabetes in Pregnancy: A Retrospective Study. J Clin Med. 2020;9:2707. doi: 10.3390/jcm9092707. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Karkia R, Giacchino T, Hii F, et al. Gestational diabetes mellitus: relationship of adverse outcomes with severity of disease. J Matern Fetal Neonatal Med. 2024;37:2356031. doi: 10.1080/14767058.2024.2356031. [DOI] [PubMed] [Google Scholar]
- 10.Amiri FN, Faramarzi M, Bakhtiari A, et al. Risk Factors for Gestational Diabetes Mellitus: A Case-Control Study. Am J Lifestyle Med. 2021;15:184–90. doi: 10.1177/1559827618791980. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Erickson AC, Arbour LT. Heavy smoking during pregnancy as a marker for other risk factors of adverse birth outcomes: a population-based study in British Columbia, Canada. BMC Public Health. 2012;12:102. doi: 10.1186/1471-2458-12-102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Epifano L, Di Vincenzo A, Fanelli C, et al. Effect of cigarette smoking and of a transdermal nicotine delivery system on glucoregulation in type 2 diabetes mellitus. Eur J Clin Pharmacol. 1992;43:257–63. doi: 10.1007/BF02333019. [DOI] [PubMed] [Google Scholar]
- 13.Maddatu J, Anderson-Baucum E, Evans-Molina C. Smoking and the risk of type 2 diabetes. Transl Res. 2017;184:101–7. doi: 10.1016/j.trsl.2017.02.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kataria Y, Gaewsky L, Ellervik C. Prenatal smoking exposure and cardio-metabolic risk factors in adulthood: a general population study and a meta-analysis. Int J Obes (Lond) 2019;43:763–73. doi: 10.1038/s41366-018-0206-y. [DOI] [PubMed] [Google Scholar]
- 15.Masalin S, Kautiainen H, Gissler M, et al. Impact of smoking on gestational diabetes mellitus and offspring birthweight in primiparous women. Acta Obstet Gynecol Scand. 2020;99:1632–9. doi: 10.1111/aogs.13924. [DOI] [PubMed] [Google Scholar]
- 16.Bar-Zeev Y, Haile ZT, Chertok IA. Association Between Prenatal Smoking and Gestational Diabetes Mellitus. Obstet Gynecol. 2020;135:91–9. doi: 10.1097/AOG.0000000000003602. [DOI] [PubMed] [Google Scholar]
- 17.Kim MK, Han K, You SY, et al. Prepregnancy smoking and the risk of gestational diabetes requiring insulin therapy. Sci Rep. 2020;10:13901. doi: 10.1038/s41598-020-70873-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Na J, Chen H, An H, et al. Passive Smoking and Risk of Gestational Diabetes Mellitus among Nonsmoking Women: A Prospective Cohort Study in China. Int J Environ Res Public Health. 2022;19:4712. doi: 10.3390/ijerph19084712. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Choi H, Kwak DW, Kim MH, et al. The Korean Pregnancy Outcome Study (KPOS): Study Design and Participants. J Epidemiol. 2021;31:392–400. doi: 10.2188/jea.JE20200055. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Syngelaki A, Wright A, Gomez Fernandez C, et al. First-Trimester Prediction of Gestational Diabetes Mellitus Based on Maternal Risk Factors. BJOG. 2025;132:972–82. doi: 10.1111/1471-0528.18110. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Kim KS, Hong S, Han K, et al. The Clinical Characteristics of Gestational Diabetes Mellitus in Korea: A National Health Information Database Study. Endocrinol Metab (Seoul) 2021;36:628–36. doi: 10.3803/EnM.2020.948. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Leonardi-Bee J, Britton J, Venn A. Secondhand smoke and adverse fetal outcomes in nonsmoking pregnant women: a meta-analysis. Pediatrics. 2011;127:734–41. doi: 10.1542/peds.2010-3041. [DOI] [PubMed] [Google Scholar]
- 23.Xu X, Liu Y, Liu D, et al. Prevalence and Determinants of Gestational Diabetes Mellitus: A Cross-Sectional Study in China. Int J Environ Res Public Health. 2017;14:1532. doi: 10.3390/ijerph14121532. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Hosler AS, Nayak SG, Radigan AM. Stressful events, smoking exposure and other maternal risk factors associated with gestational diabetes mellitus. Paediatr Perinat Epidemiol. 2011;25:566–74. doi: 10.1111/j.1365-3016.2011.01221.x. [DOI] [PubMed] [Google Scholar]
- 25.Carroll X, Liang X, Zhang W, et al. Socioeconomic, environmental and lifestyle factors associated with gestational diabetes mellitus: A matched case-control study in Beijing, China. Sci Rep. 2018;8:8103. doi: 10.1038/s41598-018-26412-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Ashford K, Wiggins A, Rayens E, et al. Perinatal Biochemical Confirmation of Smoking Status by Trimester. Nicotine Tob Res. 2017;19:631–5. doi: 10.1093/ntr/ntw332. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Korea Disease Control and Prevention Agency Korea National Health and Nutrition Examination Survey (KNHANES): Trends in current smoking rates. [30-Dec-2024]. https://kosis.kr/statHtml/statHtml.do?orgId=177&tblId=DT_11702_N001&conn_path=I2 Available. Accessed.