Timing of smoking cessation before and during pregnancy and risk of gestational diabetes: A systematic review

Abstract

Introduction

Gestational diabetes mellitus (GDM) is one of the most common metabolic complications of pregnancy, with a steadily increasing global prevalence. Cigarette smoking is a well-established modifiable risk factor for metabolic disorders; however, its association with gestational diabetes mellitus remains unclear, partly due to the widespread use of simplified, binary exposure definitions. Of particular interest is the timing of smoking cessation before or during pregnancy, as it may differentially influence the risk of glucose dysregulation during gestation.

Objective

The objective of this systematic review was to synthesize available evidence from original studies that classify maternal smoking according to cessation timing—before pregnancy, during early pregnancy, or continued smoking beyond the first trimester—and to examine the association of these exposure patterns with gestational diabetes mellitus and pregnancy-related dysglycemia.

Methods

This review was designed and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. A comprehensive systematic search was conducted in MEDLINE, Embase, Scopus, and Web of Science from inception with no date restrictions. Eligible studies were original observational investigations including pregnant women without pregestational diabetes, in which smoking exposure was clearly defined with respect to cessation timing. The primary outcome was gestational diabetes mellitus, while secondary outcomes included pregnancy-related dysglycemia. Due to substantial heterogeneity in exposure definitions and outcome assessment, results were synthesized narratively.

Results

Of 1672 identified records, four studies met the inclusion criteria. Overall, continued smoking beyond early pregnancy was consistently associated with an increased risk of gestational diabetes mellitus. In contrast, smoking cessation before pregnancy or during the first trimester appeared to attenuate or eliminate this excess risk, although some estimates did not reach statistical significance due to limited statistical power. Findings related to pregnancy dysglycemia further highlighted the complex interplay between smoking behavior, gestational weight gain, and maternal glucose metabolism.

Conclusions

The timing of smoking cessation is a critical modifier of gestational diabetes mellitus risk. Continued smoking after the first trimester is associated with increased metabolic risk, whereas early cessation before or during early pregnancy appears beneficial. These findings underscore the importance of timing-sensitive smoking exposure assessment in both epidemiological research and clinical counseling of pregnant women. However, these findings are based on a very limited evidence base, as only four observational studies met the inclusion criteria, substantially restricting the strength and generalizability of inferences.

1. Introduction

Gestational diabetes mellitus (GDM) represents one of the most common metabolic complications of pregnancy, with a steadily increasing global prevalence over recent decades. This rise parallels demographic and lifestyle shifts, including advanced maternal age, increasing rates of overweight and obesity, and changes in health-related behaviors before and during pregnancy [1]. GDM is associated with substantial short- and long-term consequences, including adverse perinatal outcomes, increased cesarean delivery rates, and a markedly elevated lifetime risk of type 2 diabetes and cardiometabolic disease in affected women, as well as metabolic vulnerability in their offspring [2,3]. Clinically, GDM is typically screened for at 24–28 weeks of gestation and diagnosed using glucose tolerance testing; the most widely used contemporary criteria include a one-step 75-g 2-h oral glucose tolerance test (OGTT), with GDM diagnosed if any value meets or exceeds fasting 92 mg/dL (5.1 mmol/L), 1-h 180 mg/dL (10.0 mmol/L), or 2-h 153 mg/dL (8.5 mmol/L), or a two-step strategy consisting of a non-fasting 50-g glucose load test followed—if positive—by a diagnostic 100-g 3-h OGTT (commonly using Carpenter–Coustan thresholds and requiring ≥2 abnormal values) [4].

Cigarette smoking is a well-established modifiable risk factor for metabolic disease in the general population [5]. In non-pregnant adults, smoking has been consistently associated with insulin resistance, chronic low-grade inflammation, oxidative stress, and an increased risk of type 2 diabetes mellitus, while smoking cessation may transiently alter glycemic control and promote weight gain. In pregnancy, however, the relationship between smoking and glucose metabolism appears more complex and less consistent, with studies reporting increased, null, or even inverse associations between smoking during pregnancy and GDM risk [[6], [7], [8]].

One important limitation of the existing literature is the predominant use of binary exposure definitions, classifying women simply as smokers or non-smokers during pregnancy. Such an approach fails to capture the dynamic nature of smoking behavior surrounding conception and early gestation. Pregnancy is a unique period during which many women substantially modify health behaviors, with a significant proportion stopping smoking either before conception or shortly after pregnancy recognition, while others continue smoking beyond the first trimester. These distinct cessation trajectories may carry different metabolic implications, particularly in the context of gestational insulin resistance, placental adaptations, and gestational weight gain.

Emerging evidence suggests that timing of smoking cessation may be critical in determining metabolic risk during pregnancy. Large observational cohorts and registry-based studies have demonstrated that women who stop smoking during early pregnancy exhibit metabolic and perinatal profiles that differ from those who continue smoking later into gestation [8,9]. In the Finnish Medical Birth Register, cessation during the first trimester was clearly distinguishable from continued smoking in terms of GDM prevalence and pregnancy outcomes, highlighting early gestation as a potentially sensitive window for metabolic risk modification [8]. Similarly, analysis stratifying smoking behavior by early versus continued exposure have suggested that continued smoking is associated with higher GDM risk, even after adjustment for maternal age and prepregnancy body mass index [10].

In addition to clinically diagnosed GDM, smoking cessation timing may also influence subclinical gestational dysglycemia, including abnormal glucose tolerance and elevated glucose screening values. Prospective cohort data with repeated smoking assessments across pregnancy indicate that preconception smoking, early pregnancy cessation, and mid-pregnancy smoking represent metabolically distinct exposure states, with divergent associations with glucose tolerance testing results [11]. These findings underscore the importance of considering gestational timing rather than static exposure classifications when evaluating smoking-related metabolic risk.

Despite these signals, the evidence remains fragmented. Existing systematic reviews and meta-analyses addressing smoking and GDM have largely focused on overall smoking exposure or passive smoking, without systematically synthesizing data according to cessation timing trajectories. As a result, critical questions remain unanswered regarding whether early smoking cessation mitigates GDM risk, whether preconception cessation confers metabolic risk comparable to never smoking, and to what extent continued smoking beyond early pregnancy drives dysglycemia independently of established confounders such as prepregnancy BMI and gestational weight gain.

The aim of this systematic review is to synthesize original human studies that explicitly allow classification of maternal smoking by cessation timing, early pregnancy, or later/continued smoking—and to examine their associations with gestational diabetes mellitus and pregnancy-related dysglycemia outcomes.

2. Materials and methods

2.1. Study design and reporting framework

The review was designed and reported in accordance with the principles of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [12]. The protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO; registration number: CRD420261285619).

2.2. Eligibility criteria

Studies were considered eligible if they met the following criteria. The study population had to consist of pregnant women without known pregestational diabetes mellitus. Exposure assessment was required to include explicit information on smoking cessation timing, allowing classification of maternal smoking behavior as cessation before pregnancy, cessation during early pregnancy (typically first trimester or after pregnancy recognition), or continued smoking beyond early pregnancy. Eligible outcomes included clinically diagnosed GDM or pregnancy-related dysglycemia, defined using oral glucose tolerance testing, medical record diagnoses, registry data, or validated screening criteria. Only original observational studies, including prospective cohorts, retrospective cohorts, register-based studies, or population-based survey analyses, were eligible.

Studies were excluded if smoking exposure was reported only as a binary variable (smoker versus non-smoker) without information on cessation timing, if they focused exclusively on passive smoking, if pregestational diabetes was not excluded or could not be disentangled, or if they were reviews, editorials, conference abstracts without full data, case reports, or animal studies.

2.3. Information sources and search strategy

A comprehensive literature search was conducted in MEDLINE (via PubMed), Embase, Scopus, and Web of Science Core Collection from database inception to the date of the final search. The search strategy combined controlled vocabulary and free-text terms related to pregnancy, gestational diabetes, cigarette smoking, and smoking cessation timing. Core concepts included “gestational diabetes” or “pregnancy dysglycemia,” “pregnancy” or “gestation,” “smoking” or “cigarette,” and terms capturing cessation timing such as “quit,” “cessation,” “stopped,” “preconception,” “first trimester,” “early pregnancy,” or “during pregnancy.”

An example MEDLINE search strategy was: (gestational diabetes OR GDM OR abnormal glucose tolerance OR hyperglycaemia in pregnancy) AND (pregnancy OR pregnant OR gestation∗) AND (smoking OR cigarette∗ OR tobacco) AND (cessation OR quit∗ OR stopped OR preconception OR first trimester OR early pregnancy). Equivalent strategies were adapted for the other databases using database-specific syntax and indexing terms.

To maximize sensitivity, no date restrictions were applied. Reference lists of relevant reviews and all included studies were manually screened to identify additional eligible articles. Citation tracking of key studies was also performed.

2.4. PRISMA process

The electronic database search initially identified 1672 records. After removal of duplicate entries, 1450 unique records remained for screening. Title and abstract screening led to the exclusion of 1363 records. The most common reasons for exclusion at this stage were the classification of smoking exposure as a binary variable without information on cessation timing, outcomes unrelated to glucose metabolism or insulin resistance, and study populations not fulfilling the predefined eligibility criteria.

A total of 87 articles were retrieved for full-text evaluation. During full-text assessment, 83 studies were excluded. The principal reasons for exclusion included failure to differentiate smoking exposure according to critical reproductive periods (preconception, early pregnancy, or continued smoking throughout pregnancy), treatment of smoking solely as a confounding variable rather than as the primary exposure of interest, and absence of gestational diabetes mellitus or pregnancy-related dysglycemia as a reported outcome. Following the application of all inclusion and exclusion criteria, four studies met the eligibility requirements and were included in the final qualitative synthesis. Fig. 1 illustrates the study selection process.

Fig. 1.

Illustrates the study selection process.

2.5. Data extraction

Data was extracted using a standardized data collection form developed specifically for this review. Extracted information included author and publication year, country and data source, study design and study period, sample size and population characteristics, smoking exposure assessment method and definitions of cessation timing, comparator groups, outcome definitions and diagnostic criteria, adjusted effect estimates with corresponding confidence intervals, and covariates included in multivariable models. For each study, data was extracted from the most fully adjusted analytical model available.

2.6. Exposure classification and harmonization

To allow conceptual comparability across heterogeneous study designs, smoking exposure was harmonized into three overarching categories: cessation before pregnancy, cessation during early pregnancy, and continued smoking beyond early pregnancy. When preconception cessation could not be separated from never-smoking, women were classified according to the most temporally precise cessation category available. Differences in exposure ascertainment and timing definitions were explicitly considered in the synthesis.

2.7. Outcome definitions

The primary outcome of interest was GDM, as defined by each study using oral glucose tolerance testing thresholds, medical record diagnoses, registry codes, or validated self-report of clinician-diagnosed GDM. A secondary outcome was pregnancy-related dysglycemia, defined as abnormal glucose tolerance or elevated glucose screening values in the absence of a formal GDM diagnosis.

2.8. Risk of bias assessment

Risk of bias was assessed qualitatively using key domains adapted from the Newcastle–Ottawa Scale for observational studies [13], including selection of the study population, ascertainment of smoking exposure and cessation timing, outcome assessment, and adequacy of adjustment for relevant confounders. Particular attention was paid to adjustment for prepregnancy body mass index and gestational weight gain, given their potential roles as confounders or mediators in the smoking–GDM association.

2.9. Data synthesis

Given heterogeneity in study design, exposure definitions, outcome ascertainment, and analytical strategies, findings were synthesized using a structured narrative approach. Results were grouped according to cessation timing category and outcome type, and patterns of association were interpreted in the context of study-specific methodological strengths and limitations. Quantitative pooling was not undertaken because cessation timing and outcome definitions were insufficiently comparable across studies.

3. Results

3.1. Study selection and characteristics

Four original observational studies met the inclusion criteria for this systematic review [8,10,11,14], all of which evaluated smoking exposure with explicit attention to cessation timing before or during pregnancy in relation to GDM or pregnancy-related dysglycemia. Two studies were large population-based analyses from national surveillance or registry data (United States and Finland) [8,10], and two were prospective cohort studies from clinical or population-based samples [11,14]. Study periods ranged from 2000 to 2015, with sample sizes varying from 1006 to 222,408 pregnancies.

Smoking exposure was assessed using self-reported data in all studies, with cessation timing categorized variably as cessation before pregnancy, cessation in early pregnancy (often first trimester or after pregnancy recognition), and continued smoking during pregnancy. Outcomes included clinically diagnosed GDM based on oral glucose tolerance testing or medical records in three studies, and abnormal glucose tolerance (AGT) defined by elevated screening glucose values in one study.

Table 1 highlights several cross-study patterns that help contextualize the direction and consistency of findings. First, across settings and designs, the most coherent signal is observed for continued smoking beyond early pregnancy, which was associated with higher odds of GDM in the two largest population-based datasets (PRAMS and the Finnish register) and showed a similar direction when compared against continued smoking in the Generation R cohort, albeit with limited precision. In contrast, early cessation (before pregnancy or during the first trimester/after pregnancy recognition) generally yielded effect estimates closer to the null, suggesting attenuation of excess risk relative to continued smoking; however, confidence intervals were wide in smaller cohorts, limiting definitive conclusions regarding risk normalization versus partial reduction.

Table 1.

Baseline characteristics of the included studies.

Study (Author, year)	Country/data source	Design/years	Population	Exposure (cessation timing)	Comparator	Outcome (definition)	Main adjusted results (GDM)	Covariates (adjusted model)	NOS score
Bar-Zeev et al., 2020 [10]	USA; PRAMS (CDC/state surveillance), linked to birth certificate data	Secondary analysis of cross-sectional postpartum survey; 2009–2015	N = 222,408 singleton pregnancies; postpartum respondents surveyed 2–6 months after delivery	Self-reported cigarettes/day 3 months before pregnancy and last 3 months of pregnancy; categories: nonsmoker ≥2y, quit before pregnancy, smoked before but quit during pregnancy, reduced, same or more	Reference: nonsmoker ≥2 years	GDM self-reported (“told by a doctor/nurse/health care worker …“)	vs nonsmoker ≥2y: Quit before pregnancy aOR 0.90 (0.74–1.10); Quit during pregnancy aOR 1.05 (0.95–1.16); Reduced aOR 1.22 (1.08–1.38); Same or more aOR 1.46 (1.25–1.71)	Maternal age, race/ethnicity, parity, prepregnancy BMI category, gestational weight gain category, hypertension in pregnancy	7/9
Haskins et al., 2010 [11]	USA; Latina Pregnancy Study, Western Massachusetts	Prospective cohort, 2000–2004	N = 1006 Hispanic (predominantly Puerto Rican) singleton pregnancies; excluded pregestational DM	Smoking assessed at three time points: pre-pregnancy, early pregnancy (∼15 wks), mid-pregnancy (∼28 wks) using PRAMS questions. Categories included non-smoker, former smoker (quit before or during pregnancy), current smoker	Non-smokers who did not previously smoke	Abnormal glucose tolerance (AGT) defined as 1-h 50 g OGTT >135 mg/dL	Pre-pregnancy smoking: aOR 1.27 (0.67–2.38); early-pregnancy former smoker: aOR 1.28 (0.68–2.41); early-pregnancy smoker: aOR 0.48 (0.21–1.10); mid-pregnancy former smoker: aOR 1.06 (0.57–2.00); mid-pregnancy smoker: aOR 0.38 (0.13–1.11)	Age, prepregnancy BMI, gestational weight gain, parity, education (pre-pregnancy models also adjusted for early-pregnancy smoking)	7/9
Masalin et al., 2020 [8]	Finland; Finnish Medical Birth Register + national health registers	Observational register-based cohort; deliveries 2009–2015	N = 4111 primiparous women, singleton term births (37–42 wks), complete OGTT data; pregestational DM excluded	Smoking status self-reported at delivery using predefined categories: non-smoker, quit during 1st trimester, continued smoking after 1st trimester	Non-smokers	Gestational diabetes mellitus, defined by ≥ 1 abnormal value on standard 2-h 75 g OGTT (fasting ≥5.3 mmol/L; 1-h ≥ 10.0; 2-h ≥ 8.6)	Quit during 1st trimester vs non-smoker: aOR 1.24 (95% CI 0.90–1.72); Continued after 1st trimester vs non-smoker: aOR 1.65 (95% CI 1.09–2.47)	Maternal age, prepregnancy BMI, educational attainment, cohabiting status	9/9
Al-Hassany et al., 2020 [14]	Netherlands; Generation R Study, Rotterdam	Prospective population-based cohort, 2001–2005	N = 7389 singleton pregnancies with ≥2 pregnancy weight measures; women with pregestational diabetes excluded	Smoking status from questionnaires: never smoked during pregnancy, smoking cessation in early pregnancy (quit after pregnancy recognition), continued smoking during pregnancy	Continued smoking during pregnancy	Gestational diabetes mellitus, obtained from medical records (ICD-based clinical diagnosis)	Smoking cessation in early pregnancy vs continued smoking: aOR 0.38 (95% CI 0.09–1.74) (maternal BMI–adjusted); Never smoking vs continued: aOR 0.97 (0.50–1.86)	Gestational age at enrolment and measurement, maternal age, ethnicity, education, parity, folic acid use, alcohol use; additional model adjusted for prepregnancy BMI	7/9

aOR – Adjusted Odds Ratio; AGT – Abnormal Glucose Tolerance; BMI – Body Mass Index; CDC – Centers for Disease Control and Prevention; CI – Confidence Interval; DM – Diabetes Mellitus; GDM – Gestational Diabetes Mellitus; ICD – International Classification of Diseases; NOS: Newcastle Ottawa Scale; OGTT – Oral Glucose Tolerance Test; PRAMS – Pregnancy Risk Assessment Monitoring System; RoB – Risk of Bias; USA – United States of America; wks – Weeks.

Second, the table illustrates that outcome ascertainment differs meaningfully across studies, ranging from self-reported clinician-diagnosed GDM (PRAMS) to OGTT-defined GDM using registry-linked clinical testing (Finland) and medical record–based diagnoses (Generation R), while one cohort assessed abnormal glucose tolerance rather than clinical GDM. These differences likely contribute to variability in effect sizes and may dilute associations when outcomes are less specific or when diagnostic thresholds differ across countries and time periods.

Third, adjustment strategies varied, particularly regarding prepregnancy BMI and gestational weight gain, variables that may act as confounders and/or mediators in the smoking–GDM pathway. Studies that included gestational weight gain in the multivariable model (e.g., PRAMS and the Latina Pregnancy Study) may have partially adjusted away smoking-related metabolic effects operating through pregnancy weight trajectories, whereas others used more limited covariate sets. Overall, Table 1 supports the interpretation that methodological heterogeneity—especially in exposure timing definitions, outcome ascertainment, and treatment of weight-related variables—likely explains much of the between-study variability, while the direction of association for continued smoking remains comparatively consistent.

3.2. Smoking cessation timing and risk of gestational diabetes mellitus

Three studies directly evaluated GDM as the outcome in relation to smoking cessation timing.

In the large U.S. Pregnancy Risk Assessment Monitoring System (PRAMS)-based analysis by Bar-Zeev et al. [10], women who continued smoking during pregnancy—either at the same or increased intensity—had significantly higher odds of GDM compared with long-term non-smokers. In contrast, women who quit smoking before pregnancy or quit during pregnancy did not exhibit a statistically significant increase in GDM risk after adjustment for maternal age, race/ethnicity, prepregnancy BMI, gestational weight gain, and hypertension.

In the Finnish register-based cohort by Masalin et al. [8], smoking cessation timing was classified with greater precision. Women who continued smoking beyond the first trimester had a significantly increased risk of GDM compared with non-smokers (adjusted OR 1.65, 95% CI 1.09–2.47). Women who quit smoking during the first trimester showed a numerically higher but statistically non-significant risk (adjusted OR 1.24, 95% CI 0.90–1.72), suggesting partial attenuation of risk with early cessation.

The Generation R cohort study by Al-Hassany et al. [14] directly compared early pregnancy cessation (defined as quitting after pregnancy recognition) with continued smoking during pregnancy. Early cessation was associated with a lower odds of GDM compared with continued smoking (adjusted OR 0.38, 95% CI 0.09–1.74), although the estimate was imprecise due to a small number of GDM cases. No meaningful difference in GDM risk was observed between never-smokers and continued smokers.

Across studies, continued smoking during pregnancy—particularly beyond early gestation—was consistently associated with higher GDM risk, whereas early cessation appeared to attenuate or eliminate this excess risk, albeit with variability in statistical precision.

3.3. Smoking cessation timing and pregnancy dysglycemia

One prospective cohort study assessed pregnancy dysglycemia rather than clinical GDM. In the Latina Pregnancy Study by Haskins et al. [11], smoking exposure was assessed at multiple time points (pre-pregnancy, early pregnancy, and mid-pregnancy), enabling a detailed cessation-timing analysis. Smoking during early or mid-pregnancy was not associated with an increased risk of AGT and was instead associated with lower mean screening glucose values. Light smoking before pregnancy showed a transient association with AGT in age-adjusted models, which was attenuated after adjustment for gestational weight gain.

Although this study did not evaluate GDM per se, its findings contribute mechanistic insight into the complex relationship between smoking exposure, cessation timing, and maternal glucose metabolism during pregnancy.

4. Discussion

This systematic review synthesizes the available evidence examining smoking cessation timing before and during pregnancy in relation to GDM and pregnancy-related dysglycemia. By moving beyond binary exposure definitions and focusing on when smoking cessation occurs, this review addresses a critical gap in the literature and provides a more clinically actionable framework for understanding smoking-related metabolic risk in pregnancy.

Across heterogeneous populations and study designs, a consistent pattern emerges: continued smoking during pregnancy, particularly beyond early gestation, is associated with an increased risk of GDM, whereas cessation before conception or during early pregnancy appears to attenuate this risk. Register-based evidence from Finland demonstrated a statistically significant increase in GDM among women who continued smoking after the first trimester, while women who quit during the first trimester showed a lower and non-significant excess risk [8]. Similarly, data from the Generation R cohort suggest that early pregnancy cessation may confer a protective effect compared with continued smoking, albeit with limited statistical precision due to small numbers of GDM cases [14].

In contrast, large surveillance data from the United States indicate that women who quit smoking either before pregnancy or during pregnancy did not experience a statistically significant increase in GDM risk compared with long-term non-smokers, whereas those who continued smoking—particularly at unchanged or increased intensity—had clearly elevated odds of GDM [9]. Taken together, these findings support the concept that smoking cessation timing is a key modifier of gestational metabolic risk, and that risk appears largely driven by persistent exposure later in pregnancy, rather than by smoking history per se.

Evidence from pregnancy dysglycemia outcomes further supports this interpretation. In the prospective Latina Pregnancy Study, smoking exposure assessed across multiple gestational windows revealed no increased risk of abnormal glucose tolerance among women who stopped smoking early, while continued smoking was associated with lower screening glucose values—findings that reflect the complex and sometimes paradoxical metabolic effects of nicotine exposure and gestational weight dynamics [10].

Variation in effect estimates across the included studies can be partly explained by important methodological differences in study design, exposure classification, outcome ascertainment, and analytical adjustment strategies. Large register- and surveillance-based analyses, such as those from the United States and Finland, benefited from substantial statistical power and comprehensive population coverage, but relied primarily on self-reported smoking status captured at limited time points, potentially leading to nondifferential exposure misclassification that may have attenuated true associations. In contrast, smaller prospective cohort studies incorporated repeated smoking assessments and more precise temporal exposure definitions, enabling finer discrimination between early cessation and continued smoking, but were limited by fewer GDM cases and wider confidence intervals.

Differences in outcome ascertainment further contribute to heterogeneity. Studies using registry-based or medical record–confirmed GDM diagnoses based on oral glucose tolerance testing likely captured clinically meaningful dysglycemia, whereas studies relying on self-reported GDM or surrogate outcomes such as abnormal glucose tolerance may have introduced outcome misclassification and diluted observed associations. Moreover, diagnostic criteria for GDM varied across countries and study periods, reflecting evolving clinical guidelines, which further complicates direct comparison of absolute risk estimates.

Adjustment strategies also differed substantially. Some studies adjusted extensively for prepregnancy body mass index, gestational weight gain, and hypertensive disorders of pregnancy, while others included a more limited set of sociodemographic covariates. Importantly, gestational weight gain was variably treated as a confounder rather than a potential mediator, which may have resulted in overadjustment and partial attenuation of smoking-related effects in some analyses. Collectively, these methodological differences likely account for a significant proportion of the observed variability in effect estimates and underscore the need for harmonized exposure definitions, standardized outcome assessment, and analytically coherent modeling strategies in future studies (see Fig. 1).

Fig. 2 provides a conceptual summary of the relationship between smoking cessation timing and gestational metabolic risk, highlighting the gradient of risk reduction associated with earlier cessation.

Fig. 2.

Timing of smoking cessation in relation to gestational metabolic risk.

This schematic summarizes evidence from the included studies on the association between smoking cessation timing and risk of gestational diabetes mellitus (GDM). Continued smoking beyond early pregnancy is consistently associated with increased GDM risk. Smoking cessation before conception or during early pregnancy appears to attenuate this excess risk, although current evidence is insufficient to confirm complete risk normalization to the level of never-smokers. Observed associations may be influenced by gestational weight gain, physiological insulin resistance of pregnancy, exposure misclassification, and variability in GDM diagnostic criteria.

Prior systematic reviews and meta-analyses examining smoking and GDM have generally relied on binary exposure classifications, yielding heterogeneous and sometimes contradictory findings. Early meta-analyses reported either null or inverse associations between smoking during pregnancy and GDM, a phenomenon often referred to as the “smoking paradox” [15,16]. More recent reviews incorporating larger datasets and passive smoking exposure have reported modestly increased risks, particularly among heavier smokers [17].

However, none of these reviews explicitly synthesized evidence according to cessation timing, nor did they distinguish between preconception cessation, early pregnancy cessation, and continued smoking into mid- or late gestation. As a result, potentially opposing risk trajectories were collapsed into single exposure categories, obscuring clinically relevant differences. By contrast, the present review demonstrates that timing-sensitive exposure definitions yield more coherent and biologically plausible patterns, helping to reconcile discrepancies in the existing literature.

Several mechanisms may underlie the observed associations between continued smoking during pregnancy and increased GDM risk. Nicotine exposure has been shown to impair insulin sensitivity, promote systemic inflammation, and induce oxidative stress, all of which may exacerbate the physiological insulin resistance of pregnancy [5,18]. Continued exposure beyond early gestation may interfere with placental development and endothelial function, further contributing to dysregulated glucose homeostasis.

At the same time, smoking is associated with reduced gestational weight gain, which may partially counterbalance insulin resistance in some women and contribute to the paradoxically lower glucose values observed in certain cohorts [5]. Importantly, cessation—particularly in early pregnancy—is frequently accompanied by accelerated weight gain, which may transiently increase glycemic load and obscure the metabolic benefits of stopping smoking if gestational weight gain is not adequately controlled for. Studies that adjusted for prepregnancy BMI and gestational weight gain generally reported attenuation of associations, underscoring the importance of considering these factors as mediators rather than simple confounders.

An important clinical distinction emerging from this synthesis is whether early smoking cessation appears to normalize the risk of gestational diabetes mellitus to that of never-smokers, or whether it merely attenuates an otherwise elevated risk. Overall, the limited available evidence suggests that early cessation—either before conception or during the first trimester—tends to substantially reduce excess GDM risk relative to continued smoking. However, complete risk normalization cannot be consistently demonstrated across studies. In some cohorts, women who quit smoking early exhibited GDM risks comparable to long-term non-smokers, whereas in others, point estimates remained modestly elevated but did not reach statistical significance, likely reflecting limited statistical power rather than true residual risk.

From a clinical perspective, this distinction is meaningful. If early smoking cessation fully normalizes GDM risk, it reinforces the potential for preconception and early-pregnancy interventions to effectively eliminate smoking-related metabolic risk. Conversely, if early cessation primarily attenuates—rather than abolishes—risk, this would suggest that prior smoking exposure or associated metabolic sequelae may exert lingering effects, underscoring the need for continued metabolic surveillance and targeted lifestyle counseling in this group. Given the small number of eligible studies and the imprecision of effect estimates, the present evidence is insufficient to definitively distinguish between these two scenarios. Nevertheless, the consistent contrast between early cessation and continued smoking supports the clinical message that earlier cessation confers progressively greater metabolic benefit, even if complete risk normalization cannot yet be assured.

Notably, the present synthesis does not support complacency regarding smoking during pregnancy based on isolated findings of lower glucose levels or inverse associations. Instead, it highlights that such observations are likely artifacts of exposure misclassification, residual confounding, and complex interactions with gestational weight gain, rather than evidence of a protective effect.

The main strength of this review lies in its timing-focused framework, which allows for a more nuanced interpretation of smoking-related metabolic risk in pregnancy. By including only studies that explicitly define cessation timing, this review minimizes exposure misclassification and enhances comparability across studies.

Importantly, the interpretation of these findings must be situated within the context of a highly limited evidence base. Despite a comprehensive search strategy across multiple databases, only four studies fulfilled the predefined inclusion criteria, reflecting the scarcity of research explicitly examining smoking cessation timing in relation to gestational diabetes mellitus. This small number of eligible studies substantially limits statistical power, restricts the ability to explore heterogeneity or subgroup effects, and precludes quantitative synthesis. Consequently, conclusions drawn from the present review should be considered hypothesis-generating rather than definitive.

4.1. Limitations

This study has several limitations that should be acknowledged. First, the diagnostic criteria used for the identification of gestational diabetes mellitus may have varied across included studies, reflecting differences in clinical guidelines, screening strategies (one-step versus two-step approaches), glucose load, and diagnostic thresholds applied over time and across settings. This heterogeneity may have introduced misclassification bias and limited the direct comparability of study populations. Second, the timing of GDM diagnosis and the degree of glycemic control achieved during pregnancy were not consistently reported, precluding detailed analyses of disease severity or treatment effects on outcomes of interest. Third, residual confounding cannot be excluded, as important maternal factors such as pre-pregnancy body mass index, gestational weight gain, ethnicity, socioeconomic status, and coexisting metabolic conditions were incompletely adjusted for in several studies. Finally, most available data were derived from observational study designs, which limits causal inference and may be subject to selection bias and unmeasured confounding. These limitations should be considered when interpreting the findings, and future well-designed prospective studies using standardized diagnostic criteria and comprehensive confounder adjustment are warranted to confirm and extend the present results.

4.2. Future directions

Future research in this field should prioritize prospective study designs that allow for repeated, time-resolved assessment of smoking behavior across the preconception period and throughout pregnancy, rather than relying on single or retrospective exposure measurements. Repeated assessments would enable more precise classification of smoking cessation trajectories, including preconception cessation, early pregnancy cessation, relapse, and continued smoking into mid- and late gestation, thereby reducing exposure misclassification and recall bias.

Wherever feasible, biochemical validation of smoking status—using biomarkers such as cotinine or other nicotine metabolites—should complement self-reported data, particularly in early pregnancy when underreporting may be common. Incorporation of objective exposure measures would strengthen causal inference and improve comparability across studies.

In addition, future analyses should explicitly model gestational weight gain as a potential mediator rather than solely as a confounder in the association between smoking cessation and gestational diabetes mellitus. Advanced analytical approaches, including mediation and causal pathway modeling, may help disentangle the complex interrelationships among smoking cessation, weight dynamics, insulin resistance, and glucose tolerance during pregnancy.

Finally, adequately powered multicenter cohorts with standardized diagnostic criteria for gestational diabetes mellitus and harmonized definitions of smoking cessation timing are needed to facilitate quantitative synthesis and subgroup analyses according to baseline metabolic risk, body mass index, and sociodemographic characteristics. Such efforts would substantially strengthen the evidence base and support the development of more precise, timing-sensitive clinical recommendations for smoking cessation in pregnancy.

5. Conclusions

In conclusion, the available evidence indicates that smoking cessation timing plays a critical role in shaping the risk of GDM. Continued smoking beyond early pregnancy is associated with increased GDM risk, whereas cessation before or during early pregnancy appears to attenuate this risk. These findings refine current understanding of the smoking–GDM relationship and underscore the need for future studies that prospectively capture cessation trajectories, incorporate biochemical validation, and explicitly model gestational weight gain as a mediator. A timing-sensitive approach to smoking exposure should be incorporated into both future epidemiological research and clinical counseling strategies aimed at reducing the burden of gestational metabolic disease.

CRediT authorship contribution statement

Areti Tsirozoglou: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Resources, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Vikentia Harizopoulou: Writing – review & editing. Angeliki Bolou: Writing – review & editing. Victoria Vivilaki: Writing – review & editing. Ermioni Palaska: Writing – review & editing. Vasiliki Epameinondas Georgakopoulou: Writing – review & editing. Maria Vlachou: Writing – review & editing. Athina Diamanti: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization.