Exploring the impact of smoking on coronary heart disease risk in women: Insights from the NHANES database

Abstract

Coronary heart disease (CHD) is a widespread chronic condition. Its risk factors are numerous and complex, with smoking being a key factor. Recently, CHD risk in women has notably risen, partly due to increased smoking and lifestyle changes. This highlights the critical need for gender-specific CHD research. This study aims to assess CHD risk in smoking and nonsmoking women, identifying crucial biochemical markers influencing this risk. Our goal is to develop personalized risk assessment tools for improved clinical decision-making. We analyzed data from 41,482 female National Health and Nutrition Examination Survey participants (2011–2020), focusing on blood markers. Logistic regression models for smokers and nonsmokers were developed to predict CHD risk, assessed by the area under the curve of the receiver operating characteristic curve. We also created nomograms to translate biochemical indicator measurements into CHD risk probabilities, supporting clinical decisions. Univariate analysis showed significant correlations between age, biochemical markers, and CHD risk. The logistic regression models were highly predictive, with area under the curves of smoking CHD model and nonsmoking CHD model being 0.813 (95% confidence interval: 0.788–0.837) and 0.829 (95% confidence interval: 0.811–0.847), respectively. The nomograms effectively assessed risk across patient groups, confirmed by accurate calibration curves. This study presents distinct CHD risk assessment models for smoking and nonsmoking women, along with an innovative visual risk assessment tool. These insights underscore the role of gender in CHD risk and inform future public health strategies and clinical practices.

1. Introduction

Globally, coronary heart disease (CHD) is a common chronic illness. The evolving societal and lifestyle changes have led to shifting epidemiological characteristics of CHD, with a growing focus on gender-related impacts. Research indicates a critical role of gender in CHD risk assessment. For example, a study^[1] revealed that continuous night shift work increases CHD incidence in women, likely due to unique physiological and psychological responses. Additionally, study^[2] noted increased sensitivity to psychosocial risk factors in female CHD patients, indicating a stronger gender-CHD risk link in the context of psychosocial factors.

Traditionally, men have been seen as the primary high-risk group for CHD,^[3,4] attributed to male-specific physiology, hormonal differences, and lifestyle factors like smoking and poor diet.^[5] However, societal changes have significantly altered women’s lifestyles, including increased work stress, poor diets, and lack of exercise, elevating their CHD risk.^[6] The rising prevalence of smoking among women^[7] further complicates their cardiovascular health challenges.

While smoking is a well-known CHD risk factor,^[8,9] the exact biological mechanisms it triggers in women remain unclear. Investigating blood biochemical indicators is a powerful approach in this area. Analyzing lipid profiles, inflammatory markers, and blood glucose provides insights into metabolic and inflammatory states affecting cardiovascular health.^[10,11] Studies^[12–14] suggest that smoking elevates CHD risk by promoting low-density lipoprotein oxidation and impairing high-density lipoprotein (HDL) function. Furthermore, cigarette smoke chemicals can cause oxidative stress, impairing endothelial function.^[15] This impairs nitric oxide bioactivity and the endothelial cells’ anti-inflammatory role, heightening arteriosclerosis risk. In women, smoking’s effect on blood biochemical markers may vary due to sex hormones’ modulating effects. Estrogen is thought to have vascular protective effects that might mitigate smoking’s adverse impacts on women.^[16,17] However, this estrogen-related protection likely diminishes postmenopause.^[18,19]

Therefore, this study aims to leverage National Health and Nutrition Examination Survey (NHANES) data to create 2 independent CHD risk assessment models for smoking and nonsmoking women. We seek to analyze the relationship between smokers’ blood biochemical markers and CHD risk, and identify risk indicators for nonsmokers, offering tailored risk assessments for both populations. This research will elucidate smoking’s biological impact on women’s cardiovascular health and inform targeted prevention strategies. It emphasizes the need to consider smoking habits in research on women’s cardiovascular health.

2. Research design and methods

2.1. Data source and preparation

The NHANES dataset (2011–2020) (https://wwwn.cdc.gov/Nchs/Nhanes/) was used, offering extensive insights into U.S. adults’ and children’s health and nutrition. A total of 41,482 female participants were included, ensuring a broad representation to enhance the study’s generalizability. Females across all age groups were included, with no exclusions. Forty-five key biomarkers pertinent to cardiovascular health were strategically selected. These included HDL, triglycerides (TRIGLY), cholesterol (CHOL), and other markers such as complete blood count, cytomegalovirus, and C-reactive protein. Any samples with missing values were removed from the analysis to ensure the reliability of the dataset. None of the selected variables had a significant proportion of missing values, and the proportion of samples removed due to missing data was relatively low. After cleaning, the dataset had 14,843 variables with a 64% retention rate. Subsequently, the sample was divided into smokers (n = 6088) and nonsmokers (n = 8755), based on the criterion of “Smoked at least 100 cigarettes in life” being “yes” or “no,” to aid further analysis.

2.2. Univariate analysis

Univariate analyses were performed separately for smoking and nonsmoking groups to assess the relationship between blood biochemical indicators and CHD risk. The outcome variable was defined as “Ever told you had coronary heart disease”= “yes.” We focused on indicators with statistical significance (P < .05).

2.3. Logistic regression model construction

Logistic regression was selected as the modeling approach due to its high interpretability and clinical relevance. The datasets were divided into training and testing sets (7:3 ratio) using “caret” R package. The training set comprised 4262 smokers and 6129 nonsmokers, while the testing set included 1826 smokers and 2626 nonsmokers. Logistic regression models, namely the nonsmoking CHD model and smoking CHD model, were developed to explore the effects of blood biochemical indicators on CHD risk. These models were tested using the respective testing sets.

Model performance was assessed using the area under the receiver operating characteristic (ROC) curve (AUC), measuring the ability to distinguish event occurrence. The “pROC” package in R was utilized to generate ROC curves, visually demonstrating the models’ discriminatory capabilities. The optimal cutoff threshold for classification was determined using Youden’s index (J = sensitivity + specificity − 1), which identifies the point that maximizes the trade-off between sensitivity and specificity. Precision, recall, accuracy, and the Matthews correlation coefficient (MCC) were calculated to provide a comprehensive assessment of model performance at the optimal cutoff point. Confusion matrices, generated using the “caret” R package, assessed the predictive accuracy of the models on the testing sets. To further assess the clinical utility of the models, decision curve analysis (DCA) was performed using the “rmda” package in R. Additionally, ten-fold cross-validation was conducted to evaluate the robustness and generalizability of the models.

2.4. Nomogram construction and validation

Nomograms were created from the logistic regression results, visually depicting each predictor’s contribution to CHD risk. The nomograms’ performance was evaluated at various probability thresholds, crucial for gauging their applicability in different clinical scenarios. To validate the predictive accuracy of the nomograms, calibration curve analysis was performed using the “rms” R package. Calibration curves illustrate the consistency between the model-predicted risk probabilities and actual event probabilities. Ideally, a calibration curve aligns closely with a 45-degree diagonal line, indicating high consistency between predicted and actual probabilities.

2.5. Statistical analysis

Univariate analyses were conducted using t-tests, chi-square tests, or nonparametric tests, depending on the distribution characteristics of the data. Continuous variables were compared using t-tests or nonparametric tests, while categorical variables were compared using chi-square tests. A P-value of <.05 was considered statistically significant. Effect sizes and confidence intervals were reported for a complete interpretation of the results. All analyses utilized R software (https://www.R-project.org/) and its associated packages.

2.6. Availability of data

Data supporting this study’s findings are available from NHANES, a National Center for Health Statistics program. NHANES data are publicly accessible via the CDC website. Researchers may apply for and, upon approval, download the datasets for analysis. This study utilized data from NHANES survey cycles between 2011 and 2020. As the supporting data comes from public database (NHANES), an ethics committee or institutional review board is no necessary need in the study.

3. Results

3.1. Baseline characteristics

A descriptive analysis was conducted on the baseline characteristics of both smoker and nonsmoker populations (refer to Tables S1 and S2, Supplemental Digital Content, https://links.lww.com/MD/P395). Most participants reported no history of CHD. Blood biochemical indicators, including basophil count, mean hemoglobin concentration, and CHOL, showed no significant differences between the training and testing sets across both smokers and nonsmokers. Statistically, the training and testing sets are comparable in baseline characteristics.

3.2. Results of univariate analysis

For smoking women, age emerged as a key variable with a significant correlation, evidenced by its low P-value (<.0001) despite a lower F-value (163.43) compared to the nonsmokers. LYM, monocyte count (MONO), and total protein (TP) in smokers showed high significance (P < .0001), reflecting the potential impact of smoking on these biomarkers. Uric acid (UA) and total bilirubin demonstrated statistical significance (P < .05). Conversely, direct HDL and CHOL were not significant in smokers (P > .05), suggesting a weaker association with smoking. Refer to Table 1 for details.

Table 1.

Results of univariate analysis of smoking population.

	Sum sq	Mean sq	F value	Pr (>F)
Age	9.122813	9.122813	163.4293	9.62E‐37
Albumin to creatinine ratio	0.015152	0.015152	0.271444	0.602392
Mean hemoglobin concentration	0.617057	0.617057	11.05418	0.000892
White blood cell count	0.578544	0.578544	10.36425	0.001295
Red blood cell count	0.29144	0.29144	5.220967	0.022366
Hemoglobin	0.010393	0.010393	0.186178	0.666139
Average cell volume	0.03471	0.03471	0.621816	0.430418
Mean cell hemoglobin	0.949433	0.949433	17.00848	3.79E‐05
Red blood cell distribution width	0.144387	0.144387	2.586592	0.107847
Platelet count	1.403789	1.403789	25.14799	5.53E‐07
Mean platelet volume	0.294765	0.294765	5.28052	0.021614
Lymphocyte percentage	0.096076	0.096076	1.721138	0.189618
Monocyte percentage	0.34777	0.34777	6.23008	0.012598
Lymphocyte count	2.464025	2.464025	44.14143	3.44E‐11
Number of monocytes	1.905193	1.905193	34.13032	5.54E‐09
Segmented neutrophil count	0.491919	0.491919	8.812417	0.003009
Eosinophil count	0.014775	0.014775	0.264691	0.606943
Basophil count	1.717885	1.717885	30.7748	3.08E‐08
Glycated hemoglobin	0.003286	0.003286	0.058875	0.808295
Direct HDL Cholesterol	0.029462	0.029462	0.527787	0.467579
Total cholesterol	0.128814	0.128814	2.307629	0.128815
Albumin	0.096237	0.096237	1.724024	0.189247
Sodium	0.270752	0.270752	4.85034	0.027695
Chloride	0.017095	0.017095	0.306251	0.580019
Creatinine	0.213543	0.213543	3.825488	0.050545
Serum blood glucose	0.148788	0.148788	2.665437	0.102624
Bicarbonates	0.806406	0.806406	14.44624	0.000146
Blood urea nitrogen	0.359334	0.359334	6.43723	0.011211
Phosphorus	0.080637	0.080637	1.444566	0.229469
Osmotic pressure	0.340819	0.340819	6.10555	0.013515
Glutamyl transferase	0.083631	0.083631	1.498198	0.221017
Alkaline phosphatase	1.070204	1.070204	19.17202	1.22E‐05
Cholesterol	0.098196	0.098196	1.759124	0.184806
Alanine aminotransferase	0.129976	0.129976	2.328442	0.127104
Total bilirubin	0.452764	0.452764	8.110983	0.004421
Uric acid	1.334458	1.334458	23.90596	1.05E‐06
Potassium	0.040022	0.040022	0.71697	0.397188
Total calcium	1.182489	1.182489	21.18354	4.30E‐06
Triglycerides	0.003198	0.003198	0.057298	0.810831
Total protein	1.679802	1.679802	30.09258	4.36E‐08
Total calcium	0.046908	0.046908	0.84033	0.359355
Aspartate aminotransferase	1.30E-05	1.30E-05	0.000232	0.987835
Lactate dehydrogenase	0.160822	0.160822	2.881027	0.089703

For nonsmoking women, age displayed the most significant correlation (F = 263.05, P < .0001), underscoring its strong association with the biomarkers. Lymphocyte percentage (LYM%) and glycated hemoglobin showed highly significant correlations (P < .0001) in nonsmokers, indicating their potential importance. Other blood parameters including total white blood cell count (WBC), platelet count (PLT), and alkaline phosphatase (ALP) were also significant (P < .05) in nonsmokers, highlighting their importance. Conversely, mean cell hemoglobin and total bilirubin were not significant (P > .05) in nonsmokers, indicating a weaker correlation with the study’s objectives. Refer to Table 2 for details.

Table 2.

Results of univariate analysis of nonsmoking population.

	Sum sq	Mean sq	F value	Pr (>F)
Age	19.5083	19.5083	263.0526	6.01E‐58
Albumin to creatinine ratio	0.872676	0.872676	11.76728	0.000607
Mean hemoglobin concentration	1.603129	1.603129	21.61681	3.40E‐06
White blood cell count	6.311369	6.311369	85.10334	3.83E‐20
Red blood cell count	2.95898	2.95898	39.89928	2.86E‐10
Hemoglobin	0.071316	0.071316	0.961633	0.326815
Average cell volume	2.244102	2.244102	30.25977	3.93E‐08
Mean cell hemoglobin	7.43E-05	7.43E-05	0.001002	0.974752
Red blood cell distribution width	0.247873	0.247873	3.342349	0.067567
Platelet count	5.051146	5.051146	68.11034	1.88E‐16
Mean platelet volume	0.855773	0.855773	11.53935	0.000686
Lymphocyte percentage	11.80504	11.80504	159.1807	4.82E‐36
Monocyte percentage	1.615449	1.615449	21.78294	3.12E‐06
Lymphocyte count	0.128034	0.128034	1.726434	0.188917
Number of monocytes	0.436318	0.436318	5.883373	0.015313
Segmented neutrophil count	2.870117	2.870117	38.70104	5.27E‐10
Eosinophil count	0.153204	0.153204	2.065828	0.150684
Basophil count	0.001746	0.001746	0.023541	0.878064
Glycated hemoglobin	7.633002	7.633002	102.9244	5.40E‐24
Direct HDL Cholesterol	0.832772	0.832772	11.22921	0.00081
Total cholesterol	0.005307	0.005307	0.071565	0.789081
Albumin	0.147755	0.147755	1.992346	0.158147
Sodium	0.424433	0.424433	5.723116	0.016773
Chloride	1.308896	1.308896	17.64932	2.69E‐05
Creatinine	0.326684	0.326684	4.40505	0.035874
Serum blood glucose	0.58812	0.58812	7.930293	0.004877
Bicarbonates	1.208333	1.208333	16.29333	5.49E‐05
Blood urea nitrogen	0.185078	0.185078	2.495619	0.114215
Phosphorus	1.108655	1.108655	14.94925	0.000112
Osmotic pressure	0.368237	0.368237	4.965351	0.025896
Glutamyl transferase	0.670671	0.670671	9.043418	0.002647
Alkaline phosphatase	2.446359	2.446359	32.98704	9.73E‐09
Cholesterol	1.174559	1.174559	15.83791	6.98E‐05
Alanine aminotransferase	0.128059	0.128059	1.72676	0.188875
Total bilirubin	0.247605	0.247605	3.338741	0.067715
Uric acid	0.118852	0.118852	1.602612	0.205582
Potassium	0.158474	0.158474	2.13688	0.143846
Total calcium	0.001583	0.001583	0.021348	0.88384
Triglycerides	1.669231	1.669231	22.50813	2.14E‐06
Total protein	1.976456	1.976456	26.6508	2.51E‐07
Total calcium	0.220223	0.220223	2.969518	0.084898
Aspartate aminotransferase	0.329387	0.329387	4.441493	0.035116
Lactate dehydrogenase	0.92848	0.92848	12.51974	0.000406

3.3. Smoking CHD model

In the study of the smoking population, the logistic regression model indicated significant correlations between biochemical indicators such as age, MONO, monocyte percentage (MON%), eosinophil count, PLT, total WBC, segmented neutrophil count, LYM, ALP, UA, and bicarbonate levels (HCO3), and the CHD risk (Fig. 1A). Hazard ratios and 95% confidence intervals for these indicators are detailed in the chart, highlighting age and segmented neutrophil count as strongly correlated with elevated CHD risk.

Figure 1.

Construction and validation of smoking CHD model. (A) Forest plot showing significant correlations between various biochemical indicators and the CHD risk in the smoking women. (B) Confusion matrix illustrating the predictive accuracy of the CHD model in smoking women. (C) ROC curve for the training set of the smoking CHD model. (D) ROC curve for the testing set of the smoking CHD model. (E) DCA of the smoking CHD model. (F) Ten-fold cross-validation of the smoking CHD model. CHD = coronary heart disease, DCA = decision curve analysis, ROC = receiver operating characteristic.

The confusion matrix (Fig. 1B) for the testing set indicated a 93.66% accuracy in CHD prediction, comprising 52 true positives and 111 false positives. The ROC curve revealed an AUC of 0.813 (95% confidence interval [CI]: 0.788–0.837) in the training set (Fig. 1C) and 0.739 (95% CI: 0.690–0.781) in the testing set (Fig. 1D), suggesting better predictive performance in the training set. The optimal cutoff value was 0.101, at which point the model exhibited a precision of 0.319, a recall of 0.8125, an accuracy of 93.66%, and a MCC of 0.480, suggesting satisfactory sensitivity and moderate overall classification performance. DCA further demonstrated that the model offered a positive net clinical benefit across a wide range of threshold probabilities (Fig. 1E). Moreover, ten-fold cross-validation confirmed the model’s stability and robustness (Fig. 1F).

3.4. Nonsmoking CHD model

For nonsmokers, a similar logistic regression model evaluated the correlation between multiple biochemical indicators, including age, LYM%, TRIGLY, phosphorus, HCO3, ALP, HbA1c, total WBC, mean platelet volume, chloride (Cl), PLT, TP, and CHOL, and CHD risk (Fig. 2A). Notably, age, phosphorus, and ALP were significantly positively correlated with increased CHD risk.

Figure 2.

Construction and validation of nonsmoking CHD model. (A) Forest plot showing significant correlations between various biochemical indicators and the CHD risk in the nonsmoking women. (B) Confusion matrix illustrating the predictive accuracy of the CHD model in nonsmoking women. (C) ROC curve for the training set of the nonsmoking CHD model. (D) ROC curve for the testing set of the nonsmoking CHD model. (E) DCA of the nonsmoking CHD model. (F) Ten-fold cross-validation of the nonsmoking CHD model. CHD = coronary heart disease, DCA = decision curve analysis, ROC = receiver operating characteristic.

The confusion matrix (Fig. 2B) for the testing set indicated a 91.41% accuracy in CHD prediction, with 73 true positives and 214 false positives. The ROC curve indicated AUCs of 0.829 (95% CI: 0.811–0.847) for the training set (Fig. 2C) and 0.817 (95% CI: 0.789–0.860) for the testing set (Fig. 2D), demonstrating the model’s strong and consistent predictive capability. The optimal cutoff value was 0.134, at which point the model exhibited a precision of 0.254, a recall of 0.709, an accuracy of 91.41%, and a MCC of 0.391, suggesting reliable discriminative ability with moderate predictive balance. DCA confirmed the clinical utility of the model, revealing a positive net benefit across a wide range of threshold probabilities (Fig. 2E). Moreover, 10-fold cross-validation confirmed the model’s stability and robustness (Fig. 2F).

3.5. Nomogram models for disease risk estimation in different populations

Nomogram models were developed to convert biochemical indicators into a composite CHD risk score. Each indicator’s value is assigned points, weighted by their impact on CHD risk. To use the nomogram, one identifies the points for each indicator’s measured value and sums them to obtain a total score. This total score correlates with a specific CHD risk, representing the patient’s likelihood of developing CHD within a defined period.

The smoking population’s nomogram model includes indicators like age, MONO, and MON%, with scores reflecting their CHD risk contribution (Fig. 3A). For instance, an older smoking female with high scores on various indicators could have a total score up to 150 points, equating to around a 50% CHD risk. The nonsmoking population’s nomogram model considers indicators such as age, LYM%, and TRIGLY (Fig. 4A). For instance, a young nonsmoking female with low scores on all indicators could total 30 points, indicating less than a 1% CHD risk. Calibration curve charts (Figs. 3B and 4B) for both models compare predicted and actual risks. The Ideal line represents perfect prediction, while the Apparent and Bias-corrected lines indicate the original and adjusted predictions’ fit, reflecting the models’ clinical reliability and accuracy.

Figure 3.

Construction and validation of visualization model for smoking women. (A) Nomogram for smoking women. (B) Calibration curve for the nomogram of smoking women.

Figure 4.

Construction and validation of visualization model for nonsmoking women. (A) Nomogram for the nonsmoking women. (B) Calibration curve for the nomogram of nonsmoking women.

4. Discussion

The rising prevalence of smoking among women has escalated the risk of CHD, posing a significant public health challenge. This study effectively developed 2 distinct models to evaluate CHD risk in smoking and nonsmoking women. Furthermore, this study innovatively visualized complex cardiovascular disease risk factor data via nomograms. This visualization enhances data interpretation clarity and enables swift, accurate CHD risk assessments by healthcare professionals.

This research is pioneering in systematically examining independent CHD risk factors in female smokers and nonsmokers, an area previously understudied. Additionally, we developed dynamic risk assessment tools, namely nomograms, which allow for both static assessment and dynamic adjustments based on new patient data, facilitating personalized risk prediction. Moreover, our study highlighted the role of blood biochemical indicators in assessing CHD risk, especially among smoking and nonsmoking females. Our analysis indicates that initial changes in biochemical indicators can act as early warning signs for cardiovascular disease, vital for clinicians.

This study identified age as a significant factor influencing CHD risk in both smoking and nonsmoking women. This aligns with studies^[20,21] that emphasize age’s role in elevating cardiovascular disease risks, including arteriosclerosis and vascular function decline. Additionally, our study found a significant correlation between CHD risk and increased ALP levels, corroborated by other research.^[22] ALP, a membrane-bound enzyme with broad substrate specificity, catalyzes the hydrolysis of various phosphomonoesters. Its activity is associated with bone health^[23] and liver function.^[24] This suggests a link between CHD risk and liver function, bone metabolism, and inflammation levels. The study also underscored the importance of HCO3 levels in CHD risk assessment. In line with other studies,^[25] elevated HCO3 levels correlate with improved cardiovascular risk factors and serum nutritional markers. Oral sodium HCO3 treatment can slow kidney function decline and potentially enhance endothelial function in CKD patients, reducing cardiovascular events.^[26] This highlights the crucial role of acid-base balance in cardiovascular health, as imbalances can significantly impact cardiovascular disease risk. Increased total WBC and PLT changes in both groups reaffirm the central role of inflammation and blood coagulation in cardiovascular events. Aligning with research by Mohammed et al.^[27] and Shah et al.,^[28] these findings show a correlation between elevated WBC and a higher risk of major adverse cardiovascular events. Furthermore, Foy et al^[29] highlight the significance of fluctuations in white blood cells and PLT during recovery from acute inflammation. Our study also identified a significant link between TP levels and CHD risk. TP levels, indicating the blood’s protein content, reflect nutritional status and serve as a biomarker for chronic inflammation and other disease states. Studies^[30,31] have suggested hypoproteinemia, linked with malnutrition, liver disease, or renal dysfunction, correlates with a heightened risk of cardiovascular diseases. Moreover, changes in TP levels may relate to chronic inflammation, potentially impairing endothelial function and accelerating atherosclerosis, increasing CHD risk.^[32] These common risk factors imply that besides smoking as a crucial lifestyle factor, other biological mechanisms also impact cardiovascular health.

This study significantly contributes by elucidating the fundamental differences in CHD risk factors between smoking and nonsmoking female populations. In smoking women, an increase in MONO and MON% was noted, indicative of an intensified inflammatory response. Previous research^[33] has confirmed that smoking exacerbates atherosclerosis and other cardiovascular diseases by enhancing the activity of inflammatory cells like monocytes. The elevated UA levels in smokers are noteworthy. AU, produced during purine metabolism, is widely recognized as an indicator of increased cardiovascular diseases risk. However, studies^[34] indicate that high UA levels may not directly harm endothelial function or hasten arteriosclerosis. This association might may arise from high UA levels’ association with various cardiovascular disease risk factors like hypertension, insulin resistance, and inflammation.^[35] This implies a potential connection between smoking and metabolic pathways like UA, thus elevating CHD risk. Moreover, a significant rise in eosinophil count was observed in smokers. Eosinophils, a type of white blood cell, are crucial in immune responses, especially in allergic reactions and immune regulation.^[36] Smoking notably impacts the immune system, including the distribution and function of immune cells. Smoking may enhance the activity of immune cells like eosinophils, altering the body’s inflammatory and infection responses. Such changes could result in an overactive or abnormal immune response, increasing CHD risk. Therefore, this underscores the need to consider immune regulatory mechanisms in evaluating smoking’s impact on cardiovascular health.

Conversely, the roles of HbA1c, LYM%, and TRIGLY in nonsmoking women highlight distinct metabolic and immune pathways in cardiovascular disease development, independent of smoking. HbA1c, a long-term blood sugar marker, is crucial in evaluating cardiovascular disease risk, particularly in diabetes or metabolic syndrome. Its significance likely reflects the link between poor blood sugar control and elevated cardiovascular disease risk.^[37,38] Its significance might reflect the connection between poor blood sugar control and increased risk of cardiovascular diseases. Moreover, LYM% variations may signify the immune system’s role in cardiovascular disease development, especially regarding chronic inflammation and immune regulation.^[39,40] Elevated TRIGLY underscore lipid metabolism’s role in cardiovascular disease risk, likely from increased arterial wall lipid accumulation, heightening arteriosclerosis and cardiovascular event risks.^[41,42] These findings indicate that in nonsmoking women, cardiovascular disease risk may be impacted by diverse metabolic and immune factors, underscoring the need to consider these elements in risk assessment.

While our study provides offers valuable insights into CHD risks in smoking and nonsmoking females, it is not without limitations. Firstly, being a cross-sectional study, our findings cannot establish causality; the association between smoking and biochemical indicator changes does not necessarily imply direct causation. Furthermore, unconsidered variables or confounders like lifestyle, genetic predispositions, or socioeconomic status might have influenced the analysis. Thus, future research should incorporate precise, long-term follow-ups in a broader demographic to delve deeper into the complex relationship between smoking and cardiovascular health.

In summary, this study systematically uncovers the CHD risk factor differences between female smokers and nonsmokers, bridging a gap in existing research. This not only enriches our understanding of cardiovascular disease risk factors but also offers significant insights for future research and clinical practice. The findings contribute crucial information for gender-specific CHD prevention and intervention strategies, emphasizing the importance of gender considerations in public health policy formulation.

Author contributions

Conceptualization: Jun Xia.

Data curation: Jun Xia.

Methodology: Yang Mu.

Project administration: Jun Xia.

Software: Yang Mu.

Supervision: Jun Xia.

Resources: Yang Mu.

Visualization: Yang Mu.

Writing – original draft: Yang Mu.

Writing – review & editing: Jun Xia.