Gender-Specific Lifestyle Risk Behaviors and Machine Learning Models for Predicting Atherosclerotic Cardiovascular Disease

doi:10.21203/rs.3.rs-4838697/v1

Download PDF

Article

Gender-Specific Lifestyle Risk Behaviors and Machine Learning Models for Predicting Atherosclerotic Cardiovascular Disease

https://doi.org/10.21203/rs.3.rs-4838697/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Atherosclerotic cardiovascular disease (ASCVD) is a leading cause of global morbidity and mortality. We aimed to assess the predictive accuracy of machine learning (ML) models incorporating lifestyle risk behaviors for ASCVD risk stratified by gender, using data from the Korea National Health and Nutrition Examination Survey. We analyzed data from 8,573 participants aged 40–79 years, excluding those with prior cardiovascular events. ASCVD risk was assessed using the American College of Cardiology/American Heart Association Pooled Cohort Equations, with a high-risk threshold of ≥ 15% over 10 years. Five ML algorithms—logistic regression (LR), support vector machine, random forest, extreme gradient boost, and light gradient boosting models—were utilized, with performance metrics including AUROC, accuracy, precision, recall, and F1 score. Among men, the support vector machine model achieved the highest AUROC of 0.952, whereas, among women, the LR model achieved the highest AUROC of 0.980. Significant predictors for men included age, smoking, BMI, and LDL cholesterol, while for women, predictors extended to household income and residential area. Comparing the significant the Shapley additive explanation variables in the ML model to the significant variables in the conventional bivariate LR model, lifestyle risk behaviors such as household income, residential area, and weight change over 1 year were identified as significant variables in both models. This analysis provides the importance of gender-specific lifestyle risk factors in ASCVD prediction. The integration of ML and lifestyle factors offers enhanced predictive capabilities over traditional models, highlighting the necessity for tailored prevention strategies.

Health sciences/Medical research

Health sciences/Medical research/Epidemiology

Atherosclerotic Cardiovascular Disease

Machine Learning

Lifestyle Risk Behaviors

Gender Differences

Predictive Modeling

Atherosclerotic cardiovascular disease (ASCVD) is the leading cause of morbidity and mortality worldwide¹. In 2019, cardiovascular diseases (CVDs) were responsible for an estimated 32% of all global deaths². Additionally, out of the 17 million premature deaths caused by non-communicable diseases in 2019, 38% were due to CVDs². The prevalence of CVDs in South Korea is estimated to be increasing recently³. While the CVD mortality rate declined until 2010, it has been steadily rising since then⁴. Considering these trends, addressing CVDs is crucial for improving public health outcomes.

The main risk factors for developing ASCVD include hypertension, smoking, and dyslipidemia⁵. Additionally, ASCVD outcomes vary across different ethnic groups based on socioeconomic status, which also influences lifestyles according to gender^6,7. Biological differences between gender also lead to different impacts on CVD risk factor profile and behaviors⁸. As the prevalence of ASCVD is increasing, the importance of primary prevention has been emphasized from a public health perspective⁹. ASCVDs are closely related with modifiable risk factors that can be controlled through lifestyle modification, which is the cornerstone of ASCVD prevention⁹. The most effective prevention strategy is a comprehensive approach that promotes a healthy lifestyle and addresses all major risk factors¹⁰.

Machine learning (ML) has been introduced to the field of medical data analysis, offering numerous benefits that enhance patient care and improve clinical outcomes. ML-based prediction using patients’ data such as demographics, clinical and laboratory data, and image data can identify patient future outcomes that might be missed by conventional prediction model¹¹. Additionally, ML enables to plan for the precision medicine, early detection of high-risk patients for the disease progression with a high accuracy¹¹. Previous studies have constructed ML model for predicting ASCVD, but few have modeled the impact of lifestyle behavior on ASCVD risk and assessed sex-specific model^11–13. ASCVD risk is associated with a variety of demographic factors, including age, race, and gender, as well as a patient's lifestyle behaviors, and these lifestyle risk behaviors differ significantly by gender¹⁴. Thus, our objective was to evaluate how accurately a prediction model that incorporates these lifestyle risk behaviors in predicting ASCVD risk by gender.

Participants

This study used nationally representative data from the Korea National Health and Nutrition Examination Survey (KNHANES). The KNHANES is a nationwide cross-sectional survey initiated by the Korea Disease Control and Prevention Agency in 1998 to assess the health and nutritional status of the Korean population. Among the 22,889 participants in the 2019–2021 KNHANES, 10,481 participants 40–79 years of age with no cardiovascular event were selected. Subsequently, 1,908 participants with one or more missing data on sodium intake, low-density lipoprotein (LDL)-cholesterol, time spent sitting, occupation, marital status, smoking, income, body mass index (BMI), and weight change over one year were excluded. A total of 8,573 participants were finally selected for analysis (Fig. 1).

Measurement of ASCVD risk score

The 10-year ASCVD risk was assessed using the American College of Cardiology/American Heart Association Pooled Cohort Equations (PCEs) risk score¹⁵. They recommended that PCEs for non-Hispanic whites may be considered for risk estimation in populations other than non-Hispanic African Americans and non-Hispanic whites¹⁵. Therefore, the estimation function of the ASCVD risk score for the white population was used in this study. This score algorithm is designed to predict both the 10-year (40–79 years of age) and lifetime (20–59 years of age) risk of ASCVD. This includes the assessment of both fatal and non-fatal coronary heart disease and stroke¹⁶. However, PCEs tend to overestimate 10-year ASCVD risk in contemporary lower-risk populations, which also may lead to overtreatment¹⁷. Therefore, we used a high-risk score for the risk stratification, defined as an ASCVD risk score of ≥ 15% over 10 years¹⁸.

Measurement of lifestyle risk behaviors

The lifestyle risk behaviors analyzed in this study included sex, age, occupation, marital status, residential area, household income, alcohol consumption, smoking, omega-3 intake, sodium intake, BMI, weight change over 1 year, LDL cholesterol, physical activity, and time spent sitting. The occupation was classified into three categories of non-manual workers, manual workers, and other workers. Marital status was separated, widowed, and divorced participants were assigned the no spouse status. Residential area was grouped as urban or rural areas. Alcohol consumption was divided into two categories, where participants drank at least 2 days per week and fewer than 2 days per week¹⁹, regardless of the amount of alcohol consumed. Individuals who had smoked at least 100 cigarettes in their lifetime and currently smoked were classified as smokers, whereas those who had smoked fewer than 100 cigarettes in their lifetime but did not currently smoke and never smoked in their lifetime were classified as non-smokers. BMI was calculated as weight in kilograms divided by height in meters squared. Weight change over 1 year was classified into three categories of no weight change, weight gain, and weight loss. Using the criteria specified in the questionnaire, weight gains and losses of 0 to less than 3 kg over 1 year were classified as no change, gains of more than 3 kg as weight gain, and losses of more than 3 kg as weight loss. Physical activity was calculated using the sum of the minutes of exercise per week and considering the strength of the exercise intensity (vigorous and moderate). The combination of vigorous intensity and moderate intensity exercise was considered by calculating the minutes exercised per week as follows: 2 × moderate activity (min/week) + vigorous activity (min/week) ²⁰. Physical activity was used as a continuous variable. We used omega-3 intake in grams per day, sodium intake in milligrams per day, and time spent sitting in minutes per day, and these variables were entered into the ML model as a continuous variable.

Statistical analysis

In the descriptive analyses, continuous variables were presented as mean ± standard deviation, and categorical variables were presented as frequencies (%). Differences between high and low ASCVD groups were compared using the t-test or Mann-Whitney U Test for continuous variables after normality test and the chi-square test for categorical variables. Multivariable logistic regression was performed to estimate the adjusted odds ratio (OR) and 95% confidence interval (CI) to investigate the association of lifestyle factors with ASCVD risk.

We used five ML algorithms: logistic regression (LR), support vector machine (SVM), random forest (RF), eXtreme gradient boost (XGB), and light gradient boosting (LGB) model for the prediction of high ASCVD risk. We divided the dataset into a training dataset and a test dataset in an 8:2 ratio. Then, the training dataset was divided 5-fold, and cross-validation was performed by composing the same ratio of classes in the divided dataset. The ML model was optimized using Optuna, a hyperparameter optimization framework based on the random search²¹. Understanding the performance of ML models is becoming increasingly important, which highlights the growing significance of explainable artificial intelligence in interpreting model results. We utilized the SHapley Additive exPlanations value to visualize the importance and relationships among the input features of the model. The area under the receiver operating characteristics curve (AUROC), accuracy, precision, recall, F1 score, specificity, area under the precision-recall curve (AUPRC) were calculated to assess the performance of the developed ML models.

Baseline characteristics according to gender

Of the 8,573 participants, 3,514 (40.99%) were men and 5,059 (59.01%) were women. Table 1 shows a comparison of the characteristics of participants according to men and women. Among both men and women in the study, the high ASCVD group had an older age, lower income, a higher proportion of urban residents, lower physical activity, and lower sodium and omega-3 intake than the low ASCVD group. Weight loss in the past year was higher in the high ASCVD group in both men and women. Among men, the high ASCVD group had a lower proportion of non-manual workers (9.6% vs 36.5%), and higher proportion of current smoking. Among women, the high ASCVD group had a lower proportion living with a spouse (53.8% vs 79.5%), and lower frequency of alcohol intake (2.9% vs. 9.9%).

Table 1

Characteristics of the study population
Variable	Categories	Men (n = 3,514)			Women (n = 5,059)
		ASCVD < 15	ASCVD ≥ 15	P-value	ASCVD < 15	ASCVD ≥ 15	P-value
Age, mean(± SD)		52.4 (± 8.0)	68.4 (± 7.2)	< 0.001	55.4 (± 9.1)	73.9 (± 3.6)	< 0.001
Job	Non-manual workers	821 (36.5%)	122 (9.6%)	< 0.001	921 (21.0%)	7 (1.0%)	< 0.001
	Manual workers	1102 (49.0%)	560 (44.2%)		1542 (35.2%)	225 (33.0%)
	Other workers	324 (14.4%)	585 (46.2%)		1914 (43.7%)	450 (66.0%)
Marital status	With spouse	1891 (84.2%)	1048 (82.7%)	0.267	3478 (79.5%)	367 (53.8%)	< 0.001
	Without spouse	356 (15.8%)	219 (17.3%)		899 (20.5%)	315 (46.2%)
Residential area	Rural	1809 (80.5%)	888 (70.1%)	< 0.001	3530 (80.6%)	464 (68.0%)	< 0.001
	Urban	438 (19.5%)	379 (29.9%)		847 (19.4%)	218 (32.0%)
Household income, 10,000 won/month		542.5 (± 327.1)	305.2 (± 274.9)	< 0.001	477.4 (± 330.4)	186.68 (± 194.5)	< 0.001
Alcohol drinking,	≥ 2/week	813 (36.2%)	445 (35.1%)	0.529	434 (9.9%)	20 (2.9%)	< 0.001
	< 2/week	1434 (63.8%)	822 (64.9%)		3943 (90.1%)	662 (97.1%)
Smoking	yes	697 (31.0%)	447 (35.3%)	0.010	4196 (95.9%)	650 (95.3%)	0.501
	no	1550 (69.0%)	820 (64.7%)		181 (4.1%)	32 (4.7%)
Omega-3 intake, g/day		2.2 (± 2.0)	1.9 (± 2.0)	0.001	1.8 (± 1.8)	1.4 (± 1.5)	0.001
Sodium intake, g/day		4.1 (± 2.1)	3.6 (± 2.0)	0.001	2.8 (± 1.6)	2.4 (± 1.5)	0.001
Body mass index, kg/m²		24.7 (± 3.2)	24.4 (± 3.0)	0.056	23.8 (± 3.5)	24.8 (± 3.4)	< 0.001
Weight change over 1 year	Weight loss	252 (11.2%)	196 (15.5%)	< 0.001	458 (10.5%)	128 (18.8%)	< 0.001
	No weight change	1592 (70.9%)	935 (73.8%)		2764 (63.1%)	464 (68.0%)
	Weight gain	403 (17.9%)	136 (10.7%)		1155 (26.4%)	90 (13.2%)
LDL, mg/dL		116.4 (± 34.3)	108.8 (± 35.6)	0.178	120.7 (± 35.1)	109.2 (± 35.8)	0.001
Physical activity, min/ week		149.04 (± 316.12)	87.85 (± 250.19)	< 0.001	82.44 (± 245.17)	25.16 (± 93.40)	< 0.001
Time spent sitting, min/day		506.8 (± 218.0)	503.3 (± 224.9)	0.418	480.1 (± 205.0)	549.1 (233.83)	0.001
ASCVD, atherosclerotic cardiovascular disease; LDL, low-density lipoprotein.

Performance of machine learning model according to gender

Table 2 shows the performance of the ML model for ASCVD risk by gender. Among 5 ML models, the highest AUROC of 0.952 for men was observed with the SVM model, while the highest AUROC of 0.980 was observed with the LR model in women. (Fig. 2)

Table 2

Predictive model performance for ASCVD risk by gender in the test set
Model	LR		SVM		RF		XGB		LGB
	Men	Women	Men	Women	Men	Women	Men	Women	Men	Women
Accuracy	0.872	0.910	0.872	0.904	0.853	0.902	0.878	0.943	0.886	0.925
Precision	0.788	0.604	0.788	0.587	0.774	0.586	0.915	0.882	0.829	0.674
Recall	0.881	0.963	0.881	0.963	0.838	0.926	0.727	0.662	0.862	0.853
F1 score	0.832	0.742	0.832	0.730	0.805	0.718	0.811	0.756	0.845	0.753
AUROC	0.952	0.979	0.952	0.979	0.929	0.965	0.951	0.977	0.952	0.973
Specificity	0.867	0.902	0.867	0.894	0.862	0.898	0.962	0.986	0.900	0.936
AUPRC	0.928	0.890	0.929	0.887	0.878	0.802	0.921	0.873	0.928	0.858
ASCVD, atherosclerotic cardiovascular disease; LR, logistic regression; SVM, support vector machine; RF, random forest; XGB, extreme gradient boosting; LGB, light gradient boosting; AUROC, area under the receiver operating characteristic curve; AUPRC, area under the precision recall curve.

Variable importance in ML model according to gender

In men, individual-level variables such as age, smoking, BMI, and LDL cholesterol were important predictors of the risk of ASCVD (Fig. 3a and b). On the other hand, in women, individual-level variables such as age, BMI, and smoking, as well as socioeconomic variables such as household income and residential area, were important predictors of the risk of ASCVD (Fig. 3c and d).

Associations of lifestyle risk factors with 10-year ASCVD risk according to gender

We compared how the significant predictors of high ASCVD risk in binary logistic regression correlated with the variable importance identified in the ML model. High ASCVD risk variables in ML models were significantly associated with a higher risk of ASCVD in bivariate LR analysis (Table 3). Age, smoking, BMI, LDL cholesterol, household income were significant factors of predicting high ASCVD risk in conventional LR and ML model for men, simultaneously. While for women, as age increases, the odds ratio is 1.68 (95% confidence interval: 1.60–1.76); as the rurality of the residential area increases, the odds ratio is 1.39 (95% confidence interval: 1.02–1.90); as household income decreases, the odds ratio is 0.999 (95% confidence interval: 0.999–1.000); as a current smoker, the odds ratio is 8.26 (95% confidence interval: 3.95–17.27); and as BMI increases, the odds ratio is 1.13 (95% confidence interval: 1.08–1.18).

Table 3

Binary logistic regression analysis for the association between lifestyle factors and ASCVD.
Variables	Categories	OR (95% CI)
		Men	Women
Age, years		1.42 (1.38–1.46)	1.68 (1.60–1.76)
Job	manual workers	1.19 (0.85–1.67)	1.12 (0.34–3.65)
	other workers	1.41(0.97–2.01)	1.19 (0.37–3.82)
	non-manual workers	1.0 (reference)	1.00 (reference)
Marital status	without spouse	1.37 (0.97–1.94)	0.85 (0.63–1.15)
	with spouse	1.0 (reference)	1.00 (reference)
Residential area	rural	1.15 (0.87–1.53)	1.39 (1.02–1.90)
	urban	1.0 (reference)	1.00 (reference)
Household income, 10,000 won/month		1.00 (1.00–1.00)	1.00 (1.00–1.00)
Alcohol drinking	≥ 2/week	1.03 (0.81–1.32)	0.54 (0.25–1.15)
	< 2/week	1.0 (reference)	1.00 (reference)
Smoking	yes	14.28 (10.39–19.63)	8.26 (3.95–17.27)
	no	1.0 (reference)	1.00 (reference)
Omega-3 intake, g/day		0.97 (0.91–1.03)	0.96 (0.88–1.05)
Sodium intake, mg/day		1.00 (1.00–1.00)	1.00 (1.00–1.00)
Body mass index, kg/m²		1.24 (1.19–1.29)	1.13 (1.08–1.18)
Weight change over 1 year	no weight change	0.72 (0.51–1.02)	0.49 (0.33–0.73)
	weight gain	0.63 (0.39–1.01)	0.49 (0.29–0.83)
	weight loss	1.00 (reference)	1.00 (reference)
LDL cholesterol, mg/dL		1.01 (1.01–1.01)	1.00 (0.99-1.00)
Physical activity, min		1.00 (1.00–1.00)	1.00 (1.00–1.00)
Time spent sitting, min/day		1.00 (1.00–1.00)	1.00 (1.00–1.00)
ASCVD, atherosclerotic cardiovascular disease; LDL, low-density lipoprotein; OR, odds ratio; CI, confidence interval.

This study used nationally representative data to develop ML models predicting high ASCVD risk based on lifestyle risk behaviors, emphasizing significant gender-specific differences in risk factor profile and model performance. Additionally, the variables that were important in the conventional statistical model for predicting high ASCVD risk were simultaneously identified as important variables in the ML model. These results suggest that predicting ASCVD risk should take into account a variety of variables, including lifestyle risk behaviors, and that different strategies should be designed for different genders for accurate prediction and intervention for prevention.

The pathophysiology of atherosclerosis shows different patterns between women and men due to inherent biological and social differences²². Noninvasive cardiovascular diagnostic imaging reveals that men develop plaques earlier and have a greater plaque burden than women, even after accounting for differences in risk factors. Plaque area, rather than the degree of stenosis, predicts adverse ischemic events. This is consistent with the greater incidence of ischemic events in males, although this relationship changes with advancing age²³. Generally, before menopause, women are relatively protected from cardiovascular disease, and then, after menopause, the risk for cardiovascular disease greatly increases in women. The decline of sex hormones has been shown to play an important role in the development of CVD with the onset of advanced age²⁴. Therefore, understanding and considering the differences by gender is important in assessing ASCVD risk factors.

Age is associated with increased oxidative stress, which leads to an increased susceptibility of CVD onset²⁴. Excessive generation of reactive oxygen species leads to a state of oxidative stress which is a major risk factor for the development and progression of atherosclerosis²⁵. Aging is an unmodifiable ASCVD risk factor, but it can be prevented if other factors could be corrected. Cigarette smoking is widely accepted as a major risk factor for the development of clinical CVD resulting from direct effects on atherosclerosis. Epidemiologic studies strongly support that cigarette smoking in both men and women increases the incidence of myocardial infarction and fatal coronary artery disease²⁶. Because the chemical constituents of smoke have high oxidant and inflammatory capacities, they can directly induce endothelial damage and potentiate an inflammatory response²⁶. Smoke is able to increase LDL levels through metabolic alterations and the induction of LDL oxidation due to the direct oxidant capacity of smoke components²⁷. Obesity has been linked to persistent inflammation and oxidative stress. Oxidative stress plays a crucial role in disorders related to obesity, such as dyslipidemia and hypertension, causing cardiovascular diseases²⁸. The prevalence of overweight and obesity has been strongly increasing over the last few decades, and it is considered to be one of the largest challenges for public health work worldwide^28,29.

Our study indicates that LDL cholesterol and moderate physical activity in men, as well as area and income as components of socioeconomic status and omega-3 intake in women, are associated with risk factors for ASCVD. The reason LDL cholesterol is a higher risk factor in men compared to women is because men tend to have higher LDL cholesterol level than women³⁰. Estrogen in premenopausal women can contribute to cardioprotection through several mechanisms, including the maintenance of a health lipoprotein profile. This is evident by the negative correlation of serum estrogen levels with total cholesterol, LDL cholesterol, triglycerides, and VLDL cholesterol while being positively correlated with HDL cholesterol³¹.

The American Heart Association has recently outlined a new framework that focuses on defining and optimizing cardiovascular health through the adoption of 8 simple health components: healthy diet, engaging in regular physical activity, avoidance of nicotine, healthy sleep, and healthy levels of blood lipids and glucose, and blood pressure³². This framework also recognizes that cardiovascular health cannot exist without addressing psychological well-being and social determinants of health. In large-scale meta-analysis with information on more than 1 million incident CVD events, indices of socioeconomic status were inversely associated with CVD risk in both sexes. Besides, the association with coronary heart disease and lower educational year was significantly stronger in women³³. Therefore, a gender-specific approach to ASCVD risk may be important for accurate prediction and prioritization of treatment strategies for prevention.

This study has several limitations. First, it is a cross-sectional study, which does not allow for establishing causal relationships. Second, some variables were measured by questionnaire, which may lead to underestimation or overestimation for the prediction of ASCVD risk. Especially in the case of nutrition, it is particularly difficult to accurately reflect daily intake. Third, the study involved a secondary analysis of data from the KNHANES, so it was not possible to include all lifestyle and socioeconomic status variables in the analysis. Nevertheless, the strength of this study is that it used representative and reliable data from KNHANES to predict modifiable lifestyle risk factors predicted by sex using a ML approach.

We identified that additional consideration of several gender-specific lifestyle risk behaviors as well as several risk factors could contribute to more accurately predict ASCVD risk than the conventional CVD risk models. These gender-specific ML strategies should be critical to understand and prevent ASCVD risk

Ethical approval and consent to participate

The study protocol was approved by the Institutional Review Board of Hallym University Chuncheon Hospital (IRB No.: 2024–03–008). This study was complied with of the Declaration of Helsinki, and the informed consent for this study was exempted because the KNHANES is a public dataset.

Competing interest:

The authors declare no potential conflicts of interest with respect to the research, authorship, or publication of this article.

Author Contribution

Conceptualization, H.J.K., H.J.A., S.H.S., C.K.; Data curation, H.C., C.K., S.H.L.; Formal analysis, H.J.K., S.H.S. ;Funding acquisition, J.J.L., C.K.; Investigation, H.C., S.H.S., J.H.S.; Methodology, H.J.A.; Software, H.C., J.H.S.; Validation, H.J.A., S.H.L.; Visualization, H.C., S.H.L.; Writing - original draft, H.J.K., C.K. Writing - review & editing, H.J.K., H.J.A., C.K., J.J.L. All authors reviewed the manuscript.

Acknowledgments

None.

Data Availability

The data that support the findings of this study are available from the corresponding author on reasonable request.

Sources of Funding: This research was supported by a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (grant number : HR21C0198), and by a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI); funded by the Ministry of Health & Welfare, Republic of Korea (grant number: HI22C1498) ; supported by "Regional Innovation Strategy (RIS)" through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (MOE) (2022RIS-005) and supported by Hallym University Research Fund.

Weir HK, et al. Peer reviewed: heart disease and cancer deaths—trends and projections in the United States, 1969–2020. Prev. Chronic Dis. 13, E157 (2016).
Roth GA, et al. Global burden of cardiovascular diseases and risk factors, 1990–2019: update from the GBD 2019 study. J. Am. Coll. Cardiol. 76, 2982–3021 (2020).
Kim HC. Epidemiology of cardiovascular disease and its risk factors in Korea. Glob. Health Med. 3, 134–141. (2021).
Lee HH, et al. Korea heart disease fact sheet 2020: analysis of nationwide data. Korean Circ. J. 51, 495–503 (2021).
Poznyak AV, et al. Hypertension as a risk factor for atherosclerosis: Cardiovascular risk assessment. Front. Cardiovasc. Med. 9, 959285 (2022).
Devareddy A, Sarraju A, Rodriguez F. Health disparities across the continuum of ASCVD risk. Curr. Cardiol. Rep. 24, 1129–1137 (2022).
Mollborn S, Lawrence EM, Hummer RA. A gender framework for understanding health lifestyles. Soc. Sci. Med. 265, 113182 (2020).
Rajendran A, et al. Sex-specific differences in cardiovascular risk factors and implications for cardiovascular disease prevention in women. Atherosclerosis. 384, 117269 (2023).
Yusuf S, et al. Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): case-control study. Lancet. 364, 937–952 (2004).
Arnett DK, Khera A, Blumenthal RS. 2019 ACC/AHA guideline on the primary prevention of cardiovascular disease: part 1, lifestyle and behavioral factors. JAMA Cardiol. 4, 1043–1044 (2019).
Weng SF, et al. Can machine-learning improve cardiovascular risk prediction using routine clinical data? PLoS One. 12, e0174944 (2017).
Nakanishi R, et al. Machine learning adds to clinical and CAC assessments in predicting 10-year CHD and CVD deaths. JACC Cardiovasc. Imaging. 14, 615–625 (2021).
Cho SY, et al. Pre-existing and machine learning-based models for cardiovascular risk prediction. Sci Rep. 11, 8886 (2021).
Connelly PJ, et al. The importance of gender to understand sex differences in cardiovascular disease. Can. J. Cardiol. 37, 699–710 (2021).
Goff DC Jr, et al. 2013 ACC/AHA guideline on the assessment of cardiovascular risk: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation. 129, S49–S73 (2014).
Wong ND, et al. Atherosclerotic cardiovascular disease risk assessment: An American Society for Preventive Cardiology clinical practice statement. Am. J. Prev. Cardiol. 10, 100335 (2022).
Lloyd-Jones DM, et al. Use of risk assessment tools to guide decision-making in the primary prevention of atherosclerotic cardiovascular disease: a special report from the American Heart Association and American College of Cardiology. Circulation. 139, e1162–e1177 (2019).
Mortensen MB, et al. A simple disease-guided approach to personalize ACC/AHA-recommended statin allocation in elderly people: the BioImage Study. J. Am. Coll. Cardiol. 68, 881–891 (2016).
Lee HK, Lee BH. The epidemiology of alcohol use disorders. J. Korean Diabetes. 13, 69–75 (2012).
Lopez JPR, et al. Do vigorous-intensity and moderate-intensity physical activities reduce mortality to the same extent? A systematic review and meta-analysis. BMJ Open Sport Exerc. Med. 6, e000775 (2020).
Akiba T, et al. Optuna: A next-generation hyperparameter optimization framework. In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, Anchorage, AK, USA, August 4–8, (2019). (pp. 2623–2631). URL https://arxiv.org/abs/1907.10902.
Jebari-Benslaiman S, et al. Pathophysiology of atherosclerosis. Int. J. Mol. Sci. 23, 3346 (2022).
Man JJ, Beckman JA, Jaffe IZ. Sex as a biological variable in atherosclerosis. Circ. Res. 126, 1297–1319 (2020).
Rodgers JL, et al. Cardiovascular risks associated with gender and aging. J. Cardiovasc. Dev. Dis. 6, 19 (2019).
Batty M, Bennett MR, Yu E. The role of oxidative stress in atherosclerosis. Cells. 11, 3843 (2022).
Howard G, et al. Cigarette smoking and progression of atherosclerosis: The Atherosclerosis Risk in Communities (ARIC) Study. JAMA. 279, 119–124 (1998).
Gambardella J, et al. Quit smoking to outsmart atherogenesis: Molecular mechanisms underlying clinical evidence. Atherosclerosis. 257, 242–245 (2017).
Haslam D. Obesity: a medical history. Obesity Rev. 8, 31–36 (2007).
Kelly T, et al. Global burden of obesity in 2005 and projections to 2030. Int. J. Obes. 32, 1431–1437 (2008).
Gavina C, et al. Sex differences in LDL-C control in a primary care population: The PORTRAIT-DYS study. Atherosclerosis. 384, 117148. (2023).
Elder P, et al. Identification of female-specific risk enhancers throughout the lifespan of women to improve cardiovascular disease prevention. Am. J. Prev. Cardiol. 2, 100028 (2020).
Rodriguez F. Sex Disparities in Prevention of Atherosclerotic Cardiovascular Disease Across the Life Course. Circulation. 147, 523–525 (2023).
Backholer K, et al. Sex differences in the relationship between socioeconomic status and cardiovascular disease: a systematic review and meta-analysis. J Epidemiol. Community Health. 71, 550–557 (2017).

No competing interests reported.

TableS1.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Gender-Specific Lifestyle Risk Behaviors and Machine Learning Models for Predicting Atherosclerotic Cardiovascular Disease

Status:

Version 1

Abstract

Figures

Introduction

Methods

Participants

Measurement of ASCVD risk score

Measurement of lifestyle risk behaviors

Statistical analysis

Result

Baseline characteristics according to gender

Performance of machine learning model according to gender

Variable importance in ML model according to gender

Associations of lifestyle risk factors with 10-year ASCVD risk according to gender

Discussion

Conclusion

Declarations

Ethical approval and consent to participate

Competing interest:

Author Contribution

Acknowledgments

Data Availability

References

Additional Declarations

Supplementary Files

Status:

Version 1