Explainable machine learning model for predicting skeletal muscle loss during surgery and adjuvant chemotherapy in ovarian cancer.
Journal of cachexia, sarcopenia and muscle
BACKGROUND:Skeletal muscle loss during treatment is associated with poor survival outcomes in patients with ovarian cancer. Although changes in muscle mass can be assessed on computed tomography (CT) scans, this labour-intensive process can impair its utility in clinical practice. This study aimed to develop a machine learning (ML) model to predict muscle loss based on clinical data and to interpret the ML model by applying SHapley Additive exPlanations (SHAP) method. METHODS:This study included the data of 617 patients with ovarian cancer who underwent primary debulking surgery and platinum-based chemotherapy at a tertiary centre between 2010 and 2019. The cohort data were split into training and test sets based on the treatment time. External validation was performed using 140 patients from a different tertiary centre. The skeletal muscle index (SMI) was measured from pre- and post-treatment CT scans, and a decrease in SMI ≥ 5% was defined as muscle loss. We evaluated five ML models to predict muscle loss, and their performance was determined using the area under the receiver operating characteristic curve (AUC) and F1 score. The features for analysis included demographic and disease-specific characteristics and relative changes in body mass index (BMI), albumin, neutrophil-to-lymphocyte ratio (NLR), and platelet-to-lymphocyte ratio (PLR). The SHAP method was applied to determine the importance of the features and interpret the ML models. RESULTS:The median (inter-quartile range) age of the cohort was 52 (46-59) years. After treatment, 204 patients (33.1%) experienced muscle loss in the training and test datasets, while 44 (31.4%) patients experienced muscle loss in the external validation dataset. Among the five evaluated ML models, the random forest model achieved the highest AUC (0.856, 95% confidence interval: 0.854-0.859) and F1 score (0.726, 95% confidence interval: 0.722-0.730). In the external validation, the random forest model outperformed all ML models with an AUC of 0.874 and an F1 score of 0.741. The results of the SHAP method showed that the albumin change, BMI change, malignant ascites, NLR change, and PLR change were the most important factors in muscle loss. At the patient level, SHAP force plots demonstrated insightful interpretation of our random forest model to predict muscle loss. CONCLUSIONS:Explainable ML model was developed using clinical data to identify patients experiencing muscle loss after treatment and provide information of feature contribution. Using the SHAP method, clinicians may better understand the contributors to muscle loss and target interventions to counteract muscle loss.
10.1002/jcsm.13282
Development and validation of an insulin resistance model for a population without diabetes mellitus and its clinical implication: a prospective cohort study.
EClinicalMedicine
Background:Insulin resistance (IR) is associated with diabetes mellitus, cardiovascular disease (CV), and mortality. Few studies have used machine learning to predict IR in the non-diabetic population. Methods:In this prospective cohort study, we trained a predictive model for IR in the non-diabetic populations using the US National Health and Nutrition Examination Survey (NHANES, from JAN 01, 1999 to DEC 31, 2012) database and the Taiwan MAJOR (from JAN 01, 2008 to DEC 31, 2017) database. We analysed participants in the NHANES and MAJOR and participants were excluded if they were aged <18 years old, had incomplete laboratory data, or had DM. To investigate the clinical implications (CV and all-cause mortality) of this trained model, we tested it with the Taiwan biobank (TWB) database from DEC 10, 2008 to NOV 30, 2018. We then used SHapley Additive exPlanation (SHAP) values to explain differences across the machine learning models. Findings:Of all participants (combined NHANES and MJ databases), we randomly selected 14,705 participants for the training group, and 4018 participants for the validation group. In the validation group, their areas under the curve (AUC) were all >0.8 (highest being XGboost, 0.87). In the test group, all AUC were also >0.80 (highest being XGboost, 0.88). Among all 9 features (age, gender, race, body mass index, fasting plasma glucose (FPG), glycohemoglobin, triglyceride, total cholesterol and high-density cholesterol), BMI had the highest value of feature importance on IR (0.43 for XGboost and 0.47 for RF algorithms). All participants from the TWB database were separated into the IR group and the non-IR group according to the XGboost algorithm. The Kaplan-Meier survival curve showed a significant difference between the IR and non-IR groups (p < 0.0001 for CV mortality, and p = 0.0006 for all-cause mortality). Therefore, the XGboost model has clear clinical implications for predicting IR, aside from CV and all-cause mortality. Interpretation:To predict IR in non-diabetic patients with high accuracy, only 9 easily obtained features are needed for prediction accuracy using our machine learning model. Similarly, the model predicts IR patients with significantly higher CV and all-cause mortality. The model can be applied to both Asian and Caucasian populations in clinical practice. Funding:Taichung Veterans General Hospital, Taiwan and Japan Society for the Promotion of Science KAKENHI Grant Number JP21KK0293.
10.1016/j.eclinm.2023.101934
Machine learning algorithms predicting bladder cancer associated with diabetes and hypertension: NHANES 2009 to 2018.
Medicine
Bladder cancer is 1 of the 10 most common cancers in the world. However, the relationship between diabetes, hypertension and bladder cancer are still controversial, limited study used machine learning models to predict the development of bladder cancer. This study aimed to explore the association between diabetes, hypertension and bladder cancer, and build predictive models of bladder cancer. A total of 1789 patients from the National Health and Nutrition Examination Survey were enrolled in this study. We examined the association between diabetes, hypertension and bladder cancer using multivariate logistic regression model, after adjusting for confounding factors. Four machine learning models, including extreme gradient boosting (XGBoost), Artificial Neural Networks, Random Forest and Support Vector Machine were compared to predict for bladder cancer. Model performance was assessed by examining the area under the subject operating characteristic curve, accuracy, recall, specificity, precision, and F1 score. The mean age of bladder cancer group was older than that of the non-bladder cancer (74.4 years vs 65.6 years, P < .001), and men were more likely to have bladder cancer. Diabetes was associated with increased risk of bladder cancer (odds ratio = 1.24, 95%confidence interval [95%CI]: 1.17-3.02). The XGBoost model was the best algorithm for predicting bladder cancer; an accuracy and kappa value was 0.978 with 95%CI:0.976 to 0.986 and 0.01 with 95%CI:0.01 to 0.52, respectively. The sensitivity was 0.90 (95%CI:0.74-0.97) and the area under the curve was 0.78. These results suggested that diabetes is associated with risk of bladder cancer, and XGBoost model was the best algorithm to predict bladder cancer.
10.1097/MD.0000000000036587
Correlation analysis of lipid accumulation index, triglyceride-glucose index and H-type hypertension and coronary artery disease.
PeerJ
Objective:The current research was designed to explore the relationship between the lipid accumulation index (LAP), coronary artery disease (CAD), and the triglyceride-glucose (TyG) index in patient with H-type hypertension. Methods:From June 2021 to January 2022, our hospital's information management system collected data on 186 patients with essential hypertension. The participants were categorized into two groups (H-type hypertension ( = 113) and non-H-type hypertension ( = 73)) based on their homocysteine levels. Both groups' general condition, lipid accumulation index, triglyceride-glucose index, and Gensini score were compared to determine the factors influencing the severity of CAD in H-type hypertension patients. Results:There were statistically significant differences ( < 0.05) in homocysteine (Hcy, GLP-1 and SAA) level, LAP, and TyG indexes, but not in body mass index (BMI), smoking, sex, age, total cholesterol (TC), triglyceride (TG), high-density lipoprotein cholesterol (HDL-C), fasting plasma glucose (FPG), diastolic blood pressure, and systolic blood pressure. Additionally, there were substantial variations between the two groups regarding the number of lesion branches, degree of stenosis, and Gensini score ( > 0.05). patient with grade III to IV lesions had substantially higher LAP and TyG indices than those with stage I to II ( < 0.05). TyG (OR = 2.687) and TyG-LAP (OR = 4.512) were the factors determining the incidence of coronary artery disease in H-type hypertension, according to multivariate logistic regression analysis. The lesion number, stenosis degree, and Gensini score ( < 0.05) varied among both groups. LAP and TyG indexes were substantially greater in patients with double and triple vessel lesions than in those without lesions or with single vessel lesions ( < 0.05); similarly, these two indexes were considerably higher in individuals with grade III to IV lesions than in patients with grade I to II lesions ( < 0.05). As per the Pearson correlation analysis, the LAP, TyG indices and SAAlevel were adversely connected to the Gensini score ( = 0.254, 0.262, 0.299, < 0.05), the GLP-1 level was negatively correlated to the Gensini score (r = -0.291, < 0.05). TyG (OR = 2.687) and TyG-LAP (OR = 4.512) were the factors determining the frequency of coronary artery disease in H-type hypertension, according to multivariate logistic regression analysis. Conclusion:In conclusion, the LAP and TyG indexes were observed to be closely related to the degree of CAD in H-type individuals with hypertension, which can better understand the pathogenesis of coronary artery disease in patients with H-type hypertension and is of great significance for guiding clinical doctors to carry out personalized treatment and management.
10.7717/peerj.16069
Application of machine learning algorithms to construct and validate a prediction model for coronary heart disease risk in patients with periodontitis: a population-based study.
Frontiers in cardiovascular medicine
Background:The association between periodontitis and cardiovascular disease is increasingly recognized. In this research, a prediction model utilizing machine learning (ML) was created and verified to evaluate the likelihood of coronary heart disease in individuals affected by periodontitis. Methods:We conducted a comprehensive analysis of data obtained from the National Health and Nutrition Examination Survey (NHANES) database, encompassing the period between 2009 and 2014.This dataset comprised detailed information on a total of 3,245 individuals who had received a confirmed diagnosis of periodontitis. Subsequently, the dataset was randomly partitioned into a training set and a validation set at a ratio of 6:4. As part of this study, we conducted weighted logistic regression analyses, both univariate and multivariate, to identify risk factors that are independent predictors for coronary heart disease in individuals who have periodontitis. Five different machine learning algorithms, namely Logistic Regression (LR), Gradient Boosting Machine (GBM), Support Vector Machine (SVM), K-Nearest Neighbor (KNN), and Classification and Regression Tree (CART), were utilized to develop the model on the training set. The evaluation of the prediction models' performance was conducted on both the training set and validation set, utilizing metrics including AUC (Area under the receiver operating characteristic curve), Brier score, calibration plot, and decision curve analysis (DCA). Additionally, a graphical representation called a nomogram was created using logistic regression to visually depict the predictive model. Results:The factors that were found to independently contribute to the risk, as determined by both univariate and multivariate logistic regression analyses, encompassed age, race, presence of myocardial infarction, chest pain status, utilization of lipid-lowering medications, levels of serum uric acid and serum creatinine. Among the five evaluated machine learning models, the KNN model exhibited exceptional accuracy, achieving an AUC value of 0.977. The calibration plot and brier score illustrated the model's ability to accurately estimate probabilities. Furthermore, the model's clinical applicability was confirmed by DCA. Conclusion:Our research showcases the effectiveness of machine learning algorithms in forecasting the likelihood of coronary heart disease in individuals with periodontitis, thereby aiding healthcare professionals in tailoring treatment plans and making well-informed clinical decisions.
10.3389/fcvm.2023.1296405
Association between triglyceride glucose index and hyperuricemia: a new evidence from China and the United States.
Frontiers in endocrinology
Background:Hyperuricemia (HUA) is a glo\bal public health problem. The etiology of HUA is complex and efficient and accurate assessment metrics are still lacking when conducting large-scale epidemiologic screening. The aim of this study was to evaluate the association of the triglyceride glucose (TyG) index, TyG-body mass index (BMI), TyG-waist-to-height ratio (WHtR) with the risk of HUA. Methods:Based on data collected from the National Health and Nutrition Examination Survey (NHANES) in the United States and the China Health and Aging Longitudinal Study (CHARLS) in China, a total of 14,286 U.S. adults and 4,620 Chinese adults were included in the analysis. The study examined the levels of TyG, TyG-BMI, TyG-WHtR, and TyG-WC. Multivariate logistic regression was utilized to investigate the relationships between these variables and hyperuricemia (HUA), separately. Additionally, the study used restricted cubic splines (RCS) to explore the linear associations of TyG, TyG-BMI, TyG-WHtR, TyG-WC, and HUA, separately. Results:The NHANES results showed that TyG [Q2, 1.58(1.26, 1.98); Q3, 2.36 (1.94, 2.88); Q4, 3.21 (2.61, 3.94)], TyG-BMI [Q2, 2.14 (1.74, 2.65); Q3, 3.38 (2.74, 4.17); Q4, 6.70 (5.55, 8.02)], TyG-WHtR [Q2, 1.92 (1.56, 2.36); Q3, 3.14 (2.56, 3.85); Q4, 6.28 (5.12, 7.69)], TyG-WC [Q2, 2.32 (1.85, 2.90); Q3, 3.51 (2.84, 4.34); Q4, 7.32 (5.95, 9.02)] were identified as risk factors for hyperuricemia (HUA). Similarly, the CHARLS results, when fully adjusted for covariates, indicated that TyG [Q4, 2.36 (1.08, 5.15)], TyG-BMI [Q3, 2.60 (1.05, 6.41); Q4, 3.70 (1.64, 8.32)], TyG-WHtR (Q4, 2.84 (1.23, 6.55), TyG-WC [Q4, 2.85 (1.23, 6.5)] were also risk factors for HUA. The predictive ability of each indicator for the risk of developing HUA was stronger in women than in men. Furthermore, there was an observed nonlinear relationship between TyG, TyG-BMI, TyG-WHtR, TyG-WC, and HUA in both the NHANES and CHARLS datasets ( < 0.05). Conclusion:These findings suggest that TyG, TyG-BMI, TyG-WHtR and TyG-WC are associated with an increased risk of HUA. They are potential indicators for screening HUA status in the general population in China and the United States.
10.3389/fendo.2024.1403858
Comparison of seven anthropometric indexes to predict hypertension plus hyperuricemia among U.S. adults.
Frontiers in endocrinology
Purpose:This study aims to compare the association of hypertension plus hyperuricemia (HTN-HUA) with seven anthropometric indexes. These include the atherogenic index of plasma (AIP), lipid accumulation product (LAP), visceral adiposity index (VAI), triglyceride-glucose index (TyG), body roundness index (BRI), a body shape index (ABSI), and the cardiometabolic index (CMI). Methods:Data was procured from the National Health and Nutrition Examination Survey (NHANES), which recruited a representative population aged 18 years and above to calculate these seven indexes. Logistic regression analysis was employed to delineate their correlation and to compute the odds ratios (OR). Concurrently, receiver operating characteristic (ROC) curves were utilized to evaluate the predictive power of the seven indexes. Results:A total of 23,478 subjects were included in the study. Among these, 6,537 (27.84%) were patients with HUA alone, 2,015 (8.58%) had HTN alone, and 2,836 (12.08%) had HTN-HUA. The multivariate logistic regression analysis showed that the AIP, LAP, VAI, TyG, BRI, ABSI, and CMI were all significantly associated with concurrent HTN-HUA. The OR for the highest quartile of the seven indexes for HTN-HUA were as follows: AIP was 4.45 (95% CI 3.82-5.18), LAP was 9.52 (95% CI 7.82-11.59), VAI was 4.53 (95% CI 38.9-5.28), TyG was 4.91 (95% CI 4.15-5.80), BRI was 9.08 (95% CI 7.45-11.07), ABSI was 1.71 (95% CI 1.45 -2.02), and CMI was 6.57 (95% CI 5.56-7.76). Notably, LAP and BRI demonstrated significant discriminatory abilities for HTN-HUA, with area under the curve (AUC) values of 0.72 (95% CI 0.71 - 0.73) and 0.73 (95% CI 0.72 - 0.74) respectively. Conclusion:The AIP, LAP, VAI, TyG, BRI, ABSI, and CMI all show significant correlation with HTN-HUA. Notably, both LAP and BRI demonstrate the capability to differentiate cases of HTN-HUA. Among these, BRI is underscored for its effective, non-invasive nature in predicting HTN-HUA, making it a superior choice for early detection and management strategies.
10.3389/fendo.2024.1301543
A high frequency of Gilbert syndrome (UGT1A1*28/*28) and associated hyperbilirubinemia but not cholelithiasis in adolescent and adult north Indian patients with transfusion-dependent β-thalassemia.
Shrestha Oshan,Khadwal Alka Rani,Singhal Manphool,Trehan Amita,Bansal Deepak,Jain Richa,Pal Arnab,Hira Jasbir Kaur,Chhabra Sanjeev,Malhotra Pankaj,Das Reena,Sharma Prashant
Annals of hematology
Hyperbilirubinemia and pigment gallstones are frequent complications in transfusion-dependent β-thalassemia (TDβT) patients. Bilirubin production and clearance are determined by genetic as well as environmental variables like ineffective erythropoiesis, hemolysis, infection-induced hepatic injury, and drug- or iron-related toxicities. We studied the frequency of the Gilbert syndrome (GS), a common hereditary cause of hyperbilirubinemia in 102 TDβT patients aged 13-43 years (median 26 years). Total and unconjugated hyperbilirubinemia were frequent (81.4% and 84.3% patients respectively). Twenty (19.6%) patients showed total bilirubin > 3.0 mg/dL; 53 (51.9%) had an elevation of either alanine or aspartate aminotransferase, or alkaline phosphatase liver enzymes. Nineteen (18.6% of the 92 tested) were positive for hepatitis B or C, or HIV. The mean total and unconjugated bilirubin levels and AST, ALT, and ALP levels in patients positive for hepatitis B or C were not significantly different from negative cases. Eighteen patients (17.7%) had GS: homozygous (TA)7/7 UGT1A1 promoter motif (the *28/*28 genotype), 48 (47.1%) were heterozygous (TA)6/7. Total + unconjugated bilirubin rose significantly with the (TA)7 allele dose. Fourteen (13.7%) patients had gallstones. There was no significant difference in total/unconjugated bilirubin in patients with/without gallstones and no significant differences in frequencies of gallstones within the three UGT1A1 genotypes. This largest study in Indian TDβT patients suggests that GS should be excluded in TDβT cases where jaundice remains unexplained after treatable causes like infections, chelator toxicity, or transfusion-related hemolysis are excluded. GS was not associated with gallstones, possibly due to a lower incidence of cholelithiasis overall, a younger age cohort, or other environmental factors.
10.1007/s00277-020-04176-2
Incidence, Risk Factors, and Outcomes of Hyperbilirubinemia in Adult Cardiac Patients Supported by Veno-Arterial ECMO.
Lyu Lin,Yao Jingxin,Gao Guodong,Long Cun,Hei Feilong,Ji Bingyang,Liu Jinping,Yu Kun,Hu Qiang,Hu Jinxiao
Artificial organs
The aims of this study were to evaluate the incidence, risk factors, and outcomes of hyperbilirubinemia in cardiac patients with veno-arterial (VA) ECMO. Data on 89 adult patients with cardiac diseases who received VA ECMO implantation in our hospital were retrospectively reviewed. All patients were divided into the following three groups: 24 in normal group (N, total bilirubin [TBIL] ≤3 mg/dL), 30 in high bilirubin group (HB, 6 mg/dL ≥ TBIL > 3 mg/dL), and 35 in severe high bilirubin group (SHB, TBIL > 6 mg/dL). lg(variables + 1) was performed for nonnormally distributed variables. The incidence of hyperbilirubinemia (>3 mg/dL) was 73%. In a multiple linear regression analysis, lg(peak TBIL + 1) was significantly associated with lg(peak AST + 1) (b-coefficient 0.188, P = 0.001), lg(peak pFHb + 1) (b-coefficient 0.201, P = 0.003), and basic TBIL (b-coefficient 0.006, P = 0.009). Repeated measurement analysis of variance revealed that the main effect for three groups in pFHb and lg(AST + 1) was significant at first 3 days during ECMO. The patients in SHB had low platelets during ECMO and low in-hospital survival rate. Hyperbilirubinemia remains common in patients with VA ECMO and is associated with low platelets and high in-hospital mortality. Hemolysis and liver dysfunction during ECMO and basic high bilirubin levels are risk factors of hyperbilirubinemia.
10.1111/aor.12979
Association between sarcopenia and sleep disorders: a cross-sectional population based study.
Frontiers in nutrition
Objective:Sleep disorders is a worldwide public health problem. We sought to examine the association between sarcopenia, a decline in skeletal muscle mass and function, and sleep disorders within the adult demographic of the United States during the period spanning 2011 to 2018. Methods:Diagnosis of sarcopenia and sleep disorders was ascertained through appropriate calculations and a structured questionnaire. The primary correlation analysis was conducted using a weighted multivariate logistic regression model. Furthermore, to confirm the presence of a potential non-linear association between sarcopenia and sleep disorders, additional analyses were performed using multivariate logistic regression and restricted cubic spline (RCS) regression with dose-response curve analysis. Subgroup analyses were also conducted to explore the influence of relevant socio-demographic factors and other covariates. Results:The final analysis encompassed 5,616 participants. Model 4, inclusive of all pertinent covariates, revealed a positive correlation between sarcopenia and sleep disorders, yielding an odds ratio (OR) of 1.732 (95% CI: 1.182-2.547; = 0.002). Further analysis, utilizing the restricted cubic spline model, indicated a decreasing trend in sleep disorders as sarcopenia indices rose. Stratified analyses across diverse variables underscored the significant impact of sarcopenia on sleep disorders prevalence in several subgroups. Specifically, males, individuals aged 40 and above, non-Hispanic whites, those with high school education or equivalent, unmarried individuals, obese individuals (BMI ≥ 30), alcohol drinkers, former smokers, diabetics, and those engaging in less rigorous recreational activities exhibited a more pronounced association between sarcopenia and sleep disorders. The incidence of sleep disorders exhibited an upward trend as the incidence of sarcopenia declined among study participants. Conclusions:In summary, our study provides evidence of an association between sarcopenia and the prevalence of sleep disorders, with a negative correlation observed between the sarcopenia index and the odds ratio of sleep disorders. These findings suggest that maintaining optimal muscle mass may have a beneficial impact on sleep-related issues. In terms of exploring the mechanisms underlying the relationship between sarcopenia and sleep disorders, more in-depth research is warranted to ascertain the definitive causal relationship.
10.3389/fnut.2024.1415743
Association between obstructive sleep apnea and visceral adiposity index and lipid accumulation product: NHANES 2015-2018.
Lipids in health and disease
BACKGROUND:Obesity refers to a significant contributor to the development of obstructive sleep apnea (OSA). Early prediction of OSA usually leads to better treatment outcomes, and this study aims to employ novel metabolic markers, visceral adiposity index (VAI), and lipid accumulation product (LAP) to evaluate the relationship to OSA. METHODS:The data used in the current cross-sectional investigation are from the National Health and Nutrition Examination Survey (NHANES), which was carried out between 2015 and 2018. To examine the correlation between LAP and VAI levels and OSA, multivariate logistic regression analysis was adopted. In addition, various analytical methods were applied, including subgroup analysis, smooth curve fitting, and threshold effect analysis. RESULTS:Among totally 3932 participants, 1934 were included in the OSA group. The median (Q1-Q3) values of LAP and VAI for the participants were 40.25 (21.51-68.26) and 1.27 (0.75-2.21), respectively. Logistic regression studies indicated a positive correlation between LAP, VAI, and OSA risk after adjusting for potential confounding variables. Subgroup analysis revealed a stronger correlation between LAP, VAI levels, and OSA among individuals aged < 60 years. Through smooth curve fitting, specific saturation effects of LAP, VAI, and BMD were identified, with inflection points at 65.684 and 0.428, respectively. CONCLUSION:This study demonstrates that elevated levels of LAP and VAI increase the risk of OSA, suggesting their potential as predictive markers for OSA and advocating for dietary and exercise interventions to mitigate OSA risk in individuals with high LAP and VAI levels.
10.1186/s12944-024-02081-5
Association between vitamin levels and obesity in the national health and nutrition examination surveys 2017 to 2018.
Journal of developmental origins of health and disease
In recent years, the rapidly increasing incidence of obesity is becoming a worldwide public health problem. Obesity is a chronic disease which may have a major negative effect on the people's quality of life. Previous studies on the comprehensive effects of multivitamins on central obesity and general obesity are relatively few. The aim of this study was to evaluate association of vitamins exposure with obesity risk and obesity-related indicators. We fitted three statistical models (linear regression model, logistic regression model, and Bayesian kernel machine regression model) to evaluate the correlation between vitamin levels and obesity in the study population. The vitamin score represents the overall level of vitamin in serum, which was mutually verified with the results obtained from statistical model. The vitamin (A, C, and D) levels were significantly higher among non-obesity group compared to the obesity group. Using the lowest quartile of vitamin level as a referent, vitamin A, C, and D levels showed significantly negative correlation with the obesity risk in both adjusted and unadjusted models. When considering all vitamin as a mixed exposure, we found a generally negative relationship between vitamin mixtures with binary outcome (obesity) and continuous outcome (BMI, waist circumference, and hsCRP). Reduced levels of vitamins (A, C and D) increased the risk of obesity. Increased levels of vitamin mixtures can significantly reduce obesity risk and obesity-related indicators. Vitamins may reduce the risk of obesity by suppressing inflammatory responses.
10.1017/S2040174423000466
An analysis of the potential association between obstructive sleep apnea and osteoporosis from the perspective of transcriptomics and NHANES.
BMC public health
BACKGROUND:Obstructive sleep apnea (OSA) and osteoporosis (OP) are prevalent diseases in the elderly. This study aims to reveal the clinical association between OSA and OP and explore potential crosstalk gene targets. METHODS:Participants diagnosed with OSA in the National Health and Nutrition Examination Survey (NHANES) database (2015-2020) were included, and OP was diagnosed based on bone mineral density (BMD). We explored the association between OSA and OP, and utilized multivariate logistic regression analysis and machine learning algorithms to explore the risk factors for OP in OSA patients. Overlapping genes of comorbidity were explored using differential expression analysis, Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis, Least Absolute Shrinkage and Selection Operator (LASSO) regression, and Random Forest (RF) methods. RESULTS:In the OSA population, the weighted prevalence of OP was 7.0%. The OP group had more females, lower body mass index (BMI), and more low/middle-income individuals compared to the non-OP group. Female gender and lower BMI were identified as independent risk factors for OP in OSA patients. Gene expression profiling revealed 8 overlapping differentially expressed genes in OP and OSA patients. KCNJ1, NPR3 and WT1-AS were identified as shared diagnostic biomarkers or OSA and OP, all of which are associated with immune cell infiltration. CONCLUSION:This study pinpointed female gender and lower BMI as OP risk factors in OSA patients, and uncovered three pivotal genes linked to OSA and OP comorbidity, offering fresh perspectives and research targets.
10.1186/s12889-024-19540-4
A machine learning model to predict the risk of depression in US adults with obstructive sleep apnea hypopnea syndrome: a cross-sectional study.
Frontiers in public health
Objective:Depression is very common and harmful in patients with obstructive sleep apnea hypopnea syndrome (OSAHS). It is necessary to screen OSAHS patients for depression early. However, there are no validated tools to assess the likelihood of depression in patients with OSAHS. This study used data from the National Health and Nutrition Examination Survey (NHANES) database and machine learning (ML) methods to construct a risk prediction model for depression, aiming to predict the probability of depression in the OSAHS population. Relevant features were analyzed and a nomogram was drawn to visually predict and easily estimate the risk of depression according to the best performing model. Study design:This is a cross-sectional study. Methods:Data from three cycles (2005-2006, 2007-2008, and 2015-2016) were selected from the NHANES database, and 16 influencing factors were screened and included. Three prediction models were established by the logistic regression algorithm, least absolute shrinkage and selection operator (LASSO) algorithm, and random forest algorithm, respectively. The receiver operating characteristic (ROC) area under the curve (AUC), specificity, sensitivity, and decision curve analysis (DCA) were used to assess evaluate and compare the different ML models. Results:The logistic regression model had lower sensitivity than the lasso model, while the specificity and AUC area were higher than the random forest and lasso models. Moreover, when the threshold probability range was 0.19-0.25 and 0.45-0.82, the net benefit of the logistic regression model was the largest. The logistic regression model clarified the factors contributing to depression, including gender, general health condition, body mass index (BMI), smoking, OSAHS severity, age, education level, ratio of family income to poverty (PIR), and asthma. Conclusion:This study developed three machine learning (ML) models (logistic regression model, lasso model, and random forest model) using the NHANES database to predict depression and identify influencing factors among OSAHS patients. Among them, the logistic regression model was superior to the lasso and random forest models in overall prediction performance. By drawing the nomogram and applying it to the sleep testing center or sleep clinic, sleep technicians and medical staff can quickly and easily identify whether OSAHS patients have depression to carry out the necessary referral and psychological treatment.
10.3389/fpubh.2023.1348803
Interpretable machine learning framework to predict gout associated with dietary fiber and triglyceride-glucose index.
Nutrition & metabolism
BACKGROUND:Gout prediction is essential for the development of individualized prevention and treatment plans. Our objective was to develop an efficient and interpretable machine learning (ML) model using the SHapley Additive exPlanation (SHAP) to link dietary fiber and triglyceride-glucose (TyG) index to predict gout. METHODS:Using datasets from the National Health and Nutrition Examination Survey (NHANES) (2005-2018) population to study dietary fiber, the TyG index was used to predict gout. After evaluating the performance of six ML models and selecting the Light Gradient Boosting Machine (LGBM) as the optimal algorithm, we interpret the LGBM model for predicting gout using SHAP and reveal the decision-making process of the model. RESULTS:An initial survey of 70,190 participants was conducted, and after a gradual exclusion process, 12,645 cases were finally included in the study. Selection of the best performing LGBM model for prediction of gout associated with dietary fiber and TyG index (Area under the ROC curve (AUC): 0.823, 95% confidence interval (CI): 0.798-0.848, Accuracy: 95.3%, Brier score: 0.077). The feature importance of SHAP values indicated that age was the most important feature affecting the model output, followed by uric acid (UA). The SHAP values showed that lower dietary fiber values had a more pronounced effect on the positive prediction of the model, while higher values of the TyG index had a more pronounced effect on the positive prediction of the model. CONCLUSION:The interpretable LGBM model associated with dietary fiber and TyG index showed high accuracy, efficiency, and robustness in predicting gout. Increasing dietary fiber intake and lowering the TyG index are beneficial in reducing the potential risk of gout.
10.1186/s12986-024-00802-2
Vitamin A Deficiency and the Lung.
Timoneda Joaquín,Rodríguez-Fernández Lucía,Zaragozá Rosa,Marín M Pilar,Cabezuelo M Teresa,Torres Luis,Viña Juan R,Barber Teresa
Nutrients
Vitamin A (all--retinol) is a fat-soluble micronutrient which together with its natural derivatives and synthetic analogues constitutes the group of retinoids. They are involved in a wide range of physiological processes such as embryonic development, vision, immunity and cellular differentiation and proliferation. Retinoic acid (RA) is the main active form of vitamin A and multiple genes respond to RA signalling through transcriptional and non-transcriptional mechanisms. Vitamin A deficiency (VAD) is a remarkable public health problem. An adequate vitamin A intake is required in early lung development, alveolar formation, tissue maintenance and regeneration. In fact, chronic VAD has been associated with histopathological changes in the pulmonary epithelial lining that disrupt the normal lung physiology predisposing to severe tissue dysfunction and respiratory diseases. In addition, there are important alterations of the structure and composition of extracellular matrix with thickening of the alveolar basement membrane and ectopic deposition of collagen I. In this review, we show our recent findings on the modification of cell-junction proteins in VAD lungs, summarize up-to-date information related to the effects of chronic VAD in the impairment of lung physiology and pulmonary disease which represent a major global health problem and provide an overview of possible pathways involved.
10.3390/nu10091132
Machine-learning-based prediction of cardiovascular events for hyperlipidemia population with lipid variability and remnant cholesterol as biomarkers.
Health information science and systems
Purpose:Dyslipidemia poses a significant risk for the progression to cardiovascular diseases. Despite the identification of numerous risk factors and the proposal of various risk scales, there is still an urgent need for effective predictive models for the onset of cardiovascular diseases in the hyperlipidemic population, which are essential for the prevention of CVD. Methods:We carried out a retrospective cohort study with 23,548 hyperlipidemia patients in Shenzhen Health Information Big Data Platform, including 11,723 CVD onset cases in a 3-year follow-up. The population was randomly divided into 70% as an independent training dataset and remaining 30% as test set. Four distinct machine-learning algorithms were implemented on the training dataset with the aim of developing highly accurate predictive models, and their performance was subsequently benchmarked against conventional risk assessment scales. An ablation study was also carried out to analyze the impact of individual risk factors to model performance. Results:The non-linear algorithm, LightGBM, excelled in forecasting the incidence of cardiovascular disease within 3 years, achieving an area under the 'receiver operating characteristic curve' (AUROC) of 0.883. This performance surpassed that of the conventional logistic regression model, which had an AUROC of 0.725, on identical datasets. Concurrently, in direct comparative analyses, machine-learning approaches have notably outperformed the three traditional risk assessment methods within their respective applicable populations. These include the Framingham cardiovascular disease risk score, 2019 ESC/EAS guidelines for the management of dyslipidemia and the 2016 Chinese recommendations for the management of dyslipidemia in adults. Further analysis of risk factors showed that the variability of blood lipid levels and remnant cholesterol played an important role in indicating an increased risk of CVD. Conclusions:We have shown that the application of machine-learning techniques significantly enhances the precision of cardiovascular risk forecasting among hyperlipidemic patients, addressing the critical issue of disease prediction's heterogeneity and non-linearity. Furthermore, some recently-suggested biomarkers, including blood lipid variability and remnant cholesterol are also important predictors of cardiovascular events, suggesting the importance of continuous lipid monitoring and healthcare profiling through big data platforms.
10.1007/s13755-024-00310-w
Machine learning-based prediction of vitamin D deficiency: NHANES 2001-2018.
Frontiers in endocrinology
Background:Vitamin D deficiency is strongly associated with the development of several diseases. In the current context of a global pandemic of vitamin D deficiency, it is critical to identify people at high risk of vitamin D deficiency. There are no prediction tools for predicting the risk of vitamin D deficiency in the general community population, and this study aims to use machine learning to predict the risk of vitamin D deficiency using data that can be obtained through simple interviews in the community. Methods:The National Health and Nutrition Examination Survey 2001-2018 dataset is used for the analysis which is randomly divided into training and validation sets in the ratio of 70:30. GBM, LR, NNet, RF, SVM, XGBoost methods are used to construct the models and their performance is evaluated. The best performed model was interpreted using the SHAP value and further development of the online web calculator. Results:There were 62,919 participants enrolled in the study, and all participants included in the study were 2 years old and above, of which 20,204 (32.1%) participants had vitamin D deficiency. The models constructed by each method were evaluated using AUC as the primary evaluation statistic and ACC, PPV, NPV, SEN, SPE, F1 score, MCC, Kappa, and Brier score as secondary evaluation statistics. Finally, the XGBoost-based model has the best and near-perfect performance. The summary plot of SHAP values shows that the top three important features for this model are race, age, and BMI. An online web calculator based on this model can easily and quickly predict the risk of vitamin D deficiency. Conclusion:In this study, the XGBoost-based prediction tool performs flawlessly and is highly accurate in predicting the risk of vitamin D deficiency in community populations.
10.3389/fendo.2024.1327058
Distribution of lipid levels and prevalence of hyperlipidemia: data from the NHANES 2007-2018.
Lipids in health and disease
BACKGROUND:Lipid-lowering therapy is important, and the distribution of lipid levels and the incidence of hyperlipidemia may vary in different subgroups of the population. We aimed to explore the distribution of lipid levels and the prevalence of hyperlipidemia in subpopulations with subgroup factors, including age, sex, race, and smoking status. METHODS:Our study used data from the National Health and Nutrition Examination Survey (NHANES) from 2007 to 2018, ultimately enrolling and analyzing 15,499 participants. A cross-sectional analysis was performed to assess the distribution of lipids and prevalence of hyperlipidemia in subpopulations, and multifactorial logistic regression analyses were performed for the prevalence of hyperlipidemia, adjusted for age, sex, race and smoking status. RESULTS:Blacks had significantly lower mean serum total cholesterol and triglycerides and higher serum high-density lipoprotein cholesterol (HDL-C) than whites (P < 0.001). In contrast, Mexican Americans had markedly higher mean serum triglycerides and lower serum HDL-C than whites (P < 0.001). Furthermore, the prevalence of hypercholesterolemia and hypertriglyceridemia was lower in blacks than in whites (P = 0.003 and P < 0.001, respectively), while the prevalence of hypertriglyceridemia was significantly higher in Mexican Americans than in whites (P = 0.002). In addition, total cholesterol and triglyceride levels were significantly higher in women aged 65 years or older and markedly higher than in men in the same age group (P < 0.001). In addition, overall mean total cholesterol, triglyceride, and low-density lipoprotein cholesterol (LDL-C) levels were higher in smokers than in nonsmokers (P = 0.01, P < 0.001, and P = 0.005, respectively). CONCLUSION:Based on NHANES data, the mean lipid levels and prevalence of hyperlipidemia differed by sex, age, race, and smoking status.
10.1186/s12944-022-01721-y
Race/Ethnicity and Other Predictors of Early-Onset Type 2 Diabetes Mellitus in the US Population.
Journal of racial and ethnic health disparities
OBJECTIVES:Among US adults aged 20 + years in the USA with previously diagnosed type 2 diabetes mellitus (T2DM), we aimed to estimate the prevalence of early-onset T2DM (onset at age < 50.5 years) and to test associations between early-onset T2DM and race/ethnicity, and other hypothesized predictors. METHODS:We pooled data from the annual National Health and Nutrition Examination Surveys (NHANES) over the years 2001 through 2018. We tested hypotheses of association and identified predictors using stepwise logistic regression analysis, and 11 supervised machine learning classification algorithms. RESULTS:After appropriate weighting, we estimated that among adults in the USA aged 20 + years with previously diagnosed T2DM, the prevalence of early-onset was 52.9% (95% confidence intervals, 49.6 to 56.2%). Among Non-Hispanic Whites (NHW) the prevalence was 48.6% (95% CI, 44.6 to 52.6%), among Non-Hispanic Blacks: 56.9% (95% CI, 51.8 to 62.0%), among Hispanics: 62.7% (95% CI, 53.2 to 72.3%). In the final multivariable logistic regression model, the top-3 markers predicting early-onset T2DM in males were NHB ethnicity (OR = 2.97; 95% CI: 2.24-3.95) > tobacco smoking (OR = 2.79; 95% CI: 2.18-3.58) > high education level (OR = 1.65; 95% CI: 1.27-2.14) in males. In females, the ranking was tobacco smoking (OR = 2.59; 95% CI: 1.90-3.53) > Hispanic ethnicity (OR = 1.49; 95% CI: 1.08-2.05) > obesity (OR = 1.30; 95% CI: 0.91-1.86) in females. The acculturation score emerged from the machine learning approach as the dominant marker explaining the race disparity in early-onset T2DM. CONCLUSIONS:The prevalence of early-onset T2DM was higher among NHB and Hispanic people, than among NHW people. Independently of race/ethnicity, acculturation, tobacco smoking, education level, marital status, obesity, and hypertension were also predictive.
10.1007/s40615-024-01980-8
Identification and optimization of relevant factors for chronic kidney disease in abdominal obesity patients by machine learning methods: insights from NHANES 2005-2018.
Lipids in health and disease
BACKGROUND:The intake of dietary antioxidants and glycolipid metabolism are closely related to chronic kidney disease (CKD), particularly among individuals with abdominal obesity. Nevertheless, the cumulative effect of multiple comorbid risk factors on the progression and complications of CKD remains inadequately characterized. METHODS:This study analyzed data from the National Health and Nutrition Examination Survey (NHANES) dat abase (2005-2018), to examine potential factors related to CKD, including glycolipid metabolism, dietary antioxidant intake, and pertinent medical history. To explore the associations between these variables and CKD, the present study used a multivariable-adjusted least absolute shrinkage and selection operator (LASSO) regression model, along with a restricted cubic spline (RCS) model. Furthermore, an optimal predictive model was developed for CKD using ten machine learning algorithms and enhanced model interpretability with the Shapley Additive Explanations (SHAP) method. RESULTS:A cohort comprising 8,764 eligible individuals (52% male, including 1,839 CKD patients) with abdominal obesity aged 20-85 years were included. The findings revealed significant positive correlations in patients with abdominal obesity between the presence of CKD and age, a history of heart failure, hypertension, diabetes, elevated lipid accumulation product (LAP) and triglyceride glucose-waist circumference (TyG-WC) levels. Conversely, negative correlations were identified between CKD and variables such as sex, high-density lipoprotein cholesterol (HDL-C) levels, and the composite dietary antioxidant index (CDAI). In parallel, RCS regression analysis revealed significant nonlinear associations between the CDAI, HDL-C, TyG-WC, and CKD among patients with abdominal obesity aged 60-80 years. The development of predictive models demonstrated that the CatBoost model surpassed other models, achieving an accuracy of 86.74% on the validation set. The model's area under the receiver operator curve (AUC) and F1 score were 0.938 and 0.889, respectively. The SHAP values revealed that age was the most significant predictor, followed by diabetes history, hypertension, HDL-C levels, CDAI index, TyG-WC, and LAP. CONCLUSION:CatBoost models, along with glycolipid metabolism indexes and dietary antioxidant intake, are effective for early CKD detection in patients with abdominal obesity.
10.1186/s12944-024-02384-7
The association between Weight-adjusted-Waist Index (WWI) and cognitive function in older adults: a cross-sectional NHANES 2011-2014 study.
BMC public health
BACKGROUND:The impact of obesity on cognitive function has engendered considerable interest. Weight-adjusted waist index (WWI) has emerged as a novel and innovative marker of obesity that reflects weight-independent abdominal obesity. However, the association between WWI and cognitive function remains unclear. To address this gap, the present study aims to explore the relationship between weight-adjusted waist index (WWI) and cognitive performance in older adults. METHODS:We conducted a cross-sectional investigation using datasets from the National Health and Nutrition Examination Survey (NHANES) 2011-2014. The study included 3,472 participants (48.59% male, 51.41% female) of various races (Mexican American, Other Hispanic, Non-Hispanic White, Non-Hispanic Black, and Other), with a mean age of 69.95 years (SD = 6.94). Multivariate regression and smoothing curve fitting were used to investigate the linear and nonlinear relationship between WWI and cognitive performance in the following domains: learning and memory, verbal fluency, and processing speed, as measured by Consortium to Establish a Registry for Alzheimer's Disease Word Learning subtest (CERAD-WL), Animal Fluency Test (AFT), and Digit Symbol Substitution Test (DSST), respectively. Subgroup analysis and interaction tests were conducted to examine the stability of this relationship across groups. Machine learning models based on random forests were used to analyze the predictive performance of WWI for cognitive function. RESULTS:A total of 3,472 participants were included in the analysis. The results revealed significant negative associations between WWI and low scores on the CERAD-WL [-0.96 (-1.30, -0.62)], AFT [-0.77 (-1.05, -0.49)], and DSST [-3.67 (-4.55, -2.79)]. This relationship remained stable after converting WWI to a categorical variable. In addition, this significant negative association was more pronounced in men than women and diminished with advancing age. Non-linear threshold effects were observed, with correlations intensifying between WWI and CERAD-WL when WWI surpassed 12.25, AFT when WWI surpassed 11.54, and DSST when WWI surpassed 11.66. CONCLUSIONS:A higher WWI, indicating increased abdominal obesity, was associated with deficits in learning, memory, verbal fluency, and processing speed among older adults. These findings suggest that abdominal obesity may play a crucial role in cognitive decline in this population. The stronger relationship observed between WWI and cognition in men highlights the need for gender-specific considerations in interventions targeting abdominal obesity. The results demonstrate the importance of interventions targeting abdominal obesity to preserve cognitive performance in older adults.
10.1186/s12889-024-19332-w
The association of hypertriglyceridemic waist phenotype with hypertension: A cross-sectional study in a Chinese middle aged-old population.
Xuan Yan,Shen Ying,Wang Sujie,Gao Ping,Gu Xi,Tang Dou,Wang Xun,Zhu Fanfan,Lu Leiqun,Chen Ling
Journal of clinical hypertension (Greenwich, Conn.)
The present study aimed to evaluate the relationship between the hypertriglyceridemic waist (HTGW) phenotype and hypertension. We undertook a cross-sectional study with a sample of 9015 adults from China. The HTGW phenotype was defined as elevated waist circumference (WC) and elevated triglyceride (TG) concentration. Logistic regression analysis was used to evaluate the association between the HTGW phenotype and hypertension. The prevalence of hypertension was significantly higher in individuals with the HTGW phenotype, than in those with the normal waist normal triglyceride (NWNT) phenotype (89.9% vs 75.3%, respectively, P < .001). After adjusting for age, sex, BMI, current smoker, and current alcohol consumption, the HTGW phenotype was associated with hypertension (Odds Ratio (OR)1.53; 95% CI 1.25-1.87). After further adjustment for potential confounders, the HTGW phenotype was still significantly associated with hypertension (adjusted OR1.28; 95% CI 1.04-1.58) regardless of sex. The subgroup analyses generally revealed similar associations across all subgroups. This study indicated that the HTGW phenotype was strongly associated with hypertension, and blood pressure should be clinically monitored in individuals with the HTGW phenotype. We suggested a combined use of hypertriglyceridemia waist phenotype in identifying participants who are at high risk of hypertension.
10.1111/jch.14424
Association Between Lifestyle and Hypertriglyceridemic Waist Phenotype in the PREDIMED-Plus Study.
Fernández-García José Carlos,Muñoz-Garach Araceli,Martínez-González Miguel Ángel,Salas-Salvado Jordi,Corella Dolores,Hernáez Álvaro,Romaguera Dora,Vioque Jesús,Alonso-Gómez Ángel M,Wärnberg Julia,Martínez J Alfredo,Serra-Majem Luís,Estruch Ramón,Lapetra José,Pintó Xavier,Tur Josep A,Garcia-Rios Antonio,García Molina Laura,Gaforio José Juan,Matía-Martín Pilar,Daimiel Lidia,Martín Sánchez Vicente,Vidal Josep,Prieto Lucia,Ros Emilio,Goñi Nuria,Babio Nancy,Ortega-Azorin Carolina,Castañer Olga,Konieczna Jadwiga,Notario Barandiaran Leyre,Vaquero-Luna Jessica,Benavente-Marín Juan Carlos,Zulet M Angeles,Sanchez-Villegas Almudena,Sacanella Emilio,Gómez Huelgas Ricardo,Miró-Moriano Leticia,Gimenez-Gracia Mariano,Julibert Alicia,Razquin Cristina,Basora Josep,Portolés Olga,Goday Albert,Galmés-Panadés Aina M,López-García Carmen M,Moreno-Rodriguez Anai,Toledo Estefanía,Díaz-López Andrés,Fitó Montserrat,Tinahones Francisco J,Bernal-López M Rosa,
Obesity (Silver Spring, Md.)
OBJECTIVE:The hypertriglyceridemic waist (HTGW) phenotype is characterized by abdominal obesity and high levels of triglycerides. In a cross-sectional assessment of PREDIMED-Plus trial participants at baseline, HTGW phenotype prevalence was evaluated, associated risk factors were analyzed, and the lifestyle of individuals with metabolic syndrome and HTGW was examined. METHODS:A total of 6,874 individuals aged 55 to 75 with BMI ≥ 27 and < 40 kg/m were included and classified by presence (HTGW ) or absence (HTGW ) of HTGW (waist circumference: men ≥ 102 cm, women ≥ 88 cm; fasting plasma triglycerides ≥ 150 mg/dL). Analytical parameters and lifestyle (energy intake and expenditure) were analyzed. RESULTS:A total of 38.2% of the sample met HTGW criteria. HTGW individuals tended to be younger, have a greater degree of obesity, be sedentary, and be tobacco users. They had higher peripheral glucose, total cholesterol, and low-density lipoprotein cholesterol levels; had lower high-density lipoprotein cholesterol levels; and had increased prevalence of type 2 diabetes mellitus. Mediterranean diet (MedDiet) adherence and physical activity were greater in HTGW patients. Age, BMI, tobacco use, total energy expenditure, hypertension, type 2 diabetes mellitus, and MedDiet adherence were associated with HTGW . CONCLUSIONS:HTGW is a highly prevalent phenotype in this population associated with younger age, higher BMI, tobacco use, and decreased MedDiet adherence. HTGW individuals were more physically active with greater total physical activity, and fewer had hypertension.
10.1002/oby.22728
The hypertriglyceridemic waist phenotype among women.
LaMonte Michael J,Ainsworth Barbara E,DuBose Katrina D,Grandjean Peter W,Davis Paul G,Yanowitz Frank G,Durstine J Larry
Atherosclerosis
BACKGROUND:Elevated plasma triglycerides (TG) and waist girth (hypertriglyceridemic waist (HTGW)) has been associated with elevated insulin, small dense low-density lipoprotein (sLDL) particles, and Apo B in men. The HTGW has not been reported for women and the effect of cardiorespiratory fitness ("fitness") on associations between HTGW and coronary risk factors is unknown. PURPOSE:To determine the prevalence of HTGW and the influence of fitness on the relationship between HTGW and coronary risk among 137 healthy women (54+/-9 year; body mass index (BMI)=28+/-6 kg/m(2)). METHODS:HTGW was defined as waist girth >88 cm and TG >150 mg/dl. The metabolic triad was defined as insulin >31 pmol/l, Apo B >69 mg/dl and LDL-C >84 mg/dl. Fitness was assessed with a maximal treadmill exercise test. RESULTS:The sample prevalence of HTGW (n=15) was 11% (95% CI=5.7-16.0%). Apo B (P=0.04) and insulin (P=0.0001) increased across quintiles of waist girth, and LDL-C (P=0.004) increased across quintiles of TG. Metabolic triad prevalence was highest (67%, n=10) among HTGW women and lowest (22%, n=26) among non-HTGW women. A trend for higher coronary heart disease CHD risk factors was observed among HTGW compared with non-HTGW women. Among the HTGW group, a trend for lower CHD risk factors was observed among fit (>or=6.5 METs, n=7) versus unfit women (<6.5 METs, n=8). Sample size limitations prohibited meaningful tests of significant differences in CHD risk factors when stratified simultaneously on HTGW and fitness status. CONCLUSIONS:HTGW is associated with increased coronary risk factors similarly among women as reported for men. Higher fitness may improve the CHD risk profile among women with HTGW.
10.1016/j.atherosclerosis.2003.07.008
Using interpretable machine learning methods to identify the relative importance of lifestyle factors for overweight and obesity in adults: pooled evidence from CHNS and NHANES.
BMC public health
BACKGROUND:Overweight and obesity pose a huge burden on individuals and society. While the relationship between lifestyle factors and overweight and obesity is well-established, the relative contribution of specific lifestyle factors remains unclear. To address this gap in the literature, this study utilizes interpretable machine learning methods to identify the relative importance of specific lifestyle factors as predictors of overweight and obesity in adults. METHODS:Data were obtained from 46,057 adults in the China Health and Nutrition Survey (2004-2011) and the National Health and Nutrition Examination Survey (2007-2014). Basic demographic information, self-reported lifestyle factors, including physical activity, macronutrient intake, tobacco and alcohol consumption, and body weight status were collected. Three machine learning models, namely decision tree, random forest, and gradient-boosting decision tree, were employed to predict body weight status from lifestyle factors. The SHapley Additive exPlanation (SHAP) method was used to interpret the prediction results of the best-performing model by determining the contributions of specific lifestyle factors to the development of overweight and obesity in adults. RESULTS:The performance of the gradient-boosting decision tree model outperformed the decision tree and random forest models. Analysis based on the SHAP method indicates that sedentary behavior, alcohol consumption, and protein intake were important lifestyle factors predicting the development of overweight and obesity in adults. The amount of alcohol consumption and time spent sedentary were the strongest predictors of overweight and obesity, respectively. Specifically, sedentary behavior exceeding 28-35 h/week, alcohol consumption of more than 7 cups/week, and protein intake exceeding 80 g/day increased the risk of being predicted as overweight and obese. CONCLUSION:Pooled evidence from two nationally representative studies suggests that recognizing demographic differences and emphasizing the relative importance of sedentary behavior, alcohol consumption, and protein intake are beneficial for managing body weight status in adults. The specific risk thresholds for lifestyle factors observed in this study can help inform and guide future research and public health actions.
10.1186/s12889-024-20510-z