{"title":"肥胖和非肥胖人群中 N 端前 B 型钠尿肽与全因死亡率和心血管死亡率的关系以及机器学习预测模型的开发:国家健康与营养调查(NHANES)1999-2004。","authors":"Han Zhou, Chen Yang, Jingjie Li, Lin Sun","doi":"10.1111/dom.15927","DOIUrl":null,"url":null,"abstract":"<p><strong>Aims: </strong>To explore the potential of N-terminal pro-B natriuretic peptide (NTproBNP) in identifying adverse outcomes, particularly cardiovascular adverse outcomes, in a population with obesity, and to establish a risk prediction model.</p><p><strong>Methods: </strong>The data for this study were obtained from the National Health and Nutrition Examination Survey (NHANES) for 6772 participants without heart failure, for the years 1999 to 2004. Multivariable Cox regression models, cubic spline restricted models and Kaplan-Meier curves were used to evaluate the relationship between NTproBNP and both all-cause mortality and cardiovascular mortality. Predictive models were established using seven machine learning methods, and evaluation was conducted using precision, recall, F1 score, accuracy, and area under the curve (AUC) values.</p><p><strong>Results: </strong>During the population follow-up, out of 6772 participants, 1554 died, with 365 deaths attributed to cardiovascular disease. After adjusting for relevant covariates, NTproBNP levels ≥300 pg/mL were positively associated with both all-cause mortality (hazard ratio [HR] 3.00, 95% confidence interval [CI] 2.48, 3.67) and cardiovascular mortality (HR 6.05, 95% CI 3.67, 9.97), and remained significant across different body mass index (BMI) strata. However, in participants without abdominal obesity, the correlation between NTproBNP and cardiovascular mortality was significantly reduced. Among the seven machine learning methods, logistic regression demonstrated better predictive performance for both all-cause mortality (AUC 0.86925) and cardiovascular mortality (AUC 0.85115). However, establishing accurate cardiovascular mortality prediction models for non-abdominal obese individuals proved challenging.</p><p><strong>Conclusion: </strong>The study showed that NTproBNP can serve as a predictive factor for all-cause mortality and cardiovascular mortality in individuals with different BMIs, including obese individuals. However, significant cardiovascular mortality correlation was observed only for NTproBNP levels ≥300 pg/mL, and only among participants with abdominal obesity.</p>","PeriodicalId":158,"journal":{"name":"Diabetes, Obesity & Metabolism","volume":null,"pages":null},"PeriodicalIF":5.4000,"publicationDate":"2024-09-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Association of N-terminal pro-B natriuretic peptide with all-cause mortality and cardiovascular mortality in obese and non-obese populations and the development of a machine learning prediction model: National Health and Nutrition Examination Survey (NHANES) 1999-2004.\",\"authors\":\"Han Zhou, Chen Yang, Jingjie Li, Lin Sun\",\"doi\":\"10.1111/dom.15927\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><strong>Aims: </strong>To explore the potential of N-terminal pro-B natriuretic peptide (NTproBNP) in identifying adverse outcomes, particularly cardiovascular adverse outcomes, in a population with obesity, and to establish a risk prediction model.</p><p><strong>Methods: </strong>The data for this study were obtained from the National Health and Nutrition Examination Survey (NHANES) for 6772 participants without heart failure, for the years 1999 to 2004. Multivariable Cox regression models, cubic spline restricted models and Kaplan-Meier curves were used to evaluate the relationship between NTproBNP and both all-cause mortality and cardiovascular mortality. Predictive models were established using seven machine learning methods, and evaluation was conducted using precision, recall, F1 score, accuracy, and area under the curve (AUC) values.</p><p><strong>Results: </strong>During the population follow-up, out of 6772 participants, 1554 died, with 365 deaths attributed to cardiovascular disease. After adjusting for relevant covariates, NTproBNP levels ≥300 pg/mL were positively associated with both all-cause mortality (hazard ratio [HR] 3.00, 95% confidence interval [CI] 2.48, 3.67) and cardiovascular mortality (HR 6.05, 95% CI 3.67, 9.97), and remained significant across different body mass index (BMI) strata. However, in participants without abdominal obesity, the correlation between NTproBNP and cardiovascular mortality was significantly reduced. Among the seven machine learning methods, logistic regression demonstrated better predictive performance for both all-cause mortality (AUC 0.86925) and cardiovascular mortality (AUC 0.85115). However, establishing accurate cardiovascular mortality prediction models for non-abdominal obese individuals proved challenging.</p><p><strong>Conclusion: </strong>The study showed that NTproBNP can serve as a predictive factor for all-cause mortality and cardiovascular mortality in individuals with different BMIs, including obese individuals. However, significant cardiovascular mortality correlation was observed only for NTproBNP levels ≥300 pg/mL, and only among participants with abdominal obesity.</p>\",\"PeriodicalId\":158,\"journal\":{\"name\":\"Diabetes, Obesity & Metabolism\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":5.4000,\"publicationDate\":\"2024-09-06\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Diabetes, Obesity & Metabolism\",\"FirstCategoryId\":\"3\",\"ListUrlMain\":\"https://doi.org/10.1111/dom.15927\",\"RegionNum\":2,\"RegionCategory\":\"医学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"ENDOCRINOLOGY & METABOLISM\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Diabetes, Obesity & Metabolism","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1111/dom.15927","RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ENDOCRINOLOGY & METABOLISM","Score":null,"Total":0}
Association of N-terminal pro-B natriuretic peptide with all-cause mortality and cardiovascular mortality in obese and non-obese populations and the development of a machine learning prediction model: National Health and Nutrition Examination Survey (NHANES) 1999-2004.
Aims: To explore the potential of N-terminal pro-B natriuretic peptide (NTproBNP) in identifying adverse outcomes, particularly cardiovascular adverse outcomes, in a population with obesity, and to establish a risk prediction model.
Methods: The data for this study were obtained from the National Health and Nutrition Examination Survey (NHANES) for 6772 participants without heart failure, for the years 1999 to 2004. Multivariable Cox regression models, cubic spline restricted models and Kaplan-Meier curves were used to evaluate the relationship between NTproBNP and both all-cause mortality and cardiovascular mortality. Predictive models were established using seven machine learning methods, and evaluation was conducted using precision, recall, F1 score, accuracy, and area under the curve (AUC) values.
Results: During the population follow-up, out of 6772 participants, 1554 died, with 365 deaths attributed to cardiovascular disease. After adjusting for relevant covariates, NTproBNP levels ≥300 pg/mL were positively associated with both all-cause mortality (hazard ratio [HR] 3.00, 95% confidence interval [CI] 2.48, 3.67) and cardiovascular mortality (HR 6.05, 95% CI 3.67, 9.97), and remained significant across different body mass index (BMI) strata. However, in participants without abdominal obesity, the correlation between NTproBNP and cardiovascular mortality was significantly reduced. Among the seven machine learning methods, logistic regression demonstrated better predictive performance for both all-cause mortality (AUC 0.86925) and cardiovascular mortality (AUC 0.85115). However, establishing accurate cardiovascular mortality prediction models for non-abdominal obese individuals proved challenging.
Conclusion: The study showed that NTproBNP can serve as a predictive factor for all-cause mortality and cardiovascular mortality in individuals with different BMIs, including obese individuals. However, significant cardiovascular mortality correlation was observed only for NTproBNP levels ≥300 pg/mL, and only among participants with abdominal obesity.
期刊介绍:
Diabetes, Obesity and Metabolism is primarily a journal of clinical and experimental pharmacology and therapeutics covering the interrelated areas of diabetes, obesity and metabolism. The journal prioritises high-quality original research that reports on the effects of new or existing therapies, including dietary, exercise and lifestyle (non-pharmacological) interventions, in any aspect of metabolic and endocrine disease, either in humans or animal and cellular systems. ‘Metabolism’ may relate to lipids, bone and drug metabolism, or broader aspects of endocrine dysfunction. Preclinical pharmacology, pharmacokinetic studies, meta-analyses and those addressing drug safety and tolerability are also highly suitable for publication in this journal. Original research may be published as a main paper or as a research letter.