列线图预测代谢综合征患者心功能减退的风险

Nomogram predicts the risk of cardiac dysfunction in patients with metabolic syndrome

目的建立列线图预测模型识别代谢综合征(MS)患者中可能出现心功能减退的高风险患者。方法将本院511例MS患者随机分为训练队列369例和验证队列142例。收集患者的基本信息、血液学检测指标及超声心动图指标。根据超声心动图结果,将训练队列患者划分为左室舒张功能减低(LVDD)组132例与no-LVDD组237例。通过多因素Logistic回归筛选与LVDD相关的独立影响因素。分别基于训练队列和验证队列对列线图模型的性能进行内部与外部验证。结果腰围、合并症数量、糖化血红蛋白(HbA1c)、整体纵向应变(GLS)是MS患者心功能减退的独立影响因素(P<0.05)。所构建的列线图预测模型在内部及外部验证中都具有良好的区分度(AUC:内部验证0.910,外部验证0.745)和一致性(Hosmer-Lemeshow检验:内部验证χ^(2)=5.261,P=0.729;外部验证χ^(2)=6.142,P=0.434)。决策曲线分析显示,在训练和验证队列中,模型均具有较高净收益。结论通过综合腰围、合并症数量、HbA1c和GLS建立的列线图模型可有效预测MS患者心功能减退的风险。

Objective To establish a nomogram to identify patients at high risk of possible cardiac dysfunction in patients with metabolic syndrome(MS).Methods A total of 511 MS patients from our hospital were selected.The patients were randomly divided into a training cohort(n=369)and a validation cohort(n=142).The basic information,hematological indicators and echocardiographic indicators of the patients were collected.Based on the echocardiography results,the patients in the training cohort were further divided into a left ventricular diastolic dysfunction(LVDD)group(n=132)and a non-LVDD group(n=237).Independent influencing factors associated with LVDD were screened by using multivariate logistic regression analysis.The performance of the nomogram model was internally and externally validated based on the training and validation cohorts,respectively.Results Waist circumference,number of comorbidities,glycated hemoglobin(HbA1c)and global longitudinal strain(GLS)were independent influencing factors for cardiac dysfunction in MS patients(P<0.05).The constructed nomogram was demonstrated to have good discrimination(AUC:0.910 for internal validation and 0.745 for external validation)and calibration(Hosmer-Lemeshow test:χ^(2)=5.261,P=0.729 for internal validation andχ^(2)=6.142,P=0.434 for external validation).Decision curve analysis showed that the model had high net benefit in both the training and validation cohorts.Conclusions The nomogram model established by integrating waist circumference,comorbidity count,HbA1c and GLS may effectively predict the risk of cardiac dysfunction in MS patients.