Risk factors of recurrence of acute ischemic stroke and construction of a nomogram model for predicting the recurrence risk based on Lasso Regression.
10.12122/j.issn.1673-4254.2024.12.13
- Author:
Jiaxin JIN
1
;
Pengzhen MA
2
;
Eryu WANG
1
;
Yingzhen XIE
1
Author Information
1. Dongzhimen Hospital, Beijing University of Chinese Medicine, Beijing 100700, China.
2. Traditional Chinese Medicine Data Center, China Academy of Traditional Chinese Medicine, Beijing 100700, China.
- Publication Type:Journal Article
- Keywords:
Lasso regression;
influencing factors;
ischemic stroke;
predictive model;
recurrence
- MeSH:
Humans;
Nomograms;
Risk Factors;
Recurrence;
Ischemic Stroke/etiology*;
Female;
Male;
Logistic Models;
Aged;
ROC Curve;
Middle Aged;
Stroke/etiology*
- From:
Journal of Southern Medical University
2024;44(12):2375-2381
- CountryChina
- Language:Chinese
-
Abstract:
OBJECTIVES:To investigate the risk factors of recurrence of acute ischemic stroke (AIS) within 1 year and establish a nomogram model for predicting the recurrence risk.
METHODS:This study was conducted in two cohorts of AIS patients (≤7 days) hospitalized in Dongzhimen Hospital (modeling set) and Fangshan Hospital (validation set) from March, 2021 to March, 2022. Lasso regression analysis was used to identify the important predictive factors for AIS recurrence within 1 year, and multivariate Logistic regression analysis was performed to analyze the independent factors affecting AIS recurrence. The recurrence risk prediction nomogram model was constructed using R studio, and its discriminating power and calibration were assessed using ROC curve analysis and Hosmer-Lemeshow goodness-of-fit test.
RESULTS:The modeling and validation sets contained 28 cases (15.22%) and 21 cases (15.00%) of AIS recurrence, respectively. In the modeling set, compared with the non-relapse group, the recurrence group had higher proportions of patients with age >65 years, diabetes, arrhythmia, constipation after stroke, and FBG >7.5 and significantly higher levels of NLR, UREA, Cr, HbA1c, FIB and TT (P<0.05). Multivariate Logistic regression analysis showed that an age >65 years, arrhythmia, constipation after stroke, FBG >7.5, NLR and Cr were all independent risk factors of AIS recurrence (P<0.05). Hosmer-Lemeshow goodness-of-fit test and calibration curve analysis showed that the risk prediction model had good fitting between the modeling set and the verification set. The ROC curve showed that for predicting AIS recurrence within 1 year, the AUC of the predictive model was 0.857 (95%CI: 0.782-0.932) in the modeling set and 0.679 (95%CI: 0.563-0.794) in the validation set.
CONCLUSIONS:The nomogram model established based on age >65 years, arrhythmia, constipation after stroke, FBG >7.5, NLR and Cr has a good predictive value for AIS recurrence within 1 year.
