A clinical prediction model for 90-day functional dependence after minor stroke/TIA with exploratory pharmacogenetic associations: a post-hoc analysis of the INSPIRES trial.
8 affiliations
- 1Beijing Tiantan Hospital, Capital Medical University, Beijing, China.
- 2Department of Neurology, Beijing Tiantan Hospital, Capital Medical University, Beijing, China.
- 3China National Clinical Research Center for Neurological Diseases, Beijing, China.
- 4Chinese Institute for Brain Research, Beijing, China.
- 5National Center for Neurological Disorders, Beijing, China.
- 6Advanced Innovation Center for Human Brain Protection, Capital Medical University, Beijing, China.
- 7Beijing Laboratory of Oral Health, Capital Medical University, Beijing, China.
- 8Laboratory for Clinical Medicine, Capital Medical University, Beijing, China.
Abstract
Dual antiplatelet therapy (DAPT) reduces stroke recurrence after minor stroke/TIA, yet a clinically meaningful proportion of patients remain functionally dependent at 90 days. We developed an interpretable clinical prediction model to define a high-risk subgroup, and separately explored pharmacogenetic associations within this subgroup to generate hypotheses for precision secondary prevention. In this post-hoc analysis of the INSPIRES trial (4,439 patients), LASSO logistic regression selected predictors for 90-day functional dependence (mRS > 2). Model performance was assessed by discrimination, calibration, and decision curve analysis. In a genotyped subcohort, CYP1A2 rs762551 and POR rs1057868 were evaluated for outcome association and enrichment in the high-risk subgroup. Logistic regression achieved the highest AUC point estimate (AUC = 0.727) among the four models evaluated (AUC range: 0.495-0.698; 95% CIs overlapped across models); it was selected as the final model for its interpretability and calibration, rather than for statistical superiority. Key predictors were admission NIHSS, age >65 years, and diabetes. The highest-risk quartile (25% of patients) carried a disproportionate outcome burden, with no effect modification by antiplatelet strategy (p = 0.793). CYP1A2 CC (23.65% vs. 16.28%; Pearson χ2p = 0.0031) and POR TC (54.58% vs. 46.22%; Pearson χ2p = 0.0105) genotypes were enriched in functionally dependent patients; CYP1A2 showed no significant allele-dosage trend (trend p = 0.0821), whereas POR demonstrated a significant allele-dosage trend (p = 0.0032). This clinical prediction model identified a high-risk subgroup accounting for most poor outcomes after minor stroke/TIA. Exploratory analyses identified CYP1A2 and POR variants enriched in this subgroup, generating pharmacogenetic hypotheses for future investigation.
Related
- Physiological Associations, Within-Session Repeatability, and Exploratory Prediction of Peak Oxygen Uptake Using the Revised 3-Minute Seated Stepping Test in Healthy Young Japanese Adults.
- CYP2C9 Pharmacogenetic Variants and Adverse Drug Reactions to Antiseizure Medications: An Exploratory Analysis in the Venezuelan Andes.
- Modeling Factors Associated with Smoking Dependence Severity Using Binary Classification-Based Machine Learning: Prediction with Hematological and Demographic Patient Data
- Associations Between Upright Cardiovascular-Autonomic Characteristics, Temporal Gait Measures, and Mediolateral Plantar Center-of-Force Measures in Adults with Developmental Disabilities: An Exploratory Cross-Sectional Study
- Restrictive Eating Patterns and Eating Disorder Risk in Female Ballet Dancers: Exploratory Associations with Self-Reported Workplace Bullying and Possible Mobbing
BibTeX
@article{wang2026clinical,
title = {{A clinical prediction model for 90-day functional dependence after minor stroke/TIA with exploratory pharmacogenetic associations: a post-hoc analysis of the INSPIRES trial.}},
author = {Yaping Wang and Baizhu Li and Hongyi Yan and Lingling Jiang and Yuesong Pan and Yilong Wang},
journal = {Frontiers in aging neuroscience},
year = {2026},
month = oct,
doi = {10.3389/fnagi.2026.1922100},
url = {https://europepmc.org/article/MED/42953014},
note = {CC BY},
}
RIS
TY - JOUR
TI - A clinical prediction model for 90-day functional dependence after minor stroke/TIA with exploratory pharmacogenetic associations: a post-hoc analysis of the INSPIRES trial.
AU - Yaping Wang
AU - Baizhu Li
AU - Hongyi Yan
AU - Lingling Jiang
AU - Yuesong Pan
AU - Yilong Wang
PY - 2026
DA - 2026/10/14
JO - Frontiers in aging neuroscience
DO - 10.3389/fnagi.2026.1922100
UR - https://europepmc.org/article/MED/42953014
AB - Dual antiplatelet therapy (DAPT) reduces stroke recurrence after minor stroke/TIA, yet a clinically meaningful proportion of patients remain functionally dependent at 90 days. We developed an interpretable clinical prediction model to define a high-risk subgroup, and separately explored pharmacogenetic associations within this subgroup to generate hypotheses for precision secondary prevention. In this post-hoc analysis of the INSPIRES trial (4,439 patients), LASSO logistic regression selected predictors for 90-day functional dependence (mRS > 2). Model performance was assessed by discrimination, calibration, and decision curve analysis. In a genotyped subcohort, CYP1A2 rs762551 and POR rs1057868 were evaluated for outcome association and enrichment in the high-risk subgroup. Logistic regression achieved the highest AUC point estimate (AUC = 0.727) among the four models evaluated (AUC range: 0.495-0.698; 95% CIs overlapped across models); it was selected as the final model for its interpretability and calibration, rather than for statistical superiority. Key predictors were admission NIHSS, age >65 years, and diabetes. The highest-risk quartile (25% of patients) carried a disproportionate outcome burden, with no effect modification by antiplatelet strategy (p = 0.793). CYP1A2 CC (23.65% vs. 16.28%; Pearson χ2p = 0.0031) and POR TC (54.58% vs. 46.22%; Pearson χ2p = 0.0105) genotypes were enriched in functionally dependent patients; CYP1A2 showed no significant allele-dosage trend (trend p = 0.0821), whereas POR demonstrated a significant allele-dosage trend (p = 0.0032). This clinical prediction model identified a high-risk subgroup accounting for most poor outcomes after minor stroke/TIA. Exploratory analyses identified CYP1A2 and POR variants enriched in this subgroup, generating pharmacogenetic hypotheses for future investigation.
SN - 1663-4365
ER -