This study developed and internally validated a multivariable model combining age, prior stroke, diabetes, sdLDL, ApoA-I, homocysteine, cystatin C, and D-dimer to estimate PSCI risk after acute ischemic stroke, showing acceptable predictive performance that may support early risk stratification.
In a prospective cohort of 198 acute ischemic stroke patients, 138 (about 70%) had post-stroke cognitive impairment 3 months after stroke.
The final multivariable prediction model included eight predictors: age, prior stroke history, diabetes mellitus, small dense LDL (sdLDL), apolipoprotein A-I (ApoA-I), homocysteine, cystatin C, and D-dimer.
The model showed acceptable discrimination, with an apparent AUC of 0.798 that decreased slightly after internal bootstrap validation.
Calibration analysis suggested reasonable agreement between predicted and observed PSCI risk.
Decision curve analysis and clinical impact curves indicated potential utility of the model for PSCI risk stratification.
The model integrates routinely available blood biomarkers (lipid, renal, metabolic, and coagulation markers) with clinical characteristics rather than relying on clinical variables alone.
This study looked at whether a combination of simple patient information and routine blood tests taken after a stroke could help predict who will develop thinking and memory problems within three months. The researchers followed 198 people who had an ischemic stroke (caused by a blocked blood vessel) at one hospital in Shanghai. About 70% (138 patients) had cognitive impairment at three months, as measured by a telephone-based cognitive test. The researchers built a prediction tool using eight factors: age, previous stroke, diabetes, and five blood markers (small dense LDL cholesterol, apolipoprotein A-I, homocysteine, cystatin C, and D-dimer). The tool correctly separated patients who did and did not develop impairment reasonably well. Its accuracy score (AUC) was about 0.76 after statistical adjustment for over-optimism, where 0.5 is chance and 1.0 is perfect. Additional analyses suggested the predicted risks matched what was observed and that the tool might be useful for identifying higher-risk patients. This research suggests that routinely available blood tests could help flag patients at higher risk of cognitive problems after stroke. However, the study was small, came from a single hospital, used a telephone-based cognitive screening, and was validated only internally, so the model needs testing in other groups of patients before it could be used in routine care.
Zhang X, Hao X, Li Y, Chen J, Zheng X, Zhang A, et al.. (2026). A clinical and biomarker model for predicting post-stroke cognitive impairment: a prospective cohort study.. Frontiers in neurology. https://doi.org/10.3389/fneur.2026.1963734