Predictors and clinical implications of post-stroke cognitive impairment: a retrospective study
Abstract Investigate prevalence and determinants of Post-Stroke Cognitive Impairment (PSCI) to guide clinical management. Retrospective study of 792 first-time stroke patients (2020–2022) categorized into PSCI (n = 437) and non-PSCI (n = 355) groups via Mini-Mental State Examination (MMSE) scores. V...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-07-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-08048-5 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract Investigate prevalence and determinants of Post-Stroke Cognitive Impairment (PSCI) to guide clinical management. Retrospective study of 792 first-time stroke patients (2020–2022) categorized into PSCI (n = 437) and non-PSCI (n = 355) groups via Mini-Mental State Examination (MMSE) scores. Variables including demographics, medical history, lesion characteristics, and clinical assessments were analyzed. LASSO regression identified significant predictors, followed by multivariate logistic regression. A nomogram model was validated using ROC, calibration, and decision curve analysis (DCA). PSCI incidence was 55.18%. Risk factors included age ≥ 60 (OR = 6.806), BMI ≥ 24 (OR = 2.69–6.53), basal ganglia/frontotemporal-parietal lesions (OR = 6.04–6.20), white matter changes (OR = 3.18), smoking (OR = 2.05), diabetes (OR = 3.42), high diastolic blood pressure variability (OR = 1.39), elevated NIHSS (OR = 1.19), and depression (OR = 1.09). Protective factors were high school (OR = 0.48) or college education (OR = 0.17) and preserved lower limb muscle strength (OR = 0.83). The nomogram achieved AUC = 0.925 (sensitivity = 88.80%, specificity = 81.70%), with robust calibration and clinical utility per DCA. Over half of stroke patients developed PSCI. Key modifiable risks (BMI, blood pressure control, diabetes management) and non-modifiable factors (lesion location, age) were identified. Education and rehabilitation targeting muscle strength may mitigate risk. The predictive model aids early high-risk patient identification, supporting tailored interventions to improve outcomes. |
|---|---|
| ISSN: | 2045-2322 |