Table 1.
General characteristics of included studies.
| No. | First author | Year of publication | Author country | Study type | Souce of patients | Stroke type | Diagnostic criteria for cognitive impairment |
|---|---|---|---|---|---|---|---|
| 1 | Yinwei Zhu (12) | 2022 | China | RCT | Multicenter | Acute ischemic stroke | MMSE < 25 |
| 2 | Fei Zha (13) | 2022 | China | Retrospective cohort study | Single center | Cerebral stroke | MMSE score ≤ 19 (illiteracy), ≤ 22 (primary education), ≤ 26 (Secondary school and above) |
| 3 | Georgios Vlachos (14) | 2023 | Norway | Retrospective cohort study | Multicenter | Mild acute stroke | the Barthel ADL index and the modified Rankin Scale (mRS). |
| 4 | Lisa R¨ohrig (15) | 2022 | Germany | Prospective cohort study | Multicenter | Right hemisphere stroke | letter cancelation test; bells cancelation test the Center of Cancelation [CoC; (35)]; The CoC |
| 5 | Ragnhild (16) | 2022 | Norway | Prospective cohort study | Multicenter | Cerebral stroke | Premorbid cognitive status based on GDS |
| 6 | Zhao-Yin Ma (17) | 2022 | China | RCT | Single center | Acute ischemic stroke | MoCA score ≥ 26 indicates normal cognitive function; < 26 indicates MCI; < 20 indicates CI |
| 7 | Reeree Lee (18) | 2021 | Republic of Korea | Prospective cohort study | Multicenter | Cerebral stroke | Objective neuropsychology tests, including MMSE and CDR |
| 8 | Yongzhe Gu (19) | 2022 | China | Retrospective cohort study | Multicenter | Ischemic stroke | MoCA score < 26 |
| 9 | Nacim Betroun (20) | 2022 | Australia | Prospective cohort study | Multicenter | Cerebral stroke | MMSE score < 27 or MoCA score < 25 |
| 10 | Xueling Yuan (21) | 2021 | China | Retrospective cohort study | Single center | Cerebral stroke | MMSE score ≤ 17 (illiteracy), ≤ 20 (Primary education), ≤ 24 (Secondary school and above), MOCA score ≤ 26 |
| 11 | Nick A Weaver (22) | 2021 | Netherlands | Retrospective cohort study | Multicenter | Cerebral stroke | Performance below the fifth percentile of local normative data in at least one cognitive domain on the Multidomain Neuropsychological Assessment or the Montreal Cognitive Assessment |
| 12 | Renaud Lopes (23) | 2021 | franc | Retrospective cohort study | Multicenter | Cerebral stroke | IQCODE 49 ± 2 |
| 13 | Youssef Hbid (24) | 2021 | UK | Prospective cohort study | Single center | First occurrence of cerebral stroke | MMSE score < 24 or AMT < 8 |
| 14 | Li Gong (25) | 2021 | China | Prospective cohort study | Multicenter | Mild acute stroke | MoCA score < 22 |
| 15 | Yi Dong (26) | 2021 | China | Retrospective cohort study | Multicenter | Acute ischemic stroke | MoCA score < 22 |
| 16 | Eva Birgitte Aamodt (27) | 2021 | Norway | Prospective cohort study | Multicenter | Cerebral stroke | TMT A and B, CERAD, COWAT, MoCA, AD-8, GDS (36), NPI Q, HADS, and the Cornell scale |
| 17 | Yueli Zhu (28) | 2020 | China | Prospective cohort study | Multicenter | First occurrence of cerebral stroke | MMSE score < 27 and MoCA score < 21 |
| 18 | Zhengbao Zhu (29) | 2019 | China | Prospective cohort study | Multicenter | Ischemic stroke with elevated blood pressure | MMSE score < 27 or MoCA score < 25 |
| 19 | Jae-Sung Lim (30) | 2017 | Republic of Korea | Prospective cohort study | Multicenter | Cerebral stroke | AHA-ASA Criteria, at least two cognitive defects |
| 20 | Nagaendran Kandiah (31) | 2015 | Singapore | Retrospective cohort study | Multicenter | Mild acute ischemic stroke | MMSE score ≤ 2 5 or MoCA score ≤ 22 |
| 21 | Sheng Ye (32) | 2022 | China | Retrospective cohort study | Single center | Lacunar infarction | MMSE score < 24 |
| Follow-up duration | Number of patients with cognitive impairment after stroke | Total sample size | Number of patients with cognitive impairment in training set | Total sample size of training set (deriving set and modeling cohort) | Validation set generation method [internal validation (k-fold cross-validation, leave-one-out method, random sampling), external validation (prospective, institutional)] | Number of patients with cognitive impairment in validation set | Total sample size of validation set | Variable screening/feature selection method | Model type |
|---|---|---|---|---|---|---|---|---|---|
| 3 m | 228 | 599 | 228 | 599 | No validation set | Multivariate | Logistic regression | ||
| 3 m | 87 | 367 | 58 | 245 | Random sampling at a ratio of 2:1 | 29 | 122 | Multivariate | Logistic regression |
| 12 m | 21 | 117 | 21 | 117 | No validation set | Multivariate | Logistic regression | ||
| 27 | 103 | 27 | 103 | No validation set | Multivariate | Logistic regression | |||
| 3 m | 91 | 589 | 91 | 589 | No validation set | No | No | Multivariate | Logistic regression |
| 94 | 161 | 94 | 161 | No validation set | No | No | Multivariate | Logistic regression | |
| 6 m | 19 | 110 | 19 | 110 | Random internal validation | 12 | 70 | Multivariate | Logistic regression |
| 6 m | 69 | 123 | 69 | 123 | External validation | 38 | 60 | Multivariate | Logistic regression |
| 1y | 77 | 327 | 62 | 262 | 5-fold cross validation | 15 | 65 | Multivariate | random forest |
| 1y | 118 | 376 | 118 | 376 | External validation | 75 | 125 | Multivariate | LASSO regression |
| 106 | 338 | 10-fold cross validation | 12 | 38 | |||||
| 15 m | 1,286 | 2,950 | 1,286 | 2,950 | External validation | 107 | 246 | Multivariate | Logistic regression |
| 1,179 | 1,704 | ||||||||
| 12-fold cross validation | |||||||||
| 36y | 9 | 72 | 9 | 72 | External validation | 40 | Multivariate | Ridge regression | |
| 5y | 1,000 | 2,468 | 1,000 | 2,468 | External validation | 204 | 940 | Multivariate | Mixed effects model |
| Internal validation | |||||||||
| 1y | 112 | 228 | 112 | 228 | External validation | No | 1,000 | Multivariate | Multivariate logistic regression |
| 6 m | 131 | 383 | 131 | 383 | External validation | 281 | Multivariate | Multivariate logistic regression | |
| Internal validation | 102 | ||||||||
| 3 m | 125 | 227 | 125 | 227 | Leave-one-out cross validation | 227 | Multivariate | Classifier model | |
| 6 m | 66 | 104 | 66 | 104 | Internal cross validation | 66 | 104 | Multivariate | Decision tree |
| 3 m | 340 MMSE scoring | 638 | No | Multivariate | Logistic regression | ||||
| 422 MoCA scoring | |||||||||
| 3 m | 50 | 308 | No | Multivariate | Logistic regression | ||||
| 6 m | 78 | 209 | 70 | 88 | 10-fold cross validation | 8 | 21 | Multivariate | Logistic regression |
| 78 | 209 | External validation | 35 | 185 | |||||
| 12 m | 52 | 313 | 38 | 219 | Random sampling at a ratio of 7:3 | 14 | 94 | Multivariate | Logistic regression |
MMSE, mini-mental state examination; CDR, clinical dementia rating; GDS, global deterioration scale; MoCA, Montreal cognitive assessment; IQCODE, informant questionnaire on cognitive decline in the elderly; AMT, abbreviated mental test; TMT A and B, neuropsychological test battery included Trail making A and B; CERAD, 10 word memory and recall test; COWAT, the controlled oral word association test; AD-8, ascertain dementia 8-item informant questionnaire; NPI Q, neuropsychiatric inventory; HADS, hospital anxiety and depression scale.