Post-Stroke Cognitive Impairment in Georgia: Diagnostic Utility of Georgian NINDS-CSN Neuropsychological Battery

Background: Post-stroke cognitive impairment (PSCI) affects approximately half of stroke survivors within the first year. Data from low- and middle-income countries, including Georgia, are limited. Although a Georgian-language version of the Montreal Cognitive Assessment (MoCA) has previously been validated as a brief cognitive screening instrument, no validated Georgian-language multidomain neuropsychological battery for vascular or post-stroke cognitive impairment is currently available. This study aimed to validate the Georgian version of the National Institute of Neurological Disorders and Stroke–Canadian Stroke Network (NINDS–CSN) neuropsychological battery by assessing its concurrent validity with the Montreal Cognitive Assessment (MoCA), and to evaluate its clinical utility for estimating PSCI prevalence. Methods: Patients with ischemic stroke attending a 6-month follow-up visit were prospectively enrolled. The control group (CG) comprised healthy adults. The Georgian NINDS–CSN 30 min battery was administered. Validity for diagnosing PSCI was assessed using receiver operating characteristic (ROC) curve analysis, with area under the curve (AUC) as the primary measure. Results: A total of 175 participants were included (73 stroke patients and 102 healthy controls); 34 of 73 stroke patients (46.6%) met the clinical reference criteria for PSCI. In demographically adjusted analyses (n = 72), the four-test NINDS-CSN composite demonstrated good discrimination for clinically defined PSCI (AUC, 0.821); the optimal threshold of −1.144 yielded 78.8% sensitivity and 82.1% specificity. Incorporation of TMT-A and TMT-B into a six-test composite modestly increased the AUC to 0.861, with an optimal threshold of −1.728, 72.7% sensitivity, and 94.9% specificity. The MoCA demonstrated higher discriminative performance (AUC, 0.978; 95% CI, 0.952–1.000); at the previously validated cutoff of ≤22, sensitivity was 97.0%, and specificity was 79.5%. The MoCA AUC was significantly higher than that of the six-test NINDS-CSN composite (AUC difference, 0.117; 95% CI, 0.031–0.204; p = 0.008). Conclusions: The Georgian NINDS–CSN battery demonstrated good discriminative performance. However, given the superior performance of the Georgian MoCA, it may be preferred for screening PSCI in routine practice, while the NINDS–CSN battery remains suitable for comprehensive cognitive assessment.