# Domain-specific cognitive screening in acute first-ever stroke: a comparative study of the Oxford Cognitive Screen (OCS) and ACE-III

## TL;DR

This study compares two cognitive tests for detecting cognitive impairment after stroke, finding that a stroke-specific tool is more reliable in the acute phase.

## Contribution

The study evaluates the diagnostic performance of ACE-III versus the stroke-specific OCS for detecting post-stroke cognitive impairment.

## Key findings

- The OCS detected PSCI in 70% of patients, while ACE-III missed five cases identified by OCS.
- ACE-III misclassified two aphasic patients as cognitively impaired, highlighting its limitations in stroke populations.
- Standard and stroke-specific ACE-III cut-offs showed suboptimal accuracy for acute cognitive screening.

## Abstract

Post-stroke cognitive impairment (PSCI) is common and often under-recognized, particularly in the acute phase. Most cognitive screening tools provide only a global score, overlooking domain-specific deficits that influence functional recovery. The Addenbrooke’s Cognitive Examination-III (ACE-III) is a comprehensive cognitive test whose utility for acute stroke patients remains under-studied. This study evaluated the diagnostic performance of the ACE-III against the stroke-specific Oxford Cognitive Screen (OCS) in detecting PSCI following first-ever stroke in the acute period.

Patients with first-ever stroke and no history of cognitive impairment were prospectively assessed within seven days of onset using both the OCS and ACE-III. PSCI was defined by impairment in one or more cognitive domains on the OCS. The discriminatory capacity of the ACE-III for detecting PSCI was examined, and associations between specific cognitive deficits and functional dependence were analyzed.

The OCS detected PSCI in 70% of the 30 patients that were recruited. The ACE-III demonstrated good discriminatory capacity (AUC = 0.897); however, it failed to detect PSCI in five patients identified by the OCS, and misclassified two aphasic but cognitively intact patients as impaired. Two patients classified as impaired on ACE-III were deemed cognitively intact by OCS, underscoring its limitations in stroke populations. Standard and stroke-specific ACE-III cut-offs demonstrated suboptimal accuracy for acute screening.

While ACE-III performs well at the group level, it may miss relevant cognitive impairment in the acute stroke setting. Domain-based, stroke-specific tools such as the OCS more reliably detect deficits and may offer greater clinical utility for early cognitive profiling and rehabilitation.