Stroke is indiscriminate and has the potential to erase a lifetime of experiences or abilities in an instant. Given that stroke sequelae can have long-lasting consequences, with many stroke survivors experiencing ongoing adverse effects several years later, there is mounting evidence supporting stroke not only as an acute but also as a chronic medical condition. Cognitive impairment is a frequently reported symptom post stroke, occurring in more than two thirds of stroke patients. Unlike physical impairments which can often recover following stroke, cognitive impairment becomes increasingly worse over time resulting in poor prognosis. Moreover, symptoms commonly associated with cognitive impairment are often overlooked, leading to increased rates of admission to private hospitals and higher rates of mortality. This is further complicated by the fact that assessment of post stroke cognitive function is not routinely carried out therefore prevalence of cognitive deficits may be underestimated. Existing data on long-term outcomes specific to stroke are scarce in NZ, particularly in terms of cognition. The purpose of this thesis was to determine prevalence, profiles, predictors, and trajectory of cognitive impairment in a cohort of four-year stroke survivors in NZ. Additionally, this thesis also sought to determine the accuracy of a screening tool to detect post stroke cognitive impairment.

This population-based follow-up study adopts a quantitative approach and includes a sample drawn from the fourth Auckland Regional Stroke Outcomes Study (ARCOS-IV). Two and hundred and fifty-seven people with stroke, completed cognitive assessments at baseline or within two weeks of stroke, and/or 1 month, 6, 12 and 48 months after stroke onset.

At four years post-stroke, the greatest proportion (84%) of stroke survivors exhibited below average cognitive functioning as assessed by the Montreal Cognitive Assessment Scale (MoCA) (M = 20.7, ± 4.70). Specific domains of cognition most affected by chronic stroke included attention, memory, information processing speed and executive functioning. While there was some recovery of cognitive function within the first 12 months, cognition steadily declined from this time-point to four-years. Regression analysis revealed baseline predictors including; older age (>75 years), male gender, not working, greater cognitive impairment at baseline and stroke-related vascular risk factors (i.e. hypertension, diabetes mellitus and arrhythmia) were associated with declining cognition over the four-year period. Vascular territory and stroke subtype as classified by the Oxfordshire Stroke Community Stroke Project (OCSP) were associated with specific-domain impairments, more than hemisphere, pathological stroke type and lesion location. ROC analysis comparing the MoCA to a more comprehensive computerized neuropsychological battery (CNS-VS), revealed the MoCA was found to have acceptable sensitivity and specificity to detect global cognitive impairment and high sensitivity to detect domain-specific impairment in executive function, visuo-spatial ability and memory.

This study identified that cognitive impairment following stroke does not resolve and continues to decline over time in the vast majority of people with stroke. The trajectory of cognitive recovery is influenced by sociodemographic and stroke-related risk factors. Additionally, assessment of post-stroke global cognitive impairment and certain domain-specific impairments could be accurately detected using the MoCA. Early identification of individual cognitive domains and risk factors that result in cognitive decline may inform the development of rehabilitation interventions targeting restoration, remediation and adaptation of cognitive function. The first step in the management of cognitive impairment is to educate health professionals working in stroke units and hospitals about the importance of recognising and diagnosing deficits prior to discharge. This can be achieved by standardisation of the MoCA in all health settings as the primary method for initial assessment of post stroke cognitive impairment. Finally, evidence gathered in this research should be used to further investigate the effects of early assessment, rehabilitation and frequent monitoring of cognitive function throughout the stroke journey.