Background: Stroke is a leading cause of death and morbidity in all countries, yet treatment options are few. Numerous agents that were successful in animal models, failed in humans. Establishing the cause of stroke in the individual patient from the heterogeneous stroke mechanisms and measurement of clinical deficit including cognitive impairment in stroke are pivotal in successful treatment. An indigenous stroke data bank was established with specific emphasis on aetiology of stroke and higher cortical function measurement. Aim: 1. Establishment of an indigenous stroke data bank using contemporary neuroinvestigative modalities to determine stroke mechanism as precisely as possible. 2. To determine in this population, the frequency and extent of cognitive disorders in the acute and subacute stroke period, using a battery of predefined higher cortical function tests applied to all patients. 3. Collation of a comprehensive array of epidemiological, clinical, investigative and prognostic variables in complete digitised storage form. Methods: The patient population was a hospital based consecutive case series with an inpatient and outpatient stroke service in association with an acute stroke unit. A three tier investigative protocol was devised to incorporate contemporary neuroinvestigative modalities. All patients had mandatory investigations of stroke relevant blood tests, electrocardiogram, chest radiograph and brain scan. All patients were evaluated with a comprehensive battery of predefined, bedside higher cortical function tests. Standardised neurological deficit, clinical stroke scales, aetiological scales and disability scales were incorporated to quantitate deficit, stroke subtype and handicap at presentation. All patients were evaluated by the author and all information digitised by the author into the computerised registry - Durban Stroke Data Bank (DSDB). Results 1. Stroke Data Bank Issues: The first 1000 patients evaluated comprised of 561 men, 439 women, 781 Whites, 103 Asian Indians, 100 Blacks, 14 of Mixed Race and 2 other race groups. All patients had either a CT brain scan (698;69.8%), MRI brain scan (426;42.6%) or both (124;12.4%). Single Photon Emission Computed Tomography scans were performed in 104 (10.4%). Among the 23 different symptoms coded for, long tract signs, vision abnormalities and speech impairment predominated but 150 (15%) had additional other symptoms not coded for. Among the 29 different risk factors coded for, hypertension (42.1%), smoking (26.7%), cardiac illness (17.7%), Diabetes Mellitus (10.4%) and carotid stenosis (25.1%) were the most numerous. Approximately 96 different causes and possible causes of stroke were identified. The clinical ischaemic stroke classification (OCSP) revealed partial anterior circulation strokes in 447 (44.7%), posterior circulation in 258 (25.8%), total anterior circulation in 185 (18.5%) and lacunar in 82 (8.2%). The aetiological classification identified a large proportion of strokes due to "other" (253;25.3%) causes as opposed to large (264;26.4%) and small vessel disease (262;26.2%) or cardioembolism (122;12.2%). In 99 (9.9%) patients no cause could be established. The haemorrhage group was small (48;4.8%). Comparison of the clinical and aetiological classifications showed a significant difference overall (Chi square p-value=0.001). Black race had relatively higher other causes (39%) and unknown (20%) causes as did the young stroke (8-49 years) population; other (46.5%) and unknown (19.1%). Final aetiological classification differed significantly in young versus old in all categories (p=0.001) except cardioembolism (p=0.884). Admission neurological deficit (CNS) score compared to admission disability score (Rankin) showed moderate correlation with a Kappa value of 0.543. 2. Cognitive issues: One or more higher cortical function abnormalities was detected in 60.7% of non drowsy (drowsy, coma or delirious n=45) patients. The most numerous categories were aphasias (25.2%), apraxias (14.5%), amnesias (11.6%) and frontal systems syndromes (9.2%). In 76 patients, neuropsychological testing, (used as the gold standard) was performed and comparison to the HCFD test revealed a sensitivity of 80.2% (CI: 72-88%) and specificity of 100%. Cognitive impairment occurred without elementary neurological deficits (motor, sensory or visual i mpairment) in 137/608 (22.5%). Univariate and multivariate analyses of risk factors and likelihood of developing a HCFD revealed that increasing age, black race, being overweight and recent infection were independent variables at a p value of 0.05. HCFD did not differ significantly in younger versus older patients (p=0.194). Frontal system syndromes were more common in subcortical (32.3%) versus cortical (23.5%) lesions and more common in younger versus older patients (p=0.001) Conclusions: I. Cognitive disturbance is present in the majority of all types of stroke. This necessitates a reliable appraisal of this form of neurological deficit in all stroke patients in order to measure the true extent of deficit and monitor treatment and rehabilitation. This has important consequences for acute treatment trials that depend on changes in quantifiable deficit. 2. At times cognitive disturbance may be the sole presentation of stroke, unaccompanied by long tract signs. Therefore inadequate HCFD assessment may miss the deficit altogether. 3. Subcortical stroke is commonly associated with cognitive impairment - usually of a frontal system impairment. Such deficits are best correlated with functional brain scanning and not anatomical brain scanning. This is consistent with the network theory of brain functioning. 4. Risk factors for developing cognitive impairment in the indigenous stroke population included increasing age, black race, overweight body habitus and recent infection. This is an important message for the local population as the latter two are amenable to preventative measures. 5. In the young stroke population, although causes of stroke were numerous, prothrombotic states, infection associated strokes and dissection were the most numerous. All are amenable to primary preventative measures and treatable in the acute phase of stroke. 6. The Durban Stroke Data Bank showed that at least two dozen symptoms in stroke are important. In some instances, the diagnosis of stroke may be missed altogether if a wide array of symptoms are not entertained on presentation. 7. There were important black white differences in stroke with black people being younger with an increasing rate of HIV associated stroke being, documented. 8. Clinical and aetiological post investigative classification is useful in the management of stroke patients with significant differences found in all subgroups. This guides early, emergent stroke investigations and management. / Thesis (M.D.)-University of Natal, Durban, 1998.