Stroke is the 4th leading cause of mortality in Hong Kong. It also carries considerable socioeconomical consequences due to disability.

Ischemic stroke can be classified by the TOAST classification, which includes large artery atherosclerosis, cardioembolism, small-artery occlusion and other causes (1). Among which, intracranial atherosclerosis (ICAS) had been major cause of acute ischemic stroke (AIS) in the Asia Pacific. It was estimated as high as 24.1% of AIS or transient ischemic attacks (TIAs) were attributed to ICAS in China (2). Management of ICAS related strokes has been challenging owing to its high rate of recurrence despite medical therapy. Recent randomized clinical trial suggested that aggressive medical therapy may result in reduction in recurrence compared with historical cohorts (3).

Our group has previously observed a 2.5-fold increase in atrial fibrillation related stroke over a 15-year period (4). The inverstigator also observed a decline in ICAS related AIS as well as its recurrent stroke risk throughout the recent years. Possible mechanisms include better management of metabolic risk factors and aggressive secondary prevention. Other possible reasons are increased atrial fibrillation (AF), small vessel disease (SVD) or other stroke mechanisms.

This study is aim to find the evolution of different stroke subtypes in relation to the characteristics of our stroke population over a 15-year period. This may influence territorial prevention strategy.

Five time points, i.e. 2004, 2006, 2008, 2010, 2012, 2014, 2016, 2018 are selected for analysis. All consecutive stroke patients included in the Prince of Wales Hospital stroke registry will be recruited.

Patients will be stratified according to their stroke subtypes with reference to the TOAST classification (1), which includes large artery disease, cardioembolism, small vessel occlusion, other determined causes, undetermined cause or incomplete investigations.

Stroke etiology is determined by stroke neurologists with reference to the history, physical signs, imaging including carotid duplex, transcranial doppler, computer tomography (CT), CT angiogram, magnetic resonance imaging (MRI), MR angiogram, electrocardiography (ECG) and echocardiography. Interobserver variability for stroke mechanism determination will be evaluated.

Pre-defined demographic data including age, gender, smoking status, vascular risk factors including hyperlipidaemia, hypertension, diabetes, prior stroke or TIA, etc. will be retrieved from the stroke registry.

The investigator will also compare the use of antiplatelet agents, anticoagulants, lipid-lowering agents, specific antihypertensives in the Shatin territory across the 15-year period using the CDARS.

Inclusion Criteria:

• All consecutive Chinese stroke patients included in the Prince of Wales Hospital stroke registry at 2-year intervals, i.e. from 2004 to 2018

Exclusion Criteria:

1. Uncertain diagnosis of stroke
2. Non-Chinese patients

Patients in the stroke registry of the Prince of Wales Hospital in the recent 15 years.