Study design
This was a retrospective cohort study performed at two major GOPCs (Lek Yuen & Ma On Shan family medicine center, NTEC, HA), which served a population of 630,000 in Shatin District (8.9% of HK population).
Inclusion and exclusion criteria
Subjects were identified from the Clinical Management System (CMS) database of the HA. Inclusion criteria were (1) Age ≥ 18; (2) Non-acute ischemic stroke patients with coding by International Classification of Primary Care, 2nd edition (ICPC-2) K90 (stroke/cerebrovascular accident) or K91 (cerebrovascular disease); (3) Only those patients diagnosed with ischemic stroke and who had at least 2 consecutive follow-up visits in the same clinic within the recruitment period were included. Patients were excluded if 1) they had a previous history of acute coronary syndrome or revascularization procedure of coronary artery; 2) hemorrhagic stroke; 3) acute ischemic stroke within 4 weeks of onset; 4) being followed up by medical specialists only. The recruitment period lasted for 6 months from 1/1/2010 -30/6/2010. Ethical approval of this study was granted by The Joint Chinese University of Hong Kong – New Territories East Cluster Clinical Research Ethics Committee (Reference No. 2019.252).
Outcome measures
The primary outcome was the recurrence of fatal and non-fatal cerebrovascular events including TIA, ischemic stroke or hemorrhagic stroke. The secondary outcomes were all-cause mortality and CHD, respectively. If death or a recurrent cerebrovascular or cardiac event was recorded as occurring simultaneously, one event was included into the analysis. The incidence of cardiovascular events was identified by the International Classification of Diseases, Ninth Edition, Clinical Modification (ICD-9-CM) and ICPC-2 codes from the CMS. The earliest date of diagnosis with ICD-9-CM of 430.x to 438.x was the time of recurrent stroke event. The earliest date of diagnosis with ICPC-2 of K74 to K76 or ICD-9-CM of 410.x, 411.x to 414.x, 36.07, 36.10 to 36.14 was defined as the time of development of CHD (angina pectoris, myocardial infarction (MI), percutaneous coronary intervention or coronary artery bypass graft surgery). The recurrent cardiovascular events and death causes were further identified by reviewing of the medical discharge summary records. Indefinite information about the cause of death was labelled as unknown.
Baseline covariates
Baseline information of patients including socio-demographics, clinical parameters, and regular medications were retrieved from CMS. The socio-demographics of patients included gender, age and smoking status. The patient was considered a smoker if he/she currently smoked. Clinical parameters including BMI, BP, lipid profile and HbA1c if diabetic were extracted from CMS. Co-morbidity including obesity, HT, DM were retrieved by ICPC code from the CMS. Hypertension (HT) was defined as a history of systolic blood pressure (SBP) ≥ 140 mmHg or diastolic blood pressure (DBP) ≥ 90 mmHg (SBP≥ 130 mmHg or DBP ≥80 mmHg among diabetic patients), or a history of treatment with antihypertensive medication. The diagnosis of DM was established with any of the following criteria: 1) when a patient presented with classic symptoms of hyperglycemia and has a random plasma glucose value of ≥11.1 mmol/L; 2) In an asymptomatic individual with fasting blood glucose level ≥ 7.0 mmol/L, HbA1c ≥ 6.5 %, a 2 hour post-oral glucose tolerance test values of ≥11.1 mmol/L. In the absence of unequivocal symptomatic hyperglycemia, the diagnosis of DM could be established on a subsequent day by repeating the same test for confirmation. Obesity was defined as BMI ≥25.0 kg/m2. Treatment modalities including the use of antihypertensive agents, statins, antiplatelet and antidiabetic medications were retrieved from CMS. Statin users were defined as patients receiving statin at baseline. Non-statin user was defined as follows: 1) patient not using statin at baseline; 2) patient who stopped statin within 6 months after initiation of treatment at baseline.
Follow-up
The entry date was the first attendance record in GOPC clinics for stroke follow-up during the recruitment period. All patients were followed up regularly at an interval of 3 to 4 months till the date of incidence of the first recurrent stroke, cardiac event, death, the censored date of lost to follow-up or 31/12/2018. Cases were censored for recurrence if they died before experiencing a recurrent event.
Statistical Analysis
All data analyses were performed using the SPSS software (version 25.0 for windows). All tests were two-sided, and P values < 0.05 were regarded as statistically significant. Baseline comparisons were made with the Student’s t-test or the chi-squared test as appropriate. Multivariate Cox proportional hazard regression analysis were performed to estimate the risks of recurrent stroke, cardiac events and death, taking the time variable into consideration. Hazard ratio (HR) and its 95% confidence interval (CI) were reported for each variable within the regression model after adjusting for other confounding factors.
Sample size calculations
Our pilot study including 100 subjects found the recurrent stroke rate and mortality rate was 0.2 and 0.3, respectively. Given a confidence level of 95% and a relative precision of 0.2, the sample size needed for observing the expected proportion of recurrent stroke and mortality among our GOPC patients was 385 and 225, respectively. Below is the formula of sample size calculation:
where p=0.2 (stroke recurrent rate) or 0.3 (all-cause mortality), z=1.96, ε=20% or 0.2.