A parametric repeated time-to-event modelling of recurrent ischemic stroke after index stroke among patients with diabetes mellitus

Objectives: Diabetes (DM) is associated with an increased risk of recurrent ischemic stroke (IS). A temporal description of an increased or decreased risk of recurrent IS among diabetic patients is not fully established. This study aimed to investigate the repeated time-to-event of recurrent IS after the index IS among patients with diabetes using a parametric approach.

Method: A total of 3493 DM patients with a history of index IS attack were extracted from the National Neurology Registry. Several parametric survival models were evaluated using nonlinear mixed-effect modelling software (NONMEM 7.5.0). The final model was determined according to the lowest objective function value, graphical evaluation, numerical diagnostics and clinical plausibility.

Results: One hundred ninety-four (5.55%) DM patients developed a recurrent IS at maximum follow up of 5.67 years. Gompertz model fitted the data best. Without influence of risk factors, the index IS attack was predicted to contribute the hazard of recurrent IS for 0.192 within the first three months after the index IS then reduced to 0.0037 three months after index IS. Additionally, the risk of recurrent IS was increased by half at two time points throughout the five-year follow up and accelerated with the presence of classical risk factors such as ischemic heart disease (HR, 2.51; 95% CI [1.82-3.47]), hyperlipidaemia (HR, 1.98; 95% CI [1.50-2.61]), and hypertension (HR, 1.90; 95% CI [1.24-2.91]).