The following analyses will be performed on:

		the R1 phase to test for the treatment difference between n-3 PUFA and placebo on the two primary endpoints.
		the R2 phase to test for the treatment difference between rosuvastatin and placebo on the two primary endpoints

The main analysis will be performed according to an intention-to-treat approach, therefore all patients randomised in the study will be included in the analysis.
Baseline characteristics will be presented by treatment groups. Any unbalance for baseline characteristics thought to be of prognostic importance will be considered for multivariate adjustment in the subsequent analyses.
Plots of the Kaplan-Meier estimates of the survival curves will be presented.
A Cox proportional hazards model with terms for fish-oil, rosuvastatin and their interaction will be fitted to the data. If the term for interaction is non significant, it will be removed from the model. Results will be presented in terms of hazards ratios.