Gissi > Gissi 1 > Introduzione
GISSI
Original GISSI 1 Logo
Newsletter Contattaci (e-mail e indirizzi)
Login per i medici del Gissi
GISSI Home Page Actual GISSI project - GISSI Heart Failure Other GISSI Projects Other Links
 

Protocol

Introduction

The therapy

The n-3 PUFA hypothesis

The statin hypothesis

Study design

Primary objectives

Other end-point measures of efficacy

Subgroup analysis

Mechanistic goals

Concomitant treatments

Sample size

Main analyses

References
  Sample size

Overall mortality and the combined end-point of all-cause mortality or hospitalizations for cardiovascular reasons are the two primary endpoints of the study.

Assumptions for overall mortality

a. the rate of death of patients with heart failure of comparable severity recruited in the treated groups of previous randomized trials is usually between 9 and 10 % per year. Thus, the total mortality rate of the study after 3 years of follow-up is assumed to be 25%;

b. the results of available clinical trials in heart failure patients suggest that about half of heart failure diagnoses are due to coronary heart disease, and about half of these patients are expected to have a possible indication to statin treatment. Conversely, only 10% of patients with heart failure not due to coronary artery disease are expected to require cholesterol-lowering treatment. Therefore, it is expected that about 70-75% of the heart failure patients enrolled in R1 (n-3 PUFA vs placebo) can also be randomized to R2 (rosuvastatin vs placebo);

c. the minimum clinically relevant beneficial effect of the tested treatments is set at a 15% reduction of all-cause mortality.


Assumptions for the combined endpoint of all-cause mortality and hospitalization for cardiovascular reasons

a. the expected cumulative rate of death or hospitalization for cardiovascular reasons during 3 years of follow-up is expected to be between 40% and 45%;

b. the minimum clinically relevant beneficial effect of the tested treatments on the cumulative endpoint is set at a 20% reduction of risk.


Number of patients to be recruited

Patients enrolled in each trial will be followed until the occurrence of a sufficient number of deaths, unless the trial is stopped early on the basis either of the interim analysis or of new scientific evidences.

To achieve an overall significance level of <0.05, an adjustment for two primary endpoints will be made by setting a a= 0.045 for all-cause mortality and a a=of 0.01 for the composite end point of all-cause mortality or hospitalization for cardiovascular reasons.

Since the trial is event driven, the number of expected deaths which are needed to allow a reliable evaluation of the efficacy of tested drugs is set for both R1 and R2 at 1252 for a study power of 90% at the significance level a=0.045. The number of patients to be recruited to test the efficacy of n-3 PUFA and rosuvastatin is set at 7000 and 5250, respectively, assuming a uniform accrual period of 18 months and a maximum follow-up period of 4.5 years (33). According to the aforementioned, the study follow-up will last at least 3 years, provided that:

the rate of death to be observed during the follow-up will allow the evaluation of the assumed differences between the treatment arms according to the event-driven study design;

if the required number of events is reached for R1 before R2, R2 will be continued until 1252 deaths have occurred in R2;

in case the level of significance at the interim analysis will be reached for one of the two phases (R1 or R2), the study will be carried on only for that phase in which the level of significance has not been reached.

Therefore, the follow-up period of each trial can be prolonged until the minimum expected number of events (i.e. with a power of 90%, 1252 fatal events in 7000 for R1 and 1252 fatal events in 5250 patients for R2) will occur.

The number of patients to be recruited in both R1 and R2 is more than adequate to test the efficacy of n-3 PUFA and rosuvastatin in terms of the cumulative endpoint. In particular, the 7000 and 5250 patients expected to be recruited in R1 and R2 will allow a study power of > 99% (with a type I error of 0.001) and > 95% (with a type I error of 0.01), respectively.