Skip to main content
Fig. 1 | BMC Medical Informatics and Decision Making

Fig. 1

From: Surrogate endpoint evaluation using data from one large global randomized controlled trial

Fig. 1

Correlation between three-component MACE (surrogate endpoint) and CV death (true endpoint)—primary analysis using country as grouping variable. a Analysis of the LEADER dataset grouped while ensuring ≥ 30 CV deaths in each resulting group; nine groups were derived. b Secondary analysis of the dataset grouped while ensuring ≥ 20 CV deaths in each resulting group; 13 groups were derived. c Secondary analysis of the dataset grouped while ensuring ≥ 40 CV deaths in each resulting group; seven groups were derived. Circles represent a by-country group and the size of the circle is proportional to the number of trial participants in the group. Dashed lines represent the 95% prediction interval. The point of intersection of the upper limits of the 95% prediction intervals and a HR of 1 for the true endpoint on the y-axis identifies the surrogate threshold effect on the x-axis (dotted lines), the estimated value (HR) of which is shown in light grey. The coefficient of determination (R2) and the associated 95% CI were derived from a weighted linear regression model of the treatment effect (hazard ratio between liraglutide and placebo) for the surrogate endpoint (MACE) vs that for the true endpoint (CV death). CI confidence interval, CV cardiovascular, HR hazard ratio, MACE major adverse cardiovascular event, STE surrogate threshold effect

Back to article page