Efficacy and safety of icosapent ethyl in hypertriglyceridaemia: a recap

Figures

• 

  Figure 1. Relationship of progressive increase in non-fasting plasma triglycerides and remnant cholesterol with progressive increase in risk of morbidity and mortality. Observational relationships are shown; however, using a genetic instrument in Mendelian randomization analysis, such associations are confirmed to represent causal relationships. For acute pancreatitis, only the relationship with plasma triglycerides is depicted, as remnant cholesterol is not thought to play a causal role in this disease. Figure produced by Professor B. G. Nordestgaard based on data from roughly 100 000 individuals participating in the Copenhagen General Population Study and the Copenhagen City Heart Study as published previously.12–17

• 

  Figure 2. From elevated triglyceride-rich lipoproteins to atherosclerotic plaque initiation, plaque progression, and finally plaque rupture causing myocardial infarction or ischaemic stroke. This mechanistic cartoon depicts the situation in the non-fasting or postprandial state with lipoprotein in plasma both of intestinal (chylomicrons and chylomicron remnants) and hepatic origin [very low-density lipoproteins, very low-density lipoprotein remnants (= intermediate-density lipoproteins), and low-density lipoproteins]. Low-density lipoproteins and remnants cross the endothelial layer of intima in a process dependent on blood pressure, lipoprotein size, and concentration, possibly involving transcytosis. Figure produced by Professor B. G. Nordestgaard. apoB, apolipoprotein B; LDL, low-density lipoproteins; VLDL, very low-density lipoproteins.

• 

  Figure 3. Dose–response effects of n-3 PUFA consumption on fasting plasma triglycerides in RCTs. Based on 55 placebo-controlled trials of n-3 PUFA consumption for two or more weeks as extracted from a prior systematic review276 as well as two additional RCTs of fish or n-3 PUFA consumption169,277,278 to provide additional dose–response information at doses of <1 g/day eicosapentaenoic acid + docosahexaenoic acid (DHA). Each point represents the change in plasma triglycerides from baseline for each individual study arm, as compared with control. The solid line represents the line of best fit calculated from linear regression. Overall, each 1 g/day increase of eicosapentaenoic acid + DHA reduced triglycerides by −5.9 mg/dL (95% confidence interval −2.5 to −9.3 mg/dL). This effect was significantly greater in trials of individuals with higher starting triglyceride levels (P interaction < 0.001). Among trials of individuals with mean baseline triglycerides below the median (<83 mg/dL), each 1 g/day eicosapentaenoic acid + DHA decreased triglycerides by −1.7 mg/dL (95% confidence interval −3.1 to −0.2 mg/dL). Among trials of individuals with mean baseline triglycerides above the median (>83…

• 

  Figure 4. Cumulative incidence of cardiovascular events. (A) The Kaplan–Meier event curves for the primary efficacy composite endpoint of cardiovascular death, non-fatal myocardial infarction, non-fatal stroke, coronary revascularization, or unstable angina in the icosapent ethyl group and the placebo group, in a time-to-event analysis. (B) The Kaplan–Meier event curves for the key secondary efficacy composite endpoint of cardiovascular death, non-fatal myocardial infarction, or non-fatal stroke in the two trial groups, in a time-to-event analysis. In each panel, the inset shows the same data on an expanded y-axis. The curves were visually truncated at 5.7 years because a limited number of events occurred beyond that time point; all patient data were included in the analyses. See Figure 1A, ref.43 with permission. CI, confidence interval.

• 

  Figure 5. Distribution of first and subsequent primary composite endpoint events in the reduced dataset for patients randomized 1:1 to icosapent ethyl versus placebo. Hazard rations and 95% confidence intervals for between treatment group comparisons were generated using Li–Lagakos modified Wei–Lin–Weissfeld method for the first, second, and third event categories. Rage ratio and 95% confidence interval for between group comparisons used a negative binomial model for additional events beyond first, second, and third occurrences, i.e. fourth event or more and overall treatment comparison. Analyses are based on reduced dataset accounting for statistical handling of multiple endpoints occurring in a single calendar day by counting as a single event. See Central Illustration, ref.64 with permission. CI, confidence interval; HR, hazard ration; RR, rage ratio.