RE:RE:RE:RE:RE:Unfortunate timing - Nicholls article on RVX - 09/15 Welcome back princeofcut, cutthecheese, onlytruthhere.
5-point MACE (death, MI, stroke, coronary revascularization, hospitalization for ACS or heart failure) was a pre-defined secondary endpoint of ASSURE, not 3-point MACE (death, MI, stroke). There was no pre-defined MACE endpoint in SUSTAIN. Combining ASSURE and SUSTAIN for pooled 5-point MACE post-hoc analysis showed the 55% RRR (p=0.02) for total population and 77% RRR (p=0.01) for diabetic population. These numbers are the ones that follow are from the ESC Cardio 2014 slide show. The 55% RRR in the total population is based on 499 patients; the 77% RRR in the diabetic population is based upon 192 patients. Despite the low numbers of patients AND only 6-month treatments period, they still recorded 35 total 5-point MACE events in the total population and 14 total 5-point MACE events in the diabetic population......AND achieved the above p-values for the RRRs. You can't argue with that. They reduced 5-point MACE....period.
You may doubt the validity/importance of 5-point MACE, but many very important, well respected, approved recent Phase 3 trials have used or continue a 5-point MACE. For example, the recently ended CETP inhibitor trial ACCELERATE that used evacetrapib used the same 5-point MACE scale. The still ongoing anacetrapib REVEAL Phase 3 trial for CETP inhibition originally had 3-point MACE comprised of death, MI and coronary revascularization; however, just recently the steering committee for REVEAL recommended replacing the coronary revascularization with stroke. The recently concluded Jardiance EMPA-REG OUTCOME Phase 3 trial in diabetic patients used a 3 point MACE (death, MI, stroke). The EXAMINE alogliptin trial in diabetic patients, which Resverlogix is using to project their BETonMACE event rates, use 3-point MACE (death, MI, stroke). So the point is that some Phase 3 event trials in recent past have used 3-point but some have used 5 point MACE. And for some, they have been modified along the way per the guidance of the steering committee. BETonMACE will be using the most strict severe 3-point MACE (death, MI, stroke).
To get an idea of what the ASSURE/SUSTAIN post-hoc MACE analysis would look like in terms of 3-point MACE (death, MI, stroke) or 5-point MACE (death, MI, stroke, cornonary revascularization, hospitalization for unstable angina or heart failure) would look like, let me releay the numbers from the ESC cardio presentation. I have the presentation right in front of me, so no question about my memory. Right off the bat, let me note that there was no reporting of stroke. I am interpreting this to mean that there was no incidence of stroke in ASSURE/SUSTAIN patients in the 6 month study period.
For the 168 total patients treated with placebo, there were 17 5-point MACE events (10.1% event rate). Of these 17, 14 were of the coronary revascularization/hospitalization for unstable angina or heart failure category. So if we drop those, there were only 3 events fitting the strict 3-point MACE; 3 deaths (1.8%), 0 MI, and assuming 0 stroke. For the 331 total patients treated with RVX-208, there were 18 5-point MACE events (5.7% event rate). Of these, 13 were of the coronary revascularization/ hospitalization for unstable angina or heart failure category. So if we drop those, there were 5 events fitting the strict 3-point MACE: 1 death (0.3%), 4 MI (1.2%), and assuming 0 stroke. The 3 events (all deaths) in the total placebo group of 168 patients equates to a 1.786% event rate on a 3-point MACE scale. The 5 events (1 death, 4 MI) iin the total RVX-208 group of 331 patients equates to a 1.511% event rate on a 3-point MACE scale. This is a 0.8459 hazard ratio for a RRR of ~15.4%. The 17 vs. 18 total 5-point MACE events led to a p=0.02 p-value. It is not surprising that the 3 vs. 5 total 3-point MACE events would have a non-significant p-value because it is underpowered.
Let's repeat this exercise for the diabetics. For the 65 diabetic patients treated with placebo, there were 9 5-point MACE events (13.8% event rate). Of these 9, 7 were of the coronary revascularization/ hospitalization for unstable angina or heart failure category. So if we drop those, there were only 2 events fitting the strict 3-point MACE; 2 deaths (3.1%), 0 MI, and assuming 0 stroke. For the 127 diabetic patients treated with RVX-208, there were 5 5-point MACE events (3.9% event rate). Of these, 4 were of the coronary revascularization/ hospitalization for unstable angina or heart failure category. So if we drop those, there was only 1 event fitting the strict 3-point MACE: 0 death, 1 MI (0.8%), and assuming 0 stroke. The 2 events (2 deaths) in the diabetic placebo group of 168 patients equates to a 3.077% event rate. The 1 event (1 MI) iin the diabetic RVX-208 group of 331 patients equates to a 0.7874% event rate. This is a 0.2559 hazard ratio for a RRR of ~74.4%. The 9 vs. 5 total 5-point MACE events led to a p=0.01 p-value. It is not surprising that the 2 vs. 1 total 3-point MACE events would have a non-significant p-value because it is underpowered.
How to interpret this?
#1: Time: It's only 6 months. Things can change for better or for worse in a longer study such as BETonMACE in which patients will be dosed an average of 18 months.
#2: Patient population: BETonMACE will only use low-HDL, diabetic patients; whereas SUSTAIN and ASSURE didn't require patients to be diabetic.
#3: Patient numbers: BETonMACE will use at least 2400 patients, much more than the ~500 patients that met the criteria for the ASSURE/SUSTAIN post-hoc analysis.
#4: Patient ACS history: This may be the most important one. BETonMACE will require patients to have had an acute coronary syndrome event within 7-90 days of randomization in the trial. ASSURE/SUSTAIN did not have this requirement. How significant is this? Re-watch the 9/25 corporate update presentation during Kausik Ray's and Michael Sweeney's part and they discuss the importance of this. Also, at around the 1 hour 30 min mark of the Q&A Kausik addresses a question related to this. Both the EXAMINE and EMPA-REG OUTCOME trials were based on diabetic patients using 3-point MACE. Only EXAMINE required an ACS event w/i 90 days of randomization. Examine had a 3-point MACE event rate of about 10.5% at 18 months (about the same for treatment and placebo groups). The event rate of EMPA-REG OUTCOME was more modest at around 6% at 18 months (as best asI can extrapolate from the graph). So long story short, diabetic pateints in BETonMACE trial will be at a much higher risk for a 3-point MACE event than those diabetic patients in ASSURE/SUSTAIN because of this very recent ACS inclusion event.
Onlytruthhere, no one is arguing that the ASSURE/SUSTAIN post-hoc analysis is underpowered for 3-point MACE. Where do you want to go from here? Are you going to play nice and contribute to a worthwhile discussion, or are you going to go back to your old Princeofcut/Cutthecheese tricks? I hope the former.
Best regards,
BearDownAZ