I'm just re-posting this thread for reference, since the futility analysis timing is front and center now. Here's the original message that I posted back in March. My forecast back then was predicting the 125 event futility analysis being triggered in September 2017, and it hasn't changed as new enrollment numbers have been announced.

Remember, the BETonMACE patient population is almost exactly the same as in the EXAMINE trial with the exception of BETonMACE having an additional low-HDL requirement. We know the EXAMINE trial event rate curves for 3-point MACE (see links in previous message). We know the approximate enrollment figures for BETonMACE announced by the company. We don't know how the low-HDL requirment will affect the event rate. Personally, I think that the low-HDL requirement will increase the event rate by ~15% in the overall BETonMACE trial population relative to EXAMINE trial. We also don't know how apabetalone will affect the event rate. If we assume a 30% reduction in event rate in the apabetalone arm, the net reduction in event rate in the entire BETonMACE trial population is 15% since half of patients will be on placebo. So the 15% increase due to low-HDL and the 15% decrease (net effect in whole BETonMACE pateint population) due to apabetalone will cancel one another out and the BETonMACE event for the entire population should be similar to the EXAMINE trial. I still have a good feeling for September.

Here's the text of the original message:

Who wants to play a little game of "When will the 125 event futility test be?"

Aside from the unknown about the rate of enrollment, there are 3 primary factors to consider, in my opinion.

1) EXAMINE Trial Event 3-point MACE Rate Estimates: Estimates of BETonMACE event rates are estimated from the EXAMINE trial in patients with diabetes and an ACS event within the past 90 days. BETonMACE patients will also be diabetic and recent ACS, but also will have low-HDL (see #2 below). The study drug, Alogliptin, in EXAMINE didn't work to increase or decrease 3-point MACE events, so both placebo and Alogliptin groups behaved more or less the same. The percentage of patients experiencing their first 3-point MACE event was ~11% over the median follow up of 18 months based upon graphs and statements from sources below. Additionally, eye-balling the primary outcome graph, I estimate a percent of: ~13% at 24 months, ~11% at 18 months, ~8% at 12 months, ~5% at 6 months, and ~2.5% at 3 months. So the event rate (first 3-point MACE event/month of follow up) is highest during the first 6 months of the follow up, and then the rate drops a bit in each of the subsequent 6 month periods.

http://www.nejm.org/doi/full/10.1056/NEJMoa1305889#t=article

https://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/EndocrinologicandMetabolicDrugsAdvisoryCommittee/UCM444148.pdf

http://secardiologia.es/images/rec/blog/EXAMINE.pdf

2) Effect of low-HDL on 3-point MACE Event Rate: EXAMINE trial did not require patients to be low-HDL. EXAMINE participants had an average HDL-cholesterol (HDL-C) level of 43 mg/dL at baseline. BETonMACE has a low-HDL-C inclusion requirement of <40 mg/dL (1.04 mmol/L) for males and <45 mg/dL (1.17 mmol/L) for females at Visit 1. The upper limit of allowable baseline HDL-C level for BETonMACE is lower than the average baseline HDL-C level for EXAMINE. Therefore, the average baseline HDL-C level for BETonMACE will be lower than that of EXAMINE. How does this affect the event rate? The best resource I could find was a recent re-visit to the Framingham Offspring Study by J. Bartlett et al. (https://www.ncbi.nlm.nih.gov/pubmed/27166203). They looked at low-HDL-C (<40 mg/dL in men and <50 mg/dL in women) compared to those with HDL-C above those levels and found:

"Compared with isolated low HDL-C, [cardiovascular disease (CVD)] risks were higher when low HDL-C was accompanied by LDL-C ≥100 mg/dL and [triglyceride (TG)] <100 mg/dL (odds ratio 1.3 [1.0, 1.6]), TG ≥100 mg/dL and LDL-C <100 mg/dL (odds ratio 1.3 [1.1, 1.5]), or TG and LDL-C ≥100 mg/dL (odds ratio 1.6, [1.2, 2.2]), after adjustment for covariates...... Compared with isolated low HDL-C, the risk of CVD was 30% to 60% higher when low HDL-C was accompanied by higher levels of TG, low-density lipoprotein cholesterol, or both.......In the Framingham Offspring Study, low and high HDL-C phenotypes are not uniformly predictive of CVD risk. TG and LDL-C represent important modifiers of incident CVD risk at both ends of the HDL-C spectrum."

Now, patients in BETonMACE will be on statins, which "should" keep LDL under 100 mg/dl. As for TG, these are diabetics, so I wouldn't be surprised if even with some diabetes medications if their TG levels are above 100 mg/dl. Also, if we are comparing EXAMINE to BETonMACE for HDL-C levels, we will be comparing pretty low-HDL-C (avg. baseline 43 mg/dl already) to very-low HDL (average ?). So I doubt that the very-low HDL-C levels BETonMACE compared to the already low HDL-C in EXAMINE will give as much of an increase in CVD risk as the Framingham Offspring Study predicted for their low- vs. normal-HDL-C cut-offs.

In my opinion, I think the added criteria of very-low HDL-C in BETonMACE may increase CVD risk by 15%. However, this is a big unknown. It might have a large, modest or no effect on the event rates modeled by the EXAMIINE trial. For example, if the EXAMINE trial modeled an 11% event rate at 18 months, the very-low HDL-C population in BETonMACE would be (11%) X (1.15) = 12.65%.

3) Relative Risk Reduction of RVX-208/Apabetalone. Management indicated that the study is powered to detect a 30% relative risk reduction. The effect of RVX-208 could be more (boo-yah), or could be less (ho-hum). For the purpose of our exercise here, let's stick with 30% relative risk reduction. Half of the patients in BETonMACE will be treated with RVX-208 and the other half with placebo. If the RVX-208 treated group has a 30% relative risk reduction compared to placebo, then the event rate of the RVX-208 group will be 30% lower than the event rate in the placebo group. However, when thinking about the event rate of the entire BETonMACE population (placebo and RVX-208 groups) as a whole relative to the event rates modeled in EXAMINE, the apparent event rate would only be 15% lower. In other words, placebo might behave like EXAMINE model, RVX-208 might show a 30% relative risk reduction, but the average effect in the entire population is 15%.

Following our same example in #2 above, if the EXAMINE trial modeled an 11% event rate at 18 months, the RVX-208 relative risk reduction of 30% in BETonMACE would be (11%) X (0.85) = 9.35%. Or thinking of it another way, the placebo group would have an event rate of 11% and the RVX-208 group would have an event rate of (11%) X (0.7) = 7.7%. But since only half are RVX-208-treated, the average of the two event rates (11% and 7.7%) is 9.35%.

My own rough estimates of enrollment with the EXAMINE event rates without a low-HDL or RVX-208 effect have the 125 events being hit in September 2017. Just to provide more wiggle room, let's expand that range to August 2017 to October 2017. We could also get different combinations of #2 and #3, which could advance or delay hitting the 125 events. Of course, this is just one person's guessing game, but here are my scenarios for hitting 125 events. Caveat is that there is A LOT of room for error here and was just one set of limited scenarios.

• No effect of low HDL, no effect of RVX-208: August-Oct 2017
• No effect of low HDL, 30% relative risk reduction of RVX-208: Oct-Dec 2017
• Low-HDL effect, no effect of RVX-208: June-August 2017
• Low-HDL effect, 30% relative risk reduction of RVX-208: August-Oct 2017

As you can appreciate, certain scenarios speed up the process or slow down the process, and in the case of the low-HDL effect and 30% relative risk reduction both occuring, these may cancel one another out and have the same predicted target date as no HDL effect and no RVX-208 effect.

Enjoy!

BearDownAZ