...We Welcome You To The Resverlogix HUB withIn The AGORACOM COMMUNITY!

Free
Message: Re: Consequences of BETonMACE Findings
1
May 10, 2016 01:26PM

Koo wrote, "The primary endpoint is the time to first occurrence of Major Adverse Cardiac Events (MACE)."

Yes that is true. The primary endpoint of BETonMACE is time to first occurrence of narrowly defined MACE (CV death or non-fatal MI or stroke) and the secondary endpoint of BETonMACE is time to first occurence of broadly defined MACE (CV death, non-fatal MI, hospitalization for CVD events, or stroke). However, this information is used to calculate the hazard ratio, relative risk reduction and absolute risk reduction. So it is all the same data. Just a matter of how it is graphed out, presented or analyzed.

EMPA-REG OUTCOME (see slides 41-62) offers a good example of how there are many ways to play with the data to get a favorable result when you're dealing with slightly hitting or slightly missing endpoints.

In slide 43 you can see that for 3-point MACE, EMPA-REG OUTCOME had a 14% RRR with a p-value of 0.0382. That's a pretty weak RRR and a very weak p-value that barely met the cut-off of p<0.05. This combined both the 10 mg and 25 mg groups and on slide 44 you can see that if the 10 mg and 25 mg groups are analyzed separately that neither one has a p<0.05. And on slide 45, you can see that if certain patients from 10mg+25mg combined groups are excluded due to either 1)Excluding events >30 days after last intake of study drug and patients who received study drug for <30 days (cumulative) or 2) Patients treated with ≥1 dose of study drug who did not have important protocol violations, that this weakens the p=0.0382 value and results in a p>0.05. So simply based on reduction of 3-point MACE, EMPA-REG OUTCOME nearly failed.

However, in slides 47-49, you can see that nearly all of the positive effect of the drug on reducing 3-point MACE was due to reduced CV death (p<0.0001 for combined 10 mg+25 groups, p=0.0016 for 10 mg group and p=0.0001 for 25 mg group), whereas there was no significant improvement in non-fatal MI (p=0.2189) and a trend towards increased, but not decreased, stroke (p=0.1638). If they added in their 4-point MACE in slide 53, which also included hospitalization for unstable angina pectoris, this reduced the p-value to a non-significant p=0.0795.

They also looked at hospitalization for heart failure in slides 56-58 (one of their secondary outcomes), which showed a strong benefit (p=0.0017 in combined 10mg+25 mg group, p=0.0044 in 10 mg group, p=0.0166 in 25 mg group). All cause mortality (slides 59-62) was strongly reduced as well (p<0.0001 in combined 10mg+25 mg group, p=0.0013 in 10 mg group, p=0.0003 in 25 mg group). This all cause mortality data indicates that other deaths and not just CV death were also decreased.

In BETonMACE, they are of course going to have the total RVX-208 treated vs. the total placebo group for 3-point MACE. But they are also going to have atorvastatin vs. rosuvastatin sub-groups. They are also going to have normal kidney function vs. impaired kidney function subgroups. And there are a lot of other secondary outcomes built in to the trial. If BETonMACE has a p<0.05 for the total population for reducing 3-point MACE, then the trial is a success barring some strange split in the CV death/non-fatal MI/stroke data. However, if BETonMACE misses this in the total population, then it wouldn't nesessarily be an absolute failure. For example, if the data shows huge benefit with rosuvastatin but not with atorvastatin, then 1) this would explain why the combined statin group was not significant and 2) bolster the claim the RVX-208 synergizes with rosuvastatin. However, if we completely miss the 3-point MACE (no reduction in death, MI or stroke) for any of the comparisons (total population, statin separated, kidney disease separated) and only have significance with the 4- or 5-point MACE (ASSURE/SUSTAIN was 5-point, I think secondary outcome for BETonMACE is 4-point), then BETonMACE is likely deemed a failure. What we need is a STRONG p-value so that we hit this out of the park and are not working at the margins of significance.

BearDownAZ

Share
New Message
Please login to post a reply