Ann: Mesoblast Phase 3 Chronic Heart Failure Results, page-492

As unimpressed as I am with HC at this point, this discussion from the US Yahoo board, deserves to make it across here:

Invest314154 hours ago

Wow... I still have to dive in , take a closer deeper read at the trial design and the results, but my impression so far is amazing. So I bet some of the investors or readers here are confused. What happened? Trial didn't meet primary endpoint?... but did improve survival?... what's going on here?...
Well ... here's an attempt for a very simple way to look at it:
The primary endpoint for the trial (and several other trials in CHF) was to measure in general if treated patients arrive LESS times to admissions due to non fatal heart failure events. Basically , in layman language , counting how many times they admitted to a hospital with some event related to their heart failure disease. There are many reasons for why such an admission can happen, some are severe, some are less. So the logic here is to simply count all of them together and assume that a "good" drug will reduce their total number.
Note, however, that this is not measuring mortality, but something else. Why? Why measure this surrogate outcome? ... well... sometimes trials don't go for the holy grail, but instead aim at a more modest result. At an outcome that is believed to be a proxy for the real measure (in this case mortality from heart problems is the real important endpoint, and the proxy endpoint is the number of HF related admissions). This is done when it is believed it will be hard to prove the holy grail result, but it may be easier to prove the proxy result. In CHF it was SO hard to prove an improvement in mortality in the last years (cross several trials), that the FDA felt OK with the proxy, thinking it's better than nothing and better than letting all CHF trials fail, while getting something at least.

So... back to our story here... what happened?...
It seems that rex-L treatment did not reduce the number of HF related admissions significantly.
However it did lower the next VERY IMPORTANT outcomes:
(1) It lowered the number of SEVERE events not ending with death: namely MI or stroke not ending with death by 60% !!!!! (p value 0.002 !!!!!).
(2) It lowered the number of any heart related death by 60% !!!!! (p value 0.037 , I'm guessing not lower than that due to sample size and overall number of deaths in the trial, but this is very significant).
(3) It very significantly slowed the disease and prevented patients moving from stage II to stage III (!!!!)

Those results, and esp, no (2) and no (1) are the HOLY GRAILs of treating CHF .... so the trial did not pass the proxy endpoint (will have to read a full report/paper to understand why and what happened), but it did pass the much more important and harder to pass mortality endpoint, and lowered the SEVERE events !!

Let me say it in another way .... the designers of this trial did not believe in rex-L as much as they could !!!, Had they known this could be what's happening here I can assure you they would have gone for the primary goal of mortality and reduction in MACE events. For sure. As this is the golden path to approval. This is what matters to physicians (and FDA) most above anything else: prevent deaths, and prevent major events. They went for a proxy outcome, as this is what people are doing in these CHF trials, when they know how hard it is to improve the real thing (mortality , MACE) and they try to improve something else.

This is by all means an amazing outcome for the trial. To me this is a surprise to the good side. I'm surprised this did not pass the primary endpoint (is it maybe because they prevented major events and deaths but that came along with a few more admissions for these patients???), but I'm even more surprised to see 60% !!!! reductions in MACE and deaths !!! 60% !!!!!! how long has it been since you've seen a drug reducing the number of MI or strokes by 60% ??? or reducing the number of HF deaths by 60% ???? I've never seen something like this with a drug for CHF. This is close to a miracle. They put it on secondary because they did not believe it could happen.

All this with caution - as we need to get more information and learn better what happened here, get a paper out , see all the tables, all the numbers, and understand. But this can be a new era for CHF patients. Nothing less than that.

GLTA

Left-e3 hours ago

I agree with the analysis. Surrogate endpoints are typically used to shorten the length of time for a study to complete. They "speed things up" and are common in cancer studies, for example. Rather than waiting for patients to die let's monitor an event/marker that can be detected sooner and suggests a higher mortality rate will ensue in that group. But death is the final arbiter followed by MACE and those low p-values in a study this size can't be denied. The results here may seem paradoxical at first, but I suspect hospitalizations for stroke and new heart attack were perhaps not included in the primary end-point category "recurrent non-fatal decompensated heart failure events". That marker apparently does not predict a higher mortality. We'll have to see the full study and it will be most interesting to see where these results are published. This could be NEJM material or a journal of comparable prestige. I suspect a full analysis of the two groups will indeed reveal marked differences in other parameters besides MACE and mortality. For example, in cardiac perfusion scans, exercise tolerance tests, biomarkers, etc. And perhaps in length of hospital stay or intensity of care, CCU vs a regular ward. As SI said in response to one question, "we're still analyzing the details of the data". I suspect a lot more will follow as the review process completes and they go to publication and the FDA. Market reaction seems knee-jerk; more thought and analysis are in order, we could be looking at a landmark medical event. Recall that p values under .01 might be considered evidence of "overwhelming" efficacy. p=.004 and p=.002 are well below that.

• mitchell22 minutes ago

  @Left-e Indeed, this case is atypical in that regard. I rather like the brief explanation shared on the call in relation to the treatment MOA. My take is this trial gave them more than they hoped for. A clear and significant survival benefit in a well defined patient population. I expect they continue to explore and learn about use across the treatment paradigm, but there is now a fixed, definitive, and rather large population subset to get going with. Very intriguing outcome. In drug development we are always learning and often surprised.

  Reply

  1
• 
  

  

  Left-e46 minutes ago

  @mitchell, agreed, there can be a disconnect between the surrogate endpoint and OS... but usually the criticism comes from the opposite extreme: the treatment leads to improvement in the surrogate marker, but if you follow patients long enough, lo and behold there's no survival difference between the two groups. That's been seen with a lot of medical interventions from cancer treatments to various screening tests. In fact, it's been seen with most previous CHF studies. Here we're seeing the opposite: no difference in the surrogate but a huge significant difference in MACE and mortality. With those p-values the surrogate MUST be re-thought. It's more of a symptom than a reliable sign for progression of the underlying illness. SI said this study should lead to a new paradigm in designing CHF studies. It should also lead to a new paradigm in treating CHF. NB: if the study had truly failed we would have seen no difference in the primary and no difference in the secondary endpoints. If the naysayers and the market in general "get it" now or not until later I couldn't care less. That's what makes investment opportunities.

  More

  Reply