Recent Mipomersen Publication Questions Company Credibility

As you may remember, I had become quite bullish about ISIS the stock for the potential of its lead drug candidate, the ApoB-antisense inhibitor mipomersen for the treatment of hypercholesterolemia, to surprise to the upside. Based on the combined phase I and II data, the recent phase III data in the homozygous and heterozygous familial hypercholesterolemia populations appeared to have been on the lower end (~25-30% LDL-c reductions) of what I thought might be the true efficacy of mipomersen and maybe influenced by the genetic background of those patients. More importantly, the rate with which patients from these earlier studies were supposed to have entered the open-label extension studies and the apparent absence of increased safety liabilities with consequent prolonged dosing made the extent of the analyst scrutiny of the liver enzyme elevations appear overly alarmist.

Well, so I thought and was waiting for the release of the data from the phase III trial in the patients with severe hypercholesterolemia, the patient population with arguably the largest commercial potential for mipomersen, when, without much, or should I say any fanfare, a paper (Akdim et al, 2010) on a phase II study for which top-line data had been announced in a press release in November 2007 came out. I have followed biotech long enough now to know that companies like to put their topline data in the best possible light which is usually followed by the subsequent revelation of a few warts here and there. In light of the published data, however, I found in this case the take-home message of the press release to have been sufficiently misleading to warrant this blog entry and take back some of my optimism with regard to mipomersen…both efficacy-wise and, more significantly, with respect to safety.

The November 2007 press release which can be found here reported on the efficacy of ApoB and LDL-cholesterol lowering following 13 weeks of treatment at the important 200mg/week dosage which forms the basis for the current registration trials to be 42% and 48% reductions, respectively. What struck me a little bit as odd at the time, but interpreted as rather a sign of strength of mipo’s safety, was that no trial-specific safety data were offered. Instead, the safety data in that press release was an ‘integrated safety analysis’ comprising all the mipomersen studies until that date which were largely short-term studies.

This analysis claimed that, while on treatment, only 3% of the more than 250 patients exposed to mipomersen had experienced ALT elevations of 3-5x upper-limit of normal (ULN), the same percentage as in the combined placebo group, and none experienced more than 5x ULN. Taking into account also the follow-up periods (‘entire study period), the numbers would increase to 5% (placebo) and 7% (mipomersen), a slight, but not worrisome increase it would seem. Only by studying the tables, one would have noticed that there was one case of >5x ULN that apparently occurred during the follow-up of dosing part of the study (‘entire study period’). Injection site reactions that, as we now know, poses a number of development challenges for ISIS/Genzyme (soon daily injections?), were mentioned as a side-effect but characterized as ‘cosmetic inconveniences’ of no medical concern.

The fact that there were 2 more patients in the Akdim et al. paper than in the press release for the 200mg cohort could be due to many things (including drop-outs), but it is quite unfortunate that with the enlarged patient group LDL-cholesterol lowering suddenly would drop from the 48% in the PR to 36% in the paper. Well, 36% should be a degree of knockdown that the company should be able to live with, but sets off alarm bells that maybe some of the other data could have been presented in an overly positive light as well.

And indeed this is my conclusion after seeing the safety data: 5 out of the 10 patients (granted a small population, but this is 50%, nothing like the 3-7% mentioned in the press release!) in the 13-week 200mg cohort had an ALT increase of >3X ULN, 4 of which on 2 consecutive occasions! One patient even experienced an increase of >5x ULN and subsequently dropped out before all planned doses had been administered (which was then included in the ‘entire study period’ group, but not on-treatment group). Also not mentioned in the original press release was the fact that the ‘cosmetic’ injection site reactions applied to almost all patients receiving mipomersen (much, much less in placebo obviously) and that these accounted for 2 drop-outs in the 10-patient 200mg cohort.

Everybody, of course, has to form his/her own opinion about how much to rely on press releases reporting top-line clinical trial data. The way these phase II data were handled, however, rocks my confidence in how much to trust management’s contention that the wide-spread mipomersen concerns are ill-founded. The company does not have to be surprised that its credibility in the investment world is what it is and will be difficult to repair without structural changes. This is unfortunate, because I believe ISIS has otherwise made an admirable contribution to the development of oligonucleotide-based therapeutics.

PS: The purpose of this blog entry is not beat up on ISIS for the sake of it. Unfortunately, attempts to resolve my concerns first with the company directly failed as an email to ISIS investors relations (info@isisph.com) two days ago remained unreplied to.