|Safety data explained for AVNR|
|By Patrick Crutcher|
|Monday, 18 October 2010 22:02|
With all the recent attention on Avanir Pharmaceuticals Inc.(NASDAQ:AVNR) and their upcoming PDUFA date of 10/30 for AVP-923(formerly Zenvia), we reached out to management in order to have them answer some questions.
These questions were about some of the safety data from their pivotal STAR trial which was conducted under a Special Protocol Assessment with the FDA. AVNR received a Complete Response Letter(CRL) in 2006 for their initial NDA with AVP-923.
The coming FDA decision has major implications for the future of AVNR, so it is important to talk about the safety data. We reached out to the VP of Communications, Eric Benevich, in order to have him answer questions which were on the minds of investors. The following is a transcript of that exclusive interview:
Biomedreports: How do you think the FDA will classify the deaths reported in the STAR trial? Do you anticipate them interpreting them in a similar way to your committee? I assume this was a conversation you had with the FDA at some point. I'm particularly interested in the death classified as "possible".
Benevich: There were 2 separate assessments done for deaths reported in the STAR trial. Consistent with Good Clinical Practices, investigators were asked to assess the study-medication-relationship for adverse events using the following explanations: Not Related, Unlikely, Possible, Probable, Highly Probable. There were 7 deaths reported in the STAR trial, all among patients with underlying amyotrophic lateral sclerosis (ALS): 3 in the AVP-923 (30/10) group, 3 in the AVP-923 (20/10) group, and 1 in the placebo group. 6 of these were reported by the investigators as not related, while the 7th was reported by the investigator as ‘possibly’ related to study medication.
Following standard procedures, detailed clinical narratives were created for each case. This information was reviewed by an independent mortality adjudication committee on a blinded basis. All deaths were classified by the committee as respiratory related and likely due to the progression of the underlying ALS. No acute decompensation of respiratory function after initiation of study drug was observed, and no deaths were ascribed to a cardiac cause.
Amyotrophic lateral sclerosis (ALS) is a relentlessly disabling neurologic disease that inevitably progresses to respiratory failure and death. When respiratory muscles weaken, a cascade of events leading to respiratory failure can be observed. Death usually occurs secondary to respiratory failure or respiratory complications and the average life expectancy of an ALS patient is about 2-5 years from time of diagnosis. All data related to the 7 reported deaths were provided to the FDA as part of the full response filing.
BioMedReports: Can you give an explanation for the difference in mortality rates seen between dosing and placebo groups?
Benevich: Since ALS is a terminal disease with an expected life span of 2-5 years from diagnosis, the company did expect to see deaths among the ALS patients. Based on the epidemiologic literature, the mortality rate among ALS patients over the course of a 3 month trial was expected to be between 4% and 6%. The actual mortality rates reported in the STAR trial were 4.6%, 4.4% and 1.6% in the 30/10, 20/10 and placebo groups, respectively.
A baseline imbalance between dosing groups in terms of time from diagnosis (TFD) of ALS may help explain differences seen in mortality rates. Mean time from ALS diagnosis for each cohort was as follows: AVP-923 22.7 months, AVP-923 (20/10) was 16.3 months, and placebo group was 13.4 months. Life expectancy for an ALS patient is expected to be between 2 and 5 years from time of diagnosis. Time from diagnosis is a significant predictor of ALS mortality.
BioMedReports: Any comment on the higher incidents of serious adverse event (SAEs) related to AVP-923 and higher discontinuation due to AE in 20/10 group relative to both 30/10 and placebo group?
Benevich: Serious adverse events were well balanced between the groups. A total of 7.3%, 8.4% and 9.2% of patients in the 30/10, 20/10 and placebo groups reported at least one serious adverse event (SAE). Two SAEs, both in the AVP-923 20/10 group, were reported as possibly treatment related. In 1 of these patients, the event was reported as respiratory depression and ALS progression. The other patient had worsening muscle spasticity. There were more treatment discontinuations in the low-dose 20/10 group possibly because the treatment benefits appeared to be greater in the 30/10 higher dose group. However, both doses were well tolerated and the overall discontinuation rates were low.
BioMedReports: Can you address how the current response to the CRL answers this question?
Benevich: In 2006, the FDA’s Approvable Letter expressed 2 primary concerns: cardiovascular safety and adverse events among vulnerable ALS patients. After meeting with the FDA we agreed to create lower quinidine dose formulations and conduct a single confirmatory trial to demonstrate efficacy and safety. The new lower dose formulations, along with dose titration for the first week of therapy, may account for the decreased reported rates of dizziness, nausea and somnolence in the STAR trial compared to previous Phase III trials utilizing the original 30/30 mg dose. In addition, there were no differences in reported falls between active treatment and placebo in the STAR trial. With a 67% lower dose of quinidine, we increased the margin of cardiac safety. In the STAR trial there were no clinically meaningful changes in QT interval, no pro-arrhythmic events or cardiovascular SAEs.
BioMedReports- Can you address any concerns with the QTc data?
Benevich: If you look at the guidelines for QT studies(ICH E14), clinical data showing mean prolongation of QT/QTc interval by more than around 5 and less than 20 ms are deemed “inconclusive” of risk of arrhythmia. The FDA looks at mean changes in QTc and any outliers of concern. What we saw in the STAR trial was mean changes from baseline to end of study between 1.0 and 4.8 ms. At all time points assessed, no AVP-923 recipient had a QTc-interval absolute value >480 milliseconds (with Fridericia correction) or a change from baseline >60 milliseconds.
BioMedReports.com: Do you have any comment on why (in the subset of MS patients) missed on some very important endpoints? Specifically the primary endpoint of the trial was missed with the MS data in the 20/10 dose missed and the secondary endpoint of the CNS-LS score improving was missed in both dosing groups?
Benevich: It’s important to recognize that the STAR trial was not powered to detect differences in efficacy within subgroups, which was also not the point of trial. The STAR trial was designed to measure differences in efficacy between each treatment cohort. The study design and statistical analyses were pre-specified in the Special Protocol Assessment (SPA) agreement with the FDA. Both doses of AVP-923 met the primary efficacy endpoint. The secondary endpoints were exploratory and not required for consideration of FDA approval.
Our initial coverage of AVNR: http://bit.ly/dzo62G
August 2009 SEC filing: http://bit.ly/aEEZ4v
STAR Trial - Annals of Neurology: http://goo.gl/LC0m
Disclosure: Long AVNR