EntreroMedics ETRM ReCharge Trial Analysis

 Uncategorized  No Responses »
Jan 272013
 

ETRM popped up on my options screener last weekend. Despite my previous post, I couldn’t contain my curiosity and decided to investigate. The high implied volatility is due to the upcoming release of phase III study results from the ReCharge trial. Here is a quick summary of my thoughts:

  • ETRM is proposing VBLOC therapy for the treatment of obesity. Electrodes attached to the vagus nerve produce a high frequency electrical signal that disrupts communication between the brain and the gastrointestinal system. This leads to early satiety (feeling full sooner) and weight loss. The exact mechanism of action is unknown.

  • The first phase III trial (EMPOWER) failed because the treatment was not statistically significantly different from the control; the control demonstrated unexpected efficacy. Interestingly, the control was not a true placebo of zero electrical signal; instead, the control group received a low frequency signal as a substitute for the treatment group’s high frequency signal. In the trial’s postmortem, the investigators hypothesized that the low frequency signal was responsible for the efficacy seen in the control group. Their hypothesis seems plausible for four reasons:

    1. The control group experienced weight loss (16% EWL) above historical norms for placebo-controlled weight loss trials (8% EWL, which I corroborated by calculating %EWL for the ARNA and VVUS placebo arms).
    2. The control group demonstrated a dose-dependent effect. While a dose-dependent placebo response is not unheard of, in my opinion it adds circumstantial evidence for low-frequency efficacy.
    3. The control group’s low frequency signal is similar to the signal used to stimulate the vagus nerve in the treatment of epilepsy. I am not an electrophysiologist, but I am guessing that whether you’re stimulating (epilepsy) or blocking (VBLOC) the vagus nerve, you’re altering the natural course and overriding the original electrical signal. If low-frequency signaling is effective in epilepsy, then it’s plausible it could have an effect on obesity as well.
    4. Topiramate is an anti-epileptic drug that is also a component of VVUS’ anti-obesity drug QNEXA. One of the proposed mechanisms for topiramate in epilepsy is the blockage of voltage-dependent sodium channels, which I presume hinders neural signaling. Although topiramate and VBLOC’s mechanisms of action are unknown, my hunch is that they’re both blocking the same signal in the nervous system, which causes the efficacy in weight loss.

Thus, ETRM’s rationale for conducting the ReCharge trial seems sound. In ReCharge, the control group will receive a true placebo of no electrical current. As with EMPOWER, trial results will be communicated as %EWL (excess weight loss). EWL = total weight loss / (total body weight – ideal body weight), where ideal body weight is the theoretical weight that would yield a BMI of 25. The primary endpoints for the ReCharge trial are:

  • Observe a 10% greater EWL from randomization with the Maestro System after 12 months of VBLOC Therapy compared to control by BMI method.
  • Observe clinically meaningful responder rates in the treatment arm of 20% and 25% EWL from implant at 12 months (not statistically based).

Regarding the first endpoint, the treatment group experienced 17% EWL in the EMPOWER trial. A subsequent refinement of the electrical algorithm led to 22.7% EWL. Thus, a 10% EWL difference compared to 8% EWL historical placebo standards should be expected in ReCharge, assuming the investigators have properly diagnosed and addressed the cause of efficacy in the EMPOWER control group. For the second end point, 22% of the treatment group in EMPOWER achieved 25% EWL, rising to 41% if subjects in the treatment arm used the device for > 9 hours. This endpoint technically cannot fail in ReCharge because it is not a statistical comparison.

How do these endpoints relate to the FDA guidance on weight loss products? The FDA sets minimum efficacy standards in terms of absolute percent weight loss, not excess weight loss. A treatment must meet either of the following conditions:

  • The difference in mean weight loss between the active-product and placebo-treated groups is at least 5 percent and the difference is statistically significant
  • The proportion of subjects who lose greater than or equal to 5 percent of baseline body weight in the active-product group is at least 35 percent, is approximately double the proportion in the placebo-treated group, and the difference between groups is statistically significant

We need to translate between EWL and absolute percent weight loss to compare EMPOWER/ReCharge results with the FDA guidance. All calculations below are based on an average 5’4″ female with a BMI of 35 entering the trial. A 5’4″ female with a BMI of 35 weighs 204 lbs; a BMI of 25 weighs 145 lbs. The percentages are approximately the same regardless of specific height and BMI because they are normalized by the BMI calculation.

  • Regarding the first FDA efficacy measure, in the EMPOWER trial, the treatment group experienced 17% EWL. A 17% EWL corresponds to a 4.9% absolute weight loss. An 8% EWL typically seen in other placebo trials corresponds to a 2.3% absolute weight loss. A difference of 5% over placebo historical norms is 7.3% absolute weight loss, which corresponds to a 25% EWL. With a normal placebo response, the EMPOWER trial would not have met the first FDA requirement for efficacy.

    Given the later improvement in the electrical algorithm to 22.7% EWL, it’s possible to achieve the first criteria of 25% EWL, but only if the subjects are compliant and use the device for more than 9 hours a day. In the EMPOWER trial, there was a dose dependent effect where 22% of the treatment group achieved 25% EWL, but this increased to 41% if subjects in the treatment arm used the device for > 9 hours. Subjects must use the device for most of the day for it to have an effect.

  • Regarding the second FDA efficacy measure, a 5% drop in absolute body weight equals 17.2% EWL. In the EMPOWER treatment group, 58% of subjects who used the device for > 9 hours a day experienced 15% EWL, while only 27% of subjects who used the device < 9 hour a day experienced 15% EWL. With the improvement in the electrical algorithm, it’s certainly possible to meet the FDA’s second efficacy criteria, but again, the subjects need to be compliant and actually use the device.

As a comparison, the VVUS clinical trial experience demonstrated 29% EWL in the treatment group and 6% EWL in the placebo group. The ARNA clinical trial experience demonstrated 20% EWL in the treatment group and 8.6% EWL in the placebo group.

Conclusions

The company has set the bar pretty low for the ReCharge trial. They will probably succeed and announce positive top-line results. The key assumptions are:

  • They have properly diagnosed the causes of efficacy in the control group of the EMPOWER trial and ReCharge placebo efficacy returns to historical norms.
  • They implemented the improved electrical algorithms.
  • Subjects are more compliant and used the device for longer periods of time than they did in EMPOWER.

However, meeting their self-imposed endpoints means nothing if they have not satisfied the FDA’s requirements. So while the company might announce positive results, here are the two key points to examine in detail:

  • If the control group shows 8% EWL, then the treatment group needs 25% EWL to satisfy the first requirement. If placebo EWL is higher, then adjust the treatment EWL accordingly for a 5% difference in absolute weight loss over placebo.
  • OR the trial needs to show >35% of treatment subjects with >17% EWL. The percentage of treatment subjects who achieved 17% EWL must be double the percentage who achieved >17% EWL on placebo and the difference must be statistically significant.

If either of these requirements are met, then hang on for the ride. Lap-band at one time was worth $1B to Allergan, although that data may be somewhat dated. This treatment should similarly be worth $1B since it’s minimally invasive. ETRM has a current market cap of $125M, so this could be an 8x return if everything works out properly over time. If the probability of the trial succeeding is >12.5%, then the expected value is in your favor.

 Posted by at 9:25 pm

The Contrast Effect Between AEGR and ISIS

 Adcom  No Responses »
Oct 112012
 

The Endocrinologic and Metabolic Drugs Advisory Committee (EMDAC) will meet on October 17 and 18 to discuss lomitapide capsules (submitted by AEGR) and mipomersen injection (submitted by ISIS), respectively. Both drugs are being evaluated for the potential treatment of homozygous familial hypercholesterolemia (hoFH). We believe the back-to-back panels will introduce a contrast effect between the two drugs, leading to a negative vote for AEGR and a positive vote for ISIS.

Homozygous familial hypercholesterolemia is a rare genetic disorder in which both copies of the LDL receptor gene are defective. People with untreated hoFH have severely elevated LDL-C levels, which leads to early onset coronary heart disease and an expected lifespan of 20 to 30 years.

Lomitapide and mipomersen both successfully lowered LDL-C levels in their clinical trial programs. LDL-C was used as a surrogate marker for improved cardiovascular risk – neither drug underwent a cardiovascular endpoints trial to see if cardiovascular morbidity and mortality were improved with use of the drug. Thus, the advisory committee will be asked if the improvement in cholesterol surrogate endpoints reduces cardiovascular risk enough to offset the potentially severe side effects.

When approving a drug based on surrogate endpoints, all surrogate endpoints should point in the same favorable direction. Although lomitapide and mipomersen both significantly lower LDL-C, lomitapide lowers HDL-C (the “good” cholesterol), while mipomersen raises HDL-C. Thus, the LDL-C and HDL-C cholesterol surrogates for cardiovascular risk contradict each other for lomitapide, while they agree with each other for mipomersen. In this regard, lomitapide looks worse than mipomersen. One caveat: lomitapide HDL levels returned to baseline levels after 78 weeks, see safety extension trial link in the table below.

Lower HDL-C levels associated with MTP inhibition have been documented in the medical literature and were not a fluke of the trial: MTP knock-out mice have 50% lower HDL compared to wildtype mice. Treatment with an MTP inhibitor such as lomitapide mimics abetalipoproteinemia, a genetic disorder in which the MTP gene is defective. HDL levels are low in abetalipoproteinemia and the HDL is qualitatively abnormal. Abetalipoproteinemia reduces life expectancy.

Treatment with mipomersen mimics hypobetalipoproteinemia, a genetic disorder in which the apo-B gene is defective. Unlike abetalipoproteinemia, hypobetalipoproteinemia is associated with elevated HDL-C levels. Hypobetalipoproteinemia increases life expectancy.

All approved cholesterol drugs listed in the ATP III guidance for the treatment of high cholesterol lower LDL-C (bad cholesterol) and raise HDL-C (good cholesterol). One drug not on that list, ezetimibe, ran into controversy when it did not improve cardiovascular outcomes in the ENHANCE trial. All of the cardiologists on the panel will have the ENHANCE trial in the back of their minds when they review lomitapide’s data because one lomitapide trial compared lomitapide administered with and without ezetimibe (see table below).

Lomitapide faces the advisory committee first, mipomersen second. Just as a gymnastics judge in the olympics saves his best score for last, it’s natural for judges to be conservative at the beginning and more liberal at the end once they’ve seen all of the competitors. Mipomersen gets a preview of the advisory committee and can tailor their presentation to address the committee’s issues.

The question is whether it is worth swapping the cardiovascular risk of hoFH with the potential symptoms and liver toxicity of synthetically induced abetalipoproteinemia/hypobetalipoproteinemia. Based on the contrast effect of HDL-C surrogate endpoints, we feel lomitapide will receive a negative vote and mipomersen will receive a positive vote.

Comparing Lomitapide and Mipomersen

Lomitapide

Mipomersen

Mechanism of Action

Microsomal triglyceride transfer protein (MTP) inhibitor – Small molecule prevents the transfer of triglyceride to nascent apoliprotein-B (apo-B), lowering production of VLDL and chylomicrons.

Apoliprotein-B100 inhibitor – Anti-sense oligonucleotide reduces apo-B100 protein production in the liver, lowering levels of apo-B-containing lipoproteins such as VLDL, LDL and lipoprotein a.


Clinical Trial Experience

Efficacy
  • 40% reduction in LDL-C from a baseline of 337 mg/dL in hoFH
  • 6-9% reduction in HDL-C in trial with and without ezetemibe
  • 10% reduction in HDL-C in phase 3 trial, although this decrease was not shown to be permanent in the safety extension study
  • 24.7% reduction in LDL-C from a baseline of 440 mg/dL in hoFH
  • 34-47.3% reduction in LDL-C in other studies
  • 15.1% increase in HDL-C in hoFH

Safety
  • Gastrointestinal symptoms
  • 4/23 (17%) patients with elevated aminotransferases between 5x and 11x upper limit of normal (ULN). 3x ULN is generally considered to be the warning threshold, so the 5x levels could be underreporting the issue.
  • Increase in hepatic fat from 1% at baseline to ~8% through week 78
  • Injection site reactions/Flu-like symptoms
  • 6-18% of patients with elevated transaminases > 3x ULN across all studies
  • Approximately 5% median increase in hepatic fat level from baseline
 Posted by at 3:47 pm

CRTX Lixivaptan FDA Advisory Committee Behavioral Analysis

 Adcom  No Responses »
Sep 112012
 

On September 13, 2012, the Cardiovascular and Renal Drugs Advisory Committee will meet to discuss new drug application (NDA) 203009, lixivaptan, submitted by Cardiokine Biopharma, LLC, for the proposed indication of the treatment of symptomatic hypervolemic and euvolemic hyponatremia associated with heart failure and syndrome of inappropriate antidiuretic hormone (SIADH), respectively. We conducted a behavioral analysis of the voters listed on the draft committee roster to provide insight into how this panel might vote based on their past voting records. For more information on how we analyze voting behavior, please read our adcom behavioral analysis.

Of the 8 voting members listed on the draft roster, 7 members have served on one or more committees in which the vote was non-unanimous. Non-unanimous votes provide the differential voting patterns that allow us to analyze voter behavior. A summary of the voting records for these 7 individuals is listed in the table below. Please click each name for a detailed voting history.

For an explanation of the table headers, please see the “individual results” section of the methods page.

Name Num Votes Proportion Yes Proportion No Proportion Abstain Num Informative Votes Num Inverse Yes Inverse Yes Num Inverse No Inverse No Liberal-Conservative Index Maverick-Conformist Index
Diane Aronson 18 0.2778 0.6667 0.0556 10 0 0 10 0.4594 0.3473 0.0561
Mori Krantz 13 0.6154 0.3077 0.0769 9 6 0.3168 3 0.7705 0.5486 -0.3397
Allan Coukell 8 0.75 0.25 0 6 5 0.2063 1 0.4 0.2286 -0.1227
Michael Lincoff 7 0.7143 0.2857 0 6 4 0.2622 2 0.0919 -0.0821 -0.0832
Scott Emerson 10 0.3 0.7 0 5 0 0 5 0.3574 0.2501 0.0868
Daniel Gillen 3 0.3333 0.6667 0 3 1 0.25 2 0.4358 0.2495 -0.1752
Vasilios Papademetriou 2 1 0 0 2 2 0.2727 0 0 -0.1729 -0.0656
Briefing Documents

Most of the issues we highlighted in our CRTX preview were discussed in the FDA briefing documents. The central vs. local serum measurements were not as important to the FDA reviewer as we predicted, perhaps because the tolvaptan NDA only included local serum measurements. The central vs. local serum discrepancy was larger for BALANCE than it was for LIBRA or HARMONY. The lack of statistical difference at all time points was also not an issue.

However, modest efficacy in the surrogate endpoint was a concern for the reviewer given the safety red flag of an imbalance of deaths in the lixivaptan arm of the BALANCE study.

Voting Clusters
  • Yes Cluster: Coukell (consumer), Papademetriou (cardiology), Lincoff (cardiology)
  • No Cluster: Emerson (biostats), Aronson (patient)
  • Unclear: Gillen (stats), Krantz (cardiology), Fried (nephrology)

Coukell, Papademetriou, and Lincoff all voted yes for CHTP back in Feb, which had questionable efficacy and safety. They will be reliably liberal. Additionally, Lincoff is the only current panel member to have served on the tolvaptan panel, where he voted yes.

Emerson and Aronson only vote yes when the committee is unanimously in favor of approval. That will not be the case with this panel, so they will vote no.

Gillen’s sample size is too small to effectively predict how he will vote. Krantz is a wildcard who abstained in the CHTP vote; he has tended to be more conservative but has been in the minority on both the yes and no side of the vote in the past. We do not have a record of Fried in out voting database.

The modest efficacy and safety signal in the BALANCE trial likely means the committee will unanimously vote no on the heart failure indication. The safety concerns are not specific to the FDA reviewer, as “The imbalance in the early death rate between the treatment groups prompted the data safety monitoring committee to issue a letter on June 9, 2010 urging the applicant to terminate the trial. The letter stated that ‘the board voted unanimously that due to safety concerns that the study should be terminated as soon as possible’ after the board reviewed data on 469 subjects with observations at 15 days and 463 subjects with observations at 30 days.” Therefore this safety opinion should be generalizable to the advisory committee as a whole.

Regarding the SIADH indication, Krantz will probably vote yes with his fellow medical colleagues and Gillen will probably vote no with the other statistician. Fried could be the deciding vote, leaving it either 5-3 if she votes with the other doctors in favor of approval, or a 4-4 tie if she votes against.

 Posted by at 2:28 pm
 Older Entries