Please Note: Blog posts are not selected, edited or screened by Seeking Alpha editors.

Lorcaserin: A Summary of Valvulopathy Data

|Includes: Arena Pharmaceuticals, Inc. (ARNA)

 This morning Adam Feuerstein tweeted the follwing:

adamfeuerstein: Jefferies note argues $ARNA did not meet statistical test for proving lorcaserin heart valve safety. Hold rating, lowers PT from $6 to $5

Now I haven’t found the alleged note and from what I’ve heard there was no mention on flyonthewall or anywhere else. Let’s assume he got the data from some sort of reliable source and maybe he’s privy to information before the rest of us. While I’m not suggesting this may be the only reason for ARNA’s stock swing today, it is the only piece of news I see to accompany the 7ish% drop currently going on.

I wanted to share with you some of the data I found on the valvulopathy issue as I was writing myprevious piece on the cardiac data for lorcaserin. Rather than launch into my own statistical analysis, I’ll give you the data and opinions I’ve compiled from around the web on the issue and you can make of it what you will.

First, an excerpt from the BLOOM and BLOSSOM trial press releases. First BLOOM (3,182 patients):

Using an ITT-LOCF analysis, the assessment of echocardiograms performed at baseline and after patients completed 6, 12, 18 and 24 months of dosing indicated no apparent drug-related effect on the development of FDA-defined valvulopathy (moderate or greater mitral insufficiency and/or mild or greater aortic insufficiency).

Lorcaserin met the primary safety endpoint of no significant difference in rates of valvulopathy at 12 months. Rates of valvulopathy at 6, 12, 18 and 24 months for lorcaserin versus placebo were 2.1% vs. 1.9%, 2.7% vs. 2.3%, 2.9% vs. 3.1% and 2.6% vs. 2.7%. At 18 and 24 months, rates of valvulopathy for lorcaserin patients crossing over to placebo were 3.6% and 1.9%, respectively.

The FDA has requested that Arena rule out a 1.5-fold or greater risk of valvulopathy with 80% power. Assuming similar results in BLOSSOM (Behavioral modification and Lorcaserin Second Study for Obesity Management), the integrated data set from the two trials will be more than sufficiently large to meet this requirement.

And BLOSSOM (4,008 patients):

The assessment of echocardiograms performed at baseline and after patients completed 6 and 12 months of dosing indicated that lorcaserin did not increase echocardiographic heart valve regurgitation. Lorcaserin met the primary safety endpoint that evaluated the rates of new FDA-defined valvulopathy in BLOSSOM at Week 52: lorcaserin 10 mg twice daily (2.0%), 10 mg once daily (1.4%) and placebo (2.0%). The integrated BLOOM (Behavioral modification and Lorcaserin for Overweight and Obesity Management) and BLOSSOM echocardiography data set rules out a risk of valvulopathy in lorcaserin patients according to criteria requested by the FDA.

New data demonstrate that similar numbers of mitral insufficiency and aortic insufficiency shifts were reported for patients on lorcaserin and placebo. In patients with pre-existing FDA-defined valvulopathy at baseline, changes in valvular regurgitant scores did not differ between the placebo and lorcaserin groups. The majority of patients experienced either no change or an improvement in valvular regurgitation.

The rates of valvulopathy were actually lower in the lorcaserin groups in both trials at the longest time frames. I am, of course, not suggesting lorcaserin protects against valvulopathy; this is a statistical anomaly of two things that both don’t affect valvulopathy (one being the placebo, the other being lorcaserin).

Also, from the NEJM article on the BLOOM data (bold mine):

Activation of the 5-HT2b receptor in cardiac valvular interstitial cells is thought to cause serotonin-associated valvulopathy, which is characterized by the thickening of heart valves (particularly the mitral and aortic valves) and valvular insufficiency. Serotonergic agents such as ergotamine, methysergide, pergolide, and cabergoline,24-27 andespecially the weight-loss drug fenfluramine, increase the risk of valvulopathy. Each of these agents has significant agonist activity in vitro at the 5-HT2b receptor. In contrast, agents that activate the 5-HT2a or 5-HT2c receptor, but not the 5-HT2b receptor, are not associated with valvulopathy. Lorcaserin caused no significant increase, relative to placebo, in the incidence of FDA-defined valvulopathy, a finding that supports the hypothesis that valvulopathy is not associated with activation of the 5-HT2c receptor. When the aortic and mitral valves were studied separately, the rates of increased and decreased valvular insufficiency did not differ significantly between the study groups. Variability in echocardiographic interpretation for individual readers (as measured with the use of blinded reads of “standard” echocardiograms) in the current trial comparedfavorably with that measured in other trials involving echocardiographic end points with two readers.

The sample size and power calculations in our trial were based on the assumption that FDA defined valvulopathy would develop in 5% of patients in the placebo group within 1 year, an assumption that was in turn based largely on data from a previous 3-month, phase 2 study of lorcaserin. Because only 2.3% of patients in the placebo group actually had a finding of FDA-defined valvulopathy at week 52, the statistical power to rule out a relative risk with lorcaserin of 1.5 is 60%, which is below the desired power of 80%.
I’ve included the last paragraph to show where the confusion might lie. On its own, and because valvulopathy was assumed to occur in 5% of patients, the statistical power to rule out relative risk is below the desired power of 80%. If you re-crunch the numbers with a different assumption, you get a higher statistical power. Additionally, in the first excerpt Arena stated they would pool the results – which, of course, is the right thing to do.

The only way this fails the FDA desired statistical power is if both of the following occur: 1) the 5% assumed rate of valvulopathy stands and no recalculation takes place and 2) the FDA refuses to take the two trials together and evaluate the patient population as a whole.

If there is a stats expert in the audience that wants to chime in I think we’d all greatly appreciate it. I’m no stats expert, but the data seems to speak pretty strongly in favor of lorcaserin being completely safe in terms of cardiac valvulopathy.

Disclosure: No positions at time of writing. Long ARNA within the day of publishing (bought on downturn).