NEW YORK (
) -- My column last week about the
generated an unexpectedly high volume of comments from professional investors. A few smart buysiders questioned two of my concerns: 1) Sarepta's control group might not accurately reflect natural progression in Duchenne muscular dystrophy (DMD), and 2) competitor
pivotal trial implied a higher-than-expected bar for accelerated approval.
Before we dig in, let me reiterate my investment conclusion: Sarepta seems like a solid, modestly sized long position for a risk-tolerant investor.
Sarepta bulls allege that I unfairly questioned the 78-meter absolute decline in the six-minute walk test (6MWT) reported after 36 weeks for control patients in the phase IIb study of eteplirsen. My view was based on a December 2012
Muscle and Nerve
(McDonald C, et al. Muscle Nerve 2010 42:966-974)
, which showed a 57-meter decline in 6MWT for a small cohort of DMD patients. Citing the same paper, the bulls suggested that I should have focused on the 115-meter decline observed for boys more than seven years old, since that more closely matched the demographics in the company's trial. Had I used this subgroup as a benchmark, Sarepta's control arm data at 36 weeks would seem reasonable.
My first thought was: "$%@#! How did I miss that?"
Step 1: Double check the median age of patients in the Sarepta Phase IIb trial. As far as I can tell, management hasn't publicly disclosed those data. However, a review of the study protocol shows that enrollment was limited to boys between 7 and 13 years old.
Tip of the hat to those investors who highlighted my oversight, and apologies to my readers.
Step 2: Review the literature in more detail. Feeling like a bit of a dope, I dug deeper through the medical literature for longitudinal data on the natural progression of 6MWT scores in DMD. After an exhaustive search, I unearthed another relevant citation that Wall Street appears to have missed.
In July 2011, investigators from 11 neuromuscular centers in Italy independently published a longitudinal study of 6MWT scores in DMD in
(Mazzone E, et al. Neurology 2011; 77:250-256.)
The investigators followed 106 patients -- a far larger group than the 18 patients in the McDonald study -- for a year. Like McDonald, the Italians noted a significant age-based difference in the 6MWT decline, with "the point of slope change...around 7 years."
The Italian physicians observed a 12-month 6MWT decline of only 42.3 meters among the 71 boys that were more than seven years old. That's far less than the decline observed in the McDonald study. Even if we exclude patients receiving continuous steroid treatment -- a high-risk therapeutic option that has modest, temporary efficacy in DMD -- the Italian cohort performed far better than the Americans, with 6MWT scores declining by 66.4 meters.
There are disadvantages to both citations. Investors should rightly worry about the quality of data from a single European country, although the large cohort size and relentless progression of DMD may lessen this concern. On the other hand, compared to the independently funded, multi-center Italian group, the American study -- paid for by PTC Therapeutics and including only patients from a single institution -- has an increased risk of bias.
My second review of the literature raises an important point. Despite the steady loss of function over time, DMD patients experience more clinical variability than I would have expected. In the Italian study, the standard deviation -- a measure of data dispersion around the mean -- exceeded the mean 6MWT decline (-42.3 meters ± 73.9 meters for the subgroup of patients more than seven years old). McDonald and colleagues also note the high degree of variability as patients approach loss of ambulation. Finally, the Phase I-IIa trial of Glaxo and Prosensa's GSK-2402968 (formerly PRO-051) show a standard deviation nearly as wide as the mean 6MWT result itself.
These data all suggest that progression in DMD can vary widely, which makes it especially difficult to tell whether Sarepta's results show a drug effect or just unexplained progression in an unblinded study cohort. I'm inclined to believe it's a drug effect, but I don't think it's as clear-cut as do the bulls. As a result, I'm standing by my recommendation that investors keep Sarepta long positions relatively modest.
I'm not sure how to think about the rebuttals to my concern that Glaxo's Phase III trial suggests global regulatory agencies might require a long-term, placebo-controlled study. On the one hand, I agree with my colleague Adam Feuerstein that the new Prescription Drug User Fee Act (PDUFA) regulations give the
FDA increased leeway to rapidly approve promising drugs
for devastating diseases. Sarepta's eteplirsen may clear this hurdle, although so might Glaxo's GSK-2402968. After all, although Glaxo doesn't have data correlating dystrophin levels with clinical outcomes, the patients enrolled were 9.2 years old on average -- beyond the point at which recovery would be expected -- and yet showed an absolute improvement in 6MWT results.
Investor pushback on this issue seems to rely largely on the interpretation of management comments from semi-public breakout sessions or other small group settings. Unfortunately, I wasn't present at those meetings.
In general, it's usually not a good sign when the smaller biotech company appears to be taking a more aggressive path to market than the large-cap pharmaceutical company. Having said that, the question of what might be acceptable to the FDA, especially in the face of what will likely be significant pressure from patient advocates, remains murky.
Some investors suggested that FDA concern about the observation of proteinuria -- protein in the urine -- among patients treated with GSK-2402968 forced Glaxo to conduct long-term randomized studies. That's certainly reasonable, and requires that we return to the literature for more information.
As a reminder, both Sarepta's eteplirsen and Glaxo's GSK-2402968 induce the body to "skip" a section of the exon (a sequence of nucleic acids) that codes for dystrophin, a protein critical to muscle repair that is dysfunctional in patients with DMD. Think of a row of puzzle pieces with one missing. Eteplirsen and GSK-2402968 facilitate creation of a semi-functional protein by hiding a mismatched puzzle piece (exon 51), which in turn enables matching puzzle pieces (sections of protein) on either side of the mutation to connect.
Although more-or-less functionally equivalent, the two drug candidates use different chemical maneuvering to outwit the body's immune system. Glaxo and Prosensa's compound belongs to an older class of molecule -- it's a 2'-O-methyl phosphorothioate antisense oligonucleotide, or 2'OMe AO, for you nerds out there -- which has the advantage of having been studied extensively in preclinical and clinical studies. The disadvantages are a relatively narrow therapeutic index and apparent uptake of filtered oligonucleotide by the proximal tubules in the kidney. This latter observation does not appear linked to kidney toxicity, although it's something to watch.
Sarepta uses newer, so-called morpholino chemistry for the design of eteplirsen. This chemistry affords eteplirsen a wide therapeutic window, good stability in vivo, and no obvious toxicity signal. The downside is the unknown, since morpholino chemistry has limited preclinical and clinical data. Those looking to dig deeper on the chemistry of these drugs should start by reading the excellent review article published in the July 2011
American Journal of Pathology
(Hoffman E, et al. Am J Pathol 2011, 179:12-22.)
Given the severity of DMD, I'm not convinced that global regulatory agencies would be sufficiently concerned by the chemistry for GSK-2402968 to mandate long-term studies for Glaxo and not Sarepta. Nonetheless, this is a very difficult issue to parse from the outside. Until I get direct access to management, I'm tentatively standing by my original view: Glaxo's choice of a lengthy development pathway is a worry, although perhaps only a modest one.
Regardless of what happens with Glaxo, Sarepta's 48-week follow-up data for eteplirsen are critical. If those results show an unambiguous correlation between expression of dystrophin and 6MWT, there is plenty of upside from current levels.
One final note: I want to thank everyone who provides constructive feedback to my columns, whether or not we agree. Despite my reputation as a skeptic, I'm always willing to change my mind if the data suggest I'm wrong.
Disclosure: Sadeghi has no positions in any of the stocks mentioned in this article.
Follow Nathan Sadeghi-Nejad on
Nathan Sadeghi-Nejad has 15 years experience as a professional health-care investor, most recently as a sector head for Highside Capital. He has worked on the sell side (with independent research boutiques Sturza's Medical Research and Avalon Research) and the buyside (at Kilkenny Capital prior to Highside). Sadeghi-Nejad is a graduate of Columbia University and lives in New York. You can follow him on Twitter @natesadeghi.