Pharma’s Cutting Edge

Here’s the Times UK take on the appointment of Jeffrey Kindler to CEO of Pfizer effective in 2007 with Hank McKinnell’s retirement. I know nothing about Mr. Kindler except what I’ve read recently. I’m sure he’s very talented and highly qualified to lead a large company. Still, I lament the decision to once again appoint a non-scientist, non-clinician to head the world’s largest therapeutics company. I know there are some people who continue to believe that scientists and physicians are too wishy-washy, too indecisive to lead the largest companies in the industry effectively. But recent experience, from Genentech to Novartis, argues otherwise. Pharma and biopharma companies are best led by individuals who understand the scientific and medical bases of their products well enough to create and champion their own visions for technological innovation. This has always been true, not just in Pharma but in every industry where technological innovations drive growth. Perhaps Pfizer’s Board is signaling that the giant’s next phase of growth will come from focusing on business-model innovation or lifecycle management more so than technological innovation?

For those readers actively engaged with or otherwise interested in drug screening, this open-access article by Inglese et al from the NIH Chemical Genomics Center is welcome news and good reading. The qHTS approach by Inglese at al, in which a 5-fold dilutional titration of each sample compound is represented across seven experimental plates (each compound at the same relative plate position to avoid positional errors), resulting in a 7x library replication, addresses many of the problems that limit the utility of HTS, particularly HTS of combichem libraries, with their typically limited structural diversity. They show, convincingly, that this multi-concentration approach yields fewer false positives and, very importantly, fewer false negatives of active compounds, including active compounds with high potency and efficacy as compared with single or double concentration screening, at least in the context of a luminsecent enzyme-based assay with high signal to noise ratio. They are now testing their method with cell-based assays. I suspect that some drug companies and discovery-only houses have quietly been using a similar approach for some time, but I believe this is the first published description of the technique in practice. Expect companies who haven’t adopted the technique to begin testing it for themselves in the near future. It promises to result in more hits, with fewer false positives, and earlier SAR, which means better drug candidates sooner. What’s not to like?

Yesterday, Pfizer announced it would be rolling out Exubera in the U.S. beginning in September, following an educational campaign to assist with early uptake. In related news, Nektar, which receives revenues from manufacturing Exubera and royalties on sales, raised its its first-year revenue guidance from Exubera by $10 million (12 to 17%). As Nektar’s guidance depends largely on orders placed by Pfizer, the raised guidance suggests that Pfizer has become more optimistic about initial uptake of the drug.

This will certainly be one of the more interesting product launches in diabetes care ever. I’m intrigued by competing factors clinicians, patients, and third-party payers will weigh when deciding to use/reimburse for the drug: convenience (multiple injections vs. multiple inhalations; syringes and needles vs. an inhalation device), dose tailoring (extreme flexibility vs. limited flexibility), cost (3x to 5x higher for Exubera), toxicity (an increased risk of lung toxicity and lung function testing before use for Exubera), and Pfizer’s powerful position in the industry. It’s clear from the comparison, that Pfizer must highlight the convenience and purported QOL benefits (which NICE rejected–see my PCE post from June) of Exubera, as these are the points of relative advantage. To do so, they will likely use the same arguments we saw used in the NICE appraisal, albeit magnified in extreme proportion by thousands of salespeople, medical affairs physicians, and “medical science liasions” in the field. Will they succeed? We shall soon see. Oh…you’d like me to put my opinion on the line? Fine. There likley exists a decent-sized pool of insulin users that is eagerly anticipating launch of the drug. These users will be identified via Pfizer’s rollout campaigns and will contribute the vast majority of initial sales. Following satisfaction of this pool in the first 2 to 3 months, uptake of Exubera will trace a curve similar to the early uptake of Lantus. The time until and magnitude of peak sales will depend largely on patient satisfaction with the device/drug relative to the injections. Again, we shall soon see.

Pfizer’s bold, but flawed, strategy to market only the combination formulation of its experimental CETP inhibitor, torcetrapib, with atorvastatin (Lipitor) has apparently been reversed late in the game. Although Pfizer never detailed its reasons for the initial strategy, it’s obvious that it was predicated on commercial rather than therapeutic considerations. Indeed, there was no therapeutic reason to limit development to a single combination tablet rather than pursue both the combination and separate tablets from the get-go. As this NYT article points out, Pfizer had earlier argued that the cost of developing both the combination and the stand-alone would be prohibitively expensive. Pfizer’s recent decision reveals this excuse for what it was.

Analogies of plausible combination tablet development scenarios abound in the cardiometabolic arena, especially in hypertension and diabetes. Typically, a stand-alone formulation is developed first, tested in combination with a number of different add-on therapies, and then a combination tablet is developed by bridging to the stand-alone data. The advantage to patients of the traditional development scheme is that the new drug is typically approved with broad combination claims (e.g. Actos plus any sulfonylurea), allowing the prescriber maximal flexibility to tailor therapy. Pfizer will be taking the opposite path now with torcetrapib, limiting prescriber flexibility initially to just the combination with Lipitor and only at the dose(s) of Lipitor Pfizer chooses to develop in combination.

As shares of Encysive are trading down 42% as I write this, I can’t help thinking again that most of this industry and its investors operate out of ignorance when it comes to managing its new-drug (and biologics) applications to FDA. And I can’t help wondering why, when information is available for guidance. I’m speaking (again!) about the purgatory known as the approvable decision.

Approvable letters are written by FDA when a drug meets minimum regulatory requirements, avoiding the dreaded “not approvable” decision, but when it lacks the necessary demonstrations of efficacy and safety to warrant marketing approval. In April, I published a report that describes comprehensively the recent history (1998 through 2005) of FDA’s approvable decisions. Over 100 approved products (small molecules and proteins) received such decisions during this time period. Of the eventually approved small molecules that were not approved following the first review cycle only 24% received an approvable letter following the second review cycle. Thus, Encysive joins an inglorious minority of sponsors that failed to achieve approval for its new molecular entity after the second review cycle (I’m assuming Thelin will eventually be approved). Will Encysive go for three approvable letters, or will they change tactics?

If there is one thing I learned when researching these approvable decisions it’s this: Argue with FDA over “matters of judgment,” and argue strenuosuly if you believe you have a good case to make, but make your best case in your NDA/BLA, well prior to the first action date for your NME. Repeating prior arguments, with minimal additional evidence in support of them, to answer an approvable letter is nearly always futile. If FDA states that it needs more data prior accepting your arguments, get FDA the data they want. Giving them data you think they need, if not the same as what they think they need, very rarely results in a winning argument. These and other lessons may be drawn from the historical record as described in the report.