For drug pipelines, more limp than leap, report says
Tim Anderson, MD
With new and novel medicines launched the last couple of years for diseases like cancer, hepatitis C virus and multiple sclerosis, to name a few, one would think that pharmaceutical pipelines are gradually improving. Think again.
A hot-off-the-press Bernstein Research report shows that, essentially, “Pipeline success rates across all phases of development have been slowly worsening or at best staying flat, depending on the phase (preclinical through registration).”
Drawing on data from consultancy KMR Group in areas like success rate across development phases and product-cycle metrics, the report by Bernstein analyst Dr. Tim Anderson illuminates what appear to be some potentially disturbing trends: across all phases, success rates more often have declined, while it's taking longer for products to go from discovery to market.
One reason: regulators and payers are prompting pharma to take more risks. “What you're seeing is a function of [the industry] pursuing more novel drugs—pushing into disease areas with unmet medical need,” Anderson told MM&M in an exclusive interview. “FDA is forcing folks, and payers are forcing folks, to get into these areas.”
The drug industry's two main problems the last decade have been lagging R&D and the patent cliff, says Anderson. Recent R&D wins—like Boehringer Ingelheim's Pradaxa and Johnson & Johnson's Xarelto in the oral anticoagulant category; Novartis' Gilenya for MS; and Bristol-Myers Squibb's Yervoy and Genentech's Zelboraf and Perjeta for various forms of cancer—suggest a productivity renaissance.
In fact, rates of success have deteriorated. This can be seen most strikingly in Phase II, where win rates went from 34% between 2003-2007, to 25% in 2005-2009, to 22% in 2007-2011. For candidates in Phase III, win rates have gone from 70% to 67% to 65% for the periods 2003-2007, 2005-2009 and 2007-2011, respectively.
The KMR data yield some other striking trends:
- The number of preclinical candidates needed to beget one marketed drug rose 2.5-fold over the past decade. While in 2003-2007, the ratio was 12 preclinical assets needed to get one approved drug, the ratio increased to 24:1 in 2005-2009, then 30:1 for the period between 2007 and 2011.
- The time span between preclinical development and regulatory approval has also inched up, from 11.4 years (for 1999-2001) to 13.7 years (for 2009-2011).
Why does this appear to be happening, especially in the midst of the apparent new-product bonanza? According to Anderson, there are some potentially confounding factors overshadowing the data, including the way that KMR clusters its data, which Anderson calls “best-in-class because it comes directly from the drug companies.” KMR groups data by several years at a time, which he says could mask a more recent potential turnaround.
Then there's the pipeline disruption of recent mega-mergers—three in 2009 (Pfizer+Wyeth, Merck+Schering-Plough and Roche+Genentech)—that have led to fall-out. Companies also may be rightly terminating assets earlier, assassinating those that exhibit incipient signs of hobbled performance. Further, with regulators and payers less apt to approve and pay for “me-too” products, industry's focus has shifted more toward higher-value and, hence, higher-risk products.
“FDA has raised the bar on me-too drugs for the past five to seven years,” Anderson said. “They have gotten more cautious in their approach to safety. That and the payer dynamic are pushing [the industry] into the territory of unmet need. And with that, the risk of failure goes up."
Taken at face value, the KMR data suggest the industry is not making pipeline progress. That runs counter to Anderson's general perception that, for the last one or two years, more novel drugs are starting to reach market. The analyst's own plot of commercially meaningful drug approvals since 2000, and forecast for 2012 and 2013, shows a potential uptick in the pipeline.
Is the industry's R&D malaise showing signs of recovery, or isn't it? After the aforementioned explanations for the discrepancy between the KMR data and Anderson's own sense that things are improving, the report points out that, “Without more granularity it is difficult to know if these factors (i.e., data clustering, R&D disruption of mega-mergers) could be influencing the appearance of [the KMR] data or not.”
At the same time, if the trends bear out and prove true over the next one to two years, this report could be the harbinger of several more lean years for the industry.
Anderson concludes, “It will probably take another year or two to more clearly say whether pharmaceutical industry R&D trends are improving. If this is not happening, then drug companies will continue to limp along, desperately seeking alternate solutions to try and deliver on their growth objectives.”