Challenges of Early Clinical Development


I joined Bristol-Myers Squibb (BMS) in 2000 soon after Dr. Leon Rosenberg, ex-Chief of Medicine at Yale, had  stepped down as Chief Scientific Officer BMS, and was warning about the dearth of physician-scientist support.  As a yeoman drug developer closely associated with both drug discovery and early clinical development teams, I learned that early clinical development consisted of template protocols for Single Ascending and Multiple Ascending (SAD and MAD), first-in-human studies, conducted by clinical pharmacologists and kineticists (Clin Pharm). These studies were referred to ungraciously as “feed’em and bleed ‘em”. They focused on acute safety/tolerability but provided little insight into possible early efficacy. Management prioritized a goal to better leverage emerging biomarkers to get more clinical value out of the these early studies. To this end one of the first initiatives was to hire more physicians into the Clin Pharm group.


At first, there was not a natural fit between the new physicians and the entrenched PK scientists who managed the departments. This also created a vacuum whereby “late development” representatives could stick our noses into things we weren’t entirely expert at. One specific example sticks with me;  our Vice-President of late development dismissed the idea of including an oral glucose tolerance test in an early study of a new DPP4-inhibitor because it “was a diagnostic test that lacked clinical relevancy”. I was too new to raise a fuss and thought that an additional “food effect” study would be good enough to look for the expected increase in post-prandial incretin hormones that could confirm the mechanism of action. I didn’t understand at the time that a “food effect” study involves drug administration within 20 minutes AFTER a meal is consumed, too late for a DPP4-inhibitor to work, far less be able to discern small potential differences between doses.  That was a big fail that required much work to get keep the program on track.   


Management moved to empower clin pharm further by changing the name to “Translational Medicine” and hiring new leaders with true academic research bone fides. This empowerment worked so well that at two different companies, separated by 10 years, I was unable to convince translational medicine (TM) to consider a stepped hyperglycemic clamp study to clarify how glucose concentrations might impact the dose-response curves for new SGLT2-inhibitors. Part of the fault lay in my inability to finesse the idea but at the time it also seemed that some   resistance reflected the new perspective that TM professionals were the only people qualified to conceive of and execute such studies. We were told to stand down on another project and let clin pharm takeover conduct of a hyperlactate clamp for a novel gluconeogenesis inhibitor even though the reason it needed to be done was because a clin pharm  study in combination had resulted in 2 cases of lactic acidosis. Noone was irrevocably hurt. I had the opportunity to talk directly with the medical reviewer at FDA because the entire clin pharm group happened to be unavailable the day he  called to put us on medical hold. We avoided this by doing an outpatient proof-of-concept study with intensive proactive safety monitoring rather than a mechanistic TM study. The drug was shown to work but not well enough.


One further management innovation was to task TM with responsibility through POC (proof-of-concept) for new agents.  By focusing on a true Go-No-Go milestone, it was thought, TM might be drawn into alignment with the overall drug development enterprise. This incentive was misguided because early clinical development does not have the budget or the staff to execute multi-center trials. In diabetes, in particular,  real POC requires at least 1-3 months with a minimum of 30 subjects/group often across more than one experimental dose. The business pressure to deliver a certain number of POCs per year (and hence receive bonuses)   incentivized even the most well-meaning colleagues to focus more on studies that were achievable as opposed to optimally instructional to the program.  


It is probably fair to say still that the voice and impact of the ethos of the “physician-scientist” is not fully leveraged in early drug development in any institutionalized way.  For one thing, there simply aren’t that many physicians or scientists with a formal background in human clinical research.   From a business perspective companies will continue to hesitate to think about early mechanistic studies out of a sense of risk of the unknown. In large companies it can be argued that the value of a risky but meaningful study is that it can inform Go-No-Go decision making, i.e. a well-designed study that yields negative results can be valuable. In contrast, for a small company, a poorly designed study that appears to yield positive results may be crucial for partnering.


My experience going from large pharma to small biotech and back is that there is much untapped engagement within development teams. Even though there are so many players, most of whom do not have specific healthcare training, there is an abundance of goodwill in terms of wanting to contribute meaningfully to important new medicines. But there is a general reluctance to exercise judgment, partly due to the hierarchical nature of business but also because there isn’t any definitive textbook and more of a tendency for participants to overweight their most recent experiences rather than extract general heuristics. In an industry where so many projects are expected to fail, people don’t always discriminate good plans from bad plans because they have no way of gauging why they succeeded or failed. Obviously that is a key reason people reach out to Kinexum, looking for a reliable authority.  Clearly this is an exciting aspect of drug development where no one has cornered the market.