Pharma R&D Today

Ideas and Insight supporting all stages of Drug Discovery & Development

Select category
Search this blog

Reducing the Risk of Late-Phase Failures

Posted on July 5th, 2016 by in Pharma R&D

microscope-772297_1920

Late-stage failures are common. The likelihood of a phase transition from Phase 2 to 3 is about 30%, and moving from Phase 3 to submission is only about 60% likely. The majority of late-stage failures are due to a lack of demonstrated efficacy (56%), followed by safety issues (28%)[1]. These abysmal stats are in the face of novel compounds coming from molecular biology.  Why so many failures, and how to avoid them?

Mismatch of drug to disease

I had a boss who believed that any compound could be developed – the key was finding the right development plan. Meaning, there should be a good match between compound properties and the disease. Thalidimide is, unfortunately, a good example. The thalidomide tragedy can be viewed as a good drug and the wrong disease. If the immune pharmacology profile was initially mapped to Leprosy or multiple myeloma as the primary indication, the anti-nausea/sedative effects of thalidomide would not have been developed for use in pregnant mothers.

Look for a good match early in clinic: The original exploratory clinical research with entanerocept (i.e. Enbrel) was done in HIV, septic shock, and rheumatoid arthritis (RA) patients. RA was chosen because there was an obvious good match between properties and disease. Often, the initial diseases are chosen for commercial value, with less concern for the drug’s profile. When the compound fails or is borderline in efficacy, and the company still progresses to Phase 3, the result is a late-phase failure or a non-approval. Bad drug or bad plan – or both?

A precocious phase transition

Louis Shreiner[2]  coined the term “Learn and Confirm” to denote the separation of early and late-phase clinical development. Early or “Learn” phases are for characterizing the drug. Describing dose-response, testing biomarkers, and patient selection are the goals of early phases. The late phases, for efficacy and safety, are for confirmation because they are expensive and long in duration. If learning drug characteristics still occurs in a late phase, there is high risk of failure. Typically, there are budget and time pressures to expedite clinical development by combining or overlapping phases and by limiting questions to only those required for phase transition.

I worked with a biotech company that was developing a drug for acute stroke. The FDA strongly suggested postponing Phase 3 in order to learn information about dose and safety. After much pleading by the biotech, the FDA allowed the transition to Phase 3, but warned that any data inconsistencies will result in a non-approval. The company announced that the FDA-endorsed going to Phase 3, its stock price went meteoric, the trials failed, the project was terminated, and the biotech is gone.

Here are practices that can be used to reduce the risk of late-phase failures[3],[4]:

  1. Sufficiently link the molecular target and the disease with PK/PD. First, use systems pathway identification and translational medicine to link the target to human disease. Second, follow the “three pillars”[5] approach, which is: demonstrate adequate drug concentrations at the receptor (pillar 1), adequate binding to the target (pillar 2), and detection of downstream pharmacologic effects (pillar 3). All three pillars will ensure a valid hypothesis-driven, proof-of-concept trial.
  2. Biomarkers need to be developed. Especially for novel mechanisms, biomarkers are used in modeling PK/PD, optimizing doses, and predicting patient outcomes. Pharmacodynamic biomarkers are specific for the compound pharmacologic actions; safety, prognostic, and diagnostic biomarkers give information about the status of the patient and disease. This knowledge is greatly valued for de-risking late-phase trials[6] .
  3. Choose the right patients who are sensitive and responsive to the drug. Using patient sub-group enrichment and stratification strategies, one can identify patient characteristics which give good drug response. The best mix of patients would give an adequate positive response rate and appease commercial needs.
  4. Be realistic about the safety margins. The first safety signals and exposure margins are constructed in the preclinical phase. Many times these are overestimated and give a false impression of a wide safety buffer. I worked on a neuropathic pain compound where preclinical data showed a 20-fold margin between analgesic effect and seizures. However, because the drug had genetically-influenced, variable PK, the margin disappeared in clinic. Additionally, developing safety biomarkers and disease systems mapping can help to avoid safety surprises in late phase.
  5. Use modeling to help optimize trial conditions and give recommendations for going forward. Modeling PK/PD, quantitative systems pharmacology (QSP), and clinical trial simulation are very powerful tools for optimizing late-phase trial design. Exposure/response, probability of meeting success criteria, differentiation from comparators, can be modeled. Models can be convincing to regulators and are routinely used to answer review questions, instead of conducting more trials. An example of QSP: I worked on an allosteric dopamine sub-receptor agonist for use in schizophrenia. In Silico[7], using literature and preclinical data, developed a QSP model that allowed me to quantify the influence of concomitant medications, choice of patient subgroups, and range of dose.

Late-phase failure rates are high, and expensive, and causes are due primarily to the pressures of rapidly moving ahead because “the patients are waiting.” If we keep in mind the tools I listed above, it’s possible to make rational decisions to raise the probability of success in late phase.

[1] Clinical Development Success Rates for Investigational Drugs Michael Hay, David W Thomas, John L Craighead, Celia Economides & Jesse Rosenthal; Nature Biotechnology; volume 32, number 1; January 2014

[2] Learning Versus Confirming In Clinical Drug Development; Sheiner LB.; Clin Pharmacol Ther; Mar; 61(3): 275-91,1997.

[3] Improving Productivity in Pharmaceutical Research and Development: The Role of Clinical Pharmacology, Duke-Margolis Center for Health Policy, June 2016.

[4] Can The Flow Of Medicines Be Improved? Fundamental Pharmacokinetic And Pharmacological Principles Toward Improving Phase II Survival; Paul Morgan1, Piet H. Van Der Graaf, John Arrowsmith, Doug E. Feltner, Kira S. Drummond, Craig D. Wegner, Steve D.A. Street; Drug Discovery Today; Volume 17, Issues 9–10, May 2012, Pages 419–424

[5] ibid

[6] Improving Productivity in Pharmaceutical Research and Development: The Role of Clinical Pharmacology, Duke-Margolis Center for Health Policy, June 2016.

[7] www.in-silico-biosciences.com

 


All opinions shared in this post are the author’s own.

R&D Solutions for Pharma & Life Sciences

We're happy to discuss your needs and show you how Elsevier's Solution can help.

Contact Sales