A few years ago, there were many failures in phase III trials due to suboptimal dosing.
PK/PD modeling tools are helping change that.
Drug development is a complex and expensive process: The cost of bringing a new drug to market is estimated to be between $800 million and $1 billion. Currently, there is a huge gap between the number of candidate drug compounds in testing and the ones that actually get approved. Less than 10% of drugs in phase I clinical trials make it to the approval phase. Two key reasons why drugs fail at late stages are a lack of understanding of the relationship between dose-concentration response and unanticipated safety events. Given this scenario, it is critical to have enabling tools that help predict how a drug will perform in vivo and assist in the success of a clinical therapeutic candidate.
Pharmacokinetics (PK) characterizes the absorption, distribution, metabolism, and elimination properties of a drug. Pharmacodynamics (PD) defines the physiological and biological response to the administered drug. PK/PD modeling establishes a mathematical and theoretical link between these two processes and helps better predict drug action. Integrated PK/PD modeling and computer-assisted trial design via simulation are being incorporated into many drug development programs and are having a growing impact.
PK/PD testing is typically performed at every stage of the drug development process. Because development is becoming increasingly complex, time consuming, and cost intensive, companies are looking to make better use of PK/PD data to eliminate flawed candidates at the beginning and identify those with the best chance of clinical success.
A few years ago, determining dose regimens for clinical trials was often empirical or semi-empirical, and there were many failures in phase III trials due to suboptimal dosing, which ultimately resulted in poor efficacy or excessive toxicity. “Such failures could be devastating to smaller companies, often putting them out of business,” says Martin Graham, PhD, President and Founder of PKPD Inc., Philadelphia. “For larger companies, there are a huge amount of resources and dollars at stake. Hence, several drug development companies started using predictive models earlier in the development process so as to have better outcomes in phase II and phase III.”
Streamlining the process
A major factor in the drug development process is time. Predictive modeling tools can provide invaluable information to better streamline the drug development process. Pharmaceutical companies traditionally perform sequential testing of drug candidates by screening and selecting the best performers at every phase of the clinical drug-development cycle. This can take as long as six to 10 years and cost several hundred million dollars. Hence, it is imperative for companies to adopt technologies that improve the quality of the drug development process and improve speed to market.
“PK/PD modeling can greatly compress timelines by enabling companies to utilize PK/PD data collected in phase I trials,” says Graham. “Instead of a sequential approach, modeling enables a parallel approach, helping jump-start phase II and phase III trial designs. This can significantly reduce the development time at every phase. Added benefits of PK/PD modeling are optimized dose regimens resulting in smarter phase II and phase III trials with a decreased risk of drug failure at the final stage.” He cites the example of a company that benefited from PK/PD modeling at the phase IIb stage by eliminating the need for an additional dose group, thereby reducing the sample size by approximately 60 patients, something that saved six months and more than $1 million.
Traditional PK/PD modeling in drug development defines parameters such as drug dose concentration, drug exposure effects, drug half-life, drug concentrations against time, and drug effects against time. “When used more broadly, quantitative techniques such as drug modeling, disease modeling, trial modeling, and market modeling can support the entire development process, which results in better decisions through explicit consideration of risk and better utilization of knowledge,” says Mark Hovde, senior vice president of marketing for Pharsight Inc., Mountain View, Calif.
However, implementing a PK/PD modeling approach can be challenging. “One has to invest time and resources up front in the drug development process. From a modeling standpoint, one needs to identify and build into the trial design frequent measures of clinical efficacy and/or toxicity as well as representative biomarkers for effective modeling,” says Graham. Researchers engaged in phase I trials often need guidance in the benefits of PK/PD modeling because they have to justify the increased cost and inconvenience associated with additional sampling during phase I/phase II trials. “This can impact the trial design and in some cases make it harder to recruit patients. However, the long-term benefits of an integrated modeling approach can be enormous.”
PK/PD modeling approaches are proving useful in determining relationships between biomarker response, drug levels, and dosing regimens. The PK/PD profile of a drug candidate and the ability to predict a patient’s response to it are critical to the success of clinical trials. Recent advances in molecular biology techniques and a better understanding of targets for various diseases have validated biomarkers as a good clinical indicator of a drug’s therapeutic efficacy.
“Biomarker assays help identify a biological response to a drug candidate. Once a biomarker is clinically validated, trial simulations can be effectively modeled. Biomarkers have the potential to achieve surrogate status that may someday substitute for clinical outcomes in drug development,” says Peter Prince, MSc, a principal scientist at MDS Pharma Services, Montreal, Canada. Some current technologies employed to develop biomarker assays are within the realm of ligand binding and liquid chromatography/tandem mass spectrometry. Future techniques could include matrix-assisted laser desorption/ionization time-of-flight mass spectrometry, array-based techniques for use in proteomics, and pharmacogenomics approaches.
Biomarker assay development offers a unique set of challenges. To make early decisions during drug development, a specific biomarker panel needs to be set up for a target disease so the multiple effects of therapeutic treatment, such as biological mechanisms, efficacy, and safety profile can be investigated. “Other challenging factors in biomarker method development are sensitivity, specificity, binding proteins, proteases, preexisting antibodies, endogenous analytes, reference standards, free versus bound, activity versus concentration, intact versus fragments of analyte, as well as pretreatment and quality of sample preparation,” says Keguan Chen, MD, a principal scientist at MDS.
The FDA and Model-Based Drug Development
The increased cost of bringing drugs to market has been mainly due to the large investment required during the “critical path” years covering preclinical trials through phase III trials. This resulted in a steadily decreasing number of new drug submissions for regulatory approvals. Another key reason has been clinical failure of drugs and the inability to predict a drug’s response early on in the development process. In the spring of 2004, the US Food and Drug Administration (FDA) released a white paper that addressed this issue and proposed the use of model-based approaches, validated biomarkers, and new trial designs to improve the drug development and decision-making process.
There is a major shift in the way the FDA is approaching model-based drug development. The agency is increasingly using modeling and simulations proactively to help design better studies and drug development programs. The FDA also wants to be involved earlier in the drug development decision-making process. A problem for the FDA has been that by the time a drug company comes to FDA for an end-of-phase II meeting, it is already too late for the FDA to have much impact on the drug development program. This led to the initiation of end-of-phase IIa meetings where the FDA representative may come prepared to the meeting with predictive models and expects the sponsor to be prepared to answer questions based on these predictive models.