Key result
Concentration-QT analyses provide a powerful complementary method to standard timepoint-based assessments for evaluating drug-induced QT prolongation, though they should not yet replace primary endpoints.
Answers are grounded in this paper and cite their sources.
Concentration-QT analyses provide valuable complementary insights for evaluating drug-induced ventricular repolarization, though questions remain regarding their routine use in drugs without apparent QT effects.
The ICH E14 guidance1 outlines a rigorous assessment on the potential for a drug to prolong ventricular repolarization in a “thorough QT/QTc study” that includes the use of a positive control, characterization of the QT/QTc response at a supratherapeutic dose, and methods for reducing variability in the measurement of the QTcf interval (replicate ECGs, use of a centralized core laboratory to measure QTcf interval, etc). The primary analysis of a “thorough QT/QTc study” is based on an assessment of the upper bound of the 95% one-sided confidence interval for the largest time-matched mean effect of the drug on the QTc interval (which needs to exclude 10 ms). With regard to an assessment of concentration-QT relationship, the guidance document suggested “establishing the relationship of drug concentrations to changes in QT/QTc interval may provide additional information to assist the planning and interpretation of studies assessing cardiac repolarization,” although it noted that “this area is under active investigation.” In this issue of the Journal, Garnett and coworkers examine the potential role of the assessment of concentration-QT relationship and provide 3 examples that illustrate how this type of assessment has been useful during regulatory review. The authors go a step further than the ICH E14 guidance and state that “regulatory reviews routinely include a characterization of the C-QT relationship,” consistent with the statement made by Norman Stockbridge, MD, PhD, Director, Cardio-Renal Drug Products, Head, Interdisciplinary Review Team for QT, CDER, FDA: “Regulatory review of QT study is not complete without an assessment of concentration-QTc relationship.”2 While the potential benefits of a concentration-QT assessment are becoming clear when there is a relationship between the concentration of a drug and its effects on QT interval, important questions remain unanswered about the role of this assessment in the majority of drugs that have no apparent effect on QT. When a drug has a discernable effect on ventricular repolarization, an assessment of the relationship between the concentration of the drug and its effect on QT/QTc interval can utilize powerful quantitative methods that clearly have a number of advantages over the standard E14 (timepoint-based) assessment. An evaluation of the concentration-QT relationship integrates data across time points as well as all the treatment groups, and takes into account individual responses (independent of variable Tmax or QT effect) instead of averaging the QT response at each time point. In addition, the assessment of concentration-QT relationships allows an analysis that can predict the magnitude of the effect of a drug on QT interval from prior data collected in other stages in development. Such analyses allow a better understanding of the uncertainty in response and assess impact on outliers. However, the cost and effort related to concentration-QT analyses are not trivial. These analyses require considerable effort in the preparation of the data, development of the population pharmacokinetics (PK) structural model, quantifying the pharmacodynamics model, and evaluating model appropriateness and identifiability. The analytic approaches to PK-QTc modeling are still evolving. The development of empiric (data-dependent) as well as mechanistic models (incorporating disease elements in addition to circadian effects) requires skilled and experienced modelers. In order to justify the resources required for concentration-QT analyses, this complex approach should provide tangible drug development and regulatory benefits that either reduce subject number, simplify study conduct, and/or provide a better understanding of problematic studies. In addition, the development of concentration-QT models should be operator independent. The analytic methodologies used in these models need to be rigorously evaluated by statisticians involved in drug development, and potential modeling issues need to be identified and their management delineated. Ultimately, the application and interpretation of these analyses should be globally harmonized. The examples provided by Garnett et al illustrate 3 clear cases in which an analysis of the concentration-QT relationship provided important insights that were useful in the regulatory review of the drug. The first example demonstrated how the analysis of the concentration-QT relationship can be used to predict QT prolongation at doses lower than those studied. The second example illustrated how the analysis of the concentration-QT relationship can clarify ambiguous results from an underpowered thorough QT study in which the effect on QT appeared greater at the lower dose than the higher dose study. In this example, the analysis of the concentration-QT relationship supported an argument for no effect on QT even though the study was nominally positive. The third example demonstrated how an evaluation of the moxifloxacin concentration-QT relationship salvaged the assessment of assay sensitivity in a thorough QT study. Garnett's work is based on certain critical assumptions, perhaps the most important of which is that an exposure/response relationship is physiologically plausible. Clinical and biological literature asserts that a linear pharmacodynamic model is nonphysiological. However, in drug development, the maximum concentration of a drug evaluated is often not large enough to ascertain the plateau of pharmacodynamic effect (for safety reasons). Thus, the commonly used therapeutic dose and the supratherapeutic dose may only be amenable to a nonphysiologic linear relationship, thus running the risk of significantly over-predicting the magnitude of QTc prolongation at concentrations/doses greater than those actually evaluated. The challenge confronting us is to determine when concentration-QT analyses should be done. Should these analyses be done for every drug even in the absence of a credible signal? If not, what are the criteria that should be used to trigger the need for a concentration-QT analysis? If these analyses are done for every drug, how will these analyses be interpreted when there is no relationship between the concentration of a drug and the change in QTc? Are there analytic and statistical methods that can clearly define the lack of such a relationship? What statistic should be used to demonstrate that there is no relationship between the concentration of a drug and QTc? To demonstrate no effect using the derived slope β, some equivalence bounds around the slope of zero would have to be defined in advance. How close to zero does the slope have to be to demonstrate that no relationship exists? If there is a weak relationship between the concentration and QT, how will regulatory authorities interpret and use this information in their review? These questions need to be addressed before concentration-QT analyses become a routine part of the review of all thorough QT studies. Merck collects samples for plasma PK in all thorough QTc studies with flexibility to analyze (for concentration/QTc analyses) when there is a perceived need to clarify the primary analyses specified in the E14 guidance. In cases where there appears to be an effect of a drug on QT interval, we often believe that there is considerable added value in characterizing the concentration-QT relationship in a quantitative fashion. However, since most compounds in development are selected in part because there is no effect of the drug on ventricular repolarization during the extensive preclinical evaluation of the drug, one would expect that the majority of drugs would have completely negative thorough QTc studies. In these cases, we tend to analyze a limited number of PK samples (without exploring concentration-QT analyses) solely to document that the study achieved the anticipated exposures of the drug. Despite the unanswered questions mentioned above and the resources involved in concentration-QT analyses, preliminary results from the initial application of concentration-QT analyses as described by Garnett et al are promising and justify further effort to clarify and develop the potential utility of concentration-QT analyses in the regulatory review of thorough QT studies. However, at the present time, due to the complexity of these analyses and the lack of clarity on how these analyses will be used and interpreted in drugs without an effect on ventricular repolarization, concentration-QT analyses should not be used as the primary endpoint analysis of thorough QT studies. Rather, these concentration-QT analyses should be encouraged, wherever possible, as a complementary method when a drug is suspected of having an effect on QTc interval. In addition, a number of sponsors have used concentration-QT analyses of early phase I studies to improve or enhance the design of “thorough QT/QTc studies” and to support internal decision-making based on concentration QT analyses. Ultimately, more experience is necessary to harvest the potential power of concentration-QT analyses and would benefit from the type of collaboration that is being developed in the Cardiac Safety Research Consortium (http://www.cardiac-safety.org), a public-private partnership and a Critical Path Initiative that includes industry sponsors, regulators, and academics.
No takes yet. Share an insight, caveat, or question.
Bloomfield et al. (2007) conducted an editorial in Drug-induced QT prolongation. Concentration-QTc analysis vs. Standard timepoint-based E14 assessment was evaluated. Concentration-QT analyses provide a powerful complementary method to standard timepoint-based assessments for evaluating drug-induced QT prolongation, though they should not yet replace primary endpoints.
Synapse has enriched 10 related papers. Consider them for comparative context: