TQT studies - this is the end!

Transcription

1 TQT studies - this is the end! Borje Darpo MD, PhD Associate Professor of Cardiology Chief Scientific Officer icardiac Technologies

2 Yes as the only way to confidently exclude that a new drug has a clinically relevant effect on ECG parameters page 2

3 The objective of the initiative was to evaluate whether QT assessment in early phase clinical studies can replace or serve as an alternative to the TQT study We therefore conducted a prospective clinical study in healthy subjects with design similarities with a standard FIH study The underlying idea is to apply exposure response (ER) analysis on robust ECG data from early phase clinical studies without change in standard design Collaborative project between the IQ-consortium, CSRC and FDA Design of study, selection of drugs and doses and statistical analysis discussed and agreed upon with FDA page 3

4 The IQ-CSRC study is a consequence of activities that have been on-going for several years Extensive experience has been gained in FDA by applying ER analysis on TQT study and patient data through centralized (IRT) review of QT studies Good correlation between predicted QT effect in the by timepoint analysis and the ER analysis Sponsors have long claimed that ECG assessment as part of early clinical development could be used, and pointed to anecdotal examples page 4

5 Probably nothing, if seen as a screening tool that must identify all drugs (100%) that prolong the QT interval, i.e. very sensitive The combination between the by timepoint analysis (IUT) and the objective to exclude a QT effect >10 ms, has however rendered the study low specificity If the same data can be generated from a standard early phase clinical pharmacology study, this would represent a more efficient approach and also have other advantages (e.g. liability known early on) The TQT study has successfully prevented drugs with unknown QT liability to gain approval page 5

6 Lomitapide Albiglutide QTcF QTcF (msec) Time (hours) 75 mg Lomitapide adjusted for Placebo 200 mg Lomitapide adjusted for Placebo 75 mg Lomitapide + Ketoconazole adjusted for Ketoconazole alone Ketoconazole adjusted for Placebo Moxifloxacin adjusted for Placebo 54 subjects, 5-way SD (incl. keto), XO study Precision: Mean SD of QTcF 6 to 7 ms 85 subjects, 4 treatment, parallel with nested XO study, 39 days of dosing Precision: Mean SD of QTcI = 6.1 ms Darpo et al. ANE 2013; Doi: /anec Darpo et al. Diabetes Ther. 2013; Doi: /s page 6

7 Vast majority of drugs that cause TdP do this based on concentration dependent QT prolongation - exposure response (ER) analysis of QT effect therefore pharmacologically makes sense If carefully studied in healthy subjects, drugs with a true QT liability will cause concentration-dependent QT prolongation Early QT assessment generates QT data with same high level of confidence as the TQT study, i.e. is not contingent upon changed non-clinical paradigm page 7

8 Median concentration quantiles Mean (90% CI) predicted QTcF prolongation QTCF (ms) Concentration (ng/ml) Linear model Intercept: 2.78 ms (-0.16 to 5.71) with intercept Slope: ms per ng/ml ( to ) Westerberg G et al. Br J Clin Pharm Br J Clin Pharmacol 2015; 79: page 8

9 Linear model with mean intercept fixed to 0 (with variability) Slope: ( ; ) Linear model with intercept: Intercept: 3.9 ms (90% CI: 1.0 to 6.8 ms) Slope ms per ng/ml (0.011 to ms) page 9

10 20 male and female healthy subjects 3 treatment periods 9 subjects were to receive each drug, 6 on placebo Target to have at least 6 on active and 5 on placebo Study drugs: 5 QT-positive drugs, well characterized from previous studies Ondansetron, quinine, dolasetron, moxifloxacin, dofetilide 1 QT negative Levocetirizine Placebo Dosing on 2 days: Day 1: Dose intended to give app. 10 to 12 ms QTc effect Day 2: Dose intended to give app. 15 to 20 ms effect ECG methodology as in TQT studies Primary analysis: Based on exposure response page 10

11 All 5 positive drugs met the prespecified criteria, i.e. the study was able to demonstrate a drug-induced QT effect at the dose identified by FDA The negative drug, levocetirizine, also met the criterion, i.e. a QT effect above 10 ms could be excluded Criteria met for all drugs in sensitivity analysis with data from Day 1 only (lower dose) for positive drugs and Day 2 only (high dose) for negative drug were used Criteria met for all drugs in analysis of parallel groups of subjects Known effect of quinine and dolasetron on cardiac conduction (PR and QRS) identified page 11

12 Red bars denote observed median (IQR) QTcF within each concentration decile Slope, mean ms per ng/ml LB 90% CI UB 90% CI Treatment effect (intercept) ms Cmax Day 1, ng/ml Predicted QTc effect mean, ms LB 90% CI UB 90% CI Criteria * ** Met *: The positive slope is statistically significant **: QTc effect above 10 ms at the Cmax of Day 1 cannot be excluded page 12

13 Slope, mean ms per ng/ml LB 90% CI UB 90% CI Treatment effect (intercept) ms Cmax Day 1, ng/ml Projected QTc effect mean, ms LB 90% CI UB 90% CI Criteria * ** Met page 13

14 Slope, mean ms per ng/ml LB 90% CI UB 90% CI Treatment effect (intercept) ms Cmax Day 2, ng/ml Predicted QTc effect mean, ms LB 90% CI UB 90% CI Criterion * Met *: QTc effect above 10 ms can be excluded at the geometric mean Cmax on Day 2 page 14

15 Drug Slope, mean ms per ng/ml LB 90% CI UB 90% CI Treatment effect ms Cmax Day 1, ng/ml Projected QTc effect mean, ms LB 90% CI* UB 90% CI* Positive drugs (Day 1) Ondansetron Quinine Dolasetron Moxifloxacin Dofetilide* Negative drug (Day 2) Levocetirizine *: Slope from linear model for comparison. Predicted effect for dofetilide using Emax model: 11.6 ms; 90% CI 7.0 to 16.0 page 15

16 Criteria for negative QT assessment: The upper bound of the 2-sided 90% confidence interval (CI) of the predicted placebo-adjusted QTcF is below 10 ms at clinically relevant plasma levels of the drug. *: Darpo B, Garnett C, Keirns J and Stockbridge N.: Implications of the IQ-CSRC prospective study - time to revise ICH E14. Drug Safety 2015; 38: page 16

17 FDA has independently analyzed and confirmed the results Results presented to ICH E14 Discussion group The ICH E14 Implementation group decided in June 2015 at the meeting in Japan that E14 should be revised, initially through a revision of the Q&A document. We are working with FDA to define quality metrics that can be applied to small sized QT studies TQT study waivers are granted as we speak page 17

18 o An drug-induced QTc effect exceeding 10 ms at high, clinically relevant plasma levels has been excluded i.e. the upper bound of the 90% CI of the predicted effect, using exposure response analysis, should be below 10 ms o In the lack of a positive control, the confidence in the data will increase with truly supratherapeutic plasma levels If an app. 4-fold margin of plasma levels as compared to exposure in patients cannot be achieved, there may be a need for a positive control (i.e. a TQT study) o The role of non-clinical assays not fully clear A clean package undoubtedly helps page 18

19 Clean non-clinical package Achieved Cmax in MAD study app. 8-fold above peak therapeutic levels TQT study waiver granted by FDA based on these data To be presented at 2015 Annual Meeting of the American College of Clinical Pharmacology (ACCP), September 27-29, 2015, San Francisco, CA, USA page 19

20 The sponsor s choice Some sponsors may choose to conduct a TQT study for drugs that make it beyond phase 2 only When a sufficient exposure margin cannot be (or were not) achieved in healthy subjects May call for the use of a positive control To confirm or more precisely characterize the QT effect for a drug with a non-negative Early QT assessment Can be done in a TQT study or in patients page 20

21 CONCLUSIONS Early QT assessment is now an acceptable alternative way to confidently exclude that a new drug has a clinically relevant effect on the QT interval The TQT study will not completely disappear as a viable option in select cases Overall, QT assessment will be done more effectively from a resource perspective page 21