077: Advanced Waveform Analysis of ECG Signals to Characterise Cardiac Drug Safety Between Caucasian & Japanese Volunteers

Statement of Purpose, Innovation or Hypothesis: Safety assessment during clinical development of candidate drugs involves evaluation of their pro-arrhythmic propensity through detailed ECG studies. Disaggregation of study data by sex and ethnicity, may uncover adverse reaction risk groups. ECG analysis is currently reduced to simplified biomarkers, such as rate and intervals, including QT. However, ECG waveforms are complex and generally sampled with high fidelity (125-1000 Hz). Current methods disregard information contained in the complete morphology and variability of these signals which could potentially detect physiologically relevant changes more robustly. Our novel mathematical method, Symmetric Projection Attractor Reconstruction (SPAR) uses every recorded data point to faithfully replots ECG waveforms generating a new 2D image termed an ‘attractor’. Attractor features amplify small changes in underlying waveforms and can be quantified, generating new metrics of cardiac function. Moxifloxacin is a reversible blocker of the rapid component of the delayed rectifier, potassium current of the cardiac IKr potassium channel and has previously been shown to cause a mean increase of the QTc interval. We hypothesised that SPAR analysis of ECG will more sensitively detect moxifloxacin induced ECG changes, to support safety evaluation during medicines development.

Description of Methods & Materials: Retrospective analysis of ECG time-series data from a Ph1 study of healthy human volunteers cross over of placebo, moxifloxacin (400 mg) in fasted Caucasian (n=13) and Japanese (n=19), male (n=18) and female (n=14) subjects, was analysed with SPAR using a bespoke MATLAB software tool. This transformed 10 second ECG time-series data into corresponding attractors, which were then disaggregated by sex and ethnicity and qualitatively and quantitatively analysed.

Data & Results: Consistent with previous studies, ECG attractors showed high interindividual variability. Females tended towards wider armed attractors compared to males in the placebo baseline state. Moxifloxacin caused ECG and subsequent attractor changes to different extents within individuals. Overall, we showed larger delta attractor changes in Japanese compared to Caucasian subjects. The nature of the observed attractor changes was dependent on gender. Overall, male responders showed a widening while female responders showed a narrowing of the attractor arms.

Interpretation, Conclusion or Significance: This pilot study illustrates the ability of SPAR to highlight changes on the ECG between Caucasian and Japanese ethnicity  may which may be less obvious from QT information alone. These findings suggest that such data should be disaggregated to understand subpopulation and individualised risks. Further analysis is now necessary to measure the visual changes observed in the attractors and quantitatively compare the sensitivity of SPAR and QTc studies on detecting moxifloxacin induced cardiac changes. In addition, future studies will correlate these attractor quantifications to blood drug concentrations, to assess the precision of SPAR at detecting drug-induced changes. The individualised nature of ECGs, which is highlighted by corresponding attractors, indicates that the method may have highest utility in personalised medicine, evaluating each individual’s response and hence risk to drug induced arrythmia.

Disclosures: No conflicts to declare.

Citations/References:

Additional Information/Authors: n/a