The FDA and international regulators require that all new chemical entities (NCEs) undergo a variety of safety tests, including cardiac safety assessments. It is common to hear members of drug development teams question why it is necessary to collect cardiac safety data on their new non-cardiac drug, which has no known preclinical or early clinical cardiac toxicities. Many drugs intended for non-cardiac indications have no obvious link to cardiac side effects. So, why the requirement for the collection of cardiac safety data?
Medication induced sudden cardiac death has been a major issue in drug development since the 1980s, when a number of drugs were removed from the market after they were linked to many unexpected deaths, often in previously healthy individuals. These deaths were ultimately linked to a rare type of ventricular tachycardia known as Torsades de Pointes (TdP), which was not detected in the standard clinical drug trials of the time. After research demonstrated the common mechanism by which these drugs produced TdP, it was shown that all of these drugs produced prolongation of the QT interval on the surface ECG. This led to the ICH E14 guidance on strategies for drug developers to follow as they bring NCEs through clinical development.1
For drug developers, the ICH E14 is currently the ultimate word for planning how to assess proarrhythmic risk. As such, the assessment of new drugs now includes dedicated cardiac studies, called Thorough QT (TQT) trials, to detect the potential for producing QT prolongation, and hence the risk of TdP. Since the adoption of the ICH E14 in 2005, no drug which has been compliant with this guidance has been removed from the market due to ventricular proarrhythmia, pointing to the success of this strategy.
On the flip side, however, there is concern that the strategy of testing for QT prolongation has had a negative effect on drug development. Questions persist about how to handle drugs that produce only a small QTc increase. Not all drugs that increase QTc produce Torsades de Pointes; and yet, the development of promising new drugs has been terminated due to concerns about modest QT prolongation, without any other evidence of proarrhythmia.
Since the QT interval is highly variable and can be difficult to measure precisely, a variety of newer strategies to screen new drugs for proarrhythmic risk are being considered—including increased use of Phase I QT data, increased use of preclinical ion channel assays, and the use of quantitative T-wave morphology assessments.2,3
T-wave morphology assessment is an innovative method for identifying drugs which can produce lethal arrhythmias by analyzing standard 12-lead ECGs for drug induced changes in T-wave shape. Besides lengthening the QT interval, drugs which produce TdP also produce characteristic changes in the shape of the T-wave which often precede the onset of TdP. As a result of this, we look beyond the measurements of the QT interval during drug development and consider the T-wave shape to observe if a drug produces T-wave flattening, notching, or multiple component T-waves. To this end, researchers at the Aalborg University in Denmark and GE Medical have developed a computer algorithm which can perform quantitative T-wave morphology analysis (QTWMA) on digital ECG files.4 The algorithm generates individual scores for T-wave flattening, asymmetry, and notching, and provides a numeric Morphology Combination Score (MCS) descriptor of T-wave morphology. The ability to generate a quantitative score for the shape of T-waves makes it possible to quantitatively analyze the effects of drugs on T-wave morphology. QTWMA research has shown that drugs which produce TdP produce large changes in T-wave morphology, while drugs which lengthen the QT, but don’t cause TdP, produce minimal changes.
Among the difficulties associated with QT measurements are the dependence of QT on heart rate, which requires that QT be corrected for heart rate (QTc), as well as the very high daily variability of QTc (the daily variability of QTc is up to 50-75 ms/day, while we are trying to detect drugs which increase QTc by as little as 5-10 ms). T-wave morphology, and QTWMA scores, fortunately do not vary with changes in heart rate (unless extreme), and therefore there is no need to perform heart rate correction.5 Furthermore, QTWMA scores do not vary as extensively over time as QTc measurements.
A number of published papers have evaluated the utility of QTWMA. One such study evaluated the changes in QTc and T-wave morphology of two different QT prolonging drugs, Moxifloxacin and Sotalol. Moxifloxacin is an antibiotic that is known to prolong the QTc, with a peak increase in QTcF of approximately 10-15ms. However, Moxifloxacin has an extremely low risk of producing TdP; the MCS score was only increased slightly by 0.07 (a typical MCS score of a single ECG ranges from 0.4 to 0.6). In contrast, Sotalol is a very potent antiarrhythmic with a large peak increase in QTcF of 40-60 ms and a high risk of TdP (1-4%). Sotalol produced much larger changes in T-wave morphology than Moxifloxacin—Sotalol increased the MCS score by 0.53 for a 160 mg dose and 1.07 for a 320 mg dose.6 This study, as well as several others which have compared QTWMA and QT assessments, suggest that T-wave morphology analysis can potentially help to better assess the arrhythmogenic potential of drugs which have a small or equivocal QT effect, and may be a very useful adjunct to traditional QT studies.7, 8
There are several roles in which QTWMA may be useful. QTWMA can be performed during typical Phase I studies and may demonstrate a proarrhythmic risk early in drug development. There is some evidence that QTWMA may be more sensitive than standard QTc analysis for detecting proarrhythmic risk. As such, QTWMA may be ideal for Phase I studies, which typically involve small numbers of subjects per cohort and which may therefore be underpowered to detect a modest QT effect. QTWMA can also be performed on ECGs collected during a standard TQT to further refine the assessment of a drug’s proarrhythmic risk. Since QTWMA can be performed on any digital ECG files which have been stored in XML format, it can be performed at the end of a trial, or using retrospective data.
In addition, an exciting potential use for QTWMA is for reexamining potentially useful drugs whose development has been delayed or idled due to an equivocal or small QT signal. The demonstration that a drug produces little or no change in T-wave morphology, despite a small increase in QTc, may be very helpful in finding a partner or the funding to continue development. Similarly, for drugs that are already approved and have adverse labeling as a result of a small QT signal (10-15 ms), the demonstration that the agent produces no change in T-wave morphology might lead to better labeling.
While there is currently no replacement for QT measurements, the FDA has indicated that they will consider quantitative T-wave morphology data with a submission. Regulators, and particularly the FDA, have expressed great interest and have in fact already sponsored and completed a trial assessing the use of T-wave morphology. Regulators have made it clear that they are extremely interested in newer strategies, such as QTWMA, which may provide alternative methods to our current paradigm of relying almost exclusively on the detection of QTc prolongation. Such new methodologies may help improve cardiac safety while reducing the number of drugs which are delayed or terminated in development due to a small QT signal—drugs which may ultimately turn out to be safe and effective.
Robert Kleiman, MD, Vice President, Cardiology and Chief Medical Officer, ERT