Design and Analysis of Thorough QT Studies

Design and Analysis of Thorough QT Studies

By Alex Dmitrienko, PhD, Charles Beasley, MD, Malcolm Mitchell, MD

Comments

By S. Stanley Young, NISS,

Use of a Positive Control Group in Thorough QT Studies

The FDA comments (1) on the possible utility of the positive control group in the thorough QT study: “It is not anticipated that a claim will be generated based on the positive control included in these studies, and the positive control is only to be included (in theory) to confirm assay sensitivity.” and “One argument in support of this requirement (positive control) is that since it is not reasonable to assume that everything will remain constant from study to study, a positive control is the best guard against all known and unknown factors.”

The argument is made that there is need to demonstrate the sensitivity of the experiment. There are a number of reasons that the use of a positive control group does not make sense. First, once an experiment is complete assay sensitivity can determine by using the estimated variability from the experiment. It is a straightforward calculation as sample size, size of the effect to be detected and variability are all known or given. These calculations go under the name of “retrospective power analysis” (2,3). Second, if we were inferring from some species such as dog to man, where the experimentation is in dog, then we would want to know the sensitivity of dog to agents known to be positive in man. Here we are clearly working in the target species, man, so the sensitivity of the organism is not in question. Third, there are few precedents for the use of a positive control group in this situation. For example, if we were testing a new HIV drug, we do not require a positive control drug to demonstrate the sensitivity of the experiment. Most importantly, fourth, suppose that a new compound does indeed increase QT, but by a different mechanism from the model positive compound, hERG. Sensitivity to the model compound would not imply any sensitivity to the new compound following the different mechanism. Here we would be relying on the inherent sensitivity of the target species, human, as we typically do for virtually all other target effects. The classic mechanism is hERG and essentially all new compounds are pre-screened for hERG and are unlikely to advance to man if there is an effect. So, in effect, the positive control is following one mechanism and any new compounds, if they have an effect, are likely to be following a different mechanism. Chemists routinely design out QT effects. For example (4), EPIX Pharmaceuticals has developed a virtual docking method to design out hERG effects, which are then tested in animals before going to man. Fifth, why subject healthy volunteers to a positive agent, if it is contended that the level of effect induced is claimed to increase risk to the individual?

The use of a positive control group in the Thorough QT Study appears to add nothing to the interpretation of the experiment beyond what is easily calculable. The inclusion of a positive control group adds additional expense and creates the chance that the entire experiment will be declared invalid.

References

1. U.S. Food and Drug Administration. TPD Preliminary Concept paper. The clinical evaluation of QT/QTc Interval Prolongation and proarrythmic potential for non-arrythmic drug products. RockvilleMD:

2. Lenth, R.V. Two sample-size practices that I don’t recommend.

[Lenth objects to retrospective power calculations mostly on the grounds that the calculations are too simple; the same information is captured in the p-values, sample sizes, etc. I spoke to Russ Lenth as some length. He agrees that retrospective power analysis adequately addresses the sensitivity question.]

3. Thomas, L. (1997) Retrospective power analysis. Conservation Biology. 11, 276-280.

[Thomas discusses retrospective power from the scientist’s point of view. Here, clearly, estimated variability, sample size, and size of the effect to be detected give the sensitivity of the experiment.]

4. Davies, K. See sidebar.

[They comment, “After synthesizing 15 compounds, a new candidate showed minimal adrenergic activity, but had gained binding to HERG – a common cause of cardiovascular side effects and a “no go” compound. So Shacham’s group had to generate another model, this time for the HERG channel. By docking freshly revised candidates to all three proteins, seeking to maintain potency while minimizing unwanted activity, the group identified PRX-00023 as a clean compound with no alpha activity or QT prolongation.]

1