gms | German Medical Science

MAINZ//2011: 56. GMDS-Jahrestagung und 6. DGEpi-Jahrestagung

Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie e. V.
Deutsche Gesellschaft für Epidemiologie e. V.

26. - 29.09.2011 in Mainz

To "correct" or to "adjust" for heart rate changes in the evaluation of QT interval prolongation

Meeting Abstract

Search Medline for

  • Arne Ring - Diabetes Trials Unit, OCDEM, University of Oxford, Oxford
  • Robert Schall - Department of Mathematical Statistics and Actuarial Science, University of the Free State, Bloemfontein

Mainz//2011. 56. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie (gmds), 6. Jahrestagung der Deutschen Gesellschaft für Epidemiologie (DGEpi). Mainz, 26.-29.09.2011. Düsseldorf: German Medical Science GMS Publishing House; 2011. Doc11gmds048

DOI: 10.3205/11gmds048, URN: urn:nbn:de:0183-11gmds0483

Published: September 20, 2011

© 2011 Ring et al.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License ( You are free: to Share – to copy, distribute and transmit the work, provided the original author and source are credited.



The assessment of QT prolongation in thorough QT (TQT) trials according to ICH E14 [1] is often based on a two-step approach [2]:

Determination of the heart rate correction (e.g. slope between log(QT) and log(RR)) based on the data of this trial;
Estimation of the placebo-adjusted heart rate corrected QTc interval (ΔQTc) following active study treatment.

Heart rate correction is generally seen as the standard method to analyse QT interval data, specifically when the study drug changes the heart rate by some relevant amount (> 3bpm). The estimation of the treatment effect ΔQTc is performed conditional on the outcome of step 1. It has been shown that even slight changes of the QT/RR slope estimate may have a substantial impact on the estimated placebo-corrected ΔQTc when the heart rate is altered during the study.

Generally, the two step approach assumes that the estimated slope is determined exactly. However, as usual with “natural” data, the heart rate correction can only be estimated with some degree of uncertainty. Therefore, the 90% confidence intervals (CI) of the treatment effect ΔQTc are anti-conservative when using the two-step approach.

One option to account for the uncertainty of the slope estimate is to involve not only a single value of the slope estimate, but to perform the two-stage approach based on the estimated distribution of the slope estimate.

The preferred option is to perform a one-step assessment of QT prolongation that adjusts for the heart rate as a covariate. It has been shown that this approach is effective and holds the alpha level for the estimation of ΔQTc [3].

We will compare both options and show their impact on the treatment estimates of the ΔQTc intervals using various statistical models.


ICH E14. The Clinical Evaluation of QT/QTc Interval Prolongation and Proarrhythmic Potential for Non-Antiarrhythmic Drugs. CHMP/ICH/2/04.
Schall R, Ring A. Mixed models for data from thorough QT studies: Part 1. Assessment of marginal QT prolongation. Pharm Statist. DOI: 10.1002/pst.463. External link
Schall R. Mixed models for data from thorough QT studies: Part 2. One-step assessment of conditional QT prolongation. Pharm Statist. DOI: 10.1002/pst.465. External link