Copyright © 2004 by the European Society of Cardiology.
Current Opinion
The data monitoring experience in the MOXCON trial
a Medical Statistics Unit, Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, Keppel Street, London WC1E 7HT, UK
b Cardiovascular Institute, Göteburg University, Sweden
c Cardiology Division, Central Hospital in Rogaland, Stavanger, Norway
d Department of Cardiology, Cleveland Clinic Foundation, Cleveland, OH, USA
e Statistics Collaborative, Washington DC, USA
Received February 17, 2004; revised July 27, 2004; accepted September 9, 2004 * Corresponding author. Fax: +44 20 7637 2853 (E-mail: stuart.pocock{at}lshtm.ac.uk).
Abstract
AIMS: This article describes a challenging data monitoring experience that occurred in a major international randomized placebo-controlled trial in patients with heart failure, in which the accumulating interim data showed an excess of deaths on the active treatment.
METHODS AND RESULTS: The MOXonidine CONgestive Heart Failure trial was a randomized comparison of moxonidine, a central sympathetic inhibitor, with placebo. It was planned to recruit 4500 patients with heart failure. The primary endpoint was all-cause mortality, and average follow-up was anticipated to be around 2.5 years until 724 deaths occurred. The trial Data Monitoring Board (DMB) was to conduct safety monitoring reviews of interim data at least every six months, and make their recommendations to the Executive Committee. Within a few months of the study starting, the Data Monitoring Board (DMB) observed an emerging trend of an excess of deaths on moxonidine compared with placebo. This article describes the sequence of events that unfolded: several DMB meetings to evaluate the accumulating evidence, a DMB recommendation to stop the trial, consequent dialogue with the Executive Committee and sponsor leading to a final decision to stop the trial. Ten months after the first patient was randomized, the study was stopped based on 46 versus 25 deaths in 990 moxonidine and 943 placebo patients, respectively, P=0.01. The final published evidence had 54 versus 32 deaths, P=0.012.
CONCLUSIONS: This study illustrates the problems faced by a DMB, and subsequently the trial Executive Committee and sponsor, in deciding how to act in the face of an emerging (and agonizing) negative trend for mortality in a major international trial.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  T. E. Lindley, D. W. Infanger, M. Rishniw, Y. Zhou, M. F. Doobay, R. V. Sharma, and R. L. Davisson Scavenging superoxide selectively in mouse forebrain is associated with improved cardiac function and survival following myocardial infarction Am J Physiol Regulatory Integrative Comp Physiol, January 1, 2009; 296(1): R1 - R8. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. J Pocock Current controversies in data monitoring for clinical trials Clinical Trials, December 1, 2006; 3(6): 513 - 521. [Abstract] [PDF]
|
|
|
|
 |
|
 S. J Pocock The simplest statistical test: how to check for a difference between treatments. BMJ, May 27, 2006; 332(7552): 1256 - 1258. [Full Text] [PDF]
|
|