Reliability of cause-specific mortality adjudication in the TORCH clinical trial
L. P. A. McGarvey, M. John, J. A. Anderson, M. T. Zvarich, R. A. Wise (Belfast, Greenford, United Kingdom; Prerow, Germany; RTP, Baltimore, United States Of America)
Source: Annual Congress 2006 - COPD and inflammation
Session: COPD and inflammation
Session type: Oral Presentation
Number: 1999
Disease area: Airway diseases
Abstract Rationale: Cause-specific mortality is an important outcome measure in COPD, but the reliability of this assessment is not known.Methods: The TORCH study, a global multi-site clinical trial of salmeterol / fluticasone propionate 50/500 inhalation, enrolled over 6,000 participants followed for three years. Cause-specific mortality, a tertiary outcome measure, was adjudicated by a committee of 3 physicians based on clinical records, death certificates, and site-investigator narratives. Each death was classified according to a set of pre-specified criteria.Results: Among the first 784 adjudicated deaths, the causes were as follows: Respiratory 36% (COPD 28%, Pneumonia 8%, Other <1%); Cancer 22 % (Lung Cancer 14%, Other 8%); Cardiovascular 27% (Sudden Death 16%, Acute MI 3%, CVA 4%, CHF 3%, Other 1%); Other Causes 10%; Unknown 6%. Deaths were further classified whether they were related to COPD: Definite 39%; Probable 2%; Possible 2%; Unlikely <1%, Not related 50%; Unknown 7%. 100 cases were blindly classified a second time. The cause of death was identical in 83% of cases, (Kappa = 0.80). The assessment of COPD-related death was identical in 84% of cases. (Kappa = 0.73). Interpretation of both Kappa values is indicative of substantial agreement. Discrepant cases were most frequently due to coding of COPD exacerbation vs. Pneumonia, or unknown vs. a specific diagnosis.Conclusion: A clinical endpoint committee can provide a standardized, reliable, and informative adjudication of COPD mortality. Many deaths are related to COPD even if COPD is not the primary cause of death. Sudden death is common in COPD although documented myocardial infarction is infrequent.
Rating:
You must login to grade this presentation.
Share or cite this content
Citations should be made in the following way:
L. P. A. McGarvey, M. John, J. A. Anderson, M. T. Zvarich, R. A. Wise (Belfast, Greenford, United Kingdom; Prerow, Germany; RTP, Baltimore, United States Of America). Reliability of cause-specific mortality adjudication in the TORCH clinical trial. Eur Respir J 2006; 28: Suppl. 50, 1999
You must login to share this Presentation/Article on Twitter, Facebook, LinkedIn or by email.
Member's Comments
Related content which might interest you:
Related content which might interest you: