Table 1

Cases with number of deaths in record that are discordant with the corresponding publication

PopulationWas death a specified outcome* DefineReporting module or recordPublication
Arm 1Arm 2Arm 1Arm 2
Case 1
Lung cancerYes Survival is a secondary outcomeFollow up: While on study drug+30 d after last dose (estimated 4 mo)Follow up: From random assignment until first day of progression or until death
SummaryBoth record [NCT00085839] and the publication [PMID18281658] reported hazards ratios for survival and mean survival in months, but not the number of deaths for the outcome. Both reported deaths under serious adverse events, but counts differed between record and report. In addition the publication reported deaths in the flow diagram, while the record did not. The total number of deaths is discrepant between record and publication; however, neither is likely to represent the total number of deaths that occurred during the study.
Case 2
Multiple myelomaNoFollow up: Up to 18 moFollow up: Enrolled 2/06-12/06, analysis through 8/2007
SummaryBoth record [NCT00259740] and publication [PMID19714603] reported 1 death per arm in the participant flow. The total number of deaths is discrepant between record and publication, however, since the publication also reported 5 deaths under SAE.
Case 3
Refractory prostate cancerYes Survival is the primary outcomeFollow up: Analyzed through 9/2009Follow up: Analyzed through 9/2009
SAE0/371 sudden death1/371 sudden death275/371227/371
SummaryThe record [NCT00417079] reported hazards ratios for survival as well as survival in months, but not the total number of deaths per arm for this outcome. The publication [PMID20888992] reported a large number of deaths per arms for the outcome of survival (as) and also a large number of deaths under SAE. The numerators and denominators differed slightly based on intention to treat analyses or per protocol analyses. The record reported only one death under SAE; although based on the survival analysis, it appeared likely that the total number of deaths in the study was higher. The total number of deaths is discrepant between record and report.
Case 4
Chronic Obstructive Pulmonary DiseaseYes Death is a secondary outcomeFollow up: 52 wkFollow up: 52 wk
SAE1/778 sudden death; 0/778 death3/790 sudden death; 2/790 death-/778-/790
SummaryThe record [NCT00297115] reported 25 per arm as number analyzed in the outcome module and defined the number analyzed as the number died. Further, the record reports deaths under SAE using two different death definitions (‘sudden death’ and ‘death’), while the publication [PMID19716960] does not report any. Assuming that the deaths reported under SAE in the record are included in those reported for the outcome of death, the total number of deaths is consistent across record and publication. The publication describes 2 trials of similar design with two separate NCT number, but only the results corresponding to the trial in the index record were compared.
Case 5
Prostate cancerYes Death is a secondary outcomeFollow up: From start of therapy up to 30 d after last doseFollow up: Duration of therapy+30 d
SummaryThe record [NCT00385580] reported results for 2 arms. The publication presents only results for Arm 2. The report shows 2 deaths in the outcome module, but none under SAE. The publication [PMID19920114] shows 2 deaths under SAE. The number of deaths reported for this arm was consistent between record and publication.
  • *In the

  • Data collection in on Feb 14 2012.

  • Abbreviations:,; D/C, discontinuation; NCT, National Clinical Trial (number); SAE, serious adverse events; – (dash), not reported.