Article type
Year
Abstract
Background: Data for clinical trials were presented by FDA reviewers and pharmaceutical companies at the Oncologic Drugs Advisory Committee of the Food and Drug Administration (FDA/ODAC) hearing on the safety of erythropoietin/darbepoetin (EPO/Darb) in May 2004.
Objectives: In a meta-analysis assessing the relative risk of thromboembolic events (TEs) from EPO/Darb in cancer patients, we compared results obtained from FDA reports and journal publications with data submitted by the pharmaceutical industry.
Methods: RCTs of EPO/Darb were identified by systematic search of MEDLINE, EMBASE, Cochrane Library (1985-2005). Additionally, we evaluated all reports and presentations submitted to the FDA/ODAC. Data from different sources were compared. Discrepant data were grouped into 2 subsets: A) data reported in abstracts, full texts or FDA reviewer's documents, B) data submitted by pharmaceutical companies.
Results: We identified 33 trials with 6,255 participants that reported TEs in cancer patients receiving EPO/Darb versus placebo/no treatment. Of all studies, data for 12 studies were available from different sources. For 2 of the 12 studies the data reported in different publications was identical. For 10 studies the results reported were discrepant. Overall the pooled estimates of subset A and B were comparable. However, although the same trials were analysed, less patients were reported in subset B compared to subset A. More important, the overall numbers of TEs reported in subsets A and B were comparable in the EPO/Darb groups, but not in the control groups, where more TEs were reported in subset B. As a consequence, the overall estimate for TEs is smaller in subset B than in subset A. One reason for the differences observed might be the use of different definitions for TEs. See table below.
Conclusion: Data for identical trials reported in different sources tend to be discrepant and may influence overall effect estimates. In practice it is difficult to assess the reliability of the data and to decide which data to include into meta-analysis.
Trials
Participants
EPO/Darb; events/sample
Control; events/
sample
RR
(95% CI)
Subset A
10
2,984
104/1,563 (6.7%)
60/1,421 (4.2%)
1.66 (1.18; 2.33)
Subset B
10
2,930
109/1,523 (7.1%)
72/1,407 (5.2%)
1.46 (1.07; 1.99)
Objectives: In a meta-analysis assessing the relative risk of thromboembolic events (TEs) from EPO/Darb in cancer patients, we compared results obtained from FDA reports and journal publications with data submitted by the pharmaceutical industry.
Methods: RCTs of EPO/Darb were identified by systematic search of MEDLINE, EMBASE, Cochrane Library (1985-2005). Additionally, we evaluated all reports and presentations submitted to the FDA/ODAC. Data from different sources were compared. Discrepant data were grouped into 2 subsets: A) data reported in abstracts, full texts or FDA reviewer's documents, B) data submitted by pharmaceutical companies.
Results: We identified 33 trials with 6,255 participants that reported TEs in cancer patients receiving EPO/Darb versus placebo/no treatment. Of all studies, data for 12 studies were available from different sources. For 2 of the 12 studies the data reported in different publications was identical. For 10 studies the results reported were discrepant. Overall the pooled estimates of subset A and B were comparable. However, although the same trials were analysed, less patients were reported in subset B compared to subset A. More important, the overall numbers of TEs reported in subsets A and B were comparable in the EPO/Darb groups, but not in the control groups, where more TEs were reported in subset B. As a consequence, the overall estimate for TEs is smaller in subset B than in subset A. One reason for the differences observed might be the use of different definitions for TEs. See table below.
Conclusion: Data for identical trials reported in different sources tend to be discrepant and may influence overall effect estimates. In practice it is difficult to assess the reliability of the data and to decide which data to include into meta-analysis.
Trials
Participants
EPO/Darb; events/sample
Control; events/
sample
RR
(95% CI)
Subset A
10
2,984
104/1,563 (6.7%)
60/1,421 (4.2%)
1.66 (1.18; 2.33)
Subset B
10
2,930
109/1,523 (7.1%)
72/1,407 (5.2%)
1.46 (1.07; 1.99)