Biased Safety Reporting in Blinded Randomized Clinical Trials: Meta-Analysis of Angiotensin Receptor Blocker Trials

BACKGROUND Cough is listed as an adverse drug reaction (ADR) on the labels of angiotensin receptor blockers (ARB). However, a causal association with cough has also been reported for angiotensin converting enzyme inhibitors (ACEI), which have frequently been used as comparator drugs in the registration clinical trials of ARBs. This prompted us to examine the possible influence of using comparator drugs with well-known ADRs on the safety reporting of investigational drugs in blinded randomized clinical trials. METHODS AND FINDINGS The double-blinded, randomized clinical trials with comparator drugs were identified in the Japanese dossiers for the new drug applications of ARBs. The risk ratios (RR) of reporting cough and headache in ARB arms were calculated for each ARB by comparing trials using ACEIs and trials using non-ACEIs, were then combined with a meta-analysis. 23 trials with a total of 6643 patients were identified, consisting 6 trials using an ACEI comparator including 819 ARB patients and 17 trials using a non-ACEI comparator including 5824 ARB patients. The combined RR of cough reporting was significantly elevated (20.77; 95% confidence interval [CI], 7.47 to 57.76), indicating more frequent reporting of cough in clinical trials using an ACEI comparator. In contrast, the combined RR of headache, a negative control, was insignificant (1.45; 95% CI, 0.34 to 6.22). CONCLUSION The use of comparators with well-known ADRs in blinded randomized trials produces potential bias in the reporting frequency of ADRs for investigational drugs. The selection of appropriate comparator drugs should be critical in unbiased safety assessment in double-blinded, randomized clinical trials and thus have relevance in reviewing the safety results from a regulatory point of view.