Types, frequencies, and burden of nonspecific adverse events of drugs: analysis of randomized placebo‐controlled clinical trials

PURPOSE: Scarce studies analyzing adverse event (AE) data from randomized placebo-controlled clinical trials (RPCCTs) of selected illnesses suggested that a substantial proportion of collected AEs are unrelated to the drug taken. This study analyzed the nonspecific AEs occurring with active-drug exposure in RPCCTs for a large range of medical conditions. METHODS: Randomized placebo-controlled clinical trials published in five prominent medical journals during 2006-2012 were searched. Only trials that evaluated orally or parenterally administered active drugs versus placebo in a head-to-head setting were selected. For AEs reported from ≥10 RPCCTs, Pearson's correlation coefficients (r) were calculated to determine the relationship between AE rates in placebo and active-drug recipients. Random-effects meta-analyses were used to compute proportions of nonspecific AEs, which were truncated at a maximum of 100%, in active-drug recipients. RESULTS: We included 231 trials addressing various medical domains or healthy participants. For the 88 analyzed AE variables, AE rates for placebo and active-drug recipients were in general strongly correlated (r > 0.50) or very strongly correlated (r > 0.80). The pooled proportions of nonspecific AEs for the active-drug recipients were 96.8% (95%CI: 95.5-98.1) for any AEs, 100% (97.9-100) for serious AEs, and 77.7% (72.7-83.2) for drug-related AEs. Results were similar for individual medical domains and healthy participants. The pooled proportion of nonspecificity of 82 system organ class and individual AE types ranged from 38% to 100%. CONCLUSION: The large proportion of nonspecific AEs reported in active-drug recipients of RPCCTs, including serious and drug-related AEs, highlights the limitations of clinical trial data to determine the tolerability of drugs.