Japanese and American reports of randomized trials: differences in the reporting of adverse effects

We sought to identify differences in the description of adverse drug experiences in reports of randomized clinical trials (RCTs) from the United States and Japan, using diclofenac and simvastatin as test drugs. Reports were identified in Medline (Index Medicus 1966-1990), EMBASE (Excerpta Medica 1974-1990), JAPICDOC (1979-1990), and JOIS-III (JMEDICINE 1980-1990). In each search keywords describing study design were paired with the drugs' generic names, chemical names, and development numbers. Twenty-seven U.S. reports (18 for diclofenac and 9 for simvastatin) and 22 Japanese reports (17 for diclofenac and 5 for simvastatin) identified in these four databases were selected for review. For each paper we identified the relation of the article to the data (preliminary, primary, and secondary reports, reviews), the means of identifying adverse reactions, the principal outcomes of the trials, and a variety of descriptive measures relating to study design, authorship, and elements of presentation. With few exceptions, Japanese reports were not indexed in English-language databases, and studies from the United States were not carried out in the Japanese databases. The Japanese literature consisted exclusively of primary reports of clinical trials, whereas the U.S. literature was dominated by review articles and secondary reports of data from trials not fully published elsewhere. Japanese reports contained more detail on adverse experiences but reported principally those attributed to the drugs by attending clinicians. U.S. reports by contrast offered little detail but tended to include all adverse experiences, whether or not clinically attributed to drugs. A preponderance of U.S. articles reported significant differences between drugs in safety or treatment efficacy, whereas only one third of the Japanese articles did so for the same agents. Reports from both countries offered few details of the methods used to gather information on adverse drug experiences, and as a result the reported absolute frequencies of such events are difficult to compare between trials or to generalize to other settings. In conclusion, the reporting of adverse reactions in clinical trials is inadequate in both the United States and Japanese literature. The shortcomings are complementary in that reports of U.S. trials contain insufficient detail and Japanese reports do not interpret or synthesize experience. Clinical research into drug safety in both countries could be improved through the adoption of simple standards of clarity and consistency in the monitoring and reporting of drug adverse effects.