Background: The U.S. Food and Drug Administration (FDA) approves new molecular entities (NMEs) based on clinical trials submitted by the sponsor. A subset of these trials are considered "pivotal" in demonstrating the NME's safety and efficacy, and are the trials described in an NME's label or package insert. There is, however, no requirement that any trials supporting approved NMEs be published.
Objective: To determine the proportion of trials supporting successful new drug applications to the U.S. FDA that are published in the medical literature, and to identify predictors of publication.
Methods: Cohort study of all trials submitted to the FDA in support of all NMEs approved in 1998 to 2000. For each trial, we abstracted the intervention details, sample size, statistical significance of primary outcome results, and randomization and double-blinding status. Full publications of these trials were identified using a search of PubMed, the Cochrane Library, and CINAHL through August 1, 2006. Trials described in the "clinical studies" section of the FDA-approved label were classified as pivotal trials. We then evaluated trial characteristics that were associated with full publication in the medical literature. Our main outcome measures were full publication in the medical literature at 2 years, 5 years and final follow-up, and time from FDA approval to full publication.
Results: Between 1998 and 2000, 909 trials were submitted to the FDA supporting 90 approved NMEs. Of the 909 trials, only 383 (42%) were published in full by the final follow-up date (August 1, 2006). Thirty-four percent of the submitted trials (309/909) were pivotal trials, of which 73% (226/309) were published by the final follow-up date. In a multivariable logistic regression model evaluating publication at 5 years, trials were more likely to be published if they reported statistically significant results (odds ratio [OR], 3.6; 95% CI, 2.1-6.4); larger sample sizes (OR, 1.34 per 2-fold increase in sample size; 95% CI, 1.16-1.55), or were pivotal trials (OR, 4.6; 95% CI, 2.6-8.2). These characteristics were also significant predictors of publication in multivariable logistic regression models at 2 years and multivariable Cox proportional models for all trials, as well as in a sensitivity analysis of the pivotal trials only.
Conclusions: Over half of the supporting trials of approved NMEs remain unpublished at least 5 years after FDA approval. Pivotal trials and trials with statistically significant results and larger sample sizes are more likely to be published. Incomplete publication exists for clinical trials supporting the safety and efficacy of newly approved drugs in the United States. These findings support the need for mandating full results reporting to ensure an unbiased public evidence base.