Randomized controlled trials in neurosurgery: an observational analysis of trial discontinuation and publication outcome.

OBJECTIVE This study aimed to determine the trial discontinuation and publication rate of randomized controlled trials (RCTs) in neurosurgery. METHODS Trials registered from 2000 to 2012 were identified on the website clinicaltrials.gov using a range of key words related to neurosurgery. Any trials that were actively recruiting or had unknown status were excluded. Included trials were assessed for whether they were discontinued early on the clinicaltrials.gov database; this included trials identified as withdrawn, suspended, or terminated in the database. For included trials, a range of parameters was identified including the subspecialty, primary country, study start date, type of intervention, number of centers, and funding status. Subsequently, a systematic search for published peer-reviewed articles was undertaken. For trials that were discontinued early or were found to be unpublished, principal investigators were sent a querying email. RESULTS Sixty-four neurosurgical trials fulfilled our inclusion criteria. Of these 64, 26.6% were discontinued early, with slow or insufficient recruitment cited as the major reason (57%). Of the 47 completed trials, 14 (30%) remained unpublished. Discontinued trials showed a statistically significant higher chance of remaining unpublished (88%) compared with completed trials (p = 0.0002). Industry-funded trials had a higher discontinuation rate (31%) compared with non-industry-funded trials (23%), but this result did not reach significance (p = 0.57). Reporting of primary outcome measures was complete in 20 (61%) of 33 trials. For secondary outcome measures, complete reporting occurred in only 11 (33.3%) of 33. CONCLUSIONS More than a fifth (26.6%) of neurosurgical RCTs are discontinued early and almost a third of those that are completed remain unpublished. This result highlights significant waste of financial resources and clinical data.