Background: Drug-drug interactions (DDIs) underlie 15-20% of adverse-drug reactions requiring hospitalisation and have caused several market withdrawals due to related adverse events (AEs). More prevalent polypharmacy in ageing populations and increasing numbers of chemical entities emphasise the need for clinical studies of DDIs.

Objectives: To review the characteristics of clinical trials on DDIs and assess their registered and published safety data in a publicly available trial register.

Methods: We performed an observational study of clinical trials retrieved from ClinicalTrials.gov by using the search term 'drug-drug interaction' (search performed on 16 October 2015). Trials were included if they were: 1) investigating the DDIs; 2) had a ClinicalTrials.gov registration number; 3) closed and completed in October 2015; 4) registered between 23 June 2005 and 16 October 2015. Publications were identified through CT.gov, PubMed and SCOPUS. Data on 8 items from the World Health Organization minimum dataset on AEs were abstracted by one author and verified by another.

Results: Among the 2059 retrieved clinical trials, 762 were excluded because of the incorrect classification as related to DDIs. Most were industry-sponsored (65%), started before registration (57%), and were primarily interventional studies (96%) in phase I (72%). Only a few studies had registered results (13%), among which 23% registered occurrence of serious and other AEs. 71 trials (5%) had both registered and published data (Table 1). Reported SAE and OAE description, frequencies and/or absolute numbers were identical to registered data for only 17 trials (24%).

Conclusions: We found a remarkably low rate of reporting of study results and AEs, as well as high discrepancy between registered and published AEs. Immediate efforts of all stakeholders to improve transparency are needed, as well as more stringent regulatory requirements for trial registration and drug marketing authorisation for DDIs.