Aims/hypothesis Clinical trials assessing interventions for treating and preventing diabetes mellitus and its complications are needed to inform evidence-based practice. To examine whether current studies adequately address these needs, we conducted a descriptive analysis of diabetesrelated trials registered with ClinicalTrials.gov from 2007 to 2010. Methods From a dataset including 96,346 studies registered in ClinicalTrials.gov downloaded on 27 September, 2010, a subset of 2,484 interventional trials was created by selecting trials with disease condition terms relevant to diabetes. Results Of the diabetes-related trials, 74.8% had a primarily therapeutic purpose while 10% were preventive. Listed interventions included drugs (63.1%) and behavioural (11.7%). Most trials were designed to enrol ≤500 (91.1%) or ≤100 (58.6%) participants, with mean/median times to completion of 1.8/1.4 years. Small percentages of trials targeted persons aged ≤18 years (3.7%) or ≥65 years (0.6%), while 30.8% excluded patients >65 years and the majority excluded those >75 years. Funding sources included industry (50.9%), NIH (7.5%) or other, with most being single-centre trials of other sponsorship (37.7%) or industryfunded multicentre studies (27.4%). A small number of trials (1.4%) listed primary outcomes including mortality or clinically significant cardiovascular complications. The distribution of trials by global region and US state does not correlate with prevalence of diabetes. Conclusions/interpretation The majority of diabetes-related trials include small numbers of participants, exclude those at the extremes of age, are of short duration, involve drug therapy rather than preventive or non-drug interventions and do not focus upon significant cardiovascular outcomes. Recently registered diabetes trials may not sufficiently address important diabetes care issues or involve affected populations.