Background: Several studies have examined the reporting of subgroup analyses (SAs) in randomized trials. All have, however, focused on the highest impact or specialty medical journals. Objectives: To investigate, in randomized trials published in high impact (i.e. NEJM, JAMA, Lancet, BMJ, and Annals of Internal Medicine) and lower impact journals, to what extent investigators report SAs, and what study characteristics are associated with reporting of SAs. Here, we report our findings of a pilot study. Methods: We included randomized trials published in 108 Core Clinical Journals defined by the National Library of Medicine that cover all clinical and public health fields. We searched Medline for 2007. We randomly selected 120 citations from 3662 hits. Teams of two reviewers, trained in health research methodology independently abstracted the data. Results: After title and abstract screening, full texts of 60 reports were further screened, and 51 reports including 52 trials were finally included, 9 published in high impact journals, 43 in lower impact journals. Table 1 shows the reporting of SAs in those trials. Trials reporting SAs have larger average sample size (median 351, range 38-1493) than those not reporting SAs (75, 12-12525, p=0.0013). Trials in high impact journals report SAs more frequently than lower impact journals (78% vs. 23%, p=0.003). Trials with funding from private for-profit organizations trend toward reporting SAs more frequently (44% vs. 26%, p=0.19). Conclusions: The pilot study suggests that reporting of SAs might differ in trials between high and lower impact journals; the difference may in part be driven by larger sample size in high impact journal publications.