An observational study analyzed 239 clinical trials submitted to The BMJ between 2019-2023 that were not prospectively registered in an ICMJE-accepted registry, compared with 239 randomly selected prospectively registered studies. Studies with larger samples (101-500 vs. 1-100; odds ratio 0.24, 95% CI 0.09-0.67), from Oceania (vs. Europe; 0.35, 0.14-0.82), with more authors (10 vs. 5; 0.71, 0.59-0.87), mentions of CONSORT standards (0.22, 0.06-0.67), more recent contributions (2019-2022; 0.42-0.95), and funding (non-profit vs. none; 0.20, 0.09-0.41); studies with a corresponding author from Asia had higher odds (vs. Europe; 1.75, 1.07-2.89). Of the 176 non-prospectively registered studies from 2019-2021, 146 (83%) were registered retrospectively (median delay 193 days), 23 (13%) were not registered and 7 (4%) were in the non-accepted registry. The majority (155, 88%) were published, 138 (89%) in journals with an impact factor (median 5.39) and 96 (62%) in journals claiming adherence to ICMJE recommendations; the median time from BMJ submission to publication was 12 months. Only about a sixth (out of 155) explicitly revealed registration deficiencies during publication. Of the 72 authors who answered the question about prospective registration, 60 (83%) incorrectly claimed compliance. Many of the rejected BMJ studies were published in high-impact journals without publication