Objective To describe study features and publication characteristics of COVID-19–related clinical trial preprints.

Design This was a cross-sectional analysis of COVID-19– related clinical trials published as preprints. Preprints were included if they were uploaded between January 1 and December 31, 2020, to any open-access preprint server indexed by the National Institutes of Health iSearch COVID-19 portfolio and described results from a clinical trial assessing an intervention related to the treatment or prevention of COVID-19. A single investigator assessed manuscripts for eligibility by reviewing the titles and abstracts for all preprints within the iSearch registry, followed by a full-text review of potentially eligible preprints. MEDLINE, Google Scholar, and Embase were then searched to identify peer-reviewed publications matching the included preprints. Two investigators, including a medical librarian, independently searched for published articles at least 14 months after the initial preprint posting date. Two investigators abstracted information from each eligible preprint, including basic study characteristics (location, funding, size, intervention type), features affecting risk of bias (prospective registration, allocation method, blinding), and trial outcome. Descriptive statistics are reported for these study features, and Cox proportional hazards regression was used to determine associations between key study characteristics and peer-reviewed publication status.

Results A total of 22,615 preprints were screened for eligibility and 145 met inclusion criteria. These included preprints uploaded to 5 different preprint servers (medRxiv, Research Square, SSRN, preprints.org, and bioRxiv). Funding sources included government for 65 (45%), industry for 32 (22%), a university or health system for 28 (19%), and a foundation for 16 (11%). The most commonly assessed interventions were drugs (86 59%), convalescent plasma (19 13%), and vaccines (16 11%). Median enrollment was 78 participants (IQR, 30-174 participants). Most trials were registered (136 94%), although less than half were registered prospectively (71 49%). Blinding was reported for participants in 35 trials (24%), investigators and research staff in 27 (19%), and outcome assessors in 34 (23%). Among 106 trials with more than 1 treatment group, 89 (84%) randomly allocated participants between treatments. Matching peer-reviewed publications were identified for 118 of 145 preprints (81%) in 77 different journals. The median time elapsed between preprint posting and peer-reviewed publication was 115 days (IQR, 55-201 days). Cox proportional hazards analysis revealed significant associations with peer-reviewed publication for government funding and study size (Table 67).

Conclusions Most COVID-19 clinical trial preprints had undergone peer-reviewed publication, although delays of several months were common. A high proportion were not prospectively registered. Significant associations were not observed between peer-reviewed publication and study characteristics related to risk of bias.

Conflict of Interest Disclosures Timothy F. Platts-Mills is an employee of Ophirex. Christopher W. Jones reported receiving grants from AstraZeneca, Abbott, Vapotherm, and Ophirex.

Funding/Support This work was supported by US Department of Health and Human Services Office of Research Integrity grant ORIIR200066-01-00.

Role of the Funder/Sponsor The sponsor had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the abstract for presentation.

Disclaimer The contents of this abstract are those of the authors and do not represent the official views of or an endorsement by the Office of the Assistant Secretary for Health, US Department of Health and Human Services, or the US government.