Patients hospitalised for COVID-19 are at risk for multiorgan failure and death. Sodium-glucose co-transporter-2 (SGLT2) inhibitors provide cardiovascular and kidney protection in patients with cardiometabolic conditions and could provide organ protection during COVID-19. We aimed to investigate whether SGLT2 inhibitors can reduce the need for organ support in patients hospitalised for COVID-19. This pragmatic, multicentre, open-label, randomised, controlled, platform trial was conducted across 63 sites in the USA, Spain, Brazil, Italy, and Mexico. Patients aged at least 18 years hospitalised for COVID-19 (moderate or severe illness) were randomly assigned (1:1), via an interactive voice system or web-response system, to receive locally available SGLT2 inhibitor (administered orally, once daily) plus standard-of-care or standard-of-care for 30 days. The primary outcome was organ support-free days evaluated through 21 days, assessed using intention-to-treat approach. This trial is registered on ClinicalTrials.gov, NCT04505774. The first patient was randomly assigned to the SGLT2 inhibitor domain on Dec 3, 2021. On March 31, 2023, at the recommendation of the data and safety monitoring board, enrolment in the SGLT2 inhibitor domain for both moderately and severely ill hospitalised patients was stopped prematurely for futility due to a low likelihood of finding a treatment benefit. The final randomised population consisted of 575 patients (mean age 72 years [SD 13], 242 (42%) female and 154 (27%) Hispanic; 504 in the moderate illness group and 71 in the severe illness group). 573 patients had a known 21-day outcome; 215 (75%) of 285 patients in the SGLT2 inhibitor plus standard-of-care group did not require respiratory or cardiovascular organ support versus 231 (80%) of 288 patients in the standard-of-care group. The adjusted odds ratio (OR) for an SGLT2 inhibitor effect on organ support-free days was 0·74 (95% Credible Interval [CrI] 0·48-1·13; where OR higher than 1 indicated treatment benefit, yielding a posterior probability of futility P(OR <1·2) of 99% and a posterior probability of inferiority P(OR<1·0) of 91%). There were 37 deaths (13%) in the SGLT2 inhibitor plus standard-of-care group and 42 deaths (15%) in the standard-of-care group at 90 days (hazard ratio 0·91 [95% CrI 0·58-1·43], probability of hazard ratio <1 of 66%). No safety concerns were observed with SGLT2 inhibitors, including no cases of ketoacidosis. SGLT2 inhibitors did not significantly increase days free of organ support or reduce mortality in patients hospitalised with COVID-19. SGLT2 inhibitors were well tolerated with no observed safety concerns. Overall, these findings do not support the use of SGLT2 inhibitors as standard care in patients hospitalised with COVID-19. National Institutes of Health.