BACKGROUND: There are limited data on the effect of hydroxychloroquine on medium term outcomes in patients with coronavirus disease 2019 (COVID-19) requiring intensive care. We aimed to evaluate the effects of hydroxychloroquine on day 90 mortality in this specific population. METHODS: This retrospective, multicenter, propensity matched cohort analysis, used data of adult patients with laboratory confirmed COVID-19 admitted to 3 university affiliated intensive care units between March 7, 2020, to April 7, 2020 in Lyon, France. Patients received either hydroxychloroquine (loading dose of 400 mg twice daily at day 1 followed by 200 mg twice daily from day 2 to day 10) or standard of care without hydroxychloroquine. We compared all-cause mortality at day-90 after ICU admission between propensity score matched groups receiving hydroxychloroquine or standard of care. RESULTS: A total of 157 patients were included with a day-28 and day-90 mortality rate of 23.6% and 32.5%, respectively. The median (interquartile) age was 67 years (56-76 years), 105 (66.9%) were men, 65 (41.4%) fulfilled criteria for acute respiratory distress syndrome, and 64 (41%) received hydroxychloroquine (HCQ) for 10 days (4-10 days). In the propensity score matched cohort (59 patients in each group), day-90 mortality was 35.6% for patients who received HCQ and 23.7% for patients who did not (P=0.23). Kaplan Meier survival analysis showed no statistically significant association between HCQ therapy and mortality (P=0.20 by log-rank test). CONCLUSIONS: In this study, off-label use of HCQ in critically ill patients with COVID-19 was not associated with any significant change in medium-term prognosis, confirming results of studies in less severe patients.