Managing the inflammatory response to SARS-Cov-2 could prevent respiratory insufficiency. Cytokine profiles could identify cases at risk of severe disease. We designed a randomized phase II clinical trial to determine whether the combination of ruxolitinib (5 mg twice a day for 7 days followed by 10 mg BID for 7 days) plus simvastatin (40 mg once a day for 14 days), could reduce the incidence of respiratory insufficiency in COVID-19. 48 cytokines were correlated with clinical outcome. Patients admitted due to COVID-19 infection with mild disease. Up to 92 were included. Mean age was 64 ± 17, and 28 (30%) were female. 11 (22%) patients in the control arm and 6 (12%) in the experimental arm reached an OSCI grade of 5 or higher (p = 0.29). Unsupervised analysis of cytokines detected two clusters (CL-1 and CL-2). CL-1 presented a higher risk of clinical deterioration vs CL-2 (13 33% vs 2 6% cases, p = 0.009) and death (5 11% vs 0 cases, p = 0.059). Supervised Machine Learning (ML) analysis led to a model that predicted patient deterioration 48h before occurrence with a 85% accuracy. Ruxolitinib plus simvastatin did not impact the outcome of COVID-19. Cytokine profiling identified patients at risk of severe COVID-19 and predicted clinical deterioration. https://clinicaltrials.gov/, identifier NCT04348695.