Patients with acute hypoxemic respiratory failure (AHRF) due to COVID-19 undergoing High Flow Therapy (HFT) before intubation presented an increased risk of mortality when intubation was delayed. We designed a prospective study seeking biomarkers for early prediction of HFT failure. An analytical value in blood could be more stable than the ROX index since it will not depend on the vital signs that the patient presents at that moment. We defined HFT failure as the need to scale the treatment to Non-Invasive Positive Pressure Ventilation (NPPV). The needs were respiratory rate >25, oxygen saturation of <90% despite being on flow of 60 l·min-1 and FiO2 1 or levels of PaO2/FiO2 ratio <100mmHg. The all-treatment population included all subjects enrolled in the trial. 139 patients were enrolled after starting HFT. The Pearson chi-squared test was used to compare the main study outcomes. These included the incidence of intubation, the cumulative incidence of mortality at 30 days, the cumulative incidence of mortality at 1 year, and the composite outcome of intubation or death at the end of the trial. Kaplan-Meyer plot was used to illustrate the time to HFT failure. The Cox regression model was used to estimate the hazard ratio for HFT failure for all the parameters. All were measured or collected at baseline. Lower levels of bicarbonate, thrombocytopenia, and higher levels of C-reactive protein (CRP), lactate dehydrogenase (LDH), creatinine, and glucose are early blood biomarkers independently associated with HFT failure.SUMMARY AT A GLANCEHigher levels of C-reactive protein and lactate dehydrogenase in patients with COVID-19 pneumonia allow us to early detect patients requiring intubation with an apparent good response to high-flow oxygen therapy.