Using Machine Learning to Determine a Suitable Patient Population for Anakinra for the Treatment of COVID-19 Under the Emergency Use Authorization

Abstract

A randomized, double-blind, placebo-controlled study (SAVEMORE trial) provided data to support an Emergency Use Authorization (EUA) of anakinra in hospitalized adults with positive results of direct severe acute respiratory syndrome-coronavirus 2 viral testing with pneumonia requiring supplemental oxygen (low- or high-flow oxygen) who are at risk of progressing to severe respiratory failure and likely to have an elevated plasma soluble urokinase plasminogen activator receptor (suPAR). Currently, the suPAR assay is not commercially available in the United States. An alternative method was needed to identify patients that best reflect the population in the clinical trial selected based on suPAR level ≥ 6 ng/mL at baseline. A machine learning approach based on data from the SAVEMORE trial was used to develop a scoring rule to identify patients who are likely to have a suPAR level ≥ 6 ng/mL at baseline. External validation of the scoring rule was conducted with data from a different trial (SAVE). This clinical scoring rule with high positive predictive value, high specificity, reasonable sensitivity, and biological relevance is expected to identify patients who are likely to have an elevated suPAR level ≥ 6 ng/mL at baseline. As such, it is included in the EUA to identify patients that fall within the authorized population for whom the known and potential benefits outweigh the known and potential risks of anakinra.