Hydroxychloroquine Fails Hospitalized Adults with Covid-19: A Smart Analysis of the Orchid Trial

ABSTRACT

INTRODUCTION: Hydroxychloroquine (HCQ) was hoped to be repurposed for treating COVID-19, especially in regions where more expensive regimens were not available. Unfortunately, several randomized clinical trials (RCT) using HCQ to treat COVID-19, ranging from simple exposure to significant illness, have not improved outcomes. In five failed sepsis RCT's and one previous COVID-19 study the SMART statistical approach identified, from pre-randomization data, cohorts among whom study drugs reduced mortality significantly. Additionally, in one study the SMART statistical approach predicted futility. Whether such sub-groups exist in HCQ COVID-19 RCT's is unknown and is what this research aims to examine. OBJECTIVE(S) To determine from pre-randomization data if HCQ-responsive patients exist within the PETAL Clinical Trials Network ORCHID RCT in hospitalized COVID-19 patients. METHOD(S) With the approval of the BioLINCC data repository and the IRB of Inspira Health Network, HIPAAcompliant data on 452 ORCHID subjects (224 HCQ, 228 placebo) were analyzed. Pre-randomization stepwise logistic regression survival models were built separately for placebo and HCQ. Baseline data from all patients were then entered into both models. Interactions of placebo and HCQ survival models with 28-day mortality HCQ treatment effects determined optimum cutoffs, incrementally excluding from efficacy analysis patients predicted as HCQ non-responders. RESULT(S) Placebo model independent variables age, VITHRH and potassium. HCQ model independent variables age, SOFA score SpO2, and D-BL-ICU. Overall, 28-day mortality was 10.71% placebo and 10.53% HCQ (p= 0.9483). Interactions of the mortality models with HCQ efficacy were optimized in a cohort of 49 patients (10.2% of total RCT), with placebo 28-day mortality 46% and HCQ mortality 24% (p=0.1085). In the non-SMART group (n=403, 89% of total), mortality was 6.5% placebo and 8.9% HCQ (p=0.3726) CONCLUSION(S) Even in a pre-randomization optimized SMART-identified cohort comprising only 10.2% of the ORCHID RCT, hydroxychloroquine did not reduce mortality significantly in hospitalized patients with COVID-19. These results predict futility for further RCTs of hydroxychloroquine in COVID-19. Hydroxychloroquine fails hospitalized COVID-19 patients.