Lead Investigator: Gus Slotman, Inspira Health Network
Title of Proposal Research: IDENTIFYING TOCILIZUMAB-RESPONSIVE COVID-19 PATIENTS FROM COVACTA, EMPACTA, AND REMDACTA RANDOMIZED CLINICAL TRIAL PRE-RANDOMIZATION DATA
Vivli Data Request: 8342, 7444
Funding Source: None
Potential Conflicts of Interest: None
Summary of the Proposed Research:
Discovering effective treatments of COVID-19 infection is vital. In the EMPACTA randomized clinical trial (RCT) (NCT04372186) and the COVACTA RCT (NCT04320615) and the REMDACTA (NCT04409262) RCTs tocilizumab, a monoclonal antibody against the interleukin-6 receptor, did not improve survival in hospitalized COVID-19 patients with pneumonia. These RCT failures may indicate that tocilizumab was not effective. However, alternatively non-specific clinical RCT entry criteria may have enrolled so many patients not responsive to tocilizumab’s IL-6 inhibition that its true treatment effects were blunted.
Hypothesis: The clinical definitions used as COVACTA, EMPACTA, and REMDACTA RCT entry criteria may have enrolled so many patients whose pathophysiology could not benefit from tocizumab that its true treatment effects were diluted to invisibility. Previously the SMART prognostic methodology identified 55% of the COVACTA tocilizumab RCT in COVID-19 among whom tocilizumab reduced mortality by over 14%, while in the other COVACTA patients, with placebo mortality 1/3 of that in the responder group, tocilizumab increased mortality by 16%. The present study proposes to determine whether or not similar cohorts exist in the REMDACTA and EMPACTA RCT’s and if the SMART predictive models are similar in these studies. If this investigation identifies such patients, then the predictive models generated and validated in the combined databases of the 3 RCTs can be used to match tocilizumab-responsive COVID-19 pneumonia patients with this life-saving treatment
Statistical Analysis Plan:
All pre-randomization data on REMDACTA RCT patients with complete data sets will be screened as possible prognostic independent variables using the statistical modeling described in our previous publication in sepsis RCT’s. Using multivariate, stepwise logistic regression with all ways elimination (simultaneous forward and backward elimination of non-weighted independent variables), survival models will be developed separately for the placebo and active drug groups. After the modeling process is completed, pre-randomization data from every patient in that study will be entered into both equations, and lengthy explorations into the relationship between the placebo and active drug models and their interactions with treatment effects will be undertaken to determine optimum cutoffs for tocilizumab. Beginning with the original clinical definition patient population, this process will test tocilizumab treatment effects in progressively smaller sub-populations, incrementally excluding, always at pre-randomization baseline, from efficacy analysis patients predicted to be unable to respond to or benefit from tocilizumab.
Requested Studies:
A Randomized, Double-Blind, Placebo-Controlled, Multicenter Study to Evaluate the Safety and Efficacy of Tocilizumab in Patients with Severe COVID-19 Pneumonia
Data Contributor: Roche
Study ID: NCT04320615
Sponsor ID: WA42380
(Note: Additional studies added as part of Data Request 8342)
A Phase III, Randomized, Double-Blind, Multicenter Study to Evaluate the Efficacy and Safety of Remdesivir Plus Tocilizumab Compared With Remdesivir Plus Placebo in Hospitalized Patients With Severe COVID-19 Pneumonia
Data Contributor: Roche
Study ID: NCT04409262
Sponsor ID: WA42511
A Randomized, Double-Blind, Placebo-Controlled, Multicenter Study to Evaluate the Efficacy and Safety of Tocilizumab in Hospitalized Patients With COVID-19 Pneumonia
Data Contributor: Roche
Study ID: NCT04372186
Sponsor ID: ML42528
Public Disclosures:
Kelter, D. and Slotman, G.J., 2022. Smart Identifies Dichotomous Drotrecogin Alfa Activated Treatment Effects on Mortality in Prowess Shock. Journal of the American College of Surgeons, 235(5), p.S296. Doi : 10.1097/01.XCS.0000895332.83940.b8