(117) Outcomes in patients hospitalized with severe COVID-19 treated with add-on immunomodulatory agents

Background: Randomized trials have demonstrated survival benefits based on addition of either an interleukin-6 receptor antagonist (IL-6RA) or a Janus Kinase inhibitor (JAKi) to corticosteroids (CS) compared to standard of care, for persons with severe COVID-19. However, little real-world evidence has been generated on this topic.

Objectives: As part of a methods-focused analysis demonstrating an adaptive approach to design and analysis, we emulated a test-case target trial evaluating the risk of inpatient death (primary outcome) and progression to invasive mechanical ventilation or extracorporeal membrane oxygenation (secondary outcome) up to 28 days after tocilizumab (TCZ, an IL-6RA) or baricitinib (BAR, a JAKi) initiation among hospitalized patients with severe COVID-19 receiving systemic CS..

Methods: Adults hospitalized with severe COVID-19 (ICD-10 U07.1 and receipt of supplemental oxygen/non-invasive ventilation/high flow oxygen) in the HealthVerity Hospital Chargemaster and Claims database from June 16th 2020 - February 1st 2022 were included. An active comparator, new-user design emulated treatment strategies for either TCZ (single IV dose) or BAR (once-daily oral dose) within 4 days of hospital admission. Follow-up began 1 day after TCZ/BAR initiation with censoring at the earliest of discharge or death (for the outcome of death or competing risk of discharge, respectively), treatment discontinuation, or end of the 28-day follow-up period (with pre-specified contingencies to evaluate shorter follow-up periods). HRs were estimated from a marginal structural Cox model weighted by the product of inverse probability of treatment weights (time-fixed, to control baseline confounding) and censoring weights (time-varying, to account for informative censoring), after truncation.

Results: No difference was found in the risk of 28-day mortality associated with TCZ versus BAR (HR 1.01, 95% CI: 0.75, 1.37). When follow-up was truncated, shorter follow-up after TCZ initiation was associated with an increasing risk of death (day 21: HR 1.45, 95% CI: 1.03, 2.04; day 14: HR 1.84, 95% CI: 1.25, 2.70; day 7: HR 2.02, 95% CI:1.20, 3.40). Similar trends for 7- and 14-day follow up periods were observed for the secondary composite outcome.

Conclusions: Results suggest no increased risk of mortality, progression to IMV/ECMO or mortality associated with TCZ at 28 days. However, we observed an increased risk of these events with TCZ in the early period after initiation. Similar time-varying trends have been observed in at least one other observational study (Sunny 2023). This finding of time varying-trends could be due to actual differences in drug pharmacology or bias (e.g., residual confounding due to unmeasured differences in COVID-19 severity between TCZ and BAR users, depletion of susceptibles). Further evaluation is warranted.