Estimating Real-world Outcomes Under a Rapidly Evolving Treatment Paradigm: Dexamethasone and Inpatient Mortality Among US Hospitalized COVID-19 Patients

Friday, August 26, 2022

5:30 PM – 5:45 PM CEST

Location: Congress Hall D4

Publication Number: 79

Background: In June 2020, the UK RECOVERY trial reported that dexamethasone (DEX) reduced 28-day mortality in hospitalized COVID-19 patients on invasive mechanical ventilation (IMV) by 36% with a smaller effect in those on oxygen (O2; 18%). Subsequently, inpatient DEX use increased in the United States (US).

Objectives: Determine the 28-day mortality risk for DEX initiation versus no treatment with systemic corticosteroids (CS) among US hospitalized COVID-19 patients on O2 and IMV to develop processes and evaluate methods for rapid cycle real-world treatment effectiveness evidence generation.

Methods: We designed an observational cohort study to emulate a hypothetical target trial of COVID-19 inpatients randomized to DEX or non-CS standard of care. Using HealthVerity chargemaster and claims data (April 2020 - March 2021), COVID-19 patients initiating DEX within 21 days following admission were matched to patients who had not yet initiated CS based on date, age, sex, comorbidity score, days since admission, and COVID-19 severity (no O2, O2, IMV) at DEX initiation (index). Patients were propensity score (PS) matched 1:1. They were followed from the day after the treatment index date until death (discharge status of expired), discharge, or a 28-day maximum to determine an initial-treatment hazard ratio (IT HR), or start of CS among non-CS for an as-treated HR (AT HR). The AT HR suggested informative censoring; thus, a post hoc IMV subgroup inverse probability of censoring weighting (IPCW) analysis was added to determine if adjustment impacted the AT HR.

Results: Of the 804 IMV PS matched pairs, the DEX versus non-CS IT and AT HRs were 1.02 (95% CI 0.89-1.18) and 1.16 (95% CI 0.91-1.48), respectively; 67% of non-CS patients initiated CS within 7 days after index (48% within 1 day). The AT HR was slightly attenuated with IPCW ranging from 1.09 (95% CI 0.84-1.42) for the full model to 1.02 (95% CI 0.79-1.32) for a model with only data driven covariates. Of the 8008 O2 pairs, IT and AT HRs were 1.32 (95% CI 1.21-1.45) and 1.76 (95% CI 1.50-2.07); 63% of non-CS patients initiated CS within 7 days after index (47% within 1 day).

Conclusions: While post hoc analyses show that the extent of differential censoring may have obscured an underlying protective effect, the results differed from the RECOVERY trial. Substantial crossover of the referent patients during follow-up was not anticipated. When routine care is rapidly evolving, drug utilization during follow-up should be monitored (keeping treatment-specific outcomes blinded) to increase interpretability of the study findings.

Presenting Author(s)

Elizabeth M. Garry, MPH, PhD

Head of Scientific Research
Aetion, Inc
Boston, Massachusetts, United States

Slides

Co-Author(s)

Imaani Easthausen

Aetion, Inc
New York, New York, United States
Sarah Vititoe, MPH

Aetion, Inc
Brooklyn, New York, United States
AW

Andrew Weckstein

Aetion, Inc
New York, New York, United States
AC

Aloka Chakravarty, PhD, MStat

Senior Statistical Advisor
US Food and Drug Administration, United States
Tamar Lasky, PhD

Senior Advisor on RWE
US Food and Drug Administration
Silver Spring, MD, United States
Marie C. Bradley, PhD, MPharm, MPH

Senior Pharmacoepidemiologist
Office of Surveillance and Epidemiology, CDER, FDA
Silver Spring, Maryland, United States
Silvia Perez-Vilar, PhD

Senior Epidemiologist
US Food and Drug Administration
Silver Spring, Maryland, United States
Donna R. Rivera, PharmD, MSc, FISPE

Associate Director of Pharmacoepidemiology
US Food and Drug Administration
Silver Spring, Maryland, United States
KQ

Kenneth Quinto, MD, MPH

Senior Medical Advisor for Real World Evidence Analytics
US Food and Drug Administration
Silver Spring, Maryland, United States
DZ

Di Zhang, PhD

US Food and Drug Addministration, MD, United States
YZ

Yueqin Zhao, PhD

US Food and Drug Administration, United States
AR

Anil Rajpal, MD, MPH

US Food and Drug Administration, United States
AB

Aidan Baglivo, BS

Aetion, Inc, United States
Jeremy A. Rassen, DSc, FISPE

Aetion, Inc
Sharon, Connecticut, United States
SL

Sandy Leonard, MPH

SVP, Partnerships and Real World Data
HealthVerity
Philadelphia, PA, United States
Nicolle Gatto, PhD, FISPE

Chief Science Officer
Aetion, Inc., United States