(103) Impact of chronic oral glucocorticoid treatment on mortality in patients with COVID-19 - An analysis from a population-based cohort within the SCIFI-PEARL project

Publication Number: 440

Background: Oral glucocorticoids (GC) are essential in managing inflammatory diseases and may prevent acute respiratory distress syndrome in COVID-19 patients. Prolonged GC use may, however, increase thromboembolic events and mortality. Nevertheless, research on cause-specific death in GC-treated patients with COVID-19, where thromboembolic complications and sepsis are common, is limited.

Objectives: To investigate overall death in COVID-19 patients who have long-term exposure to GC, as well as death from pulmonary embolism (PE), myocardial infarction, stroke, sepsis, and COVID-19.

Methods: Using nationwide register data from the SCIFI-PEARL database, we studied all patients infected with COVID-19 in Sweden from Jan 2020 to Nov 2021 (N=1200153) and followed them from COVID-19 infection date (index) until death or end of study. Long-term oral GC use was defined as A) High exposure: ≥2 prescriptions of GC with total dose >750mg of prednisolone or equivalent amount of other oral GCs during the 6 months before index, and B) Any exposure: ≥1 GC prescription during the 12 months before index. GC users were compared to non-exposed patients that had not received GC treatment within 1 year before index. We estimated hazard ratios (HRs) and 95% confidence intervals (CI) for the average treatment effect for everyone (ATE) and average treatment for the treated (ATT) for GC exposure on the outcomes using: 1. Adjusted Cox proportional hazard models (ATE); 2. Cox models with 1:2 matching on a propensity score (ATT). A competing risk regression was also performed on each outcome, considering other causes of death. The same broad set of confounders was used in all analyses.

Results: 3378 deaths occurred in the GC exposure group (N=48806; 6.9%) compared to 14850 deaths in the non-exposed group (1151347; 1.3%). Both high and any exposure to GC were associated with overall death; HR 1.98 (95%CI 1.87-2.09) and 1.58 (1.52-1.65) respectively. In the ATE analysis, deaths from sepsis, PE and COVID-19 were associated with high GC exposure [2.90 (2.37-3.54), 1.37 (1.09-1.74) and 1.96 (1.85-2.08)] and similarly but slightly weaker with any exposure, while death in stroke was only associated with high GC exposure [1.33 (1.05-1.67)]. The ATT analysis showed consistent results as the ATE. Competing risk models showed stronger effects for death from sepsis than non-sepsis causes.

Conclusions: COVID-19 patients on chronic GC treatment face increased mortality, particularly from sepsis, pulmonary embolism and COVID-19. Close monitoring may prevent pulmonary embolism death in instances where the effect of GC is of interest. Clinicians should weigh GC benefits and monitor high-risk patients closely.