(B-282) Risk of upper gastrointestinal (GI) bleeding associated with initiation of oral systemic corticosteroids among patients with and without any pre-existing GI bleed

Background: Oral corticosteroids (OCS) are used to treat inflammation and autoimmune flares in many health conditions, including asthma. Upper gastrointestinal (GI) bleeding is a potential adverse effect of OCS. While OCS labeling advises caution with preexisting peptic ulcers (PPU), published evidence of any increased risk of OCS-associated upper gastrointestinal (GI) bleeding in patients with a history of PPU is limited, conflicting, and includes small numbers of patients. Rigorous pharmacoepidemiological investigation of this risk is problematic due to lack of direct treatment alternatives to OCS.

Objectives: To estimate the risk of upper GI bleeding after initiation of OCS among asthma patients with and without PPU using other drugs for insufficiently controlled asthma (long-acting beta agonists [LABAs] or leukotriene receptor antagonists [LTRAs]), not known to cause upper GI bleeding, as active comparators to minimize confounding including by factors associated with asthma control (e.g., stress).

Methods: We identified patients with asthma ≥65 years in US Medicare (2008-2019) continuously enrolled ≥365 days before newly initiating OCS or LABA/LTRA. Crude incidence rates (IR) of upper GI bleeding among OCS versus comparator drugs were stratified by PPU. Inverse probability of treatment weighted Cox regression using propensity scores estimated the risk of upper GI bleeding, overall and stratified by PPU, adjusting for sociodemographics, diagnostics, medications, and healthcare utilization.

Results: 73,267 patients with asthma initiated a study drug (mean age 73.8). Most were female, white, and had hypertension.
In patients without PPU (N=66,202), upper GI bleed occurred in 41 patients with IRs of 3.38/1,000 person-years (kpy) for OCS and 2.36/kpy for LABAs/LTRAs. In patients with PPU (N=7,048), upper GI bleed occurred in 12 patients with IRs of 14.4/kpy for OCS and 5.5/kpy for LABAs/LTRAs.
Higher risk of upper GI bleed was seen for OCS compared to LABA/LTRA [adjusted hazard ratio (aHR) 1.59 (95% CI 0.65-3.85]. Risk for upper GI bleed among those without PPU was similar between initiators of OCS and LABA/LTRA [aHR 1.06 (0.45-2.5)]. In patients with PPU, however, OCS initiators had a greater than 4-fold risk of upper GI bleeds [aHR 4.35 (0.47-50.0)]. However, confidence intervals (CIs) were wide and crossing 1.0 in all analyses.

Conclusions: Interpretation of these findings is difficult due to few events and wide CIs. However, if replicated, these results may suggest that risk of OCS-associated upper GI bleeding (when used relatively short term as in asthma treatment) is more likely in patients with PPU.