2.4 Study variables
We considered the types of NSAID, high-dose NSAID use, concomitant
medications, demographics, and baseline comorbidities as candidate risk
factors. Baseline comorbidities and past medical history were confirmed
within 6 months of NSAID initiation (see Table S1 for the diagnosis
code). NSAID types were classified as selective COX-2 inhibitors or
nsNSAIDs. High-dose NSAID use was
defined as the total daily administration of NSAIDs, including the use
of multiple NSAIDs, in excess of the daily recommended dose based on the
ATC/DDD system from the World Health Organization collaborating center
[10]. Medications known to affect GI complications [11] were
identified by reviewing prescriptions within 30 days prior to the index
date: GPA, corticosteroids, selective serotonin reuptake inhibitors
(SSRI), HMG-CoA reductase inhibitors, calcium channel blockers, and oral
anticoagulants, including vitamin K antagonists and direct oral
anticoagulants, as well as antiplatelet agents, including ASA,
P2Y12 inhibitors, and PDE3 inhibitors. GPAs include
proton pump inhibitors (PPIs), misoprostol, H2-receptor antagonists
(H2RAs), rebamipide, and Artemisia herb soft extract [12].
Concomitant medications were considered for concomitant use with NSAIDs
when used for at least 80% overlap with NSAID treatment [6].
The customized HIRA research data used as the derivative cohort included
detailed age, whereas the HIRA sample data (HIRA-APS 2019) for external
validation included age information as age groups in 5-year increments,
and 80+ was provided as one age group.