Patients receiving direct oral anticoagulants (DOACs) who undergo high‑risk endoscopic procedures face competing peri‑procedural hazards: bleeding if anticoagulation is continued or resumed too early, and thromboembolism if anticoagulation is interrupted. The study aimed to quantify rates of peri‑procedural adverse events (AEs) and the associated incremental healthcare costs within 90 days of high‑risk endoscopy among anticoagulated patients.
The investigators conducted a retrospective cohort study using the TriNetX Linked Claims database covering 2015–2025, a US multi‑payer claims network. The analytic sample comprised 2,486 unique adults on DOAC or warfarin therapy who underwent a high‑risk endoscopic procedure. Linked cost data were available for 1,933 patients and used for the cost analyses.
Peri‑procedural bleeding and thromboembolic events (stroke or transient ischemic attack [TIA]) were identified using principal‑diagnosis codes recorded during acute care encounters (emergency department or inpatient). Events occurring within 90 days of the index procedure were included, with additional sub‑timepoint reporting at 7 days.
Unadjusted and multivariable‑adjusted incremental 90‑day costs were estimated using generalized linear models. The models compared patients with and without each AE type. Costs were derived from the linked claims and reflect the incremental healthcare spending associated with each event type within 90 days of the index procedure.
Among the 2,486 anticoagulated patients, median age was 65 years, 61.5% were male, and 81% had atrial fibrillation. The source reports these baseline descriptors for the overall cohort; additional patient‑level detail was not provided in the abstract.
Bleeding events were substantially more frequent than thromboembolic events in the peri‑procedural window. Bleeding occurred in 18.0% of patients by 7 days and in 22.9% by 90 days after high‑risk endoscopy. Stroke or TIA occurred in 0.5% by 7 days and in 1.7% by 90 days.
The study reports marked differences in the incremental costs associated with each AE type. On an unadjusted basis, a peri‑procedural stroke/TIA added $8,874 in 90‑day costs (p<0.001). After multivariable adjustment, stroke/TIA was associated with an adjusted incremental cost of $5,798 and a cost ratio of 1.56 (95% CI 1.11–2.20; p=0.011).
For peri‑procedural bleeding, the unadjusted incremental 90‑day cost was $3,471 and the adjusted incremental cost was $3,157 (cost ratio 1.33; 95% CI 1.18–1.50; p<0.001).
Patients who experienced both bleeding and a subsequent stroke/TIA incurred the highest mean 90‑day costs reported in the study: $23,716.
Although thromboembolic events were much less frequent than bleeding after high‑risk endoscopy in this anticoagulated population, each thromboembolic event imposed substantially higher incremental healthcare costs—approximately 2.5‑fold greater per event than bleeding in this dataset. The authors emphasize that these economic differences underscore the clinical and systems‑level importance of peri‑procedural DOAC management and support the need for randomized trials to define optimal timing for anticoagulant resumption after high‑risk endoscopy.
This work is a retrospective claims analysis and has the inherent limitations of observational claims research, including reliance on principal‑diagnosis codes to identify events and potential unmeasured confounding. The report is a preprint and has not undergone peer review. The authors note that individual patient‑level data cannot be publicly shared because of the data use agreement with TriNetX; however, the study was pre‑registered on the Open Science Framework and analytic code is publicly available at that registry.
In a US multi‑payer claims cohort of anticoagulated patients undergoing high‑risk endoscopy, peri‑procedural bleeding was common whereas stroke/TIA was uncommon. On a per‑event basis, thromboembolic events resulted in substantially higher 90‑day incremental healthcare costs than bleeding. The findings highlight the economic as well as clinical trade‑offs involved in peri‑procedural anticoagulant interruption and motivate randomized evidence on optimal DOAC resumption strategies.