Warfarin and GLP-1 Receptor Agonist Interaction Effects on Time in Therapeutic Range

Background

Warfarin is a commonly prescribed anticoagulant with many dietary and medication interactions. Warfarin's therapeutic efficacy is monitored using the International Normalized Ratio (INR), a standardized measure of prothrombin time. The time in therapeutic range (TTR) is based on INR data and maximizing the TTR provides maximum benefit for preventing stroke, major hemorrhage, and death.

Glucagon-like peptide-1 receptor agonists (GLP1-RA), such as semaglutide and exenatide, are a newer class of medications commonly employed in the management of type 2 diabetes and, increasingly, to promote weight loss. GLP1-RA's enhance glucose-dependent insulin secretion, suppress glucagon release, increase satiety, and slow gastric emptying. Given delayed gastric emptying, patients taking GLP1-RA may have altered absorption of the medication or changed intake of vitamin k rich foods and therefore could have altered international normalized ratio (INR), which could adversely affect coagulation in this patient population.

On the other hand, no known pharmacokinetic or pharmacodynamic interactions have been reported between warfarin and GLP1-RA. The potential for secondary interactions, however, cannot be ruled out, especially due to the gastrointestinal side effects of GLP1-RAs. Given the increasing prevalence of type 2 diabetes and the common co-occurrence of cardiovascular comorbidities requiring anticoagulation, understanding the potential interactions between GLP-1 receptor agonists and warfarin is of paramount importance.

Objective: To determine the impact of initiation of GLP-1 therapy on INR control in patients receiving warfarin.

METHODS

This was a single center retrospective cohort study reviewed patients concurrently prescribed warfarin and GLP1-RA between 1/1/2008 to 12/1/2023. Patients were excluded if they were not taking medications concurrently or discontinued either medication in the 90 days after initiation. Furthermore, patients with fewer than four INR measurements within 90 days before and 90 days after GLP-1 agonist initiation. When both conventional venous and POCT values were available, the conventional test. The primary outcome measure was TTR values were calculated in the 90 days pre and post GLP1-RA initiation. The percentage of INR in range and time above and below were secondary outcome measures. Statistical analysis was performed in Microsoft Excel, with two-tailed t-tests comparing the pre- and post-GLP-1 initiation periods.

Results

Of 81 patients screened, 28 patients met the above criteria to be included in the analysis. Thirteen patients (46%) were male. The average age of participants was 59.4 (minimum 38, maximum 80). Most (9, 32%) of the patients were on warfarin for mechanical valve, followed by afib/aflutter (7, 25%), venous thromboembolism (7, 25%), and other (5, 18%). At baseline, 22 (79%) of patients had a diagnosis of diabetes. Semaglutide was the most prescribed GLP1-RA (10, 36%), followed by exenatide (8, 28.5%), dulaglutide (5, 17.5%), liraglutide (4, 14%), and tirzepatide (1, 3.5%). Average TTR was 59% and 60% for the 90 days before and after GLP1-RA initiation, respectively (p=0.97). The percentage of in-range INRs were 56.6% and 51.9% for the pre-and post-initiation groups, respectively (p=0.51). No patients had bleeding post initiation of GLP1-RA. Among the sub-group of patients with baseline TTR greater or equal to 80%, post-GLP1-RA TTR decreased to 70%.

Discussion/Conclusion

Initiation of GLP1-RA therapy in patients taking warfarin did not predictably impact TTR. A subgroup of patients who were well controlled on warfarin had significant changes in their INR in the months following GLP1-RA initiation, however without bleeding events reported. These findings suggest that GLP-1 RAs can be co-administered with warfarin without major impacts on anticoagulation stability, however, this study was limited as it was a retrospective single-center cohort, and further prospective research may be warranted.

No relevant conflicts of interest to declare.