Publication Date



genomics, warfarin, stroke, atrial fibrillation


Background: With over 40 years of demonstrated clinical efficacy, warfarin remains the world’s most used pharmaceutical to prevent ischemic stroke in patients with atrial fibrillation (AF). However, warfarin has many challenges. Thus, despite known effectiveness, warfarin is a leading cause to drug-induced morbidity and mortality. Over 50 different warfarin therapy protocols, including a number of pharmacogenomic-based (PG) protocols, with as many as 14 independent variables, have been developed to improve safety and efficacy, thereby reducing ischemic strokes and intracranial hemorrhages (ICH).

Purpose: To conduct a preliminary cost-effectiveness study to determine the price point at which using warfarin PG dosing to prevent ischemic stroke and ICH would provide a neutral cost difference for the AF patient population at Aurora Health Care.

Methods: Using a 15-year retrospective electronic medical record, we generated a large enough simulated AF population using Bayesian Network modeling to conduct a series of simulated warfarin therapies. We used five different PG and non-PG warfarin therapy protocols. The protocols with various levels of personalization used different PG and non-PG dosing algorithms for initial, adjustment and maintenance warfarin dosing. The simulation platform was able to predict daily international normalized ratio values and the rate of ischemic stroke and ICH in each simulated patient over 90 days for each of the five warfarin therapy protocols. Using nationwide estimates derived through literature review, we estimated the total cost of administering warfarin using the five different protocols and subsequent acute and 5-year care costs (adjusted according to annual Medical Care Component of the Consumer Price Index).

Results: In 2000–2015, a total of 48,006 patients, or ~3,000 patients annually, initiated warfarin therapy with a primary diagnosis of AF at Aurora. We found that Aurora’s current best practice warfarin therapy protocol had the highest predicted costs for ICH, resulting in $163,462.55 for acute care costs (averaged over 1,000 patients) and $171,279.60 for 5-year ongoing care related to ICH. The current Aurora protocol also had the highest associated costs for ischemic stroke at $51,333.45 for acute care and $58,507.21 for 5-year ongoing care. In contrast, PG-protocol 3, which incorporated patient’s genotype into the warfarin dosing protocol, had the lowest predicted acute care and ongoing costs associated with ICH at $119,823.78 for acute care and $125,462.75 for ongoing care. For 1,000 AF patients initiating warfarin therapy in Aurora, switching to a PG protocol would save $55,299.94 in acute care costs and $59,198.50 in 5-year ongoing care.

Conclusion: At a cost of $59.20 per patient, warfarin PG dosing is cost neutral. With substantial decreases in genotyping cost in recent years, it is likely this price is currently achievable. Therefore, although there is only incremental clinical benefit from warfarin PG dosing, it is likely cost-effective.