Effects of INR Variability and Time in the Therapeutic Range on the Rates of Gastrointestinal Bleeding in Patients Supported by Left Ventricular Assist Devices

Abstract

Listen

Translate article

Purpose Patients supported with left ventricular assist devices (LVADs) require chronic anticoagulation with warfarin to prevent thrombus formation within the device. While a goal INR range of 2 to 3 is generally accepted as appropriate, higher time in the therapeutic range (TTR) has been associated with improved outcomes. This analysis aimed to also determine the impact of INR variability on gastrointestinal bleeding (GIB) events. Methods This was a retrospective, longitudinal cohort study at a single center from January 2015 to December 2017. Standard INR goal was 2 to 3 but could be adjusted on a patient-specific basis. GIB was defined per the standard INTERMACS definition. INR TTR was calculated using the Rosendaal method, and coefficient of variation (CV) was calculated as a ratio of standard deviation to the mean. For the primary analysis, patients who experienced a GIB were compared to those who had not. Results Baseline characteristics were similar between the two groups (Table 1). Of the 87 subjects analyzed in the study, 23 (26.4%) experienced a GIB (average time to bleed 363 days) with 9 patients developing a second GIB. The overall TTR for the population was 65.0%, and the overall CV was 27.3%. Patients who had a GIB had a lower mean TTR than those who did not bleed (58.8% vs 67.2%, p=0.03). For those that bled, the TTR in the 30 days prior to the GIB event (50.4%) was lower than the overall TTR (p=0.032). Patients who experienced a GIB also had significantly more variability in their INR values than those who did not bleed (31.2% vs 25.8%, p=0.035) (Table 2). Conclusion Both increased INR variability and decreased TTR are associated with increased rates of GIB in patients supported by LVADs. This is the first study to explore whether more INR variation puts patients with LVADs at greater risk of developing a GIB. Further prospective studies are warranted to further test this hypothesis.