Blasting the myth of predictive INR changes related to plasma transfusion: an academic institution's experience

Abstract

Introduction: Plasma transfusion is a common therapeutic strategy used to lower international normalized ratio (INR) values in the non-emergent setting. However, due to lack of evidence of its efficacy, standardized guidelines for this practice have not been well established. Methods: This retrospective observational cohort study analyzed 276 inpatient encounters that involved plasma transfusions focusing on change in INR values from pre- to post-transfusion, with respect to the following predictor variables: vitamin K co-administration, number of plasma units transfused, order indication and body mass index (BMI). Results: The overall average change in the INR was 1.35. Patients who received vitamin K showed an average change of 2.51, while patients that did not receive vitamin K demonstrated an average change of 0.70. Increased numbers of plasma units transfused showed benefit up to three-unit orders. Greater decreases in the INR were observed for patients requiring plasma for anticoagulation reversal or active bleeding. There was no significant difference in the change in INR based on the BMI. By multivariate and regression analyses, the stepwise addition of each successive predictor variable demonstrated an increase in the shared variance in the outcome of the post-transfusion INR: the pre-transfusion INR and vitamin K co-administration alone was not significant (p = 0.45); the additional number of plasma units transfused was significant (R² = 0.13, p < 0.001), and; the subsequent additional plasma order indications (R² = 0.19, p < 0.001) and BMI (R² = 0.18, p < 0.001) were increasingly significant. Conclusion: Taking into consideration the combination of multiple predictive factors may aid in a more efficient use of plasma products.