Is the International Normalized Ratio (INR) Enough? Reevaluating INR-Based Decision-Making for Four-Factor Prothrombin Complex Concentrate in Geriatric Warfarin-Associated Intracranial Hemorrhage: A Multicenter Cohort Study

Background/Objectives: Warfarin-associated intracranial hemorrhage (WA-ICH) remains one of the most devastating anticoagulant-related emergencies, particularly among older adults, with substantial morbidity and mortality. Although current guidelines recommend that reversal strategies should be guided primarily by bleeding severity and clinical presentation rather than the international normalized ratio (INR) alone, real-world treatment decisions often continue to rely on INR thresholds. This study investigated whether INR is an adequate determinant for four-factor prothrombin complex concentrate (4F-PCC) administration in geriatric patients with non-traumatic WA-ICH. Methods: In this multicenter retrospective cohort study, consecutive patients aged ≥65 years presenting with non-traumatic WA-ICH to five tertiary-care centers in Turkey were evaluated. Patients were categorized according to reversal strategy into a 4F-PCC group (n = 30) and a heterogeneous non-PCC management group (n = 30), which included supportive care, intravenous vitamin K, FFP, combinations of these approaches, or no specific reversal treatment. Subgroup analyses were performed using multiple INR thresholds to examine the relationship between INR levels, treatment allocation, and in-hospital mortality. The primary outcome was all-cause in-hospital mortality. Results: Patients who received 4F-PCC had higher admission INR values than those managed without PCC, whereas admission Glasgow Coma Scale (GCS) was similar between treatment groups. Admission GCS was substantially lower among non-survivors than survivors. In the revised Firth-penalized logistic regression model, higher admission GCS was independently associated with lower odds of in-hospital mortality (aOR 0.37 per 1-point increase; 95% CI 0.10–0.62; p < 0.001), whereas 4F-PCC administration and admission INR values were not independently associated with mortality. Conclusions: Higher admission INR was associated with 4F-PCC administration and was not independently associated with mortality, whereas admission GCS was independently associated with in-hospital mortality. These findings support consideration of neurological, radiological, and laboratory information together rather than interpretation of admission INR in isolation. Given the retrospective design, small sample size, baseline differences in anticoagulation intensity, and potential residual confounding, the findings should be considered hypothesis-generating.