Background: Prothrombin complex concentrates (PCCs) are frequently used to reverse the effect of vitamin K antagonists (VKAs) in patients with non-traumatic intracerebral hemorrhage (ICH). However, information on the rate of thromboembolic events (TEs) and allergic events after PCC therapy in VKA-ICH patients is limited. Methods: Consecutive VKA-ICH patients treated with PCC at our institution between December 2004 and June 2014 were included into this retrospective observational study. We recorded international normalized ratio (INR) values before and after PCC treatment, baseline clinical characteristics including the premorbid modified Rankin Scale (pmRS) score, TE and allergic event that occurred during the hospital stay. All events were classified by 3 reviewers as being ‘related', ‘probably related', ‘possibly related', ‘unlikely related' or ‘not related' to treatment with PCC. To identify factors associated with TEs, log-rank analyses were applied. Results: Two hundred and five patients were included. Median INR was 2.8 (interquartile range (IQR) 2.2-3.8) before and 1.3 (IQR 1.2-1.4) after PCC treatment and a median of 1,500 IU PCC (IQR 1,000-2,500) was administered. Nineteen TEs were observed (9.3%); none were classified ‘related' but 9 were classified as ‘possibly' or ‘probably related' to PCC infusion (4.4%). One allergic reaction (0.5%), ‘unlikely related' to PCC, was observed. In the whole cohort, PCC doses >2,000-3,000 IU, ICH volumes >40 ml, National Institute of Health Stroke Scale values >10 and a pmRS >2 were associated with the development of TEs (p = 0.031, p = 0.034, p = 0.050 and p = 0.036, respectively). Conclusions: Overall, INR reversal with PCC appears safe. Though no clear relationship between higher PCC dosing and TEs was observed, PCC doses between >2,000 and 3,000 IU and higher morbidity at ICH onset were associated with TEs. Hence, individual titration of PCC to avoid exposure to unnecessarily high doses using point-of-care devices should be prospectively explored.