In this study, we found wICH patients have a high prevalence of baseline thromboembolic risk factors and high incidence of post-ICH troponin elevation regardless of rFVIIa exposure. The overall incidence of clinically-significant MI (troponin-I > 1.0 ng/dL) ranged from 6-13% and was not significantly different between rFVIIa-treated compared to standard-therapy wICH cohorts. Recombinant FVIIa use also did not appear to increase the incidence of post-ICH extra-cardiac thromboembolic events in our cohorts. In addition, thromboembolic complications did not contribute to any in-hospital deaths in either wICH treatment group; rather, all deaths were attributable to the primary neurological disease. While a prior history of CAD and abnormal EKG at baseline may confer a higher risk of clinically significant MI following wICH, rFVIIa use did not appear to increase the risk of post-ICH troponin elevation or clinically significant MI after adjusting for risk factors.
Compared with the Factor Seven for Acute Hemorrhagic Stroke (FAST) study [14
], we found a higher incidence of troponin elevation (41-47%) regardless of rFVIIa use in these wICH patient cohorts. This difference likely arose from the fact that FAST included non-anticoagulated ICH patients whose baseline risk for thromboembolism may be very different from wICH patients. As we found in this study, wICH patients tend to have a high prevalence of baseline risk factors for thromboembolism. The FAST cohort had an average age of 65, 12% baseline thromboembolism history, and only 5% history of prior MI. In comparison, our wICH patients were older (mean age 73–77 years) and had a much higher prevalence of prior thromboembolic diseases, including a 32-38% prevalence of pre-existing CAD. FAST found that rFVIIa use in non-anticoagulated ICH patients was associated with a dose-dependent increase in the combined incidence of non-ST elevation MI and ST-elevation MI (STEMI) (6.4% in placebo group – 12.1% in 80μg/kg FVIIa group). Despite higher prevalence of baseline risk factors in our wICH cohort, we did not observe any STEMI in this patient cohort.
Clinical outcome analysis suggest that the rFVIIa treated cohort had lower INR at 3 and 6 hours, which is a known effect of rFVIIa on INR measurements and may not reflect reversal of warfarin-associated coagulopathy. The fact that rFVIIa treated cohort also received fewer units of FFP transfusion and were more likely to undergo surgical evacuation compared to conventional therapy cohort may be secondary to their lower INR measurements, or to clinical evidence of hemostasis. While surgical evacuation is not associated with overall mortality, we observe an interesting association between surgical evacuation and higher rate of survival in the rFVIIa-treated cohort but not in the cohort treated by conventional therapy. This association may suggest that rFVIIa-treated cohort derived more benefit from surgical hematoma evacuation while the conventional-therapy cohort did not. A prospective, randomized study would be necessary to this postulated association between rFVIIa treatment and potential benefit from surgical hematoma evacuation.
To our knowledge, this is the largest study comparing thromboembolic complications of rFVIIa use to FFP and IV vitamin K therapy in warfarin-related ICH. Compared to a large case series by Robinson and colleagues, where 101 patients treated with variable doses of rFVIIa for warfarin-associated hemorrhages of the brain and spinal canal had no cardiac thromboembolic events [19
], we found a much higher incidence of cardiac thromboembolic events following wICH regardless of rFVIIa use. Differences in location of hemorrhages, methods use for cardiac thromboembolism screening, and differences in rFVIIa dosage likely have contributed to the difference in results between our study and the study by Robinson et al. Our results on rFVIIa-related extra-cardiac thromboembolism are more consistent with that reported by Robison et al. (DVT 10%, ischemic stroke 3%) and by the FAST study (DVT 4-5%, PE 1%, ischemic stroke 3-6%). These results suggest a relatively lower overall incidence of extra-cardiac thromboembolic complications in ICH regardless of warfarin or rFVIIa use.
Consistent with prior studies [20
], the rFVIIa-treated cohort in this study achieved lower mean INR values than the standard-therapy wICH group. This effect on INR was sustained for at least 6 hours following rFVIIa administration, and the rFVIIa-treated cohort received less FFP. While recombinant FVIIa is known to normalize INR rapidly in warfarin-associated systemic bleeding [21
], whether such INR normalization reflects full reversal of coagulopathy remains controversial [22
] though several prior studies of rFVIIa use in wICH did report potential clinical therapeutic benefits [20
]. Because this study was not designed to measure the therapeutic effect rFVIIa in hemostasis in wICH, we must interpret these data with caution. The fact that rFVIIa-treated group required less FFP may reflect that rFVIIa-treated patients had clinical evidence of hemostasis with less FFP use, but a prospective study would be necessary to determine whether rFVIIa promotes improve outcome in wICH.
This study has several limitations. Although one of the larger studies of rFVIIa use in wICH, our overall sample size is small and this limits our power to detect small effects. The observed rate of 13% versus
6% clinically significant MI may reflect a true difference in cardiac ischemia risk associated with rFVIIa use which may be detected in a larger cohort. The retrospective design subjects this study to potential biases such as selection and data ascertainment bias. To minimize patient-selection bias, we chose a comparison standard-therapy wICH cohort from a time period when rFVIIa was not available on formulary. This avoids the most serious systematic errors in selecting patients with lower thromboembolic risks for rFVIIa therapy. Existence of a standardized clinical protocol for cardiac and extra-cardiac thromboembolism screening in all wICH patients limits the potential for data ascertainment bias in this particular study. This study does not address the comparative thromboemboilc complication rates and efficacy of warfarin reversal of rFVIIa compared to other potential agents such as PCC. Though we have established the prevalence of thromboembolic complications from rFVIIa use in wICH, the thromboembolic risks of PCC use in this population is still unknown. While rFVIIa as an activated factor capable of direct binding to tissue factor suggest it may achieve faster hemostasis, PCC contains all vitamin-K dependent coagulation factors and may be a more efficacious agent for warfarin reversal. Further studies are necessary to address both of these important questions. Finally, we chose to analyze all patients with intra-cerebral hemorrhage as a single cohort in this study, as has been done in other studies of rFVIIa use in wICH [19
]. Though this design increases generalizability of our data on systemic thromboembolic effects of rFVIIa use in intracranial hemorrhages in general, it introduces heterogeneity in the neurological conditions and their subsequent treatments, thereby limiting our ability to study the specific effects of rFVIIa on neurologic outcomes of wICH. Despite these limitations, this study is the first to report the overall incidence of cardiac and extra-cardiac thromboembolic complications with rFVIIa use in wICH, and to provide a comparison with thromboembolic risks of wICH patients treated with standard therapy.