A Real-World Study of Pre-Post Annualized Bleed Rates and All Cause Costs Among Non-Inhibitor Patients with Hemophilia a Switching from FVIII Prophylaxis to Emicizumab

Background Hemophilia A (HA) is a rare genetic disease characterized by a deficiency in clotting factor VIII (FVIII). Persons with HA suffer from spontaneous and traumatic bleeds which significantly impact short- and long-term quality of life. Prophylaxis treatment with FVIII replacement or non-factor replacement (e.g. emicizumab) intends to prevent bleeding episodes. To date, clinical comparisons between FVIII and emicizumab are limited to non-interventional studies and indirect comparisons. Comparisons of costs are limited to cost-effectiveness models or observational studies that include patients with and without inhibitors. An increase in availability of real-world data since emicizumab's approval in 2018 has created opportunity for comparative outcomes research in the non-inhibitor HA population. Objective To compare billed annualized bleed rates (ABR b) and all-cause costs (ACC) among non-inhibitor HA patients switching from prophylaxis with FVIII replacement to emicizumab. Methods This retrospective, observational, pre-post study used the IQVIA PharMetrics® Plus database (2015-2020)-a large longitudinal US commercial health plan database with over 190 million lives. International Classification of Diseases codes (ICD-10), National Drug Codes, and Healthcare Common Procedure Coding System were used to identify diagnoses, therapies, and procedures. Males with ≥1 claim for emicizumab who were on prophylaxis treatment with FVIII prior to initiating emicizumab were included in the analysis. Patients who received bypassing agents, immune tolerance induction, or rituximab were assumed to have inhibitors and were excluded. Patients with ≥2 occurrences of any of the following diagnoses were excluded: von Willebrand disease, hemophilia B, acquired HA, or other coagulation disorders. Annualized bleed rate was defined as billed ABR and represents bleeding episodes that required evaluation, treatment, or procedure resulting in an ICD-10 claim. Therefore, bleeds treated at home and untreated bleeds were not captured. A clinical review of ICD-10 codes resulted in a list of 535 codes used to identify HA-related bleeding episodes (e.g. hemarthrosis). The ACC were calculated as the mean cost per patient per year in 2020 US dollars actually paid by the insurer. Descriptive statistics were used to summarize, and Bayesian models were developed to compare, ABR b and ACC in the pre- and post-switch periods. Bayesian inferences estimated the population mean difference in A