Incidence of hemophilia is commonly cited in the literature as 1/5,032 male births (Soucie et al., Am J Hematol 1998). Estimates of the true population prevalence of hemophilia A (HA; factor VIII deficiency) and B (HB; factor IX deficiency) are limited, and prevalence is influenced by the longer patient life expectancy now versus previous decades. We calculated updated epidemiologic estimates of HA and HB in the USA based on a literature review and other data sources.
Data were collected from a systematic literature search (1970–2018) in Medline, EMBASE, conference proceedings, and other secondary data sources including registries (e.g., Registry for Bleeding Disorders Surveillance [part of Community Counts]). Keywords included: hemophilia, bleeding disorder, factor VIII or IX deficiency, incidence, prevalence, mortality, diagnosed, undiagnosed, severity. Eighteen references contributed to the USA analysis. Variables assessed comprised: incidence, prevalence (diagnosed, undiagnosed, total), and disease severity (mild, moderate, severe). Lastly, a simplified Markov model was developed to calculate the annual incoming patients, deaths and prevalence of HA or HB in a given year, including projections to 2020.
As most literature references the 1/5,000 male birth incidence rate for HA and HB combined, we used this and assumed it remains constant. Based on assumptions from 2015, and an incidence rate of 15.49 for HA and 4.24 for HB per 100,000 male births there were 321 incident patients with HA and 88 with HB in the USA. Total USA prevalence was estimated to be 22,118 for HA and 6,058 for HB. Diagnosed and undiagnosed prevalence rates per 100,000 males were: 8.69 and 4.87 for HA, and 2.57 and 1.14 for HB, respectively. The corresponding percentage of patients with mild, moderate, and severe disease, respectively was 34, 16, and 51% for diagnosed HA and 39, 33, 28% for diagnosed HB. In addition, we assumed severity rates for undiagnosed HA were 70, 30, and 0% and 60, 40, 0% for undiagnosed HB. Given current treatment rates, we assumed prevalence rates plateaued around 2015. Therefore, 2015 prevalence rates were used to project prevalence in 2020. Based on the model, predicted total prevalence of hemophilia A and B in 2020 is estimated to be 22,913 and 6,276 patients. (Table)
Overall, the model assumes constant incidence rates of HA and HB. Due to exclusion of undiagnosed patients in previous literature, this model’s total prevalence estimates are higher. In the USA, thousands of patients remaining undiagnosed suggest room for improvement in diagnosis and/or reporting.