Abstract
BACKGROUND AND OBJECTIVES: Prevalence estimates for classical homocystinuria (HCU) are variable and likely underestimated due to underdiagnosis. Claims data represent a strong but seldom used resource to analyze prevalence of HCU. The aim of this study was to estimate a prevalence range of HCU in the US utilizing a combination of diagnosis codes, total homocysteine levels, and clinical presentations indicative of HCU. METHODS: This was a non-interventional retrospective cohort study, using Optum's de-identified Market Clarity Data, with a patient identification period from January 01, 2016, through September 30, 2021. An algorithm was developed to identify 2 cohorts of patients using broad and strict definitions of HCU. The index date was the date within the identification period on which the first criterion was met for the inclusion criteria. Baseline demographics, clinical characteristics, and complications were assessed and summarized using descriptive statistics. Crude and standardized prevalence estimates were calculated. RESULTS: There were 3880 and 633 patients that met the relevant inclusion criteria for the broad and strict cohorts, respectively. The projected US prevalence of HCU was calculated to be 17,631 and 3466 based on the broad and strict definitions, respectively. The average annual standardized prevalence across 2016-2020 was 5.29 and 1.04 per 100,000 people for the broad and strict cohorts, respectively. CONCLUSIONS: Prevalence estimates of HCU vary depending on databases or datasets used and identification criteria. Many patients with clinical presentations suggesting a diagnosis of HCU did not have an associated diagnosis, potentially indicating underdiagnosis or underreporting. Future research should study alternative methods, such as the identification algorithm in our analysis, to better diagnose and understand the true prevalence of HCU.