Prevalence and Time Trends of Oral Anticoagulant Utilization in Adults with Sickle Cell Disease

Abstract

Background: Sickle cell disease (SCD) is an inherited red blood cell disorder, caused by a genetic mutation in the β-hemoglobin chain. SCD is associated with chronic activation of coagulation and an increased risk of venous thromboembolism (VTE). Although the American Society of Hematology guidelines recommend indefinite anticoagulation for unprovoked or recurrent provoked VTE, no guidance is given regarding choice of anticoagulant due to a paucity of data. Existing evidence on the use of anticoagulants in patients with SCD has been mostly focused on oral anticoagulants, was limited by small sample sizes, and did not clearly demonstrate the temporal utilization patterns. Real-world evidence on oral anticoagulant use in SCD from large population-based studies is largely lacking. This study examined the prevalence and time trends of oral anticoagulant utilization in patients with SCD in the United States (US). Methods: This retrospective, repeated cross-sectional study analyzed the Cerner Health Facts® database from January 1, 2011, to December 31, 2016, which captures longitudinal electronic medical records (EMR) of patient data from participating facilities in the US. To be included, patients were required to (1) have ≥ 1 inpatient or emergency department encounters with SCD, or 2 outpatient encounters that were 30 days apart in a calendar year from 2012 to 2016, (2) have ≥ 1 prescription records on the encounters with SCD, (3) aged ≥ 18 years at index, and (4) have ≥ 1 encounter with SCD in a year prior to the index. For patients with multiple encounters with SCD in the calendar year, we randomly selected one encounter with SCD and defined it as the index. In this design, a patient contributed only one observation per year but may contribute to multiple years of observations if meeting the selection criterion. The outcome was the use of anticoagulants, defined as any prescription drug encounter for an oral anticoagulant (Vitamin K antagonist [VKA]: warfarin; direct oral anticoagulants [DOACs]: apixaban, rivaroxaban, dabigatran, edoxaban) within a grace period of 7 days discharged from the index. The independent variable was each calendar year measured as dummy variables. Covariates were measured during the 12-month baseline period prior to the index, which included patient demographics, comorbidities and complications of SCD, and facility characteristics. Time trends on the proportions of eligible patients with anticoagulant use against the calendar years of the SCD diagnosis were plotted, and the significance on the crude trend was examined using Cochran-Armitage trend test. In the main analysis, a modified Poisson regression by using generalized estimating equations (GEEs) estimated the adjusted prevalence ratios (APRs) for the trends and covariates associated with any anticoagulant utilization. In subgroup analysis, we estimated the APRs of covariates for the use of DOACs versus VKA. Results: The cohort included 8,824 adult patients with SCD in 182 facilities using automated Cerner EMR system for the years 2012 to 2016. Prescription rates of all oral anticoagulants gradually increased from 2012 to 2014; the rates of VKA prescriptions substantially dropped from 11.1/1000 to 4.1/1000 persons between 2014 and 2016, whereas prescription rates of DOACs increased significantly from 12.0/1000 to 29.1/1000. Prevalence of VTE was stable at approximately 6.2/1000 to 7.5/1000 between 2012 and 2016. Prescription rates of DOACs were lower in patients with (versus without) VTE (6.1/1000 vs 7.3/1000), while a reversed prescription pattern was observed for VKA (VTE [yes vs no]: 4.4/1000 vs 1.8/1000). Controlling for covariates, our main analysis demonstrated that patients with a diagnosis of acute chest syndrome, pulmonary hypertension, and depression were 1.51, 2.30, and 1.57 times more likely to have oral anticoagulant prescription, respectively (all p <0.05). Subgroup analysis showed that having a diagnosis of VTE was associated with a decreased likelihood of DOACs prescription relative to VKA (APR= 0.75, 95%CI: 0.62-0.91). Conclusion: Prescription of oral anticoagulants in adults with SCD has significantly increased from 2012 to 2016 and the increase was dominated by DOACs. The treatment paradox of a lower relative rate of DOACs utilization versus VKA in adults with SCD requires more study to understand the reasons and additional clinical outcomes associated with the utilization patterns.