(P-431) Designing Epidemiological Studies to Estimate the Incidence of Thrombosis in Patients with Antiphospholipid Syndrome with History of Recurrent Thrombotic Events

Background: Due to the scarcity of data on the incidence of thrombosis in antiphospholipid syndrome (APS) patients with history of recurrent thrombotic events, we planned and explored new epidemiological studies to estimate the incidence in a cost-effective manner.

Objectives: To design studies to (1) explore the feasibility of utilizing TriNetX electronic medical record (EMR) database derived from U.S. health care organizations to estimate thrombotic risk in APS patients with recurrent thrombosis and (2) explore alternative designs to improve the validity of study.

Methods: First, the TriNetX database as of June 2020 was used to extract records of patients with APS, identified by two occurrences of ICD-10 diagnosis code D68.61 at least 30 days apart. APS patients with any of the two distinct thrombotic events, such as deep vein thrombosis (DVT), pulmonary embolism (PE), ischemic stroke (IS), acute myocardial infarction (AMI) and/or renal vein thrombosis (RVT) occurring at least 30 days apart were identified by presence of relevant ICD diagnosis codes. A patient’s index date was defined as Day 31 starting on the date of the second thrombotic event. Patients were followed for new thrombotic events from the index date until censoring of up to 1 year or June 30, 2020. Next, based on the results of the database analysis, alternative study designs were proposed.

Results: A total of 1,182 APS patients with history of recurrent thrombosis were included in the analysis. During one-year follow-up period, there were 798 new thrombotic events: 302 DVT, 379 PE, 70 IS, 45 AMI and 2 RVT. Comparing these results to the reported incidence of thrombosis in APS patients is challenging as previous studies were not restricted to patients with history of recurrent thrombosis. One limitation of the database study is that the validity of using ICD codes to identify thrombotic events in APS patients has not been assessed, though the algorithms for identifying DVT, PE, IS and AMI have been validated in other patient populations. Therefore, we proposed to augment the database study by including outcome adjudication through medical chart review. Another approach is to conduct a full medical chart review study to collect additional clinical data beyond diagnosis codes of thrombotic events, however it is known to be more costly and time-consuming.

Conclusions: Considering the limitations of an analysis using EMR databases, we concluded that including medical chart review in outcome ascertainment could improve the validity of our study.