Feasibility of Using Finnish Real-World Data to Build an External Control Arm for the Factor Xia Inhibitor Asundexian Pacific-AF Trial

OBJECTIVES: To evaluate whether Finnish-Real World Data (F-RWD) can be used as an External Control Arm (ECA) to the PACIFIC-AF trial. PACIFIC-AF (phase 2b) compared the safety of the oral FXIa inhibitor asundexian to apixaban in patients with atrial fibrillation (AF).

METHODS: The Finnish healthcare system was used as data source to develop an ECA and test novel approaches for clinical trials. PACIFIC-AF randomized participants to one of two asundexian arms (20mg or 50mg) or to the Internal Control Arm (ICA) for a 12-week period. The primary endpoint for the PACIFIC-AF trial was the ratio of the proportions of the composite of International Society of Thrombosis and Haemostasis (ISTH) major and clinically relevant non-major bleeding within 3 months. Participants treated with apixaban and fulfilling eligibility for PACIFIC-AF were identified in F-RWD; participants were then matched to the PACIFIC-AF ICA using propensity-score (PS) matching with a common-support region. Standardized mean differences (SMDs) across 28 pre-specified baseline confounders, descriptive statistics and Aalen-Johansen curves were used to compare ECA and PACIFIC-AF ICA.

RESULTS: ICA (n=250) and ECA (n=239) were established from an eligible F-RWD pool (n=1763). Confounders were well-balanced between ICA and ECA, with the largest discrepancy being prior use of heparins (22.0% vs 27.6%), respectively. The composite bleeding outcome occurred in 6 participants compared to 9 participants in the ECA, the cumulative incidence was 2.4 (90% CI 1.16 - 4.42) versus 3.88 (90% CI 2.12 - 6.26). Sensitivity analyses, consisting of PS-weighting approaches, provided similar but improved matching results.

CONCLUSIONS: The F-RWD was successfully used to build an ECA that showed baseline and outcome overlap with the PACIFIC-AF ICA. Sensitivity analyses showed improvement over the pre-specified primary approach, suggesting a gate-keeping approach to PS-matching and/or weighting methods to be used to establish the most efficient ECA while maintaining pre-specification of methods.