(196) Validation of Four Post-Operative Atrial Fibrillation (POAF) Scoring Tools Following Cardiac Surgery

Introduction: Post-operative atrial fibrillation (POAF) is common following cardiac surgery and is associated with worse clinical outcomes. Several risk scoring tools have been proposed to assess the risk of developing POAF. However, the predictive ability of these tools is suboptimal and lacks robust external validation. This study aims to assess the performance of four POAF scoring tools at an institutional level.

Methods: This retrospective validation study assessed patients who underwent CABG, valve replacement, or combination CABG and valve replacement from January to December 2023. Patients with a history of atrial fibrillation were excluded. Data collected from the STS Adult Cardiac Surgery Database and chart review were used to calculate risk scores using four scoring tools: CHADS2 score, HATCH score, a tool developed by Chen et al., and the POAF score. The validity of these tools was assessed using sensitivity, specificity, and area under the receiver operating curve (AUC).

Results: A total of 483 patients were analyzed: 299 isolated CABG procedures, 132 isolated valve procedures, and 52 combination valve and CABG procedures. The CHAD2 score had a sensitivity of 56.6%, specificity of 48.8%, and AUC of 0.54 (95% confidence interval [CI], 0.48-0.59) overall, and sensitivity of 58.3%, specificity of 49.3%, and AUC of 0.55 (CI 0.48-0.63) in isolated CABG. The HATCH score had a sensitivity of 57.9%, specificity of 52.8%, and AUC of 0.57 (CI 0.51-0.62) overall, and sensitivity of 55.9%, specificity of 59%, and AUC of 0.59 (CI 0.53-0.66) in isolated CABG. The tool developed by Chen et al. had a sensitivity of 58.5%, specificity of 43.5%, and AUC of 0.51 (CI 0.45-0.57) overall, and sensitivity of 61.6%, specificity of 46.6%, and AUC of 0.53 (CI 0.45-0.61) in isolated CABG. The POAF score had a sensitivity of 27.6%, specificity of 70.1%, and AUC of 0.52 (CI 0.46-0.59) overall, and sensitivity of 26.1%, specificity of 79.2%, and AUC of 0.50 (CI 0.43-0.57) in isolated CABG.

Conclusions: In a surgical population with a POAF incidence comparable to average, the validity of four scoring tools was poor. The tool developed by Chen et al. had the highest sensitivity both overall and for isolated CABG, and the HATCH score had the highest AUC. Further research is warranted to develop and assess POAF scoring tools.