Development and validation of a novel prediction model for new-onset atrial fibrillation after lung resection.

BACKGROUND

Postoperative atrial fibrillation (POAF) is the most prevalent and potentially life-threatening arrhythmia following thoracic surgery. This study aimed to construct and validate a predictive model for assessing POAF risk.

METHODS

A meta-analysis was conducted to rank risk factors associated with POAF based on their respective risk ratios (RRs). Significant risk factors identified from the meta-analyses were incorporated into the model and assigned weights. External validation was performed using a retrospective cohort from China. Receiver operating characteristic (ROC) curves, calibration plots and decision curve analysis (DCA) were employed to assess the model's predictive performance, calibration and clinical utility.

RESULTS

We screened 40 cohort studies involving 58,899 patients. We developed a risk model that incorporated age ≥ 70 years (RR 2.10, 95% CI 1.34-3.30;  < 0.05), male sex (RR 1.46, 95% CI 1.34-1.60;  < 0.05), COPD (RR 2.28, 95% CI 1.81-2.89;  < 0.05), CAD (RR 1.72, 95% CI 1.49-1.99;  < 0.05), heart failure (RR 1.62, 95% CI 1.12-2.35;  < 0.05), pneumonectomy (RR 2.32, 95% CI 2.01-2.67;  < 0.05) and lobectomy (RR 1.86, 95% CI 1.38-2.51;  < 0.05) and thoracotomy (RR 1.46, 95% CI 1.30-1.64;  < 0.05). Validation was performed in an external cohort of 1546 participants, demonstrating strong discrimination with an area under the receiver operating characteristic curve (95% CI) of 0.89 (95% CI 0.81-0.83). The calibration curve and DCA curve results demonstrated good concordance and applicability.

CONCLUSIONS

This model, built with easily accessible clinical variables, could accurately predict the risk of POAF. This holds promise for improving clinical decision making and guiding early interventions.