Predicting atrial fibrillation after acute abdominal surgery
Abstract Introduction Post-operative atrial fibrillation (POAF) is currently considered a phenomenon rather than a definite diagnosis. Nevertheless, POAF is associated with an increased rate of complications, including stroke and mortality. The incidence of POAF in acute abdominal surgery has not been reported and prediction of patients at risk has not previously been attempted. Purpose We aim to report the incidence of POAF after acute abdominal surgery and provide a POAF prediction model based on pre-surgery risk-factors. Methods Designed as a prospective, single-centre, cohort study of unselected adult patients referred for acute, general, abdominal surgery. Consecutive patients (>16 years) were included during a three month period. No exclusion criteria were applied. Follow-up was based on chart reviews, including medical history, vital signs, blood samples and electrocardiograms. Chart reviews were performed prior to surgery, at discharge, and three months after surgery. Atrial fibrillation was diagnosed either by specialists in Cardiology or Anaesthesiology on ECG or cardiac rhythm monitoring (≥30 seconds duration). Multiple logistic regression with backward stepwise selection was used for model development. Receiver operating characteristic curves (ROC) including area under the curve (AUC) was produced. The study was approved by the Regional Ethics committee (H-19033464) and comply with the principles of the Declaration of Helsinki of the World Medical Association. Results In total, 466 patients were included. Mean (±SD) age was 51.2 (20.5), 194 (41.6%) were female, and cardiovascular comorbidity was present in ≈10% of patients. Overall incidence of POAF was 5.8% (27/466) and no cases were observed in patients <60 years. Incidence was 15.7% (27/172) for patients ≥60 years. Prolonged hospitalization and death were observed in 40.7% of patients with POAF vs 8.4% patients without POAF (p<0.001). Significant age-adjusted risk-factors were previous atrial fibrillation odds ratio (OR) 6.84 [2.73; 17.18] (p<0.001), known diabetes mellitus OR 3.49 [1.40; 8.69] (p=0.007), and chronic kidney disease OR 3.03 [1.20; 7.65] (p=0.019). A prediction model, based on age, previous atrial fibrillation, diabetes mellitus and chronic kidney disease was produced (Figure 1), and ROC analysis displayed AUC 88.26% (Figure 2). Conclusions A simple risk-stratification model as the one provided, can aid clinicians in identifying those patients at risk of developing POAF in relation to acute abdominal surgery. This is important, as patients developing POAF are more likely to experience complications, such as prolonged hospitalization and death. Closer monitoring of heart rhythm and vital signs should be considered in at-risk patients older than 60 years. Model validation is warranted. FUNDunding Acknowledgement Type of funding sources: None.