Potentially modifiable risk factors for atrial fibrillation following lung resection surgery: a retrospective cohort study

Abstract

Atrial fibrillation is the most frequent arrhythmia after thoracic surgery and is associated with increased hospital costs, morbidity and mortality. In this study, we aimed to identify potentially modifiable risk factors for postoperative atrial fibrillation following lung resection surgery and to suggest possible measures to reduce risk. We retrospectively reviewed the medical records of 4731 patients who underwent lobectomy or more major lung resection over a 6-year period. Patients who developed atrial fibrillation postoperatively and required treatment were included in the postoperative atrial fibrillation group, while the remaining patients were assigned to the non-postoperative atrial fibrillation group. Risk factors for postoperative atrial fibrillation were analysed by multivariate analysis and propensity score matching. Overall, 12% of patients developed postoperative atrial fibrillation. Potentially modifiable risk factors for postoperative atrial fibrillation were excessive alcohol consumption (odds ratio (OR) = 1.48, 95% CI 1.08-2.02, p = 0.0140), red cell transfusion (2.70(2.13-3.43), p < 0.0001), use of inotropes (1.81(1.42-2.31), p < 0.0001) and open (vs. thoracoscopic) surgery (1.59(1.23-2.05), p < 0.0001). Compared with inotrope use, vasopressor administration was not related to postoperative atrial fibrillation. Use of steroids or thoracic epidural anaesthesia did not reduce the incidence of postoperative atrial fibrillation. We conclude that high alcohol consumption, red cell transfusion, use of inotropes and open surgery are potentially modifiable risk factors for postoperative atrial fibrillation. Pre-operative alcohol consumption needs to be addressed. Avoiding red cell transfusion and performing lung resection via video-assisted thoracoscopic surgery may reduce the incidence of postoperative atrial fibrillation and the administration of vasopressors rather than inotropes is preferred.

Prevalence and risk factors of atrial fibrillation during lung and esophageal surgery

“…Alcohol consumption was defined as consuming eight alcohol drinks or more per week for women and 15 or more per week for men. [ 9 ]…”

“…In previous studies, many risk factors were reported to be associated with the development of postoperative AF (POAF) such as older age, male sex, hypertension, obesity, smoking, alcohol consumption, and history of chronic obstructive pulmonary disease (COPD). [ 9 – 12 ] Furthermore, some surgical procedures are reported to be related to the occurrence of POAF. [ 13 – 16 ] Muranishi et al reported that mediastinal lymph dissection for patients with early-stage lung cancer is associated with a high risk for the development of POAF.…”

“…[ 17 ] Lee et al reported that open thoracotomy was a more significant risk factor of POAF than video thoracoscopy-assisted thoracic surgery (VATS). [ 9 ] AF is a potentially preventable complication, as some risk factors such as smoking or surgical procedures are modifiable. We can evaluate the risk of surgical AF accurately and take actions to prevent the development of AF by defining more risk factors especially in surgical procedures.…”

Major trauma and the need for massive transfusion

“…This is an increasingly popular method used to improve causal inferences when analysing observational data in retrospective studies. For example, propensity score matching was used in two recent retrospective studies published in Anaesthesia , one looking at the effect of prolonged propofol infusion and one investigating potentially modifiable risk factors for atrial fibrillation .…”

Regression: How much data do I really need?

“…Since I am only including five clinically relevant variables in my model, does that mean I only have to collect data on 50 patients? This sounds too good to be true, since most retrospective studies (like the ones published in Anaesthesia ) I have seen include hundreds, if not several thousand patients. So how do I really calculate the sample size, and why have these published studies included data on so many patients?…”