Preoperative AF and CV Outcomes After Noncardiac Surgery

Quick Takes

  • In a very large population of Medicare database patients undergoing noncardiac surgery, pre-existing atrial fibrillation (AF) was independently associated with greater risk of all-cause mortality, heart failure (HF), and ischemic stroke within 30 days of surgery.
  • Conversely, pre-existing AF was independently associated with lower risk of myocardial infarction within 30 days of surgery.
  • Addition of pre-existing AF to the six covariates of the Revised Cardiac Risk Index (RCRI) model (stroke or TIA, coronary artery disease, HF, diabetes treated with insulin, ≥ Stage 3 chronic kidney disease, and planned high-risk surgical procedure) led to significant improvement in prediction of postoperative major adverse cardiovascular events compared to use of the RCRI alone.

Study Questions:

What is the impact of nonvalvular atrial fibrillation (AF) on risk of postoperative major adverse cardiovascular events after noncardiac surgery?


Medicare beneficiaries aged ≥40 years with ≥3 years of Medicare coverage and no history of mitral stenosis, undergoing noncardiac surgery between January 1, 2015–October 1, 2019, were included in this analysis. Patients with pre-existing AF were identified, and after unadjusted analysis, underwent propensity-score matching by age, sex, race and ethnicity, CHA2DS2-VASc, surgical urgency, planned surgical procedure, and Revised Cardiac Risk Index (RCRI) to patients without a history of AF. The primary outcome was 30-day postoperative all-cause mortality. Secondary outcomes included 30-day postoperative ischemic stroke, myocardial infarction (MI), heart failure (HF), major bleeding, and hospital length of stay (LOS).


Among the 8,635,758 subjects, 1,411,955 (16.4%) had pre-existing AF. Unadjusted analysis showed that patients with pre-existing AF were older (77.9 ± 9.0 vs. 73.1 ± 9.2 years), more often male (54 vs. 43%), and had greater prevalence of comorbid medical conditions.

After propensity matching against 1,923,438 patients without pre-existing AF, patients with pre-existing AF showed the following outcomes in the first 30 postoperative days:

  • Greater risk of all-cause mortality (8.3% vs. 5.8%, odds ratio [OR] 1.31 [1.30, 1.32])
  • Greater risk of HF (4.44% vs. 2.85%, OR 1.31 [1.30, 1.34])
  • Greater risk of stroke (1.70% vs. 1.13%, OR 1.40 [1.37, 1.43])
  • Greater risk of major bleeding (3.76% vs. 3.14%, OR 1.14 [1.13, 1.16])
  • Longer postoperative hospital LOS (median 5 [3, 9] vs. 4 [2, 8] days)
  • Lesser risk of 30-day nonfatal postoperative MI (1.75% vs. 1.93%, OR 0.81 [0.79, 0.82])

Subgroup analysis showed that among the patients with pre-existing AF, the higher risks of 30-day postoperative mortality, HF, and stroke were observed consistently across AF risk strata (CHA2DS2-VASc score). However, the decreased risk of 30-day postoperative MI was observed only in higher risk AF (CHA2DS2-VASc score of ≥2), whereas increased risk of 30-day postoperative MI was observed in lower risk AF (CHA2DS2-VASc score 0-1).

The addition of pre-existing AF to the RCRI resulted in improved discrimination for determining risk of adverse postoperative cardiovascular events, increasing the model’s C-statistic of the logistic regression model from 0.73–0.76 compared to use of RCRI alone.


Pre-existing AF was independently associated with greater risk of 30-day postoperative all-cause mortality, HF, and ischemic stroke in this very large database population. Inclusion of pre-existing AF as a risk factor in addition to all covariates in the RCRI, compared to use of the RCRI alone, may improve preoperative risk discrimination.


The growing prevalence of AF in the aging US population is likely to increase its relevance as a risk factor for adverse cardiovascular outcomes after noncardiac surgery. The apparent improvement in risk assessment afforded by pre-existing AF, when added to the RCRI, will now require confirmation by a prospective trial.

Clinical Topics: Arrhythmias and Clinical EP, Geriatric Cardiology, Heart Failure and Cardiomyopathies, Prevention, Atrial Fibrillation/Supraventricular Arrhythmias, Acute Heart Failure

Keywords: Atrial Fibrillation, Diabetes Mellitus, General Surgery, Geriatrics, Heart Failure, Hemorrhage, Ischemic Attack, Transient, Ischemic Stroke, Length of Stay, Myocardial Infarction, Primary Prevention, Renal Insufficiency, Chronic, Risk Assessment, Risk Factors, Stroke, Vascular Diseases

< Back to Listings