Differential risk of systemic complications and mortality in HFrEF and HFpEF patients undergoing total knee arthroplasty: A nationwide propensity matched study with cox regression analysis

Abstract

Purpose: Heart failure (HF), classified as preserved (HFpEF) or reduced ejection fraction (HFrEF), is associated with heightened perioperative risk. This study assessed postoperative outcomes after total knee arthroplasty (TKA) in HF patients, by ejection fraction (EF), comorbidities, laboratory values, and pharmacotherapy on complications and mortality. Methods: Using the TriNetX database (2005–2025), adults undergoing elective primary TKA with HFpEF or HFrEF were propensity score–matched to healthy controls. Matching included demographics, comorbidities, medication use, and labs: BMI, INR, hemoglobin A1c, troponin, and B-type natriuretic peptide (BNP). Outcomes included 90-days and 1-year systemic complications and mortality. Cox regression evaluated independent risk predictors. Results: Final cohorts included 2611 HFpEF and 554 HFrEF patients with matched controls. Both subtypes demonstrated elevated 90-days mortality (OR 7.48 HFpEF; OR 7.43 HFrEF). HFpEF patients had significantly increased risks of MI (OR 13.96), atrial fibrillation (OR 19.31), stroke (OR 9.93), sepsis (OR 10.36), pulmonary embolism (OR 6.55), and AKI (OR 5.25). HFrEF patients showed similar patterns, including atrial fibrillation (OR 17.01), AKI (OR 9.17), and MI (OR 6.91). Cox regression identified age, male sex, elevated BNP, dialysis, chronic kidney disease, and prior MI or stroke as predictors of mortality. Medication analysis revealed ACE inhibitors, ARBs, SGLT2 inhibitors, GLP-1 agonists, and beta blockers as protective, whereas calcium channel blockers and loop diuretics increased risk. Conclusion: HF patients undergoing TKA face increased complications and mortality. Stratifying risk by EF, comorbidities, and laboratory markers, particularly BNP and INR, along with tailored pharmacotherapy, may improve perioperative management and outcomes.