Perioperative Management of Anticoagulation in Patients With Mechanical Heart Valves

Importance: Perioperative interruption of vitamin K antagonist (VKA) therapy in patients with mechanical heart valves (MHVs) is complex, and contemporary data to inform management are limited. Objectives: To describe perioperative anticoagulation management and to estimate 30-day risks of arterial thromboembolism (ATE) and bleeding after VKA interruption in adults with left-sided MHVs. Design, Setting, and Participants: This retrospective cohort study included consecutive adult patients (aged ≥18 years) with left-sided aortic, mitral, or dual MHVs undergoing planned invasive procedures requiring VKA interruption from January 1, 2016, to December 31, 2023, with 30-day follow-up. The study was conducted at the thrombosis clinic at The Ottawa Hospital in Ontario, Canada. Exposure: Planned invasive procedure requiring temporary interruption of VKA therapy. Main Outcomes and Measures: Primary outcomes were 30-day postoperative ATE and major bleeding. Secondary outcomes included clinically relevant nonmajor bleeding (CRNMB) and all-cause mortality. Major bleeding and CRNMB were defined according to the International Society on Thrombosis and Hemostasis. Results: The cohort included 373 patients (median [IQR] age, 67 [60-73] years; 217 [58.2%] male) contributing 613 interruptions. Therapeutic-dose bridging was used preoperatively in 516 interruptions (84.2%) and postoperatively in 193 (31.5%) (99 of 215 [46.0%] in mitral or dual MHV vs 94 of 398 [23.6%] in aortic MHV interruptions). Patients with mitral or dual (vs aortic) valve position and prior thromboembolism had higher estimated odds of receiving postoperative therapeutic-dose bridging (mitral or dual vs aortic: adjusted odds ratio [aOR], 2.90; 95% CI, 1.91-4.41; prior thromboembolism: aOR, 1.91; 95% CI, 1.05-3.46). Estimated 30-day risks of ATE and major bleeding were 1.5% (95% CI, 0.8%-2.8%) and 2.1% (95% CI, 1.2%-3.6%), respectively. Clinically relevant bleeding (composite of major bleeding and CRNMB) occurred in 3.9% (95% CI, 2.6%-5.8%) of interruptions. Three deaths occurred (0.5%; 95% CI, 0.2%-1.5%), 1 attributed to fatal ischemic stroke. Omission of postoperative bridging was associated with higher estimated ATE risk (4.9% vs 0.9%; subdistribution hazard ratio [sHR], 5.30; 95% CI, 1.45-19.40; P = .01); this finding was no longer statistically significant in landmark analysis (sHR, 3.26; 95% CI, 0.62-17.21). Conclusions and Relevance: In this retrospective cohort study, patients with MHVs experienced clinically meaningful risks of perioperative ATE and bleeding. These data do not establish a causal protective effect of postoperative bridging. However, omitting bridging in patients with MHVs was associated with a higher estimated risk of ATE in exploratory analyses and should be approached cautiously, pending stronger evidence from representative cohorts.