Chek Kuan Lam, Ka Lai Cally Ho, Oswald Joseph On-Jing Lee, Shing Lung Wong, Yue Fan, Ka-Lam Wong
Background: Perioperative anticoagulation management for left ventricular assist device (LVAD) recipients undergoing elective non-cardiac surgery (NCS) remains uncertain, particularly in Asian populations. The safety of direct warfarin interruption without heparin bridging in procedures of different bleeding risks is a key knowledge gap. This study aimed to quantify 30-day thrombotic and bleeding outcomes after elective NCS in a predominantly Chinese LVAD cohort and identify predictors of perioperative bleeding. Methods: -tests and Chi-squared/Fisher's exact tests. Results: Fifty-five patients (96.4% Chinese; 94.5% male) underwent 85 procedures [HeartMate II (HMII) 12.7%, HeartMate 3 (HM3) 76.4%, HeartWare 10.9%]; mean 2.52±2.0 years from LVAD implantation to NCS. Warfarin was interrupted without bridging in 84.7%, interrupted with bridging in 7.1%, and continued in 8.2%. Thrombotic events occurred in 1/85 (1.2%). Any bleeding (BARC ≥1) occurred in 10.6%; major bleeding (BARC ≥3a) in 8.2% (3.5% BARC 3a; 4.7% BARC 3b). Thirty-day mortality was 1.2% and readmission 8.2%. Predictors of bleeding included higher creatinine (P<0.001), lower haemoglobin (P=0.02), higher alanine aminotransferase (ALT) (P=0.03), angiotensin-converting enzyme inhibitor/angiotensin receptor blocker/angiotensin receptor-neprilysin inhibitor (ACEi/ARB/ARNI) use (P=0.02), and heart failure etiology (P=0.03). Warfarin continuation (P=0.56), LMWH bridging (P=0.12), LVAD type (P=0.48), and procedural bleeding risk (>0.9) were not associated with bleeding. Conclusions: Elective NCS in LVAD recipients under locally lower international normalized ratio (INR) targets showed low thrombotic and acceptable bleeding rates. Direct warfarin interruption without bridging, predominantly for minor/low-risk procedures, appeared safe. Renal dysfunction, anemia, elevated ALT, ACEi/ARB/ARNI use, and heart failure etiology may guide individualized perioperative strategies.