医学
背景(考古学)
肺移植
单中心
肺
移植
随机对照试验
四分位数
外科
比例危险模型
队列
内科学
人口
灌注
存活率
临床试验
生存分析
人口统计学的
作者
Sara Sakowitz,Syed Shahyan Bakhtiyar,Yas Sanaiha,Deep Mehta,Richard Jay Shemin,Peyman Benharash
标识
DOI:10.1097/sla.0000000000007203
摘要
OBJECTIVE: Ex vivo lung perfusion (EVLP) has been championed as a disruptive strategy to expand the lung transplantation (LT) donor pool, yet contemporary evidence remains limited. Using a real-world cohort, we emulated a randomized controlled trial (RCT) to characterize the causal effect of EVLP on post-LT survival. METHODS: LT recipients ≥18 years were identified from the 2018-2025 Organ Procurement and Transplantation Network. Annual hospital LT volume was computed; centers in the top quartile (>54 LT/y) were designated high-volume centers (HVCs). Using target trial emulation with clone-censor-weight methodology, we simulated a pragmatic, open-label RCT comparing EVLP with standard allografts at HVCs and non-HVCs. Survival was evaluated using Kaplan-Meier and multivariable Cox models. RESULTS: Among 20,627 LTs, 1419 (7%) received EVLP allografts. The proportion of EVLP allografts increased across all centers, from 5% in 2018 to 10% in 2024 and 8% in 2025 (P trend <0.001). In the Lung Allocation Score era, LT using EVLP allografts yielded similar 2-year outcomes at HVC, but inferior outcomes at non-HVC. In the Composite Allocation Score period, however, EVLP was associated with significantly worse survival at non-HVC. In the emulated RCT among CAS-era recipients, EVLP remained independently associated with similar outcomes at HVC, but increased 2-year hazards of patient mortality (HR: 2.43, CI: 1.36-4.34, P=0.003) and allograft failure at non-HVC (HR: 2.52, CI: 1.31-4.84, P=0.005). CONCLUSIONS: EVLP was associated with comparable survival at HVC but inferior outcomes at non-HVC. These findings highlight the critical role of hospital context and center experience in determining the effectiveness and safety of EVLP and call for standardization of protocols across centers.
科研通智能强力驱动
Strongly Powered by AbleSci AI