Implications of ECMO Bridging and Salvage Strategies on Mortality and PGD

Academic Article


  • PURPOSE: ECMO is increasingly used for bridging and salvage therapies in lung transplantation, with favorable short-term results. Mid- to long-term outcomes of ECMO use for bridging and salvage are incompletely understood, as is effect on PGD risk; therefore, we sought to define the effects of these strategies on mortality and PGD. METHODS: The Lung Transplant Outcomes Group study is a multi-center prospective cohort study designed to identify risk factors for PGD and mortality. Patients were enrolled from August 2011 to June 2018. The effect of pre-operative (bridging) and post-operative (salvage) ECMO on Grade 3 PGD incidence and long-term survival were assessed using time-to-event analyses adjusted for center. RESULTS: 1,537 subjects were enrolled with a 17.0% incidence of PGD on day 3 and an 18.9% mortality overall. PGD was associated with mortality (p=0.0001, Fig. 1a). ECMO bridging and salvage strategies increased over time (Fig. 1b). 138 (9%) subjects had bridging ECMO and 275 (17.9%) had salvage ECMO. PGD was associated with bridging strategies (p=0.001). PGD incidence in bridged patients (50.7%) exceeded non-bridged (42.2%). Bridging was not associated with mortality overall (p=0.70, Fig. 1c); however, if PGD developed after transplant using a bridging strategy, there was a significant increase in mortality (HR 1.8 [1.3; 2.61]; p=0.0007). About 45% of patients bridged with ECMO required salvage after transplant. ECMO use for salvage was highly associated with mortality (HR 2.1 [1.6; 2.8]; p<0.0001, Fig. 1d). CONCLUSION: Use of the ISHLT Revised definition identifies Grade 3 PGD as highly associated with mortality in the ECMO bridging and salvage era. ECMO use is increasing for both bridging and salvage strategies. ECMO bridging strategies do not seem to increase mortality in lung transplant patients unless Grade 3 PGD subsequently develops. ECMO salvage after transplant is associated with increased risk of mortality; however, it is likely used only in the severest cases, thus indication bias is likely present.
  • Authors

    Digital Object Identifier (doi)

    Author List

  • Cantu E; Bermudez C; Cevasco M; Suzuki Y; Buckley T; Galati V; Majeti N; Benvenuto L; Anderson M; Wille KM
  • Start Page

  • S103
  • End Page

  • S104
  • Volume

  • 39
  • Issue

  • 4