Clinical Outcomes of the Lund, Toronto, and Organ Care System Protocols for Normothermic Ex Vivo Lung Perfusion: A Systematic Review and Pursuit of Network Meta-Analysis

Purpose Ex Vivo Lung Perfusion (EVLP) is an excellent method to preserve and recondition donor lungs, thus increasing the lung donor pool. This systematic review summarizes studies comparing both primary graft dysfunction (PGD) and survival between EVLP and conventional lung transplantation (c-LTx), or between different EVLP protocols. The objective is to evaluate outcomes of the three main EVLP protocols: ‘Lund’, ‘Toronto’ and ‘Organ Care System’ (OCS), through a network meta-analysis (NMA). Methods A systematic search was performed in Pubmed, Embase, the Cochrane library and Web of Science. Studies performing EVLP on initially discarded donor lungs were independently assessed for eligibility by two researchers. Data on study design, participants, and EVLP methods were extracted, together with outcomes such as survival and PGD. The generated database was assessed on meeting the assumptions for NMA. Results Nine studies were included: two non-randomised clinical trials (n = 240), and four prospective (n = 1360) and three retrospective observational cohort studies (n = 540). The study protocols could be divided into: Toronto (5/9), Lund (3/9) and OCS (1/9). Most frequently reported outcomes were PGD grade 3 at 72 hours post-LTx, and one-month and one-year survival. All included studies compared EVLP to c-LTx, providing solely indirect evidence for comparison between protocols and thus no loops within the NMA model. In the absence of loops, consistency within the model cannot be statistically assessed. Thus, thorough examination for differences that could modify the treatment effect is required. Substantial variations were found in study design, type of LTx, and participant characteristics. It was therefore not feasible to construct a subset of studies that met the similarity and consistency assumptions of NMA. Conclusion Currently, no substantiated statistical comparison can be made between different EVLP protocols. Further attempts at NMA are dependent on the introduction of studies comparing multiple EVLP protocols, or an increase in randomised controlled trials performing EVLP on standard donor lungs. Alternatively, collaborative data management between EVLP centres would be a promising way to explore the effect of different protocols, and their individual components, on patient outcomes.