The Conversional Efficacy of Ex Vivo Lung Perfusion and Clinical Outcomes in Patients Undergoing Transplantation of Donor Lungs by Ex Vivo Lung Perfusion: A Meta-Analysis

Background

Lung transplantation (LTx) is an effective treatment for advanced lung disease, which improves patients’ quality of life and extends their life expectancy [1]. However, a profound shortage of donors and underutilization of donor lungs remains a significant challenge in performing LTx [2,3]. In addition to the use of marginal donors and living donors to expand the donor organ pool, ex vivo lung perfusion (EVLP) technology can reduce receptor waiting list mortality by improving donor lung utilization and increasing LTx activity [4,5].

EVLP is a relatively new technology for the procurement of donor lungs which was initially developed as a method for assessing graft quality and improving cardiac death (DCD) donor lung function [6,7]. The first successful use of EVLP to assess and recondition LTx in donor lungs was reported by Steen et al. in 2001, which was the starting point for the “Lund protocol” [7]. In 2008, Cypel et al. [8] in Toronto reported the use of a novel strategy to expand the EVLP assessment of lung function, which laid the foundation for the “Toronto protocol”. In 2012, Warnecke et al. in Hanover reported the first-in-human experience using the portable Organ Care System (OCS) lung device for concomitant preservation, assessment, and transport of donor lungs, which was known as “OCS protocol” [9]. EVLP has evolved to demonstrate that marginal donor lungs could be assessed and treated to achieve similar early outcomes as the standard criteria donor lungs [5]. In addition, due to the process of EVLP not being regarded as “ischemic time”, EVLP might play an essential role in expanding the procurement time and contributing to the long-distance transportation, especially using the portable OCS technique [9]. EVLP technology has attracted more and more attention from transplant centers around the world, but there are still serious concerns about the poor results after transplantation. Although several comparative analyses of clinical outcome between EVLP and traditional cold storage have been reported and some multicenter randomized control clinical trials (RCTs) are being conducted, there are still a lot of uncertainty about EVLP clinical application. Thus, this meta-analysis was performed to determine the short- mid- to long-term results of EVLP compared with that of standard cold storage.

Material and Methods

LITERATURE SEARCH STRATEGY:

The meta-analysis was performed according to the recommendations of the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement [10]. The electronic databases, including PubMed, PMC, EMBASE, and Ovid, were comprehensively searched for relevant articles published until March 1, 2019. Search terms included the following: “EVLP or ex vivo lung perfusion” and “lung transplantation”. All references reported in the identified articles were also scanned to identify potentially relevant reports.

CRITERIA FOR INCLUSION AND EXCLUSION:

The studies included in this meta-analysis need to meet the following criteria: 1) RCTs or cohort studies studying lung transplantation; and 2) studies comparatively analyze the post-transplantation results between EVLP technique and traditional cold storage. The following studies were excluded: 1) articles about animals; 2) single-arm analysis about EVLP technique; and 3) review articles without original data. For duplicate articles reporting the same case population, only the most complete or up-to-date one was included. Two reviewers independently selected eligible studies. Disagreements were settled by discussion.

DATA EXTRACTION AND QUALITY ASSESSMENT:

The following data was collected independently by 2 reviewers using a predesigned form, comprising first author, publication year, study period, country, study design, sample size (donors and recipients), age (donors and recipients), gender (donors and recipients), type of donor, time in ventilation (donors and recipients), chest x-ray abnormalities, PaO2/FiO2 100% (P/F, donors and recipients), indication for LTx, type of LTx, lung allocation score (LAS), extracorporeal membrane oxygenation (ECMO) bridge to LTx, intraoperative extracorporeal circulation (ECC)/ECMO, reason for EVLP, technological type of EVLP, EVLP solution, EVLP duration, the number of accepted donor after EVLP, severe primary graft dysfunction (PGD) after LTx, post-LTx ECMO, residence time in the intensive care unit (ICU), total length of hospital stay, FEV1 of the predicted value (FEV1%), FVC of the predicted value (FVC%), follow-up time and survival data after LTx. The hazard ratio (HR) and 95% confidence intervals (CI) of the EVLP group compared to the traditional cold storage group for OS were primarily collected. If the HR and 95% CI were not explicitly provided, we used Tierney’s methods to extract survival data from the original study data or Kaplan-Meier curve [11]. If the aforementioned items were not reported in the original study, the items were labeled as “not available (NA)”. Inconsistencies in the process were solved by consultations.

The quality of the included cohort studies was independently assessed by 2 reviewers according to the Newcastle-Ottawa Scale (NOS) [12]. The NOS evaluated a study with the score ranged from 0 to 9. A study with a score of 6 or high was considered as a high-quality one. The quality of RCT reports was measured by the Jadad scale [13]. The Jadad scale evaluated a study from 3 perspectives, including randomization, blinding, and withdraw, with scores ranging from 0 to 5. A study achieving a score of 3 or more was identified as a high-quality one.

STATISTICAL ANALYSIS:

Interest results included improvement of P/F ratio in donor lung after EVLP, total cold ischemic time (CIT) and preservation time in donor lung, P/F ratio after LTx, extubation time, severe incidence of PGD after LTx, requirement of ECMO, residence time in the intensive care unit (ICU), total length of hospital stay, FEV1%, FVC%, survival rate at 30 days, 90 days and 1-year after LTx, and accumulative survival. The HR and 95% CI were used as an effect size (ES) to assess the impact of EVLP on accumulative survival outcomes. Relative risk (RR) with its 95% CI and Mantel-Haenszel model were used to measure the effect of EVLP on categorical variables. For continuous outcomes, weighted mean difference (WMD) was used as the ES to assess the difference between the EVLP group and the traditional group. Heterogeneity across the studies was tested using I-squared statistics [14]. I2≤50% indicated no or moderate heterogeneity, in which case a fixed-effect model was used. I2>50% showed statistically significant heterogeneity, in which case a random-effect model is chosen. To explore the difference among Lund, Toronto, and OCS protocols, subgroup analyses based on different protocols were adopted. Sensitivity analysis by omitting a single study to confirm the robustness of the combined results. By convention, an observed ES>1 implied a more unsatisfactory outcome for the EVLP group compared with the traditional cold storage group. Assessment of potential publication bias was conducted through Begg’s funnel plot and Egger’s test. Data are presented as mean±standard deviations (SD), median (ranges), or median (inter-quartile range, IQR). Stata software version 12.0 (Stata Corporation, TX, USA) was used in the meta-analysis. All the tests were 2-sided, and P<0.05 was considered statistically significant.

Results

STUDY SELECTION AND CHARACTERISTICS:

The results of articles selection are shown in Figure 1. A total of 789 articles were identified initially using the search strategy described. We excluded 738 articles because they were duplicate documents, review articles, or irrelevant studies. Afterward, 51 articles were read in full text. Finally, 20 articles were considered suitable for inclusion in the meta-analysis. Among the eligible study, the NOVEL trial [15] is expected to end in 2020, and its 1-year results have been reported in summary form. Considering that the NOVEL trial is an RCT study and can provide some available information, such as necessary patient information, donor lung conversion rate, 30-day survival, and 1-year survival, it was included in our study.

Table 1 listed the main features of the 20 eligible articles [4,15–33] Three RCTs, 3 prospective cohort studies, and 14 retrospective cohort studies were included, and the publication year ranged from 2011 to 2018. The study included a total of 2574 donors and 2567 recipients. Table 1 and Figure 2 show the features of the donor, and Table 2 and Figure 3 show the characteristics of the recipient. There was no significant difference in the donors’ age (Figure 2A), gender (Figure 2B), donor type (Figure 2C), and mechanical ventilation time (Figure 2D) between the EVLP group and the traditional cold storage group. However, compared with the non-EVLP group, the EVLP group donors had more chest X-ray abnormalities (RR 1.39, 95% CI 1.03–1.87, P<0.05, Figure 2E) and more inferior P/F ratio (WMD −106.06, 95% CI −150.78–61.33 mmHg, P<0.001, Figure 2F). There was also no significant difference for recipients’ age (Figure 3A), LAS (Figure 3C), mechanical ventilation pre-LTx (Figure 3D), ECMO bridging to LTx (Figure 3E), type of LTx (Figure 3F), or total CIT (Figure 4A). Still the EVLP group had more female patients (Figure 3B) and showed more intraoperative ECC/ECMO needs (RR 1.34, 95% CI 1.01–1.78, P<0.05, Figure 3G) and longer preservation time (WMD 379.54, 95% CI 271.16–487.91 minute, P<0.001, Figure 4B) compared with the traditional cold storage group. In the subgroup analysis based on different protocols, the OCS subgroup exhibited equivalence between the 2 groups but the shorter total CIT in the OCS-EVLP group (Figure 4A).

THE EFFICACY OF EVLP IN IMPROVING DONOR LUNGS:

The parameters of EVLP and its role in conversing marginal donor lungs are summarized in Table 3 and Figure 5. Compared with the P/F pre-EVLP, the P/F after EVLP was significantly improved (WMD 184.38, 95% CI 130.17–238.59 mmHg, P<0.001, Figure 5). However, the OCS subgroup did not show significant improvement in P/F (Figure 5), which might be because 1 study in the OCS subgroup involved only standard criteria donors [16]. The conversion rate of donor lungs by EVLP ranged from 34% to 100%. Among those included studies, a total of 1985 cases received traditional cold storage LTx, and 582 cases received EVLP LTx, so it can be said that EVLP made a 29.3% contribution to the LTx activity.

THE EFFECT OF EVLP ON OUTCOMES OF RECIPIENTS:

As shown in Figures 6 and 7, there was no significant difference about P/F after LTx (Figure 6A), time to extubation (Figure 6B), postoperative ECMO requirement (Figure 6D), length of hospital stays (Figure 6F), FEV1% (Figure 6G), FVC% (Figure 6H), survival rate at 30 days (Figure 7A), 90 days (Figure 7B) and 1 year (Figure 7C) after LTx, and accumulative survival after LTx (Figure 7D) between the EVLP group and the non-EVLP group. However, compared with the non-EVLP group, the EVLP group showed a lower incidence of PGD 3 (Figure 6C) after LTx, but the longer length of ICU stays (Figure 6E).

SENSITIVITY ANALYSES AND PUBLICATION BIAS:

The corresponding pooled ESs did not alter significantly during the sensitivity analysis process, suggesting robustness of the results. Publication bias was tested using Begg’s funnel plot and Egger’s test. No significant publication bias was observed in either the visualization of the funnel plot (Figure 8, P=0.381) or Egger’s test (P=0.272).

Discussion

EVLP as a new technique not only could preserve donor lungs but also could be used to assess and recondition/improve the borderline lungs, so it has a great potential to replace the standard cold storage in the procurement of donor lungs. However, synthetic comparative analysis of EVLP technique and standard cold storage in LTx is limited, especially for low-quality donor lungs. This present meta-analysis systematically evaluated the impact of EVLP on LTx outcomes compared with standard cold storage. In the 2 RCTs included, donors in the EVLP group were standard criteria donors [16,17]. Still in other studies, donors in the EVLP group were expanded criteria donors, marginal donors, or initially rejected donors. Combined analyses about donor features showed that the EVLP group had more chest x-ray abnormalities and a poorer P/F ratio than the traditional cold storage group. After the process of EVLP, the poor P/F ratio in the EVLP group was significantly improved with the conversion rate of marginal/rejected donor lungs ranging from 34% to 100%, which promoted the LTx growth by about 29.3%. Luc et al. [18] and Machuca et al. [19] only involved DCD donors and reported the comparison between DCD lungs that underwent EVLP and those transplanted without the use of EVLP, which indicated that EVLP could improve the utilization of extended criteria DCD lungs.

The preservation time in donor lung was much longer in the EVLP group than that in the traditional cold storage group, especially in the Toronto and Lund subgroups. Although the total CIT was similar between the EVLP group and the traditional cold storage group, Toronto and Lund subgroups exhibited longer total CIT in the EVLP group, and the OCS subgroup exhibited shorter total CIT in the EVLP group. Thus, the extra part of preservation time in the EVLP group consisted primarily of the EVLP process, and a longer total CIT. The wide gap between WMD of preservation time (379.54 minutes) and WMD of total CIT (73.28 minutes) could be more approximate to the duration of EVLP, which indicated that EVLP could play an essential role for the expansion of the procurement time [16]. In addition, the OCS protocol based portable EVLP device may allow a significantly shorter CIT and more extended distance transport for donor lungs [16].

The clinical-pathologic features of the recipients between the EVLP and the non-EVLP groups were equivalent, except the EVLP group had more female composition and required more intraoperative ECC/ECMO than the non-EVLP group. There were no significant differences between the 2 groups in using mechanical ventilation/ECMO support after LTx. PGD was graded based on the International Society for Heart and Lung Transplantation (ISHLT) criteria, with grade 3 representing P/F ratio <200 within 72 hours and radiographic infiltrates [34,35]. The EVLP recipients had less incidence of PGD3 throughout the initial 72 hours after LTx than the non-EVLP recipients. However, the length of ICU stays of the EVLP group was longer than the non-EVLP group. This may be probably because the OCS subgroup contributed less incidence of PGD3 (Figure 6C), and the Lund subgroup donated more extended ICU stays (Figure 6E). The peak pulmonary function (FEV1% and FVC%) after LTx, and the short-to long-term survival outcomes were all similar between the 2 groups.

Despite our efforts to conduct a comprehensive analysis, there are still some limitations that need to be recognized. First, most of the included studies in our analysis were retrospective cohort studies that provided only weaker statistical power. Second, some studies have shown a relatively small number of patients, which may affect the validity of the statistics. Third, several ESs and its 95% CI were calculated by extracting the data from Kaplan-Meier curves, which might bring statistical deviations inevitably. Finally, donor/recipient characteristics, EVLP processes, and follow-up showed significant heterogeneity. Although random-effect models, subgroup analyses, and sensitivity analyses were performed to address this heterogeneity, these results should still be interpreted with caution.

Conclusions

In our study, EVLP can be used to assess and improve the quality of high-risk donor lungs to expand lung supply and improve donor lung utilization. Additionally, the application of EVLP is non-inferior to standard cold storage regarding postoperative outcomes. Considering an RCT designed for improving low-quality donor lung with EVLP might be problematic from an ethical point, this study can be a rationale for further work.