Role of Extracorporeal Life Support in Bridging Patients to Pulmonary Transplantation

Wenjun Mao,Weiya Xia,Jingyu Chen

DOI: https://doi.org/10.1097/tp.0b013e31825ba218

2012-01-01

Transplantation

Abstract:We read with great interest the recently published article in Transplantation by Lang et al. (1) regarding the utility of extracorporeal membrane oxygenation (ECMO) as a bridging strategy to primary pulmonary transplantation. They studied 38 patients undergoing ECMO as a bridge to transplantation with 26 of the 34 transplant patients discharged from the hospital and concluded that transplantation of patients after bridging with ECMO is reliable with acceptable outcome. Bridging patients with ECMO represents optimal therapy for circulatory and respiratory support when traditional therapy failed. The authors should be congratulated for their great effort. However, we would like to raise some questions that need clarifications. A total of 102 postoperative complications were reported in 88 patients bridged with extracorporeal life support (ECLS) in the article’s Table 4, wherein 13 were neurologic complications, which is quite different from that reported by Lang et al. (1) (13/102 vs. 17/44, two-sided χ2 test, P<0.01). It is intriguing that why profuse complications associated with neurology developed in 34 patients. The most frequent complications encountered during ECLS were hemorrhage, cannulation site complications, renal failure, sepsis, and neurologic complications (2, 3). After some thorough review, it is difficult to precisely explain the difference of complications between the anterior published articles and that of Lang et al. (1) indeed, although some potential risk factors may contribute to the disparity. Hartwig et al. (4) reported that patients on arteriovenous (V-A) ECLS showed significantly higher rates of neurologic complications when compared with venovenous ECLS; never theless, there are 12 (n=34, 35.3%) patients receiving V-A ECMO support to lung transplantation (1), whereas 35 (n=88, 39.8%) patients reported in Table 4 underwent transplantation after V-A ECMO bridge (two-sided χ2 test, P>0.05). As a consequence, modality is inadequate to determine the development of neurologic complications while on ECMO bridge, which may be largely ascribed to some confounding factors including patient management and selection, more critical illness, patient’s unwillingness to accept the invasive intervention, and lung transplantation per se. Moreover, preoperative ECMO had been proven to be a strong risk factor for mortality after transplantation in multivariable analysis (5). Thus, highly selected patients considered under ECLS bridge were likely to benefit from optimal support device when its support efficacy and adverse effects were taken into account. Twenty-six complications were found before donor lungs became available when patients in Table 4 were bridged with ECLS, which may have an adverse effect on outcome. However, no description was provided by Lang et al. (1) to clarify the adverse effects of patients who underwent transplantation while awaiting on the support device. In addition, the authors omitted the morbidity during the long-term follow-up period, which represented the quality of life in these recipients discharged from the hospital. The authors only mentioned causes of death of these patients, whereas the adverse effects are particularly helpful for the readers to judge ECLS under current medical circumstances and draw some conclusions. Bridging device mode and configuration should be individualized according to different clinical situations as reported by the authors (1). V-A ECMO is recommended when patients complicated with hypoxemic respiratory failure and hemodynamic compromise; venovenous ECMO is preferred while hypoxemia with PCO2 elevation occurs but with stable hemodynamics; and pumpless lung assist device Novalung becomes an useful strategy for isolated hypercapnemia, resulting in sufficient carbon dioxide removal (6). All the three modalities reflect invasive interventions that introduce different risks of complications to patients at disparate situations. The authors included in the retrospective study patients undergoing different bridging modalities to arrive at combined results, by which significant bias may be produced by confounding factors not accounted for. Namely, the combined analysis only represents the global results instead of separate result for each modality, respectively; therefore, the comparison by different modalities is essential to clarify the bridging efficacy objectively. In the statistical analysis, the authors used the two-sided chi-square test for the association of the clinical characteristics among three different groups (gender, diagnosis, and bridge type); however, the questions were raised because the choice of statistical method for categorical characteristics in Table 1 was confusing. From our perspective, Fisher exact test for the comparison of gender (fourfold table) is more suitable than two-sided chi-square test because the total frequency is less than 40 by calculation; moreover, the comparison of diagnosis and bridge type (2×4 table) should also be conducted with Fisher exact test, given the reason that some theoretical frequency was less than 1. These data analyzed by different statistical methods may turn to disparate results. Wenjun Mao Wei Xia Jingyu Chen Department of Cardiothoracic Surgery Wuxi People’s Hospital Nanjing Medical University Wuxi, China