Patients with end-stage pulmonary emphysema have limited therapeutic options. Lung volume reduction (LVR) and lung transplantation (LuTX) have been shown to improve lung function, quality of life (QOL) and survival, despite an associated perioperative risk (1, 2). Lung volume reduction surgery (LVRS) has gained popularity in the nineties, however, remained undervalued and underused after the large NETT trial (3–5). In more recent years, endoscopic lung volume reduction (ELVR) techniques by implantation of valves or coils, thermal ablation, or instillation of chemical sealants emerged (6). However, the effect of all LVR procedures is counteracted by the progression of disease usually leading to a decay in lung function after several months to years following treatment, leaving LuTX as the last option (7). Due to the protracted clinical course of COPD in comparison to other end-stage lung diseases, the best timing for LuTX referral remains debated. Guidelines for LuTX selection recommend the use of the BODE score, which is a good predictor for pre-LuTX mortality and post-LuTX survival benefit (8).

With increased use of LVR in highly impaired “low FEV1” patients (9, 10), and the more apparent overlap of patients eligible for both LVR and LuTX, the use of LVR procedures as a “bridge to transplant” has gained acceptance (7).

Simultaneous referral of patients for both LVR and LuTX is always recommended, and the decision should be taken in an interdisciplinary emphysema board with access to all treatment options (8). Those therapies are not mutually exclusive, and most combinations have been reported, of which, in most cases, LuTX was considered the last resort when all other previous therapies failed.

In a previous review, twelve published reports on LVRS and one on ELVR preceding LuTX were identified (11). North American papers showed that LVRS before LuTX can negatively affect survival (12, 13). Other publications demonstrated only an increased perioperative risk with no impact on survival (14–16). Nevertheless, these conclusions were not supported by the most recent and largest single institution report, in which no adverse effect of a previous LVRS was observed (17).

The increased surgical risk after LVRS can potentially be attributed to a higher occurrence of adhesions and thereby a longer operation time, more bleeding complications, and a higher need for blood transfusions. Also, a higher risk of injury to the phrenic nerve during adhesiolysis has been hypothesized but was never confirmed in the available reports (11).

For ELVR prior to LuTX, available data is even more scarce, with only a single institutional analysis available (18). In 20 ELVR patients, outcome was comparable to a matched control group, although ELVR was associated with a higher occurrence of bacterial airway colonization. This observation has been recently confirmed in a larger cohort outside the scope of LuTX (19).

Given the rarity and controversy of available evidence on the impact of LVRS and ELVR on a later lung LuTX, this study aimed to further determine the short- and long-term outcomes in these patients.

We conducted a multicenter retrospective analysis of post-LuTX outcomes in patients suffering from emphysema. Data was collected and anonymized before transmission between participating centers. This study was approved by the ethics committee of the Medical Faculty of the University of Duisburg-Essen (21-9856-BO).

615 patients [320 (52%) female and 295 (48%) male] were included in this study. Of those, 26 (4.2%) underwent LVRS before LuTX and 60 (9.8%) had ELVR prior to transplantation. Mean age was 58 ± 5.9 years. Indications for LuTX were COPD in 572 (93%) cases and alpha-1 antitrypsin deficiency emphysema in 43 (7%) cases. In 24 (92.3%) LVRS patients, surgery took place before listing (x̃: 3.8 years; range: 0.6–32.7) whereas in two cases (7.7%), LVRS was performed while patients were on the waitlist for LuTX (2.2 and 4.4 years after listing). Median time from LVRS to LuTX was 4.0 (1.1-32.7) years. LVRS was either unilateral (15; 58%) or bilateral (11; 42%). Surgical access for LVRS was VATS in 11 (44%) cases and open surgery via thoracotomy or sternotomy in 14 (56%) cases. In one patient this information was missing. Most LVRS were performed by parenchymal stapling (n = 22; 85%). The remaining 4 patients underwent lobectomy (15%).

Out of 60 patients with previous ELVR, 54 (90%) had the intervention before being listed for LuTX (x̃: 2.2 years range: 0.5–6.0) and 6 (10%) on the waiting list (x̃: 2.5 years; range 0.4–6.8 after listing). Time from ELVR to LuTX was 2.7 (0.02–7.4) years. 50 (83.3%) patients had unilateral interventions and 10 (16.7%) bilateral. The procedures were: valves: 50 (83.3%), coils: 9 (15%) and a combination of both in 1 (1.7%) case. One of the patients treated with valves had hydrogel foam instilled in the contralateral apical lobe. In 20 out of 51 (39%) patients with valves, later re-intervention became necessary to either reposition or remove the valves because of unsuccessful treatment.

All 26 patients with a history of LVRS (T_LVRS) were matched to 328 weighted controls (C_LVRS). In patients with ELVR, 56 patents (T_ELVR) remained in the analysis (n = 4 unmatched by lack of partners) and 270 weighted controls (C_ELVR). Metrics on matching are presented in the supplementary file (Table 1).

Pre-LuTX characteristics of treatment groups and matched controls are presented in Table 1. After coarsened exact matching, groups were balanced throughout demographic variables. Patients undergoing LuTX after previous LVRS had a lower median LAS at time of LuTX (32.4 vs. 33.4; p = 0.038) and lesser need for oxygen (2 vs. 3 L/min; p = 0.033); however, this was not considered of clinical relevance.

14.5% of all allocated grafts were size reduced during LuTX to match the recipient chest. Size reduction was performed either by wedge resection (n = 83; 13.5%) or by lobar transplantation (n = 6; 1.0%). 44.3% of patients had intraoperative extra corporeal support (ECS), either by ECMO (42.7%) or to a lesser extent by means of cardiopulmonary bypass (1.6%). The use of size reduction or ECS was comparable between groups. More detailed intraoperative data of both treatment groups and their weighted controls are presented in Table 2.

Verify that all the equations and special characters are displayed correctly.In patients with previous LVRS (T_LVRS) duration of LuTX (348 vs. 323 min; p = 0.296), as well as total ischemic time (TIT) of both donor lungs (465 vs. 403 min; p = 0.101) was statistically comparable to controls. Patients with previous ELVR did not exhibit significant differences in transplant duration and ischemic times, compared to their matched controls (283 vs. 288 min; p = 0.703 and 360 vs. 380 min; p = 0.180).

Short-term perioperative results (reoperation rates, intubation time, time on ICU and time in hospital) were excellent throughout treatment groups (T_LVRS and T_ELVR). When compared to controls no clinically relevant differences were observed (Table 3). Nevertheless, in patients with previous LVRS a significantly higher rate of post-LuTX ECMO use was recorded (15.4% vs. 3.8%; p = 0.006).

In T_LVRS complications occurred in 8 (30.8%) and in T_ELVR in 13 (23.2%) patients. Slight differences in the spectrum of complications were identified as LVRS patients had a higher occurrence of post-operative empyema (n = 2; 7.7% vs. 1.2%; p = 0.014). On the other hand, ELVR patients had a higher rate of wound infections (n = 5; 8.9% vs. 2.5%; p = 0.018). In deceased patients, graft failure was reported more often as cause of death in both treatment groups compared to controls (LVRS: 16.7% vs. 1.9; p = 0.044 | ELVR: 16.7 vs. 1.5%; p = 0.034). All recorded complications and causes of death are presented in Table 4.

In patients with previous LVR treatment, microbiologic colonization or infection was more often detected and the spectrum of positive cultures differed significantly (LVRS: 42.3% vs. 31.1%; p = 0.009 | ELVR: 39.3% vs. 32.4%; p = 0.01) as shown in Table 5. After LuTX and immunosuppression, the rate of positive cultures increased (LVRS: 54.5% vs. 72.8%; p = 0.005 | ELVR: 76.9% vs. 76.8%; p = 0.021) and as expected certain species became more prevalent as the microbiome changed (enterococcus spp., yeasts, mycobacteria, aspergillus; Table 6).

Short and long-term survival after LuTX was excellent across groups and controls and consistently comparable with the control groups (1-/5-year: LVRS: 92.3%/77.1%/p = 0.583|ELVR: 98.3%/91.0%/p = 0.174). Median survival after LuTX was: LVRS: 9.95 years and ELVR: 7.56 years. Median survival for controls was not calculated as more than 50% of those patients were still alive at time of analysis (Figure 1).

This multi-center matched retrospective cohort study, assessed post-operative outcomes of emphysema patients, who underwent surgical or endoscopic LVR prior to LuTX. After matching, baseline indicators and intraoperative modalities were comparable between groups and controls. The analysis showed that both previous LVRS and ELVR were associated with a different spectrum of bacterial colonization prior to LuTX which must be considered to prevent infectious complications. LVR did not impact short- and long-term survival, which was equally good in all groups.

Interestingly we observed that patients with previous LVRS had a higher rate of post-operative ECMO need after LuTX. This might be explained by a longer and more difficult preparation due to pleural adhesions. Hence, potentially leading to longer surgery durations/ischemic times and an increased need for blood transfusions, all known risk factors for early mortality and PGD (22–24). Three cases of PGD 3 within 72 h occurred in the LVRS group. In previous publications, an association of previous LVRS with a higher pulmonary arterial pressure and an increased risk of phrenic nerve palsy was postulated (13, 15). These assumptions are not corroborated by our data as median PAP was 32 mmHg and not a single case of phrenic nerve injury was observed in the LVRS group. Additionally, it is noteworthy that in this cohort, the outcome in LVRS patients was statistically comparable to the outcome in patients with previous non-LVRS intrathoracic surgery.

To the best of our knowledge, this study is the largest published series about ELVR prior to LuTX. Those patients had equally good perioperative outcomes as controls although a higher occurrence of wound infections (8.9% vs. 2.5%; p = 0.018) was observed. Out of those five patients, four had positive sputum cultures (Candida/Aspergillus/Klebsiella) before LuTX. After LuTX, these patients suffered also from empyema n = 3 and pneumonia n = 1. Although the association between ELVR and post-LuTX wound infection is not fully understood, we hypothesize that a different spectrum of pre-transplant colonization and an increased exposure to antibiotics might make them more susceptible to hospital acquired infections and multi-drug resistant bacteria.

This study is the first to assess extensively airway colonization of LuTX recipients who underwent previous LVRS or ELVR. It showed that LVR was associated with a distinct airway microbiome both before and after LuTX. In comparison to other LuTX-indications COPD has a lower risk of bacterial infections (25). It was unexpected that the number of colonized patients were as high even before LuTX (LVRS: 42.3%; ELVR: 39.3%). Unfortunately, data on colonization after LuTX was only available for one of the three participating centers. In this subgroup, colonization rates of 54% in LVRS patients and 76.9% in ELVR patients were seen after LuTX.

In LVRS patients a higher rate of slow growing mycobacteria was observed after LuTX in comparison to the control group (27.3% vs. 2.5%; p < 0.001), an observation which cannot be readily explained. On the other hand, ELVR has been previously associated with pathological colonization as implanted valves and coils impede mucus clearance (26). Although a predominance of stenotrophomonas maltophilia (40%) after ELVR and LuTX has been described (18), this relationship could not be confirmed by our data in which only 7.7% patients presented with s.maltophilia after LuTX.

The strength of this study and its conclusion is given by its design. Three high volume centers experienced in both LVR and LuTX provided data on all patients recently transplanted for emphysema at their institution. To correct for selection bias and differences in patient characteristics, LVRS and ELVR patients were matched to weighted control groups. This led to highly balanced groups.

The observations of this study are mostly in line with a recent single-center analysis comparing 52 LVRS+LuTX patients to 65 unmatched controls (17). However, our findings differed markedly from those of a recent UNOS-database analysis (12) which included 106 LVRS+LuTX patients (from 37 LuTX centers), propensity matched to 106 controls without previous intrathoracic surgery (from 67 LuTX centers). This UNOS analysis identified a significantly increased risk of death (HR: 1.72; CI: 1.13-2.6; p = 0.01) after LVRS+LuTX, which was surprisingly not associated with the total number of LuTX (HR: 0.99) or the total number of LVRS+LuTX (HR: 0.99) of individual centers. Furthermore, the observed median survival was significantly worse in LVRS patients in comparison to matched controls (3.4 vs. 6.5; p = 0.038).

The present study has several limitations. First, no donor specific characteristics apart from ischemic times of the donor lungs were taken into consideration. Secondly, there is substantial heterogeneity in how LVRS was performed. LVRS nowadays is routinely performed by a bilateral video-thoracoscopic approach and we can hypothesize that such a minimally invasive approach would have a lesser associated risk in a latter LuTX. In this cohort, sternotomy, thoracotomy, VATS, pleurectomy, pleurodesis and pleural tenting were in use and inevitably lead to pleural adhesions, albeit to different extent. By the sample size of 26 LVRS patients, this cannot be sufficiently considered. A similar limitation applies to different ELVR approaches (valves, coils, foam, vapor) having a different risk profile (pneumonia, exacerbation and pneumothorax) (27) and suggesting that their impact on a later LuTX may differ. A possible observer bias must be addressed with regards to the microbiological cultures. Although recipient bronchi were all sampled during LuTX, patients who underwent previous ELVR had supposedly more bronchoscopies and therefore more samples taken before transplant.

Most patients had LVR before being listed for transplantation (n = 78). “Bridging to LuTX” only took place in 8 patients who were already on the waiting list. Waiting time for LuTX was comparable throughout groups as presented in Table 1 (LVRS 180 days vs. 203 days in controls; ELVR 156 vs. 176 days in controls).

Additionally, critically ill patients who underwent LVR and were not later referred to LuTX (because of improvement, complications, or clinical misjudgment) were not considered. Hence, this study cannot predict the impact of LVR as an alternative to LuTX. It did not account for functional improvements while waiting, nor for changes in LAS scores and impact on waiting times. The crucial question about QOL, functional/survival benefits, and timing of LVR before LuTX cannot be answered and the authors recommend further prospective investigation to answer it.

This study clearly demonstrates that patients who underwent previous surgical or endoscopic LVR can safely be considered for later LuTX. Although a marginally increased risk of specific complications and differences in airway colonization after LuTX were observed, short- and long-term survival was very good.