Impact of donor, recipient and matching on survival after high emergency lung transplantation in France

Summary of the main results

Studying the impact of donor, recipient and matching on long-term outcome following HELT in France, we found that 1) the HELT procedure was associated with an increased instantaneous risk of death whatever the underlying diagnosis; 2) donor P_aO2, ABO mismatch and size mismatch differed between the HELT and standard lung transplantation groups; and 3) the adverse outcome associated with the HELT procedure was mainly related to the severity status of the recipients rather than donor or matching.

Waiting list and post-transplant outcome

The French lung allocation system is based on a rotational list that was historically associated with significant mortality during the waiting period, especially for idiopathic pulmonary fibrosis patients [15]. Introduced in 2007, the French HELT procedure has been designed to minimise mortality on the waiting list by rescuing the sickest patients through the allocation of any graft available on the national territory for 2 weeks, but its implications for graft utilisation and long-term outcome had not been anticipated. The French HELT allocation is indeed an effective strategy to reduce mortality on the waiting list as demonstrated in previous studies [16–18]. Waiting list data were not available for this study, but the 2017 Annual Report of the French OPO Agence de la Biomédecine showed a decrease of the cumulative incidence of death on the waiting list by taking into account the competing risk of transplantation from 1995 to 2017 (from 12% to 3% at 3 months post-registration and from 22% to 5% at 1 year post-registration) [19]. However, this cumulative incidence of death on the waiting list by taking into account the competing risk of transplantation was still higher in HELT patients than in standard lung transplantation patients (7% versus 3% at 3 months post-registration and 9% versus 6% at 1 year post-registration, respectively, between 2012 and 2017) [19].

Since the implementation of HELT in France, two multicentric studies showed a significantly lower survival in HELT compared with standard lung transplantation patients, even though survival rates after HELT were improving over time, from 51% at 2 years in 2012 (n=32 patients) to 59% at 3 years in 2014 (n=95 patients) [9, 16]. Unfortunately, these studies were based on retrospective questionnaires and were not exhaustive. Two unicentric studies focusing on certain diagnoses and a longer timespan found no significant difference between HELT and standard lung transplantation in CF patients, with 2-year survival rates of 73% and 75%, respectively (n=201; p=0.128), nor in pulmonary hypertension, with 5-year survival rates of 50% and 72%, respectively (n=234; p=0.053) [17, 18]. These studies were exhaustive, but might have lacked statistical power to allow adequate comparisons. Conversely, we found the HELT procedure to be associated with an increased instantaneous risk of death and a high rate of prolonged extracorporeal membrane oxygenation (ECMO) post-transplant, whatever the indication for lung transplantation. This adverse outcome is of particular concern given the relatively low proportion of mechanical ventilation/ECMO patients. A recent simulation study suggested that HELT should maintain a limited impact on the regular waiting list given that the graft supply is adequate [20]. Albeit beneficial to the sickest patients, the impact of HELT on long-term outcome is detrimental and its impact on lung utilisation should be assessed carefully.

Donor and matching

Whether the national priority could affect the quality of allocated grafts and matching was unclear before this study [18], with two hypotheses. On the one hand, as the lung pool is divided between 11 centres without HELT and directed toward one patient with HELT, the national priority could bias graft allocation toward offering better grafts to HELT compared with non-HELT patients (i.e. matching the best grafts to the sickest patients). On the other hand, the emergency situation might lead transplant physicians to decrease the threshold to accept a graft, because of the fear that no other graft will be available in the following days (i.e. matching the worst grafts to the sickest patients). Interestingly, we found that lung grafts accepted through the HELT procedure were similar to those accepted through the regular allocation procedure, except for a small difference in the mean P_aO2 before harvest. Donor P_aO2/F_IO2 ratio is known to be associated with early gas exchange and mortality after lung transplantation, and although it is debated, the threshold of 350 mmHg is commonly accepted [21]. The difference between HELT and standard lung transplantation grafts has therefore limited clinical implications, as the proportion of patients with P_aO2<350 mmHg remains similar between groups. Even if the donor score was significantly different between the HELT and standard lung transplantation groups, its prognostic impact remains limited.

There were more ABO mismatches in the HELT than in the standard lung transplantation group. However, ABO-compatible and ABO-identical lung transplantations have been reported to be associated with a similar prognosis in a large study of 6655 lung transplantations included in the United Network for Organ Sharing (UNOS) database between 2005 and 2011 [22]. The prognostic impact of size mismatches between donor and recipient is more difficult to analyse, as definition and indication varied across studies. Eberlein et al. [23] studied the Lung Transplant Outcome Group database, defining oversizing as a donor-to-recipient pTLC ratio >1.0, and found that patients undergoing bilateral lung transplantation with oversized lungs experienced improved survival, an effect that appears most apparent in non-COPD patients. Conversely, a recent analysis of the UNOS database focusing on restrictive lung disease found that moderate oversizing, defined as a donor-to-recipient pTLC ratio between 1.1 and 1.2, was associated with an adverse prognosis [24]. Similarly, recipients of undersized allografts, defined by Eberlein et al. [23] as a donor-to-recipient pTLC ratio ≤1, were more likely to experience primary graft dysfunction, tracheostomy and had higher resource utilisation [25]. Following International Society for Heart and Lung Transplantation guidelines [26], we defined size mismatch as oversizing if the donor-to-recipient pTLC ratio was >1.2 and as undersizing if the donor-to-recipient predicted TLC ratio was <0.8, and found it did not affect overall survival in the frame of the HELT procedure. Taken together, our data suggest that the adverse outcome associated with HELT is associated with recipient severity status, but not with donor or matching.

Allocation policy

Transplantation is the only medical therapy in which there is a life-sustaining therapy that is in short supply relative to the need [27], making allocation policy a critical point from both the individual and the population perspective. The main ethical principles associated with organ allocation include equity to eliminate bias or discrimination, justice to render each individual what is due to him/her and beneficence to expect a greater good over harm for the patient, but also utility to make the best use of a limited resource [27, 28]. The relative weights of theses principles differ between countries. In particular, the principle of utility implies scaling up from the patient–physician relationship to a population level and might be difficult to balance with the first three principles from a physician perspective. The lung allocation system is a direct reflection of the public investment into ethical considerations.

Besides France, other countries have chosen to prioritise patients based on their gravity only, and therefore to allocate lungs based on urgency and not transplant benefit. The Scandiatransplant (Denmark, Estonia, Finland, Iceland, Norway and Sweden) urgent lung allocation system (ScULAS) was implemented in 2009, giving a supranational priority to patients considered urgent [29]. No predefined criteria for listing a patient as urgent are required, but Priority 0 includes patients on extracorporeal support or mechanical ventilation, and Priority 1 includes patients not on life support but with a rapid progression of lung failure and poor short-term prognosis as defined by the responsible centre. However, only three urgent calls per year are granted for each centre. Therefore, patients on life support might not benefit from an urgent status. Since the implementation of ScULAS, 30-day graft survival and 90-day graft survival have been significantly lower among patients listed as urgent, but there was no difference in 1-year graft survival. However, 81 patients were on life support and among them, only 39 patients were listed as urgent and 15 patients (19%) died on the waiting list (five patients listed as urgent status and 10 patients listed as regular status). Moreover, graft survival was found to be significantly lower in patients on life support compared with other patients [29]. The Italian urgent lung transplant programme was introduced in November 2010, giving a national priority to patients ≤50 years old requiring mechanical ventilation and/or extracorporeal lung support. Results from the first 14 months showed 30-day, 6-month and 1-year survival rates of 82%, 76% and 71%, respectively, and a mortality rate on the urgent waiting list of 11%, while another 11% of the patients were excluded because of worsening conditions [10]. The Super-Urgent and Urgent Lung Allocation Schemes were introduced in the UK in 2017, in order to give a national priority to patients who are at most risk of dying on the waiting list [30]. Criteria for Super-Urgent Lung Allocation Scheme registration include extracorporeal life support with ECMO or interventional lung assist (Novalung), while criteria for Urgent Lung Allocation Scheme registration depend on the underlying disease [31]. The 2018 Annual Report on Cardiothoracic Organ Transplantation in the UK showed that there were 33 urgent lung registrations and seven super-urgent lung registrations between April 2017 and March 2018 [32]. Results after the implementation of this new allocation system in the UK have not yet been reported. Taken together, these data suggest that HELT and other gravity-based urgent allocation systems are associated with suboptimal outcome. Given that the whole allocation system is not destabilised, this might be an acceptable trade-off for minimising waiting list mortality.

Conversely, the LAS has been designed to maximise the benefits of transplantation, and therefore allocates lungs based on both urgency and transplant benefit. The implementation of the LAS in the USA has been reported to be associated with fewer waiting list deaths, more transplants performed, a change in the distribution of recipient diagnoses to patients more likely to die on the waiting list and an improvement in 1-year survival [6]. However, the LAS has been designed in the USA and does not take into account two urgent status parameters, i.e. extracorporeal support and mechanical ventilation. Its applicability to other transplant populations and to the most severe patients was therefore questionable. The impact of the LAS in urgent and highly urgent lung transplantation candidates was assessed in 2011 in Eurotransplant (Austria, Belgium, Croatia, Germany, Hungary, the Netherlands, Luxembourg and Slovenia). Smits et al. [28] showed that the LAS and the LASplus (encompassing extracorporeal support and noninvasive ventilation) were significantly associated with 1-year post-transplant mortality and overall mortality. Since then, Germany was the first European country to adopt the LAS in 2011, followed by the Netherlands in 2014. As patients on mechanical ventilation or extracorporeal support were not accounted for in the LAS, specific allocation rules were introduced in Eurotransplant, as well as proposing an exceptional LAS when the calculated LAS does not reflect the transplant benefit for a particular patient [33]. Since the introduction of the LAS, the German experience reported a 26% decrease in waiting list mortality and an improved 1-year survival rate from 76% (2009–2011) to 81% (2012–2014) [34], suggesting that the benefits associated with the LAS could be applicable outside the USA and in very severe patients.

Limitations

Our study has several limitations. The number of covariates (mean pulmonary arterial pressure, pre-operative cardiopulmonary support and graft ischaemic time) with >10% of missing data could bias the results. We tried to take these into account by using a multiple imputation method by chain equations. We also included recipient, donor and surgical characteristics as potential cofounding factors in the multivariate analysis. Still, we cannot exclude that other factors not studied in our analysis or unidentified may have biased our results. Our survival rates are somewhat lower than those reported in the International Thoracic Organ Transplant (TTX) Registry (https://ishlt.org/registries/ttx-registry), but represent real-life data as depicted in an exhaustive administrative database that does not include COPD patients.

Conclusions

This exhaustive French national study suggests that the adverse outcome associated with the HELT procedure is related to the severity status of the recipients rather than donor or matching characteristics.