Heart transplantation typically relies on the use of brain-dead donor hearts, which constitutes the main source of donation [1]. However, continuous large imbalances between organ supply and demand for heart transplantations has led to the investigation of new strategies to expand the donor pool [2]. Among these strategies, transplantation with donor hearts after controlled circulatory death has been of particular interest in recent years [3–9]. Increasing consideration of this strategy has been driven by the evolution of technologies, such as ex situ organ perfusion systems, allowing for reanimation of the heart after circulatory death and for evaluation of its suitability for transplantation [10]. Heart transplantation using circulatory-dead donor hearts has been proved non-inferior to standard care transplantation using brain-dead donor hearts in terms of risk-adjusted survival at 6 months in a prospective randomized controlled trial [6]. A protocol to assess the feasibility of heart transplantation following Donation after controlled Circulatory Death (DCD) was conducted in France (Protocol PFS20-004, French Agency of Biomedicine) [11]. However, data on the cost-effectiveness of this donor pool expansion strategy are lacking. Indeed, DCD heart procurements are expensive (54,000 euro per normothermic ex-situ perfusion consumable) but may improve outcomes for patients with end-stage heart failure by increasing their chances to receive a graft with comparable post-transplantation outcomes [12]. Herein, we sought to model the cost effectiveness of introducing DCD heart transplantation alongside the current standard practice of Donation after Brain Death (DBD) with static cold storage from the French healthcare system perspective and to outline the factors enabling an efficient DCD heart program.

A cohort discrete-time state transition model with a time-inhomogeneous Markov process was developed [13]. In these decision models, a hypothetical cohort of individuals transition over discrete time periods (i.e., cycles) between mutually exclusive health states representing the clinical course of the disease [14]. The time inhomogeneous approach relaxes the Markov assumption, allowing transitions, rewards, or both to vary over time [13]. A cost utility analysis framework was used to compare costs and effects (expressed as quality-adjusted life years or QALYs) from the French healthcare system perspective between standard care and the innovative strategy [15]. The model was built and run in R (version 4.2.2) within the R Studio software and using the R package “heemod” [16]. We followed the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) reporting guidelines [17].

Over a 15-year time horizon, patients in the DBD strategy experienced 5.53 QALYs at a cost of €184,110 whereas patients in the DCD + DBD strategy experienced 5.67 QALYs at a cost of €204,728. These outcomes resulted in an incremental gain of 0.14 QALYs at an incremental cost of €20,618, producing an ICER of €146,373 per QALY gained. The results of the DSA are presented in Figure 2. The main drivers of the ICER estimate were the utility of patients on the waiting list and after 3 years post-transplantation. Results from the scenario analyses were reported in Table 2. The application of a utility decrement and an increase in the cost of heart transplantation in a pessimistic scenario of major impact of PGD with DCD resulted in an ICER of €180,763 per QALY at 15 years.

Results of the PSA are presented in Figures 3, 4. The mean probabilistic ICER of the DCD + DBD strategy compared to DBD alone was €142,423 per QALY with a mean incremental cost of €20,609 (95% credible interval: €11,220–€37,453) and a mean incremental QALY gain of 0.14 (95% credible interval: –0.03–0.3). The innovative strategy reached 50% and 80% chances of being cost-effective respectively at a WTP of €139,200 and €261,000 per QALY.

The independent threshold analyses are presented in Figures 5, 6 and can be interpreted according to different willingness-to-pay thresholds. For example, taking a WTP of €100,000 per QALY, the DCD + DBD strategy is cost-effective if (1) an 8% relative reduction in waiting list mortality is reached with the donor pool expansion (i.e. 0.8% absolute reduction, from 10.2% to 9.4% probability of death 1 year after listing) and (2) a 35% relative reduction in consumables costs is negotiated (i.e., moving from €54,000 to €35,500 per consumable).

To our knowledge, this is the first study assessing the cost-effectiveness of implementing heart transplantation following DCD, according to a protocol with direct procurement procedure (i.e., involving an ex-situ perfusion machine).

The assumptions underlying our model structure and parameters are primarily drawn from the existing literature in this area: our main hypothesis is that the introduction of DCD heart transplantations could increase the odds of a candidate to receive a graft [12, 28], while achieving outcomes comparable to standard care transplantation [3–6].

These assumptions are further supported by recent international experience, including long-term Australian data showing comparable post-transplant survival between DCD and DBD recipients up to 10 years using direct procurement followed by ex-situ machine perfusion [39].

For aspects where evidence was limited or uncertain, we adopted a conservative approach in our assumptions. In particular, we assumed that an increase in the supply of donors would not affect the mortality risk on the waiting list in the base case analysis. Sensitivity analyses were also performed. We specifically tested a pessimistic alternative scenario in which PGD, a commonly feared complication with DCD [36, 37], had a major impact on costs and quality of life during the heart transplantation state in the DCD + DBD strategy, which led to an ICER of €180,763 per QALY gained over 15 years. Also, if implementing DCD heart transplantations were to reduce waiting list mortality, the strategy would offer greater value for money. In our model, an 8% relative reduction of probability of death 1 year after listing corresponded to an ICER of €100,000 per QALY. In France, an average of 500 patients are newly registered on the waiting list each year, which would translate to approximately four additional lives saved per year for the DCD program to be cost-effective at this willingness-to-pay threshold. Additionally, lowering consumable costs could further improve the cost-effectiveness of DCD heart transplantation.

This health economic evaluation aims to provide valuable insights for decision-making in countries where DCD heart transplantation procedures are not yet authorized and/or reimbursed. In some countries such as France, no predefined cost-effectiveness threshold is yet used as a formal decision-making criterion [40]. In practice, decisions are often influenced implicitly by both the cost-effectiveness ratio and the clinical severity or burden associated with the condition [41]. End-stage heart failure is an advanced stage of disease known to be associated with poor prognosis and impaired quality of life [34, 42, 43]. Given the persistent and worldwide shortage of heart donors [44], it may be regarded as falling within the scope of unmet (or partially unmet) medical needs [45]. These aspects therefore also need to be considered in the decision-making process. To help the decision, threshold analyses were also provided. They could serve as a foundation for value-based negotiations between decision-makers and manufacturers to support fair (yet sustainable) pricing of machine perfusion consumables. They could also help align heart transplantation practices with value-based objectives for efficient DCD heart programs, including reducing waiting list mortality, as this parameter appears to be pivotal for the cost-effectiveness of the strategy.

Direct procurement followed by ex-situ machine perfusion is not the only approach currently used for DCD heart transplantation [1]. Alternative strategies, particularly thoraco-abdominal NRP followed by static cold storage, have demonstrated favorable outcomes, notably in the recent Spanish national experience [8].

Direct procurement followed by ex-situ machine perfusion was considered in the present analysis as this approach was developed to introduce heart retrieval within the existing French DCD protocol based on abdominal NRP for abdominal organ procurement [11]. Therefore, it should be acknowledged that our findings are specific to this procedure, reflecting the French setting, and should not be extrapolated to strategies based on thoraco-abdominal NRP, which involve different procurement procedures and resource requirements.

Nevertheless, from an international perspective, countries such as Australia, the United States, and the United Kingdom initially implemented direct procurement followed by ex-situ machine perfusion for DCD hearts without combining heart procurement with abdominal NRP. Considering the favorable outcomes associated with abdominal-NRP for liver and kidney transplantations in the French experience [46, 47], recent efforts to promote the international standardization of DCD practices involving the use of NRP [48], and the health economic considerations outlined in the present analysis, our findings may provide useful information for healthcare systems considering the integration of abdominal NRP alongside direct heart procurement and ex-situ machine perfusion.

The main strength of this study was its use of epidemiological and survival data from the French organ transplantation agency to parametrize the model. Also, individual patient cost data in a French context were used as model inputs, retrieved from a cost of illness study and a randomized controlled trial with cost data collection (i.e., they were not reconstituted based on practice recommendations). Economic models are, by definition, simplifications of the reality, but we implemented a time-inhomogeneous Markov process to the decision model to help reflect the clinical course of the disease and associated costs and outcomes. Moreover, the model could be adapted to support the evaluation of DCD implementation in different healthcare settings globally.

Our study does have limitations. One limitation is that we could not perform subgroup analyses based on patient characteristics and pathways, which would have been interesting given the inherent heterogeneity of this population [18]. Due to the scarcity of data in this indication, we had to assume the candidate population at the entry of the model as homogenous. It might be possible that specific populations with different levels of graft priority slightly differ in terms of transplantation access, cost, and consequences. The analysis was limited to hospital costs and may therefore underestimate total costs, although hospital care was assumed to be the main cost driver in this indication. Additional implementation costs related to logistics, coordination, and learning curves were not explicitly modeled due to limited available data, but uncertainty related to procedure costs was partially explored in the DSA. Indeed, these costs remain difficult to estimate accurately in the French setting, given the very recent authorization of DCD heart transplantation in the country (March 2026). Accurately estimating these additional costs would require dedicated prospective micro-costing analyses of the procedure, which were beyond the scope of this early economic evaluation. However, we assumed that DCD heart procurement would not adversely affect abdominal organ procurement, consistent with the recently reported French national feasibility study, in which adding heart retrieval in the setting of abdominal NRP did not compromise the procurement of other organs [11]. Another limit is the absence of utility estimates on the pre-transplant period and after 3 years in a French context, which would be interesting to collect as being main drivers of uncertainty in the DSA.

Finally, in the present model, the impact of donor pool expansion was captured among listed candidates. However, increasing the availability of donor hearts may affect listing practices and outcomes in patients with advanced heart failure who are currently not listed because of persistent donor shortage. These broader effects were not modelled due to their inherent uncertainty but may further improve the overall benefits associated with DCD heart transplantation.

Future research should aim to evaluate the potential change of the candidates’ waiting list mortality risk with the implementation of DCD procedures. Organizational aspects (perfusionists, coordination of extended heart procurement, etc.) should also be investigated in real-world settings. A budget impact analysis could be of interest to assess the sustainability of this strategy for healthcare payers.

In conclusion, our results suggest that, under the assumption of a 20% annual increase in transplant activity—improving the odds of a candidate receiving a graft without affecting waitlist mortality and assuming equivalent post-transplant outcomes—implementing DCD heart transplantation alongside DBD would result in an ICER of €146,373 per QALY gained over a 15-year time horizon from the French healthcare system perspective. Lowering consumable costs and reducing waitlist mortality are key factors enabling an efficient DCD heart program.