How the Warm Ischemia Model Shapes the Message in Lung Transplant Research

When a donated lung is assessed for transplantation, the details of the laboratory model used to study it can quietly determine what the experiment appears to show. That is the central argument of a Forum article published in Transplant International by Jitte Jennekens and colleagues from University Medical Center Utrecht and KU Leuven, who respond to a study by Premachandra and co-workers on the resilience of lung grafts exposed to two hours of warm ischemia. Their commentary, titled “When the model defines the message: methodological considerations in experimental lung warm ischemia,” does not dispute the original data. Instead, it dissects the experimental choices behind those data and warns that conclusions about how lungs tolerate warm ischemia may depend more on the model than on the biology itself.

The original study examined the effects of prolonged warm ischemia on lung graft function and biological responses during ex vivo lung perfusion, a technique in which donor lungs are circulated with oxygenated perfusate outside the body before transplantation. Premachandra and colleagues reported that lungs subjected to two hours of warm ischemia showed lower initial compliance and increased pulmonary vascular resistance compared with control lungs. Notably, these functional differences resolved over the course of the perfusion period. By the end of ex vivo lung perfusion, no significant differences remained in molecular or histological markers of injury, with the exception of lactate and PECAM concentrations, which were paradoxically higher in the control lungs.

At first glance, these findings suggest that lungs are remarkably resilient to two hours of warm ischemia, a result with obvious appeal for expanding the donor pool in lung transplantation. But Jennekens and colleagues argue that the way warm ischemia was induced in the study differs in important respects from the clinical scenarios it is meant to represent. In clinical practice, extended periods of warm ischemia typically arise during a prolonged agonal phase in donation after circulatory death, or in uncontrolled donation settings where the lungs are mechanically ventilated after failed resuscitation attempts until the mandatory no-touch period has elapsed. In the experimental model used by Premachandra and colleagues, by contrast, the lungs were left inflated inside the cadaver for the entire two-hour ischemic interval.

That distinction matters because the mode of death has been shown to influence lung graft quality. Previous work has demonstrated that hypoxic cardiac arrest is associated with inferior graft quality compared with immediate circulatory arrest or exsanguination. In the study under discussion, the animals were euthanized by aortic clamping with simultaneous exsanguination, a death mode that may not reproduce the physiological conditions of the donation scenarios the study aims to inform. The Utrecht and Leuven authors contend that a clinically relevant donation after circulatory death model must address these procedural and timeline factors, because they directly shape how warm ischemia injures the lung.

The second methodological issue concerns whether the lungs are inflated or deflated during the ischemic period. Earlier studies, including work by Raemdonck and colleagues, established that ventilated or inflated lungs are protected from warm ischemic injury, whereas deflated lungs develop pulmonary edema and impaired gas exchange during subsequent normothermic reperfusion. Alveolar expansion itself, rather than continuous oxygen supply, appears to enhance postmortem preservation of pulmonary grafts. This means that an inflated warm ischemia model, as used in the original study, is intrinsically a gentler model than one in which the lungs are allowed to collapse.

This distinction becomes critical when interpreting the inflammatory findings. The authors of the original study noted that several experimental groups had reported enhanced inflammatory responses with extended periods of warm ischemia lasting one to two hours during ex vivo lung perfusion. They attributed the discrepancy in their own results primarily to the absence of methylprednisolone in those earlier experiments, and indeed steroids have been shown to reduce warm ischemic reperfusion injury in porcine donation after circulatory death models. Jennekens and colleagues acknowledge this explanation but point out another key difference: the studies cited by the original authors used deflated warm ischemia, a model associated with more extensive lung injury than the inflated model employed in the new work. The absence of inflammatory differences, they argue, may therefore reflect not only the protective effect of methylprednisolone but also the specific experimental model itself.

The third and perhaps most technically intricate concern involves the split-lung ex vivo perfusion design. In this approach, the two lungs from a single animal are perfused separately, allowing one to serve as a control for the other. The commentators were surprised that only left lungs were allocated to the warm ischemia group while right lungs served as controls, meaning that treatment assignment was confounded by anatomical laterality, lung size, and the perfusion and ventilation parameters that follow from them. Following the Toronto protocol, the investigators adjusted perfusion to 24 percent and 16 percent of the theoretical cardiac output, and tidal volumes to 3.5 and 2.5 milliliters per kilogram, for the right and left lungs respectively, corresponding to a right-to-left perfusion distribution of 60 to 40 percent. In humans, a perfusion distribution of 55 to 45 percent is commonly assumed.

Adjusting flow and ventilation for size differences between the two lungs, the commentators argue, does not guarantee that the resulting ex vivo perfusion is physiologically comparable between the groups. There is a potential risk that one group receives a more protective perfusion regimen than the other, independent of the warm ischemia intervention being tested. A similar principle applies to perfusate-derived biomarkers. The higher lactate and PECAM concentrations observed in the control lungs at the end of perfusion were interpreted by the original authors as possibly reflecting a protective priming effect of warm ischemia. Jennekens and colleagues propose an alternative explanation: the greater lung mass and potentially greater amount of perfused tissue in the right lungs of the control group may have contributed to the differences in absolute marker release. Without normalizing perfusate markers to lung mass or to another measure of perfused tissue amount, they caution, differences in marker concentrations should be interpreted with care.

The commentators also point toward a practical solution. Future split-lung ex vivo perfusion studies could randomize treatment assignment between the right and left lungs across animals, breaking the link between laterality and experimental condition. Split-lung perfusion models are increasingly valued as time- and cost-effective platforms for evaluating therapeutic interventions in donor lungs, and recent work on interventions such as ferroptosis inhibition has demonstrated their utility across organ systems. Randomization within the split design would preserve the statistical power that makes these models attractive while eliminating a systematic bias that could otherwise skew results.

Taken together, the Forum article concludes that the original study demonstrated relatively limited functional and biological effects in porcine lungs exposed to two hours of inflated warm ischemia under a steroid-containing ex vivo lung perfusion protocol. That conclusion may well be accurate for the model as constructed. But the specific experimental conditions used in ex vivo lung perfusion studies, from the mode of death and the inflation state of the lung to the laterality of treatment assignment, should be carefully considered when assessing the clinical relevance of the findings. For a field racing to expand the boundaries of organ acceptance, the message is clear: before a model’s results can be translated to the operating room, the model itself must be scrutinized as rigorously as the data it produces.

Subject of Research: Methodological considerations in experimental models of lung warm ischemia and ex vivo lung perfusion for transplantation research

Article Title: When the model defines the message: methodological considerations in experimental lung warm ischemia

Article References: Jennekens, J., Imholz, D., Braithwaite, S., de Heer, L., & Neyrinck, A. (2026). When the model defines the message: methodological considerations in experimental lung warm ischemia. Transplant International, 39, Article 17803. https://doi.org/10.3389/ti.2026.17803

Image Credits: AI Generated

DOI: 10.3389/ti.2026.17803

Keywords: lung transplantation, warm ischemia, ex vivo lung perfusion, donation after circulatory death, ischemia reperfusion injury, EVLP, split-lung model, lung graft quality, methylprednisolone, porcine model, DCD donors, transplantation research