Evaluation of heart grafts using non-invasive techniques before transplantation into the recipient appears as a useful addition to the current standard clinical practice, in which no objective examination of the graft is performed immediately before transplantation. A large variety of destructive and non-destructive biomarkers have been proposed in both clinical and preclinical settings [1]. Despite emerging machine-perfusion beating-heart storage devices [2–4], cardioplegic arrest followed by cold storage of the graft currently remains the most widely used method. In both warm and cold storage situations, an objective evaluation before transplantation could contribute to increasing transplant safety on the one hand and to widening the donor pool on the other. One important indicator of graft quality is given by high-energy phosphate compound (HEPC) concentrations [5, 6] that decrease in the absence of nutriment and oxygen supply via perfusion, and it is one purpose of preservation solutions to prevent their rapid depletion. These metabolite concentrations can be assessed in a completely non-destructive way using phosphorus-31 Magnetic Resonance Spectroscopy (³¹P MRS). Using MRS, the quality of graft preservation with different cardioplegic solutions has been assessed in animal hearts [3, 7–10], and metabolite preservation before graft implantation has been analyzed in human hearts [5, 6] during cold storage and, more recently, during warm storage under machine perfusion [4]. However, little is known about the influence of storage time on the metabolic preservation quality of human heart grafts in a clinical context, where an estimation of metabolic depletion rates in human grafts could help strengthen safety. The purpose of this work was to assess the time course of HEPC alterations during cold storage of human heart grafts. This was possible using hearts that were discarded from transplantation based on clinical donor criteria but that underwent the same retrieval and cold storage procedure as transplanted grafts for the purpose of this study.

No rupture of sterility or low temperature was caused to the grafts by this non-destructive assessment. The study was approved by the University Hospital’s ethics committee “Comité consultatif de protection des personnes dans la recherche biomédicale—Marseille 2” (Authorization #99/16) and conducted in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. The donors’ next of kin gave their written informed consent to the heart graft retrieval. This retrospective dynamic data analysis was carried out in 2024 on Magnetic Resonance data acquired from 1999 to 2001. According to current rules confirmed by the French Agence de Biomédecine, this analysis is therefore exempted from additional specific ethical approvals.

An example stack plot of spectra from the graft observed during 13 h as a function of time is displayed in Figure 1. Figure 2 shows the signal amplitudes of PCr, P_i, γATP and phosphomonoesters (PME) as a function of time after cardioplegic arrest for all examined grafts. The graphs account for the variations in arrival time on site. For visualization, three data points from misfitted spectral resonances of graft number 2 were removed. Figure 2 also shows the intracellular pH as long as PCr was measurable.

In the seven grafts, PCr signal, normalized to that at the time of retrieval, decreased by 10.3% ± 1.9%/hour and P_i increased by 12.6% ± 5.0%/hour. From the graft followed for 13 h (shown in Figure 1) it can further be seen that ATP levels were stable until PCr was depleted, and PCr depletion did not occur in the other observed grafts, i.e., before 9 h of storage. PCr decrease and P_i increase were well represented by the linear regression over 9 h. The concentration ratios PCr/ATP and Pi/ATP can, therefore, be expressed as a linear function of the ischemic time T _is [min]: PCr/ATP = 1.68-0.0028·T _is, Pi/ATP = 1.38 + 0.0029·T _is. Intracellular pH obtained from the chemical spectral shift of Pi also showed a shallow linear decrease over time: pH = 7.43–0.0012·T _is, indicating good stability of the acidic level.

We have measured the time course of changes in HEPC metabolite concentrations in seven human heart grafts during cold storage. HEPC metabolites are known contributors to the graft quality during transplantation. Depletion of high energy phosphates during ischemia is linked with harmful injury in the ischemic myocardium [6], and levels of high energy phosphate have been correlated to the recovery of function after transplantation both in animal models and human studies [6, 14–16]. In a more recent study by Föll et al. [17], the ischemic storage time was also shown to correlate with regional wall motion after transplantation.

Signal variations show that on a timescale typical for cold storage duration in heart transplantation (maximum 4 h), HEPC metabolites in human heart grafts are well preserved. Although PCr was decreasing rapidly, the ATP levels in all six example heart grafts were stable for at least 9 h following cardioplegic arrest and storage with Celsior. This indicates that in current clinical practice using these conditions of preservation, the level of ATP is well preserved over a long cold storage duration. ATP is indeed related to numerous cell structure mechanisms and to contractility, and its depletion associated with irreversible loss of precursors. These results also show a moderate acidosis of the grafts at pH 6.8 after 9 h of ischemia. Earlier studies carried out in animal hearts during ischemia with cardioplegia have shown that the intracellular pH is dependent on the pH and on the buffering capacity of the preservation solution [10]. While the heart is able to recover from mild acidosis, severe acidosis would induce cell necrosis.

We have used ³¹P Magnetic Resonance Spectroscopy for this analysis. This tool proved to be particularly useful, since the entire exam could be accomplished with preserved sterility and temperature with in situ determination of metabolites and intracellular pH. The dynamic measurement allowed us to derive change rates for the major compounds ATP, PCr and P_i and for the intracellular pH. These rates can be considered as characteristic for the particular group of grafts assessed here. As a difference with earlier work by our group and others [5, 6], all grafts studied here were discarded from transplantation based on their clinical score. This leaves some uncertainty on the initial metabolic status of the grafts before retrieval, since grafts with lower clinical scores were shown earlier to have lower PCr/P_i concentration ratios [5]. This likely explains the relatively low initial PCr/ATP and high Pi/ATP ratios found even when linearly extrapolating them back to the moment of retrieval.

The influence of ischemic time on HEPC metabolism in heart grafts has been evaluated earlier by van Dobbenburgh and coworkers [6] in a study correlating functional performance and metabolic status before transplantation in 25 heart grafts arrested with St Thomas cardioplegic solution. One has to keep in mind that van Dobbenburgh et al. assessed grafts that were actually transplanted at relatively short ischemic times (<2 h), allowing them to acquire only one time point per graft. The authors also reported a correlation of metabolism and the variable ischemic time (T _is [min]) of each graft, although only one time point per graft was available. Their constructed decreases over time were PCr/ATP = 1.31−0.0039·T _is, Pi/ATP = 0.26 + 0.0064·T _is and pH = 7.66-0.0040 ·T _is and can be compared with our regressions: PCr/ATP = 1.68−0.0028·T _is, Pi/ATP = 1.38 + 0.0029·T _is, and pH = 7.43−0.0012·T _is. During our nine-hour observation period, PCr, Pi and pH, therefore, decreased slower than in the <2 h period observed in van Dobbenburgh’s mult-organ study. Initial Pi, as calculated by linear extrapolation to the moment of explantation, was, however, found much higher in our study. These differences may be related to the low clinical scores reported for the grafts used here. In our study, time courses were obtained from each graft individually over time periods of 9 h. The change rates may also vary as a function of the storage solution (here: Celsior), which was shown earlier to have an influence on metabolite levels during reperfusion in rat hearts [14].

Similarly to observations reported by van Dobbenburgh, we observed a relatively large variation in both PCr/ATP and, especially, Pi/ATP ratios of the donor hearts as well their change rates. These differences may be caused by the condition of the donor, the use of pharmacological therapy, and the quality of cardioplegic arrest and hypothermic preservation before arrival on site. The low number of grafts available for this study, however, did not allow us to evaluate statistical correlations with these parameters.

For obvious reasons, the number of grafts at our disposal was limited by the availability, since only a single center contributed to this study. The standard deviation of metabolite change rates was therefore relatively high (in the 20% range). We report these results to give an estimate of the order of magnitude of metabolic changes during cold storage. Further observations will be needed to strengthen these results and to generalize the conclusions with sufficient confidence.

We have non-destructively analyzed the long-term behavior of high-energy phosphate compounds in human heart grafts. Phosphorus MRS proved to be a useful tool for such a non-destructive analysis. Although many parameters affect graft quality during cold storage, the dynamics of HEPC and intracellular pH may be helpful in the development of strategies aiming at extending the ischemic time. Beyond cold storage, MRS may also provide interesting additional information for evaluating grafts during or after machine-perfused graft storage in comparison.