All procedures were approved by the University of Bristol Animal Welfare and Ethical Review Body and were performed in accordance with the UK Animals (Scientific Procedures) Act 1986. Anaesthesia was standardised between all piglets (38 female Landrace piglets, around 12 kg in weight—Table 1). The piglets were fasted overnight and sedated using an intramuscular injection of ketamine (15 mg/kg), morphine (0.2 mg/kg) and midazolam (0.3 mg/kg) before induction of anaesthesia (propofol 2 mg/kg, i.v.). The piglets were then intubated using a cuffed endotracheal tube and mechanically ventilated; anaesthesia was maintained using isoflurane in an O₂/air mix. Fentanyl (10 µg/kg/h, i.v.) and pancuronium (0.1 mg/kg/h i.v.) were continuously infused to provide analgesia and muscle relaxation respectively. Standard physiological monitoring was performed throughout as was monitoring of neuromuscular blockade using train-of-four stimulation.

A median sternotomy was performed to expose the heart and great vessels and the thymus was partially excised. Heparin (400 IU/kg, i.v.) was administered, the right atrium and ascending aorta were cannulated and cardiopulmonary bypass (CPB) commenced. A cannula with Y adaptor was inserted into the ascending aorta to administer blood cardioplegia and act as a supra-aortic vent. An aortic cross clamp was applied and cardioplegic arrest was induced and maintained for 60 min. The blood pressure on cardiopulmonary bypass was maintained around 40–50 mmHg, using bolus doses of metaraminol as needed. Cardioplegic protocols followed clinical practice used in Bristol Children's Hospital.

In the cold cardioplegia group, arrest was induced by administration of 110 ml/min/m² cold blood cardioplegia for four minutes, with blood: cardioplegia (sterile concentrate for cardioplegia infusion, Martindale Pharmaceuticals, Romford, UK) ratio of 4:1. Arrest was maintained after 30 min by a maintenance dose (110 ml/min/m²) for two minutes, with additional maintenance doses given if the heart resumed activity between doses of cardioplegia. For all doses the concentration of both MgCl₂ and KCl in the cardioplegia arriving at the heart was 8 mmol/L. Cardiac temperature was monitored with a flexible temperature probe inserted into the left ventricular wall to monitor the cooling of the heart.

In the warm cardioplegia group, arrest was induced by administration of 205 ml/min/m² warm blood cardioplegia over two minutes, maintained every twenty minutes with a maintenance dose (130 ml/min/m²) for one minute, with additional maintenance doses given if the heart resumed activity between doses of cardioplegia. For all doses the concentration of both MgCl₂ and KCl in the cardioplegia arriving at the heart was 18.7 mmol/L.

For both the warm and cold protocols, the cardioplegia solution arriving at the heart contained a sildenafil (Revatio, Pfizer, New York) concentration of either 0 or 10 nmol/L, a concentration which is well below the dosage used clinically in paediatric patients for the treatment of pulmonary arterial hypertension and reported therapeutic plasma concentrations (36, 37). The surgeon, anaesthetist and perfusionist were all blinded as to whether the cardioplegia contained sildenafil or not (but it was not possible to blind whether the cold or warm cardioplegia protocol was being used).

Following 60 min of arrest, the aortic cross-clamp was released. If fibrillation occurred, up to 5 electrical shocks were given to induce sinus rhythm. Ventilation was then resumed, and the heart weaned off cardiopulmonary bypass 20 min after releasing the cross clamp. Protamine (5 mg/kg) was administered to reverse heparinisation and the animal swiftly decannulated. Following a total of one hour of reperfusion, the experiment was be terminated via an injection of pentobarbitone (0.1 mg/kg i.v.). If the pig could not be weaned off bypass it remained on bypass until termination.

Arterial blood samples were collected immediately before CPB commenced, immediately before cardioplegia was administered, immediately before the cross-clamp was removed, at 1, 10, 20 and 60 min reperfusion. Samples were collected by the perfusionist whilst the pig was on CPB and by the anesthetist at all other timepoints. Blood was spun immediately (1,000 g for 10 min at 4°C) and the plasma stored at −80°C. Cardiac tissue biopsies were collected from the left ventricle by the surgeon at four timepoints (immediately before CPB, immediately before release of the aortic cross-clamp, at 20 min reperfusion and at termination) using a trucut needle. Biopsies were immediately flash frozen in liquid nitrogen before storage at −80°C. Arterial blood gases (EPOC portable blood gas machine) were collected throughout the protocol by the anesthetist and perfusionist as considered necessary for clinical practice. For analysis, the blood gases were grouped into three time periods (before CPB, whilst the aortic cross-clamp was on and after cross-clamp release). If a pig had more than one blood gas measured during a time period, an average was calculated.

The tissue biopsies were crushed whilst still frozen in liquid nitrogen using a pestle and mortar. The resulting tissue powder was immediately transferred into a weighed LP3 tube containing 4.8% perchloric acid at 4°C. Wet tissue weight was calculated by reweighing the tubes following addition of the tissue. The mixture was vortexed and centrifuged (4,000 g, 10 min, 4°C) before the supernatant was neutralized by mixing with 0.44 M K₂CO_3. The volumes used were chosen to generate a neutral pH, with the volumes noted for subsequent calculation of the dilution factor. The mixture was then re-centrifuged before the supernatant was collected, aliquoted and stored at −20°C freezer for HPLC analysis.

High-performance liquid chromatography (HPLC) was used to separate and identify metabolites as reported previously (11, 38–40). The signals generated from HPLC were displayed as chromatograms using (OpenLab Software, Aligent Technologies, USA). Peaks of each metabolite in the sample chromatograms were identified using standards of known concentration. Data analysis was completed by measuring the area under the curve for each peak, and the values were corrected to account for the dilution factor and the weight of each biopsy to produce the concentration of each metabolite. Data for one pig in the sildenafil group of the warm cardioplegia experiment was removed due to having an extremely high baseline value of troponin (∼13 ng/ml). Energy charge for adenylates has been used as a measure of metabolic flux rate (41) and is calculated as follows:Energycharge=ATP+0.5ADPATP+ADP+AMP

Concentrations of troponin I were measured in the pig plasma using a commercially available Elisa kit (CTNI-9-HSP, Life Diagnostics, West Chester, USA). In brief, samples were diluted and incubated in a 96 well plate (150 rpm, room temperature) with two antibodies (one for solid phase immobilization and one for HRP conjugation). Wells were washed to remove excess HRP conjugate, and TMB added. In the presence of cTNI a blue colour develops, which was converted into yellow by addition of stop solution. Absorbance was measured at 450 nm, and concentrations calculated by comparison to a standard curve.

For comparisons between two groups, the data was first tested for normality using the Shipiro–Wilk test. Normal data was analyzed with unpaired, 2-tailed t-tests whilst non-normal data was analyzed with Mann–Whitney U test. Categorical data was analysed with Chi squared tests. Levels of metabolites and troponin were analyzed by 2-way mixed model ANOVA with one repeated measures factor (time) and one between subjects factor (treatment group). A one-way ANOVA with post-hoc Dunnett's test where appropriate was also performed for each treatment group to compare which timepoints were significantly different from baseline. All statistics were performed in SPSS (Version 28, IBM Analytics, New York, USA) and significance was accepted when p < 0.05.

There was no significant difference between the body weights of piglets used in the control and sildenafil groups in either the cold cardioplegia or the warm cardioplegia experiments (Table 1). Measurements of cardiac temperature in the cold cardioplegia group confirms that the cardiac tissue was effectively cooled by administration of the cardioplegia, and that there was no difference in the temperatures obtained between the control and sildenafil groups (Table 2). Blood gases throughout the protocol for each group of animals are shown in Supplementary Tables S1, S2.

In our warm cardioplegia protocol, addition of sildenafil to the cardioplegia significantly increased the success rate of being able to successfully wean the hearts from CPB compared to the control group, with the piglets in the sildenafil group also requiring significantly fewer electrical shocks to return to sinus rhythm. Previous investigations into whether sildenafil has a cardiac inotropic effect have proved controversial. Several human studies have found no evidence of an inotropic effect of sildenafil, in vivo or in vitro (42–44), whilst others have reported that sildenafil administration is associated with increases in cardiac index, haemodynamic changes in the response to exercise and alterations in the concentrations of cGMP and cAMP in the heart (45–47). Studies in animal models have also suggested sildenafil may have inotropic effects under certain conditions, with some studies demonstrating an increase in cardiac inotropy with infusion of sildenafil in anaesthetized animals (48), or a sildenafil induced decrease in the rate of cAMP breakdown (49). Clearly whether sildenafil does cause changes in cardiac contraction is dependent on many factors including the species, dosage, route of administration and any underlying pathology. Interestingly there was no evidence of this when cold cardioplegia was used; it is likely that any inotropic effects were offset by the cold temperature, which is known to depolarize the heart (50). Any inotropic influence of drugs added to cardioplegia (a solution which aims to arrest the heart) is undesirable, and this effect requires further investigation.

In both the warm and cold cardioplegia protocols, sildenafil had some beneficial effects on cardiac metabolism. For both protocols, ATP levels decreased over the period of ischaemia and early reperfusion, and in both control groups the concentration of ATP at 20 min reperfusion was significantly reduced compared to the basal levels. This is in concordance with previous work which has demonstrated that, upon reperfusion, cardiac function recovers faster than ATP levels (51). There is also a known association between decreased ATP/ADP and ATP/AMP ratios and worse cardiac function (52). Both the warm and cold cardioplegia protocols demonstrated signs of improved ATP metabolism with sildenafil treatment, with a significantly greater ATP concentration at 20 min reperfusion in the sildenafil group compared to the control group of the cold cardioplegia protocol and a significantly improved ATP/ADP ratio at 20 min reperfusion in the sildenafil group of the warm cardioplegia protocol (a similar trend was seen in the cold protocol). Sildenafil treatment in the cold cardioplegia protocol also significantly improved energy charge at 20 min reperfusion, with a similar trend in the warm cardioplegia protocol. Energy charge is calculated from the concentrations of ATP, ADP and AMP and has been found to correlate well with the levels of enzymes involved in ATP utilization—hence it has been described as a measure of metabolic flux (41).

In the control groups of both protocols and the sildenafil group of the warm protocol, there was a significant increase in the concentrations of hypoxanthine and inosine compared to basal levels by the end of ischaemia. Hypoxanthine and inosine are purine catabolites, known to be accumulated during ischaemia and released during reperfusion (53, 54). In the warm cardioplegia protocol, both metabolites were significantly lower compared to their respective control groups at 20 min reperfusion (and similar trends were measurable at the end of ischaemia). Whether this is driven mainly by reduced accumulation during ischaemia or faster release during reperfusion requires more investigation, especially as blockade of the release of the purine metabolites has been linked to cardioprotection (54).

Piglets in all treatment groups demonstrated an increase in plasma troponin concentration throughout the experiment, as expected. There is no apparent effect of sildenafil treatment. However, previous experiments of I/R injury in pigs have suggest that troponin levels will continue to rise significantly following this initial 60 min reperfusion period, likely peaking at around 12 h (55, 56). Therefore, our results likely only show a small portion of the troponin response to surgery.

The current study did not investigate the mechanisms of action by which sildenafil is cardioprotective. This was due to a limitation in the amount of cardiac tissue that could be collected mid protocol without compromising cardiac function. Further experiments investigating changes in PKG, PKC and NO signalling pathways are required, and direct measurements of cardiac function would also be beneficial. As mentioned above, the study was also limited by the short reperfusion period. Use of a longer reperfusion period or even recovery of the animals would be beneficial to investigate the longer-term effects of cardioplegia supplementation with sildenafil. Another limitation is that these are healthy pigs, whilst cardiac surgery is never performed on a healthy heart. Study of how PDE-5 expression is affected by different pathologies and testing of sildenafil in an animal model with a diseased heart would be highly beneficial. Dose response curves and testing of the timing of sildenafil administration would also be informative—for example it is possible that sildenafil may be more effective if given prior to surgery. Another limitation is the fact that only female pigs were used, meaning that sex was controlled for but not investigated. Further study in male pigs is required to investigate any intersex differences.