Key words : Early allograft dysfunction, Hyperoxia, Ischemia-reperfusion syndrome, Postreperfusion syndrome

Introduction

Liver transplantation (LT) is the definitive treatment for end-stage liver disease and unresectable hepatic malignancies. The demand for donor organs has grown as a result of more recipients with greater frailty, more multisystem comorbidities, and increasingly severe complications of liver disease being accepted as candidates for transplant.¹ This growing need has been met by the use of more marginal grafts, including from donors after cardiac death and steatotic grafts from older patients.^2,3 The use of these marginal grafts has been linked to greater instability at graft reperfusion and more postoperative graft dysfunction.^4,5 This instability, known as postreperfusion syndrome (PRS), is a common and serious intraoperative complication following reperfusion of the transplanted liver that presents as hemodynamic instability and metabolic derangement. With a reported incidence of 12% to 77%,^5-7 PRS was first defined by Aggarwal and colleagues in 1987 as a decrease in mean arterial pressure > 30% below the baseline value, lasting for at least 1 minute, and occurring during the first 5 minutes after reperfusion of the liver graft.^2,6 The pathophysiology of PRS is complex, multifactorial, and remains to be fully elucidated but is interlinked with generation of reactive oxygen species (ROS) and radicals.⁷

In part, PRS represents the systemic response to the immediate phase of ischemia-reperfusion injury,⁷ a combined hepatic insult derived from the sudden cessation of perfusion during clamping of the graft’s vascular supply at retrieval and its abrupt restoration after unclamping at reperfusion.⁸ Multiple coinci-dental pathways in which ROS play a fundamental role can lead to liver injury with hepatocellular death and subsequent graft dysfunction that negatively affects long-term graft outcomes.^9-11 Ischemia-reperfusion injury also plays a fundamental role in the development of early allograft dysfunction (EAD), affecting up to 40% of organs,¹² describing poor function in the first week post-transplant that may necessitate retransplant and increasing morbidity and mortality risk.

Oxygen is no longer regarded as a benign drug, and both hyperoxia and unnecessary normoxia have been associated with poor outcomes throughout medicine, surgery, and intensive care.¹³ Evidence to date regarding oxygen therapy during transplant has focused on either the case-average inspired oxygen fraction¹⁴ or a snapshot arterial oxygen tension at reperfusion,¹⁵ both of which are associated with worse outcomes. However, hepatic and systemic oxygen exposure over time is likely to be the key determinant of oxygen’s effect and may occur in a dose-dependent manner. This requires serial arterial blood gas data to ascertain the arterial oxygen tension throughout surgery. Hence, we have assessed the effects of hepatic oxygen exposure throughout surgery, both before and after reperfusion, on PRS, EAD, and other outcomes. Our findings could help clarify optimal perioperative oxygen therapy for anesthetists and intensive care physicians.

Materials and Methods

Study design, population, and setting
This retrospective, observational cohort study reported on adult patients undergoing deceased donor LT at the Royal Free NHS Foundation Trust. All sequential patients undergoing LT between February 2017 to June 2019 were included; however, patients undergoing super-urgent transplant and retransplant procedures were excluded because of higher oxygen requirements, disease-related respira-tory failure, surgical complexity, and longer procedural times. Patients with incomplete data preventing analysis were also excluded. Surgical and anesthetic management techniques were as per departmental standard practice for the duration of our data collection.

Data sources and collection
Demographic and perioperative details of all patients undergoing LT had been prospectively collected as part of an ongoing quality improvement program since February 2017 using an electronic case report form created in REDCap (Vanderbilt University).

For analyses, we included baseline recipient characteristics (age, sex, body mass index), disease severity and causation (UK Model for End-Stage Liver Disease [UKELD] score, liver disease etiology, transplant indication), comorbidities, and routine medications. We also included donor data on age and donation type (cardiac or brainstem death). For organ data, we included whether the procedure was whole or split graft, the cold ischemia time, and the degree of steatosis. For operative data, we included surgical technique (piggy-back technique vs caval replacement); duration of each surgical phase; total crystalloid, colloid, and red cells infused; oxygena-tion data (fraction of inspired oxygen, Pao2); cardiovascular metrics (systolic blood pressure, lactate, inotrope/vasopressor dose, heart rate); and severity of reperfusion syndrome. For postoperative data, we included liver function (bilirubin, alanine aminotransferase, aspartate aminotransferase, inter-national normalized ratio) to postoperative day 7 and length of intensive care unit (ICU) and hospital stays.

Definitions and study endpoints
Time-weighted oxygen exposure was calculated by the area under the curve (AUC) method, similar to previous studies.^16,17 The arterial partial pressure of oxygen on arterial blood gas analysis was deter-mined to have lasted from the time of that arterial blood gas sampling to the subsequent sampling, from the start of the particular stage of the LT until the subsequent sampling, or from the time of sampling until the end of the LT stage (as per (Figure 1)). We calculated the AUC for each discrete arterial blood gas sample period. We calculated the sum of the AUCs for specific procedure stages or periods of interest (total prereperfusion, total whole operation) to generate the oxygen exposure for each period.

We have developed a definition of PRS based on physician intervention rather than an absolute or relative fall in systolic blood pressure. Because it is increasingly accepted that few if any anesthetists will await severe instability before intervention, profound falls in blood pressure are less commonly seen with this proactive approach. We have proposed a grading of severity of PRS dependent on the choice and quantity of drug required to maintain hemodynamic stability (Table 1).

Endpoints were determined a priori. Primary endpoints of interest were severity of PRS and EAD. We allocated patients to PRS categories based on the peak hemodynamic support required within 20 minutes of reperfusion (Table 1). We calculated severity of EAD by using the Model for Early Allograft Function (MEAF) score¹⁸ and used criteria laid out by Dhillon and colleagues¹⁹ for incidence. Secondary endpoints of interest were intensive care unit and hospital lengths of stay. We defined a long critical care (intensive care unit) or hospital length of stay as those in the upper quartile versus the lower 3 quartiles (6 vs 23 days).

Data handling and statistical analyses
We exported REDCap data to Excel (Microsoft Corporation) and added further perioperative data via review of anesthesia charts by a single investigator. We used SPSS version 25 (IBM) for data analyses. We presented categorical data as number and percent. All continuous data distributions were non-Gaussian (Kolmogorov-Smirnov test with Lilliefors correction, P > .05). P values (2-tailed) with an α = 0.05 were considered as significant. We made comparisons by using the Pearson X², Fisher exact, or Mann-Whitney U tests, as appropriate.

We used binary logistic regression to assess associations between independent variables and nominal outcomes. We assessed possible associations between continuous outcome measures and independent measures in which a linear relationship was feasible by multiple linear regression. We initially included all independent variables and optimized the model via backward conditional elimination, with entry and removal probabilities of 0.05 and 0.1, respectively. For analysis of associations with PRS, we included preoperative and graft risk factors, prereperfusion oxygenation factors, and anhepatic cardiovascular variables. For analysis of associations with EAD, we included all patient, preoperative, and graft factors as well as reperfusion cardiovascular and oxygenation risk factors. We calculated goodness of fit by using the Hosmer-Lemeshow test for logistic regression and overall model r² for linear regression.

Ethical approval
The local research and development office did not require ethical approval and consent as the database was prospectively registered as a quality impro-vement dataset. We adhered to the STROBE checklist for the reporting of observational studies²⁰.

Results

Study population
During the data capture period, 255 patients underwent deceased donor LT and were considered for inclusion; among these, 36 patients received a super-urgent transplant, 17 underwent retransplant, and 16 had incomplete data and were excluded from the analysis. The final dataset included 185 transplant recipients (Table 2).

Transplant recipients were predominately male (66.3%) with a median age of 57 years and a UKELD score of 53; 93 patients (50.3%) had PRS, with 39 mild cases, 38 moderate cases, and 16 severe cases (Table 1). The incidence of EAD was 26.0%. No significant differences in baseline and preoperative characteristics were shown between patient groups with moderate or severe PRS or EAD and patients without these complications (Table 2). Preexisting lung disease was uncommon; the overall prevalence of asthma was 9.2% and chronic obstructive pulmonary disease was 3.2%. Patients with any form of lung disease were not more likely to have PRS or EAD, did not receive higher oxygen exposures or fractions, did not have lower Pao2, and did not have longer ICU or hospital lengths of stay.

Most organs were from donors after brain death (75.1%, (Table 3)) with few split grafts (4.9%) and few grafts with moderate steatosis (10.7%). Median cold ischemia time was 8.5 hours, in keeping with international standards,²¹ with most procedures performed with piggy-back technique. Median operative time was 454 minutes. Median ICU and hospital lengths of stay were 3 and 15 days, respectively. Patients who had moderate or severe PRS were more likely to have a higher oxygen exposure during the anhepatic phase (1.5 vs 1.4 kPa. min; P .029). No other significant differences were shown between patents with and without moderate or severe PRS.

Patients with EAD were more likely to have received a graft with moderate steatosis (21.7% vs 10.7%; P = .018), had a longer cold ischemia time (median 9.6 vs 8.5 h; P < .001), and had a higher maximum noradrenaline requirement during the anhepatic phase (0.18 vs 0.11 µg/kg/min; P = .021). These patients also had higher arterial lactate concentrations before the anhepatic phase (1.84 vs 1.62 mmol/L; P = .012); however, these levels did not persist into the anhepatic phase. No differences were shown between the groups in terms of raw oxygen exposures or other surgical factors. Postoperative ICU (5.3 vs 3.0 days; P < .001) and hospital (21.4 vs 15.3 days; P < .001) lengths of stay were prolonged for patients with EAD.

Primary endpoints
Total oxygen exposure during the anhepatic phase was found to be independently associated with presence of moderate or severe PRS (OR = 1.041; P = .007) (Table 4). Other markers of oxygen exposure, duration of surgical phases, and metrics of cardiovascular instability (maximum noradrenaline dose or lactate) were not retained in the model. Donor age (OR = 1.022; P = .042) and undergoing transplant for cirrhosis (OR = 4.528; P = .013) were also associated with increased risk of moderate or severe PRS. Patients taking beta-blockers regularly before transplant had an independently lower risk of PRS (OR = 0.414; P = .027). Model goodness of fit was reasonable (X² test = 7.000; P = .537).

The total oxygen exposure after reperfusion was independently associated with the severity of EAD (beta = 0.174; P = .011) (Table 5). Additional independent risk factors included cold ischemia time (beta = 0.268; P < .001), donor age (beta = 0.242; P = .001), caval replacement technique (beta = 0.174; P = .011), receiving an organ from a donor after cardiac death (beta = 0.149; P = .031), and use of a split liver graft (beta = 0.170; P = .030). The overall model goodness of fit was poor (r² = 0.203).

Secondary endpoints
Independent associations with a prolonged posto-perative ICU length of stay were recipient history of diabetes (OR = 2.230; P = .043), total operative time (OR = 1.006; P = .003), and presence of EAD (OR = 3.045; P = .005) (Table 6). Similarly, independent associations with prolonged hospital stay included a higher recipient UKELD score (OR = 1.096; P = .014), frailty (OR = 7.681; P = .011), and a caval replacement surgical approach (OR = 3.835; P = .003), whereas higher body mass reduced the risk (OR = 0.971; P = .035). The presence of EAD was independently associated with prolonged hospitalization (OR = 7.738; P < .001) as was the presence of PRS at reperfusion (overall P = .008), particularly severe PRS (OR = 10.074; P = .001) (Table 7).

Discussion

Main findings
We have shown for the first time that higher intraoperative oxygen exposure independently associates with worse intra- and postoperative outcomes in patients undergoing deceased donor LT. The total anhepatic oxygen exposure (kPa.h), using the AUC of serial blood gas measurements, was independently associated with an increased risk of moderate or severe PRS with an odds ratio of 1.041 (P = .007). Similarly, the total oxygen exposure throughout surgery (kPa.h) was independently associated with an increased severity of posto-perative EAD, based on the MEAF score, with a coefficient of 0.174 (P = .011). Furthermore, the presence of EAD was found to be independently associated with prolonged ICU (OR = 3.045; P = .005) and hospital stay (OR = 7.738; P < .001).

Strengths and limitations of this study
Our study had several limitations. First, this study was limited by its single-center design with a time-limited sample rather than an a priori sample size calculation, which reduced its transferability and power to draw conclusions. By definition, asso-ciations can only be inferred from work of this nature, and these do not demonstrate causation. Second, although the risk factors included in our analyses were broad and encompassed many of those recognized to affect graft and patient outcomes, some unaccounted variables and confounding by association will remain. This is more likely in the analysis for EAD (r² = 0.203) than PRS (P= .537).

Strengths of this study include the prospective nature of the data collection and the a priori definition of outcomes. Furthermore, data quality was good with only 6.3% of cases removed before analysis because of lack of data; there were no losses to follow-up. The sample size of 185 consecutive patients undergoing LT at a major transplant center aligns with other similar studies.^14,15 The transplant population at the Royal Free Hospital is reasonably representative of the UK transplant population, although we did not describe ethnicity in our patient data, which could have supported this assertion.

Our use of multiple serial blood gas data throughout surgery and detailed calculations of the AUC to determine patient oxygen exposure at various phases is an important strength of this study. This strength also helps to mitigate some of this confounding due to more unwell patients at the time of transplant who were likely to have received higher fractions of oxygen and have a higher or abnormal oxygen extraction. Analysis of PRS severity is difficult as there are no standardized or agreed-on criteria on which to define severity, only incidence. However, the approach taken here allows a simple and reproducible method to categorize PRS. Clearly, the oxygen exposure in each surgical phase will correlate with surgical time; however, the duration of each surgical phase was included in the analysis and did not associate with PRS or EAD incidence in either univariate or multivariate analyses.

Interpretation of findings
The incidence of moderate and severe PRS in our dataset was 29.2%, comparable with the 25% to 30% incidence described previously.^6,22 The only other study to investigate risk factors for PRS identified the absence of a porto-caval shunt and the duration of cold ischemia time as risk factors.²² Interestingly, neither of these risk factors were found to be independently associated with moderate or severe PRS in our analysis, although this study only included 20 patients with PRS from a cohort of 75 transplant procedures. This study defined PRS in the traditional manner by changes in mean arterial pressure, while we have focused on clinical management of PRS. The lack of association between cold ischemia time and PRS may be explained by the growing use of extracorporeal machine perfusion, which is likely to reduce the effect of cold ischemia in this cohort. More marginal organs with a longer cold ischemia time would also have been more likely to receive machine perfusion, further reducing the effect of this risk factor. The effect of cold ischemia time may be more substantial after 7 hours of cold storage with additional risk per hour beyond that point²³; the distribution of cold ischemia time in our population was quite narrow (8.5 [25% and 75% quartile, 7.0, 10.0]) such that cold ischemia time-related risk may be relatively even across the population. Interestingly, however, cold ischemia time remained as an independent association with EAD.

The negative association between long-term recipient use of beta-blockers and the presence of moderate or severe PRS may imply either a protective effect from beta blockade or that patients with a history of cirrhotic complications were at lower risk. The finding that patients with cirrhosis as their primary indication for transplant were at higher risk of PRS (OR = 4.528; P = .013) would suggest that the latter explanation is less likely. Livers donated by older donors appeared to carry a higher risk of PRS, although with a somewhat weak significance (P = .042), which may reflect a lower ability of the older liver to tolerate the donation process and hence induce a greater metabolic insult on reperfusion.

The finding that total anhepatic oxygen exposure was a key independent association with the presence of moderate or severe PRS was surprising but may fit with the concept of free radical generation as part of the process. Hypoxic hepatocytes in the donor organ are unable to regenerate ATP from adenosine diphosphate, which is instead further degraded through several intermediate molecules until xanthine is generated. On restoration of oxygen delivery, this molecule is degraded by xanthine oxidase to uric acid, releasing the free radical superoxide and triggering a cascade leading to further production of ROS.²⁴ The resultant oxidative stress causes cell membrane disruption, microvascular dysfunction, and further inflammatory activation. Hence, in the presence of a hyperoxic milieu, the generation of ROS may be either more rapid or enhanced, driving more profound cardiovascular instability. This supposition is in keeping with retrospective data showing that hyperoxic patients undergoing LT had higher lactate levels 15 minutes after reperfusion.¹⁵

Hyperoxia is thought to worsen outcomes in a variety of clinical scenarios in which ischemia-reperfusion injury plays a role, including cardiopu-lmonary bypass with cardioplegia²⁵ and post-cardiac arrest syndrome.²⁶ Hyperoxia exacerbated liver injury in a murine model of liver ischemia,²⁷ and an assessment of normothermic ex situ liver perfusion indicated that grafts exposed to hyperoxic conditions preimplantation led to a greater incidence of PRS than those perfused at normal oxygen tensions.²⁸

The incidence of EAD (26.0%) was identical to that described by Dhillon and colleagues.¹⁹ Previous analyses of EAD have demonstrated independent associations with donor age and split LT and a trend toward a higher intraoperative inspired oxygen fraction as potential risk factors,¹⁴ in keeping with our findings. This study also demonstrated that a higher inspired oxygen concentration during transplant surgery was associated with lower graft survival up to 10 years postoperatively. Patients who are hyperoxic have been found to have higher rates of hepatic cytolysis within the first 7 days post-transplant with an associated higher incidence of EAD.¹⁵ Interestingly, the hyperoxic group in this study also had a longer period of intubation on ICU and higher incidence of ileus, in keeping with our data showing that EAD associates with longer ICU and hospital stay, although we did not find any direct association between higher oxygen exposure and length of stay. Ischemia-reperfusion injury is a recognized determinant of the development of EAD,¹¹ and hyperoxia at the time of reperfusion and during the subsequent injury cascade is likely to worsen outcomes.²⁹ The generation of ROS is enhanced by cellular and mitochondrial hyperoxia, leading to an earlier and more profound saturation of cellular antioxidant processes and reduced ROS compartmentalization. This is compounded in the reperfusion and post-transplant phase due to hepatocyte death and dysfunction.

Conclusions

Higher oxygen exposure during LT may lead to worse patient and organ outcomes, and reducing unnecessary hyperoxia may lead to improved intra- and postoperative outcomes. However, unlike other settings where benefit-risk clearly lies toward the minimum safe arterial oxygen tension during surgery, LT can have complex and life-threatening complications; as such, a buffer in available oxygen is frequently warranted. At reperfusion, hyperoxia is likely to be detrimental to short- and long-term outcomes; however, hypoxemia and deranged oxygen metabolism (especially in the liver graft) are common, and a preemptive increase in oxygen fraction is an appropriate and simple therapy to minimize or prevent this. However, minimizing the arterial oxygen tension during the surgery before preparing for reperfusion is a simple intervention that may lead to improved outcomes.

References:

1. Kriss M, Biggins SW. Evaluation and selection of the liver transplant candidate: updates on a dynamic and evolving process. Curr Opin Organ Transplant. 2021;26(1):52-61. doi:10.1097/MOT.0000000000000829
   CrossRef - PubMed
2. Bodzin AS, Baker TB. Liver transplantation today: where we are now and where we are going. Liver Transpl. 2018;24(10):1470-1475. doi:10.1002/lt.25320
   CrossRef - PubMed
3. Jadlowiec CC, Taner T. Liver transplantation: current status and challenges. World J Gastroenterol. 2016;22(18):4438-4445. doi:10.3748/wjg.v22.i18.4438
   CrossRef - PubMed
4. Mazilescu LI, Kotha S, Ghanekar A, et al. Early allograft dysfunction after liver transplantation with donation after circulatory death and brain death grafts: does the donor type matter? Transplant Direct. 2021;7(8):e727. doi:10.1097/TXD.0000000000001182
   CrossRef - PubMed
5. Chapman WC, Barbas AS, D'Alessandro AM, et al. Normothermic machine perfusion of donor livers for transplantation in the United States: a randomized controlled trial. Ann Surg. 2023;278(5):e912-e921. doi:10.1097/SLA.0000000000005934
   CrossRef - PubMed
6. Aggarwal S, Kang Y, Freeman JA, Fortunato FL, Pinsky MR. Postreperfusion syndrome: cardiovascular collapse following hepatic reperfusion during liver transplantation. Transplant Proc. 1987;19(4 Suppl 3):54-55.
   CrossRef - PubMed
7. Jeong SM. Postreperfusion syndrome during liver transplantation. Korean J Anesthesiol. 2015;68(6):527-539. doi:10.4097/kjae.2015.68.6.527
   CrossRef - PubMed
8. Cannistra M, Ruggiero M, Zullo A, et al. Hepatic ischemia reperfusion injury: a systematic review of literature and the role of current drugs and biomarkers. Int J Surg. 2016;33(Suppl 1):S57-S70. doi:10.1016/j.ijsu.2016.05.050
   CrossRef - PubMed
9. Wu MY, Yiang GT, Liao WT, et al. Current mechanistic concepts in ischemia and reperfusion injury. Cell Physiol Biochem. 2018;46(4):1650-1667. doi:10.1159/000489241
   CrossRef - PubMed
10. Abu-Amara M, Yang SY, Tapuria N, Fuller B, Davidson B, Seifalian A. Liver ischemia/reperfusion injury: processes in inflammatory networks--a review. Liver Transpl. 2010;16(9):1016-1032. doi:10.1002/lt.22117
    CrossRef - PubMed
11. Ito T, Naini BV, Markovic D, et al. Ischemia-reperfusion injury and its relationship with early allograft dysfunction in liver transplant patients. Am J Transplant. 2021;21(2):614-625. doi:10.1111/ajt.16219
    CrossRef - PubMed
12. Zhou J, Chen J, Wei Q, Saeb-Parsy K, Xu X. The role of ischemia/reperfusion injury in early hepatic allograft dysfunction. Liver Transpl. 2020;26(8):1034-1048. doi:10.1002/lt.25779
    CrossRef - PubMed
13. Nakane M. Biological effects of the oxygen molecule in critically ill patients. J Intensive Care. 2020;8(1):95. doi:10.1186/s40560-020-00505-9
    CrossRef - PubMed
14. Miyachi Y, Kaido T, Hirata M, et al. Intraoperative high fraction of inspiratory oxygen is independently associated with worse outcome after living-donor liver transplantation: a retrospective study. World J Surg. 2022;46(7):1776-1787. doi:10.1007/s00268-022-06544-7
    CrossRef - PubMed
15. Reydellet L, Le Saux A, Blasco V, et al. Impact of hyperoxia after graft reperfusion on lactate level and outcomes in adults undergoing orthotopic liver transplantation. J Clin Med. 2023;12(8). doi:10.3390/jcm12082940
    CrossRef - PubMed
16. Elmer J, Wang B, Melhem S, et al. Exposure to high concentrations of inspired oxygen does not worsen lung injury after cardiac arrest. Crit Care. 2015;19(1):105. doi:10.1186/s13054-015-0824-x
    CrossRef - PubMed
17. Ketharanathan N, De Jonge RCJ, Klouwen I, et al. Hyperoxia in pediatric severe traumatic brain injury (TBI): a comparison of patient classification by cutoff versus cumulative (area-under-the-curve) analysis. Brain Inj. 2020;34(7):958-964. doi:10.1080/02699052.2020.1765021
    CrossRef - PubMed
18. Jochmans I, Fieuws S, Monbaliu D, Pirenne J. “Model for early allograft function” outperforms “early allograft dysfunction” as a predictor of transplant survival. Transplantation. 2017;101(8):e258-e264. doi:10.1097/TP.0000000000001833
    CrossRef - PubMed
19. Dhillon N, Walsh L, Kruger B, et al. A single nucleotide polymorphism of Toll-like receptor 4 identifies the risk of developing graft failure after liver transplantation. J Hepatol. 2010;53(1):67-72. doi:10.1016/j.jhep.2009.12.044
    CrossRef - PubMed
20. von Elm E, Altman DG, Egger M, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Int J Surg. 2014;12(12):1495-1499. doi:10.1016/j.ijsu.2014.07.013
    CrossRef - PubMed
21. Lozanovski VJ, Dohler B, Weiss KH, Mehrabi A, Susal C. The differential influence of cold ischemia time on outcome after liver transplantation for different indications---who is at risk? A collaborative transplant study report. Front Immunol. 2020;11:892. doi:10.3389/fimmu.2020.00892
    CrossRef - PubMed
22. Paugam-Burtz C, Kavafyan J, Merckx P, et al. Postreperfusion syndrome during liver transplantation for cirrhosis: outcome and predictors. Liver Transpl. 2009;15(5):522-529. doi:10.1002/lt.21730
    CrossRef - PubMed
23. Pan ET, Yoeli D, Galvan NTN, et al. Cold ischemia time is an important risk factor for post-liver transplant prolonged length of stay. Liver Transpl. 2018;24(6):762-768. doi:10.1002/lt.25040
    CrossRef - PubMed
24. Cowled P, Fitridge R. Pathophysiology of reperfusion injury. In: Fitridge R, Thompson M, eds. Mechanisms of Vascular Disease: A Reference Book for Vascular Specialists. University of Adelaide Press; 2011.
    CrossRef - PubMed
25. Peng YW, Mohammed A, Deatrick KB, et al. Differential effects of normoxic and hyperoxic reperfusion on global myocardial ischemia-reperfusion injury. Semin Thorac Cardiovasc Surg. 2019;31(2):188-198. doi:10.1053/j.semtcvs.2018.09.018
    CrossRef - PubMed
26. Binks A, Nolan JP. Post-cardiac arrest syndrome. Minerva Anestesiol. 2010;76(5):362-368.
    CrossRef - PubMed
27. Zangl Q, Martignoni A, Jackson SH, et al. Postoperative hyperoxia (60%) worsens hepatic injury in mice. Anesthesiology. 2014;121(6):1217-25. doi:10.1097/ALN.0000000000000447
    CrossRef - PubMed
28. Watson CJE, Kosmoliaptsis V, Randle LV, et al. Normothermic perfusion in the assessment and preservation of declined livers before transplantation: hyperoxia and vasoplegia-important lessons from the first 12 cases. Transplantation. 2017;101(5):1084-1098. doi:10.1097/TP.0000000000001661
    CrossRef - PubMed
29. Checa J, Aran JM. Reactive oxygen species: drivers of physiological and pathological processes. J Inflamm Res. 2020;13:1057-1073. doi:10.2147/JIR.S275595
    CrossRef - PubMed