Telephone: (973) 972-7218; FAX: (973) 972-6227
Ischemic preconditioning in deceased donor liver transplantation: A prospective randomized clinical trial of safety and efficacy
Article first published online: 21 JAN 2005
Copyright © 2005 American Association for the Study of Liver Diseases
Volume 11, Issue 2, pages 196–202, February 2005
How to Cite
Koneru, B., Fisher, A., He, Y., Klein, K. M., Skurnick, J., Wilson, D. J., de la Torre, A. N., Merchant, A., Arora, R. and Samanta, A. K. (2005), Ischemic preconditioning in deceased donor liver transplantation: A prospective randomized clinical trial of safety and efficacy. Liver Transpl, 11: 196–202. doi: 10.1002/lt.20315
- Issue published online: 21 JAN 2005
- Article first published online: 21 JAN 2005
Ischemic preconditioning (IPC) has the potential to decrease graft injury and morbidity after liver transplantation. We prospectively investigated the safety and efficacy of 5 minutes of IPC induced by hilar clamping in local deceased donor livers randomized 1 : 1 to standard (STD) recovery (N = 28) or IPC (N = 34). Safety was assessed by measurement of heart rate, blood pressure, and visual inspection of abdominal organs during recovery, and efficacy by recipient aminotransferases (aspartate aminotransferase [AST] and alanine aminotransferase [ALT], both measured in U/L), total bilirubin, and international normalized ratio of prothrombin time (INR) after transplantation. IPC performed soon after laparotomy did not cause hemodynamic instability or visceral congestion. Recipient median AST, median ALT, and mean INR, in STD vs. IPC were as follows: day 1 AST 696 vs. 841 U/L; day 3 AST 183 vs. 183 U/L; day 1 ALT 444 vs. 764 U/L; day 3 ALT 421 vs. 463 U/L; day 1 INR 1.7 ± .4 vs. 2.0 ± .8; and day 3 INR 1.3 ± .2 vs. 1.4 ± .3; all P > .05. No instances of nonfunction occurred. The 6-month graft and patient survival STD vs. IPC were 82 vs. 91% and median hospital stay was 10 vs. 8 days; both P > .05. In conclusion, deceased donor livers tolerated 5 minutes of hilar clamping well, but IPC did not decrease graft injury. Further trials with longer periods of preconditioning such as 10 minutes are needed. (Liver Transpl 2005;11:196–206.)
The shortage of livers remains a critical limiting factor in applying transplantation therapy to more patients in need of such a treatment, and is evidenced by the huge disparity between the number of patients on the waiting list and the number of transplantations performed each year.1 Utilization of livers from marginal deceased donors has increased, to partially alleviate the shortage of donors.2 Utilization of livers from marginal donors incurs an increase in the risks of initial poor function of the liver grafts and donor disease transmission.3–6 Initial poor function and primary nonfunction of liver allografts increase morbidity and resource utilization following transplantation.7, 8 Thus measures to decrease preservation / reperfusion (RP) injury and improve the initial function of liver grafts would not only decrease the morbidity and resource utilization after liver transplantation, but would also have the potential to increase the number of livers available for transplantation.
Ischemic preconditioning (IPC) is a counterintuitive phenomenon induced by brief periods of ischemia, which protects diverse organs from subsequent episodes of more protracted or sustained ischemia. Initially described by Murry et al.9 in a canine model of myocardial ischemia, IPC occurs in many species, including man, and in diverse organs including liver, lungs, intestine, and kidneys.9–15 In experimental animals, IPC of the liver affords protection against both warm and cold ischemia, including RP injury, following liver transplantation.11, 16, 17 Recently, the clinical efficacy of IPC has been reported in humans undergoing hepatic resection under inflow occlusion.15, 18, 19
To our knowledge, the feasibility and efficacy of hepatic IPC has not been reported in deceased donor liver transplantation in man except in a very recent abstract.20 Deceased organ donors are hemodynamically labile and frequently require 1 or more vasopressors. How hilar clamping is tolerated and whether hilar clamping induces IPC in such donor livers is not known. Therefore, a prospective randomized clinical study was undertaken to examine the safety of hepatic hilar clamping in deceased donor livers and its efficacy to induce IPC and decrease liver graft injury after transplantation.
Materials and Methods
Between March 2002 and May 2003, all deceased donors in the local organ procurement organization whose livers were allocated to adult (≥18 years) recipients at our center were eligible to be included in the study. The study was approved by the local human research committee. The following donor groups were excluded: 1) combined liver and intestine donors; 2) donors for pediatric liver recipients; 3) donors for recipients at another transplant center; and 4) donor livers recovered outside the local organ recovery network for use by local recipients. Deceased donors were randomized after liver allocation was completed to either standard organ recovery (standard [STD] group) or to 5 minutes of hilar clamping performed during liver recovery (IPC group) in a 1 : 1 ratio. Randomization was performed according to a computer-generated table of simple random numbers enclosed in numbered and sealed envelopes, which were opened only after donor recovery time was scheduled.
In donors randomized to the IPC group, the hepatic hilum was clamped with an atraumatic vascular clamp for 5 minutes after laparotomy, after inspection and mobilization of the liver. After hilar clamping, the peritoneal cavity was inspected for any bleeding and severe congestion and for edema of the pancreas and small intestine. In the 1st 2 donors in the IPC group, hilar clamping was performed 15–20 minutes before aortic cross-clamping (before heparinization). This resulted in considerable intraperitoneal bleeding and marked edema of the small intestine. Therefore, hilar clamping was subsequently performed soon after laparotomy and before any significant intraperitoneal dissection was performed. Wedge and needle biopsies of the liver were performed from the right lobe immediately after the abdomen was opened. Liver recovery was performed in a standard manner with dissection of the structures at the base of the hilum. Allografts were perfused in situ with 2 L of University of Wisconsin solution via both the portal vein and the infrarenal abdominal aorta. Allografts were stored in 1 L of University of Wisconsin solution on ice until transplantation. The following donor variables were collected: age, gender, cause of death, occurrence of cardiac arrest, use of vasopressors, number of hospital days, body mass index, presence of macrosteatosis, and serum sodium, aspartate aminotransferase (AST), and alanine aminotransferase (ALT) levels before organ recovery.5, 21–26
Recipient operation was performed in a standard manner with preservation of the inferior vena cava. No veno-venous bypass was utilized in a majority of the recipients. Graft RP followed completion of suprahepatic caval and portal vein anastomoses. After completion of the arterial and biliary anastomoses, a needle biopsy was taken from the right lobe immediately before the abdomen was closed. Immunosuppression consisted of tacrolimus and steroids. Interleukin (IL)-2 receptor antibody was used in recipients without hepatitis C. Recipient variables collected were age, gender, model for end-stage liver disease score, units of blood transfused during surgery, and graft cold and warm ischemia times. The end points of the safety part of the study were heart rate and blood pressure during organ recovery in all donors and visual inspection of the abdominal organs during and after hilar clamping in IPC donors. The primary endpoint of the efficacy part of the study was graft RP injury and function assessed by serum AST, ALT, total bilirubin, and international normalized ratio of prothrombin time (INR) on days 1, 3, and 7 and the histological score of RP biopsies (see below). The secondary end points were graft and patient survival at 6 months and length of hospital stay after the transplant. Based on earlier data, we expected STD group recipients to have peak AST and ALT levels, with standard deviations approaching the mean.26 Compared to the STD group, we expected to show a 50% decrease in serum AST and ALT levels in the IPC group.
A power analysis indicated that 126 patients (63 in each group) would provide 80% power with a 5% probability of type I error. The human research committee of our institution and the organ recovery network's physician committee approved the conduct of the study with a requirement to review the data on a periodic basis for any trends of potential harm to the recipients. Informed consent was obtained from the prospective liver recipients. An exemption from consent by the donor's next of kin, apart from that for organ donation, was granted by the human research committee.
Formalin-fixed and paraffin-embedded liver tissue sections were stained with hematoxylin and eosin and examined under a light microscope by a hepatopathologist who had no knowledge of patient assignment. The primary focus in the review of donor liver biopsies was to determine the presence and extent of steatosis. All of the biopsied tissue was scanned at a magnification of 40× and the percentage of the liver occupied by fat was determined in a semiquantitative manner. Tissue from postperfusion liver biopsies was examined at a magnification of 60× and RP injury was assessed semiquantitatively (0-10 by increasing severity), based on the histological features of hepatocyte swelling, the extent of hemorrhage, and necrosis around central veins.27 Apoptotic cells were identified by the morphological criteria of cell shrinkage, acidophil staining of the cytoplasm, and nuclear pyknosis (acidophil / Councilman bodies), and the cells were enumerated.28 At a minimum, 3 fields were examined around each of 3 different central veins per biopsy. A mean of scores of each biopsy was used in data analysis.
All study data were collected prospectively. All statistical comparisons were performed as 2-tailed tests at alpha = .05. Preliminary descriptive analyses were performed to indicate distributional features of risk factors and outcomes (i.e., approximate normality, extreme values, or skewing). T tests were performed to compare the 2 treatment groups by donor and recipient characteristics. Chi squared or Fisher's exact tests were used to compare the treatment groups by gender distribution of donors and recipients and proportion of donors with steatosis. Steatosis grades in donor liver biopsies, RP biopsy histology scores, AST and ALT values, and duration of hospital stay of the STD and IPC patients were compared by rank-sum tests. Graft and patient survival times were compared by log-rank tests for Kaplan-Meier curves. Two-way analysis of variance (general linear models) was used to test for interaction of donor age, presence of steatosis, or donor gender on the effect of IPC on peak INR. After peak ALT values were transformed to a logarithmic scale, interactions of risk factors were also tested by 2-way analysis of variance.
A total of 62 patients were enrolled during the 15 months of the study; 28 patients were in the STD group and 34 were in the IPC group. All enrolled subjects completed the study. The data was reviewed after enrollment was 50% complete, and at that time it became apparent that although the secondary efficacy variables showed a trend to a slight benefit in the IPC group, the primary efficacy variables indicated the possibility that RP injury was greater in IPC vs. STD recovery. Therefore, it was considered unethical to continue with the study. The clinical characteristics of both groups of donors are shown in Table 1. Of note, donor age, gender distribution, cause of death, vasopressor use, donor hospital days, body mass index, proportion with steatosis, and peak serum sodium were similar between the 2 groups.
|Donor variable||STD Group (N = 28)||IPC Group (N = 34)|
|Age (years)||46.4 ± 16||46.6 ± 18.1|
|Gender: M/ F||17/ 11||17/ 17|
|Donor death CHI/ Asphyxia/ CVA||8/ 2/ 17||7/ 4/ 23|
|Hospital days||3.1 ± 4.2||2.5 ± 1.8|
|Body mass index||25.4 ± 5.2||26.9 ± 6.5|
|Peak sodium (mEq/L)||158 ± 12||154 ± 11|
|AST (U/L)||36 (23–58)||28 (39–77)|
|ALT (U/L)||22 (30–43)||24 (38–62)|
Safety of IPC in the donor was assessed by the hemodynamic parameters of heart rate and blood pressure and the visual inspection for congestion of the intestines and pancreas and intraperitoneal bleeding. Preconditioning was performed 78.2 ± 31 minutes before aortic cross-clamping and organ perfusion. There were no significant differences between the IPC and STD groups, either in the heart rate or blood pressure during organ recovery. The mean heart rate and the lowest heart rate during organ recovery in the IPC group were 103 ± 20 and 91 ± 16 beats/minute, respectively, in comparison to 97 ± 14 and 85 ± 14 beats/minute, respectively, in the STD group. The mean blood pressure and the lowest systolic blood pressure in the IPC group were 89 ± 11 and 70 ± 11 mm of Hg, respectively, in comparison to 85 ± 13 and 66 ± 10 mm of Hg, respectively, in the STD group. None of the donors in the IPC group required an increase in vasopressors during the period of hilar occlusion. No instances of visceral congestion or bleeding occurred except in the 1st 2 donors in the IPC group, in whom the hilar clamping was performed 15–20 minutes before organ perfusion, as described previously. The donor and recipient data of these 2 donors are included in the analyses. There were no reports of increased edema of pancreata recovered from IPC donors. Additional data on outcomes of pancreata will need institutional review board approval at 4 pancreas transplant centers in our organ recovery organization.
The demographic data of the liver recipients in both groups are shown in Table 2. The recipients in the IPC group were slightly younger when compared to the STD group, but this difference did not reach statistical significance. The model for end-stage liver disease scores and units of blood transfused during the transplant operation were similar between the 2 groups. The cold ischemia times were slightly shorter in the IPC group, but the difference did not reach statistical significance. The mean warm ischemia time in the IPC group was 4 minutes longer than in the STD group, and this difference was statistically significant (P < .01).
|Recipient Variable||Standard Group (N = 28)||IPC Group (N = 34)||P Value|
|Age (years)||58 ± 8||49 ± 10||.08|
|Gender: M/ F||18/ 10||24/ 10||.79|
|MELD score||19.7 ± 7.0||18.8 ± 7.7||.64|
|Transfusions (RBC units)||5.0 ± 4.5||4.4 ± 4.7||.60|
|Cold Ischemia (minutes)||415 ± 87||384 ± 92||.17|
|Warm Ischemia (minutes)||37 ± 5.6||41 ± 5.8||<.01|
The end points for comparison of the primary outcomes between the 2 study groups, serum AST, serum ALT, total bilirubin, and INR are shown in Table 3. None of the differences between the 2 groups were statistically significant. Also, the histology scores of the postperfusion biopsies were similar between the 2 groups. There were no instances of primary nonfunction in either group. Graft failure requiring retransplantation occurred in 1 recipient in the STD group in the 2nd week after transplant from an unclear etiology. An imported graft, from a donor not enrolled in the study, was used to retransplant this patient. The 6-month patient and graft survival rates were 91% in the IPC group in comparison to the 82% in the STD group. These differences were not statistically significant. The hospital stay after liver transplantation was slightly shorter for the recipients in the IPC group in comparison to the STD group (median 8 vs. 10 days), but this difference was not statistically significant.
|Outcome Variable||Standard Group (N = 28)||IPC Group (N = 34)||P Value|
|AST Day 1||696 (362–1,343)||841 (477–1,532)||.43|
|AST Day 3||183 (108–316)||183 (126–2,311)||.42|
|AST Day 7||50 (34–82)||55 (33–108)||.78|
|ALT Day 1||444 (323–1,313)||764 (388–1,510)||.42|
|ALT Day 3||421 (213–741)||463 (297–866)||.43|
|ALT Day 7||217 (82–332)||223 (15–334)||.63|
|TB Day 1||2.3 (1.6–3.9)||2.7 (1.7–4.1)||.82|
|TB Day 3||1.8 (1.2–3.1)||1.7 (1.2–2.9)||.84|
|TB Day 7||1.8 (1.3–3.1)||1.8 (1.2–3.1)||.95|
|PT INR Day 1||1.7 ± .4||2.0 ± .8||.19|
|PT INR Day 3||1.3 ± .2||1.4 ± .3||.46|
|PT INR Day 7||1.2 ± .2||1.3 ± .2||.77|
|Apoptosis||.2 ± .2||.3 ± .4||.06|
|Hepatocyte swelling||1.1 ± .9||1.6 ± 1.1||.07|
|Hemorrhage||.06 ± .22||.08 ± .2||.76|
To address the possibility that the anticipated benefit of IPC in decreasing RP injury may not be uniform for all donor subgroups, further analysis of the dataset was performed with reference to donor age (≤50 years or >50 years), presence or absence of steatosis, and gender. The results of these analyses are shown in Table 4. Donor age, steatosis, and gender did not appear to impact the efficacy of IPC in this study.
|Donor age > 50||835 (513, 1,019)||2.3 ± .5||1,027 (553, 1,686)||2.4 ± .8|
|N = 13||N = 16|
|Donor age ≤ 50||600 (409, 1,632)||2.1 ± .4||922 (532, 1,542)||2.2 ± .9|
|N = 15||N = 18|
|Steatosis†||1,030 (573, 2,228)||2.4 ± .5||1,686 (1,356, 2,842)||2.6 ± 1.2|
|N = 10||N = 9|
|No steatosis||657 (414, 1,019)||2.0 ± .5||660 (429, 1,481)||2.2 ± .5|
|N = 17||N = 23|
|Female donors||573 (409, 787)||2.2 ± .6||802 (429, 1,267)||2.4 ± .7|
|N = 12||N = 16|
|Male donors||1,019 (430, 1,898)||2.1 ± .4||1,440 (660, 2,260)||2.2 ± .9|
|N = 17||N = 17|
Considerable research in experimental animals has shown that a brief period of hepatic inflow occlusion effectively induces hepatic IPC and is protective against RP injury after both cold and warm ischemia.11, 15–17 Except for a very recent abstract of a presentation, there is no data in the literature regarding the clinical application of this protective phenomenon in human deceased donor liver transplantation.20 The salient findings of our study are: 1) deceased organ donors, despite being hemodynamically labile and frequently requiring vasopressors, tolerated 5 minutes of hepatic hilar clamping well, when this was performed before the performance of extensive intraperitoneal dissection; and 2) in deceased donor liver transplantation, 5 minutes of hilar clamping during liver recovery, as performed in this study, did not induce IPC.
In patients undergoing hepatic resections Clavien et al.15 showed that 10 minutes of hilar clamping followed by 10 minutes of RP before hepatic inflow occlusion induced IPC and decreased hepatic parenchymal injury as manifested by elevated postoperative hepatic transaminase levels. The benefit of IPC was more apparent in younger patients and in those with hepatic steatosis and longer periods of inflow occlusion, both of which are important risk factors for increased RP injury.15 In a series of live partial liver donors, intermittent hilar clamping of the donor hemiliver was reported to show a slight (not statistically significant) IPC benefit.29 It is noteworthy that in the preceding studies, patients undergoing hepatic resection were otherwise apparently healthy individuals requiring no vasopressors for hemodynamic support.15, 18, 19, 29 Deceased organ donors, on the other hand, are often hemodynamically labile and require vasopressors to maintain acceptable blood pressure. Furthermore, how the splanchnic congestion inherent in hepatic inflow occlusion impacts overall organ recovery, and specifically intraabdominal organs, is not known. Clamping of the hepatic hilum immediately before donor heparinization resulted in severe congestion of the small intestine and intraperitoneal bleeding. Conversely, hilar clamping was well tolerated when performed prior to extensive intraperitoneal dissection and did not result in edema of pancreata recovered from the same donors.
Why 5 minutes of hepatic hilar clamping in deceased donors did not induce IPC and decrease graft injury after liver transplantation in our study is not clear. Some of the potential reasons are: 1) the preconditioning period of 5 minutes is too brief; 2) the interval between induction of IPC and organ removal is too long; 3) the RP injury in the control group is not severe; and finally, 4) there is the remote possibility that IPC may not be effective in the context of brain death.
When the effectiveness of different periods of hilar clamping that varied between 2 and 30 minutes were examined in mice, 5 and 10 minutes provided the optimal IPC effect against subsequent warm ischemic injury. Not unexpectedly, a further increase in the duration of the ischemia to 20 and 25 minutes led to more severe injury rather than to any benefit from IPC.30 Although 5 minutes of hilar clamping provided some beneficial IPC of the liver in rats, 10 minutes of ischemia induced the optimal IPC.16, 17 Thus, it is very likely that the 5 minutes of hilar clamping utilized in our study is not the optimum period to fully induce IPC. Although these data were taken into consideration during the design phase of this study, concern over donor stability and splanchnic congestion resulted in the selection of 5 minutes as the appropriate initial clamping interval. We have recently begun a new study with 10 minutes of hilar clamping. However, the need for tempering of expectations is demonstrated by the very recent report of Amador et al.20 of a study of 53 deceased liver donors randomized either to standard organ recovery or to 10 minutes of hilar clamping. Preliminary data from that study showed no induction of IPC.
In reports of hepatic IPC in both animals and in man, the 2nd and the principal ischemic event begins 10–15 minutes after completion of IPC. In our study, aortic cross-clamp and organ perfusion began approximately 1.3 hours after hilar occlusion to induce IPC. It is possible this longer delay could explain the lack of IPC induction in our study. However, data showing that protection afforded by IPC in rat livers lasts up to 24 hours make it unlikely that lack of IPC induction in our study is due to the longer interval between hilar clamping and organ perfusion.16
The extent of RP injury in our STD group is considered mild based on both the laboratory and histological data. Therefore, in the context of this milder RP injury in the STD group, benefit from IPC would not be as clearly apparent as when the RP injury in the STD group was more severe. This explanation would be more plausible if the levels of hepatic transaminases in the IPC group in our study were slightly lower than in the STD group. Although longer warm ischemia time was statistically significantly in the IPC group, the cold ischemia time was shorter by 30 minutes on average. It appears unlikely that the 4-minute increase in warm ischemia time would have nullified an otherwise effective IPC response. Vasopressors including phenylephrine are known to induce beneficial effects similar to ischemic preconditioning.31 However, it is very unlikely that vasopressor use in donors had any confounding effects in our study, because their use was similar in both groups of donors. One clinical and 1 experimental study indicate that IPC is beneficial in decreasing warm ischemia RP injury in fatty livers.15, 32 Our further analyses—to examine if 5 minutes of IPC had more beneficial effects in some subgroups of donors—showed no evidence of interaction between IPC, age, steatosis, and gender. However, our study suffers from insufficient numbers in the various subgroups to allow any definite conclusions. In future studies, stratification to balance donor risk factors will help assess the effects of IPC in various donor subgroups.
Finally, experimental data suggest that brain death may abrogate or blunt the induction of IPC. Conflicting evidence exists regarding the influence of brain death on IPC in experimental animals. From 3–5 minutes of myocardial IPC was ineffective in providing protection from infarction in 1 study of brain dead rabbits.33 The possibility that this might be related to the duration of preconditioning ischemia rather than the effects of brain death is suggested by the findings of another study in the same model, in which 2 cycles of 5 minutes of IPC separated by 5 minutes of RP induced a robust IPC response.34 Thus, the ineffectiveness of IPC in our study most likely derives from an insufficient duration of preconditioning ischemia.
In conclusion, 5 minutes of hepatic hilar occlusion during liver recovery in deceased donors, when performed soon after the opening of the abdomen, did not increase hemodynamic instability or result in splanchnic venous congestion. However, 5 minutes of hilar clamping did not induce effective IPC in deceased donor livers. Future clinical trials of IPC with longer or intermittent periods of ischemia and donor risk stratification are needed.
We thank the staff of the New Jersey Organ and Tissue Sharing Network and the transplant physicians of New Jersey for having facilitated this study.
- 1Data from United Network for Organ Sharing. Available at http://www.unos.org. Last accessed May 2004.
- 6Transmission of hepatitis B by transplantation of livers from donors positive for antibody to hepatitis B core antigen. The National Institute of Diabetes and Digestive and Kidney Diseases Liver Transplant Database. Gastroenterology 1997; 113: 1668–1674., , , , , , et al.
- 20Ischemic preconditioning during donor procurement in orthotopic liver transplantation. Preliminary results [Abstract]. 10th Annual Congress of the International Liver Transplantation Society, June 9–12, 2004, Kyoto, Japan., , , , , , et al.
- 28Liver biopsy interpretation. Philadelphia: W.B. Saunders, 1994: 63–67., .