Ischemic Preconditioning (IP) of the Liver as a Safe and Protective Technique against Ischemia/Reperfusion Injury (IRI)


* Corresponding author: Alessandro Franchello,


The aim of the study was to evaluate safety and efficacy of IP in LT, particularly in marginal grafts. From 2007 to 2008, 75 LT donors were randomized to receive IP (IP+) or not (IP–). Considering the graft quality, we divided the main groups in two subgroups (marg+/marg–). IP was performed by 10-min inflow occlusion (Pringle maneuver utilizing a toruniquet). Donor variables considered were gender, age, AST/ALT, ischemia time and steatosis. Recipient variables were gender, age, indication to LT and MELD/CHILD/UNOS score. AST/ALT levels, INR, bilirubin, lactic acid, bile output on postoperative days 1, 3 and 7 were evaluated. Histological analysis was performed evaluating necrosis/steatosis, hepatocyte swelling, PMN infiltration and councilman bodies. Thirty patients received IP+ liver. No differences were seen between groups considering recipient and donor variables. Liver function and AST/ALT levels showed no significant differences between the main two groups. Marginal IP+ showed lower AST levels on day1 compared with untreated marginal livers (936.35 vs. 1268.23; p = 0.026). IP+ livers showed a significant reduction of moderate-severe hepatocyte swelling (33.3% vs. 65.9%; p = 0.043). IP+ patients had a significant reduction of positive early microbiological investigations (36.7% vs. 57.1%; p = 0.042). In our experience IP was safe also in marginal donors, showing a protective role against IRI.


Ischemia/reperfusion injury (IRI) is one of the main causes of liver dysfunction, such as primary nonfunction (PNF) or delayed nonfunction (DNF), and it is responsible for increased morbidity and risk of death (1,2).

Each graft is subjected to IRI, but the so-called marginal grafts are more susceptible to this particular type of damage. The shortage of donation forced to extend selection criteria, accepting poor quality grafts (1), and consequently early nonfunction became a relevant clinical problem.

Protective strategies against IRI are different, varying from medical to surgical techniques. Ischemic preconditioning (IP) is a surgical method described for the first time by Murry in 1986 on a canin model of ischemic myocardium, consisting of a brief period of ischemia followed by reperfusion, prior to a sustained ischemic insult (3). The preconditioning effect was thereafter demonstrated in other organs, including the liver (4,5); its real efficacy improving liver transplantation (LT) outcome is still under discussion.

The aim of our study was to evaluate the safety and efficacy of IP in the setting of LT with particular regard on grafts considered marginal for age or macrovesicular steatosis.

Material and Methods

We designed a randomized clinical trial to evaluate the clinical and histological effects of IP against IRI in liver transplantation. From September 2007 to May 2008, 75 deceased liver donors were randomized to receive IP (IP+) or not (IP−) at the time of recovering, through a computer-generated number sequence, creating two different study groups.

The main groups of IP+ and IP− were divided in two subgroups considering the quality of the graft (marg+ and marg−); marginal grafts were defined according to our previously described findings (6) as: donor age >65 years and/or macrovacuolar steatosis >15%. Four subgroups were created: marginal livers subjected to preconditioning (IP+ marg+), marginal livers not treated (IP− marg+), nonmarginal grafts subjected to preconditioning (IP+ marg−) and nonmarginal grafts not treated (IP− marg−).

IP was performed during the procurement procedure by 10-min inflow occlusion (Pringle maneuver utilizing a tourniquet) followed by 30 min of reperfusion prior to the start of cold ischemia. Pediatric donors and split LTs were excluded from the study.

LT was performed with preservation of the inferior vena cava (piggy-back technique) in all cases.

A liver biopsy was collected at the end of each transplantation, just before the closure of the abdomen wall, to evaluate graft quality and to perform histological analysis.

Donor-related variables considered were gender, age, use of vasopressors, occurrence of cardiac arrest, ICU stay expressed in days, Na+, AST and ALT levels, total and warm ischemia time and macrovacuolar steatosis.

Recipient-related variables considered were gender, age, indication to LT, liver function score such as MELD, CHILD and UNOS at transplantation time and immunosuppressive drugs used.

During the intraoperative phase, 5 min after graft reperfusion, a blood sample was collected to evaluate any difference on cell count and early liver function in both groups; the need for blood supply was also noticed.

Liver injury and function were analyzed comparing the two main groups (IP+ and IP−) but also the four subgroups: AST and ALT levels, prothrombine time expressed as INR, total bilirubin, lactic acid and bile output on days 1, 3 and 7 after surgery were evaluated.

As secondary endpoints, we compared the incidence of infections during the first postoperative month, and graft survival after 6 months post-LT.

Histological analysis was performed in a blind method by a single pathologist evaluating the fine needle liver biopsy recovered at the end of each transplantation. Liver specimens were formalin-fixed and paraffin-embedded for hematoxylin and eosin staining. We use the NIDDK score for the evaluation of graft rejection.

A semiquantitative score method was utilized to evaluate the quality of the graft (necrosis and particularly macrovacuolar steatosis).

IRI was evaluated analyzing the following features:

  • 1Hepatocyte swelling, caused by passive ions shift across cell membranes as a result of homeostatic disregulation and ATP loss after the ischemic phase (1). It was graduated as low when focally detectable on the specimen, mild when plurifocal lesions were present and severe when it was diffuse on liver parenchyma.
  • 2PMN infiltration, expression of the inflammatory response triggered by cytochines production during the reperfusion phase (7). It was defined as low when a minimal neutrophils accumulation on portal and/or lobular spaces was evident; mild, if some lobules were involved focally, especially on portal spaces, inflammatory infiltration; severe, when a diffuse and above all center-lobular infiltration was present.
  • 3Kupffer cell (KC) activation, one of the earliest inflammatory responses in the reperfused liver and the main source of reactive oxygen species formation during the initial reperfusion period (7). Morphological signs of KC activation were graduated in low, mild and severe grade depending on the extension on liver parenchyma.
  • 4Sinusoidal endothelial cell (SEC) detachment and rounding, a particular form of injury inherent to transplantation, probably due to actine disassembly and metalloproteinases activation (8). We noticed its presence or absence.
  • 5Councilman bodies (or red bodies), representing hepatocyte degeneration after severe injury as a morphological sign of on-coming apoptosis; the shrinkage of hepatocyte with deep red color and chromatin condensation are typical features of these lesion. If few cellular elements in some lobules were involved, the degree of hepatocyte degeneration was defined low, mild if most of them were involved and severe if red bodies were detectable diffusely on liver parenchyma.

To further analyze the grade of apoptosis, we decided to use the ‘Terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling’ assay (TUNEL) showing pycnotic nuclei in situ.

Liver sections were deparaffinized and rehydrated through three changes between xylene and graded alcohol, then washed in PBS for 5 min and treated with microwave 350 W for 5 min in a citrate buffer. The Cell Death Detection Kit-POD (Roche, Indianapolis, IN) was then used according to the manufacturer's instructions.

Apoptotic cell total number was counted with high power field (400×, HPF) for each liver sample, and an ‘apoptotic index’ was then calculated as positive cell total number upon total HPFs for each liver specimens, expressed in percentage.

This clinical trial was approved by the local ethical committee.

Statistical analysis was performed using the SPSS program (SPSS Inc. Chicago IL, version 12.02). Continues variables were described as mean ± SD, and the difference between groups was verified with a nonparametric test (Mann–Whitney, Kruskal–Wallis or Wilcoxon). Categorical variables were described as incidence, and the difference between groups was estimated with the chi-square test (Pearson), or when requested with Fisher's exact test, when feasible Odd Ratio was calculated as risk estimation. Survival was analyzed with Kaplan–Meier curves and the log-rank test; a survival analysis with Cox regression model was performed. The significance level (α) was fixed as p-value < 0.05.


Seventy-two patients were enrolled in the study, and 75 transplantations were analyzed. Thirty patients received a preconditioned liver, and 45 LT were assigned to the control group.

In two cases, it was not possible to perform the IP because of a severe hypotension developed in donor after 5 min of Pringle maneuver.

Donor-related criteria were confrontable in both groups, and no significant differences were seen (Table 1).

Table 1.  Donor-related variables
VariablesIP+IP−p-ValueOR (95% CI)
 Mean ± SD58.73 ± 14.8260.89 ± 14.99n.s. 
 >62 years14 (46.7%)25 (55.6%)n.s.0.7 (0.277–1.770)
 Male : female18 : 1221 : 24n.s. 
Cause of death
 Cerebrovascular20 (66.7%)31 (68.9%)n.s. 
 Trauma 5 (16.7%)2 (4.4%)  
 Others 5 (16.7%)12 (26.7%)  
ICU stay (days)
 Mean ± SD5.4 ± 5.54.89 ± 4.19n.s. 
 Mean ± SD82.6 ± 159.5454.96 ± 103.77n.s. 
 Mean ± SD80.8 ± 237.9966.09 ± 166.65n.s. 
HCV+0 (0%)2 (4.4%)n.s. 
 Mean ± SD150.11 ± 7.98151.31 ± 9.01n.s. 
 Dopamine >10 μg/kg/min19 (90.5%)15 (83.3%)n.s  1.9 (0.281–12.869)
 Others16 (53.3%)21 (46.7%)n.s. 
Cardiac arrest 5 (16.7%)12 (26.7%)n.s.0.550 (0.171–1.764)
Total ischemia time (min)
 Mean ± SD518.77 ± 108.91519.62 ± 78.67n.s. 
Warm ischemia time (min)
 Mean ± SD25.17 ± 6.3122.73 ± 3.93n.s. 
Macrovacuolar steatosis
 ≥15% 4 (13.3%)9 (20%)n.s.0.615 (0.171–2.216)
 <15%26 (86.7%)36 (80%)n.s 
Preservation solution infection16 (53.3%)18 (40%)n.s.1.714 (0.674–4.357)
Positive hemoculture 5 (16.7%) 7 (15.6%)n.s.1.086 (0.310–3.804)
Biliar infection0 (0%)3 (6.7%)n.s.0.575 (0.472–0.701)
Peritoneal infection0 (0%)0 (0%)  
BAL positive culture1 (3.3%)2 (4.4%)n.s.0.741 (0.064–8.559)
BA positive culture 7 (23.3%)4 (8.9%)n.s. 3.12 (0.825–11.799)
Urinary tract infection2 (6.7%) 5 (11.1%)n.s.0.571 (0.103–3.158)

The mean recipient age in the preconditioned patients was 55.18 (±6.42) and 51.47 (±11.39) years in the control group, and no significant differences were seen in both groups considering recipient-related variables (Table 2). The most frequent LT indication was viral cirrhosis (72%), and above all the HCV-related cirrhosis (45.3%); eight patients were transplanted for alcoholic cirrhosis, and in 22.7% of the case a hepatocellular carcinoma (HCC) was also present. Three patients received a combined liver–kidney transplantation, in one case for hepatorenal polycystosis, in two cases for an associated kidney failure in a heart transplanted patient and in a patient affected by Berger syndrome.

Table 2.  Recipient-related variables
VariablesIP+IP−p-ValueOR (95% CI)
 Mean ± SD55.19 ± 6.4251.47 ± 11.39n.s. 
 Male : female23 : 734 : 10n.s.0.966 (0.321–2.907)
 Viral cirrhosis C17 (56.7%)17 (37.8%)n.s 
 Viral cirrhosis B 4 (13.3%)14 (31.1%)  
 Viral cirrhosis B+δ1 (3.2%)1 (2.2%)  
 Alcoholic cirrhosis 4 (13.3%)4 (8.9%)  
 Primitive biliar cirrhosis1 (3.3%)1 (2.2%)  
 Schlerosant Cholangitis0 (0%)1 (2.2%)  
 Unknown cirrhosis1 (3.3%)1 (2.2%)  
 Biliary athresia0 (0%)1 (2.2%)  
 Autoimmune cirrhosis0 (0%)1 (2.2%)  
 Primary nonfunction0 (0%)1 (2.2%)  
 Delayed nonfunction0 (0%)2 (4.4%)  
 Hemangioendothelioma0 (0%)1 (2.2%)  
 Hepatorenal polycystosis1 (3.3%)0 (0%)  
 Glicogenosis IVa1 (3.3%)0 (0%)  
 HCC 8 (26.7%)9 (19.6%)  
 A 5 (16.7%)9 (20.9%)n.s. 
 B15 (50%)22 (51.2%)  
 C10 (33.3%)12 (27.9%)  
 Mean ± SD14.55 ± 4.6116.88 ± 4.95n.s. 
 Creatinine 1.22 ± 1.25 0.96 ± 0.58  
 INR  1.5 ± 0.41 1.63 ± 0.46  
 Total bilirubin 4.96 ± 7.57 6.06 ± 8.43  
 1-2A1 (3.4%)4 (9.8%)n.s. 
 2B-328 (96.6%)37 (90.2%)  
Cyclosporine16 (53.3%)16 (35.6%)  
Tacrolimus14 (46.7%)27 (60%)  

During the study period, three patients were retransplanted. In one case, the graft was lost for a severe acute rejection resistant to steroid therapy, developed in a patient who received a preconditioned liver and was affected by primitive biliar cirrhosis; the other two patients, both assigned to the control group, developed a DNF. Each of them received grafts not treated with IP, at the time of the re-LT.

Only one case of PNF was observed in the control group, after intrahepatic disseminated venous thrombosis, with intraoperative death of patient.

Analyzing the intraoperative phase, no differences were detectable on cell counts 5 min after reperfusion. The mean operation time was also similar in both groups.

Median AST, ALT and LDH levels were lower in the IP+ group just 5 min after restoring the blood flow, but the difference was not statistically significant (Table 3).

Table 3.  Intraoperative variables
Operation time (h)
 Mean ± SD 6.28 ± 1.566.35 ± 1.45n.s.
Total blood infusion (mL)
 Mean ± SD2066.67 ± 1713.22138.64 ± 1704.23n.s.
WBC (*109/L)
 Mean ± SD2.76 ± 1.62.84 ± 2.22n.s.
RBC (*1012/L)
 Mean ± SD 3.04 ± 0.462.95 ± 0.46n.s.
Hb (g/dL)
 Mean ± SD15.36 ± 8.3615.47 ± 8.56 n.s.
Htc (%)
 Mean ± SD20.54 ± 9.6419.56 ± 9.07 n.s.
PLTS (*109/L)
 Mean ± SD 44.57 ± 23.8550.02 ± 51.25n.s.
 Mean ± SD 647 ± 482.39928.13 ± 963.07n.s.
 Mean ± SD 496.07 ± 419.34750.92 ± 820.66n.s.
 Mean ± SD 2202.29 ± 1889.262956.03 ± 3144.78n.s.

The evaluation of AST and ALT during the early postoperative phase showed lower levels in patients who received a preconditioned liver on day 1 concerning AST values, in days 1 and 3 concerning ALT values (p = n.s.). They were similar on day 7 (Table 4).

Table 4.  Postoperative related variables
AST mean ± SD (UI/L)
 1815.43 ± 636.8 1046.8 ± 946.1 n.s.
 3270.17 ± 193.74281.46 ± 292.94n.s.
 748.52 ± 30.3544.84 ± 23.73n.s.
ALT mean ± SD (UI/L)
 1 631.8 ± 437.15743.66 ± 581.55n.s.
 3568.97 ± 372.84674.09 ± 705.02n.s.
 7155.93 ± 75.92 156.71 ± 109.83n.s.
INR mean ± SD
 11.78 ± 0.391.78 ± 0.36n.s.
 31.4 ± 0.31.47 ± 0.37n.s.
 71.22 ± 0.17 1.3 ± 0.21n.s.
Tot bilirubin mean ± SD (mg/dL)
 16.79 ± 5.2 7.24 ± 5.1 n.s.
 37.08 ± 4.887.86 ± 6.22n.s.
 79.15 ± 6.69 9.1 ± 8.33n.s.
Lactic acid mean ± SD (mmol/L)
 11.46 ± 0.621.37 ± 0.57n.s.
 30.92 ± 0.4 1.3 ± 1.6n.s.
 71.03 ± 0.371.08 ± 0.43n.s.
Biliar output mean ± SD (mL)
 1143.19 ± 150.5 122.02 ± 106.39n.s.
 3 166.8 ± 167.23141.18 ± 108.29n.s.
 7189.95 ± 151.21150.36 ± 127.02n.s.
Rejection (NIDDK score)
 Slight acute3 (30%)2 (25%) n.s.
 Moderate acute5 (50%)5 (62.5%) 
 Severe acute1 (10%)0 (0%)  
 Clinic rejection1 (10%)1 (12.5%) 
Onset (days)
 Mean ± SD15.22 ± 20.4215.38 ± 16.35n.s.
Hospital stay (days)
 Mean ± SD13.4 ± 6.6 18.12 ± 12.92n.s.

Considering the donor criteria expressed before, 64% grafts were considered as marginal liver: 14 donors were older than 65 years and four livers had macrovescicular steatosis exceeding 15%. The four subgroups were composed by

  • 1IP− marg−: 14 patients (18.7%)
  • 2IP− marg+: 31 patients (41.3%)
  • 3IP+ marg−: 13 patients (17.3%)
  • 4IP+ marg+: 17 patients (22.7%)

Liver function showed no differences between the main two groups: median prothrombine time (INR), total bilirubin or lactic acid. Biliar output was higher in grafts subjected to preconditioning in each postoperative day considered, but no significant values were reached on statistical tests (Table 4).

Marginal liver treated with IP showed significant lower AST levels on day 1 when compared with untreated marginal liver (p = 0.026) (Table 5, Figure 1).

Table 5.  Comparison of the postoperative variables between the four subgroups
VariablesIP + marg−IP+ marg+IP− marg−IP− marg+p-Value
AST mean ± SD (UI/L)
 1657.31 ± 524.71936.35 ± 701.89572.29 ± 521.021268.23 ± 1022.790.026
 3179.39 ± 130.93339.59 ± 208.23267.29 ± 372.30288.07 ± 254.880.039
 752.54 ± 38.1842.25 ± 22.9738.57 ± 15.9747.77 ± 26.33n.s.
ALT mean ± SD (UI/L)
 1599.54 ± 474.61656.47 ± 419.44484.86 ± 302.5 864.43 ± 612.49n.s.
 3507.46 ± 422.31  616 ± 335.81632.86 ± 975.80693.33 ± 555.84n.s.
 7142.23 ± 69.74 167.06 ± 81.08 138.21 ± 129.73165.33 ± 100.51n.s.
INR mean ± SD
 1 1.6 ± 0.311.91 ± 0.4 1.75 ± 0.461.79 ± 0.31n.s.
 31.28 ± 0.171.49 ± 0.341.48 ± 0.491.47 ± 0.31n.s.
 71.19 ± 0.151.26 ± 0.181.26 ± 0.241.32 ± 0.2 n.s.
Tot bilirubin mean ± SD (mg/dL)
 16.24 ± 4.1 7.21 ± 8.1 6.11 ± 3.127.76 ± 5.77n.s.
 36.87 ± 6.657.24 ± 3.166.39 ± 4.198.55 ± 6.93n.s.
 78.19 ± 8.369.93 ± 5.11 6.2 ± 5.4210.44 ± 9.16 n.s.
Lactic acid mean ± SD (mmol/L)
 11.23 ± 0.491.64 ± 0.661.26 ± 0.421.42 ± 0.63n.s.
 30.94 ± 0.4 0.91 ± 0.421.73 ± 2.74 1.1 ± 0.49n.s.
 7 1.1 ± 0.470.98 ± 0.3 1.22 ± 0.611.02 ± 0.31n.s.
Biliar output mean ± SD (mL)
 1188.18 ± 207.63112.25 ± 89.51 163.62 ± 164.94 104 ± 63.44n.s.
 3  230 ± 235.63117.14 ± 53.74 171.83 ± 154.54128.04 ± 81.31 n.s.
 7221.25 ± 130.13169.08 ± 165.92168.58 ± 138.28139.95 ± 122.42n.s.
Figure 1.

Postoperative AST level compared between the four subgroups.

Good quality grafts showed a lower AST level than marginal livers, with minimal differences when preconditioned or not. The same trend was seen considering ALT values. No differences were seen on day 7 in any of the subgroups considered. IP therefore showed a biochemical significant reduction of IRI severity, in particular on marginal grafts; its effect was manifested essentially during the first three postoperative days.

Even when analyzing the four subgroups, no differences were detectable concerning postoperative liver function indicators as prothrombine time or lactic acid.

Marginal grafts showed lower biliar bilirubin secretion than good quality livers, with increased sieric values (Table 5); furthermore, preconditioned marginal grafts showed lower bilirubin blood level than same quality but untreated grafts (IP+ marg+: 7.21 ± 8.1 mg/dL vs. IP− marg+: 7.76 ± 5.77 mg/dL on day 1; IP+ marg+: 9.93 ± 5.11 mg/dL vs. IP− marg+: 10.44 ± 9.16 mg/dL on day 7; p = n.s.) (Table 5). Moreover, analyzing daily biliar output, we found that marginal grafts produced lower bile quantity in each postoperative day considered (Table 5, Figure 2).

Figure 2.

Postoperative biliar output compared between the four subgroups.

The results about the histological blinded evaluation are summarized in Table 6; preconditioned livers showed a significant increase of low grade (53.3% vs. 29.5%) and a significant reduction of moderate-to-severe grade hepatocyte swelling (33.3% vs. 65.9%) in comparison with untreated grafts (p = 0.043).

Table 6.  Histological analysis
Hepatocyte swelling  0.043
 Slight16 (53.3%)13 (29.5%) 
 Moderate 7 (23.3%)23 (52.3%) 
 Severe3 (10%)  6 (13.6%) 
 No 4 (13.4%)2 (4.6%) 
Neutrophils infiltration  n.s
 Slight20 (66.7%)21 (47.7%) 
 Moderate6 (20%) 18 (40.9%) 
 Severe3 (10%) 4 (9.1%) 
Endothelial cell detachment  n.s.
 Yes2 (6.7%) 7 (15.9%) 
 No28 (93.3%)37 (84.1%) 
Kupffer's cell activation  n.s.
 Slight22 (88%)  25 (78.1%) 
 Moderate3 (12%)  7 (21.9%) 
Councilman bodies  n.s.
 Slight12 (40%)  19 (43.2%) 
 Moderate3 (10%)  7 (15.9%) 
 Severe1 (3.3%)2 (4.5%) 
Hepatocyte apoptotic index (%)  n.s.
 Mean ± SD0.76%± 1.950.59%± 0.78 
 SEC apoptotic index (%)  n.s.
 Mean ± SD0.58%± 1.420.47%± 0.78 

We observed the same trend on all the other lesions considered but the values were not statistically significant. The most frequent neutrophilic infiltration grade was low in both groups, but in the IP+ group it was possible to see a more frequent presentation of that grade; on the contrary, in the IP− group there was a higher mild grade rate. KC activation of mild grade was also higher in control groups and reduced in preconditioned liver. Sinusoidal cell detachment was detected more frequently in untreated grafts, even if it was a less represented lesion in both groups. Councilman bodies' most frequent grade was low both in cases and controls, but percentage values were again lower in preconditioned graft.

Comparing the apoptotic index, evaluated with TUNEL technology, we observed a lower rate of hepatocyte apoptotic cell death in liver subjected to IP (0.22% vs. 0.32%), but no statistical significance was reached. Analyzing the apoptotic index screened for liver quality, we found a significant increase of apoptosis in marginal liver (0.85% vs. 0.30%, p = 0.029) (Table 7).

Table 7.  Hepatocyte apoptotic index compared for graft quality
Hepatocyte apoptotic index
 Mean ± SD0.84%± 1.650.30%± 0.420.029
  IP+1.09%± 2.470.25%± 0.26n.s.
  IP−0.70%± 0.860.35%± 0.53n.s.

In IP+ recipients, we observed a significant reduction of positive microbiological investigations during the first month post-LT (p = 0.042; O.R.: 3.1; 95% CI: 1.1–3.1), and the incidence of clinical infections was lower (Table 8).

Table 8.  Microbiological positivity and clinical infection within 1 month post-LT
VariablesIP+IP−p-ValueOR (95% CI)
Microbiological positivity  0.033.1 (1.1–3.1)
 Yes9 (36.7%)24 (57.1%)  
 No21 (63.3%)18 (42.9%)  
Clinical infection  0.60.7 (0.2–2.1)
  Surgical site infection20  

The overall graft survival was 94% at 6 months after surgery. Screening the study population for quality graft, marginal grafts reached lower survival rate when considered altogether (92% vs. 100%, p = n.s.) or when considering only donor age or macrovacuolar steatosis grade; in particular, the last comparison showed a significant reduction of survival in marginal grafts for steatosis, already 6 months after transplantation (75% vs. 98%, p = 0.0016) (Figure 3).

Figure 3.

Graft survival compared for macrovacuolar steatosis.

Comparing the survival rate in the two main groups, we observed higher percent values in the preconditioned group, even if no statistical significance was reached (96% vs. 93%, p = n.s.).

Good quality livers reached 100% survival at 6 months, with no differences when preconditioned or not. Analyzing marginal grafts instead, it was possible to observe a higher survival rate when the liver was subjected to IP than when untreated (94% vs. 90%; p = n.s.) (Figure 4).

Figure 4.

Graft survival compared between the four subgroups.

The Cox regression model was performed considering variables well known to be related to worse survival, together with the presence of IP (Table 9). Only macrovacuolar steatosis was able to significantly and independently impact graft prognosis, with increased risk of graft loss (p = 0.022; RR = 14.05; C.I. 95%= 1.452–135.878).

Table 9.  Cox regression model
 Equation variables
BSEWaldDfSig.R.R.95% CI for R.R.
Macro steatosis ≥15% 2.6421.1585.20710.02214.0451.452135.878
 HCV+−0.6741.1570.33910.560 0.5100.053  4.922
 IP+−0.4031.1570.12110.728 0.6680.069  6.455


IRI is one of the major causes of liver nonfunction after LT (1,2). The shortage of donors forced to extend selection criteria, accepting poor graft quality associated with a worse outcome; in our study, 64% of donors were considered marginal for age or for macrovescicular steatosis (6), and we evaluated the effect of IP with particular regard on these marginal grafts. In our knowledge, this is the first prospective study that evaluated extensively the use of IP in suboptimal livers and showed a potential protective role in this kind of grafts. Histological analysis also showed for the first time that in marginal livers treated with IP there were significant differences about the pathological lesions typically secondary to IRI (in particular hepatocyte swelling and apoptosis).

Many techniques were developed in the past years to prevent or reduce the IRI, using medical or surgical procedure (9). Most of them are still under evaluation for efficacy and feasibility in the setting of LT. In our experience, IP was an easy technique, feasible to perform during liver recovery, without any added risk for patients undergoing LT. In fact no complications directly related to IP were detected after OLT, neither in optimal nor in marginal grafts. Considering the efficacy, two different windows of cytoprotection were described; the first one is called ‘classic preconditioning’ that corresponds to the effect described by Murry and it lasts at most 4 h (10). The second one is called ‘delayed preconditioning’, which was described by Kuzuya and Marber and is less powerful than the first, but its effect is more lasting (72 h at most) even if not all organs present this kind of cellular protection (11).

IP efficacy is under discussion because different clinical trials reported contrasting results; IP first application on liver resection showed the potential benefits of its use as shown by Clavien (4,12), and IP use in hepatic surgery is quite established, but the role of preconditioning on deceased liver donor in the setting of LT is not completely clear. Jassem in his trial found a significant reduction of transaminase level between the two groups in the postoperative period (13); a significant reduction on lactic acid was also found in preconditioned livers. Azoulay found the same trend in his series but without differences on liver function indicators (14). Koneru showed instead, in his second trial, a paradox effect of preconditioning with a higher transaminase peak in preconditioned liver, counterbalanced by lower dysfunction and rejection rate (15). From our experience, IP showed a protective role against IRI, consisting essentially on its severity reduction as shown by lower mean AST and ALT levels 5 min after reperfusion and in the first three postoperative days. The reduction of enzyme levels did not correlate to a faster recovery of liver function after transplantation, and no differences were seen in the main two groups.

None of the trials considered before selectively analyzed the quality of the liver subjected to preconditioning. Amador in his evaluation supposed that IP could be more useful on marginal grafts showing lower AST, ALT, total bilirubin levels and above all lower liver dysfunctions on preconditioned marginal livers (16). To our knowledge, this is the first study considering extensively marginal grafts, comparing four subgroups. Our findings seem to confirm Amador's hypothesis. In particular, on the first and third days after surgery, marginal donor subjected to IP had a significant AST and ALT level reduction when compared with same quality but untreated graft. But even when graft quality was screened between the groups, no significant variation was detected in liver function indicators, such as prothrombine time (INR) or lactic acid.

We considered interesting the finding that IP+ livers produced more bile in the first postoperative week; this could reflect an improvement on bile output, also ameliorating late biliary complications. A further analysis about this kind of complications could be interesting.

However, in our experience all marginal livers had a slower recovery of bile synthesis function, with a higher total bilirubin blood level and lower daily biliary output; IRI was probably more important on marginal grafts.

A recent methanalysis involving different clinical trials showed no statistical differences between the groups in mortality, graft dysfunction or retransplantation, nor differences in the transaminase activity, bilirubin level and prothrombine time (17). The authors concluded that there is currently no evidence to support or refuse the use of IP in donor liver recovery and suggested the necessity of other studies to further analyze preconditioning efficacy. Clavien instead, in his recent paper, expressed his personal viewpoint suggesting that there is no evidence yet to deny IP protective effects in LT in spite of the limited number of trials reporting ambiguous results (18).

Few clinical trials considered and analyzed the histological findings typical of IRI. Jassem focused his evaluation on the neutrophils deposition and platelets activation/aggregation in liver parenchyma, showing a lower inflammatory response in liver treated with preconditioning (13). Cescon analyzed the grade of apoptosis and neutrophils infiltration, showing no differences between the two groups (19). We decided instead to perform a histological evaluation with a blind method, considering each lesion suggestive of IRI compared in both groups.

Our analysis confirmed the hepatoprotective action of IP; a significant reduction of hepatocytes suffering was seen in IP+ graft when analyzing cellular swelling after homeostatic disregulation. Preconditioning significantly increased low-grade and reduced mild-grade swelling.

The other histological features analyzed showed a similar trend, even without statistical significance; in each of them the incidence of low or mild grade lesions was doubled in the IP− group. IRI severity seemed reduced by the IP also when a histological analysis was performed.

There is evidence in literature showing how the final damage of IRI leads to hepatocyte and SEC apoptotic death (7), but actually the opinion, in which marginal grafts undergo on necrosis after IRI, is well accepted (20). Apoptotic death infrequently exceeds 1–2% and is always lower than 5%; in our series, using TUNEL technology, apoptotic index was about 0.6%. In the control group, the apoptotic index was higher than those seen in IP+ graft, showing again a protective role of preconditioning. Comparing apoptotic index to the graft quality, we observed a significant increase of apoptosis in marginal grafts. According to our knowledge, this is the first study showing that marginal grafts are subjected to a significant increase of apoptosis after preservation; poor quality grafts are therefore more susceptible to IRI.

Furthermore, no significant differences were seen on marginal graft apoptosis, when the treatment given was analyzed; the poor quality was more important than the protective effect shown by IP.

Really interesting was the observation that IP was associated with a reduction of early infective episodes in the first month post-LT. Patients receiving a LT are exposed to a great risk of infections, because they are highly immunosuppressed especially in the first periods after surgery; the mechanism of this kind of protection should be evaluated in other studies, but these data are important because they could have a significant impact on patient and graft survival. It is possible to hypothesize that KC in liver treated with IP were less damaged by IRI, receiving a higher protection, so that the whole reticular-endothelial system was more ready to prevent bacterial translocation and infections (21).

The overall 6-month patients’ survival in our study was 99%, and the overall 6-month graft survival was 94%; a comparison between the main groups did not show significant values on the log-rank test, but graft survival was higher in preconditioned patients. Considering graft quality, we observed also an improvement of survival in marginal graft subjected to preconditioning in comparison to those untreated, even if again no statistical significant value was reached.

Considering the incidence of PNF and initial nonfunction (INF) after LT, Azoulay showed that the preconditioning was the only variable significantly associated with graft nonfunction (14); in our experience indeed, two DNF and one PNF were observed only in IP− and marginal grafts. Furthermore, the only graft loss observed in the IP+ marg+ group was related to a severe acute rejection resistant to steroid therapy in a patient affected by primitive biliary cirrhosis. We think that this graft loss was not directly caused by IRI, but was worsened by the immunomediated damage and that IP did not take part in its failure.

Moreover, in a multivariate analysis, IP was not associated with an increase risk of graft loss.

Patient and graft survival analysis did not reach significant differences probably because of the number of patients enrolled, but there were statistically significant results concerning histological, biochemical and clinical variables that led to suggest that preconditioned marginal grafts had better function and lower infectious risk after LT. Survival analysis would have probably reached significant differences between the groups if more patients were enrolled in this prospective trial; we have the intent to continue our study to improve the knowledge about IP. However, to our knowledge, this is the first prospective study evaluating the efficacy of IP in a large number of marginal donors.

In conclusion, from our experience, IP was a safe technique also in marginal donors and showed a protective role against IRI; these findings led us to suggest that IP is an easy device that could be performed in LT independently from the quality of the graft. However, these results should be furthermore investigated, with particular regard on marginal donors, in a larger multicenter trial to better understand the underlying protective mechanism of IP.