Hepatic steatosis is not always a contraindication for cadaveric liver transplantation

Abstract

Background

Macrovesicular steatosis is assumed to be an important risk factor for early allograft dysfunction (EAD) after orthotopic liver transplantation (OLT).

Aim

To evaluate the impact of steatosis in combination with other risk factors on the outcome of OLT.

Methods

The degree of steatosis was analysed in 165 consecutive OLTs and was classified by histological examination as non (M0), mild (<30%, M1), moderate (30–60%, M2) or severe steatosis (>60%, M3). Recipients were analysed for EAD.

Results

EAD was observed in 28% of patients with M0, 26% with M1, 53% with M2 and 73% with M3 (P < 0.001). Patients with EAD had a significantly shorter graft survival after liver transplantation (P = 0.005) but did not correlate with survival. In multivariate regression analysis, the grade of steatosis, donating after cardiocirculatory death (DCD) grafts and duration of cold ischaemia time were significantly associated with EAD (P < 0.001, P = 0.01 and P = 0.001, respectively).

Conclusion

Livers with severe (M3) steatosis from DCD donors, combined with a prolonged CIT have a high risk for developing EAD which is correlated with shorter graft survival. Therefore M3 livers should only be considered for OLT in selected recipients without the presence of additional risk factors.

Introduction

The increasing waiting list mortality in orthotopic liver transplantation (OLT) as a result of ongoing organ shortage has led to the use of organs from extended criteria donors, including those with hepatic steatosis or non-alcoholic fatty liver disease (NAFLD). Steatosis is considered to be one of the most important variables in multivariate analysis of factors determining graft function after transplantation.¹ The presence of steatosis is correlated with the development of primary non-function or early allograft dysfunction (EAD).^2,3 Marsman et al. propagate discarding livers with more than 30% fatty infiltration owing to the potential for EAD.⁴

Although several centres report encouraging results after OLT with even severely fatty infiltrated livers, studies have not been unanimous in their conclusions about the impact of steatosis on the outcomes after liver transplantation.^5–19 Some centres claim that fatty infiltrated livers should not be considered for liver transplantation,^1,4,13,15 whereas others state that severely steatotic livers can be transplanted safely in selected cases.^{5,6,12,14,18,19}

The difference in point of view is reflected in the outcomes of a survey determining acceptance policy. Only 6% would accept an over 60% fatty infiltrated liver in the United States whereas 42% would do so in the United Kingdom.²⁰

There are multiple reports suggesting that livers with microvesicular steatosis have a better outcome after OLT compared with livers with macrovesicular steatosis.^10,21 The aim of the present study was to evaluate the impact of donor hepatic steatosis on graft function and survival after liver transplantation.

Materials and methods

Data were analysed on consecutive patients who underwent a cadaveric liver transplantation between January 2000 and December 2004. Information was extracted from a prospectively designed database which was updated with post-operative data until August 2010.

Histological examination

The presence, type and severity of steatosis were determined in all grafts. Steatosis was evaluated by biopsies at three standardised time points at the end of cold ischaemia, immediately after reperfusion and after 1 h of reperfusion. The grade of liver macrovesicular steatosis was determined according to international standards:^1,7,18

• M0: no macrovesicular steatosis,

• M1: mild focal macrovesicular steatosis (<30% hepatocytes involved),

• M2: moderate, zonal macrovesicular steatosis (30% < 60% hepatocytes involved) and

• M3: severe, panlobular macrovesicular steatosis (>60% hepatocytes involved).

All microscopic hematoxylin and eosin (H&E)-stained sections were evaluated by two pathologists in a double-blinded fashion. As the presence of microvesicular steatosis has not been reported to adversely affect graft function, this was not captured separately.¹⁰

Clinical outcomes

To assess the degree of liver damage after liver transplantation, the latest validated definition of early allograft dysfunction (EAD)²² was used: the incidence of EAD was determined when one or more of the following criteria were met: (i) bilirubin >10 mg/dl on post-operative day 7, (ii) international normalised ratio (INR) ≥ 1.6 on post-operative day 7 and (iii) aspartate aminotransferase (AST) or alanine transaminase (ALT) > 2000 IU/ml within the first post-operative week.

The influence of the pre-operative situation on outcome after transplantation was measured using the model of end-stage liver disease (MELD) score.²³ Cold ischemic time (CIT) was measured to determine its influence on outcome after transplantation.

The total hospital length of stay (LOS) was calculated as time admitted to the intensive care unit (ICU) and time admitted to the clinical ward (CW) after transplantation.

Statistical analysis

Data were analysed with the Student's t-test, chi-squared test, Fisher's exact test, ANOVA table's and Spearman's correlation tests. Survival curves were analysed using the Kaplan–Meier method and compared with the log-rank test. Logistic univariate analysis and Cox regression were performed with all relevant baseline factors: MELD score, level of steatosis, length of CIT and donation after cardiocirculatory death (DCD). Influence on EAD, graft survival (GS) and overall survival (OS) was expressed in odds ratios and hazard ratios, respectively. Multivariate analysis was executed with variables reaching a level of 10% significance in univariate analysis. A stepwise back algorithm to exclude factors without a predictive value was used at a level of 5% significance.

Results

Patient population

The study group consisted of 185 consecutive cadaveric liver transplantation patients, 80 (43.2%) of whom were female. The median age of the patients was 48 years (mean 46, range 16–69). The donor population had a median age of 47 years (mean 44, range 11–68).

Eight patients were not analysed as a result of intra-operative death. There were four transplants with a primary non-function, two with a M0 liver and two with a M3 liver. One of the M0 livers was retrieved from a DCD donor. One patient died before a re-transplantation could be performed, all others underwent a re-transplantation. Critical pre-operative data or donor liver data were missing in eight cases. As a result, data for 165 patients were left for analysis of which 17 cases (10%) were re-transplants. The median follow-up was 86 months (mean 75, range 0–127).

Prevalence of steatosis

The donor livers revealed M0 in 92 cases (55.8%), M1 in 39 cases (23.6%), M2 in 19 cases (11.5%) and M3 in 15 cases (9.1%) (Table 2). There was no significant difference in age, MELD score, body mass index (BMI) and pre-operative creatinine between the recipients in those groups based on the level of steatosis (Table 1). The underlying disease of those receiving a M3 liver was primary sclerosing cholangitis in four cases (26.7%), cyptogenic cirrhosis in three cases (20.0%), idiopathic acute liver failure in three cases (20.0%), hepatitis B in four cases (26.7%) and hepatitis C in one case (6.7%).

Table 2.

Results after transplantation

	M0	M1	M2	M3	Total
n	92 (55.8%)	39 (23.6%)	19 (11.5%)	15 (9.1%)	165 (100%)
EAD
EAD	26 (28.3%)	10 (25.6%)	10 (52.6%)	11 (73.3%)	57 (34.5%)
No EAD	66 (71.7%)	29 (74.4%)	9 (47.4%)	4 (26.7%)	108 (65.5%)
P-value		0.759	0.039	0.001	0.001
LOS
ICU (median)	4	5	5	6	5
CW (median)	19	19	19	16	19

ICU, intensive care unit; LOS, hospital length of stay.

Table 1.

Patient characteristics and biochemical profile stratified for severe steatosis

	Normal livers (<60%)	Sever steatotic livers (≥60%)	P
Patient characteristics
Age	46.0 ± 1.05*	48.4 ± 2.57*	0.490
Gender (M)	84 (56.0%)	11 (73.3%)	0.195
BMI	24.9 ± 0.40*	24.0 ± 0.51*	0.498
Diabetes	13 (8.7%)	0	0.611
Malignancy	13 (8.7%)	2 (13.3%)	0.630
Biochemical profile
Bilirubine	175.8 ± 19.37*	183.3 ± 51.81*	0.905
INR	2.26 ± 0.14*	2.13 ± 0.27*	0.781
Creatinine	91.4 ± 5.71*	75.1 ± 10.90*	0.370
MELD	17.9 ± 0.99*	17.5 ± 2.80*	0.889
Underlying disease
Hepatitis B	29 (19.3%)	4 (26.7%)	0.409
Hepatitis C	22 (14.7%)	1 (6.7%)
Alcoholic cirrhosis	12 (8.0%)	0
Cryptogenic cirrhosis	17 (11.3%)	3 (20%)
PSC	16 (10.7%)	4 (26.7%)
PBC	12 (8%)	0
Autoimmune hepatitis	13 (8.7%)	0
Idio. Or iatr. Acute liver failure	20 (13.3%)	3 (20%)
Other	9 (6%)	0

^*Mean ± SEM.

BMI, body mass index; iatr, iatrogenic; Idio, idiopathic; INR, international normalised ratio; MELD, model of end-stage liver disease; PBC, primary billiary cirrhosis; PSC, primary sclerosing cholangitis.

Post-operative AST and ALT values

ALT values during the first post-operative week were significantly higher when M3 livers were transplanted. The mean difference was significant from the day of transplantation (P < 0.001) through the sixth post-operative day (P = 0.022). Only the day of transplantation (P = 0.010) and the fourth (P = 0.007) and the fifth (P = 0.005) post-operative day showed significant different AST values when stratified for steatosis. Towards the end of the first week after transplantation, AST and ALT values return to comparable levels for all types of donor livers (Fig. 1).

Figure 1.

First week post-operative values of aminotransferase (AST) and alanine transaminase (ALT). *P < 0.05

Prevalence of EAD

Fifty-seven patients (34.5%) were diagnosed with EAD (Table 2). The prevalence of EAD after transplantation was significant higher when donor livers with moderate (M2) or severe (M3) steatosis were used (P = 0.039 and P < 0.001, respectively). As expected, recipients of M1 livers performed as well as those of M0 livers (P = 0.759). Additionally, in univariate analysis both CIT and DCD impacted EAD significantly (Table 3). DCD reached significance in spite of its low prevalence (n = 13). The EAD rate within the DCD group was 63% compared with 39% for those receiving a non-DCD liver. The mean CIT in the DCD group was 8.4 h compared with 7.2 h for the non-DCD group (P = 0.13). Two patients received a M3 DCD liver; one did not experience EAD, whereas the other did develop EAD and died 15 days post-operatively as a result of multi-organ failure. All DCD grafts utilised came from controlled, Maastricht classification III and IV, donors. Severe steatosis remains a highly significant prognostic baseline factor for the prevalence of EAD. After multivariate analysis macrovesicular steatosis, DCD and CIT remain significant (Table 4).

Table 3.

Univariate analysis of prognostic factors for EAD, GS and OS

Variable		EAD OR (P)	95% CI	GS HR (P)	95% CI	OS HR (P)	95% CI
Meld score		1.03	1.00–1.06	1.03	0.99–1.06	0.99	0.97–1.02
		(0.081)		(0.120)		(0.651)
Steatosis	M0*	1		1		1
	M1	0.88	0.37–2.05	0.48	0.14–1.64	1.44	0.74–2.78
		(0.759)		(0.240)		(0.284)
	M2	2.82	1.03–7.73	0.33	0.04–2.46	1.19	0.49–2.92
		(0.044)		(0.277)		(0.703)
	M3	6.98	2.04–23.91	0.42	0.06–3.19	1.32	0.50–3.46
		(<0.001)		(0.402)		(0.572)
Cold ischemic time	In hours	1.21	1.07–1.37	1.07	0.91–1.25	1.07	0.97–1.19
		(0.003)		(0.417)		(0.193)
DCD	No*	1		1		1
	Yes	3.36	1.05–10.82	2.44	0.71–8.36	1.19	0.43–3.30
		(0.042)		(0.155)		(0.744)
Liver injury	No EAD*	NA		1		1
	EAD	NA		3.26	1.33–7.99	1.19	0.67–2.13
				(0.010)		(0.557)

^*Index.

DCD, donating after cardiocirculatory death; EAD, early allograft dysfunction; GS, graft survival; HR, hazard ratio; NA, not applicable; OR, odds ratio; OS, overall survival.

Table 4.

Multivariate analysis for liver dysfunction (EAD)

Variable		n	Haz ratio	P	95% CI
Steatosis	M0	92	1
	M1	39	0.99	0.978	0.40–2.43
	M2	19	3.59	0.021	1.21–10.62
	M3	15	8.40	0.001	2.29–30.81
Cold ischemic time	in hours		1.27	0.001	1.10–1.46
DCD	No	152	1
	Yes	13	5.67	0.011	1.50–21.44

CI, confidence interval; DCD, donating after cardiocirculatory death; HR, hazard ratio.

Graft survival

Five-year graft survival was 87.4% (SE ± 2.7%). Twenty out of 165 liver transplantations required re-transplantation (12.1%). Three patients underwent multiple re-transplantations (2–3). There was no significant difference in GS between steatotic livers and non-steatotic livers (P = 0.38), as can be seen in the Kaplan–Meier curve (Fig. 2). Patients experiencing EAD had a 5-year GS of 76.4% whereas patients without EAD had a 5-year GS of 92.8% (P = 0.006, Fig. 2). In univariate analysis no additional significant prognostic baseline factors for GS were identified.

Figure 2.

(a) Graft survival vs. macro vesicular steatosis. (b) Graft survival vs. early allograft dysfunction

As most graft loss was in the first 6 months, a subanalysis was performed of this group. Aside from age no other patient or disease characteristics with a prognostic value for early graft loss were identified. Furthermore, M3 livers were not overrepresented in this early graft loss group (Table 5). Six patients (42.9%) lost their graft early as a result of billiary duct ischaemia, five (35.7%) owing to hepatic artery thrombosis, two (14.3%) related to small-for-size graft function and one related to chronic rejection (7.1%).

Table 5.

Sub analysis early graft loss

	Early graftloss (within 6 months) n = 14	No or late graft loss n = 151	P
Patient characteristics
Age	38.8 ± 2.85*	47.0 ± 1.02*	0.019
Gender (M)	6 (42.9%)	89 (58.9%)	0.244
BMI	24.1 ± 2.29*	24.9 ± 0.36*	0.590
Diabetes	0	13 (8.6%)	0.605
Malignancy	0	15 (9.9%)	0.368
Biochemical profile
Bilirubine	257.2 ± 76.83*	169.4 ± 18.54*	0.187
INR	2.32 ± 0.40*	2.24 ± 0.14*	0.859
Creatinine	72.7 ± 9.54*	91.2 ± 5.65*	0.354
MELD	18.9 ± 3.10*	17.8 ± 0.97*	0.734
Underlying disease
Hepatitis B	2 (14.3%)	31 (20.5%)	0.743
Hepatitis C	1 (7.1%)	22 (14.6%)
Alcoholic cirrhosis	0	12 (7.9%)
Cryptogenic cirrhosis	1 (7.1%)	19 (12.6%)
PSC	3 (21.4%)	17 (11.3%)
PBC	2 (14.3%)	10 (6.6%)
Autoimmune hepatitis	1 (7.1%)	12 (7.9%)
Idio. Or iatr. acute liver failure	3 (21.4%)	20 (13.2%)
Other	1 (7.1%)	8 (5.3%)
Steatosis
M0	9 (64.3%)	83 (55.0%)	1.00
M1	3 (21.4%)	36 (23.8%)
M2	1 (7.1%)	18 (11.9%)
M3	1 (7.1%)	14 (9.3%)
Peri-operative blood loss
Packed cells administered	6.3 ± 0.67	7.3 ± 0.50	0.543

^*Mean ± SEM.

BMI, body mass index; iatr, iatrogenic; Idio, idiopathic; INR, international normalised ratio; MELD, model of end-stage liver disease; PBC, primary billiary cirrhosis; PSC, primary sclerosing cholangitis.

Patient survival

Overall actuarial 1-, 5- and 10-year patient survival rates were 84.8% (SE ± 2.8%) 75.0% (SE ± 3.4%) and 67.2% (SE ± 4.4%), respectively. Patients with M0 livers had 78.1% 5-year OS, whereas patients with M3 livers showed 66.7% 5-year OS. In spite of this seemingly substantial difference in survival rates, significance was not achieved (P = 0.732, Fig. 3), probably related to the low prevalence of M3 donor livers and subsequent type two error. Patients with EAD after transplantation showed no significant difference in survival rates, 70.0% vs. 77.7% (P = 0.552, Fig. 3). In univariate analysis no significant factors influencing OS after transplantation could be identified.

Figure 3.

(a) Survival vs. macro vesicular steatosis. (b) Graft survival vs. early allograft dysfunction

Hospital length of stay

The median hospital admission time after liver transplantation was 24 days; the median ICU and ward stays were 5 and 19 days, respectively (Table 2). There was no significant difference between mean LOS for patients with a steatotic donor liver compared with a M0 liver [P = 0.383 (ICU)/P = 0.258 (CW)].

The median LOS for patients with EAD was 27 days; patients without EAD were admitted for 23 days (Table 1). This mean difference was not significant [P = 0.326 (ICU)/P = 0.178(ward)] even when corrected for in-hospital deaths after transplantation.

Discussion

The widespread shortage of donor organs has led to a more liberal acceptance of donor livers for transplantation with more centres transplanting severely steatotic livers.^{5,6,12,14,18,19} A strong association between macrovesicular steatosis and early allograft dysfunction or even primary non-function has been described previously.^{1,4,5,7–9,11,12,15,18,19}

The present study confirms the correlation between severe steatosis and early allograft dysfunction after transplantation. Although severity of steatosis did not correlate with graft or patient survival, the positive predictive value of EAD for shorter graft survival was significant. EAD was not correlated with survival which is in contrast with earlier reported results by Olthoff et al.²² who found a relative risk of 10.7% for death in the case of EAD.

The accuracy of pre-transplant frozen sections has been the subject of much debate for a number of years. Recently, Allesandro et al.²⁴ reported that pre-transplant frozen sections were reliable when a cut-off point of 60% of steatosis was used whereas it became less accurate when a cut-off point of 30% was used.

In our series we noticed a prevalence of 28% of EAD in the M0 group, which is comparable with the M1 livers. This percentage seems high when compared with other studies with a similar study design. Afonso et al. reported a prevalence of initially poor function of the graft after transplantation of 22%,⁵ whereas this prevalence was as low as 9% in a study reported by Canello et al. The lower prevalence in Canello's study can be related to overlap with M1 livers, in which the prevalence of initially poor function was as high as 44%.⁸ It is difficult to report on these differences as the present study included additional variables in the analysis such as ischaemia time and DCD.

Recipients of M3 livers had a more complicated recovery period in the first months after transplantation reflected in the fact that they developed significantly more EAD which is correlated with significantly more graft loss in the first post-operative months. However, after approximately 6 months, they had the same prognosis as patients without EAD. Therefore, based on this study we cannot recommend that M3 livers should be discarded routinely.

As has frequently been described in the literature, other risk factors should also be taken into account when considering the suitability of a liver for transplantation. An important risk factor associated with a less favourable outcome after transplantation is using livers from DCD donors. The use of such livers is correlated with an increased prevalence of primary non-function, a higher re-transplantation rate and shorter graft and overall patient survival.^25–30 In the present study, DCD was correlated significantly with EAD and remained significant in multivariate analysis.

Other known risk factors for poor outcome after liver transplantation are prolonged cold and warm ischaemic times. Prolonged CIT is strongly associated with poor graft function after transplantation and a high prevalence of graft loss, particularly when the graft was harvested from a DCD donor.²⁸ In addition, warm ischaemia times need to be minimised, preferably to less than 30 min.²⁷ Our results were in concordance with these studies. In our series, severe steatosis, a DCD liver and a prolonged CIT were significant in multivariate analysis and therefore all these variables contribute to the prevalence of EAD. It is evident therefore, that care should be taken when M3 livers are transplanted, especially in the presence of other risk factors such as DCD livers and prolonged CIT.

Although beyond the scope of the present study, recipient characteristics are probably equally important when deciding if an organ is suitable for transplantation. Hepatitis C (HCV) is increasingly recognised as one of the major causes leading to end-stage liver failure and transplantation. Steatosis of the graft is associated with viral recurrence.³¹ Although mildly steatotic livers have no impact on HCV recurrence after transplantation,³² a combination of severe fatty livers with prolonged CIT appears to have a negative effect on graft survival.³³

In conclusion, based on our data, steatotic livers, including severely (>60%) steatotic livers, should not be routinely discarded as they are in most transplant centres. Those livers could be used in selected patients typically underserved by the current MELD system, for instant in patients with a relatively long preserved liver function such as primary sclerosing cholangitis (PSC). Avoiding additional risk factors such as a prolonged CIT and livers from DCD donors, these livers can be transplanted successfully.

Conflicts of interest

None declared.