Non-Alcoholic Fatty Liver Disease, Non-Alcoholic Steatohepatitis and Orthotopic Liver Transplantation


*Corresponding author: Michael R. Lucey,


Obesity, non-alcoholic fatty liver disease (NAFLD) and non-alcoholic steatohepatitis (NASH) are becoming increasingly common medical problems in the developed world, often in the setting of the metabolic or insulin resistance syndrome (IRS). It is predicted that by the year 2025 > 25 million Americans may have NASH-related liver disease. NASH and NAFLD also affect the donor population. The use of steatotic donor livers for liver transplantation (LT) is associated with an increased risk of primary nonfunction (PNF) in the allograft. There is particular reluctance to use steatotic livers for living donor LT. There is indirect evidence to suggest that patients undergoing LT for cirrhosis resulting from NASH may have poorer outcome, despite careful selection of LT candidates. Indeed it is likely that many potential LT candidates with NASH are excluded from LT due to co-morbid conditions related to IRS. The post-LT patient is at risk of several components of IRS, such as diabetes mellitus, hypertension, hyperlipidaemia and obesity and there is increasing recognition of de novo and recurrent NAFLD and NASH after LT. Thus NAFLD and NASH affect all aspects of LT including donors, patients in evaluation and the LT recipient.


Non-alcoholic fatty liver disease (NAFLD) and non-alcoholic steatohepatitis (NASH) are becoming increasingly common in the developed world population. NAFLD and NASH are commonly seen in conjunction with features of the metabolic or insulin resistance (IR) syndrome (IRS), which include obesity, diabetes mellitus, hypertension, hypercholesterolemia and hyperlipidemia. We shall discuss herein the impact of these conditions on orthotopic liver transplantation (LT) addressing NAFLD and NASH as causes of end-stage liver disease requiring LT, the evaluation of NAFLD/NASH patients for LT, the use of steatotic livers as donor livers; the evaluation of potential live donors for steatosis; and recurrence or de novo NAFLD and NASH post-LT.


NAFLD and NASH as causes of end-stage liver disease leading to transplantation

Accurate data for the prevalence of NAFLD and NASH in the US population are not available. Based on autopsy and imaging studies, is estimated that NAFLD and NASH could affect 20% and 3% of the general population (1–3). Although not the only risk factor (4), obesity is the most prevalent risk factor for NAFLD, up to 20% of the obese population may have NASH (1). Additional risk factors for hepatic injury are increased age (5,6) and the presence of type II diabetes mellitus (7,8). In 2002, the National Health and Nutrition Examination Survey (NHANES) reported that 35% of the US adult population were overweight [body mass index (BMI) 25–29.9 kg/m2] and that an additional 30% were obese (BMI > 30kg/m2) (9). The Center for Disease Control (CDC) estimates from 1999 were only slightly more favorable, suggesting that 27% of the adult population were obese (10). It is predicted that by 2025, up to 45–50% of the adult population could be obese (11). Given the current estimates of the prevalence of NAFLD and NASH in normal weight and obese persons, the expected prevalence of NASH in the USA, in 2025 could be > 25 million affected persons. This prevalence would exceed by 10-fold the current USA prevalence of hepatitis C virus (HCV) (12).

The frequency with which NASH progresses to end-stage liver disease is uncertain. The reported prevalence of cirrhosis in case series of NASH patients varies from 3% to 15% (13–15). Studies of sequential liver biopsies showed 0–35% of cases progressing (increase of fibrosis of at least 1 level) with 0–7% progressing to cirrhosis over 3–8.3 years (Table 1). Of note, one of these studies showed a regression in the disease in 16% of patients over a mean of 3 years (16), and one showed progression of fibrosis in 7 out of 27 (26%) with a NAFLD index biopsy over a median of 8.3 years (17). This is of concern as NAFLD in the absence of NASH has previously been considered to be a relatively benign condition.

Table 1.  Histologic progression of NAFLD and NASH




n (%)

n (%)
(13)NASH424.53 (7)3 (7)
(14)NASH78.20 (0)0 (0)
(78)NASH224.37 (32)0 (0)
(17)NASH298.33 (10)1 (3)
 NAFLD277 (26)0 (0) 
(16)NAFLD/83329 (35)5 (6)

Accurate data on the percentage of LTs performed for NASH-related cirrhosis is not available. A single center study (18) noted that 2.9% of their primary liver transplants were performed for NASH. Two recent case–control studies suggest that up to 50% of cases identified as cryptogenic cirrhosis may in fact have arisen from NASH (7,19). Cryptogenic cirrhosis currently accounts for 7% of US LTs (20) and 8% of European LTs (21). In keeping with our understanding of IRS, obesity and diabetes were considerably more common among patients with cryptogenic cirrhosis than among cirrhotic controls or the general population (Table 2).

Table 2.  Prevalence of obesity and diabetes in patients with cryptogenic cirrhosis awaiting liver transplantation
Study referenceNo. of cases populationNo. of controlsaPrevalence obesitybGeneral ControlsDiabetescGeneral Controls
  1. aControls for Caldwell exclude those with diagnosis of NASH.

  2. bObesity defined as BMI > 31.1 (males), >32.3 (females) [Caldwell (7)], BMI > 30 [Poonawala (19)].

  3. cOverall prevalence of diabetes in USA is 6.2%, but 20% in those >65 years of age.


Furthermore, obesity may accelerate the progression of liver disease in HCV- (22) or alcohol-related liver disease (23). Rates of cirrhosis-related death are also increased among the obese (24).

If we estimate that 3% of LTs are undertaken for NASH-related cirrhosis, we calculate that 160 LTs were performed in 2002 for NASH from a total of 5329 (25). This approximates well to half of the 378 LTs performed for cryptogenic cirrhosis in 2002 and represents 0.002% of the population with NASH. If the current increase in prevalence of obesity and associated NASH continues, 0.002% of the > 25 million who could have NASH in 2025 would suggest a need of 500 LTs for NASH. This would not include those with end-stage liver disease excluded from LT due to co-morbid conditions. It is not clear what percentage of potential LT candidates with NASH are precluded from LT due to the presence of cardiovascular disease or other co-morbid conditions, but the IR syndrome is associated with a clustering of coronary artery disease risk factors, and even as an independent risk factor, obesity is associated with a 4.6 relative risk of cardiac mortality (26).


The pathophysiology of NAFLD and NASH has not been elucidated, although the ‘two-hit’ theory proposed by James and Day has received considerable acceptance (27). They proposed that insulin resistance (IR) (the first hit) leads to steatosis of the liver, and the presence of oxidative stress (OS) (the second hit) in conjunction with IR leads to NASH. Here is a brief summary of some of the data that underscores the ‘two-hit’ hypothesis.

IR was demonstrated in 54 of 55 patients with NASH compared with 3 of 36 patients with hepatitis C and 0 of 20 normal subjects (28). IR has also been documented in 19 lean patients with histologically documented NASH compared with 19 age-, sex- and BMI-matched controls (29). Furthermore, the hyperinsulinemia was shown to stem from pancreatic hyper secretion, not decreased hepatic insulin clearance, which can be seen in cirrhosis (30). IR was also demonstrated in patients with NASH (n = 6) and NAFLD (n = 6) when compared with normal controls (n = 6) (31). All subjects in these studies were retrospectively chosen.

Regarding OS, in situ staining for 5-nitrotyrosine (31) and 8-hydroxygeoxyguanosine (32) (semi-quantitative markers for OS) were increased in liver biopsies from NASH patients when compared with control livers and obesity has been associated with increased excretion of urinary isoprostanes (33) (a marker of OS), even when controlling for diabetes mellitus – another known association with OS (34).

Despite the putative association between IR and NAFLD/NASH, the mechanism whereby IR and OS might predispose to NAFLD and NASH are not entirely clear. Tumor necrosis factor-alpha (TNF-α) may be a mediator (35). Its levels are increased in the obese, particularly with visceral rather than peripheral obesity (36). It inhibits the action of insulin by deactivating insulin receptor substrates (37) and is associated with increased oxidative stress perhaps via increased mitochondrial injury (38). Regardless, excessive delivery of free fatty acid to the liver from peripheral sources and or impaired utilization or excretion of fatty acids from the liver will lead to hepatic steatosis. Likewise, free radicals from a variety of sources, including free fatty acids or impaired mitochondrial function, or inadequate antioxidant protection, could lead to increased OS in the liver (3) (Figure 1).

Figure 1.

A schematic of the development of non-alcoholic fatty liver disease (NAFLD) and non-alcoholic steatohepatitis (NASH). Insulin resistance with or without other factors leads to the accumulation of hepatic fat. Oxidative stress or other agents in the presence of steatosis lead to cellular injury. Other environmental and genetic factors may influence the likelihood of disease progression. Adapted from Nonalcoholic Steatohepatitis: Summary of an AASLD single topic conference. Reproduced by permission of Wiley-Liss, Inc., a subsidary of John Wiley & Sons, Inc., from: Neuschwander-Tetri B, Caldwell S. Nonalcoholic steatohepatitis: summary of an AASLD single topic conference. Hepatology 2003; 37: 1202–1219.

Pre-transplant implications

The discrepancy between supply and demand for donor livers

The demand for LT continues to outstrip the supply of available cadaveric donor livers. Each year, over 1000 patients on the waiting list die because of a lack of a suitable donor (20). Steatosis of the donor liver is associated with an increased risk of primary non-function (PNF) in the allograft. PNF has a high mortality and usually requires re-listing and emergent re-transplantation. Because outcomes for re-transplantation are less favorable than for primary transplantation (67–69% vs. 86–87% 1-year survival) (20, 39), there is a tension between the wish to put every potential donor liver to use and the wish to avoid graft loss from PNF.

The use of steatotic livers for transplantation

Prevalence of NAFLD/NASH in donors. As mentioned above, it is estimated that 20% of the population have hepatic steatosis. In a survey of 94 LT surgeons from the UK and US, 65% of the respondents estimated that at least 20% of recovered livers were steatotic, 17% estimated that greater than 40% of retrieved livers were affected (40). Steatosis is typically non-uniform and is difficult to assess accurately, particularly when it is less severe (41). Only 13% of the respondents biopsied all livers, thus the above data may underestimate the prevalence of NAFLD.

Living donor liver transplantation

In order to optimize donor and recipient outcome, most centers aim to select living donors with less than 10% hepatic steatosis and are very reluctant to use livers with more than 30% steatosis (42,43). One-third to one-half of potential living liver donors have some degree of steatosis on liver biopsy and in 23–35% of cases the steatosis represents more than 10% of the liver tissue (44–46). Furthermore, given the importance of avoiding morbidity in the donor, the increasing prevalence of co-morbid conditions with increasing BMI, some centers automatically exclude potential donors with a BMI > 28 (42,44), thus further limiting the pool of potential donors. The role of liver biopsy in the evaluation of potential donors remains under debate with practices varying widely between centers (43). MRI and CT may fail to detect mild (<10%) degrees of steatosis having a 45–55% false–negative rate (44,45). This decreases to 8–20% for more severe (>30%) steatosis. Yet, given the need to exclude such individuals from live liver donation, it is probably prudent to biopsy the majority rather than the minority of potential living liver donors.

Outcomes after transplantation using steatotic donor livers

The use of grafts with macrovesicular steatosis is associated with increased rates of PNF and poorer outcome (47). When 59 patients receiving livers with up to 30% macrovesicular steatosis were compared with 57 patients receiving livers without fatty infiltration, the recipients of the steatotic livers had higher rates of PNF (5.1 compared with 1.8%), and worse 2-year patient (77% compared with 91%) and graft (70% compared with 82%) survival. Recipient status was similarly distributed between the varying grades of steatosis, refuting the concern that poorer outcomes among recipients of fatty livers are confounded by greater severity of illness pre-OT. In contrast, microvesicular steatosis does not seem to carry the same risk of PNF, as does macrovesicular steatosis. A single center study of 426 liver transplants compared the outcome of LT using donor livers with at least 30% microvesicular steatosis (n = 40) to that of the remaining 386 livers with < 30% microvesicular steatosis (48). Rates of PNF (5.0% and 5.1%), and 1-year patient survival (80% compared with 79.8%) and graft survival (72.5% compared with 68.4%) were similar in both groups.

There is no consistent threshold of estimated macrovesicular fat deposition above which LT is precluded. Additional factors such as diabetes mellitus in the donor or poor health in the recipient may influence the outcome of transplantation. Thus, whereas livers with more than 60% macrosteatosis should probably be excluded automatically, livers with more moderate steatosis (30–60%) may be utilized in the absence of additional risk factors in the donor or recipient, or when the recipients circumstances are judged to be sufficiently critical to justify the risk (49).

Mechanisms of PNF or poor early graft function

The mechanisms whereby steatosis of the donor organ leads to primary non-function are not understood completely, but several have been proposed (41,49). They include impaired metabolism in the steatotic hepatocytes (50–52); the physical effects of lipid (53,54), particularly during cold ischemia (55); diminished portal blood flow (54,56) and increased sensitivity to OS on reperfusion (57–59) (Table 3).

Table 3.  Proposed mechanisms of poor graft function in steatotic livers (41,49)
Diminished portal flowBallooned hepatocytes distort sinusoidal lumen causing increased resistance, reduced blood flow and secondary ischemia(54,56)
Inefficient anerobic metabolismSteatotic hepatocytes has increased uncoupling protein, and decreased mitochondrial ATP production(50–52)
 Hepatocyte energy level correlates with post-LT outcome 
Physical properties of lipidAltered plasma membrane fluidity in the sinusoids of steatotic livers leads to increased Kupffer cell adhesion and activation on reperfusion. Alternatively,lipid solidifies during cold preservation and may cause physical disruption to hepatocytes(53–55)
Oxidative stressThe steatotic liver is predisposed to OS at baseline. OS is a key component of reperfusion injury of the newly perfused graft. Administration of tocopherol, an oxygen radical scavenger, improves survival of Zucker rats (a model of NAFLD) exposed to ischemia/reperfusion injury(57–59)

Impact of the insulin resistance syndrome on morbidity and mortality after liver transplantation

There are no data directly addressing the impact of IRS on outcome after LT and inferences must be made by evaluating the impact of some of its component features. Diabetes prior to LT is an independent risk factor for poor outcome. Shields et al. (60) compared the outcome of LT in patients with and without diabetes (defined according the World Health Organization criteria). They found survival of only 64% in recipients who were diabetic compared with 91% in non-diabetic recipients. The data assessing the impact of obesity on outcome are more mixed. Data from a single-center transplant series suggest that pre-LT obesity is not a risk factor for post-LT morbidity and mortality (61). In contrast, a review of the UNOS database, evaluating 3,877 patients with a BMI > 30, showed poorer survival at all time points (immediate to 5-years post-LT) largely as a result of cardiovascular events (62). The increase in mortality is seen despite careful pre-operative evaluation of LT candidates to exclude those with significant co-morbid conditions and in at least one study, LT recipients whose underlying liver disease was attributed to NASH were younger than those in the control groups (7).

Evaluation for OLT of patients with NASH or cryptogenic cirrhosis

Assessment of severity of liver disease. Evaluation of patients with NASH or cryptogenic cirrhosis for LT should follow the same guidelines as that of patients with other forms of liver disease. In each case, their liver disease should be of such severity so as to warrant LT and the patient should be free of co-morbidities that would mitigate against a good outcome.

Obesity is associated with an increased risk of cardiovascular disease, due to its association with hypertension, diabetes, hypercholesterolemia and sedentary lifestyle. A poor functional capacity, e.g. inability to walk up a flight of stairs has been shown to predict increased peri-operative and long-term risk (63).

Through its links with obstructive sleep apnea, restrictive lung disease and thromboembolic disease, obesity is a risk factor for pulmonary hypertension, independent of the porto-pulmonary hypertensive syndrome. Pulmonary hypertension leads to an increased risk of peri-operative mortality in the setting of LT. At the time of reperfusion, the venous drainage from the newly perfused liver has negative inotropic and arrhythmogenic effects. In the setting of pulmonary hypertension, this leads to acute right heart failure, circulatory collapse and congestion and ischemia of the engrafted liver. Even when patients with severe pulmonary hypertension survive the liver transplant operation, ongoing right ventricular failure may ensue with approximately 50% 1-year mortality (64). Thus, a mean pulmonary artery pressure of 50mmHg, although considered ‘moderate’ by the pulmonologists, is a contra-indication to LT (64,65). Obese patients should also be screened for obstructive sleep apnea as the associated daytime somnolence may be mistaken for hepatic encephalopathy. Previously unrecognized diabetes mellitus should be ruled out and evaluated if found.

Fatty liver disease after liver transplantation

There are now several case reports of NASH developing de novo or recurring in the LT recipient (61,66,67). Steatosis has been reported to occur within 6 months and cirrhosis within 2 years of transplantation in those receiving LT for NASH (66,67). In a case–control study comparing patients transplanted for cryptogenic cirrhosis (cases) with those transplanted for alcohol or cholestatic liver disease (controls), the development of post-LT steatosis was significantly more frequent in the case compared with the control group (50% and 25% respectively) (61).

Liver transplant recipients demonstrate an increased prevalence of several components of IRS (Table 4) (68). Both corticosteroids and calcineurin inhibitors promote hypertension and hypercholesterolemia. Prednisone is a risk factor for fatty liver. Tacrolimus and cyclosporine A, (but particularly tacrolimus) are diabetogenic (69).

Table 4.  Risk factors for insulin resistance syndrome/fatty liver disease after liver transplantation and in the general population

Risk factor
Prevalence post-
(79,80) Rate in
US population
  1. HDL: high-density lipoprotein.

  2. Modified from Table 2, reference 68.

Hypertension (BP > 140/90)41–81%15.7%
Hypercholesterolemia (240 mg%)20–66%14.9%
HDL < 35 mg%52%12%
Diabetes mellitus21–32%6.2%
Obesity (BMI > 30)39–43%16.1%

The metabolic aberrations seen after LT, described in Table 4, are associated with hepatic steatosis. Hypercholesterolemia (70) and diabetes mellitus (71,72) have been associated with increased OS. In addition, we have shown that the post-LT state is associated with OS (73) (Figure 2). Fifty LT candidates had urinary isoprostane levels measured pre-LT, intra-operatively and during the first post-operative year. When compared with healthy volunteers, urinary isoprostanes were markedly elevated prior to LT. Levels increased further during reperfusion of the allograft. Then fell significantly although never reached normal levels. Over the following year, levels increased steadily until they were no longer significantly different from the pre-LT levels. There was no significant difference among those with or without hepatitis C, acute cellular rejection or a composite of adverse events. Thus IR and OS, components of the ‘two-hit’ hypothesis for the development of NADLD and NASH, may emerge after LT. Therefore, it is of little surprise that NAFLD and NASH are observed after LT also.

Figure 2.

Urinary dinor-dihydro iPF-III excretion in liver transplant recipients (LT) and healthy volunteers. Dinor-dihydro iPF-III levels in LT subjects at all time points were significantly different to healthy volunteers (dark grey bar). Pre-operative levels are almost fourfold those of normal volunteers. Levels increase further at reperfusion. Levels fall significantly in the early post-operative period but remain twice normal. Dinor-dihydro iPF levels climb steadily until at 6–12 months post-LT they are not significantly different to pre-LT levels (*significantly different to pre-operative levels, p < 0.05).

Management of the insulin resistance syndrome after liver transplantation

Remarkably little has been written on the management of IRS or metabolic syndrome after LT. De novo research on the topic is almost non-existent. However with a 3-year survival of approximately 80%, increasing recognition of the long-term medical problems of the liver transplant recipient and a likely increase in the rates of LT for NASH there is need for research in this area.

In the non-transplant population, the mainstay of management is weight reduction, ideally through modification of lifestyle, and pharmacological management of the various components, e.g. hypertension or diabetes, as they arise.

There have been several pilot studies on pharmacological intervention with ursodeoxycholic acid, antioxidants and insulin-sensitizing agents. These were recently reviewed elsewhere (3). Many of these studies showed promising results; however the studies were limited by small numbers of subjects, lack of randomization, use of surrogate markers that have not been validated and short duration of therapy. Furthermore, concern remains with the use of these agents with the potential for hepatoxicity in a liver already at risk of injury. Among the more promising is a recent publication suggesting that the PPARγ agonist rosiglitazone may improve NASH. In an open label study, 13 out of 22 (59%) patients showed histologic improvement after 48 weeks of therapy (74). Unfortunately this was accompanied by significant weight gain. Thus although medications especially antioxidants are commonly used in patient with NAFLD and NASH, it should be noted that no specific pharmacological treatment has been shown to be effective in the treatment of NAFLD (75).

Physical fitness has been shown to be as powerful a modulator of insulin resistance as body weight, each independently accounting for approximately 25% of the differences in insulin-mediated glucose disposal in non-diabetic individuals (76). Exercise has been shown to be associated with better blood glucose, blood pressure and lipid levels and decreased cardiovascular mortality in the non-transplant population. Unfortunately, only a minority (24%) of long-term liver transplant survivors achieve the level of physical activity recommended by the Surgeon General (77). It has been suggested that the metabolic benefits of exercise are related to overall energy expenditure rather than intensity of exercise and can occur without significant changes in cardiovascular fitness. Thus, while recognizing that fatigue is a major limitation in functional capacity for patients with end-stage liver disease and post-LT, patients should be encouraged to engage in regular physical activity. Moreover, it is intuitive that early application of a physical exercise program in conjunction with a calorie-controlled nutritionally adequate diet would decrease the rates of post-LT hypertension, diabetes, obesity and hypercholesterolemia. However, it is not known how best to achieve these aims nor how effective they might be. It is unlikely that lifestyle measures alone will be sufficient for control of IRS syndrome post-LT but should be combined with antihypertensive and lipid lowering medications, glycemic control and smoking cessation.

Diet and exercise has also been shown to decrease weight and improve transaminases in non-LT patients with NASH and several studies are underway to assess the role of antioxidants and insulin-sensitizing agents in the treatment of NASH. The role of these therapeutic modalities in the transplant setting remains to be investigated.