Hepatology over the years


  • Steven Schenker,

  • D. Montgomery Bissell,

  • Andres T. Blei

  • Potential conflict of interest: Nothing to report.

An overview of the material published in Hepatology over 25 years is not a minor task. The breadth and depth of original articles are impressive, encompassing multiple areas of research, from basic to clinical, from classic physiology to molecular biology. We chose two approaches for a brief summary.

  • 1The accompanying tables list widely cited articles. Table 1 (on page S5) illustrates the most quoted research articles, some of which will be easily recognized by card-carrying hepatologists, such as the data that led to the development of the “Knodell score.” Hepatitis C dominates the list, and Hepatology has been a vehicle for the dissemination of knowledge in this area, even in the days of nonA-nonB hepatitis. Table 2 (on page S5), the most cited review articles, illustrates the influence of transplantation, and especially Thomas Starzl, on the field. Table 3 (on page S5), most cited articles in pathobiology, covers the gamut of research interests reflected in our pages.
  • 2We also provide brief commentary on a few highly cited articles. The length of the text varies, not the richness of the experience at being at the helm of the journal.

Orthotopic Liver Transplantation for Acute and Subacute Hepatic Failure in Adults

Peleman RR, Gavaler JS, Van Thiel DH, Esquivel C, Gordon R, Iwatsuki S, Starzl T

Hepatology 1987;7:484–489


The role of liver transplantation in 29 patients with fulminant and subacute hepatic failure due to a variety of different causes was examined by comparing the outcome and a variety of “hospitalization” variables. Transplanted patients (n = 13) were more likely to survive (p less than 0.05), were younger (p less than 0.05) and spent more time in the hospital (p less than 0.025) than did those who were not transplanted (n = 16). Despite spending a much longer time in the hospital, transplanted patients spent less time in the intensive care unit (p less than 0.05) in coma (p less than 0.01) and on a respirator (p less than 0.01) than did those not transplanted. Most importantly, the survival rate for transplanted patients was significantly improved (p less than 0.05) as compared to those not transplanted. We conclude that liver transplantation can be applied successfully to the difficult clinical problem of fulminant and subacute hepatic failure.


The conclusion of this pioneering report has passed the test of time. By purely scientific criteria the Starzl paper had many limitations. It dealt with a small number of patients, the composition of the group was unusual in the large number of patients with Wilson's disease, and the individuals transplanted versus those not having a transplant were not consecutive or apparently matched as to severity and duration of disease. The latter is understandable as the prognosis of this type of liver failure with supportive care alone was dismal, with a mortality of about 90%.1 Despite the shortcomings of a retrospective study, the survival improvement with transplantation was very impressive, about 3-4-fold that of nontransplanted individuals (Fig. 1).

Figure 1.

Actuarial life table analysis of the 29 patients evaluated with acute fulminant and subacute hepatic failure divided into two groups: those transplanted (solid squares) and those not transplanted (open circles). Peleman RR et al. Hepatology 1987;7:484.

On comparing the patients in this study with those now considered for transplantation, there are key similarities. The current definition of acute liver failure is based on the presence of encephalopathy, coagulopathy and brief duration of liver disease.2 This corresponds to the patients in the Starzl study, except for the patients with Wilson's disease who clearly had prior liver disease. This group, however, is often included in the fulminant category, given an appropriate clinical presentation. Current spontaneous survival depends greatly on the cause of liver failure, with the best prognosis in patients with acetaminophen-induced acute liver failure, for which a specific antidote is available.3 For those failing this therapy and patients with other causes of acute liver failure, liver transplantation is still the treatment of choice, with a one year survival of about 70% as compared with less than 20% in the absence of a new liver.2

Major issues in getting the appropriate candidates to a timely transplantation remain. These are the overall scarcity of organs, the need to obtain a donor liver over a short time consistent with the rapid progression of acute liver failure, the procurement of requisite funds and the identification of individuals who may survive without a transplant versus those who will die without it. The complexity of managing such patients requires specialized units within a center offering liver transplantation.

There has been considerable discussion recently about the optimal approach to assessing the prognosis of patients with acute liver failure. Outcomes for liver failure due to acetaminophen-induced damage and that due to other causes vary. One approach (Clichy) employs the level of factor V,4 but this has not been prospectively evaluated. A more widely accepted set of criteria derive from studies done at Kings College.5 This approach uses the prothrombin time (INR) alone, or a combination of INR, age and etiology of liver failure. These criteria have been shown to have good specificity in identifying the need for transplant but low sensitivity (failing to pick up some patients in need of transplant).6 A number of new tests to enhance the sensitivity of King's College criteria have been suggested — an elevated lactate,7 a high serum phosphate8 and most recently, a major decrease of Gc-globulin.9 All require further validation. There is also a need for more donor organs and for development of bridging devices to maintain liver function until transplantation. In the interim, hepatic transplantation for acute liver failure, first reported to be effective some 20 years ago in Hepatology ,10 has now been accepted as an established and life-saving procedure.

Increased Tumor Necrosis Factor Production by Monocytes in Alcoholic Hepatitis

Craig J. McClain and Donald A. Cohen

Hepatology 1989;9:349–351


Tumor necrosis factor is a cytokine that mediates many of the biologic actions of endotoxin. Recent studies have shown that tumor necrosis factor administration may cause liver injury and that tumor necrosis factor may mediate the lethality of the hepatotoxin galactosamine. One of the most potent inducers of tumor necrosis factor production is endotoxin. Because patients with alcoholic liver disease frequently have endotoxemia and because many of the clinical manifestations of alcoholic hepatitis are known biologic actions of tumor necrosis factor, we thought it would be important to evaluate tumor necrosis factor activity in patients with alcoholic hepatitis. Basal and lipopolysaccharide-stimulated tumor necrosis factor release from peripheral blood monocytes, a major source of tumor necrosis factor production, was determined in 16 patients with alcoholic hepatitis and 16 healthy volunteers. Eight of 16 alcoholic hepatitis patients and only two of 16 healthy volunteers had detectable spontaneous tumor necrosis factor activity (p less than 0.05). After lipopolysaccharide stimulation, mean monocyte tumor necrosis factor release from alcoholic hepatitis patients was significantly increased to over twice that of healthy controls (25.3 ± 3.7 vs. 10.9 ± 2.4 units per ml, p less than 0.005). We conclude that monocytes from alcoholic hepatitis patients have significantly increased spontaneous and lipopolysaccharide-stimulated tumor necrosis factor release compared to monocytes from healthy volunteers. We suggest that some of the metabolic abnormalities and possibly some of the liver injury of alcoholic hepatitis may be due to enhanced tumor necrosis factor production.


The article cited above1 was among the first to suggest a role for tumor necrosis factor α (TNFα) in alcoholic liver disease. This concept (and the closely related area of cytokine regulation and balance) has evolved greatly since then, both as to the complexity of cytokine biology, the pathogenesis of this disease and approaches to its treatment.2,3

Alcoholic liver disease is an important cause of liver damage in the United States, often coexisting with chronic hepatitis C. Whereas most early studies on the pathogenesis of alcoholic liver injury focused on the relative roles of nutrition versus alcohol or acetaldehyde, and on the importance of excess fat on subsequent fibrosis, those of the last 10-15 years have been more directed toward the effects of inflammation and cell damage on the progression of the disease. This, in part, was due to development of animal models of alcohol-induced liver injury4, 5 and of techniques for assay of cytokines.

A possible role for TNFα in alcoholic liver injury surfaced with reports that TNFα concentration correlated with severity of disease and mortality.5,7,8 There was also an increase in cytokines (interleukin-6, -8 and -18) that are induced by TNFα,, particularly in more severe liver disease.9 The findings appeared to link the systemic manifestations of alcoholic liver disease with the known metabolic effects of some cytokines.5

Since the original contribution of McClain and Cohen,1 studies in experimental animals have confirmed a role for TNFα in the pathogenesis of this disorder. In the intragastric alcohol feeding model, liver injury correlated with increased TNFα mRNA in the liver4,10 and a decrease in the protective IL-10 cytokine production.2,5 TNFα is a low molecular weight polypeptide with 157 amino acids generated by macrophages, primarily Kupffer cells and monocytes. Endotoxin from the intestinal tract and reduced oxygen intermediates are potent stimuli of TNFα production by Kupffer cells. A variety of techniques have been designed to interrupt/modify this sequence including antibiotics to decrease endotoxin production in the gut, gadolinium chloride to “block” Kupffer cell function, and antioxidants to protect the mitochondria from oxidative stress.5 All decreased liver injury. Moreover, liver damage was largely prevented in animals receiving TNFα antibody11 and in those without a TNFα receptor.12 As discussed in a recent review,2 in alcoholic liver disease Kupffer cells are apparently primed by endotoxin, reduced oxygen intermediates and possibly other mediators to generate TNFα. Subsequent sensitization of hepatocytes by a variety of factors such as depletion of reduced glutathione, lipid peroxidation products and/or decreased proteasomal function may result in cell death and continuing inflammation. A simplified schematic of mechanisms of alcohol-induced liver injury is shown in Fig. 2.

Figure 2.

Key tissues, cells, and mediators involved in alcoholic liver cell injury. Kaplowitz, AASLD, 1991.

With regard to human liver disease from alcohol, until the present, abstention was the only proven approach of value,13 along with enteral nutrition for selected groups.14 However, mechanism-based therapy was lacking. Efforts to target TNFα started with the use of corticosteroids15 in patients with severe alcoholic hepatitis and no contraindications, although it was recognized that corticosteroid have broad effects.3,16 The treatment improved survival over 28 days by about 20%.16 Since corticosteroid use for a month can lead to an increase in sepsis, it has recently been suggested that a fall in serum bilirubin within 7 days of corticosteroid therapy may identify early responders.17 Pentoxyphylline, an inhibitor of TNFα synthesis,18 also has been shown to benefit patients with severe alcoholic hepatitis, primarily by decreasing the development of hepatorenal syndrome.19 While the drug has minimal side effects, its benefits require confirmation. Initial reports are appearing on the use of infliximab, a chimeric human/mouse antibody that blocks TNFα. The drug has been used effectively in patients with rheumatoid arthritis, Crohn's disease and recently in psoriasis. At present two pilot and two small randomized studies have been done.20–23 These vary with regard to patient selection, drug dose, the use of corticosteroids and the end point. Hence, comparisons are not possible. Some increase in sepsis was observed with a high dose of infliximab. Measurements of cytokines in one study suggested a beneficial response.21 In another pilot study using etanercept, an agent that binds and neutralizes soluble TNFα, interleukin-6 (a surrogate marker of TNFα) was decreased, but some side effects were seen leading to a premature discontinuation of the drug in 23% of patients.24 Clinical benefit could not be assessed because of the small patient number, which will mandate further studies if the safety issues are resolvable. Whatever the future may bring, the initial study of McClain and Cohen1 expanded the potential targets for therapeutic intervention in alcohol-related injury and, more generally, encouraged exploration of cytokine balance in liver disease.

Up-regulation of the Mmultidrug Resistance Henes, Mrp1 and Mdr1b, and Down Regulation of the Organic Anion Transporter, Mrp2, and the Bile Salt Transporter, Spgp, in Endotoxemic Rat Liver

Vos Ta, Hooiveld GJE, Konong H, Childs S, Meijer DKF, Moshage H, Jansen PLM, Muller M

Hepatology 1998;28:1637–1644


Endotoxin-induced cholestasis is mainly caused by an impaired canalicular secretion. Mrp2, the canalicular multispecific organic anion transporter, is strongly down-regulated in this situation, and canalicular bile salt secretion is also reduced. We hypothesized that other adenosine triphosphate–binding cassette (ABC) transporters may compensate for the decreased transport activity to protect the cell from cytokine-induced oxidative damage. Therefore, we examined the expression of ABC-transport proteins in membrane fractions of whole liver and of isolated hepatocytes of endotoxin-treated rats and performed reverse-transcriptase polymerase chain reaction (RT-PCR) on mRNA isolated from these livers. In addition, the localization of these transporters was examined using confocal scanning laser microscopy. By 6 hours after endotoxin administation, we found a clear increase of mrp1 mRNA and protein, whereas mrp2 mRNA and protein were decreased. This was confirmed in isolated hepatocytes. In addition, mdr1b mRNA was strongly increased, whereas mdr1a and mdr2 mRNA did not change significantly. Both the mRNA and protein levels of the sister of P-glycoprotein (spgp), the recently cloned bile salt transporter, decreased. After endotoxin treatment, the normally sharply delineated canalicular staining of mrp2 and spgp had changed to a fuzzy pattern, suggesting localization in a subapical compartment. We conclude that endotoxin-induced cholestasis is caused by decreased mrp2 and spgp levels, as well as an abnormal localization of these proteins. The simultaneous up-regulation of mrp1 and mdr1b may confer resistance to hepatocytes against cytokin-induced metabolic stress.


The initial characterization of cholestasis, based on morphological parameters denoting bile stasis in cellular elements of the liver, has been replaced by a functional definition centered on the impairment of bile flow. This concept, well articulated by Hans Popper in an Editorial in volume 1 of Hepatology ,1 has been greatly expanded lasting recent years. The discovery, molecular characterization and functional assessment of hepatocellular transporters have completely changed our views of the physiology of bile formation and its derangement in disease.2 The widely quoted paper from Vos et al. illustrates the new biology that has emerged from this successful research effort.

It was well known that liver cells exhibit two polar domains, one involved in the uptake of endo/xenobiotics from the bloodstream and the other in efflux into the bile canaliculus. What has revolutionized the field is the dissection of its components.3,4 On the sinusoidal side, sodium dependent bile salt uptake occurs through the action of a series of transporters from the SCL10 family.5 Sodium-independent organic anion uptake occurs through another family of organic anion transport proteins (OATP)6 with broad substrate specificity.

At the other pole, canalicular transport systems include multidrug resistance P-glycoproteins, which mediate the efflux of endogenous and exogenous compounds, and whose activity is ATP dependent (hence the term ABC transporters, ATP-binding cassette). While such transporters may be present in multiple cell types, the bile salt transporter, BSEP (bile salt export protein, previously termed Spgp) is the predominant conduit for bile salt efflux from hepatocytes.7 The driving force for bile salt-independent flow, via the excretion of reduced glutathione, is MRP2,8 a member of the multidrug resistance-associated proteins.

Only 20 years ago, endotoxin-induced cholestasis was being examined in the isolated perfused liver,9 with the uptake and excretion of BSP and ICG being used to infer mechanistic pathways for the action of endotoxin. Such was the state of the field at that time. Now, as demonstrated by Vos and colleagues, the molecular effects of endotoxin are defined entirely as alteration of defined transporters: downregulation of the two transporters involved in bile salt dependent and independent bile flow, and upregulation of transporters conferring resistance to the effects of cytokines. Additional possibilities include the ability of MRP1 and Mdr1b to increase the efflux of organic anions across the basolateral hepatocyte membrane, a pathway that teleologically could protect the hepatocyte from the accumulation of toxic substances in cholestasis.10 Pathways involved in this adaptive change have also been dissected, with different FXR-mediated effects that lead to the upregulation of the activity of compensatory transport systems (Fig 3.).11

Figure 3.

Upregulation of the Multidrug Resistance Genes, Mrp1 and Mdr1b, and down regulation of the organic anion transporter, Mrp2, and the bile salt transporter, Spgp, in endotoxemic rat liver. Basolateral uptake systems are shown in blue, basolateral ATP-binding (ABC) transporters are shown in red. Canalicular ATP transporters are shown as green ovals while presumptive ABC lipid translocases (flippases) are symbolized by green circles. Figure reproduced from Pauli-Magnus C, Meier PJ. Hepatocellular transporters and cholestasis. J Clin Gastroenterol. 2005;39(4 Suppl 2):S103–S110, copyright © 2006, Lippincott Williams & Wilkins. Reprinted with permission from the authors and publisher.

Cholestasis of sepsis is one example of several clinical entities where our understanding of pathogenesis has been fundamentally altered by knowledge of hepatocellular/canalicular transporters. Another is Dubin-Johnson syndrome (familial conjugated hyperbilirubinemia), which is due to genetic variation in MRP2.12 Other types of familial intrahepatic cholestasis involve deficiency or dysfunction of FIC-1 (a P-type ATPase), BSEP and MDR3, respectively,13 Cholestasis of pregnancy also may be linked to mutations in the MDR3 gene.14 Benign recurrent intrahepatic cholestasis, in its two forms, is linked to dysfunction of FIC-1 and BSEP.15 The entire spectrum of acquired pathology, from sepsis-induced to drug-induced cholestasis to primary biliary cirhosis, is being re-examined under this new light, where it is likely that endo- and xenobiotics interact with different genetic backgrounds and result in different susceptibility to disease.

The concluding sentence of Dr. Popper's editorial was “Rational therapy of intrahepatic cholestasis remains elusive.”1 While that remains the case, a deeper understanding of the genetic and molecular underpinnings of the nature of bile transport is the basis for rational therapies and constitutes a major advance.

Model to Predict Poor Survival in Patients Undergoing Transjugular Intrahepatic Portosystemic Shunts (TIPS)

Malinchoc M, Kamath PS, Gordon FD, Peine CJ, Rank J, ter Borg PCJ

Hepatology 2000;31:864–871


Transjugular intrahepatic portosystemic shunts (TIPS) may worsen liver function and decrease survival in some patients. The Child-Pugh classification has several drawbacks when used to determine survival in such patients. The survival of 231 patients at 4 medical centers within the United States who underwent elective TIPS was studied to develop statistical models to (1) predict patients survival and (2) identify those patients whose liver-related mortality post-TIPS would be 3 months or less. Among these elective TIPS patients, 173 had the procedure for prevention of variceal rebleeding and 58 for treatment of refractory ascites. Death related to liver disease occurred in 110 patients, 70 within 3 months. Cox proportional-hazards regression identified serum concentrations of bilirubin and creatinine, international normalized ratio for prothrombin time (INR), and the cause of the underlying liver disease as predictors of survival in patients undergoing elective TIPS, either for prevention of variceal rebleeding or for treatment of refractory ascites. These variables can be used to calculate a risk score (R) for patients undergoing elective TIPS. Patients with R > 1.8 had a median survival of 3 months or less. This model was superior to both the Child-Pugh classification as well as the Child-Pugh score, in predicting survival. Using logistic regression and the same variables, we also developed a nomogram that indicates which patients survive less than 3 months. Finally, the model was validated among an independent set of 71 patients from the Netherlands. This Mayo TIPS model may predict early death following elective TIPS for either prevention of variceal rebleeding or for treatment of refractory ascites.


Clinical hepatologists are called to provide optimal management of portal hypertension and its complications, especially variceal bleeding. This was the case 50 years ago1 and it is still the case today, when managing such complications has the additional challenge of maintaining patients on the liver transplant waiting list.2 Progress has been made, as witnessed by the improved outcomes after variceal bleeding, with a clear reduction in inpatient mortality over a 20 year period.3

The two major tools to predict outcome in liver failure have stemmed from procedures designed to correct portal hypertension. The Child score4 and its subsequent iterations5 were designed to predict surgical mortality after portal decompression. The chosen variables included signs of portal hypertension as well as laboratory markers of the synthetic and elimination capacity of the liver. The variables were chosen empirically, and limitations of the Child score have been recently discussed.6 These include the arbitrary nature of the cutoff values and a similar weight given to each of 5 parameters. Nonetheless, the score has been applied to estimating prognosis in a wide range of liver conditions, an evolution reflected in the journal.7–10

The Mayo TIPS model, presented in this publication, was also a product of necessity for predicting outcomes in a nonsurgical portal decompressive procedure.11 The score was obtained from a cohort of more than 200 patients in the United States. Univariate analysis and the subsequent multivariate analysis determined the relative importance of serum creatinine, plasma bilirubin and INR; etiology of liver disease was also a factor. An independent validation cohort in the Netherlands confirmed the findings from the American experience (Fig. 4). The paper stands as a model on how models should be developed.

Figure 4.

Survival of 71 independent TIPS patients from the Netherlands who were stratified according to their risk score into two risk groups, namely a high risk group with a median survival less than 3 months (R > 1.8) and a low risk group with a median predicted survival more than 3 months (R < 1.8). Actual (Kaplan-Meier) and unexpected survival using the Mayo model were compared using the one sample log rank test. For the low- and high-risk patients, the observed and expected survival were similar (P = .88 and P = .41, respectively).

As in the case of the Child score, the Mayo TIPS score acquired a life of its own. In 1998 the Department of Human and Health Services called for the development of an instrument to ensure equitable allocation of organs based on parameters of disease severity. After a quick adaptation, the original system metamorphosed into the MELD score (Model for End Stage Liver Disease), as a reasonably accurate tool for predicting survival in patients with end-stage liver disease, including those awaiting liver transplantation.12 Recent examination of the performance of the MELD score for patients on the UNOS list has shown remarkable correlation between mean values and the risk of death.13 As in the case of the Child classification, the usefulness of MELD as a prognostic tool in several clinical entities was reflected in the pages of the journal.14–16

MELD has changed the optic through which liver failure is viewed, emphasizing the role of kidney dysfunction in the short-term prognosis. The mechanisms that lead to renal dysfunction in liver failure are complex, but a central theme is the response of the kidney to arterial vasodilatation, a hypothesis well articulated in the pages of Hepatology .17 Further refinements of MELD are expected, with serum sodium the prime candidate for addition to the score18–20; hyponatremia is also the result of the body's response to arterial vasodilatation. In only 5+ years since its introduction, MELD has become the primary bedside tool for tracking progression in end-stage liver disease.