Liver failure and liver disease


  • Potential conflict of interest: Nothing to report.

In any stewardship of Hepatology the Editor will dream of announcing discovery of the cause of chronic liver diseases such as primary biliary cirrhosis (PBC) and primary sclerosing cholangitis (PSC). The likelihood of this happening during the next quinquennium must be high with a slightly reluctant nod in the direction of genomic and proteomic screening which threatens the preserve of the Claude Bernard clone of traditionalist researchers, haughtily insistent on a working hypothesis. Despite containing no such equivalent of the “eureka” experience Hepatology publications in the past few years have steadily increased our understanding of adult liver diseases, most significantly in the area of non alcoholic fatty liver disease (NAFLD).


PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis; NAFLD non alcoholic fatty liver disease; NASH, non alcoholic steatohepatitis; VLDL, very low density lipoprotein; BMI, body mass index; MRC, mitochondrial respiratory chain; WHVPG, wedged hepatic venous pressure gradient; MELD, model for end-stage liver disease; HRS, hepatorenal syndrome; TIPS, trans jugular intrahepatic portal-systemic shunt; ALF, acute liver failure; ICP, intracranial pressure; TCA, tricarboxylic acid.


Despite the many new insights into disease mechanisms, efforts to identify a screening tool which detects non alcoholic steatohepatitis (NASH) with high degrees of sensitivity and specificity have been unsuccessful1 (2003;37:1286). In promotion of fatty change and in comparison with controls, NASH patients were shown to have higher ratios of saturated: polyunsaturated fats in their diet, impaired very low density lipoprotein (VLDL) secretion following a fatty meal2 (2003;37:909), and reduced synthetic rates of apolipoprotein B-100, the key apoprotein for export of lipid as VLDL from the liver3 (2002;35:898). Activity of CYP2E1, a centrizonal cytochrome with particular propensity for promoting lipid peroxidation via free radical formation, was increased in NASH patients4 (2003;37:544); its activity was correlated to body mass index (BMI) and to the degree of steatosis on liver biopsy5 (2003;38:428). Another major source of oxidative stress in NASH was demonstrated on liver biopsy as reduced activity in each component of the mitochondrial respiratory chain (MRC) complex. Statistically significant correlations were shown between the reduced function of MRC complexes, BMI, serum tumour necrosis factor α (TNFα) and insulin resistance expressed as HOMAIR6 (2003;38:999). Several studies confirmed that insulin resistance is a key factor in the aetiology of NAFLD7,8(2002;35:373, 2002;35:497). Splanchnic adipocytes have long been known to be several-fold more resistant to insulin than their peripheral counterparts. One of the new understandings to emerge is that these splanchnic fat storage cells constitute a major endocrine influence on the liver and that imbalance of their hormonal products, “adipokines” (adipocyte cytokines) influences hepatic fat deposition, insulin resistance and progression to NASH. The adipokines include TNFα, leptin, resistin and adiponectin. TNFα is a proinflammatory cytokine with profound influence in promoting insulin resistance, inflammation and fibrosis in the liver. Hepatic steatohepatitis is promoted when TNFα is increased in proportion to leptin and adiponectin whose actions it antagonises. Serum concentrations of TNFα, TNF recetor 2, a marker of activation of the TNFα system, and leptin were increased in NAFLD with the highest levels being found in NASH9 (2002;36:403). It remains sub judice whether in such patients the hyperleptinemia is adverse or beneficial in its net effect on hepatic steatohepatitis. Leptin lowers hepatic expression of sterol regulatory element-binding protein 1 (SREBP-1) which upregulates enzymes involved in fatty acid oxidation and downregulates lipogenesis. In patients with NASH and lipodystrophy, an experiment of nature in which adipocytes are deficient, a dramatic reduction in steatosis and hepatocyte ballooning was observed in paired liver biopsies following correction of leptin deficiency by replacement therapy with recombinant methionyl human leptin10 (2005;41:753). Recent articles in Hepatology have highlighted the inverse relationship of TNFα levels to those of adiponectin, the new kid on the block. The severity of necroinflammation and fibrosis in NASH was associated with reduced serum adiponectin levels11 (2004;40:46) and following a fatty test meal NASH subjects experienced enhanced post-prandial lipemia in association with a further fall in adiponectin which was contrary to the anticipated rise seen in controls12 (2005;42:1175).

The role of insulin resistance in promotion of NASH provided the rationale for two controlled clinical trials of its treatment with the thiazolidinedione derivatives rosiglitazone13 (2003;38:1008) and pioglitazone14 (2004;39:188). Both studies reported significant histological improvement but with two drawbacks. A proportion of patients experienced weight gain, a no-no in these patients, and benefits were poorly sustained following cessation of therapy.

There may be culture shock that among the high science of modern pharmacy the gold medal for correcting the insulin resistance, cytokine imbalance, steatosis, lipid peroxidation and stellate cell activation of NASH goes to stomach stapling (laparoscopic adjustable gastric banding)15 (2004;39:1647). Weight loss was sustained and accompanied by complete resolution of the histological changes of NASH in a majority of patients. This is evidence we can quote to our patients — that in NASH, as in the alcoholic, the steatohepatitis and steatofibrosis is a direct consequence of excessive oral consumption and that in only a disadvantaged minority such as those with partial lipodystrophy will this liver pathology persist and progress in the face of strict moderation of intake.

The Rise and Fall of WHVP —Utility and/or Futility.

Measurement of wedged hepatic venous pressure gradient (WHVPG) provides a measure of hepatic resistance to portal vein inflow, a key factor generating portal hypertension. Numerous studies have used this clinical measurement to increase our understanding of the pathogenesis of portal hypertension and its complications. Increased hepatic resistance to portal blood flow stimulates reflexes designed to overcome it as dictated by Ohms' law. Increased portal pressure is achieved via renal reflexes which expand the circulatory volume and adjustments in vascular tone which favour the splanchnic circulation. Elegant experiments by Ming et al.16 (2002;35:167) demonstrate that a neural reflex is triggered from the space of Mall where hepatic blood flow is autoregulated via reciprocal adjustments to hepatic arterial and portal venous supply. Evidence that the renal response to reduced portal inflow is stimulated by accumulation of adenosine in the space of Mall was provided by demonstration that the normal renal response of salt and water retention could be abolished by denervation of hepatic afferent or renal efferent nerve fibres or alternatively by portal infusion of an adenosine antagonist.

A high WHVPG has prognostic significance and was found to have an independent effect on survival when added to the model for end-stage liver disease (MELD) score but not to such an extent that would justify its use in that context17 (2005;42:793). When patients with cirrhosis and tense ascites were studied prospectively those who progressed to the hepatorenal syndrome (HRS) over the subsequent 2 years were proven to have had a higher WHVPG and higher plasma renin levels at baseline18 (2005;42:439). As HRS progressed the WHVPG rose precipitously, a sign from the liver of desperate thirst for improved portal inflow. Although escalation of portal pressure may, at first, compensate for progressive hepatic resistance, sooner or later the mechanisms by which this is achieved become futile and even counter productive. Thus the renal vasoconstrictor response when pushed towards its limits results in the HRS. This report emphasized the importance of a depressed cardiac response in the pathogenesis of HRS.18

Other studies have further challenged us to consider inclusion of WHVPG measurement as a tool to optimise our treatment of patients with bleeding oesophageal varices. Failure to reduce the WHVPG to <12mm Hg or by less than 20% of baseline values was predictive of rebleeding and this prognostically important identification of non-responders to β-blockade could not be derived from any other measure such as reduction of pulse rate19 (2002;36:1361). In a headline deserving report, a WHVPG measurement of >20mm Hg was used to define a high risk group of patients with bleeding varices who were randomly allocated to standard therapy or urgent trans jugular intrahepatic portal-systemic shunt (TIPS); urgent TIPS resulted in a significant reduction in 1 month (11% vs. 38%) and 1 year (31% vs. 65%) mortality20 (2004;40:793). The hepatological fraternity may turn a blind eye to these findings because of the seemingly unattainable resource allocation that would be required to enable such management to be provided routinely. It seems to me that as a minimum we should mount a further trial that would definitively confirm or refute these impressive results.

The Brain's Highs and Lows in Acute and Chronic Liver Failure

Safeguarding cerebral function continues to challenge the hepatologist managing patients with both chronic and acute liver failure (ALF). In ALF raised intracranial pressure (ICP) remains a major cause of morbidity and mortality and Hepatology papers have highlighted the interrelatedness of increased cerebral blood flow and cerebral edema in its causation. Terlipressin is a pressor agent widely used internationally in the treatment of bleeding varices and hepatorenal syndrome. On a cautionary note, its administration in ALF was found to produce a significant increase in ICP attributed to increased cerebral blood flow despite lack of any detectable effect on the systemic circulation21 (2004;39:471).

Hyperammonemia-induced accumulation of glutamine/glutamate in the brain remains a focus of research in acute and chronic liver failure. Following an amino acid meal designed to replicate a gastrointestinal bleed, hyperammonemia was shown to coincide with an increase the brain glutamine: creatine ratio on MR spectroscopy22 (2003;37:931). It may surprise the reader that the source of hyperammonemia was shown to be the kidney (not the gut) and for its removal skeletal muscle was more important than liver. Under baseline conditions renal ammonia production is largely glutamine based but following the simulated bleed alanine uptake was selectively increased in the kidney where alanine is a substrate for renal gluconeogenesis and coincident ammoniagenesis. The cirrhotic liver's maximal potential for ammonia removal was no more than equivalent to the functioning of skeletal muscle from a single leg23 (2003;37:1277).

Semipermeable dialysis of the subdural space in ALF facilitated dynamic monitoring of cerebral metabolism with results from measurement of glucose, lactate, pyruvate and glutamate concentrations being made available within 20 minutes. Patients who developed raised ICP had higher initial glutamate levels but no increase that was temporally related to surges of ICP. Rather these surges coincided with sharp increases in lactate concentrations in the absence of cerebral hypoxia. It was hypothesized that the lactate induced an abrupt rise of ICP via a vasodilatory effect on the cerebral vasculature24 (2002;36:1333). In a rat model in which brain lactate and glutamine were monitored by nuclear magnetic spectroscopy it was similarly found that raised ICP was associated with a continuing increase of brain lactate but not of glutamine25(2003;37:420). Since ammonia is known to inhibit α-ketoglutarate dehydrogenase, a rate-limiting tricarboxylic acid (TCA) cycle enzyme both publications propose that increased glycolysis and impaired function of the TCA cycle account for the lactate excess.

In both acute and chronic liver disease evidence has been presented that hyponatremia has deleterious effects upon the brain. In stable patients with cirrhosis studied by magnetic resonance spectroscopy serum sodium concentration correlated with reduced brain concentrations of myo-inositol and creatine/phosphocreatine. It was proposed that these osmolytes leak from the brain as an adaptive osmoregulatory response that compensates for the increased concentrations of glutamine/glutamate which, left unbalanced, would result in cerebral edema26 (2004; 39:1613). Murphy and colleagues prevented hyponatremia in ALF by infusion of hypertonic saline (30%w/v) to maintain serum sodium concentrations at 145-155mmol/l and observed reduced ICP in comparison with control subjects27 (2004;39:464). The adrenal response to synacthen stimulation was found to be subnormal in patients with ALF who required inotropic support suggesting that a relative adrenal insufficiency was contributing to the severity of the illness and that supraphysiologic doses of hydrocortisone which can augment the pressor response to noradrenaline should be considered in these patients28 (2002;36:395).