Hepatology highlights

Two studies have examined the effect that therapeutic maneuvers have on intracranial pressure in patients with acute liver failure. Both cerebral edema and increased cerebral blood flow are major determinants of intracranial pressure, and herniation of the midbrain due to elevated intracranial pressure remains a common cause of death.

Hypertonic Saline Ameliorates Brain Swelling in Acute Liver Failure

Cerebral edema is one of the most serious and sinister complications of acute liver failure (ALF). Water movement across the blood–brain barrier is more dependent on diffusion across osmotic gradients than on pressure filtration which is the main determinant in organs where endothelia are fenestrated. Consequently, much experimental work has focused on the contribution of intracerebral osmolytes—including glutamine—to brain swelling.

It has been reported that hypertonic saline successfully reduces intracranial pressure (ICP) following traumatic brain injury. Murphy and colleagues report on a randomized, controlled trial in which the incidence of elevated ICP was compared between groups of ALF patients randomized to receiving conventional treatment versus similar treatment with the addition of intravenous infusion of hypertonic saline (30% w/v NaCl) to maintain serum sodium concentrations at 145 to 155 mmol/L. The potentially deleterious effect on cerebral edema of allowing hyponatremia to develop is illustrated by the negative correlation between initial ICP and serum sodium concentration at presentation to the intensive care unit. Hyponatremia was corrected such that the two groups had similar values prior to entry into the trial. Continuous veno-venous hemofiltration with a median of 4 L/h was also employed in 27 of the 30 cases studied, and the regimens should not be reproduced without close attention to this context. The treatment group (n = 15) received an average of 780 mmol of sodium chloride during the first 24 hours, achieving a mean peak serum sodium concentration of 153 mmol/L. In the treatment group over the first 24 hours, ICP was reduced significantly compared with admission levels but tended to rise (albeit insignificantly) in controls. At 42 hours, ICP in the treatment group became significantly less than in the controls (P = .04; see figure). The primary end point of a rise in ICP to above 25 mm Hg for 10 minutes or longer was seen in seven of the control group and three of the treatment group (NS). Many cases were censored because of death, transplantation, or premature removal of ICP monitoring devices, and no conclusion can be drawn regarding the effect of treatment on survival. The authors discuss several mechanisms by which hypertonic saline may diminish the risk of intracranial hypertension, including reduction of brain swelling through an osmotic effect (as occurs with mannitol administration) and improvement of cerebral perfusion pressure due to a dehydrating effect on swollen vascular endothelial cells. Similar studies are required to address the challenge of maintaining homeostasis of the chemical and physical environment of the brain during ALF. (See HEPATOLOGY 2004;39:464–470.)

Terlipressin Raises ICP in ALF

In many countries, intravenous bolus terlipressin is used with obvious benefit for treatment of the hepatorenal syndrome in patients with chronic liver disease. Under those circumstances the aim is to constrict the splanchnic circulation and improve renal perfusion through improvement systemic hemodynamics. In a recent study by Shawcross and colleagues, terlipressin was administered to six patients with ALF in a single small bolus dose that had no detectable effect on the systemic circulation. Nevertheless, ICP increased significantly because of a significant rise in cerebral blood flow. The concern thus raised by Shawcross and colleagues illustrates the fundamentally different pathophysiologies of acute and chronic liver failure and the pitfalls of failing to distinguish between them in therapy. (See HEPATOLOGY 2004;39:471–475.)

Nonalcoholic Steatohepatitis: Whom to Biopsy and How to Treat?

Clinical research in nonalcoholic steatohepatitis (NASH) has reached an interim phase wherein many of the issues have been clarified and clear pointers have been provided. The two major clinical dilemmas for the clinician confronted by nonalcoholic fatty liver disease (NAFLD) remain “Should we biopsy?” and, when the biopsy is diagnostic of NASH, “How to treat?”

The aim of biopsy is mainly to detect steatohepatitis and cirrhosis-threatening fibrosis. Bugianesi and colleagues report on a large Italian study in which 263 patients were enrolled with NAFLD but no overt symptoms of diabetes. Univariate analysis identified older age, female gender, obesity, increased aspartate aminotransferase–alanine aminotransferase ratio, higher serum ferritin level, increased fasting glucose, and decreased insulin sensitivity as predictors of severe fibrosis; the presence of these factors strengthens the case for biopsy. The importance of insulin resistance in the cause of NASH is once more confirmed by the finding that reduced oral glucose insulin sensitivity (OR = 0.53; CI = 0.33–0.87; P = .0113) was an independent predictor of severe fibrosis. The hyperferritinemia should not be interpreted as an indicator of iron overload, which in the past has been proposed as a common cofactor in the pathogenesis of NASH. There was no increased prevalence of genetic hemochromatosis (C282Y or H69D) in the NAFLD cohort above that found in the background population. Histologic evidence of hepatic iron overload was present in only 9%, but hyperferritinemia occurred in 21% and constituted an independent predictor of severe fibrosis (OR = 1.77; 95% CI = 1.21–2.58; P = .0032); therefore, in this context hyperferritinemia appears to be an index of hepatic injury rather than of iron stores. (See HEPATOLOGY 2004;39:179–187.)

Consolidation of the view that insulin resistance plays a key role is provided by another report by Promrat and colleagues of improvement in NASH during therapy to enhance sensitivity to insulin, this time with pioglitazone. As highlighted in the December 2003 issue of HEPATOLOGY in a closely parallel study with Rosiglitazone (HEPATOLOGY 2003:38:1008–1017), these agents improve serum transaminases, hepatic steatosis, and histologic features of steatohepatitis, presumably as a direct consequence of improving sensitivity to insulin. Both trials of treatment of NASH with these thiazolidinedione derivatives reported that the main adverse event was weight gain. (See HEPATOLOGY 2004;39:188–196.)

An important negative result is reported by Lindor and colleagues, who found no benefit in 166 patients with biopsy-proven NASH randomized to 2 years' treatment with ursodeoxycholic acid (13–15mg/kg/d) or placebo. (See HEPATOLOGY 2004;39:770–778.)

Study of Bile Canalicular Transporters Reveals No Genetic Defect in Primary Biliary Cirrhosis or Primary Sclerosing Cholangitis

The identification of a group of adenosine triphosphate–binding cassette transporters in the bile canalicular membrane and the recognition that inherited mutations of their encoding genes are responsible for a group of inherited progressive familial intrahepatic cholestatic (PFIC) diseases begged the question of whether or not they may have a role in the chronic cholestatic diseases primary biliary cirrhosis (PBC) and primary sclerosing cholangitis (PSC). This speculation was fueled by the realization that different point mutations in the same gene account for the phenotypically polarized progressive familial intrahepatic cholestasis of childhood and benign remitting recurrent intrahepatic cholestasis. Pauli-Magnus and colleagues report on their findings from sequencing the promoter, coding, and flanking intronic regions of genes encoding for the bile salt export pump and phospholipid flippase, mutations of which account for PFIC2 and PFIC3, respectively. One hundred and forty-nine healthy Caucasians acted as a control group and were compared with 76 PBC patients and 46 PSC patients. Although many variants were defined, there were no disease-specific differences in haplotype distribution or linkage disequilibrium, making it very unlikely that these transporters have any causal role in the initiation of these diseases. (See HEPATOLOGY 2004;39:779–791.)

Classification of Budd Chiari Syndrome According to a New Prognostic Index May Refine Selection Criteria for Anticoagulation and Porto-systemic Shunting

A report by Darwish Murad and colleagues combines data on a cohort of 237 Budd Chiari syndrome (BCS) patients diagnosed between 1984 and 2001 and treated at tertiary referral centers in France, Holland, and the United States. The rarity of the disease stimulated the international multicenter collaboration as a means of overcoming bias inherent in small series from single institutions. Despite the scatter of place and time, the cohort was sufficiently homogeneous to pool data. The study includes cases previously studied by the group from Clichy (France) that defined a prognostic index (PI) based on responsiveness of ascites to diuretics, the Child-Pugh score, age, and serum creatinine. The need to determine responsiveness to diuretics necessarily delayed calculation of that PI. In the current analysis with a stepwise backward elimination technique, encephalopathy (P < .001), ascites (P = .08), prothrombin time (P = .02), and serum bilirubin (P = .07) were shown to be independent prognostic markers for survival in the 205 patients in whom a complete dataset was available. This PI can be calculated without delay at the time of diagnosis. The new PI was used to stratify the cohort into three prognostic categories, class I (5-year survival 89%, CI = 79–99) and class III (5-year survival 42%; CI = 28–56) representing the upper and lower quartile (see figure). The authors went on to test for therapeutic benefit from anticoagulation and porto-systemic shunting (PSS) and found an insignificant trend toward benefit from anticoagulation only in class I patients. Porto-systemic shunts were performed in 117 patients (49%) for the indications of refractory ascites and/or deteriorating liver function. The interval from diagnosis of BCS to PSS varied from 0 to 132 months (median: 1 month); a delay of more than 6 months between diagnosis and PSS was associated with increased mortality (RR, 4.15; CI = 1.42–12.12; see table ). On time-dependent Cox analysis of the three different prognostic classes, only class II patients showed a tendency toward improved survival following PSS. (See HEPATOLOGY 2004;39:500–508.) 

Table  . Survival in Relation to PSS (n = 117), According to the Interval Between Diagnosis and Shunting Using a Cox Analysis Testing for Proportionality

Interval Between Diagnosis and PSS (mos.)	P Value	Relative Risk	95% CI
<1 (n = 63)	.83	1.07	0.57–2.00
1–6 (n = 39)	.002	3.05	1.53–6.09
>6 (n = 15)	.009	4.15	1.42–12.12