Hepatology highlights


  • Potential conflict of interest: Nothing to report.

Can Cirrhosis Resolve?

Of all liver diseases, hemochromatosis has provided the most anecdotal evidence that cirrhosis may be a reversible condition. A clear etiology and curative treatment provide the ideal test case of the potential for fibrosis to resolve and normal architecture to be restored. Falize et al. addressed this issue in their series of patients with biopsy-proven hemochromatosis (all C282Y homozygotes) who were successfully venesected. A second liver biopsy was available at least 2 years after completion of the initial course of phlebotomy as indicated by serum ferritin <50 ng/mL. In 36 patients paired liver biopsies, taken 3-28 years apart, were deemed adequate (≥1 cm in length and 6 portal tracts) and reviewed by 2 pathologists. Fibrosis was staged as severe (F3) or cirrhosis (F4) according to the METAVIR system. The 2 pathologists provided concordant grading on 83% of the initial biopsies but only 42% of second biopsies, conferring later before coming to an agreement. Fibrosis decreased 85% with stage F3 initially (n = 13) and 44% with initial cirrhosis (n = 23). Of those with F4, 1 resolved to stage 0, 4 to stage 1, and 3 to stage 2. Overall fibrosis regressed by at least 2 METAVIR units in 47% of the sample. The goal of preventing or reversing cirrhosis would be anticipated to diminish many complications including portal hypertension and hepatocellular carcinoma. In this series, all 3 patients with hepatocellular carcinoma and 1 of 2 with cholangiocarcinoma showed no evidence of fibrosis regression. Lack of regression was also associated with higher serum gammaglobulin, lower platelet counts, and prolonged prothrombin time at diagnosis. One is left with the impression that F4 with METAVIR comprises a mixed group including those in whom fibrosis may indeed regress leaving a normal architecture, but also those with more advanced disease at diagnosis in whom damage has destroyed the normal architecture and passed the point of no return. (See HEPATOLOGY 2006;44:472-477.)

When Is a Problem Not a Problem?

It seems curious that statins continue to attract such attention to liver function testing when many other drugs which have a vastly greater potential for hepatotoxicity are not subjected to any routine surveillance. Given that patients with the metabolic syndrome have a common predisposition to atheroma and nonalcoholic fatty liver disease (NAFLD) the scares associated with abnormal LFTs in patients taking statins has probably resulted in many being unnecessarily denied the benefits of cholesterol reduction. Concerns about a direct drug toxicity are accompanied by theoretical concerns that statin usage in patients with NAFLD might aggravate steatosis via both increased de novo lipogenesis and increased uptake of lipid from plasma due to induced expression of LDL receptors. In the Dallas Heart Study reported by Browning, the hepatic triglyceride content (HTGC) of 2,287 subjects was measured by proton magnetic resonance spectroscopy. The prevalence of hepatic steatosis, defined as HTGC > 5.5% ranged from 24% with isolated hypercholesterolemia to 83% in those with mixed hyperlipidemia combined with elevation of serum ALT (see Fig.). In a comparison between 140 subjects taking a statin as their sole lipid-lowering agent and paired controls matched for ethnicity, gender, age, waist circumference, and insulin resistance there was no evidence of statin-induced hepatotoxicity in terms of ALT, HTGC, or any other measure. Those on statins stood to benefit from the reduced risk of cardiovascular events conferred by having significantly lower levels of LDL cholesterol in plasma. (See HEPATOLOGY 2006;44:466-471.)

Illustration 1.

Should Complicated Liver Surgery Be Restricted to High-Throughput Units?

In Britain, our surgical colleagues have somewhat reluctantly subscribed to the practice of restricting complicated operations with high morbidity and mortality to units where case throughput is sufficient to maintain optimal skills and favorable outcomes. Accordingly, surgery for biliary atresia in the form of Kasai operations and liver transplantation are only available at 3 designated centers. The French, as ever, have asserted their independence and entrusted progress to careful audit and feedback of information to any unit providing such a service. Serinet et al. report on the survival of two cohorts of children with biliary atresia born between 1986-1996 (cohort A; n = 472) and 1997-2002 (cohort B; n = 271). Reduction in actual mortality from 30.5% in cohort A to 12.2% in cohort B applied across all phases of management, whether no surgery, post-Kasai, or post-transplant. Four-year actual survival for cohorts A and B, respectively, was 98% versus 100% for successful Kasai, 65% versus 79% for failed Kasai, and 50% versus 93% for those without Kasai. These dramatic improvements are a tribute to improved collaboration and feedback from audit. Without compulsion, there has been voluntary centralization of activity in France with a reduction from 30 to 22 centers providing the Kasai operation and a reduction from 35% to 8% in the proportion of liver transplants performed in centers performing ≤5 pediatric liver transplants per year. Centers performing ≤2 Kasai operations per year achieved no improvement of infraoptimal results in successive cohorts. Vive la différence? (See HEPATOLOGY 2006;44:75-84.)

Some Are Very Good and Some Not So Good But Should We Care?

Polymorphic variants of a host of liver functions are described, but for the vast majority they have no known disease association. For others, risk is associated with nonphysiological challenge as with idiosyncratic drug reactions. Meier et al. analyzed the concentration of four canalicular bile transporters in normal liver removed during partial hepatectomy in 110 white Caucasian patients. Expression of the bile salt export pump (BSEP, ABCB11), multidrug resistance-associated protein 2 (MRP2,ABCC2), multidrug resistance protein 3 (MDR3, ABCB4), and multidrug resistance protein 1 (MDR1, ABCB1) all showed significant interindividual variation (see Fig.) with a unimodal distribution. The significant within-patient similarity in expression of transporters suggests a common mode of regulation such as that known to occur via the farnesoid X receptor (FXR). No correlation was found in this cohort between transporter expression and clinical data such as markers of cholestasis or current medication. Nevertheless, the finding of low or very low levels of expression of these transporters in 1%-15% of the study sample gives credence to the hypothesis that polymorphic variants which impair the functioning of these transporters may contribute to variable susceptibility to disease between individuals. The authors also describe very high levels of expression of MRP2 which, somewhat perversely, given its ability to excrete estrogen and progesterone metabolites into bile from where they are able to trans-inhibit BSEP, may conceivably be associated with cholestasis of pregnancy. A restricted number of single-nucleotide polymorphisms (SNPs) were shown to determine most very high and very low results, making screening for SNPs a practical option for larger population studies in the future. (See HEPATOLOGY 2006;44:62-74.)

Illustration 2.

It's Tough Sticking to Your Job as a Platelet When There's Cirrhosis About

Reduced platelet numbers and reduced platelet function may contribute to bleeding in patients with cirrhosis, but Lisman et al. show that compensatory mechanisms are also active, including elevated levels of von Willebrand factor (VWF). VWF is a large multimeric protein that responds to damage to vessel endothelium by an initial interaction with exposed collagen, which then enables it to interact with platelet glycoprotein 1b to slow down the platelet and thus facilitate direct bonding between platelet and collagen. Lisman et al. found that VWF levels were increased by a median of 380%, 500%, and 760% in plasma from patients with Child's A, B, and C cirrhosis, respectively, when compared to healthy controls. The ability of VWF to interact with platelet glycoprotein 1b was measured as VWF ristocetin cofactor activity and was shown to be diminished in terms of its specific activity (see Fig.). Similarly, the collagen-binding capacity of VWF was reduced in cirrhosis possibly due to a reduction of multimeric aggregates of VWF which most potently support hemostasis by virtue of having the highest affinity for both platelet glycoprotein 1b and collagen. Despite the diminished activity of VWF in these two functional assays, any deficiency should be more than compensated for by the increased concentrations of VWF found in plasma from patients with both cirrhosis and acute liver failure. The ability of platelets from cirrhotic patients and healthy controls to adhere to collagen type III was tested in an in vitro assay. Plasma from patients with cirrhosis was significantly better than normal plasma at supporting platelet adhesion in this assay, and platelets from patients with cirrhosis functioned just as well as platelets from normal controls. Various questions arise regarding the mechanisms by which plasma concentrations of VWF are increased but whatever those influences, this study shows that highly elevated levels of VWF which occur in patients with cirrhosis compensate for its reduced functional properties and will attenuate the effects of reduced platelet numbers and activity. (See HEPATOLOGY 2006;44:53-61.)

Illustration 3.

In addition to reduced platelet numbers and function, coagulation defects associated with cirrhosis include defects in fibrin formation, fibrinolysis, and thrombin generation. Thrombin generation had until recently been assayed only in platelet free plasma, but use of fluorophore substrates has enabled Tripodi et al. to simulate in vivo thrombin generation in platelet-rich plasma. When platelet numbers were adjusted to 100 × 109/L, thrombin generation was found to be normal but when platelet concentrations were unadjusted so as to represent the patients' thrombocytopenic condition, thrombin generation was significantly reduced. The authors conclude that deficient thrombin generation in patients with cirrhosis is primarily a consequence of thrombocytopenia and propose that in patients with cirrhosis and thrombocytopenia who suffer from a troublesome bleeding tendency, trials should be conducted of different methods which elevate the platelet count. (See HEPATOLOGY 2006;44:440-445.)