Hepatology highlights


Combination Therapy for Primary Sclerosing Cholangitis — Modest Improvements

Primary sclerosing cholangitis (PSC) occurs in association with inflammatory bowel disease in 80% of cases. The disease has proved resistant to treatment with a number of drugs, mainly immunomodulatory agents, as noted in clinical trials. Urso-deoxycholic acid has been promising, markedly improving serum liver chemistries on its introduction but, with only anecdotal exceptions, failing to improve histological or cholangiographic outcomes. Färkkilä et al. report the results of a 36 month trial in which 80 PSC patients were randomized to treatment with either ursodeoxycholic acid in combination with metronidazole (UDCA/MTZ) or with ursodeoxycholic acid and placebo(UDCA/placebo). Both groups responded to URSO as previously described and overall there was a trend toward greater improvement with combination therapy. Yet again, there was no difference in major end-points, 3 patients in each group leaving the study due to transplantation or cholangiocarcinoma. Serum alkaline phosphatase fell to a significantly greater degree with UDCA/MTZ (P < .05) and the New Mayo Risk Score (computed from age, serum bilirubin, albumin, aspartate aminotransferase and variceal bleeding) was significantly reduced only in the group on combination therapy. Although the number of patients who had an improvement on histological grading (44% vs. 17%) and staging (34% vs. 14%) was greater in the UDCA/MTZ group, there was no statistically significant difference between the two groups in the overall histological assessment. Similarly, comparison of before and after ERCP images showed a nonsignificant trend toward benefit with URSO/MTZ; cholangiographic findings worsened in 17% compared to 32% in the UDCA/placebo group. (See HEPATOLOGY 2004;40:1379–1386.)

Gilbert's Good News

Unconjugated hyperbilirubinemia, in the absence of any other abnormality of liver function, is found in 3% to 10% of the general population in various series. In most cases this is attributable to Gilbert's syndrome caused by deficient functioning of the enzyme UDP-glucuronosyltransferase. In the study of Zucker et al. serum bilirubin was analyzed on 16,665 individuals who were selected to provide representative data for the entire US population during the third National Health and Nutrition Examination Survey (NHANES III). According to data obtained from patient-based questionnaire responses, serum bilirubin concentrations were significantly lower in those with a history of previous gastrointestinal (GI) malignancies (0.52 ± 0.017 mg/dL) compared with those with no history of malignancy (0.62 ± 0.003 mg/dL; P = .01). Those giving a history of previous non-GI malignancies had intermediate values. Accordingly, a 1.0-mg/dL increase in serum bilirubin is associated with an almost 4-fold lower prevalence of colorectal cancer (OR, 0.257; CI, 0.254-0.260). The authors postulate that the antioxidant activity of bilirubin may account for this observation. Data are cited that demonstrate that unconjugated bilirubin is present in the colonic lumen at concentrations which experimentally are known to decrease colon cancer cell survival by inducing apoptosis. Intracellular redox cycling of bilirubin to biliverdin catalysed by biliverdin reductase is known to have cytoprotective action against oxidative stress. The article also documents differences between sexes and different racial groups that provide a series of reference values for fasting serum bilirubin levels. (See HEPATOLOGY 2004;40:827–835.)

Antiandrogens of No Benefit (Possibly Harmful) in Hepatocellular Carcinoma (HCC)

The trial reported by Trinchet at al. was justified by the many clinical and experimental observations that suggested that testosterone promotes primary liver cancer. For example, HCC occurs with a male-female ratio of between 5:1 and 9:1 and has a worse outcome in males, and serum testosterone has been shown to be a predictive factor of HCC occurrence in HCV-cirrhosis and HBsAg-positive carriers. Small nonrandomized trials had suggested benefit from antiandrogenic blockade. Despite these persuasive pointers, the trial showed that there was no benefit on survival from antiandrogenic treatment. Tamoxifen, an agent with recent evidence suggesting that it is neither beneficial nor harmful in HCC, was given to both groups, making it a specifically antiandrogen trial. Effective treatment was ensured by combining leuproprelin, an LH-RH agonist, and flutamide, a nonsteroid peripheral androgen receptor antagonist, in doses that are effective for prostatic cancer. Rather than benefit, there was a distinct trend toward the converse outcome in that those who survived more than 3 months on antiandrogenic treatment had a tendency for shorter survival and significantly more rapid tumor growth as assessed by tumor size and alpha-fetoprotein rise. This counterintuitive finding reinforces the merits of adequately powered randomized trials. Mechanisms postulated to account for enhancement of tumor growth in response to antiandrogenic treatment include a switch from antagonist to agonist effect of antiandrogens in advanced tumors, as has been described with prostatic cancer. Alternatively, a deleterious effect which promotes apoptosis in the nontumor liver with cirrhosis but favors growth of androgen-insensitive tumor cells is also possible. (See HEPATOLOGY 2004;40:1361–1369.)

Curing Hepatocellular Carcinoma (HCC) by Percutaneous Ablation: Criteria Predicting Success

A 15-year experience with the treatment of inoperable hepatocellular carcinoma using percutaneous ablation is reported by Sala et al. Both surgical criteria for resectability and methods of ablation evolved during this time and in the last years, ablation was combined with transarterial chemoembolization (TAE). Despite this heterogeneity of case-mix, significant findings emerge to inform future studies and current practice. Patients treated with percutaneous ethanol injection (PEI) or radiofrequency ablation (RF) were assessed for residual viable tumor by CT scan 1 month after completion of the first treatment cycle. Absence on CT of enhanced tumoral areas was taken to represent initial complete response (CR) with complete tumor necrosis and was seen in 73% with PEI, 65% with RF, and 57% with TAE-PEI. After a median follow-up period of 25 months, Kaplan-Meier curves show overall survival at 1, 3, and 5 years of 87%, 51%, and 27%, respectively, with a median survival of 36 months. The most favorable outcome was seen in Child-Turcotte-Pugh Class A patients who had a single tumor < 2 cm in diameter, in whom initial CR had been documented who achieved a 5-year survival rate of 63%. The authors concede that the accuracy of imaging in the definition of initial CR is currently not sufficiently precise to suggest that patients who are eligible for surgical resection should be counselled to view percutaneous ablation as a rival alternative approach. Nevertheless, the authors view the results as sufficiently encouraging to rule out on ethical grounds a randomized trial comparing ablative therapy with no treatment for HCC in patients who are ineligible for surgical resection. (See HEPATOLOGY 2004;40:1352–1360.)

Modifying MELD to Reduce Mortality on the Waiting List

The Model for End-stage Liver Disease (MELD) scores are used to prioritize patients for allocation of donor livers on the basis of “sickest first.” Inclusion of serum bilirubin along with serum creatinine and international normalized ratio (INR) as the components of MELD has a tendency to weigh organ allocation in favor of cholestatic liver disease. Heuman et al. investigated the utility of MELD in predicting death while on the waiting list for liver transplantation in successive cohorts with largely noncholestatic liver disease who had been listed for transplantation within the US Department of Veterans Affairs Health System. Analysis of 296 patients (training group) listed before adoption of MELD in February 2002 revealed that 61 had died of cirrhosis within 180 days of referral when only 18 (4%) had been transplanted. On multivariate analysis only MELD score, persistent ascites and low serum sodium (<135 meq/L) emerged as independent predictors of death within 180 days. For MELD scores higher than 21 only MELD was an independent predictor, whereas for MELD scores lower than 21 persistent ascites and low serum sodium were the sole predictors. A subsequent cohort of 211 patients referred in the MELD era (validation group) confirmed these findings but with lower accuracy as predictors of pretransplant death, which may be explained by improved prioritization of recipients due to MELD-based organ allocation. With the weighting coefficient of MELD normalized to 1, logistic regression assigned coefficients of 4.46 and 4.53 to persistent ascites and low serum sodium (<135 meq/L) respectively; the authors propose that these coefficients be added to MELD in a modified score termed “MELD-AS” (MELD-ascites-sodium). In their patients with predominantly noncholestatic cirrhosis, Heuman et al. found that the predictive accuracy of the MELD-AS logistic regression model outperformed the MELD score overall, with significantly greater accuracy in those with MELD scores lower than 21 (P < .05) and with accuracy comparable to MELD in those with scores higher than 21. (See HEPATOLOGY 2004;40:802–810.)

Illustration 1.

HVPG Tips the Balance

Acute bleeding from esophageal varices continues to have a high mortality, especially in those without control of the acute bleeding episode and/or early rebleeding, factors that define treatment failure (TF). It has been reported that a high hepatic venous pressure gradient (HVPG) is predictive of TF. Monescillo et al. measured the HVPG in 116 patients shortly after endoscopic sclerotherapy and within 24 hours of their admission with variceal bleeding. HVPG of 20 mm Hg or greater defined a high-risk (HR) group, which had a higher incidence of active bleeding at urgent endoscopy, a greater transfusion requirement, a higher proportion of Child's class C patients, and a higher risk of TF (OR, 7.00; 95% CI, 2.40-20.36) than the low-risk (LR) group with HVPG less than 20 mm Hg. Consenting HR patients were randomized to early TIPS (HR-TIPS) or conventional management (HR–non-TIPS) and TIPS was performed 22.5 ± 20.4 hours after admission. Comparison of outcomes showed a lower rate of TF (12% vs. 50%), and reduced in-hospital (11% vs. 38%) and 1-year mortality (31% vs. 65%) in the HR-TIPS group when compared with HR–non-TIPS patients. Confirmation that the approach adopted by Monescillo and colleagues produces a significant improvement in survival rates from acute variceal bleeding would have enormous resource implications if provision of urgent access to endoscopic treatment, HVPG measurement, and early TIPS in HR patients were to become the norm. (See HEPATOLOGY 2004;40:793–801.)

Illustration 2.