First-line therapies for H. pylori infection in Italy: a pooled-data analysis.

Background: Curing H. pylori infection remains challenging, and the use of most effective first-line therapy represents a therapeutic cornerstone. To monitor the efficacy of first-line therapies in Italy, we designed a systematic review with pooled- data analysis of data published in the last 15 years. Methods: The search was focused on standard regimens and adult patients. Studies that included modified therapy regimens, pediatric patients, case series with less than 5 patients, and those in language other than English were excluded. Results: A total of 40 studies, with 74 therapeutic arms and 13,539 patients were evaluated. Among the 14-day triple therapies, the combination with proton pump inhibitor (PPI), clarithromycin and amoxicillin achieved the highest (77.9%) success rate, whilst the lowest success rate (62.7%) was observed following the 14-day PPI, clarithromycin and tinidazole regimen. The overall efficacy of triple therapies significantly decreased from 75.7% to 72.1% in the last decade. Sequential (88.3% on 3431 patients), concomitant (88.8% on 376 patients), and the bismuth-based quadruple therapy with three-in-one capsule, containing bismuth subcitrate potassium (140 mg), metronidazole (125 mg), tetracycline (125 mg) (90.4% on 999 patients) achieved similarly high eradication rates, but data on concomitant are still limited. The bismuth-based was associated with the higher (38.7%) incidence of side-effects. Conclusions: Data found that all triple therapies, irrespective of drug combination and therapy duration, should be abandoned in Italy due to their unacceptable low success rates. Monitoring the efficacy of standard first-line therapies in other countries could be clinically useful for both patients and clinicians.

Not all abolished lung sliding are pneumothorax: the case of a particular lung atelectasis.

INTRODUCTION: Lung ultrasound (LUS) is expanding from the field of emergency medicine, also to the pneumological specialist field, becoming part of the diagnostic procedure of lung consolidation. CASE PRESENTATION: A 78-year-old male was admitted to our emergency department for exertional dyspnea. LUS was performed, thus showing at right hemitorax air interface, A lines pattern, pleural sliding abolished on the whole hemitorax, thus suggesting a pneumothorax, but no evidence of lung point. A scan of lower lung segment showed an absence of the diaphragmatic excursion, suggestive for hemiparalysis of the diaphragm muscle, then confirmed by a subcostal scan. Moreover, at the lower segment of right hemitorax there was mild pleural effusion allowing the visualization of a round-shaped parenchymal consolidation with the absence of air bronchograms. CONCLUSIONS: LUS allowed the visualization of a particular and rare disease such as anthracosis-associated rounded atelectasis, thus leading to a more correct and faster patient management.

Serum autoantibodies in chronic hepatitis C: comparison with autoimmune hepatitis and impact on the disease profile.

Antibodies to nuclei (ANA), smooth muscle (SMA), and liver/kidney microsomes type 1 (anti-LKM1) may occur in chronic hepatitis C. Distinct subspecificities, including ANA with the homogeneous pattern (ANA-H) and SMA with antiactin specificity (SMA-AA), are found in autoimmune hepatitis (AIH). This study was performed to characterize the hepatitis C virus (HCV)-associated autoantibodies and to evaluate their influence on the profile of the disease. Two hundred ninety consecutive patients with chronic hepatitis C and 35 control cases with AIH were screened for autoantibodies by indirect immunofluorescence (IFL) at 1:40 serum dilution. The ANA pattern was defined by IFL on HEp-2 cells and the SMA-AA identified by the presence of at least two of the following elements: 1) SMA(T) or SMA(G) pattern by IFL on kidney sections; 2) XR1 precipitating system by counterimmunoelectrophoresis; or 3) typical pattern by IFL on liver sections from phalloidin-intoxicated rats. ANA, SMA, and anti-LKM1 occurred in 9%, 20%, and 6% of chronic hepatitis C cases, respectively. The overall prevalence of autoantibodies was 30% (87 of 290). Compared with AIH, HCV-associated ANA and SMA exhibited ANA-H and SMA-AA at a lower prevalence (38% vs. 71%, P = .04 and 8% vs. 87%, P < .000001, respectively) and had a lower median titer (1:80 vs. 1:320, P < .001 and 1:40 vs. 1:320, P < .000001, respectively). The concomitant positivity for ANA-H and SMA-AA was detected in none of the HCV cases, but in 46% of AIH sera (P < .000001). Two parameters were independently associated with the autoantibodies in chronic hepatitis C: high alanine transaminase (ALT) serum levels (F = 14.04) and female gender (F = 5.03). At the univariate analysis, patients with autoantibodies had a more severe portal-periportal necroinflammation (median Scheuer's score: 2.05 vs. 1.64, P = .003). The presence of autoantibodies did not influence the response to interferon (IFN). In chronic hepatitis C, serum autoantibodies are common, but their subspecificities are distinct from those occurring in AIH. Whereas the absence of ANA-H and/or SMA-AA does not exclude AIH, the characterization of ANA and SMA may help to discriminate between the two conditions. As compared with the seronegative counterpart, autoantibody-positive chronic hepatitis C is more common in females and exhibits a more severe biochemical and histological activity. The response to IFN therapy, however, is similar.

Clinical implications of GBV-C/HGV infection in patients with "HCV-related" chronic hepatitis.

BACKGROUND/AIMS: To evaluate the clinical, biochemical and histological implications of a concomitant HGV infection in "HCV-related" chronic liver disease. METHODS: Eighty-three HCV-RNA positive patients with chronic liver disease were tested for GBV-C/HGV coinfection by heminested PCR. RESULTS: Twenty-two (26.5%) patients were found to be positive for GBV-C/HGV RNA. GBV-C/HGV+ patients differed significantly from GBV-C/HGV- ones for younger age, higher frequency of history of drug addiction, which in turn might favor coinfection with interferon-sensitive HCV genotypes (3a), and increased probability of long-term response to interferon. GBV-C/HGV infection appears to have no responsibility for specific aspects of HCV infection such as biochemical or histological cholestatic features, lymphoid follicles, symptomatic cryoglobulinemia or presence of serum autoantibodies, including LKM1. It does not worsen the HCV-related disease (ALT levels and histological activity) and does not significantly interfere with HCV infection, as explored by the number of hepatocytes positive for HCV antigens. The amount of steatosis (mean score) was shown to be higher in GBV-C/HGV+ patients. A virological follow up was performed in 17 interferon-treated GBV-C/HGV+ patients On the whole, GBV-C/HGV seems to be as sensitive to IFN treatment as HCV, but recurrence after withdrawal is more frequent. In spite of this, ALT levels often remain normal after treatment withdrawal. CONCLUSIONS: The present data suggest that GBV-C/HGV infection, apart from more marked liver steatosis, does not modify the overall picture of chronic hepatitis due to HCV infection.

Quantitative liver parameters of HCV infection: relation to HCV genotypes, viremia and response to interferon treatment.

BACKGROUND/AIMS: This study aimed to evaluate the relation between the number of hepatocytes positive for HCV antigens and the amount of HCV RNA in the liver and to evaluate the relationship between the above parameters and viremia levels, HCV genotype and response to interferon treatment. METHODS: This was a retrospective study on 31 consecutive patients with chronic HCV-related liver disease, selected on the basis of the availability of frozen liver tissue for both liver HCV antigens detection and liver HCV RNA quantitation. HCV antigens (immunohistochemistry), liver and plasma HCV RNA (competitive RT-PCR), and HCV genotype (commercial kit) were studied. RESULTS: A significant correlation (p=0.0005) was found between the amount of liver HCV RNA (log 10 copy/microg of extracted RNA) and the number of HCV-infected hepatocytes (scored from 0 to 3). These parameters were not significantly correlated with viremia levels. The highest liver HCV RNA levels and HCV antigen scores were found in patients infected with genotype 1b. Liver HCV RNA (median 541 x 10(3) vs 118 x 10(3) copy number/microg, p=0.031) and liver HCV antigens (mean score 2.3 vs 1.3, p=0.018) but not plasma HCV RNA (median 14956 x 10(3) vs 2885 [correction of 2.885] x 10(3) copy number/ml, ns) were significantly higher in patients not responding to interferon treatment compared to responders. CONCLUSIONS: The tissue parameters tested in this study were significantly correlated, shared the same clinical implications and predicted short-term response to interferon treatment better than viremia levels. We suggest that these tests should be included in the study protocol of patients under evaluation for interferon treatment, basing the choice on local facilities.

Ultrasound-detected abdominal lymphadenopathy in chronic hepatitis C: high frequency and relationship with viremia.

BACKGROUND/AIMS: This study aimed to investigate the prevalence and significance of ultrasound-detected deep abdominal lymphadenopathy in chronic hepatitis due to C virus. METHODS: One hundred and thirty-four consecutive patients with various liver disorders were examined with portable real-time equipment. RESULTS: In 25 (19%), the procedure failed because of excessive meteorism. Deep nodes, mainly located in the hepato-duodenal ligament, were detected in 62 of the remaining 109 patients (57%), reaching the highest prevalences in primary biliary cirrhosis (5/7, 71%), chronic hepatitis C (44/66, 67%) and autoimmune hepatitis type 1 (2/3, 67%). For all patients, including those with liver diseases with multiple etiology, lymphadenopathy was more frequent in anti-HCV positive (51/81, 63%) than in negative cases (11/28, 39% p=0.02). In chronic hepatitis C, serum HCV RNA was detected by nested polymerase chain reaction in all 31 patients with, but in only 75% (12/16) of those without nodes (p=0.018). No other distinct clinical or laboratory feature was found in association with lymphadenopathy; in particular, its incidence was similar in cases with and without liver cirrhosis. CONCLUSIONS: Enlarged deep abdominal lymph nodes are frequently detected by ultrasound in patients with chronic hepatitis C. This feature may be of diagnostic utility, especially in early cases, when liver cirrhosis has not yet developed and therefore no other ultrasound sign of the underlying disease can be detected. Lymphadenopathy may be of biological significance, marking hepatitis C virus infection in a replicative, viremic stage. These observations support the existence of a close interaction between hepatitis C virus and the lymphatic system.

Quantitative analysis of hepatitis C virus activity in vivo in different groups of untreated patients.

Highly sensitive competitive PCR (cPCR) and competitive reverse transcription PCR (cRT-PCR) methodologies were recently developed and applied for quantifying viral DNA and RNA species (including HCV RNA) present in clinical samples at low concentration. In this study, we used cRT-PCR to compare the viral load of 118 untreated patients with HCV infection and different clinical conditions (80 patients with chronic hepatitis, 18 infected subjects with persistently normal ALT levels and various degrees of liver injury, 10 HCV infected subjects that tested positive for anti-LKM1 antibodies, and 10 patients with HCV infection and cryoglobulinemia). The results indicate that while great individual variability of HCV viremia is detectable even among patients with similar clinical conditions, the mean HCV RNA copy number in samples from patients with different clinical conditions was similar in all groups with the single exception of patients that tested positive for anti-liver-kidney microsomal auto-antibodies type 1 (anti-LKM1); interestingly, lower HCV viremia levels were revealed in these anti-LKM1-positive cases with liver disease of uncertain pathogenesis.

Low hepatitis C viremia levels in patients with anti-liver/kidney microsomal antibody type 1 positive chronic hepatitis.

BACKGROUND/AIMS: The majority of adult patients positive for anti-liver-kidney microsomal antibody are also positive for anti-hepatitis C virus and serum HCV RNA. In these patients the role played by hepatitis C virus infection in the progression of liver damage and its relationship with anti-liver-kidney microsomal antibody are, however, still a matter of debate. METHODS: To clarify this point we have compared hepatitis C viremia in sera from 31 hepatitis C virus-related chronic hepatitis patients positive for anti-liver-kidney microsomal antibody with that of 31 patients with hepatitis C virus-related chronic hepatitis without autoantibodies using a newly developed competitive reverse transcription-polymerase chain reaction technique. Reverse transcription-polymerase chain reaction was performed using a synthetic competitor of a length similar to that of wild template (71 bp vs 86 bp). RESULTS: The results obtained have been related to hepatitis C virus genotypes. Anti-liver-kidney microsomal antibody/anti-HCV positive patients show a median value of hepatitis C virus genome molecules (626829/ml, range 9780-25651424), significantly lower than anti-liver-kidney microsomal antibody negative/anti-HCV positive patients (10158314/ml, range 101822-67429974) (p < 0.001). No hepatitis C virus genotype was significantly associated with anti-liver-kidney microsomal antibody, although a predominance of genotype 1 (subtypes a and b) has been observed in these patients. CONCLUSIONS: Since a low hepatitis C viremia has been observed in anti-liver-kidney microsomal antibody positive patients with disease severity comparable to that of patients without autoantibodies, it is conceivable that in them autoimmune mechanisms may cooperate with viral infection in sustaining disease activity.

Impact of international autoimmune hepatitis group scoring system in definition of autoimmune hepatitis. An Italian experience.

We have reclassified 110 patients with autoantibody-positive cryptogenic chronic hepatitis according tot he aggregate scoring system proposed by the International Autoimmune Hepatitis Group for signs of hepatitis C virus (HCV) infection and the newly proposed terminology of "unclassified" chronic hepatitis. Anti-HCV and HCV viremia were assessed by second-generation assays and reverse transcription-polymerase chain reaction. Immunomorphological and immunochemical characterizations of antinuclear, smooth, muscle, liver-kidney microsomal type 1, and liver cytosol type 1 autoantibodies were also performed. All 45 anti-HCV negative patients fulfilled the score criteria for the diagnosis of "definite" or "probable" autoimmune hepatitis (AIH). Eight anti-HCV-positive cases reached the score of "probable" AIH, whereas the remaining 57 cases were diagnosed as unclassified chronic hepatitis. The scoring system allows the correct identification of all autoimmune cases without HCV infection. Autoimmune hepatitis runs a more severe disease course than unclassified chronic hepatitis, whose clinical and histological features are similar to those of autoantibody-negative chronic hepatitis C.

Viral load in samples from hepatitis C virus (HCV)-infected patients with various clinical conditions.

Molecular methods for the absolute quantitation of nucleic acids present in biological samples have recently been developed and applied in basic and in medical virology; these studies indicated that competitive polymerase chain reaction (PCR) and competitive reverse transcription PCR (cRT-PCR)-based methodologies are currently the methods of choice for quantifying DNA and RNA species present in clinical samples at low concentration. Recently, quantitative molecular techniques were developed to study the hepatitis C virus (HCV) pathogenic potential, the natural history of HCV-infected patients and the efficiency of antiviral therapies in real time. The pilot study reported here was carried out using a cRT-PCR application for the direct quantitation of HCV RNA molecules in plasma samples of infected individuals which was recently developed in our laboratory. Although sharp individual variability of viral load was documented in this study, the mean HCV RNA copy number detected in samples from untreated HCV-infected patients with various clinical conditions (chronic active hepatitis, cirrhosis, cryoglobulinaemia and chronic hepatitis) was substantially similar, with only one exception: in samples from patients tested positive for anti-liver-kidney microsomal (anti LKM1) auto-antibodies, a significantly lower HCV viraemia level was revealed. Additionally, HCV viraemia was monitored in four patients with sustained biochemical and histological response (at least 12 months) following interferon-alpha discontinuation.

Hepatitis C virus (HCV) genotype, tissue HCV antigens, hepatocellular expression of HLA-A,B,C, and intercellular adhesion-1 molecules. Clues to pathogenesis of hepatocellular damage and response to interferon treatment in patients with chronic hepatitis C.

To obtain information on the mechanisms of hepatocellular damage and the determinants of response to interferon, hepatitis C virus (HCV) genotype, tissue HCV antigens, hepatocellular expression of HLA-A,B,C and intercellular adhesion-1 molecules, and the number of lobular T lymphocytes were studied in 38 anti-HCV-positive patients. 14 patients did not show a primary response to interferon treatment. HCV genotype 1b was detected in 11 of them. They displayed higher scores of HCV-positive hepatocytes, HLA-A,B,C, and ICAM-1 molecules expression than with the responders. HCV-infected hepatocytes maintained the capacity to express HLA-A,B,C and ICAM-1 molecules. CD8-positive T cells in contact with infected hepatocytes and Councilman-like bodies were observed. A significant correlation was found between the number of lobular CD8-positive T cells and alanine amino transferase levels. No differences were observed in clinical, biochemical, and histological features between patients with high and low number of hepatocytes containing HCV antigens. These data suggest a prominent role of T cell-mediated cytotoxicity in the genesis of hepatocellular damage. The high expression of interferon-inducible antigens like HLA-A,B,C molecules suggests the presence of strong activation of the interferon system possibly related to high HCV replication in nonresponder patients infected with genotype 1b.

Hepatocellular codistribution of c100, c33, c22, and NS5 hepatitis C virus antigens detected by using immunopurified polyclonal spontaneous human antibodies.

Hepatitis C virus (HCV) antigens in liver biopsy have been detected by immunohistochemistry using both spontaneous human IgG and murine monoclonal or rabbit polyclonal monospecific reagents. Conflicting results have been obtained in different studies. This was probably because of the incapacity of single experimental antibodies, raised against synthetic or recombinant peptides, to recognize native tissue antigens. To overcome this possibility, we immunopurified monospecific spontaneous polyclonal human Ig, therefore induced by native antigens, from the single antigen-containing bands of RIBA 3 strips. Antibodies to c100, c33, c22, and NS5 antigens were obtained from the serum of a patient affected by chronic hepatitis C. The IgG fraction of this serum had proved to stain tissue HCV antigens. Eight biopsies were selected on the basis of strong hepatocellular reactivity with the whole IgG fraction in a variable number (from 5% to 75%) of cells. The four antigens were detected in all biopsies; a clear cellular codistribution was observed on serial sections. These data demonstrate that the possibility to identify HCV antigens in liver biopsies is higher when using human antibodies induced by native antigens rather than experimental antibodies. The approach of immunopurification of human antibodies can be extended to other HCV-related epitopes to obtain reagents useful for the selection and optimization of monoclonal or polyclonal antibodies.

Hepatocellular expression of HLA-A, B, C molecules predicts primary response to interferon in patients with chronic hepatitis C.

Primary response rate to alpha-interferon (IFN) is about 50% in patients with chronic hepatitis C. Criteria for predicting a positive primary response are lacking. HLA-A,B,C molecule expression is known to be stimulated by viral infections. In 36 consecutive interferon-treated anti-HCV positive patients with an available frozen liver biopsy sample, the predictive value of liver HLA-A,B,C expression, and of histologic, clinical, and biochemical parameters was evaluated. Response to treatment was defined by normalization of transaminases, and disappearance of serum HCV-RNA within 3 months. According to these criteria, 17 patients were classified as nonresponders and 19 were classified as responders. The pattern of HLA-A,B,C hepatocellular positivity varied from normal (negative or occasional faint staining of hepatocellular membranes) to diffuse, strong "honeycomb" positivity. The highest scores of positivity were found in nonresponder patients. The discriminant capacity of HLA-A,B,C scores of positivity was compared with clinical, biochemical and histologic parameters by discriminant analysis. HLA-A,B,C expression was found to be the main discriminant parameter, in addition to alkaline phosphate (ALP) and gamma-glutamyl-transpeptidase (GGT) which added little additional information. The higher hepatocellular expression of class I MHC molecules in nonresponder cases may reflect a different viral effect on hepatocytes, which is induced by different HCV genotypes or levels of viremia. From a clinical point of view, the pretreatment HLA-A,B,C pattern of positivity represents a powerful tool in the selection of patients for interferon treatment.

Analysis of the hepatitis C virus genome in patients with anti-LKM-1 autoantibodies.

Hepatitis C virus genotypes have been characterized in 22 patients with anti-LKM-1 positive chronic hepatitis C. Following the Simmonds classification, 77% of patients were infected by hepatitis C virus genotype 1, 18% by genotype 2 and 5% by genotype 3, thus excluding the association of the autoimmune reaction with a particular viral type. Prevalences of genotype 1 and 2 were significantly different from those obtained in 79 patients with chronic hepatitis C who were negative for anti-LKM-1, as these were more rarely infected by genotype 1 and more frequently by genotype 2. Clinical findings of anti-LKM-1 positive patients were similar in all three groups. Sequence analysis of the amplified 5'UTR provided identification of peculiar and identical nucleotide substitutions in two out of four patients with genotype 2. The analysis of the secondary structure of this region showed that the observed nucleotide mutations increased the stability of the stem formed in this position.

Interferon therapy in liver/kidney microsomal antibody type 1-positive patients with chronic hepatitis C.

The association between liver/kidney microsomal antibody type 1 and adult cases of hepatitis C virus-related chronic liver disease has been firmly established. In the presence of both markers, evidence of autoimmunity (liver/kidney microsomal antibody type 1) and actual viremia (serum HCV RNA), the therapeutic dilemma arises between steroids, which are beneficial to autoimmune but deleterious to viral diseases, and interferon-alpha, which may exacerbate an autoimmune disorder. Six patients with liver/kidney microsomal antibody type 1 and serum HCV RNA were given interferon-alpha: three showed a response pattern similar to that observed in autoantibody-negative chronic hepatitis C cases; the other three developed a sharp transaminase peak, which was not followed by HCV RNA clearance. Considering the brisk flare-up of liver cell necrosis, interferon-alpha treatment proved to be dangerous in the above three liver/kidney microsomal antibody type 1/HCV RNA positive cases. Subsequent steroid administration reduced alanine aminotransferase peaks, but may be harmful in viral infections. Therapeutic alternatives are needed: they will probably include pure antivirals (exerting no immunostimulatory effects) with or without immunosuppressive drugs.

Quantitation of hepatitis C virus genome molecules in plasma samples.

A competitive reverse transcription PCR (cRT-PCR)-based assay for the quantitative detection of hepatitis C virus (HCV) viremia was developed, optimized, and applied to the direct molecular analysis of clinical samples from nine patients with persistent HCV infection. As for other competitive PCR-based applications, this method consists of the reverse transcription and subsequent amplification of two RNA species in the same tube: the wild-type template (to be quantified) and a known amount of a modified synthetic template. These templates have identical primer recognition sites and very similar (but not identical) sizes, thus allowing direct detection of both template species after gel electrophoresis and ethidium bromide staining. The results obtained by this cRT-PCR application for testing clinical samples from HCV-infected patients mainly indicate that the competitive approach reaches the degree of sensitivity (fewer than 5 HCV RNA molecules per 100 microliters) necessary to evaluate viral load in all HCV-infected patients, independently of clinical conditions, and that this technique is flexible enough to quantify highly divergent levels of cell-free HCV genome copy numbers in biological samples. Interestingly, we observed a sample-to-sample variation in the loss of detectable HCV genome molecules in serum in comparison with that in plasma from the same patient, thus indicating that serum specimens, although widely used in the past few years for qualitative molecular investigation of HCV-infected patients, cannot be used to obtain reliable quantitative data on HCV viremia from these patients.

Testing for hepatitis C virus sequences in peripheral blood mononuclear cells of patients with chronic hepatitis C in the absence of serum hepatitis C virus RNA.

Hepatitis C virus (HCV) is able to replicate in peripheral blood mononuclear cells (PBMC) of HCV-infected patients. Few data are available on PBMC testing for HCV RNA in serum HCV RNA negative patients, positive for anti-HCV and with histological evidence of chronic hepatitis. Twenty such patients were studied; of these, 11 were tested during interferon alpha (IFN) treatment, at the time of serum HCV RNA clearance and ALT normalisation: only one was found to be positive for HCV sequences in PBMC. Within 3 months of IFN withdrawal all 11 patients relapsed with high ALT and recurrence of serum HCV RNA. Of nine serum HCV RNA negative patients with chronic hepatitis C who were not receiving IFN when tested (four untreated patients and five patients who had already completed IFN schedule), PBMC HCV RNA was detected in four. Evidence of active HCV replication (presence of the minus strand genome) in PBMC was also observed in two cases. Thus, five of the 20 patients without detectable serum HCV RNA turned out to be carriers of HCV sequences in PBMC. These data indicate that: 1. PBMC are an extrahepatic replication site of HCV; this is true also in the absence of serum HCV RNA; 2. the role of PBMC as a "viral reservoir" after IFN-induced serum HCV RNA clearance is questioned; 3. the absence of both serum and PBMC HCV RNA in patients under IFN is not predictive of sustained viral loss; 4. testing for PBMC viral sequences might enhance the chances of detecting HCV infection.

Increased risk of hepatocellular carcinoma development in patients with cirrhosis and with high hepatocellular proliferation.

The immunohistochemical determination of the accessory protein of DNA-polymerase delta (PCNA), a marker of an early S-phase of the cell cycle, was used to evaluate cell proliferation retrospectively in formalin-fixed, paraffin-embedded liver biopsy sections in a group of patients with cirrhosis of similar age and duration of follow up, and with no evidence of hepatocellular carcinoma (41), including 17 patients with and 24 without hepatocellular carcinoma appearance during follow up. Proliferation was expressed as total (PCNA-TOT) and strongly (PCNA-STRO) positive nuclei per 1000 hepatocytes. The presence of dysplasia was also recorded. Histological findings and biochemical data, at the time of liver biopsy, were compared in the two groups. While total PCNA positivities were not significantly different in the two groups, strong reactivity was significantly higher in patients who eventually developed hepato-cellular carcinoma (median 0.7 vs 2.6). Univariate analysis of histological and biochemical data at the time of biopsy, followed by a stepwise regression study, showed that the significant parameters for a time-dependent disease-free state were, in decreasing order: cholesterol, PCNA-STRO, PCNA-TOT and alpha foeto-protein. Other clinical, biochemical and histological parameters, including dysplasia, provided no further information. From these data, hepatocellular proliferation can be evaluated in patients with cirrhosis with a currently available technique. Patients with high cell proliferation are at increased risk of developing hepatocellular carcinoma and may require differentiated follow up.