In our geographic area 2% to 3% of the general population

Transcription

1 Identification of Chronic Hepatitis C Patients Without Hepatic Fibrosis by a Simple Predictive Model Xavier Forns, 1 Sergi Ampurdanès, 1 Josep M. Llovet, 1 John Aponte, 2 Llorenç Quintó, 2 Eva Martínez-Bauer, 1 Miquel Bruguera, 1 Jose Maria Sánchez-Tapias, 1 and Juan Rodés 1 Liver biopsy is required for staging hepatic fibrosis in patients with chronic hepatitis C, but it is an expensive procedure with occasional complications and poor patient acceptance. This cohort study was designed to assess the accuracy of a noninvasive method aimed to discriminate between patients with and without significant liver fibrosis (stages 2-4 versus 0-1). Clinically relevant variables were analyzed in a cohort of 476 consecutive untreated patients (estimation group, 351 patients; validation group, 125 patients) with chronic hepatitis C who underwent a liver biopsy. Multivariate analysis identified age, gamma glutamyl transpeptidase (GGT), cholesterol, platelet count, and prothrombin time as independent predictors of fibrosis. We constructed a model and a score system combining age, GGT, cholesterol, and platelet count that proved useful to identify patients without significant hepatic fibrosis. The area under the ROC curve was 0.86 for the estimation group and 0.81 for the validation group. Using the best cutoff score (less than 4.2), presence of significant fibrosis (F2 to F4) could be excluded with high accuracy (negative predictive value of 96%) in 125 (36%) of 351 patients. Similarly, it could be excluded with the same certainty in 49 (39%) of the 125 patients of the validation group. Only 2 patients with liver fibrosis stage 2 were incorrectly classified. In conclusion, a combination of easily accessible variables accurately predicts the absence of significant fibrosis and might render liver biopsy unnecessary in more than one third of patients with chronic hepatitis C. (HEPATOLOGY 2002;36: ) In our geographic area 2% to 3% of the general population is chronically infected with the hepatitis C virus (HCV). 1 Chronic hepatitis C is the most prevalent disease in hepatology clinics, accounting for more than 30% of the visits. In general, it is accepted that the diagnostic protocol of chronic hepatitis C includes a liver biopsy, particularly in patients with elevated amino transferase levels. 2-5 Liver histology and, especially, fibrosis staging provide prognostic information 6,7 and may be useful in deciding on therapeutic strategies in individual cases. 2,3 However, liver biopsy is an invasive procedure Abbreviations: HCV, hepatitis C virus; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma glutamyl transpeptidase; ROC, receiver operating characteristic; AUC, area under the curve. From the 1 Liver Unit, Institut de Malalties Digestives, and the 2 Statistics and Epidemiology Unit, Hospital Clinic, Institut d Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), University of Barcelona, Catalonia, Spain. Received March 21, 2002; accepted July 17, Josep M. Llovet has a contract from Programa Ramon y Cajal (Ministerio de Ciencia y Tecnología). Address reprint requests to: Xavier Forns, M.D., Liver Unit, Institut de Malalties Digestives, Villarroel 170, Hospital Clinic, Barcelona 08036, Spain. xforns@clinic.ub.es; fax: (93) Copyright 2002 by the American Association for the Study of Liver Diseases /02/ $35.00/0 doi: /jhep that may cause undesirable events, such as pain in 20% to 30% of the cases, major complications in 0.5%, and even death. 8 Other than the complications derived from the procedure and the frequent poor patient acceptance, the direct cost of such procedure is high. 2,3 Nowadays, chronic hepatitis C is often recognized at an early stage of the disease, and liver fibrosis is mild or absent in about 80% of the patients undergoing a liver biopsy. 6 Thus, the finding of surrogate markers of liver fibrosis/absence of fibrosis would be relevant to reduce the number of liver biopsies in patients with chronic hepatitis C. Imbert-Bismuth et al. 9 have recently shown that a combination of biochemical markers of liver fibrosis can be useful to predict the presence or absence of fibrosis and, therefore, to reduce the number of liver biopsies in patients with chronic hepatitis C. However, some of these markers, such as 2 macroglobulin, haptoglobin, or apolipoprotein A1 are not routinely used in clinical practice. This is also the case for other proposed predictors of fibrosis, such as hyaluronate concentration or type III procollagen Using variables easily available to clinicians, we have constructed and validated a model and a score system aimed to discriminate patients with substantial fibrosis 986

2 HEPATOLOGY, Vol. 36, No. 4, 2002 FORNS ET AL. 987 from those without significant fibrosis. This model may show liver biopsy to be unnecessary in a considerable proportion of patients with chronic hepatitis C. Patients and Methods Patients. The cohort study included 502 consecutive patients with chronic hepatitis C admitted to our Liver Unit to undergo a liver biopsy between July, 1996 and December, Clinical, biochemical and hematologic data were recorded from each patient at the time of liver biopsy. All patients gave oral informed consent to use these data for scientific purposes and the study was approved by our Ethical Committee. Diagnosis of chronic hepatitis C was established by elevation of alanine aminotransferase (ALT) (greater than 40 IU/L) in at least 2 separate determinations, a positive third-generation anti-hcv test (Ortho Diagnostic Systems, Raritan, NJ), and presence of HCV-RNA in serum determined by either qualitative (Amplicor HCV 2.0; Roche Diagnostics, Branchburg, NJ, USA) or quantitative assay (Cobas Amplicor HCV Monitor 2.0; Roche Diagnostics, Branchburg, NJ). HCV genotype was determined by restriction fragment length polymorphism after amplification of the 5 noncoding region of the HCV genome, as previously described. 13 Exclusion criteria for the study were age greater than 65 years, regular alcohol intake higher than 30 g/d, morbid obesity, infection with the hepatitis B or the human immunodeficiency viruses, active intravenous drug abuse, previous interferon treatment, liver transplantation, and clinical or ultrasonographic evidence of cirrhosis. 6 Histologic Staging. Ultrasonographic-guided liver biopsy was performed according to a standardized protocol and after patient s written informed consent. Specimens were fixed, paraffin-embedded, and stained with hematoxylin-eosin and Masson s trichrome. A minimum of 6 portal tracts in the specimen were required for diagnosis. Fibrosis stage was determined using the Scheuer s classification. 14,15 All liver biopsy samples were evaluated by the same pathologist (M.B.), who was blinded for clinical data. Liver fibrosis was considered significant when it spread out the portal tract (stages 2, 3, or 4), whereas it was considered nonsignificant when it was absent or restricted to the portal tract (stages 0 or 1, respectively). 9 Twenty-six patients were excluded from the study because of inadequate histologic samples. Thus, the final study cohort included 476 patients. Statistical Analysis. The main endpoint was the identification of the presence or absence of significant fibrosis with a combination of clinically relevant variables at the time of biopsy. Most of these variables had been previously identified as potential predictors of fibrosis or the absence thereof 10-12,16-21 : age, sex, fasting glucose, ALT, aspartate amino transferase (AST), gamma glutamyl transpeptidase (GGT), bilirubin, cholesterol, albumin, leucocyte and platelet counts, and prothrombin time, measured as percentage of the daily internal control. Viral load and genotype were systematically recorded since Data from a randomly generated split-sample of 351 patients (70%) were used to estimate the model, and data from the remaining 125 patients were used to validate the model. Each of the variables included in the model was analyzed to rule out any significant differences between the estimation and the validation groups. Continuous variables were analyzed in a logarithmic scale to improve its normal distribution, except for prothrombin time that was categorized by its median value. In the estimation group, univariate analysis identified predictors of fibrosis by using the Fisher s exact test for categoric variables and the Student s t test for quantitative variables. Thereafter, all variables were included in a multivariate forward stepwise logistic regression analysis to determine the independent predictors of the absence or presence of significant fibrosis. Significant fibrosis was considered as positive result and absence of significant fibrosis as a negative result. A predictive model was constructed by modeling the values of the independent variables and their coefficient of regression. The diagnostic value of the model was assessed by calculating the areas under the receiver operating characteristic (ROC) curves. An area under the curve (AUC) of 1.0 is characteristic of an ideal test, whereas 0.5 indicates a test of no diagnostic value. The diagnostic accuracy was calculated by sensitivity, specificity, positive and negative predictive values, considering significant fibrosis as the disease. To simplify the model, a score system was constructed, from 0 (absence of fibrosis) to 10 (cirrhosis). The best cutoff points were selected from the ROC curve to identify presence and absence of significant fibrosis. For that purpose, we selected cutoff points with a 95% certainty of presence and absence of fibrosis, thus assuming a 5% of false negative results as clinically acceptable. The predictive accuracy of the model was then tested in the validation group. Statistical analysis was performed by STATA (StataCorp Stata Statistical Software: Release 6.0.College Station, TX: Stata Corporation). Results Patients Characteristics. As stated above, the final study cohort included 476 patients. The model was con-

3 988 FORNS ET AL. HEPATOLOGY, October 2002 Table 1. Baseline Characteristics of the 476 Patients With Chronic Hepatitis C at the Time of Liver Biopsy: Comparison Between the Estimation and the Validation Groups Variable Estimation Group (n 351) Validation Group (n 125) All Patients (n 476) Age (y) 39 (33-51) 38 (32-51) 39 (33-51) Male gender n, (%) 226 (64.4%) 80 (64%) 306 (64.3%) AST (IU/L) 57 (41-84) 53 (41-74) 56 (41-84) ALT (IU/L) 97 (66-149) 90 (64-135) 95 (66-149) Gamma-GT (IU/L) 38 (24-68) 41 (22-68) 38 (24-68) Bilirubin (mg/dl) 0.8 (0.7-1) 0.9 (0.7-1) 0.8 (0.7-1) Glucose (mg/dl) 92 (86-100) 92 (85-98) 92 (86-100) Cholesterol (mg/dl)* 173 ( ) 178 ( ) 175 ( ) Albumin (g/l) 44 (43-47) 45 (43-47) 45 (43-47) Leucocytes (10 9 /L) 6.2 (5-7.5) 6.0 ( ) 6.1 (5-7.5) Platelets (10 9 /L) 185 ( ) 188 ( ) 187 ( ) Prothrombin time (%) 100 (97-100) 100 (98-100) 100 (97-100) Viral load (IU/mL 10 3 ) 754 ( ) 746 ( ) 750 ( ) HCV Genotype (86%) 83 (84%) 325 (85%) Stage of fibrosis n, (%) (50.7%) 65 (52%) 243 (51%) 1 88 (25.1%) 27 (21.6%) 115 (24.1%) 2 31 (8.8%) 16 (12.8%) 47 (9.9%) 3 33 (9.4%) 13 (10.4%) 46 (9.7%) 4 21 (6%) 4 (3.2%) 25 (5.3%) NOTE. Quantitative variables are expressed as median (centile 25; centile 75); categoric variables are expressed as n, (%); there were no significant differences between the estimation and the validation groups in any of the variables. *Two patients received medication to reduce cholesterol levels. Viral load and genotype were available in 282 and 99 patients from the estimation and the validation groups, respectively. structed with data from 351 patients (estimation group) and was validated in the remaining 125 patients (validation group). Patients characteristics at the time of liver biopsy are shown in Table 1. There were no significant differences between the estimation and the validation groups in any of the assessed variables or regarding the date of liver biopsy. Half of the patients were classified as stage 0 (no fibrosis) and 24% as stage 1 (fibrosis restricted to the portal tract). The prevalence of significant fibrosis was 25%, 10% as stage 2 (few septa extending beyond the portal tract but with intact architecture), and 15% as stages 3 and 4 (bridging fibrosis or cirrhosis, respectively). Predictors of Fibrosis and Estimation of the Model. In the estimation group, all variables except for sex, viral load, and genotype were identified as predictors of fibrosis by univariate analysis (Table 2). By multivariate analysis, none of these 3 variables had independent predictive value; because genotype and viral load were not available in a significant proportion of patients, we did not include them in the final analysis. Five variables were identified as independent predictors of fibrosis (Table 2). These markers were in a decreasing rank: age (P.0001), GGT levels (P.0001), platelet count (P.0001), prothrombin time (P.003), and cholesterol levels (P.008). We constructed a model combining 4 of these variables. Prothrombin time was excluded because it is not expressed as international normalized ratio (in our center, it is calculated as the percent coagulation time with respect to coagulation time of a daily control). In addition, it did not increase the accuracy of the model (AUC with 5 variables of 0.87). The diagnostic value of this model was assessed in the estimation group by ROC curve, showing an AUC of 0.86 (Fig. 1A). When the model was applied to the validation group, the AUC remained high (0.81) (Fig. 1B). We constructed a simple score system applying a constant to the obtained formula: ln platelet count ln GGT ln age cholesterol Two cutoff values were chosen to identify absence (less than 4.21) and presence (greater than 6.9) of significant fibrosis (Fig. 1A). Applying the lower cutoff (score below 4.2), 120 (45%) of the 266 patients without significant fibrosis in the liver biopsy were correctly identified (Table 3). This represents more than one third of the entire population cohort. More importantly, the presence of significant fibrosis could be excluded with high certainty, because only 5 (4%) the 125 patients with a score below 4.21 had significant fibrosis in the liver biopsy (96% negative predictive value) (Table 3). None of these patients was cirrhotic (3 patients had a fibrosis stage 2 and 2 had a fibrosis stage 3). In the validation group, 47 (51%) of 92 patients without significant fibrosis in the liver biopsy

4 HEPATOLOGY, Vol. 36, No. 4, 2002 FORNS ET AL. 989 Table 2. Variables Associated With the Presence of Significant Fibrosis (Stage 2-4) in the Estimation Group (351 Patients) by Univariate and Multivariate Analysis Variable No Significant Fibrosis (n 266) Significant Fibrosis (n 85) P Value (Univariate) Odds Ratio (95% CI)* (Multivariate) Age (y) 37 (32-47) 51 (40-57) (3.8-23) Male gender n, (%) 174 (65.4%) 52 (61.2%).45 AST (IU/L) 51 (40-72) 74 (49-124).001 ALT (IU/L) 92 (65-139) 111 (77-204).001 Gamma-GT (IU/L) 33 (23-57) 52 (31-117) ( ) Bilirubin (mg/dl) 0.8 (0.7-1) 0.9 ( ).003 Glucose (mg/dl) 91 (86-99) 95 (91-106).02 Cholesterol (mg/dl) 178 ( ) 162 ( ) (0.98-1) Albumin (g/l) 45 (43-47) 44 (41-45).004 Leucocytes (10 9 /L) 6.5 ( ) 5.3 ( ).001 Platelets (10 9 /L) 193 ( ) 148 ( ) ( ) Prothrombin time (%) 100 (98-100) 98 (88-100) ( ) Viral load (IU/mL 10 3 ) 696 ( ) 1030 ( ).31 HCV Genotype (86%) 55 (85%).9 NOTE. Quantitative variables are expressed as median (centile 25; centile 75); categoric variables are expressed as n, (%). *Odds ratio and 95% CI of variables independently related with the presence of significant fibrosis. Viral load and genotype were available in 282 patients from the estimation group. Patients in whom genotype/viral load were available were not different from patients in whom these variables were not available in any of the analyzed variables. were identified correctly by applying the same score (Table 4). Similarly, hepatic fibrosis could be ruled out with a high negative predictive value: 47 (96%) of the 49 patients with a score below 4.21 were correctly classified (Table 4). The only 2 patients misclassified had a fibrosis stage 2 in the liver biopsy. Therefore, significant fibrosis could be excluded in more than one third of the patients undergoing liver biopsy with high accuracy (Tables 3, 4, and 5). Considering the validation and the estimation groups, 7 patients with a score lower than 4.21 showed significant fibrosis in the liver biopsy. Their score (median, 3.6; interquartile range, 2.7-4) did not differ significantly from the score of patients correctly classified (median, 3; interquartile range, ). Applying the high cutoff (score greater than 6.9), 37 (44%) of the 85 patients belonging to the estimation group with significant fibrosis in the liver biopsy were correctly identified. Thirty-seven of the 47 patients with a score higher than 6.9 showed significant fibrosis in the liver biopsy (79% positive predictive value) (Table 3). In the validation group, significant fibrosis could be predicted in only 10 (30%) of 33 patients with significant fibrosis in the liver biopsy. The positive predictive value of the model was low, because only 10 (66%) out of 15 patients with a score higher than 6.9 had significant fibrosis (Tables 4 and 5). Therefore, the diagnostic accuracy of the model decreased significantly in the validation group (Table 5). Considering the validation and the estimation groups, 15 patients with a score higher than 6.9 showed mild fibrosis (12 patients) or no fibrosis (3 patients) in the liver Fig. 1. ROC curves of a model aimed to discriminate patients with chronic hepatitis C with and without significant liver fibrosis. The model combines 4 variables: age, platelet count, GGT, and cholesterol. (A) ROC curve in an estimation group (n 351). AUC, Cutoff scores (4.2 and 6.9) and an intermediate score (5) are depicted. (B) ROC curve in a validation group (n 125). AUC 0.81.

5 990 FORNS ET AL. HEPATOLOGY, October 2002 Table 3. Predictive Value of the Model Obtained From the Estimation Group Fibrosis Stage Predicted by the Model biopsy. The score was significantly lower (and therefore closer to the cutoff) in these patients (median, 7.3; interquartile range, 7-8) compared with patients correctly classified (median, 8; interquartile range, ) (P.009). We did not find additional differences between the groups. Discussion All Patients (n 351) Fibrosis Stage Seen in Liver Biopsies Stage 0-1 (n 266) Stage 2-4 (n 85) Low cut-off point: Stage 0-1 (score 4.2) 125 (36%) 120 (45%) 5 (6%) Stage 2-4 (score 4.2) 226 (64%) 146 (55%) 80 (94%) High cut-off point Stage 0-1 (score 6.9) 304 (87%) 256 (96%) 48 (56%) Stage 2-4 (score 6.9) 47 (13%) 10 (4%) 37 (44%) Table 4. Predictive Value of the Model Obtained From the Validation Group Fibrosis Stage Predicted by the Model Total (n 125) Fibrosis Stage Seen in Liver Biopsies Stage 0-1 (n 92) Stage 2-4 (n 33) Low cut-off point Stage 0-1 (score 4.2) 49 (39%) 47 (51%) 2 (6%) Stage 2-4 (score 4.2) 76 (61%) 45 (49%) 31 (94%) High cut-off point Stage 0-1 (score 6.9) 110 (88%) 87 (95%) 23 (70%) Stage 2-4 (score 6.9) 15 (12%) 5 (5%) 10 (30%) Table 5. Diagnostic Accuracy of the Score System in the Validation Group (n 125); Prevalence of Significant Fibrosis, 26% Cutoff Score Accuracy S Sp PPV NPV LR Population Involved Interpretation % Absence of fibrosis (96% certainty) % Presence of fibrosis (66% certainty) Abbreviations: S, sensitivity; Sp, specificity; PPV, positive predictive value; NPV, negative predictive value; LR, likelihood ratio. Liver biopsy is a valuable tool in assessing the natural history of chronic hepatitis C, particularly when the data of onset of infection is known or when more than 1 biopsy is available. 2-5 Because current antiviral therapy is expensive, has many adverse effects, and often is ineffective, liver biopsy also may influence treatment decisions. In fact, the presence of mild disease in the liver biopsy might permit deferring treatment, particularly when virologic variables (viral load and genotype) are predictive of low response probability. On the contrary, the presence of advanced disease might recommend initiating antiviral therapy, even in patients with a poor response profile. However, liver biopsy is an invasive procedure with associated morbidity that carries a significant cost. 2,3,8 For these reasons, several attempts have been made to find accurate noninvasive markers of disease activity and fibrosis. Most of these attempts have focused on the noninvasive diagnosis of patients with advanced liver disease (bridging fibrosis or cirrhosis). 9,12,16,19 Age and platelet count, 16,19 ALT/AST ratio, and prothrombin index, 12,16 as well as serum fibrosis markers such as hyaluronate, 2 macroglobulin, or haptoglobin 9-12 appear to identify correctly a significant proportion of HCV-infected patients with advanced hepatic fibrosis; however, currently most patients with chronic hepatitis C are diagnosed at an early stage of the disease and have only mild lesions in the liver biopsy specimen. Refraining from an invasive procedure appears to be more rational in these subjects. 6 Our study was aimed at discriminating patients with and without significant fibrosis with a noninvasive method. The model identified around half of patients without fibrosis or fibrosis restricted to the portal tract (stages 0 and 1) with a high predictive value. The model consisted of a combination of 4 variables identified by comparative analysis of patients with and without significant fibrosis: age, GGT, platelet count, and cholesterol level. With these variables we constructed a simple score addressed to select patients at very low risk of having significant fibrosis. In our series, a score below 4.2 identified these patients with a 96% of certainty. This high level of confidence for discarding significant fibrosis has not been obtained with any proposed model with conventional markers. In fact, the accuracy of this model resembles other recently proposed models based on more sophisticated variables. In the study from Imbert-Bismut, 9 the variables proposed are useful tools for research purposes and provide high levels of certainty of the presence or absence of significant fibrosis, but its usefulness may be curtailed because some of the predictive markers (such as 2 macroglobulin, haptoglobin, or apolipoprotein A1) are not used in clinical practice in most centers. In addition, the patient population analyzed in this study included a large proportion of subjects with advanced hepatitis C infection, as indicated by the presence of significant fibrosis in up to 40% of the subjects. In our series, significant fibrosis was present in only 25% of the cases, a figure possibly closer to the spectrum of the disease in the

6 HEPATOLOGY, Vol. 36, No. 4, 2002 FORNS ET AL. 991 community. Even considering that our cohort also reflects a selection bias toward an advanced disease, liver biopsy may be avoidable in more than 1 third of the patients. Therefore, our model may be particularly useful and costeffective in community-based series of patients with HCV infection. 22 The value of age as a marker of fibrosis seems obvious as fibrosis progression is time-dependent. 6,21 Age at infection also has been shown to influence the outcome of hepatitis C and patients infected after the fourth decade have a higher risk of disease progression. 20,21,23,24 It is evident that duration of HCV infection would be a more precise indicator of fibrosis than age. However, in our patient population the mechanism of acquisition of HCV infection remains unknown in around half of the patients, 6 and, therefore, duration of HCV infection is difficult to establish in many cases. The prognostic value of a low platelet count as a marker of severe fibrosis has been reported previously. 16,19 Thrombocytopenia in patients with advanced liver disease seems to be related to the development of portal hypertension and the decreased production of thrombopoietin. 25 The relationship between cholesterol levels and liver fibrosis observed in our study is difficult to explain. The decrease in cholesterol levels seen in patients with advanced cirrhosis is caused by a reduction in cholesterol synthesis and, therefore, it is unlikely that the synthetic capacity of the hepatocytes can be altered before liver cirrhosis is established. However, other mechanisms might influence the lipid profile in patients with a persistent infection. In a study by Fabris et al., 17 patients with chronic HCV infection had lower cholesterol levels than patients infected with HBV. The lower cholesterol levels in HCV-infected patients were not dependent on age, sex, or liver function and the authors speculated that the lipid profile might vary depending on the type of viral infection because of different cytokine profiles. The use of cholesterol as a variable may be a concern in areas with a high prevalence of hypercholesterolemia, where therapy will artificially modify the serum levels of cholesterol. The value of GGT as a marker of liver fibrosis already has been described by others. 18,26 In our study its prognostic value was independent from the bilirubin and transaminase levels. HCV infection frequently causes bile duct damage and steatosis. 27 Bile duct lesions can partially explain increased GGT values and it appears that patients with bile duct lesions have significantly higher fibrosis scores. In a large cohort of Italian patients with chronic hepatitis C, Giannini et al. 18 showed that in patients with chronic hepatitis C bile duct damage was present in more than one third of patients with substantial fibrosis (stages 2-4) in comparison with less than 10% in patients with fibrosis stage 0-1. In this study, increased GGT was an independent predictor of bile duct damage. Our study has limitations related to the selection of the variables and the lack of accuracy to detect significant fibrosis. The inclusion of body mass index, for instance, might have been helpful in replacing other variables that can be artificially modified such as cholesterol and fasting glucose. Viral load and genotype were not included in the final analysis because they were not available in a significant number of patients. However, we did not find any relationship between these variables and liver fibrosis either in the univariate or in the multivariate analysis. 21,24,28 In addition, virologic variables are not always easily accessible to clinicians and HCV genotype geographic distribution would make the results of this study inapplicable elsewhere. The model was not accurate enough to detect significant fibrosis. From a clinical point of view, and despite the high likelihood ratio, the validation group provided a nonacceptable positive predictive value of fibrosis (0.66). Although more sophisticated markers may provide higher accuracy for this purpose, there are several published models that predict the presence of bridging fibrosis or cirrhosis with acceptable accuracy. 16,19 In summary, our results indicate that half of the patients with chronic hepatitis C without significant liver fibrosis can be identified with high accuracy with a combination of a few clinical, biochemical and hematologic variables. More importantly, our model might render liver biopsy unnecessary in a significant proportion of the patients with chronic hepatitis C referred to hepatology clinics. Although our results need to be validated by other centers, we believe that this model might be widely used in clinical practice, especially in a community-based approach. References 1. Garcia-Bengoechea M, Emparanza JI, Sarriugarte A, Cortes A, Vega JL, Gonzalez F, Arenas JI. Antibodies to hepatitis C virus: a cross-sectional study in patients attending a trauma unit or admitted to hospital for elective surgery. Eur J Gastroenterol Hepatol 1995; 7: Perrillo RP. The role of liver biopsy in hepatitis C. HEPATOLOGY 1997; 26(Suppl 1):57S-61S. 3. Saadeh S, Cammell G, Carey WD, Younossi Z, Barnes D, Easley K. The role of liver biopsy in chronic hepatitis C. HEPATOLOGY 2001;33: National Institutes of Health Consensus Development Conference Panel statement: management of hepatitis C. HEPATOLOGY 1997; 26(Suppl 1): 2S-10S. 5. EASL International Consensus Conference on hepatitis C. Paris, February Consensus statement. J Hepatol 1999;3(Suppl 1): Forns X, Ampurdanes S, Sanchez-Tapias JM, Guilera M, Sans M, Sanchez-Fueyo A, Quintó L, et al. Long-term follow-up of chronic hepatitis C in patients diagnosed at a tertiary-care center. J Hepatol 2001;35:

7 992 FORNS ET AL. HEPATOLOGY, October Yano M, Kumada H, Kage M, Ikeda K, Shimamatsu K, Inoue O, Hashimoto E, et al. The long-term pathological evolution of chronic hepatitis C. HEPATOLOGY 1996;23: Cadranel JF, Rufat P, Degos F. Practices of liver biopsy in France: results of a prospective nationwide survey. For the Group of Epidemiology of the French Association for the Study of the Liver (AFEF). HEPATOLOGY 2000; 32: Imbert-Bismut F, Ratziu V, Pieroni L, Charlotte F, Benhamou Y, Poynard T. Biochemical markers of liver fibrosis in patients with hepatitis C virus infection: a prospective study. Lancet 2001;357: Guechot J, Loria A, Serfaty L, Giral P, Giboudeau J, Poupon R. Serum hyaluronan as a marker of liver fibrosis in chronic viral hepatitis C: effect of alpha-interferon therapy. J Hepatol 1995;22: Guechot J, Laudat A, Loria A, Serfaty L, Poupon R, Giboudeau J. Diagnostic accuracy of hyaluronan and type III procollagen amino-terminal peptide serum assays as markers of liver fibrosis in chronic viral hepatitis C evaluated by ROC curve analysis. Clin Chem 1996;42: Oberti F, Valsesia E, Pilette C, Rousselet MC, Bedossa P, Aube C, Gallios Y, et al. Noninvasive diagnosis of hepatic fibrosis or cirrhosis. Gastroenterology 1997;113: Lopez-Labrador FX, Ampurdanes S, Forns X, Castells A, Saiz JC, Costa J, Bruix J, et al. Hepatitis C virus (HCV) genotypes in Spanish patients with HCV infection: relationship between HCV genotype 1b, cirrhosis and hepatocellular carcinoma. J Hepatol 1997;27: Scheuer PJ. The nomenclature of chronic hepatitis: time for a change. J Hepatol 1995;22: Scheuer PJ. Classification of chronic viral hepatitis: a need for reassessment. J Hepatol 1991;13: Bonacini M, Hadi G, Govindarajan S, Lindsay KL. Utility of a discriminant score for diagnosing advanced fibrosis or cirrhosis in patients with chronic hepatitis C virus infection. Am J Gastroenterol 1997;92: Fabris C, Federico E, Soardo G, Falleti E, Pirisi M. Blood lipids of patients with chronic hepatitis: differences related to viral etiology. Clin Chim Acta 1997;261: Giannini E, Ceppa P, Botta F, Fasoli A, Romagnoli P, Cresta E, Venturino V, et al. Steatosis and bile duct damage in chronic hepatitis C: distribution and relationships in a group of Northern Italian patients. Liver 1999;19: Poynard T, Bedossa P. Age and platelet count: a simple index for predicting the presence of histological lesions in patients with antibodies to hepatitis C virus. METAVIR and CLINIVIR Cooperative Study Groups. J Viral Hepat 1997;4: Poynard T, Bedossa P, Opolon P. Natural history of liver fibrosis progression in patients with chronic hepatitis C. The OBSVIRC, METAVIR, CLINIVIR, and DOSVIRC groups. Lancet 1997;349: Poynard T, Ratziu V, Charlotte F, Goodman Z, McHutchison J, Albrecht J. Rates and risk factors of liver fibrosis progression in patients with chronic hepatitis C. J Hepatol 2001;34: Freeman AJ, Dore GJ, Law MG, Thorpe M, Von Overbeck J, Lloyd AR, Marinos G, et al. Estimating progression to cirrhosis in chronic hepatitis C virus infection. HEPATOLOGY 2001;34: Kage M, Shimamatu K, Nakashima E, Kojiro M, Inoue O, Yano M. Long-term evolution of fibrosis from chronic hepatitis to cirrhosis in patients with hepatitis C: morphometric analysis of repeated biopsies. HEPA- TOLOGY 1997; 25: Seeff LB. Natural history of hepatitis C. Am J Med 1999;107:10S-15S. 25. Adinolfi LE, Giordano MG, Andreana A, Tripodi MF, Utili R, Cesaro G, Ragone E, et al. Hepatic fibrosis plays a central role in the pathogenesis of thrombocytopenia in patients with chronic viral hepatitis. Br J Haematol 2001;113: Hwang SJ, Luo JC, Chu CW, Lai CR, Lu CL, Tsay SH, Wu JC, et al. Hepatic steatosis in chronic hepatitis C virus infection: prevalence and clinical correlation. J Gastroenterol Hepatol 2001;16: Scheuer PJ, Davies SE, Dhillon AP. Histopathological aspects of viral hepatitis. J Viral Hepat 1996;3: De Moliner L, Pontisso P, De Salvo GL, Cavalletto L, Chemello L, Alberti A. Serum and liver HCV RNA levels in patients with chronic hepatitis C: correlation with clinical and histological features. Gut 1998;42: