SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References

Diagnosis of the stage of liver fibrosis in chronic hepatitis C is essential for making a prognosis and deciding on antiviral therapy. In the present study a simple model consisting of routine laboratory tests was constructed and then validated in cross-sectional and longitudinal investigations. Consecutive treatment-naive patients with chronic hepatitis C who had undergone liver biopsy were divided into 2 cohorts: an estimation set (n = 240) and a validation set (n = 120). A longitudinal set consisted of 30 patients who had undergone a liver biopsy twice, before and after IFN treatment. The FibroIndex was derived from the platelet count, AST, and gamma globulin measurements in the estimation set. The areas under the ROC curves of the FibroIndex for predicting significant fibrosis were 0.83 and 0.82 for the validation set, better than those of the Forns index and the aminotransferase-to-platelet ratio index (APRI). Using the best cutoff values, whether significant fibrosis was present was diagnosed with high positive predictive values, and 35% of patients could avoid liver biopsy. In the longitudinal set, there was a significant decrease in the FibroIndex of 14 patients whose fibrosis stage improved, and a significant increase in that of 5 patients whose fibrosis stage deteriorated. Change in the FibroIndex correlated significantly with variation in fibrosis stage. There was no such correlation with the Forns index or the APRI. Conclusion: The FibroIndex is a simple and reliable index for predicting significant fibrosis in chronic hepatitis C and could also be used as a surrogate marker during antifibrotic treatment for chronic hepatitis C. (HEPATOLOGY 2007;45:297–306.)

Information on the stage of fibrosis in chronic hepatitis C is essential for making a prognosis and deciding on antiviral treatment. Liver biopsy is the gold standard for fibrosis staging in chronic hepatitis. However, liver biopsy is invasive and costly and requires an experienced hepatologist. Between 0.6% and 5% of patients who have undergone liver biopsy have had complications including death.1, 2 Moreover, recent studies have reported that inadequate sample size and sampling error frequently lead to underestimation of fibrosis stage and a high rate of inter- and intraobserver discrepancies.3, 4 Given these disadvantages, the usefulness of liver biopsy for dynamic surveillance and follow-up is limited. Noninvasive markers as surrogates of liver biopsy need to be developed. Several investigators have reported noninvasive approaches for the quantitative diagnosis of liver fibrosis, such as routine laboratory tests,5–9 serum markers of fibrosis,10, 11 radiological imaging,12, 13 and elastography.14, 15

For judging the accuracy, simplicity, and economy of these approaches, an index comprising routinely available laboratory tests that could predict significant fibrosis or cirrhosis in patients with hepatitis C would be satisfactory. The aminotransferase-to-platelet ratio index (APRI)9; the Forns index,8 based on age, platelets, AST, and cholesterol; and the Fibrotest, based on bilirubin, gamma-glutamyl transferase, apolipoprotein A1, alpha-2-macroglobulin, and haptoglobin, are indices that have been proposed for use as noninvasive markers. These markers have been examined in different validation studies that included many patients with cirrhosis. Therefore, the composition of these validation sets has seemed to increase the diagnostic performance. The present study excluded patients with cirrhosis (fibrosis stage F4) in order to investigate diagnostic performance among patients with chronic hepatitis (stages F0-F3) only. Few published studies have evaluated whether noninvasive markers of fibrosis reflect fluctuations in fibrosis stage following treatment or during the natural course of the disease. Consequently, we developed a single model, the FibroIndex, consisting of AST, platelets, and gamma globulin measurements from 240 consecutive patients with chronic hepatitis C, and compared its diagnostic accuracy with that of the APRI and the Forns index using a validation set from 120 subsequent patients and the validation set that included patients with stage F4 fibrosis (n = 162). Furthermore, we assessed the relation between change in fibrosis stage with changes in the 3 indices in a longitudinal set composed of 30 patients who had undergone IFN therapy.

Patients and Methods

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References

Patients.

This study included 402 consecutive patients with chronic hepatitis C admitted to our hospitals to undergo a liver biopsy between April 1994 and March 2004. Clinical data were collected at the time of liver biopsy, and blood samples were collected no more than 3 days before the biopsy. HCV infection was defined by a positive second-generation anti-HCV test and detection of HCV RNA on either a quantitative or a qualitative assay. Criteria for exclusion from the study were coinfection with HIV or the hepatitis B virus, regular alcohol intake higher than 10 g/day, other liver disease, previous interferon treatment, or clinical evidence of cirrhosis (gastroesophageal varices, ascites, hepatic encephalopathy).

Histological Staging.

Liver biopsy was performed under sonography or laparoscopy. A Silverman needle (13-gauge) under laparoscopy or a Tru-cut needle (14-gauge) under sonography was used to obtain an enough hepatic tissue. The mean length ± SD of the biopsy samples was 17.7 ± 4.9 mm. Two hundred and ninety-nine samples had 1 fragment, 49 samples had 2 fragments; and 12 samples had 3 fragments. The samples were all at least 10 mm in length. Biopsy specimens were stained with hematoxylin-eosin and silver. Fibrosis stage, determined according to the METAVIR group scoring system,16 was classified as F0, no fibrosis; F1, portal fibrosis without septa; F2, few septa; F3, numerous septa without cirrhosis; or F4, cirrhosis. One pathologist (Y.M.), who was unaware of patient characteristics, assessed and scored the specimens.

Statistical Analysis.

The main endpoint was discriminating patients with significant fibrosis (F2, F3) from those without significant fibrosis (F0, F1) using a combination of relevant biochemical or hematological variables. Omitting the 42 patients classified as F4, 360 of the 402 patients were randomly assigned by computer to either the estimation set group (240 patients) or the validation set group (120 patients). Concretely, all 360 patients with chronic hepatitis were assigned a different number from 1 to 360. The 240 patients who were randomly assigned numbers between 1 and 240 by computer software (Microsoft Excel, Microsoft Japan, Tokyo, Japan) were enrolled in the estimation set. The other 120 patients were enrolled into the validation set. An additional validation set that included patients classified as stage F4 consisted of the 120 patients in the validation set plus the 42 patients classified as F4.

Variables that differed significantly according to fibrosis stage in the estimation group were identified by univariate analyses (analysis of variance). Therefore, all variables were included in a multivariate forward stepwise regression analysis to determine the independent predictors of fibrosis stage. A predictive index was constructed by modeling the values of the independent variables and their coefficient of regression. The diagnostic value of the index was assessed by calculating the area under the receiver operating characteristic (ROC) curves. Diagnostic accuracy was calculated by sensitivity, specificity, positive and negative predictive values, and likelihood ratio. The cutoffs selected from the ROC curve were those that best identified significant fibrosis (F2 + F3 vs. F0 + F1 or F3 vs. F0-F2).

We selected cutoffs with a 95% specificity for the presence and absence of significant fibrosis, because a positive result tends to rule in the diagnosis when the index has an extremely high specificity (more than 95%). The predictive accuracy of the index was then tested in the validation group. Spearman correlation coefficients were used to evaluate whether change in fibrosis stage was correlated with changes in fibrosis markers. Statistical analysis was performed with Stat View version 5.0 (SAS Institute Inc.,Cary, NC).

Results

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References

Patient Characteristics.

The study cohort had 402 patients. The index was constructed with data from 240 patients (the estimation set) and was validated in the remaining 120 patients (the validation set) and in a validation set containing 162 patients (the patients in the first validation set plus the 42 patients with stage F4 fibrosis). Patient characteristics at the time of liver biopsy are shown in Table 1. There were no significant differences between the estimation and validation sets in any clinical or biochemical variable or in fibrosis stage.

Table 1. Characteristics of the 360 Patients in the Estimation Set, the Validation Set, and the Validation Set Containing Patients with Fibrosis Stage F4
VariableAll Patients Except Those at Stage F4 (n = 360)Estimation Set (n = 240)Validation Set (n = 120)P valueValidation Set Containing Patients at Stage F4 (n = 162)
  1. NOTE. P values are of comparisons between the estimation set and the validation set.

Age (years)51.6 ± 11.551.6 ± 11.550.5 ± 11.4ns53.7 ± 12.7
WBC (/mm3)229/131152/8877/43ns98/64
Male/female5,385 ± 1,6345,388 ± 1,6705,490 ± 1,401ns5,322 ± 1,586
Platelets (× 104/mm3)16.7 ± 6.016.9 ± 5.617.6 ± 6.1ns15.9 ± 6.4
Bilirubin (mg/dl)0.80 ± 0.390.78 ± 0.350.74 ± 0.31ns0.83 ± 0.37
AST (IU/L)71 ± 5268 ± 5171 ± 53ns76 ± 49
ALT (IU/L)100 ± 8196 ± 76112 ± 93ns104 ± 83
ALP (IU/L)236 ± 101236 ± 102220 ± 84ns248 ± 106
GGT (IU/L)69 ± 6663 ± 5573 ± 72ns77 ± 75
Albumin (g/dl)4.10 ± 0.414.11 ± 0.404.17 ± 0.34ns4.03 ± 0.43
γ-gl (g/dl)1.60 ± 0.421.57 ± 0.391.61 ± 0.43ns1.73 ± 0.48
Cholesterol (mg/dl)169 ± 33169 ± 33173 ± 31ns168 ± 34
Prothrombin time (%)92.2 ± 18.793.5 ± 17.891.5 ± 20.0ns88.0 ± 19.0
Stage of fibrosis     
 F028 (7.8%)18 (7.5%)10 (8.3%)ns10 (6.2%)
 F1149 (41.4%)99 (41.3%)50 (41.7%) 50 (30.9%)
 F270 (19.4%)47 (19.5%)23 (19.2%) 23 (14.2%)
 F3113 (31.4%)76 (31.7%)37 (30.8%) 37 (22.8%)
 F4    42 (25.9%)

Derivation of FibroIndex.

In the estimation group, all variables except white blood cells and cholesterol were identified as predictors of fibrosis stage in univariate analysis (Table 2). From these variables (age, platelets, bilirubin, AST, ALT, ALP, GGT, albumin, gamma globulin, prothrombin time), 3 variables (platelets, AST, gamma globulin) were identified as independent predictors in multivariate forward stepwise regression analysis. We constructed the FibroIndex as a simple score system:

  • equation image

The parameters AST, gamma globulin, and platelets were measured by a TBA 200FR (Toshiba Medical System Corporation, Tochigi, Japan), an Olympus AES 630 (Olympus Corporation, Tokyo, Japan), and a Beckman Coulter LH 750 system (Beckman Coulter Inc., CA), respectively. Their coefficients of variation were 1.47% for low-level samples and 1.04% for high-level samples for AST, 5.72% for normal-level samples and 2.98% for abnormal-level samples for gamma globulin, and 2.06% for low-level samples, 1.67% for normal-level samples, and 2.09% for high-level samples for platelet count. Therefore, these parameters maintained sufficient accuracy.

Table 2. Variables Associated with Fibrosis Stage in the Estimation Set (240 Patients) in Univariate Analysis
 F0 (n = 18)F1 (n = 99)F2 (n = 47)F3 (n = 76)P value
Age (years)46.2 ± 12.950.0 ± 13.052.1 ± 9.354.9 ± 9.30.006
WBC (/mm3)5,844 ± 2,0145,353 ± 1,7555,436 ± 1,7435,296 ± 1,422ns
Platelets (× 104/mm3)22.7 ± 6.519.1 ± 4.815.2 ± 5.013.8 ± 4.6<0.0001
Bilirubin (mg/dl)0.61 ± 0.190.74 ± 0.310.79 ± 0.380.86 ± 0.400.029
AST (IU/L)36 ± 1852 ± 3664 ± 39100 ± 64<0.0001
ALT (IU/L)71 ± 7075 ± 5691 ± 72132 ± 90<0.0001
ALP (IU/L)256 ± 68216 ± 85231 ± 109262 ± 1180.022
GGT (IU/L)61 ± 5252 ± 5767 ± 5175 ± 540.045
Albumin (g/dl)4.30 ± 0.414.20 ± 0.404.04 ± 0.393.99 ± 0.400.0005
γ-gl (g/dl)1.38 ± 0.281.43 ± 0.321.65 ± 0.451.73 ± 0.39<0.0001
Cholesterol (mg/dl)166 ± 30175 ± 36166 ± 35164 ± 26ns
Prothrombin time (%)97.6 ± 18.799.9 ± 16.094.1 ± 15.683.8 ± 18.6<0.0001

Comparison of Performance of FibroIndex with That of Forns Index and APRI in the Estimation Set.

Diagnostic performance of the FibroIndex, Forns index, and APRI in the estimation set was compared. Figure 1 shows box plots of the FibroIndex (F = 46.1, P < 0.0001), Forns index (F = 27.5, P < 0.0001), and APRI (F = 27.1, P < 0.0001) for each fibrosis stage, and Fig. 2 shows ROC curves for different degrees of fibrosis (Fig. 2A for F0-F1 versus F2-F3, Fig. 2B for F0-F2 versus F3). The areas under the ROCs (AUCs) of the FibroIndex, Forns index, and APRI for predicting significant fibrosis (F0-F1 vs. F2-F3) were 0.828 (95% confidence interval [CI], 0.777-0.879), 0.788 (95% CI, 0.731-0.845), and 0.794 (95% CI, 0.738-0.850), respectively. Similarly, AUCs for prediction of severe fibrosis (F0-F2 vs. F3) were 0.814 (95% CI, 0.758-0.870), 0.766 (95% CI, 0.703-0.829), and 0.802 (95% CI, 0.744-0.860), respectively.

thumbnail image

Figure 1. Relationship between fibrosis index and fibrosis stage in the estimation set: (A) FibroIndex, (B) Forns index, (C) aspartate aminotransferase–to–platelet ratio index. The top and bottom of each box represent the 25th and 75th percentiles, giving the interquartile range. The line through the box indicates the median, and the error bars indicate the 10th and 90th percentiles.

Download figure to PowerPoint

thumbnail image

Figure 2. ROC curves of the APRI, Forns index, and FibroIndex in predicting (A) significant fibrosis (F2 or F3) and (B) severe fibrosis (F3) in the estimation set. The areas under the ROC curves of the APRI, Forns index, and FibroIndex were 0.79, 0.79, and 0.83, respectively, in predicting F2 or F3 and 0.80, 0.77, and 0.82, respectively, in predicting F3.

Download figure to PowerPoint

Comparison of Performance of FibroIndex with That of Forns Index and APRI in the Validation Set.

Application of these indices to the validation set that contained patients at stage F4 showed these indices significantly increased according to fibrosis stage (FibroIndex: F = 34.7, P < 0.0001; Forns index: F = 33.1, P < 0.0001; APRI: F = 15.3, P < 0.0001; Fig. 3). The AUCs of the FibroIndex, Forns index, and APRI for predicting significant fibrosis (F2-F3 or F2-F4) in the validation set without and with F4 were 0.826 (95% CI, 0.753-0.898) and 0.864 (95% CI, 0.810-0.919), 0.778 (95% CI, 0.695-0.861) and 0.835 (95% CI, 0.774-0.896), and 0.776 (95% CI, 0.694-0.857) and 0.820 (95% CI, 0.755-0.885), respectively (Figs. 4A and 5A). Similarly, AUCs for predicting severe fibrosis (F3 or F3-F4) were 0.813 (95% CI, 0.733-0.892) and 0.848 (95% CI, 0.788-0.908), 0.760 (95% CI, 0.675-0.846) and 0.831 (95% CI, 0.769-0.892), and 0.772 (95% CI, 0.688-0.857) and 0.810 (95% CI, 0.744-0.876), respectively (Figs. 4B and 5B). The accuracy of the FibroIndex in predicting significant fibrosis and severe fibrosis in the validation set without or with F4 was better than the Forns index or the APRI. Predictive values of the FibroIndex in the validation set without F4 were similar to those in the estimation set.

thumbnail image

Figure 3. Relationship between fibrosis index and fibrosis stage in the validation set containing F4: (A) FibroIndex, (B) Forns index, (C) aspartate aminotransferase–to–platelet ratio index. The top and bottom of each box represent the 25th and 75th percentiles, giving the interquartile range. The line through the box indicates the median, and the error bars indicate the 10th and 90th percentiles.

Download figure to PowerPoint

thumbnail image

Figure 4. ROC curves of the APRI, Forns index, and FibroIndex in predicting (A) significant fibrosis (F2 or F3) and (B) severe fibrosis (F3) in the validation set. The areas under the ROC curves of the APRI, Forns index, and FibroIndex were 0.78, 0.78, and 0.83, respectively, in predicting F2 or F3 and 0.77, 0.76, and 0.81, respectively, in predicting F3.

Download figure to PowerPoint

thumbnail image

Figure 5. ROC curves of the APRI, Forns index, and FibroIndex in predicting (A) significant fibrosis (F2-F4) and (B) fibrosis (F3-F4) in the validation set containing F4. The areas under the ROC curves of the APRI, Forns index, and FibroIndex were 0.82, 0.84, and 0.86, respectively, in predicting F2-F4 and 0.81, 0.83, and 0.85, respectively, in predicting F3-F4.

Download figure to PowerPoint

In general, when a test has an extremely high specificity (more than 95%), a positive result tends to rule in the diagnosis. Therefore, we selected cutoff values for each index that achieved approximately 95% specificity for the diagnosis of F0-F1 or F2-F3 and diagnostic accuracy for each index in the estimation set (Table 3). Two cutoffs for the FibroIndex were chosen in order to determine if significant fibrosis was (≥2.25) or was not (≤1.25) present. Applying the lower cutoff (≤1.25) correctly identified 47 of 117 patients without significant fibrosis (40.2%) from liver biopsy. Furthermore, only 7 of the 54 patients with a score of 1.25 or less (13%) had significant fibrosis. Similarly, applying the higher cutoff (≥ 2.25) correctly identified 44 of the 123 patients with significant fibrosis (35.8%). Forty-four of the 47 patients with a score of at least 2.25 (93.6%) had significant fibrosis on liver biopsy (93.6% positive predictive value). Taken together, a total of 101 patients (42.1%), 54 patients with a score of 1.25 or less and 47 patients with a score of at least 2.25, could avoid liver biopsy.

Table 3. Diagnostic Accuracy of 3 Indices in the Estimation Set (n = 240)
  Cutoff ValueSensitivity (%)Specificity (%)Positive Predictive Value (%)Negative Predictive Value (%)Likelihood RatioBiopsy Elimination Rate
APRIF0-1≤0.3626.595.183.857.65.430.4
 F2-3≥0.8534.195.789.458.07.9 
Forns indexF0-1≤4.525.697.690.958.010.625.0
 F2-3≥8.724.396.688.254.97.1 
FibroIndexF0-1≤1.2540.294.387.062.47.137.9
 F2-3≥2.2535.897.494.359.113.8 

In the validation set, applying a cutoff of ≤1.25, 24 of the 60 patients without significant fibrosis (40.0%) were correctly identified, and only 2 of the 26 patients with a score of ≤1.25 (7.6%) had F2 (Table 4). None of the patients at stage F3 or F4 in the validation set not containing F4 had a score of 1.25 or less. Applying a cutoff of ≥2.25, 18 of the 60 patients with significant fibrosis (30.0%) were identified, and 18 of the 20 patients with a score of at least 2.25 (90.0%) had significant fibrosis. Two patients at stage F1 and no patients at stage F0 in the validation set containing F4 had a FibroIndex score of 2.25 or greater. These patients had low platelet counts (6.9 × 104 and 12.2 × 104), and their biopsies were performed under sonography. Therefore, these discordances might be a result of biopsy errors. Comparison of the FibroIndex, Forns index, and APRI showed that liver biopsy could be avoided in 37.9% of those in the estimation group by FibroIndex (cutoffs: ≤1.25 and ≥2.25), in 25.0% by the Forns index (≤4.5 and ≥8.7) and in 30.4% by the APRI (≤0.36 and ≥0.85). Similarly, 35.0%, 23.3%, and 31.6%, respectively, of those in the validation set could have avoided liver biopsy.

Table 4. Diagnostic Accuracy of Three Indices in the Validation Set (n = 120)
  Cutoff ValueSensitivity (%)Specificity (%)Positive Predictive Value (%)Negative Predictive Value (%)Likelihood RatioBiopsy Elimination Rate
APRIF0-F1≤0.3631.698.395.059.06.331.6
 F2-F3≥1.8531.691.779.257.23.8 
Forns indexF0-F1≤4.525.693.378.955.43.723.3
 F2-F3≥8.721.798.392.955.712.8 
FibroIndexF0-F1≤1.2540.096.792.361.712.135.0
 F2-F3≥2.2530.096.790.058.09.1 

Comparison of FibroIndex with Forns Index and APRI in Patients with Normal Serum ALT.

Seventy-three patients in the estimation set had normal serum ALT (<47 IU/L), of which 10 were fibrosis stage F0, 40 were fibrosis stage F1, 15 were fibrosis stage F2, and 8 were fibrosis stage F3. Of the 39 patients in the validation set containing patients at stage F4, 6 were fibrosis stage F0, 16 were fibrosis stage F1, 7 were fibrosis stage F2, 1 was fibrosis stage F3, and 9 were fibrosis stage F4. We compared diagnostic performance in the patients with normal ALT. The AUCs of the FibroIndex, Forns index, and APRI for predicting significant and severe fibrosis in the 73 patients with normal ALT in the estimation set were 0.774 (95% CI, 0.654-0.892) and 0.763 (95% CI, 0.578-0.947), 0.741 (95% CI, 0.620-0.862) and 0.737 (95% CI, 0.551-0.923), and 0.715 (95% CI, 0.595-0.835) and 0.639 (95% CI, 0.439-0.839), respectively (Fig. 6). Similarly, AUCs in the 39 patients with normal ALT in the validation set containing F4 were 0.864 (95% CI, 0.743-0.984) and 0.931 (95% CI, 0.847-1.015), 0.813 (95% CI, 0.666-0.960) and 0.903 (95% CI, 0.793-1.014), and 0.882 (95% CI, 0.762-1.003) and 0.922 (95% CI, 0.822-1.023), respectively (Fig. 7).

thumbnail image

Figure 6. ROC curves of the APRI, Forns index, and FibroIndex in predicting (A) significant fibrosis (F2-F3) and (B) severe fibrosis (F3) in 73 patients with normal ALT of the estimation set. The areas under the ROC curves of the APRI, Forns index, and FibroIndex were 0.72, 0.74 and 0.77, respectively, in predicting F2-F3 and 0.64, 0.74, and 0.76, respectively, in predicting F3.

Download figure to PowerPoint

thumbnail image

Figure 7. ROC curves of the APRI, Forns index, and FibroIndex in predicting (A) significant fibrosis (F2-F4) and (B) severe fibrosis (F3-F4) in 39 patients with normal ALT of the validation set containing F4. The areas under the ROC curves of the APRI, Forns index, and FibroIndex were 0.88, 0.81, and 0.86, respectively, in predicting F2-F4 and 0.92, 0.90, and 0.93, respectively, in predicting F3-F4.

Download figure to PowerPoint

Longitudinal Validation of FibroIndex.

We evaluated whether these indices varied in parallel with longitudinal variation in histological fibrosis. Thirty patients were enrolled (Table 5). These patients were treated by IFN-alpha monotherapy for 6 months and underwent liver biopsy twice, before and at least 1 year after IFN therapy. Subjects then were divided into 3 groups on the basis of changes in fibrosis stage, that is, the deteriorated group (1 or more increases in fibrosis stage compared to at pretreatment biopsy), the unchanged group (the same fibrosis stage at times of pretreatment and posttreatment biopsies), or the improved group (1 or more decreases in fibrosis stage compared to at pretreatment biopsy). We compared the alterations in these indices among these 3 groups. Alterations in the FibroIndex were 0.24 ± 0.55 in the deteriorated group, 0.04 ± 0.36 in the unchanged group, and −0.47 ± 0.45 in the improved group (F = 5.6, P = 0.009). There were significant differences among the groups. However, alterations in the APRI and Forns index were 0.59 ± 1.97 and 1.40 ± 2.68 in the deteriorated group, 0.21 ± 1.15 and 0.42 ± 0.97 in the unchanged group, and −0.34 ± 0.94 and −0.19 ± 1.17 in the improved group (F = 1.51, P = 0.239; and F = 2.26, P = 0.123), respectively. There were no significant differences among the groups in either of these indices.

Table 5. Characteristics of 30 Patients in the Longitudinal Set
CharacteristicValue
  1. Abbreviations: NR, nonresponder; TR, transient responder; SR, sustained responder.

Age (years)51 ± 9
Male/Female24/6
Fibrosis stage at first biopsy 
 F18
 F27
 F315
Effect of IFN therapy 
 NR8
 TR12
 SR10
The interval of biopsies (months)49 ± 36
Alteration of fibrosis stage 
 Deteriorated5
 Unchanged11
 Improved14

Longitudinal variations in the 3 indices according to the variation in histological stage of fibrosis are shown in Fig. 8. There was a significant decrease in the FibroIndex (from 1.82 ± 0.45 at baseline to 1.35 ± 0.56 at second biopsy, P = 0.0043) in the improved group and a significant increase in the FibroIndex in the deteriorated group (from 1.70 ± 0.66 at baseline to 2.09 ± 0.81 at second biopsy, P = 0.043). Although a significant decrease was observed for the APRI in the improved group (from 1.24 ± 0.77 to 0.69 ± 0.87, P = 0.043), no significant difference for the APRI from baseline was found in the deteriorated group (from 1.29 ± 0.97 to 2.22 ± 1.67, P = 0.225) or for the Forns index (from 7.15 ± 1.74 to 6.96 ± 2.11, P = 0.435) in either the improved group (from 7.33 ± 0.87 to 8.66 ± 2.31, P = 0.225) or the deteriorated group.

thumbnail image

Figure 8. Variations between the first and second biopsies of the deteriorated and improved groups in the FibroIndex, Forns index and APRI.

Download figure to PowerPoint

Furthermore, we investigated the relation between variation in histological stage of fibrosis with variation in the 3 indices. Changes in histological fibrosis stage correlated with changes in the FibroIndex (Spearman r = 0.500, P = 0.0072), but not with changes in the APRI and or in the Forns index (Spearman r = 0.244, P = 0.190; Spearman r = 0.361, P = 0.052).

Discussion

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References

In this study, we created a simple equation called the FibroIndex, using platelet count, AST, and serum gamma globulin. These parameters are not related to one another, which means there is redundancy, and different abnormalities are being explored. Previous studies also confirmed AST and platelet count as independent predictors.6, 8, 9, 17 The APRI was the simplest method and showed an accuracy comparable to that of other methods. The Forns index measures GGT and total cholesterol, which are affected by medication, drinking, and diet. Serum gamma globulin is associated with liver fibrosis and portosystemic shunts.18 Imbert-Bismut et al.7 reported that serum gamma globulin was higher in patients with scores of F2 or F3 than in those with scores of F0 or F1. Ikeda et al.19 reported it was a significant parameter in differentiating cirrhosis from chronic hepatitis. Some patients with chronic hepatitis C are known to have high titers of autoantibodies along with hypergammaglobulinemia. Serum gamma globulin concentration reflects chronic inflammation in the liver and autoimmune phenomenon, which lead to hepatic fibrosis. Because testing for AST, platelet count, and gamma globulin is routine in most hospitals and laboratories, their measurements are reliable and sufficiently accurate. Therefore, the FibroIndex would be available to any laboratory.

The present study showed that the FibroIndex was more accurate in predicting significant fibrosis (both F ≥ 2 and F ≥ 3) with an AUC of 0.826, 0.813 in the validation set, than either the APRI (0.776, 0.772) or the Forns index (0.778, 0.760), respectively. Furthermore, the FibroIndex (0.864 for F ≥ 2 and 0.848 for F ≥ 3) also was better than the APRI (0.820, 0.810) or the Forns index (0.835, 0.831) in the validation set containing patients at stage F4. The AUCs of the APRI and the Forns index for predicting significant fibrosis in this study were almost equal to those (APRI, 0.88; Forns index, 0.81) in their respective original reports.8, 9 Therefore, we are confident that our study samples are appropriate. Inclusion of many patients with obvious cirrhosis would have artificially improved the predictive value. Moreover, we included 38 patients without fibrosis (F0), who usually would not undergo treatment.

Recently, there has been a debate about how to manage patients with normal ALT. Most published studies20 have concluded that two thirds of HCV patients with normal ALT had no or minimal fibrosis. However, that left one third of patients with significant fibrosis (≥F2). The present study showed that the FibroIndex performed well in the prediction of significant fibrosis not only in patients with elevated ALT have but also in patients with normal ALT. Therefore, candidates for antiviral treatments can be selected at least in part by results of the FibroIndex.

Several fibrotic markers, such as serum hyaluronic acid, type IV collagen, type IV collagen 7s domain, and P III P, have been reported to be useful for diagnosing liver fibrosis and cirrhosis.10, 21–24 However, their measurement is less standardized and expensive. Although serum hyaluronic acid, which has been reported to be the most useful of these fibrotic markers, has considerable diagnostic value for the diagnosis of cirrhosis, its diagnostic value for stage F2-F3 fibrosis is not significantly higher than the FibroIndex.10 The weakness of hyaluronic acid is affected by diet25 and history of gastrectomy.26 Therefore, the Fibrometer,27 Rosenberg's discriminant score,11 and Patel's 3-marker panel,28 which include hyaluronic acid, have the same problem.

FibroIndex had high specificity and positive predictive value in identifying patients with significant or severe fibrosis. However, its sensitivity was limited and not sufficient to recruit patients who needed treatment. One cause of its limited sensitivity may be variation in the laboratory test results. Because laboratory test results differ from day to day, even in a limited range and transient exacerbation of chronic hepatitis C sometimes occurs, these indices can under- or overestimate fibrosis stage. Therefore, it is recommended that laboratory test means of multiple measurements or for a certain period be used.

The present study evaluated not only cross-sectional estimation but also longitudinal variation. Variations in the FibroIndex more closely paralleled variations in fibrosis stage after IFN therapy than did the Forns index or the APRI. Because the Forns index includes the parameter of patient age, its score increases according to patient age and therefore may not accurately reflect improvement in fibrosis after treatment. Several previous studies29, 30 showed significant concordance between the Fibrotest and variation in fibrosis stage. Recently, primary therapy for chronic hepatitis C has been a combination of Peg-IFN and ribavirin. Because ribavirin can induce hemolysis and therefore a decrease in haptoglobin and an increase in indirect bilirubin, the Fibrotest, which includes measurement of haptoglobin and total bilirubin, could overestimate fibrosis stage. Similarly, the FibroIndex, Forns index, and APRI all include measurement of platelet count. Because a decrease in platelet count during IFN therapy is a well-known side effect, these indices also can overestimate fibrosis stage. These indices would not be useful during IFN or combination treatment. However, because these side effects disappear immediately after the end of treatment, these indices could be used after treatment as surrogate markers of antifibrotic effect. Using the FibroIndex should decrease the number of liver biopsies necessary during follow-up of patients with hepatitis C and could safely provide longitudinal data on the progression of liver fibrosis. Although the number of patients in our longitudinal study was still small, and study of a large number of patients may be necessary, this simple index is very useful for providing important information on the clinical course of patients with hepatitis C and for evaluating the antifibrotic effect of several hepatitis C treatments.

In conclusion, we have created and validated a simple and reliable biochemical fibrosis index, the FibroIndex, and demonstrated that this index has greater diagnostic value than either the APRI or the Forns index. Furthermore, it could be used as a surrogate marker to determine the effects of hepatitis C treatment on change in fibrosis stage.

References

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References
  • 1
    Perrault J, McGill DB, Ott BJ, Taylor WF. Liver biopsy: complications in 1000 inpatients and outpatients. Gastroenterology 1978; 74: 103106.
  • 2
    Wong JB, Koff RS for the International Hepatitis Interventional Therapy Group. Watchful waiting with periodic liver biopsy versus immediate empirical therapy for histologically mild chronic hepatitis C. Ann Intern Med 2000; 133: 665675.
  • 3
    The French METAVIR Cooperative Study Group. Intraobserver and interobserver variations in liver biopsy interpretation in patients with chronic hepatitis C. HEPATOLOGY 1994; 20: 1520.
  • 4
    Rousselet MC, Michalak S, Dupré F, Croué A, Bedossa P, Saint-André JP, et al.; Hepatitis Network 49. Sources of variability in histological scoring of chronic viral hepatitis. HEPATOLOGY 2005; 41: 257264.
  • 5
    Poynard T, Bedossa P, METAVIR and CLINIVIR cooperative study groups. Age and platelet count: a simple index for predicting the presence of histological lesions in patients with antibodies to hepatitis C virus. J Viral Hepat 1997; 4: 199208.
  • 6
    Ono E, Shiratori Y, Okudaira T, Imamura M, Teratani T, Kanai F, Kato N, et al. Platelet count reflects stage of chronic hepatitis C. Hepatol Res 1999; 15: 192200.
  • 7
    Imbert-Bismut F, Ratziu V, Pieroni L, Charlotte F, Benhamou Y, Poynard T; MULTIVIRC Group. Biochemical markers of liver fibrosis in patients with hepatitis C virus infection: a prospective study. Lancet 2001; 357: 10691075.
  • 8
    Forns X, Ampurdanès S, Llovet JM, Aponte J, Quintó L, Martínez-Bauer E, et al. Identification of chronic hepatitis C patients without hepatic fibrosis by a simple predictive model. HEPATOLOGY 2002; 36: 986992.
  • 9
    Wai CT, Greenson JK, Fontana RJ, Kalbfleisch JD, Marrero JA, Conjeevaram HS, et al. A simple noninvasive index can predict both significant fibrosis and cirrhosis in patients with chronic hepatitis C. HEPATOLOGY 2003; 38: 518526.
  • 10
    Murawaki Y, Ikuta Y, Okamoto K, Koda M, Kawasaki H. Diagnostic value of serum markers of connective tissue turnover for predicting histological staging and grading in patients with chronic hepatitis C. J Gastroenterol 2001; 36: 399406.
  • 11
    Rosenberg WM, Voelker M, Thiel R, Becka M, Burt A, Schuppan D, et al. Serum markers detect the presence of liver fibrosis: a cohort study. Gastroenterology 2004; 127: 17041713.
  • 12
    Colli A, Fraquelli M, Andreoletti M, Marino B, Zuccoli E, Conte D. Severe liver fibrosis or cirrhosis: accuracy of US for detection—analysis of 300 cases. Radiology 2003; 227: 8994.
  • 13
    Hirata M, Akbar SM, Horiike N, Onji M. Noninvasive diagnosis of the degree of hepatic fibrosis using ultrasonography in patients with chronic liver disease due to hepatitis C virus. Eur J Clin Invest 2001; 31: 528535.
  • 14
    Castéra L, Vergniol J, Foucher J, Le Bail B, Chanteloup E, Haaser M, et al. Prospective comparison of transient elastography, Fibrotest, APRI, and liver biopsy for the assessment of fibrosis in chronic hepatitis C. Gastroenterology 2005; 128: 343350.
  • 15
    Ziol M, Handra-Luca A, Kettaneh A, Christidis C, Mal F, Kazemi F, et al. Noninvasive assessment of liver fibrosis by measurement of stiffness in patients with chronic hepatitis C. HEPATOLOGY 2005; 41: 4854.
  • 16
    Bedossa P, Poynard T for The METAVIR Cooperative Study Group. An algorithm for the grading of activity in chronic hepatitis C. HEPATOLOGY 1996; 24: 289293.
  • 17
    Attallah AM, Shiha GE, Omran MM, Zalata KR. A discriminant score based on four routine laboratory blood tests for accurate diagnosis of severe fibrosis and/or liver cirrhosis in Egyptian patients with chronic hepatitis C. Hepatol Res 2006; 34: 163169.
  • 18
    Sakumoto I, Kikuchi N, Takane T, Yoshida H, Hosaka H. Studies on the relationship between the histological changes and liver cell function or plasma proteins in hepatic diseases. Gastroenterol Jpn 1976; 11: 224236.
  • 19
    Ikeda K, Saitoh S, Kobayashi M, Suzuki Y, Tsubota A, Suzuki F, et al. Distinction between chronic hepatitis and liver cirrhosis in patients with hepatitis C virus infection. Practical discriminant function using common laboratory data. Hepatol Res 2000; 18: 252266.
  • 20
    Alberti A. Towards more individualized management of hepatitis C virus patients with initially or persistently normal alanineaminotransferase levels. J Hepatol 2005; 42: 266274.
  • 21
    Murawaki Y, Ikuta Y, Koda M, Nishimura Y, Kawasaki H. Clinical significance of serum hyaluronan in patients with chronic viral liver disease. J Gastroenterol Hepatol 1996; 11: 45965.
  • 22
    Murawaki Y, Ikuta Y, Koda M, Kawasaki H. Serum type III procollagen peptide, type IV collagen 7S domain, central triple-helix of type IV collagen and tissue inhibitor of metalloproteinases in patients with chronic viral liver disease: relationship to liver histology. HEPATOLOGY 1994; 20: 780787.
  • 23
    Murawaki Y, Ikuta Y, Nishimura Y, Koda M, Kawasaki H. Serum markers for connective tissue turnover in patients with chronic hepatitis B and chronic hepatitis C: a comparative analysis. J Hepatol 1995; 23: 145152.
  • 24
    Gabrielli GB, Capra F, Casaril M, Squarzoni S, Tognella P, Dagradi R, et al. Serum laminin and type III procollagen in chronic hepatitis C. Diagnostic value in the assessment of disease activity and fibrosis. Clin Chim Acta 1997; 265: 2131.
  • 25
    Idobe Y, Murawaki Y, Ikuta Y, Koda M, Kawasaki H. Post-prandial serum hyaluronan concentration in patients with chronic liver disease. Intern Med 1998; 37: 568575.
  • 26
    Murawaki Y, Idobe Y, Ikuta Y, Koda M, Yamamoto H, Maruyama S, et al. Influence of a history of gastrectomy for gastriccancer on serum hyaluronan concentration in normal individuals and patients with chronic liver disease. Hepatol Res 1998; 10: 248254.
  • 27
    Calès P, Oberti F, Michalak S, Hubert-Fouchard I, Rousselet MC, Konaté A, et al. A novel panel of blood markers to assess the degree of liver fibrosis. HEPATOLOGY 2005; 42: 13731381.
  • 28
    Patel K, Gordon SC, Jacobson I, Hézode C, Oh E, Smith KM, et al. Evaluation of a panel of non-invasive serum markers to differentiate mild from moderate-to-advanced liver fibrosis in chronic hepatitis C patients. J Hepatol 2004; 41: 935942.
  • 29
    Poynard T, Imbert-Bismut F, Ratziu V, Chevret S, Jardel C, Moussalli J, et al. Biochemical markers of liver fibrosis in patients infected by hepatitis C virus: longitudinal validation in a randomized trial. J Viral Hepat 2002; 9: 128133.
  • 30
    Poynard T, McHutchison J, Manns M, Myers RP, Albrecht J. Biochemical surrogate markers of liver fibrosis and activity in a randomized trial of peginterferon alfa-2b and ribavirin. HEPATOLOGY 2003; 38: 481492.