A Comparison of Four Fibrosis Indexes in Chronic HCV

Development of New Fibrosis-cirrhosis Index (FCI)

Waqar Ahmad; Bushra Ijaz; Fouzia T Javed; Sana Gull; Humera Kausar; Muhammad T Sarwar; Sultan Asad; Imran Shahid; Aleena Sumrin; Saba Khaliq; Shah Jahan; Asim Pervaiz; Sajida Hassan

Disclosures

BMC Gastroenterol. 2011;11 

In This Article

Discussion

Hepatocellular carcinoma and hepatic cirrhosis are consequences of chronic hepatitis C. The mean infection time to onset of cirrhosis is approximately 30 years, but cirrhosis may occur within a range of 10–50 years.[20] Fibrosis and its extension in hepatic tissue is most common evidence of cirrhosis. Several indexes are available to predict cirrhosis but no method or score is available on exclusive basis to diagnose earlier fibrosis stages.

Genotype 3a is the most common one followed by 1a in Pakistan [,[3,21] and[22] and same was also observed in this study. Almost, 86% patients had genotype 3a while remaining 14% had genotype 1a. A recent study also reported high prevalence of genotype 3 in HCC patients in Pakistan.[23] Patients with none or initial stage (F0-F1) of fibrosis showed a remarkable difference of age with advanced stages (F2 and F3) of fibrosis and cirrhosis. Most patients with age more than 40 years showed severe fibrosis and cirrhosis. These results confirmed the previous studies that patients with mild fibrosis stage were younger than the moderate and severe disease grade and stage is independent of gender.[24–26]

Our results showed positive correlation of ALT with APRI and FI, and negative correlation with AAR and platelet, however, no correlation was established between ALT levels with disease severity and fibrosis stages. Our observation is in agreement with previous reports that serum ALT levels do not accurately predict the presence of hepatic liver damage.[27,28] Several authors reported persistently normal ALT levels (< 42 IU/l) in patients with chronic HCV. Almost, 30% of patients with chronic HCV infection reflect steadily normal serum ALT levels,[29–32] however, in our data only 10% (n = 16) patients showed normal ALT levels.

Our data showed gradual increase in serum ALP and bilirubin levels (Table 2) in fibrosis stages when compared to early infection. Both ALP and bilirubin showed strapping significant correlation with disease progression. These results lead them to an important predictor of disease severity. An increased ALP is usually associated with liver metastasis, extraheptic bile obstruction, intraheptic cholestasis, infiltrative liver disease and hepatitis.[33,34] According to Lee et al, elevated serum ALP levels were common in liver abscess patients.[35] High bilirubin levels are associated with liver metastases and liver tumor involvement leading to hepatocellular carcinoma and liver cirrhosis by active or non-active HCV or HBV.[36] Limited literature is available on the role of elevated ALP and bilirubin levels in liver fibrosis stages. However, according to Imbert-Bismut et al.,[37] bilirubin may be used as marker of liver injury, while a change in ALP levels greater than 120 U/L can be indicative of advanced disease progression.[12] These findings suggest that serum ALP and bilirubin may be used as serum markers to assess the disease progression and fibrosis stages in chronic HCV patients.

Many studies supported that platelet count alone may be clinically valuable as a non-invasive serum marker for liver fibrosis and cirrhosis.[38,39] Platelets not only predict fibrosis but also correlate with fibrotic stages.[40–42] Lackner et al,[43] showed high AUROC of 0.89 for predicting cirrhosis at platelet value < 150 × 109/L and AUROC of 0.71 for non-cirrhotic patients at a cutoff value > 150 × 109/L. Our data is also in accordance with these results as platelet count showed high AUROC (≥ 0.900) to differentiate different liver fibrosis stages as given in Table 5 and Figure 4. In our study, platelet count was significantly low in cirrhotic patients. At a cutoff value of platelet, < 100 × 109/L has an AUROC of 0.990 for prediction of cirrhosis with 81% sensitivity and 98% specificity. Ginnani et al, reported platelet < 130 × 109/L for prediction of cirrhosis in HCV patients with 91.1% sensitivity, 88.3% specificity, PPV 81.2% and NPV 94.7%.[44]

We also examined the ability of AAR, APRI, FIB-4 and F-Index for staging liver fibrosis and to differentiate them from cirrhosis. Giannini et al, reported a high diagnostic accuracy of AAR > 1.16 with 81.3% sensitivity and 55.3% specificity for the prediction of cirrhosis.[44] However, AAR was not able to differentiate among liver fibrosis stages in our sample data. At value of > 1.0, AAR has 43% sensitivity and 70% specificity for differentiating fibrosis from cirrhosis (Table 3). This poor performance of AAR is similar to that reported by Lackner et al.[43]

We observed comparatively high APRI (1.24 ± 0.8) and FIB-4 (1.76 ± 1.35) values in F0-F1 patients. The group F0-F1 contains two subgroups, patients with no fibrosis (F0) and with minimal fibrosis (F1). The mean value of APRI and FIB-4 in F0 was 1.04 and 1.21, and in F1 1.39 and 2.17, respectively (Table 2). It is reported that APRI < 0.42 predict mild fibrosis and APRI > 1.2, significant fibrosis in HCV patients with 90% NPV for absence of fibrosis and 91% PPV for fibrosis presence.[45–47] Our results showed that APRI > 1.5 could predict fibrosis with 55% sensitivity, 67% specificity. Moreover, by using same cutoff value of APRI > 1.5 in a recent study by Macias et al,[48] found that it has 28% sensitivity, 92% specificity, 79% PPV and 55% NPV for predicting significant fibrosis, and for absence of fibrosis APRI < 0.5 has 78%, 44%, 59% and 66% sensitivity, specificity, PPV and NPV, respectively.

FIB-4 was developed by Sterling et al in 2006 for diagnosis of fibrosis and cirrhosis in HIV/HCV co-infected patients. We examined this index only for HCV infected patients. A cutoff value of < 1.45 FIB-4 has a NPV for the exclusion of advanced fibrosis of 90%, while a cutoff value > 3.25 has a PPV for the diagnosis of extended fibrosis of 65%.[49] At a cutoff value of < 1.45, Vallet-Pichard observed a high NPV of 94.7% with a sensitivity of 74.3% to exclude severe fibrosis. Where as, for confirming the presence of advanced fibrosis at cutoff value > 3.25, FIB-4 had a PPV of 82.1% with specificity of 98.2%.[18] Our results are not in agreement with Sterling or Vallet-Pichard, as we observed a low NPV (70%) for excluding significant fibrosis, however, we detected a PPV of 83% with specificity of 45% for the presence of advanced fibrosis at cutoff value > 3.25. Trang et al,[50] proposed new cutoff values of FIB-4 ≤ 1.39 for F0-F1 and ≥2.05 for F2-F4 stage in HCV/HIV co infected patients. At these cutoffs, we observed sensitivity 52%, specificity 76%, PPV 63% and NPV 68% for no/minimal fibrosis and 60%, 63%, 68% and 55% for advanced fibrosis, respectively. Although, we observed low statistical values, our results were in accordance to advance stage prediction. The cut off values proposed by Trang et al better predict fibrosis stages in co infected patients and we applied on only HCV infected patients.

Fibrosis index (FI) showed high sensitivity, specificity, PPV, NPV and AUROC for discriminating different fibrosis stages. Ohta developed this simple index in 2006. At cutoff value < 2.1 FI showed sensitivity and specificity for predicting F0–1 stage 66.8% and 78.8% in initial cohort and 68.5% and 63.6% in validation cohort, respectively.[17] At same cutoff, our data showed 100% sensitivity and 58.4% specificity with AUROC 0.939 for the prediction of none/minimal fibrosis. While for predicting cirrhosis in HCV patients, FI value > 3.30 has sensitivity and specificity 67.7% and 75% in initial cohort, and 70.8% and 81% in validation cohort, respectively. However, at this value we observed 33% sensitivity and 100% specificity for predicting cirrhosis (Table 3). We proposed that a new cutoff value of FI > 2.5 can better predict cirrhosis with 95.2% sensitivity and 94% specificity.

The readily available indexes are associated with some limitations like population discrepancy, not able to distinguish all fibrosis stages individually or some primarily developed for co-infected patients. So there is a need to develop a new index that can distinguish minimal fibrosis (F0-F1) from significant (F2-F4) and advanced (F2-F3) from cirrhosis (F4). While considering substantial relationship of routinely applied tests; serum ALP, ALT, AST, Hb level, bilirubin, albumin and platelet count with liver fibrosis stages, we found that four serum markers ALP, bilirubin, albumin and platelet count have high potential to differentiate different fibrosis stages and cirrhosis at given cutoff values (Table 5 and Figure 4). We also observed that combination of these serum markers could better differentiate among fibrosis stages with high sensitivity, specificity, PPV and NPV.

Our newly derived index FCI showed better performance for discriminating between fibrosis stages as compared to AAR, APRI and FI. In initial cohort, the AUROC for predicting F0-F1 stage for FCI was 0.932 when compared to recently used non-invasive serum markers like AAR (AUROC = 0.570),[15] APRI (AUROC = 0.880),[16] FI (AUROC = 0.741),[17] FIB-4 (AUROC = 0.793),[18] Forn's index (AUROC = 0.860),[51] and Fibrotest (AUROC = 0.870).[52] Moreover, FCI (AUROC = 0.996) showed better performance for predicting cirrhosis than above mentioned serum indexes. Although in our study, platelet count showed high AUROC to predict fibrosis stages, systematic literature reviews consistently shown that panel of fibrosis markers are more accurate than single marker. Combination of two or more serum markers in a mathematical algorithm provide better chance of predicting phase of disease progression instead of individual one.[37,53–57] This analysis showed that FCI has tendency to reflect respective fibrosis stages from no/minimal to cirrhosis with great accuracy (Table 5, Figure 5 and 6). However, several studies are needed to verify these results. Secondly, because of poverty and fear of biopsy, we are not yet able to get enough patient data for verification of our FCI results in new cohort.

Comments

3090D553-9492-4563-8681-AD288FA52ACE
Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as:

processing....