Background and aims: The ability of noninvasive methods to predict the development of cirrhosis has not been established. We evaluated the ability of three noninvasive methods [the Forns index, the aspartate aminotransferase-to-platelet ratio index (APRI), and the Non-Invasive Hepatitis-C-related Cirrhosis Early Detection (NIHCED) score] to determine the risk of developing cirrhosis in chronic hepatitis C.
Methods: Consecutive patients with chronic hepatitis C who had undergone liver biopsy between 1998 and 2004 were eligible. We used the three methods to evaluate patients at baseline and at follow-up (4-10 years later). When these methods yielded discordant or indeterminate results, a second liver biopsy was performed. Logistic regression models were fitted for each method to predict whether cirrhosis would appear and to predict long-term mortality from cirrhosis.
Results: We included 289 patients in our study. The mean scores at baseline and at follow-up, respectively, were as follows: Forns, 5.47 ± 1.95 and 6.56 ± 2.02; APRI, 1.1 ± 2.33 and 1.4 ± 1.53; and NIHCED, 7.79 ± 11.45 and 15.48 ± 15.28. The area under the receiver operating characteristic curve for predicting cirrhosis was 0.83 for Forns, 0.79 for APRI, and 0.76 for NIHCED. The sensitivity and specificity for predicting cirrhosis, respectively, were 75 and 71% for Forns (cutoff 4.7), 86 and 42% for APRI (cutoff 0.48), and 41 and 82% for NIHCED (cutoff 0). The area under the receiver operating characteristic curve for predicting mortality was 0.86 for Forns, 0.79 for APRI, and 0.84 for NIHCED.
Conclusion: Indirect noninvasive markers could help identify patients with chronic hepatitis C at risk of progression to cirrhosis.