Validation of noninvasive methods for the assessment of liver fibrosis in patients with recurrent hepatitis C after transplantation

Much effort is being devoted to finding noninvasive methods for assessing fibrosis in transplant patients with recurrent hepatitis C virus (HCV) infections. This would help us to avoid some biopsy procedures, monitor fibrosis progression more strictly during posttransplant follow-up, and better predict graft prognosis and the right time for starting antiviral therapies. A prospective validation of a noninvasive index based on laboratory data [the Hospital Universitario La Fe (HULF) index] failed to reproduce the high accuracy in identifying significant fibrosis found in the initial retrospective analysis from which the HULF index was initially developed [the area under the receiver operating characteristic curve (AUROC) was 0.68 in the new validation study1 versus 0.80 and 0.84 in the training and testing groups, respectively, in the initial study2]. Previously, we prospectively tested various noninvasive indexes for the detection of significant fibrosis, including the HULF index, and transient elastography. Our findings showed good results with the HULF index [AUROC = 0.799, 95% confidence interval (CI) = 0.626-0.973] but significantly better results with transient elastography (AUROC = 0.943, 95% CI = 0.846-0.987).3 Recently, an English collaborative group proposed a new noninvasive index based on laboratory variables, the London Transplant Centres (LTC) score, that easily estimates the presence of significant fibrosis in patients with recurrent HCV infections.4 This index was derived retrospectively from their patient series and provided an AUROC of 0.86 for METAVIR fibrosis ≥ F2. Because all variables required to calculate the LTC score were available in our prospectively collected database,3 we calculated the AUROC of the LTC score for the diagnosis of METAVIR fibrosis ≥ F2 in patients included in our recent work. The AUROC calculation and comparison with other indexes were made with MedCalc 9 (MedCalc Software, Mariakerke, Belgium). The LTC score showed an AUROC of 0.761 (95% CI = 0.587-0.888) with a best cutoff of 8.25 (not very dissimilar to the value of 6.40 found by Cross and coworkers4), a sensitivity of 83.3%, a specificity of 60.9%, positive and negative likelihood ratios of 2.13 and 0.27, respectively, and positive and negative predictive values of 52.6% and 87.5%, respectively. The AUROC of the LTC score was not statistically significantly different from the AUROC of the HULF index but was significantly inferior to that of transient elastography (P < 0.02), as the AUROC of the HULF index also was (P < 0.05). Poynard and coworkers5 recommended that, when AUROCs derived from different populations are being compared, they should be adjusted for the prevalence and extent of fibrosis with the difference between the mean fibrosis stage of advanced fibrosis and the mean fibrosis stage of nonadvanced fibrosis (DANA). Our DANA was 1.46, and all AUROCs can be adjusted for this, but this adjustment would not affect the comparison of different AUROC values calculated in the same series, as in the present case; thus, we reported only the original AUROC values. In conclusion, the LTC score performs well, but it does not seem to improve the accuracy of the HULF index in this Italian series of liver transplant patients with recurrent HCV infections. Transient elastography, already shown to be a very valid technique in this setting,3, 6, 7 appears to perform better at identifying significant liver fibrosis than indexes derived from laboratory variables. Fabio Piscaglia M.D., Ph.D.*, Alessandro Cucchetti M.D. , Eleonora Terzi M.D.*, Alice Gianstefani M.D.*, * Divisions of Internal Medicine, St. Orsola-Malpighi University and General Hospital, Bologna, Italy, Divisions of Liver and Multivisceral Transplantation, St. Orsola-Malpighi University and General Hospital, Bologna, Italy.