Diagnostic accuracy, reproducibility and robustness of fibrosis blood tests in chronic hepatitis C: A meta-analysis with individual data
Introduction
Several blood tests have been proposed to evaluate fibrosis in liver diseases [1]. They may have strong clinical applications, such as initial staging of fibrosis, assessment of anti-fibrotic treatments, prognosis indicators and general screening [2], [3], [4]. Blood tests are composed of direct [2] or indirect [3] markers of fibrosis depending on their relationship with the connective tissue. Some tests such as the aspartate aminotransferase (AST) to platelet ratio index (APRI) are simple and easily available in clinical practice [4]. Other tests are more complex and were constructed from multivariate regression analyses. They variably include several indirect markers – Fibrotest – [5], direct markers – European liver fibrosis test – [6] or both direct and indirect markers – FibroMeter, and Hepascore – [7]. Only a minority of tests including APRI, Fibrotest and more recently FibroMeter and Hepascore have so far been evaluated by external and independent analysis and they have rarely been compared to one another [8], [9], [10]. In addition, their diagnostic performance was evaluated using small sample sizes, straightforward and limited statistical tools such as the area under the receiver operating characteristic (AUROC), whereas more accurate and stringent indexes of performance could be useful to precisely define their clinical application [11].
Therefore, we conducted a meta-analysis with individual data aiming primarily at comprehensively characterizing and comparing the diagnostic accuracy of four major blood tests of fibrosis including FibroMeter, Hepascore, APRI and Fibrotest in patients with chronic viral hepatitis C. The secondary aims were to evaluate the factors influencing this diagnostic accuracy, especially the reproducibility (center effect) and robustness (population effect) by the means of new statistical methods.
Section snippets
Data source
We systematically reviewed the literature from 1997 to 2007 for studies comparing FibroMeter and Fibrotest in patients with chronic viral hepatitis C in the Medline data base and by hand searching. Hepascore and APRI, whose markers are included in the two former tests, could thus be calculated. Three independent publications were retrieved. The first study including one center, Angers, was the original publication that described the FibroMeter [7]. The second study included two independent
Characteristics of patients
The main characteristics of the 825 patients are presented in Table 2. Some significant differences were observed between centers. The most important one concerned the prevalence of histological fibrosis stages. The prevalence of severe fibrosis ranged from 9% to 30%, and the prevalence of cirrhosis ranged from 1% to 19%. Biopsy length also was significantly different among centers. Therefore, comparisons between centers were adjusted on these variables.
Overall performance of blood tests
The main indices of diagnostic accuracy
Performance indexes
We focused in this multicenter study on four tests of liver fibrosis, namely FibroMeter, Fibrotest, Hepascore and APRI. Other easily available tests such as Forns' index [22] and Fibroindex [23] were not evaluated. The performance of Forns' index was indeed found to be inferior to Fibrotest or FibroMeter in several studies [7], [10], [23] including 3 centers involved in the present study. The interlaboratory reproducibility of gammaglobulins included in Fibroindex is not optimal in our
Acknowledgments
We thank other investigators from Angers: Frédéric Oberti, Sophie Michalak, Isabelle Hubert-Fouchard, Anselme Konaté, Catherine Ternisien, Alain Chevailler, Françoise Lunel, Dermot O'Toole; PACA: Denis Ouzan, Albert Tran, Danielle Botta, Christophe Renou, Christophe Sattonnet, Marie-Christine Saint-Paul, Thierry Benderitter, Stéphane Garcia, Henri Pierre Bonneau ; Grenoble: Marie-Noelle Hilleret, Jean-Pierre Zarski, Patrice Faure, Jean-Charles Renversez, Francoise Morel; Clichy: Claude Degott,
References (34)
- et al.
A simple noninvasive index can predict both significant fibrosis and cirrhosis in patients with chronic hepatitis C
Hepatology
(2003) - et al.
Biochemical markers of liver fibrosis in patients with hepatitis C virus infection: a prospective study
Lancet
(2001) - et al.
Serum markers detect the presence of liver fibrosis: a cohort study
Gastroenterology
(2004) - et al.
Comparison of test performance profile for blood tests of liver fibrosis in chronic hepatitis C
J. Hepatol.
(2007) - et al.
Prospective comparison of six non-invasive scores for the diagnosis of liver fibrosis in chronic hepatitis C
J. Hepatol.
(2007) - et al.
Performance of serum marker panels for liver fibrosis in chronic hepatitis C
J. Hepatol.
(2006) - et al.
Prospective comparison of transient elastography, Fibrotest, APRI, and liver biopsy for the assessment of fibrosis in chronic hepatitis C
Gastroenterology
(2005) - et al.
Reproducibility of blood tests of liver fibrosis in clinical practice
Clin. Biochem.
(2008) - et al.
The diagnostic odds ratio: a single indicator of test performance
J. Clin. Epidemiol.
(2003) - et al.
Impact of liver biopsy size on histological evaluation of chronic viral hepatitis: the smaller the sample, the milder the disease
J. Hepatol.
(2003)