Does noninvasive staging of fibrosis challenge liver biopsy as a gold standard in chronic hepatitis C?

Authors


Does Noninvasive Staging of Fibrosis Challenge Liver Biopsy as a Gold Standard in Chronic Hepatitis C?

To the Editor:

Liver biopsy is viewed as the gold standard for staging fibrosis in chronic hepatitis. An ongoing great search for noninvasive diagnostic tests aims at replacing this inconvenient and costly invasive procedure. Several models have been proposed1–3-; however, most of them require special laboratory parameters that are not available in clinical practice. In a recent article, Wai et al. proposed a simple model on the basis of routinely available laboratory test results (aspartate aminotransferase–to-platelet ratio index, [APRI]), which was shown to predict with high sensitivity and specificity liver fibrosis in patients with hepatitis C virus infection.4 These results were recently challenged by Giannini and Testa, who pointed out that the ratio of aspartate aminotransferase to alanine aminotransferase assessed fibrosis more accurately than the APRI.5

We would like to comment on the applicability and validity of the APRI test, which we have applied to a large cohort of 484 treatment naive patients with chronic hepatitis C (271 males; mean age, 46 ± 0.4; range, 18-68 years). All patients underwent liver biopsy as part of the screening evaluation within the course of national or international clinical trials. Staging of fibrosis was done according to the Scheuer6 score, which categorizes 4 different stages of fibrosis (F0-F4), as compared to the Ishak score used by Wai et al. referring to 6 stages (F0-F6).

Our results are in agreement with the data presented by Wai et al., although the overall sensitivity and specificity of the APRI as well as the positive predictive value and the negative predictive value for the certain cutoffs were found to be lower (Table 1). As mentioned, the different histological scoring system (i.e., Scheuer vs. Ishak score) could be responsible for these differences. Additionally, our results have been evaluated by several pathologists, not by just one as in the Wai et al. study. Interobserver differences certainly could have played a role in this respect, but in clinical practice one is confronted with this kind of situation.

Table 1. Accuracy of the AST-to-Platelet Ratio Index (APRI) in Predicting Significant Fibrosis and Cirrhosis
APRI*All Patients (n = 484) n (%)Actual FibrosisSensitivity (%)Specificity (%)PPV (%)NPV (%)
Stage 0–1 (n = 231) n (%)Stage 2–4 (n = 253) n (%)
For prediction of significant fibrosis       
≤0.50168 (35)122 (53)46 (18)82 (91)53 (47)66 (61)73 (86)
>0.50316 (65)109 (47)207 (82)    
≤1.50375 (77.5)215 (93)160 (63)37 (41)93 (95)85 (88)57 (64)
>1.50109 (22.5)16 (7)93 (37)    
  Actual Fibosis    
Stage 0–2 (n = 357)Stage 3–4 (n = 127)
For prediction of cirrhosis       
≤1.0325 (67)286 (80)39 (31)69 (89)§80 (75)§55 (38)§88 (98)§
>1.0159 (33)71 (20)88 (69)    
≤2.0409 (84.5)332 (93)77 (61)39 (57)§93 (93)§67 (57)§81 (93)§
>2.075 (15.5)25 (7)50 (39)    
  Actual Fibrosis    
Stage 0–3 (n = 422)Stage 4 (n = 62)
  • Abbreviations: AST, aspartate aminotransferase; PPV, positive predictive value; NPV, negative predictive value.

  • *

    Results are given according to the APRI cutoff points proposed by Wai et al. to predict the absence (APRI ≤ 0.50) or presence (APRI >1.50) of significant fibrosis and the absence (APRI ≤1.00) or presence (APRI >2.00) of cirrhosis.

  • Staging of fibrosis was done according to the Scheuer score, which categorizes 4 different stages of fibrosis (F0, absent; F1, mild portal fibrosis without septa; F2, moderate portal fibrosis with few septa; F3, numerous septa [bridging fibrosis] without cirrhosis; F4, cirrhosis), as compared to the Ishak score used by Wai et al. referring to 6 stages (F0–F6). Significant fibrosis was defined as ≥F2. Data for prediction of cirrhosis were given for patients histologically classified as either F3–F4 or F4.

  • Respective data for sensitivity and specificity as well as the positive and negative predictive value assessed in the study by Wai et al. for patients with Ishak fibrosis scores F0–2 vs. F3–6‡ and 5–6 vs. F0–4§ are shown in parentheses.

  • §

    Respective data for sensitivity and specificity as well as the positive and negative predictive value assessed in the study by Wai et al. for patients with Ishak fibrosis scores F0–2 vs. F3–6‡ and 5–6 vs. F0–4§ are shown in parentheses.

≤1.0325 (67)310 (73.5)15 (24)76 (89)§73.5 (75)§30 (38)§95 (98)§
>1.0159 (33)112 (26.5)47 (76)    
≤2.0409 (84.5)377 (89)32 (52)48 (57)§89 (93)§40 (57)§92 (93)§
>2.075 (15.5)45 (11)30 (48)    

From these studies it appears that fibrosis stage assessed by invasive and noninvasive approaches differ to some extent. In this respect, sampling variability and the size of liver biopsies have to be considered as important contributors to false fibrosis staging,7 and they raise the question of whether liver biopsy can still be regarded as the gold standard for fibrosis assessment. Thus, the interpretation of liver biopsy results on the background of a noninvasive fibrosis prediction method should be proven as a rational approach to improve accuracy of fibrosis staging. One should, however, be aware that a reliable noninvasive assessment of the different stages of fibrosis cannot be made in all patients. Data by Wai et al. as well as our data are in accordance, showing that a prediction concerning presence or absence of significant fibrosis was possible only in 57% and 51%, respectively, of the patients examined in both studies.

Finally, the need to confirm histologically the stage of fibrosis is suspect when clinical experience and laboratory data provide all necessary information to judge with great certainty the severity of patients' chronic liver disease.

Acknowledgements

Supported in part by the German BMBF Network of Competence for Viral Hepatitis (Hep Net).

Thomas Berg*, Christoph Sarrazin†, Holger Hinrichsen‡, Peter Buggisch§, Tilman Gerlach¶, Reinhart Zachoval¶, Stefan Zeuzem†, * Medizinische Klinik mit Schwerpunkt Hepatologie und Gastroenterologie, Charité - Universitätsmedizin Berlin, Campus Virchow-Klinikum, Berlin, Germany, † Medizinische Klinik und Poliklinik – Innere Medizin II, Universitätskliniken des Saarlandes, Homburg/Saar, Germany, ‡ Medizinische Universitätsklinik I, Klinik für Allgemeine Innere Medizin, Christian-Albrechts-Universität Kiel, Germany, § Medizinische Kernklinik/Gastroenterologie, Universitätsklinik Eppendorf, Hamburg, Germany, ¶ Medizinische Klinik II, Klinikum Großhadern, Ludwig-Maximilians-Universität, München, Germany.

Ancillary