Accuracy of international guidelines for identifying significant fibrosis in hepatitis B e antigen--negative patients with chronic hepatitis

Clin Gastroenterol Hepatol. 2013 Nov;11(11):1493-1499.e2. doi: 10.1016/j.cgh.2013.05.038. Epub 2013 Jun 28.

Abstract

Background & aims: Differing threshold levels of hepatitis B virus (HBV) DNA and alanine aminotransferase (ALT) are recommended by international guidelines for commencement of antiviral therapy. These guidelines advocate therapy for patients with significant fibrosis (METAVIR score ≥F2); we assessed the accuracy of these guideline-defined thresholds in identifying patients with ≥F2 fibrosis.

Methods: We applied the European (European Association for the Study of the Liver [EASL] 2012), Asian-Pacific (Asian-Pacific Association for the Study of the Liver [APASL] 2012), American (American Association for the Study of Liver Diseases [AASLD] 2009), and United States Panel Algorithm (USPA 2008) criteria to 366 consecutive hepatitis B e antigen-negative patients with liver biopsy samples: EASL, ALT >laboratory-defined upper limit of normal (ULN) and HBV DNA ≥2000 IU/mL (n = 171); APASL, ALT >2-fold laboratory-defined ULN and HBV DNA ≥2000 IU/mL (n = 87); AASLD, ALT >2-fold the updated ULN (0.5-fold ULN [corresponding to ≤19 U/L] for women and 0.75-fold the ULN [corresponding to ≤30 U/L] for men) and HBV DNA ≥20,000 IU/mL (n = 53); and USPA, ALT >updated ULN (>0.5-fold ULN for women and >0.75-fold ULN for men) and HBV DNA ≥2000 IU/mL (n = 173).

Results: Overall, 113 patients (30.9%) had ≥F2 fibrosis, which was more frequent among patients who fulfilled any guideline criteria (45.7% vs 17.9% for those who did not fulfill any criteria, P < .0001). In applying the EASL, AASLD, APASL, and USPA criteria, sensitivity and specificity values for detection of ≥F2 fibrosis were 45.6%, 58.5%, 56.3%, and 45.7% (P = .145) and 82.1%, 73.8%, 77.1%, and 82.4% (P = .366), respectively. The EASL criteria (area under the receiver operating characteristic [AUROC] curve, 0.66; 95% confidence interval [CI], 0.61-0.71) and USPA criteria (AUROC, 0.66; 95% CI, 0.58-0.73) performed better than APASL (AUROC, 0.64; 95% CI, 0.59-0.69; P = .421) and significantly better than the AASLD criteria (AUROC, 0.59; 95% CI, 0.54-0.64; P = .013).

Conclusions: In hepatitis B e antigen-negative patients with chronic hepatitis, the EASL, AASLD, APASL, and USPA criteria identify patients with ≥F2 fibrosis with low levels of accuracy. However, the EASL and USPA criteria are the most accurate for identification of these patients.

Keywords: AASLD; AFP; ALT; APASL; AST; AUROC; Alanine Aminotransferase; American Association for the Study of Liver Diseases; Asian-Pacific Association for the Study of the Liver; CI; Diagnosis; EASL; European Association for the Study of the Liver; Fibrosis; HBV; HBV DNA; HBeAg; HBsAg; Hepatitis B; LR; Liver Disease; Management; NPV; PPV; Sensitivity; Specificity; ULN; USPA; United States Panel Algorithm; alanine aminotransferase; area under the receiver operating characteristic; aspartate aminotransferase; confidence interval; hepatitis B e antigen; hepatitis B surface antigen; hepatitis B virus; likelihood ratio; negative predictive value; positive predictive value; upper limit of normal; α-fetoprotein.

Publication types

  • Evaluation Study

MeSH terms

  • Adult
  • Alanine Transaminase / blood*
  • Biopsy
  • DNA, Viral / blood*
  • Female
  • Health Services Research*
  • Hepatitis B e Antigens / blood*
  • Hepatitis B, Chronic / complications*
  • Humans
  • Liver / pathology
  • Liver Cirrhosis / diagnosis*
  • Liver Cirrhosis / pathology
  • Male
  • Middle Aged
  • Practice Guidelines as Topic*

Substances

  • DNA, Viral
  • Hepatitis B e Antigens
  • Alanine Transaminase