Validation of simplified diagnostic criteria for autoimmune hepatitis in Italian patients


  • Potential conflict of interest: Nothing to report.

Validation of Simplified Diagnostic Criteria for Autoimmune Hepatitis in Italian Patients

To the Editor:

A simplified score for the diagnosis of autoimmune hepatitis (AIH) using only four criteria has been recently published by the International Autoimmune Hepatitis Group on Hepatology1; the included paramenters were represented by autoantibodies, immunoglobulin G, histology, and exclusion of viral hepatitis (Table 1). The authors reached 88% sensitivity and 97% specificity with a score ≥ 6 and 81% sensitivity and 99% specificity with a score ≥ 7.

Table 1. Simplified Diagnostic Criteria for Autoimmune Hepatitis1
  1. AIH, autoimmune hepatitis; ANA, antinuclear antibody; LKM1: Liver Kidney Microsome type 1; SLA, soluble liver antigen; SMA, smooth muscle antibody; ULN, upper limit of normal.

ANA or SMA≥ 1:401
ANA or SMA≥ 1:802
or LKM1≥ 1:402
or SLAPositive2
 ≥ 1:10 × ULN2
Liver histologyCompatible with AIH1
 Typical of AIH2
Absence of viral hepatitisYES2
  ≥ 6 probable AIH
  ≥ 7 definite AIH

The aim of our study was to test this simplified score in our Italian AIH series and controls. Our study population was represented by 132 well-characterized Italian patients with AIH who were diagnosed on the basis of the revised score2 and 15 patients with a diagnosis of AIH/primary biliary cirrhosis (PBC) overlap syndrome (AIH/PBC OS) diagnosed in accordance with the reported criteria3; as pathological controls, we used 120 patients with “pure” PBC4 and 143 patients who were positive for hepatitis C virus (HCV), 38 of whom had a seropositivity for Liver Kidney Microsome type 1 (LKM1) autoantibody.

A total of 121 of 132 (91.6%) Italian patients with AIH reached a score ≥ 6 and 114 of 132 (86.3%) ≥ 7.

Among the 15 patients with AIH/PBC OS, 14 of them (93.3%) had a score ≥ 6 and 12 of 15 (80%) had a score ≥ 7. In the HCV series, we found six patients (all positive for LKM1) of 143 (4%) who reached a score ≥ 6 and none with ≥ 7; when we considered patients with “pure” PBC, we observed nine of 120 (8%) with a score ≥ 6 and one (1%) with a score ≥ 7.

The overall sensitivity and specificity for the AIH diagnosis with a score ≥ 6 was 91.8% (95% confidence interval [CI]: 87.9-94.6) and 94.3 (95% CI: 92.1-95.9), with a positive and negative likelihood ratio of 16.1 (95% CI: 11.1-22.8) and 0.08 (95% CI: 0.05-0.13), respectively.

When we considered a score ≥ 7, the overall sensitivity and specificity was 87.1% (95% CI: 84.5-87.6) and 99.6% (95% CI: 98.2-99.9), with a positive and negative likelihood ratio of 229 (95% CI: 46-1294) and 0.13 (95% CI: 0.12-0.15), respectively.

The diagnosis of AIH is often difficult and needs the evaluation of several parameters not always available at the bedside of the patient. The proposed simplified score is easy to apply with high sensitivity and specificity; in particular, in our experience, a score ≥ 7 allows a raising of the specificity without a significant loss of sensitivity. Furthermore, this simplified score seems to be efficacious also in the diagnosis of AIH/PBC OS starting from patients with PBC who have signs of AIH.

As expected, we found some patients (six) with the coexistence of HCV positivity and LKM1 seropositivity who satisfied the score for the diagnosis of AIH; in these cases, determining if the viral or autoimmune component is predominant or an overlap of viral/autoimmune syndrome is present is difficult. The score can help the physician, but the therapeutical approach should be tailored for each patient.

Paolo Muratori*, Alessandro Granito*, Georgios Pappas*, Luigi Muratori*, * Department of Clinical Medicine, Alma Mater Studiorum University of Bologna, S. Orsola-Malpighi Hospital, Bologna, Italy.