Background/aims: Antimitochondrial antibodies (AMAs) are the serological hallmark of primary biliary cirrhosis (PBC). We evaluated the sensitivity and specificity of a new M2 enhanced performance enzyme-linked immunosorbent assay (ELISA) (MIT3) for the detection of IgG- and IgA-specific isotypes of AMA in PBC patients including a number of PBC patients negative for AMA by indirect immunofluorescence (IIF) as well as in patients with diverse, non-PBC disorders. We also investigated the clinical significance of IgG and IgA AMA in PBC.
Methods: One hundred and three Greek PBC patients including 27 with AMA IIF-negative at the time of the investigation, 29 with autoimmune hepatitis-1 (AIH-1), 12 with primary sclerosing cholangitis (PSC), 26 with hepatitis C virus (HCV), 15 with hepatitis B virus (HBV), and 29 healthy were investigated for AMA (IgG and IgA) using the MIT3-based ELISAs (INOVA Diagnostics, San Diego, CA). The samples were also tested by conventional anti-M2 ELISA (INOVA Diagnostics, Inc.).
Results: The IgG MIT3-based ELISA significantly increased AMA detection in the cohort of PBC patients, over 26% of whom were AMA IIF-negative, from 63.1% by the conventional anti-M2, and 73.7% by IIF to 79.6% by MIT3-based ELISA (p<0.001). IgA AMAs were detected in 47.6% patients. Overall, IgG/IgA AMAs were detected in 84/103 (81.6%). IgG MIT3-based ELISA detected 12/27 IIF AMA-negative samples (44.4%), while IgG/IgA MIT3-based ELISAs detected 13/27 IIF AMA-negative patients (48.1%). The specificities of MIT3-based ELISAs (IgG and IgA) were 82.8% and 89.7%, respectively, in AIH-1, 100% and 93.3%, respectively, in HBV, 100% in PSC, and 96% and 93.3%, respectively, in HCV. Patients positive for IgG AMA had significantly more severe disease as shown by worse histology and elevated biochemical markers; IgG and IgA AMA titers were associated positively with the Mayo risk score but none of the isotypes were able to predict disease outcome.
Conclusions: The new IgG and IgA MIT3-based ELISAs seem to have higher specificity and sensitivity for AMA detection than IIF and the conventional anti-M2. Interestingly, these assays were able to unmask AMA presence in almost half of the AMA-negative samples by IIF. These findings may suggest the use of MIT3-based ELISAs as first-line investigation for AMA detection, particularly, when the laboratories are unfamiliar with the use and interpretation of the IIF patterns of AMA. The presence of IgG AMA seems to characterize PBC patients with more severe disease, but both IgG and IgA isotypes of AMAs were not predictive markers of disease outcome.