Background: The reported prevalence of antineutrophil cytoplasmic antibodies (ANCA) in primary sclerosing cholangitis (PSC) varies considerably (26-85%). Part of this may reflect methodological differences but part may reflect the differences in the patient groups analysed. To resolve this issue we compared the sensitivity and specificity of the immunoalkaline phosphatase (IALP) and immunofluorescence (IF) techniques in four different populations.
Method: Sera from four centres were tested blind on alcohol-fixed neutrophils using both techniques.
Patients: USA: 14 PSC, 14 primary biliary cirrhosis (PBC); Sweden: 32 PSC, 3 autoimmune hepatitis (AIH), 14 PBC, 11 chronic liver disease; Norway: 32 PSC, 14 AIH, 13 PBC, 1 hepatitis C. Italy: 8 PSC, 14 PBC, 8 viral hepatitis. Thirty-six normal healthy volunteers from Oxford, together with positive and negative controls, were also tested.
Results: The healthy controls were all ANCA negative. The diagnostic sensitivity and specificity, respectively, of ANCA for PSC using the IALP technique for the different test sera were: USA 71% and 93%, Sweden 66% and 96%, Norway 69% and 46%, Italy 50% and 95%. The diagnostic sensitivity and specificity, respectively, of the IF technique on the same sera were: USA 50% and 86%, Sweden 56% and 86%, Norway 47% and 61%, Italy 50% and 91%. Overall, combining all four groups, detection of ANCA using the IALP technique gave a diagnostic sensitivity of 66% with a specificity of 74% for PSC. In contrast, the IF technique gave an overall diagnostic of only 51% (P = 0.044, compared with IALP) with a specificity of 73%. Although overall the IALP technique was more sensitive than IF, the differences in sensitivity and specificity between the two techniques did not reach statistical significance for any individual group. Furthermore, the small differences in sensitivity between the four groups using either technique were not significant. However, the IALP technique had greater specificity in the US, Swedish and Italian groups compared with the Norwegian group (P < 0.05) whereas no statistically significant differences in specificity were noted between the groups using the IF technique.
Conclusion: This study shows that the IALP method of ANCA detection is at least as sensitive as IF for the serological diagnosis of PSC. Indeed, combining data from all four centres, the IALP technique was significantly more sensitive than IF. We therefore recommend the use of the IALP technique, which is also easier to interpret and does not require the use of a specialist fluorescent microscope. The lack of a wide variation in sensitivity between IALP and IF for any individual patient group reported in this study suggests that the previously reported regional differences in ANCA prevalence in PSC of between 26% and 85% may be patient, related, rather than due to ethnic or methodological differences in ANCA detection, perhaps reflecting possible disease heterogeneity within PSC, or case selection bias. Further studies are needed to investigate this intriguing possibility. Such differences, if confirmed, will need to be taken into account when assessing the use of ANCA as a serological marker of PSC.