Background: Investigation of a 61-year-old Caucasian male suffering from fatigue and weight loss led to the finding of a carbohydrate antigen 19-9 (CA 19-9) concentration of 80 kU/L using an ADVIA Centaur analyser. Determination of CA 19-9 on Vidas, AxSYM and Architect i2000 systems gave normal results. His rheumatoid factor concentration was very high (900 kIU/L) and assay interference was suspected.
Methods: Besides using several laboratory procedures to show the cause of the interference, we tried to estimate the frequency of the suspected interference. Therefore, two studies were performed. The first was carried out in a multicentre setting using four different CA 19-9 methods on 51 randomly selected samples with high rheumatoid factor concentrations and ten samples containing no or very low rheumatoid factor. In the second study we used heterophilic blocking tubes for 68 routinely analysed samples with CA 19-9 concentrations ranging between 37 and 250 kU/L using an ADVIA Centaur analyser.
Results: In the multicentre study we found eight discrepant CA 19-9 results, but only one was clearly due to interference. We showed that the interference detected, just as in the index case, was caused by rheumatoid factor. The other discrepancies could not be explained, but are probably related to method-dependent differences. In the 68 routinely analysed samples, no interference could be shown using the heterophilic blocking tubes.
Conclusions: Although interferences in the CA 19-9 assay are not frequent, the ADVIA Centaur system appears to be more sensitive to rheumatoid factor interference. The lack of standardisation remains an important issue for this assay. The determination of CA 19-9 during the follow-up of patients should be performed using a single method. If, however, there is any clinical doubt about a result, CA 19-9 should be determined using another method to exclude possible interferences.