Background Antibody profiling encompassing rheumatoid factor (RF) and anti-citrullinated peptide antibodies (ACPA) supports diagnosis in patients with Rheumatoid arthritis (RA). However, RF and ACPA are not specific for RA, and predictive values of tests depend heavily on the selection of individuals in which such tests are performed. Because testing for these antibodies is frequently ordered by non-Rheumatologists, at substantial costs, we sought to determine the predictive values of such testing in patients of a large tertiary hospital.
Objectives To evaluate the positive and negative predictive values (PPV, NPV) of RF and ACPA tested in a pre-selected population of all patients of the Vienna general hospital between 2006 and 2012.
Methods Results of all RF and ACPA tests performed between 2006 and 2012 were obtained from the Department of Laboratory Medicine and the ordering departments were determined. Diagnoses were extracted from the hospital-wide database.
Results Between 2006 and 2012 more than 45,000 RF and ACPA tests in 5496 patients were performed. Among these, 21% were positive for RF, 3% for ACPAs and 22% were positive for both antibodies. For 2250 patients (41% of all patients in whom RF/ACPA was tested) the tests were not ordered by the Department of Rheumatology. The tests were requested by the Departments of Infectiology (32.5%), Angiology (11.1%), Orthopedics (8%), Nephrology (8%), and Dermatology (7%). 1572 of these 2250 patients had a documented diagnosis. Figure 1 shows the distribution of diseases according to the ICD-10.
Among the 2250 patients 134 (6%) were positive for RF, 72 (3.2%) for ACPA and 31 (1.4%) patients tested positive for both antibodies. PPV and NPV for the presence of musculoskeletal diseases were 22.4% and 89.9% for RF and 26.4% and 88.8% for ACPA testing. However, for presence of chronic inflammatory musculoskeletal diseases (ICD-10-Codes M05.X to M09.X and M30.X to M35.X) PPV was only 11.9% for RF and 16.7% for ACPA. NPV was 95.3% and 95.2%, respectively.
Conclusions RF and ACPA testing was frequently ordered by non-Rheumatologists. In this patient group, we found a relatively high NPV (95%) but a very low PPV of 11-17%. Thus, >83% of positive tests did not contribute to a diagnosis of inflammatory musculoskeletal disease. This observation underscores the necessity to use such testing only in the appropriate clinical context.
Disclosure of Interest None declared