Rheumatoid Factor: How often is it a False Positive? The Answer is Often.

Rheumatoid factors are antibodies against the Fc portion of IgG (also an antibody). They can belong to any isotype of immunoglobulin – eg, IgM, IgG and IgE – and any of these can be detected in the blood test.[1] Enzyme-linked immunosorbent assay (ELISA) can differentiate between the different subtypes.
Any patient who is suspected of having rheumatoid disease. However, keep in mind it may be negative in rheumatoid disease and thus it is important to refer before serology is available.[2] Many patients with rheumatoid disease are seronegative to begin with but 80% seroconvert (become positive).[3] Disease severity is often worse in those who are seropositive.[4]
Venous blood is taken in a ‘clotted’ tube – usually the same one as U&E.
Results can be reported in titres (normal <1:20) but more commonly as units (normal <23 IU/ml but see local guidelines, as this may vary from laboratory to laboratory). The sensitivity and specificity of rheumatoid factor for rheumatoid disease is low and thus it is not a good screening test. However, the predictive value of rheumatoid factor in patients with symmetric polyarticular joint swelling is 80%.
  • Rheumatoid arthritis – sensitivity in established disease is only 60-70% with a specificity of 78%.[5] The higher the level in rheumatoid disease the worse the joint destruction and the greater the chance of systemic involvement.
  • False positives occur in 5% of healthy individuals and in any inflammatory condition – eg, Sjögren’s syndrome, systemic lupus erythematous and mixed connective tissue disorder.

Disease associations of rheumatoid factor (sensitivity in brackets)[1][6]

  • Rheumatoid arthritis (60-70%).
  • Sjögren’s syndrome (85-95%).
  • Felty’s syndrome (>95%).
  • Systemic sclerosis (~30%).
  • Infective endocarditis.
  • Systemic lupus erythematous (~25-35%).
  • Infectious mononucleosis.
  • Hepatitis.
  • Juvenile rheumatoid arthritis.
  • Tuberculosis.
  • Dermatomyositis.
  • Syphilis.
  • HIV.
  • Influenza.
  • Malignancy.
  • Sarcoidosis.
  • Leukaemia.
  • Healthy individuals (5% increasing to 20% over the age of 65 years).
Rheumatoid factor does not generally help in monitoring rheumatoid disease, although it may help with the use of newer agents such as etanercept and infliximab. In patients on etanercept or infliximab and DMARDs the levels of rheumatoid factor reduce, which is associated with reduced clinical disease activity.[78]
One study in Denmark found that people with elevated rheumatoid factor have up to 26-fold greater long-term risk of rheumatoid arthritis, and up to 32% 10-year absolute risk of rheumatoid arthritis.[9]
Rheumatoid factor can also predict disease outcome in some patients.[1] One example of this is that radiological progression – ie changes in hand X-rays – is worse in those who are seropositive.[10]

 1992 Dec;152(12):2417-20.

How useful is the rheumatoid factor? An analysis of sensitivity, specificity, and predictive value.



The rheumatoid factor (RF) is frequently ordered in an effort to detect disease, yet its diagnostic utility has not been thoroughly examined. To determine the test’s sensitivity, specificity, positive predictive value, and negative predictive value, we analyzed tests ordered in our institution.


We performed a retrospective analysis of all 86 patients with a positive RF over a 6-month period identified consecutively soon after the test was ordered. A similar analysis was applied to 86 seronegative patients selected at random from a total seronegative population of 477 during the same period. The patients represented the primary care and subspecialty practices and inpatient wards of a 504-bed university teaching hospital.


A positive RF result was strongly associated with rheumatoid arthritis or another rheumatic disease. For rheumatoid arthritis, sensitivity = 0.28 and specificity = 0.87, while for any rheumatic disease, sensitivity = 0.29 and specificity = 0.88. The positive predictive values for rheumatoid arthritis and any rheumatic disease were 0.24 and 0.34, respectively, and the negative predictive values were 0.89 and 0.85, respectively. Seropositive patients were slightly older (55 vs 49 years old), but the incidence of false-positive RFs among the elderly (69%) was not significantly higher than among younger patients (65%). The cost per true-positive RF result was $563.


In this study, most positive RF results were not helpful since the majority represented false-positive results. The low positive predictive value of the RF casts doubt on the utility of the RF in the diagnostic evaluation of patients. Contrary to traditional clinical expectations, the diagnostic utility of the RF may be greatest when it is negative. However, the subset of patients with seronegative rheumatic disease reduces the test’s power to exclude such disorders even when the RF is negative. Given the test’s limitations, clinicians should reconsider their expectations when ordering an RF. The utility of the RF may improve if it is ordered more selectively.

Sensitivity and specificity are statistical measures of the performance of a binary classification test, also known in statistics as classification function:
  • Sensitivity (also called the true positive rate, the recall, or probability of detection[1] in some fields) measures the proportion of positives that are correctly identified as such (e.g. the percentage of sick people who are correctly identified as having the condition).
  • Specificity (also called the true negative rate) measures the proportion of negatives that are correctly identified as such (e.g. the percentage of healthy people who are correctly identified as not having the condition).
Another way to understand in the context of medical tests is that sensitivity is the extent to which true positives are not missed/overlooked (so false negatives are few) and specificity is the extent to which positives really represent the condition of interest and not some other condition being mistaken for it (so false positives are few). Thus a highly sensitive test rarely overlooks a positive (for example, showing “nothing bad” despite something bad existing); a highly specific test rarely registers a positive for anything that is not the target of testing (for example, finding one bacterial species when another closely related one is the true target); and a test that is highly sensitive and highly specific does both, so it “rarely overlooks a thing that it is looking for” and it “rarely mistakes anything else for that thing.” Because most medical tests do not have sensitivity and specificity values above 99%, “rarely” does not equate to certainty. But for practical reasons, tests with sensitivity and specificity values above 90% have high credibility, albeit usually no certainty, in differential diagnosis.

Sensitivity therefore quantifies the avoiding of false negatives, and specificity does the same for false positives. For any test, there is usually a trade-off between the measures – for instance, in airport security since testing of passengers is for potential threats to safety, scanners may be set to trigger alarms on low-risk items like belt buckles and keys (low specificity), in order to increase the probability of identifying dangerous objects and minimize the risk of missing objects that do pose a threat (high sensitivity). This trade-off can be represented graphically using a receiver operating characteristic curve. A perfect predictor would be described as 100% sensitive, meaning all sick individuals are correctly identified as sick, and 100% specific, meaning no healthy individuals are incorrectly identified as sick. In reality, however, any non-deterministic predictor will possess a minimum error bound known as the Bayes error rate.
Shopping Cart