Page 31 - THE EVOLUTION OF EARLY ARTHRITIS AND CARDIOVASCULAR RISK Samina A. Turk

PREDICTION OF FUTURE RHEUMATOID ARTHRITIS
Three other studies investigated the development of RA in ACPA-positive and/or RF- 2 positive patients with arthralgia.(121,145,166) The patients were partly recruited in
primary care, and partly in the rheumatology clinic. The models were based on clinical characteristics, symptoms, and antibody characteristics, in 1 study supplemented by ultrasonographic power Doppler signal (see Table 2).(166) All 3 models provide good discrimination between persons who do or do not develop RA. However, they require
ongoing validation as other studies select and follow such cohorts of people at risk for RA. Similar studies from North America designed to predict RA in first-degree relatives of patients with RA are underway but have not yet gathered enough arthritis cases to enable the construction of prediction models.(149,173) These studies are hampered by the low frequency of autoantibodies or of increased biomarkers in relatives of patients with RA.
Measuring the risk of RA is also a matter of timing. During the early at-risk stage, before the onset of autoimmunity, clinicians can only measure genetic susceptibility and environmental factors (see the left part of Fig. 1). The predictive capability of models in this situation is becoming good, with areas under the curve of 72% to 77% for the prediction of ACPA-positive RA.(174) However, the measured risk is a lifetime risk, which makes it an abstract figure for the individual person at risk. Prediction including a time frame becomes possible nearer to the onset of clinical RA, when the aspects of symptoms, autoimmunity, and inflammation can be taken into account. In the Amsterdam risk model, points can be gathered for clinical characteristics, symptoms, and serology, with more points for high levels of ACPA or positivity for both ACPA and RF.(145) The more points, the higher the risk and the sooner the onset of arthritis can be expected (Fig. 2). This prediction reflects studies in pre-RA blood donors, in which autoantibody levels increase during the 1 to 3 years before the onset of clinical arthritis.(2,138) In an US cohort of 81 patients with clinical RA from whom stored serum was available from 1 to 12 years before disease onset, a biomarker profile including autoantibodies and cytokines was identified that predicts the imminent onset of clinical arthritis within 2 years.(160) Autoantibody epitope spreading by itself in the preclinical phase also predicts progression to classifiable RA.(143)
SCREENING STRATEGIES
Many medical, ethical, and economic issues need to be addressed before screening for risk of future RA can be offered to certain categories of unaffected persons. Basic requirements for screening groups of people to predict a disease are (1) a defined population to test; (2) the existence of an asymptomatic (or nonspecific symptomatic) phase; (3) the availability of a test with good accuracy, low rates of side effects, and low cost; and (4) the availability of a cost-effective intervention in the at-risk phase. Only the second requirement of an asymptomatic phase is clearly fulfilled at present. Regarding items 3 and 4, no single test can identify those at risk for RA and no intervention exists
29