Page 27 - THE EVOLUTION OF EARLY ARTHRITIS AND CARDIOVASCULAR RISK Samina A. Turk

PREDICTION OF FUTURE RHEUMATOID ARTHRITIS
INTRODUCTION 2
Rheumatoid arthritis (RA) on average becomes clinically manifest around the age of 55 years. During the healthy part of life, the risk of future RA is determined by genetic, reproductive and environmental factors (Fig. 1, green bar). Over time, people at risk for RA may pass through a phase of autoimmunity, accompanied by subclinical inflammation,(1) followed by a symptomatic phase, which may last a few months to several years. In the symptomatic phase, markers of autoimmunity and inflammation increase before the onset of clinical arthritis.(2) Therefore, prediction can be based on different characteristics in the asymptomatic phase and in the symptomatic phase.
The expectation that intervening in the preclinical phase of RA could be beneficial is based on the success of treatment of RA within 1 to 2 years after onset of clinical disease. (3,4) The new criteria for RA from 2010 with a focus on early signs such as involvement of even only a few small joints together with serology and acute phase reactants facilitate treatment in the earliest clinical phase(5,6) and the further characterization of the preclinical phase offers new opportunities for intervention studies even before clinically apparent arthritis occurs. Because RA is the most prevalent inflammatory rheumatic disease, with a high burden for the patient and society, it seems the ideal candidate rheumatic disease for screening and intervention programs. However, a lot of steps need to be taken before such programs can be offered to persons at risk.
This article summarizes the present knowledge on risk factors for RA, including genetic, reproductive, and hormonal factors; environmental exposures; biomarkers; personal characteristics and symptoms; and how these can be combined in risk models attempting to increase the accuracy of the prediction of RA. Genetic risk and gene-environmental interactions are dealt with elsewhere in this issue and are only mentioned here in relation to their roles in prediction models. Risk scores from such models require further validation, but could be used to select candidates for intervention studies.
METHODS
We searched the PubMed database on January 29, 2014, for the terms risk, predic- tion, and development in relation to RA. After excluding articles not directly related to prediction of RA, such as studies on prevalence, diagnosis, treatment, outcome, or comorbidities of RA, more than 200 articles remained on this topic after screening 2000 abstracts. Additional articles were added that were found after the search date until May 1, 2014, by screening rheumatologic journals.
25