Clinical Risk Prediction Modeling in Rheumatoid Arthritis

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): This proposal is submitted as a competing continuation of NIH R01 AR49880-03, Hormone, Cytokine and Genetic Risks for RA in Women, in which we identified reproductive factors including breastfeeding, early menarche, and irregular menses, inflammatory markers including anti-CCP antibodies, and TNFR2 levels and a novel RA risk allele in the prolactin gene as significant risk factors for RA. Extending our work to develop clinical risk prediction models, this proposal builds on our strong track record of studying RA epidemiology in the Nurses' Health Studies, the largest prospective rheumatic disease cohorts in the world. Recent whole genome association studies in RA from our co-investigators have identified novel risk loci. However, despite rapid advances in understanding the genetic basis of RA, it is unclear how to utilize this information clinically for RA prediction. Identification of autoantibodies and cytokines present many years prior to RA onset provides an exciting opportunity to intervene during the pre-clinical phase. However, it is critical to understand the role of RA risk factors for the targeting of potentially toxic therapies at highest risk individuals. Predictive modeling is critical in the progress towards an RA prevention clinical trial. We propose to build a RA clinical risk prediction models incorporating RA genetic susceptibility alleles and environmental risk factors and their interactions, with validation in large U.S. and Swedish cohorts. Further validation in a unique high risk RA cohort, representing a target group for prevention trials, will lead to understanding of whether the models predict development of pre-clinical RA, essential information for future RA prevention trials. We propose the following aims: 1) Using validated RA susceptibility alleles, derive a Genetic Risk Score (GRS) and examine associations between GRS and RA risk in general, and with seropositive RA risk specifically, in 700 RA cases and 700 matched controls from the Nurses' Health Study (NHS) and in 2000 cases and 1150 matched controls from the Epidemiologic Investigation of RA (EIRA) cohort; 2) Develop two RA clinical prediction models to predict 5-year RA risk for all RA and for subsets defined by sex, immune phenotype, and family history: (a) an "environmental" model using behavioral factors, environmental exposures, and clinical factors , and (b) an "environmental + genetic" model with environmental factors, GRS, and gene-environment interaction terms; and 3) Examine the goodness of fit of the prediction models developed and validated in Aim 2 for predicting an intermediate endpoint, pre-clinical RA defined autoantibodies, or RA symptoms, in a unique high risk RA cohort, the Studies of the Etiologies of Rheumatoid Arthritis (SERA) comprised of 2100 first-degree relatives of RA cases and of 800 individuals enriched with HLA-DR4 alleles (total N=2900). The ability to accurately predict an individual's 5-year risk of developing clinical RA based on a simple genetic risk score, behavioral, environmental and clinical risk factors would be an enormous advance, enabling risk factor modification and earlier introduction of effective therapies to abrogate the destruction and disability of this disease. PUBLIC HEALTH RELEVANCE: This study will describe the characteristics of patients with rheumatoid arthritis (RA) who successfully engage in physical activity and those who do not. This information will be used to develop personally tailored physical activity counseling to promote health and reduce cardiovascular risk of patients with RA.