A PROGNOSTIC MODEL FOR PROGRESSIVE DISEASE IN EARLY RHEUMATOID ARTHRITIS

Background: Establishing the prognosis of a patient with early rheumatoid arthritis is an important part to their evaluation and treatment. To date, no reliable prognostic algorithm at early stage disease has been developed to accurately predict long term functional outcome. Such an algorithm could improve the targeting of aggressive therapies such as TNF inhibitors.Objectives: A number of prognostic models were developed using the ERAS data to find an algorithm useful for clinical decision-making.Methods: The Early Rheumatoid Arthritis Study (ERAS) is a well-established dataset that has followed over 450 patients annually for up to 10 years. Previous studies have used the study for predictive models (1). Clinical markers measured include rheumatoid factor, Larsen radiological score, functional grade, number of swollen and tender joints, grip strength, Health Assessment Questionnaire (HAQ), ESR, haemoglobin(Hb), and Disease Activity Score (DAS). Baseline and 1yr clinical, laboratory and socioeconomic data (using the Carstairs deprivation index) were used to predict the functional outcome (HAQ score either severe ≥1.5 or not <1.5). Logistic regression models were used to estimate the risk of long-term outcome. Models were internally validated by discrimination (Area under ROC or c-index and Nagelkerke R2) and calibration (by estimates of slope shrinkage).Results: Using the whole cohort, HAQ at baseline and 1 years as expected were the most significant prognostic variables (ORs =1.7 and 2.4 respectively). Additional covariates were Hb, Larsen score, Carstairs index and functional grade at baseline, DAS28 and Functional Grade at 1 year. The resulting calibration indicated little need for recalibration. Reasonably large values for the c-index and the Nagelkerke R2 (=0.5) indicate that the set of prognostic factors is explaining the variation in outcome reasonably well, and this implies good prediction for individual patients. A nomogram will be presented for simple clinical calculation.Conclusion: The results suggest an opportunity to identify patients with rapid disability and progression at an early stage. Additionally the value of collecting the HAQ is demonstrated in making clinical decisions. Internal validation tests indicate the equation is generalisable, but external validation is necessary to fully determine its clinical use. Future research is required to identify whether more aggressive therapy in these identified subgroups of early RA patients will improve long-term prognosis.References: 1. D. James, Orthopaedic intervention in early rheumatoid arthritis. Occurrence and predictive factors in an inception cohort of 1064 patients followed for 5 years. Rheumatology 2004: In press. http://www.rheumatology.oupjournals.org/cgi/reprint/keh059v1Citation: , volume , supplement , year 2004, page Session: Advances in RA clinical picture