Prediction of radiographic progression in early rheumatoid arthritis (RA) using time-integrated C-reactive protein (TICRP): a population-based approach

Abstract

AIMS Various rheumatic diseases, including RA, are known to be associated with elevated levels of acute-phase proteins, particularly CRP. Studies suggest that there is a significant correlation between elevated CRP level and increased radiographic progression. The purpose of this analysis was to develop a population model to predict radiographic progression using TICRP. METHODS: Both CRP and Total Sharp Score (TSS) data were collected for up to 24 months in 560 patients with early RA (183: 25 mg etanercept twice weekly (biw); 184: 10 mg etanercept biw; 193: methotrexate once weekly) in a clinical study. A Freundlich model was used to depict the relationship between CRP and TSS using NONMEM. The potential impact of covariates, including treatment, sex, age, baseline TSS, baseline CRP, baseline erosion, and baseline joint-space narrowing, on model parameters (Freundlich exponent constant [1/N], Freundlich constant [K]) was assessed. RESULTS: The population model demonstrated that TICRP predicts the radiographic progression of RA quantitated by TSS. The model's predictive performance was confirmed by 8000 bootstrapping runs. The typical (SE) model parameters were 7.82 (1.35) for K and 0.77 (0.06) for 1/N. No important covariates were identified. CONCLUSIONS: Radiographic progression in RA can be successfully predicted by TICRP using a population-based approach. The utility of this model needs further investigation. Clinical Pharmacology & Therapeutics (2004) 75, P53–P53; doi: 10.1016/j.clpt.2003.11.202