D H Solomon1, J Greenberg2, J R Curtis3, M Liu4, M E Farkouh5, P Tsao1, J M Kremer6, C J Etzel7. 1. Brigham and Women's Hospital, Boston, Massachusetts. 2. New York University School of Medicine and New York University Hospital for Joint Diseases, New York, New York, and CORRONA, Southborough, Massachusetts. 3. University of Alabama at Birmingham. 4. CORRONA, Southborough, Massachusetts. 5. Mount Sinai School of Medicine, New York, New York, and University of Toronto, Toronto, Ontario, Canada. 6. Albany Medical College and Center for Rheumatology, Albany, New York, and CORRONA, Southborough, Massachusetts. 7. University of Texas MD Anderson Cancer Center, Houston, and CORRONA, Southborough, Massachusetts.
Abstract
OBJECTIVE: Cardiovascular disease (CVD) is the leading cause of mortality in rheumatoid arthritis (RA), but CV risk prediction scores derived from the general population do not accurately predict CV risk in RA patients. The goal of these analyses was to develop and internally validate an expanded CV risk prediction score for RA. METHODS: Study participants were patients with RA and no known CVD from the Consortium of Rheumatology Researchers of North America registry. Two-thirds of the cohort were used to derive the CV risk prediction score, and one-third for internal validation. Traditional CV risk factors were included in the base Cox regression model, and RA-related variables were assessed in an expanded model predicting confirmed CV events. Fit and utility of the expanded model were evaluated. RESULTS: The study cohort included 23,605 RA patients with 437 CV events over a median followup of 2.2 years. The RA variables found to be significant in the regression models and included in the expanded risk model were disease activity (Clinical Disease Activity Index >10 versus ≤10), disability (modified Health Assessment Questionnaire disability index >0.5 versus ≤0.5), daily prednisone use (any versus none), and disease duration (≥10 years versus <10 years). The expanded model had good fit (Hosmer-Lemeshow goodness of fit P = 0.94) and a lower Akaike's information criterion than the base model. In the internal validation cohort, the c-statistic for model discrimination was significantly improved from the base model to the expanded model (from 0.7261 to 0.7609; P = 0.0104). The net reclassification index of CV risk in models using a 4-category CV risk prediction tool was 40% (95% confidence interval 37-44%). CONCLUSION: This newly developed, expanded risk score for CV outcomes in RA performs well and improves the classification of CV risk in comparison to a risk prediction score in which only traditional risk factors were included.
OBJECTIVE:Cardiovascular disease (CVD) is the leading cause of mortality in rheumatoid arthritis (RA), but CV risk prediction scores derived from the general population do not accurately predict CV risk in RApatients. The goal of these analyses was to develop and internally validate an expanded CV risk prediction score for RA. METHODS: Study participants were patients with RA and no known CVD from the Consortium of Rheumatology Researchers of North America registry. Two-thirds of the cohort were used to derive the CV risk prediction score, and one-third for internal validation. Traditional CV risk factors were included in the base Cox regression model, and RA-related variables were assessed in an expanded model predicting confirmed CV events. Fit and utility of the expanded model were evaluated. RESULTS: The study cohort included 23,605 RApatients with 437 CV events over a median followup of 2.2 years. The RA variables found to be significant in the regression models and included in the expanded risk model were disease activity (Clinical Disease Activity Index >10 versus ≤10), disability (modified Health Assessment Questionnaire disability index >0.5 versus ≤0.5), daily prednisone use (any versus none), and disease duration (≥10 years versus <10 years). The expanded model had good fit (Hosmer-Lemeshow goodness of fit P = 0.94) and a lower Akaike's information criterion than the base model. In the internal validation cohort, the c-statistic for model discrimination was significantly improved from the base model to the expanded model (from 0.7261 to 0.7609; P = 0.0104). The net reclassification index of CV risk in models using a 4-category CV risk prediction tool was 40% (95% confidence interval 37-44%). CONCLUSION: This newly developed, expanded risk score for CV outcomes in RA performs well and improves the classification of CV risk in comparison to a risk prediction score in which only traditional risk factors were included.
Authors: Dionicio A Galarza-Delgado; Jose R Azpiri-Lopez; Iris J Colunga-Pedraza; Jesus A Cardenas-de la Garza; Raymundo Vera-Pineda; Griselda Serna-Peña; Rosa I Arvizu-Rivera; Adrian Martinez-Moreno; Martin Wah-Suarez; Mario A Garza Elizondo Journal: Clin Rheumatol Date: 2017-02-01 Impact factor: 2.980
Authors: Cynthia S Crowson; Sherine E Gabriel; Anne Grete Semb; Piet L C M van Riel; George Karpouzas; Patrick H Dessein; Carol Hitchon; Virginia Pascual-Ramos; George D Kitas Journal: Rheumatology (Oxford) Date: 2017-07-01 Impact factor: 7.580
Authors: Narendra N Khanna; Ankush D Jamthikar; Deep Gupta; Matteo Piga; Luca Saba; Carlo Carcassi; Argiris A Giannopoulos; Andrew Nicolaides; John R Laird; Harman S Suri; Sophie Mavrogeni; A D Protogerou; Petros Sfikakis; George D Kitas; Jasjit S Suri Journal: Curr Atheroscler Rep Date: 2019-01-25 Impact factor: 5.113