Article Text

SAT0069 Performance of the Framingham Cardiovascular Risk Prediction Model with and without C-Reactive Protein or Erythrocyte Sedimentation Rate in RA: Analysis of US Electronic Medical Records Database
  1. V. Rajagopalan1,
  2. E. Alemao2,
  3. H. Kawabata2,
  4. D. Solomon3
  1. 1Med Data Analytics, Milltown
  2. 2Bristol-Myers Squibb, Princeton
  3. 3Brigham and Women's Hospital, Boston, United States


Background RA is associated with an increase in risk of cardiovascular (CV) disease, leading to reduced lifespan and quality of life, and increased healthcare costs.1 Besides the issue of accelerated atherosclerosis early in the RA disease course, correlation between RA and some of the modifiable CV risk factors has been reported.

Objectives To evaluate the role of inflammatory markers, C-reactive protein (CRP) or erythrocyte sedimentation rate (ESR) in predicting CV risk in RA patients (pts).

Methods Pts aged ≥18 years and with an RA diagnosis after 01/01/2001 were identified in GE Centricity® Electronic Medical Records database. Individuals were followed from first RA diagnosis date (index) until death, disenrollment, study end period (05/30/2012) or first CV outcome (fatal and non-fatal myocardial infarction, angina, stroke, transient ischaemic attack, intermittent claudication and heart failure). Pts with a history of CV events or pts with missing values on CV risks were excluded from the analysis. Distributions of Kaplan–Meier curves of time to first CV events by baseline (BL) CRP/ESR tertiles were compared using log rank test. The base-case prediction model was the Cox model with traditional predictors of CV events similar to the Framingham Heart Study (FHS) model.2 The role of CRP or ESR on CV risk was evaluated by assessing the discriminative properties of the model without versus with CRP or ESR using the area under the receiver operating characteristic curve (Δ AUC) via C-Index and net reclassification improvement (NRI). Pts with low CRP and ESR at BL (12 months prior to index date) were categorized as “Low CRP or ESR”, pts with high CRP or ESR were categorized as “High CRP or ESR” and the remaining were classified as “Medium CRP or ESR”.

Results 5300 RA pts were included in the cohort with median follow-up of 2.0 yrs. The mean (SD) age was 59.7 (14.2) yrs, with 74.8% being female. At BL, 51.0% were taking typical antihypertensive medications, 28.7% had diagnosed dyslipidaemia, 26.2% had diagnosed diabetes and 18.3% were smokers. Only 70% of the study population had BL CRP values (mean [SD] of 1.5 [8.2] mg/L) and 92.7% had BL ESR values (mean [SD] of 36.6 [24.6] mm/h). CV event-free survival was significantly lower in pts with a high CRP or ESR range (Figure; log rank test: p<0.001). In the base multivariate Cox model, the traditional risk factors of age, antihypertensive meds, dyslipidaemia, smoking and diabetes significantly increased the hazard ratio (HR) for CV events. In the enhanced model, “High CRP or ESR” as well as 'Medium CRP or ESR' was associated with increased hazard of CV events (High vs Low CRP or ESR HR=2.05, 95% CI 1.19, 3.52; Medium vs Low CRP or ESR HR=1.86; 95% CI 1.02, 3.38). The FHS model with (vs without) CRP or ESR had improved discrimination with Δ AUC of 0.004 and NRI of 5.4%; discrimination in the female cohort was even better with Δ AUC of 0.006 and NRI of 6.5%.

Conclusions Baseline inflammatory markers (CRP or ESR) were significantly associated with an increased risk of CV outcomes and the inclusion of CRP/ESR in the Framingham Heart Study model resulted in a moderate improvement in the reclassification of CV risk.


  1. Kramer HR. Arthritis Care Res 2011;63:484–99; 2. Wilson PWF. Circ Cardiovasc Qual Outcomes 2008 November

Disclosure of Interest V. Rajagopalan Consultant for: BMS, E. Alemao Shareholder of: BMS, Employee of: BMS, H. Kawabata Shareholder of: BMS, Employee of: BMS, D. Solomon: None declared

DOI 10.1136/annrheumdis-2014-eular.1833

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.