Vertebral Fracture Assessment-detected Abdominal Aortic Calcification Enhances Cardiovascular Disease Risk Stratification of Rheumatoid Arthritis Patients

Background/Purpose: Osteoporosis and cardiovascular disease (CVD) are major comorbidities and CVD is the leading cause of death among patients with rheumatoid arthritis(RA). Traditional CVD risk prediction tools i.e., Framingham risk score(FRS) under-estimate the risk of CVD in RA. Novel biomarkers and better risk prediction tools are needed. Many RA patients undergo DXA today as part of an osteoporosis assessment, which may include a vertebral fracture assessment(VFA). VFA technology has been shown to reliably detect and quantify abdominal aortic calcification(AAC). VFA-detected AAC is an independent robust marker of CVD in other populations. It is unknown whether VFA-detected AAC is a useful marker of CVD in RA. We aimed to determine whether VFA-detected AAC is independently associated with CVD in RA patients and compared its utility to the FRS for CVD risk assessment in RA patients.

Methods: A cross-sectional study of our RA cohort at a University Hospital. We included RA patients aged ≥40 years who met 1987 ACR criteria for RA, had a DXA and VFA scan available for analysis and access to their medical records to ascertain their CVD risk factors and details. The study was approved by our local I.R.B. Two blinded consultant musculoskeletal radiologists independently reviewed all VFA scans to determine AAC using an established 24-point scale. We determined if AAC was independently associated with prevalent CVD using multivariable logistic regression. The ability of the FRS and AAC for determining the presence of CVD was assessed using ROC curve analyses.

Results: 1330 patients were screened. 603 met the inclusion criteria: mean age 56 years, 74% female, 76% sero-positive and 43% smokers. 230 had ≥1documented CVD event. Overall AAC was present in 211 of subjects: 11% was mild(<9 points), 57% moderate(9-16) and 32% severe(>16 points). The proportion of patients with AAC was substantially greater in subjects with CVD than those without(76% Vs 10%, p <0.05). VFA-detected AAC was significantly better than traditional risk factors for determining the presence of CVD. In multivariable analyses both the presence and severity of AAC was significantly and independently associated with prevalent CVD (OR 2.70; 95% CI 1.8 to 3.2). Both the FRS (AUC 0.58) and AAC (AUC 0.85) were significant predictors of CVD events (Fig. 1). The addition of VFA-detected AAC to the FRS significantly enhanced the performance of the FRS for determining CVD (AUCs increased from 0.58 to 0.79, p<0.001).

Conclusion: VFA-detected AAC is an important marker of CVD risk in RA patients, and out-performs traditional risk prediction tools. Further studies are needed to examine the utility of DXA-detected AAC in CVD assessment in RA patients.

Fig 1: Additive Effect of AAC on FRS for Predicting CVD in RA Patients*