Multibiomarker disease activity test predicts 3-year cardiovascular disease risk in RA
Click Here to Manage Email Alerts
Using clinical data, serum biomarkers and multi-biomarker disease activity, researchers have developed and validated a prognostic score that can predict 3-year cardiovascular risk in patients with rheumatoid arthritis, according to findings.
Writing in Arthritis Research & Therapy, the researchers concluded that their score demonstrates good accuracy and outperforms clinical models with and without C-reactive protein.
“CVD risk predictors developed for the general population tend to underestimate CVD risk in RA patients,” Jeffrey R. Curtis, MD, MS, MPH, of the University of Alabama at Birmingham, and colleagues wrote. “EULAR guidelines recommend that CVD risk predicted by tools such as the Framingham Risk Score (FRS) or the American College of Cardiology and American Heart Association (ACC/AHA) pooled cohort risk equation be multiplied by 1.5 to account for the effect of RA on CVD risk. A limitation of this approach is that it treats all RA patients the same, regardless of the level of disease activity.”
“The multi-biomarker disease activity (MBDA) test assesses RA disease activity by measuring 12 serum protein biomarkers to provide a validated score on a scale of 1-100 that correlates with the Disease Activity Score in 28 joints with CRP (DAS28-CRP),” they added. “In 2019, the American College of Rheumatology disease activity measures working group concluded that the MBDA score was one of 11 measures of RA disease activity that met the minimum standard for regular use. The MBDA score is predictive of future radiographic damage, independently of other measures.”
To develop and validate a biomarker-based model for predicting cardiovascular disease risk among patients with RA, Curtis and colleagues created a retrospective RA disease cohort by linking 2006-2016 Medicare claims data to MBDA test results in the Vectra database. A total of 30,751 patients with RA, aged 40 years or older, who underwent MBDA testing were included. These patients were assigned 2:1 to either the training or validation groups.
The researchers evaluated clinical and RA-related variables and MBDA score, as well as the score’s 12 biomarkers, all as predictors of a composite cardiovascular outcome that included myocardial infarction, stroke or death related to cardiovascular disease within 3 years. Construction of the model featured Cox proportional hazard regression with backward elimination. Curtis and colleagues then internally validated their MBDA-based risk score, comparing it to four other prediction models that use clinical data.
According to the researchers, there were 904 cardiovascular disease events among patients in the cohort. Covariates in the final MBDA-based risk score were age, diabetes, hypertension, tobacco use, history of cardiovascular disease excluding myocardial infarction or stroke, MBDA score, leptin, MMP-3 and TNF-R1. In their internal validation, the researchers found the MBDA-based risk score to be a strong predictor of 3-year risk for cardiovascular disease events, with a hazard ratio of 2.89 (95% CI, 2.46-3.41).
Among the cohort patients, the predicted 3-year cardiovascular disease risk was low for 9.4%, borderline for 10.2%, intermediate for 52.2%, and high for 28.2%.
The model demonstrated good fit, with mean a predicted 3-year CVD risk of 4.5%, compared with a mean observed risk of 4.4%. In addition, the MBDA-based score significantly improved risk discrimination by the likelihood ratio test, compared with four the models that used clinical data alone.
The MBDA-based score also improved prediction, reclassifying 42% of patients compared with the simplest clinical model — which used age and sex only — with a net reclassification index of 0.19 (95% CI, 0.1-0.27). It likewise reclassified 28% patients compared with the most comprehensive clinical model — which used age, sex, diabetes, hypertension, tobacco use, history of cardiovascular disease and CRP — with a reclassification index of 0.07 (95% CI, 0.001-0.13). The MBDA-based score’s C-index was 0.715, compared with a range of 0.661-0.696 for the four clinical models.
“We have developed and internally validated an MBDA-based CVD risk score that predicts risk for MI, stroke, or fatal CVD in the next 3 years for RA patients,” Curtis and colleagues wrote. “It is novel because it accounts for the contribution of RA inflammatory disease activity by including the MBDA score and three biomarkers that are independently associated with CVD. It performed better than prediction models that used only clinical data.”
“The MBDA-based CVD risk prediction score provides rheumatologists with a feasible tool for assessing CVD risk to inform the management of traditional CVD risk factors and RA inflammation,” they added. “Further validation with more extended time frames and more heterogeneous cohorts of RA patients will be helpful to assure its robustness as a prediction model.”
Collapse
Click Here to Manage Email Alerts