Background/Purpose:

The multi-biomarker disease activity (MBDA) score [Vectra]
has been evaluated in a number of settings, yet has only limited data evaluating
whether it is influenced by comorbidities common to rheumatoid arthritis (RA)
patients.

Methods:

Using data from the CORRONA registry through (December, 2014),
we conducted a cross-sectional analysis of RA patients who had received at
least one MBDA test performed as part of standard of care, concurrent with C
reactive protein (CRP) ordered on the same day. Patients were characterized
descriptively by MBDA category (low, <=29; moderate 30-44; and high, >=
45) and by CRP, ESR, swollen joint count and CDAI category. LOESS plots of MBDA
score by SJC were stratified by common comorbidities (smoking, fibromyalgia,
COPD, diabetes, body mass index [BMI]) to visually evaluate the influence of
the comorbidity on the MBDA score. Ordinary least squares regression was used
to evaluate the association between MBDA score and age, sex, comorbidities and
RA-related factors.  The comorbidities were considered confounders if adjusting
for them changed the MBDA-SJC association by more than 10% (change-in-estimate
criterion).

Results:

A total of 585 patients had at least 1 MBDA test concurrent
with a CRP performed for standard of care reasons. Patient characteristics were
mean (IQR) 65 (54,73) years, 78% female, 66% white, 27% Hispanic. Mean (IQR)
BMI was 28.0 (24.5, 32.1); comorbidity prevalence was 10% (diabetes), 3%
(COPD), 5% (fibromyalgia); 12% current, 39% former, and 49% never smokers. A
total of 72% (n=423) of patients had normal CRP (<10 mg/L); of these, 31%
had high, 44% had moderate, and 25% had low disease activity as classified by
MBDA (Figure 1).

The MBDA score was significantly associated with SJC
(correlation coefficient 0.20, p < 0.0001), and was minimally different
between patients who did and did not have each of the comorbidities . In
bivariate models, there was no significant association between MBDA score and
fibromyalgia, diabetes, or COPD, and these factors did not yield confounding
between either MBDA score and SJC, nor MBDA and CDAI. CRP alone explained 33%
of the variance (R²) in MBDA score. In the multivariable model, MBDA
remained significantly associated with both SJC and CDAI even after adjusting
for age, sex, comorbidities, BMI, current steroid use, and other factors.

Conclusion:

In this real-world RA registry, the MBDA score was
associated with RA disease activity and was negligibly affected by
comorbidities common to RA patients. Approximately one-third of patients with
normal CRP levels had high MBDA scores, suggesting the potential that MBDA may
provide an opportunity to use an objective lab measure to identify RA patients
with active disease even when CRP is normal.

Figures 1: Distribution of CDAI according to MBDA
score category for RA patients

with normal CRP (< 10mg/L) [n=423]