Lack of strong association between SAA and CRP levels, less marked differential effects using biologic DMARDs and traditional DMARDs, coupled with their superior modeling of RA disease activity, suggests that SAA may be a better biomarker for RA disease activity than CRP, mainly during treatment with TNF antagonists.

Serum levels of systemic acute-phase reactants (APR) such as C-reactive protein (CRP), and serum amyloid A (SAA) are augmented during active synovitis. APR is known to be significantly related to joint impairment and disability and this is the reason why serum APR act as an indicator in disease severity, progression and prognosis in rheumatoid arthritis (RA). The widely used disease activity score of 28 joints (DAS28) is based on a count of 28 swollen and tender joints, a measure of general health or global disease and an inflammatory marker. CRP and SAA have many advantages over ESR. Therefore, an alternative formulation of the DAS28 based on CRP (DAS28-CRP) was developed and validated against radiographic progression and physical function. However, there is a considerable variation in CRP levels in patients with RA, and a substantial proportion of patients with RA have a clinically insignificant range of values. It has been recognized that DAS28-CRP tends to yield lower values of disease activity than the DAS28-ESR, resulting in substantial classification differences. In addition, as DAS28 based on ESR or CRP integrates various aspects of the disease into a single numerical value, there can be a great discrepancy between patient and provider assessments of disease activity. In contrast to ESR or CRP, less is known about the usefulness of SAA in RA. SAA is an acute-phase protein linked to the pathogenesis of various diseases, such as atherosclerosis, diabetes, Alzheimer's disease, and RA. Growing evidence suggests that acute-phase SAA is sensitive to change, reaches much higher levels than CRP, declines rapidly, and may therefore accurately reflect disease activity.

Rationale behind research:

There is a lack of clinical studies in large groups of patients with RA to assess the value of SAA in monitoring disease activity and predicting treatment response with various treatments, such as traditional disease-modifying anti-rheumatic drug (DMARD) combination therapy or biologic therapy.

Thus, this study was conducted to understand the correlation between SAA and CRP with disease activity during treatment and whether SAA is better or the same as a biomarker of disease activity during therapy.

Objective:

To compare the effect of etanercept (ETN) and conventional synthetic disease-modifying anti-rheumatic drug (DMARD) therapy on serum amyloid A (SAA) levels and to determine whether SAA reflects rheumatoid arthritis (RA) disease activity better than C-reactive protein (CRP).

Note: Spearman correlation coefficients (rho) were used to evaluate the relationship between SAA and CRP and mixed effects models to determine whether ETN and methotrexate (MTX) treatment compared to triple DMARD therapy differentially lowered SAA. Akaike information criteria (AIC) were used to determine model fits.

Study outcomes

CRP Measurement: CRP (mg/L) was measured using a high-sensitivity immunoturbidimetric assay. High CRP was defined as >3 mg/L. Plasma CRP was measured rather than serum CRP because of the availability of plasma specimens in our study population and because the measurement of CRP in plasma and serum are comparable.

SAA (mg/L) was measured in serum samples. The lowest detection limit for SAA was 0.97 mg/L, determined by the lower limit of the reference curve.

Time points 

Efficacy: Baseline, 24 weeks, 48 weeks and 102 weeks

Baseline: There were no baseline differences in age, SAA, CRP, RF status, or disease duration between the different treatment arms of the TEAR trial.

Outcomes

• Patients with RA treated with ETN/MTX had a greater reduction in SAA than patients treated with oral DMARD combination therapy even after correcting for disease activity. SAA was lower by an average of 66 ranks following treatment with ETN/MTX compared to triple oral therapy, using the mixed-effects model. The results were similar for serum CRP both by visit (p = 0.0254) and by treatment arm (p < 0.0001). Serum CRP was also lower following treatment with ETN/MTX versus triple oral therapy. With an even more pronounced mean difference than SAA, serum CRP was lower by an average of 144 ranks following treatment with ETN/MTX compared to triple oral therapy
• Across all patients and time points, models of the DAS28-ESR using SAA were better than models using CRP; lower values of the AIC and BIC indicated a better model fit. In particular, AIC takes into account the goodness of fit and the number of parameters required to achieve this. The AIC is based on the likelihood function and provides the goodness of fit for the model. comparing the AIC values for the overall model ESR versus SAA (AIC = 6854) with ESR versus CRP (AIC = 7159.3) indicates that the former model provides the better fit (the ΔAIC between the models was 305). The model of DAS28-ESR using SAA was associated with an approximately sixfold better fit versus the CRP model for patients treated with ETN/MTX, and an approximately fivefold better fit versus the CRP model for patients treated with oral DMARDs (ΔAIC = 159 versus ΔAIC = 137, respectively).

Table 1: Determining model fit for rheumatoid arthritis disease activity (as measured by the DAS28-ESR) by serum SAA and CRP

For patients with RA in the TEAR trial, both CRP and SAA serum levels decreased over time with treatment, particularly in the MTX/ETN group. Patients with RA treated with ETN/MTX had a greater reduction in SAA than patients treated with oral DMARD therapy, even after correcting for disease activity. This suggests that SAA may be a better biomarker for RA disease activity than CRP, especially during treatment with TNF antagonists.

The findings of the study are consistent with previous studies that showed greater reduction in SAA with different combination therapies such as golimumab (golimumab + MTX vs MTX alone) or tofacitinib (tofacitinib + MTX vs tofacitinib monotherapy). In an observational study of 50 consecutive patients with RA treated with long-term leflunomide, the reduction in SAA was transient, as it was no longer observed after 6 months, in spite of reduced disease activity.