Background/Purpose:

Triggering Receptor Expressed on Myeloid cells-1 (TREM-1) is a cell surface receptor expressed mainly on monocytes and neutrophils, known to amplify inflammatory response. We aimed to investigate whether serum level of soluble form of TREM-1 (sTREM-1), which reflects TREM1 activation, is associated with disease activity (DA) features and/or can predict response to biologic agent in rheumatoid arthritis (RA).

Methods:

In this ancillary analysis of the Rotation or Change trial in which 300 patients with an inadequate response to a first line anti-TNF agent were randomized to receive either a 2^nd anti-TNF treatment or a non-anti-TNF agent, baseline serum level of sTREM-1 was assessed using ELISA kit. Levels of sTREM-1 were compared by non-parametric tests for categorical variables as gender, body mass index (BMI) and Disease Activity Score 28-C reactive protein (DAS28-CRP). Spearman correlation coefficients were calculated between sTREM-1 level and DA measures (DAS28-CRP, CRP level, tender joint count (TJC), swollen joint count (SJC) and patient global assessment (PGA)). Univariable then multivariable logistic regression analyses were used to assess whether baseline sTREM-1 level was associated with EULAR response at 24 weeks (W24) in each group of treatment.

Results:

sTREM-1 was available in 272 patients: 83.1% female, mean age (standard deviation) 56.9 (12.2) years, rheumatoid factor (RF) positive in 81.0% and anti-cyclic citrullinated peptide antibody positive in 81.6%. CRP level was >5mg/L in 59.8% patients and mean DAS28-CRP was 4.8 (1.0). W24 good or moderate EULAR response was achieved in 51.9% patients in 2^nd anti-TNF group vs 66.9% patients in non-anti-TNF group (p=0.01).

sTREM-1 was detectable in all patients with a mean level of 471.1 (242.0) pg/mL, was higher in men (585.0 (240.1) pg/mL men, 447.9 (236.3) pg/mL women, p=0.0004) but was not associated with seropositivity status nor BMI. The mean sTREM-1 level was higher in patients with DAS28-CRP>5.1 (542.5 (279.6) pg/mL) than those with DAS28-CRP<5.1 (433.3 (212.5) pg/mL; p<0.01). sTREM-1 was also positively correlated with DAS28-CRP score (R=0.25, p<0.001), due to its correlation with CRP level (R=0.38, p <0.0001), but also to specific assessments of RA (PGA R=0.14 and SJC R=0.20, p < 0.05).

Mean baseline sTREM-1 levels did not differ between W24 good and moderate EULAR responders vs non-responders (459.9 (217.0) vs 487.6 (275.0) pg/mL overall, 450.8 (210.2) vs 502.7 (291.6) pg/mL in 2^nd anti-TNF group and 466.7 (22.7) vs 466.6 (251.7) pg/mL in non-anti-TNF group, all p>0.05), nor between W24 good EULAR responders vs non-responders.

Conclusion:

Serum sTREM-1 may be a new DA marker in this large cohort of RA patients. Interestingly, sTREM-1 did not only reflect systemic inflammation (i.e. CRP level) but also clinical joint inflammation, suggesting a specific role in RA synovitis pathogenesis. In this RA population with first anti-TNF treatment failure, sTREM-1 was not associated with W24 EULAR response to treatment.