Background/Purpose: RA is characterized by synovial inflammation, cartilage loss and erosion of subchondral bone. It is critical to diagnose and effectively treat the disease early to suppress inflammation and stop destruction of the joints. Inflammation results in the release of tissue fragments that reflects disease activity which are measurable in serum. Identification of the patients that would most likely respond and not respond to treatment would bring optimal benefits for patients. The objective of the study was to investigate whether early changes (within 4 weeks) in serum biomarkers were predictive of ACR50 response at week 52 to 4 and 8 mg/kg tocilizumab (TCZ) treatment of RA patients.

Methods: The prospective biomarker substudy of the LITHE trial included 693 patients. The LITHE trial (Roche WA17823) was a 2-year phase III, 3-arm randomized, double-blind, placebo-controlled, parallel group, including patients with moderate/severe active RA, and with IR to DMARDs. The patients were randomized: TCZ8mg/kg (n=244), TCZ4mg/kg (n=207) or placebo (PBO, n=242) on a background of MTX. Biomarkers were measured in serum from baseline and 4 weeks: C2M (cartilage degradation), C3M (synovial inflammation), MMP3, total CRP, CRPM (MMP-degraded CRP), VICM (Citrullinated and MMP-degraded Vimentin), and ICTP (MMP destroyed type I collagen), osteocalcin (bone formation) and CTX-I (Bone resorption). Cut-off values for dichotomized analysis of the markers and the ACR50 response was determined by ROCs by best likelihood ratio for i) the individual markers at baseline, ii) changes from 4 weeks, and iii) ratio between markers. Decision trees (CART) were constructed using the cut-offs values and the predictive values for ACR50 response were calculated by 2×2 diagnostic tests.

Results: The proportion of ACR50 responders at week 52 for TCZ4 and PBO were 29% and 8.2%, corresponding to a number need-to-treat ratio (NNT) of 3.5 and 12, respectively. CART#1 (baseline C2M/CRP and C3M/CRPM ratios): A positive test resulted in identification of 68 TCZ4 cases which led to a PPV of 40% and a NPV of 76%. PBO; 1 out 19 cases with a positive test were responder. CART#2 (baseline MMP3 and C3M/CRPM ratio); A positive test resulted in identification of 86 TCZ4 cases, which led to a PPV of 43% and a NPV of 80%. PBO; 2 out 27 cases with a positive test were responder (NNT 2.3). CART#3 (CART#2 + 4-week change in osteocalcin and CTXI/C2M ratio); A positive test resulted in identification of 30 TCZ4 cases, which led to a PPV of 63% and a NPV of 79% (NNT 1.6). No PBO cases were selected. Similar data was obtained independently for TCZ 8mg/kg.

Conclusion: We identified more than 76% of the patients that did not respond to TCZ4 treatment. This increased the response rate from 29% up to 63%. We found that measurement of baseline biomarkers the need-to-treat ratio could be reduced from 3.5 to 2.3. If early changes in biomarkers were included then this need-to-treat ratio could be further reduced to 1.6. Whether this is an IL6 pathway-specific serological profile need to be investigated in additional clinical settings.

---

« Back to 2012 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/serum-based-biomarkers-of-joint-destruction-can-identify-responders-and-non-responder-to-tocilizumab/