Background/Purpose: The osteoarthritis (OA) population is heterogeneous and it is therefore difficult to target all patients with the same intervention. Thus there is a clinical need for diagnostic tools, which can aid in identification of patients of different phenotypes. Urinary CTX-II is a biomarker of cartilage degradation, which previously have been associated with disease severity. Serum C3M is a biomarker of type II collagen remodeling, which is released from the synovial membrane when induced with pro-inflammatory cytokines. We hypothesize that patients with low cartilage degradation (low CTX-II), but a high synovial turnover (high C3M) have a synovial-driven OA, whereas patients with a high level of cartilage degradation and a low level of synovial turnover have a cartilage-driven phenotype, and that these phenotypes are markedly different. The aim of the study was to compare these two phenotypes on clinical parameters and thereby test above hypothesis.

Methods: CTX-II and C3M were measured in the urine/serum of knee OA patients (n=497) from the placebo arms of the two phase III RCTs (NCT00486434 and NCT00704847). Baseline pain questionnaires (WOMAC), as well as radiography and MRI of both knees (target (T), non-target (NT)) were recorded for each patient. Biomarker cut-offs were determined at the 75%-quartile limit. The synovial phenotype (n=89) was defined as having C3M levels above 7.3ng/mL and CTX-II levels below or equal to 345pmol/mL (creatinine corrected). The cartilage phenotype (n=99) was defined as the inverse. Difference between clinical measures and biomarkers in the two patient groups were analyzed by Mann-Whitney.

Results: There were no difference in age and BMI, nor gender distribution, in the two phenotype groups. Joint space width were significantly lower in both knees the cartilage phenotype (NT, p=0.0083; T p=0.042). The sum of Kellgren-Lawrence score was significant higher in the cartilage phenotype (p=0.0001). This corresponded to a significantly higher cartilage volume in the synovial phenotype (NT, p=0.0026; T, p=0.046). Interestingly, WOMAC pain was significantly higher in the cartilage-driven phenotype as compared to the synovial phenotype (NT, p=0.045; T, p=0.014).

Conclusion: We found the two phenotypes, defined by two biomarkers reflecting synovial and cartilage turnover, respectively, where significantly different when comparing standard clinical measures. This may indicate that non-invasive biomarkers may aid in the identification of patient phenotypes and thereby, in the future, aid in guiding treatment of patients.

« Back to 2015 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/identification-and-characterization-of-two-osteoarthritis-phenotypes-applying-two-clinically-tested-biomarkers/