To identify the differences in inflammatory profiles between hip OA, knee OA and non-OA control cohorts and investigate the association between cytokine expression and clinical outcome measurements, specifically pain.

A total of 250 individuals were recruited in three cohorts (100 knee OA, 50 hip OA, 100 control). Serum was collected and inflammatory profiles analysed using the Multiplex Human Cytokine Panel (Millipore) on the Luminex 100 platform (Luminex Corp., Austin, TX). The pain, physical function and activity limitations of hip OA cohort were scored using the WOMAC, SF-36, HHS and UCLA scores. All cytokine levels were compared between cohorts individually using Mann–Whitney–Wilcoxon (MWW) test with Bonferroni multiple comparison correction. Within hip OA cohorts, the effect of hip alignment (impingement and dysplasia) and radiographic grade (Kellgren and Lawrence grade, K/L grade) on cytokine levels were accessed by MWW test. Spearman's rank correlation test used to assess the association between cytokines and pain levels.

The three cohorts showed distinct inflammatory profiles. Specifically, EGF, FGF-2, MCP-3, MIP-1a, IL-8 were significant different between knee and hip OA; FGF-2, GRO, IL-8, MCP-1, VEGF were significant different between hip OA and control; Eotaxin, GRO, MCP-1, MIP-1b, VEGF were significant different between knee OA and control (p-value < 0.0012). For hip OA cohorts, cytokines do not differ between K/L grade three and K/L grade four or between patients that displayed either impingement or dysplasia. Three cytokines were significant associated with pain: IL-6 (p-value = 0.045), MDC (p-value = 0.032) and IP-10 (p-value = 0.038).

We have demonstrated that differences in serum inflammatory profiles exist between hip and knee OA patients. These differences suggest that OA may include different inflammatory subtypes according to affected joints. We also identified that the cytokine IL-6, MDC and IP-10 are associated with pain level in hip OA patients. These cytokines might help explain the inconsistent of presentation of pain with radiographical severity of OA joints. Future studies are needed to validate our findings and then to understand the following questions: (1) how differently affected joints are reflected in systematic biomarkers; (2) how these cytokines are biologically involved in the OA pain pathway.