Background Metabolic syndrome-related osteoarthritis (OA) is now considered as a peculiar subgroup of OA involving systemic factors as insulin resistance and adipokines. However, few studies searched for a relationship between serum adipokines and/or insulinemia and OA in a large group.
Objectives To investigate whether insulinemia, the serum leptin/adiponectin (L/A) ratio (known to increase in insulin resistant states)1 and individual serum adipokines may be independently associated with hip OA clinical symptoms in a OA population-based cohort study.
Methods Subjects aged 40-75 years, with uni or bilateral symptomatic hip and/or knee OA (ACR criteria), Kellgren and Lawrence ≥2, were recruited in the multicenter KHOALA (Knee and Hip Osteo-Arthritis Long-Term Assessment) cohort2. Exclusively patients included for symptomatic hip OA were investigated here. Beyond demographic data, clinical evaluation of OA at inclusion was based on WOMAC score for pain, stiffness and function (higher values in more symptomatic patients) and on Osteoarthritis Knee and Hip Quality of Life (OAKHQOL) questionnaire (higher values in less symptomatic patients). A blood sample at inclusion was collected to assess adipokines (i.e. total and high molecular weight [HMW] adiponectin, leptin, visfatin), insulinemia as well as potential confounders (i.e., ultrasensitive CRP level, creatininemia, total cholesterol, triglycerides, apoA1 and ApoB100 levels). Multivariate linear regression models were used to determine whether insulinemia, serum adipokines or L/A ratio were associated with hip OA symptoms.
Results 284 patients were included for uni- or bilateral symptomatic hip OA (66% female, mean age ± SD: 61.8±8.8 years, mean body mass index 27.6±4.9 kg/m2). Serum visfatin, total adiponectin, HMW adiponectin, leptin, L/A ratio and insulinemia were 4.2±2.1 ng/mL, 6.6±3.7 μg/mL, 3.7±2.5 μg/mL, 22.8±25.4 ng/mL, 1.42±1.1 and 91.1±87.7 pmol/L, respectively, without significant difference between uni- and bilateral involvement. Multivariate analyses showed that the L/A ratio, but not each adipokine considered separately or insulinemia, was positively correlated with total WOMAC (β=3.1; p=0.002). This correlation was mainly due to the correlation with the function WOMAC subscore (β=3.3; p=0.002). Again, only the L/A ratio was independently correlated with OAKHQOL pain (β=-3.4; p=0.01) but not OAKHQOL physical activity. In addition, serum leptin level was independently correlated with OAKHQOL physical activity (β=-2.9; p=0.05).
Conclusions In hip OA, a higher L/A ratio seems to be associated with a poorer clinical hip OA outcome, highlighting the association of insulin resistance with more severe hip OA symptoms. This association was independent of all confounders including BMI, age and other metabolic parameters, suggesting a potential direct impact of adipose tissue and related insulin resistance on OA.
FM Finucane et al Diabetologia 2009;52:2345–2349.
F Guillemin et al Joint Bone Spine 2012;21:597-603.
Disclosure of Interest None declared