Background Multiple epidemiological and experimental data indicate that diabetes may be one of osteoarthritis risk factors. In spite of this, there is a lack of data on whether diabetes is an independent predictor of knee OA incidence and progression.
Objectives To assess influence of diabetes on incidence of radiographic knee OA and to evaluate diabetes effects on OA structural progression using radiological and MRI measures.
Methods For the current study we used longitudinal data obtained from the Osteoarthritis Initiative (OAI) progression (n=1390) and incidence (n=3284) subcohorts, which are publically available at http://oai.epi-ucsf.org. Incident knee OA was defined either as the occurrence of (a) new or enlarging definite osteophyte (OSP) or (b) new combination of a definite OSP and joint space narrowing (JSN). Incident OSP and JSN were defined as a development of OARSI grade >0. Radigraphic outcomes were assessed at baseline and years 1,2,3, and 4. For the analysis of diabetes effects on OA progression we assessed changes in bone marrow edema (BME) volume (baseline, year 1) and cartilage volume (baseline, years 1,2), each measured at six separate knee anatomical locations using MRI. The information on previously diagnosed diabetes was collected from the self-reported Charlson Comorbidity Index. Cox proportional hazards models and generalized estimating equations (GEE) were used to evaluate diabetes effects on dichotomized and continuous outcomes, respectively. We used both univariate and multivariate models, the latter were adjusted for demographic characteristics, prior knee surgery, Physical Activity Scale for the Elderly, body mass index (BMI), and family history of knee OA.
Results On baseline, 151 (10.9%) patients in progression subcohort and 210 (6.4%) patients in incidence subcohort reported diabetes. Diabetes effects on radiographic knee OA, individual radiographic OA feature incidence, cartilage volume and BME volume were analyzed in 4794, 1130, 600, and 160 knees, respectively. In univariate analyses diabetes was associated with increased risk of medial JSN and medial femur osteophyte incidence. Correction for BMI and other predictor variables reversed the effect and diabetes became associated with decreased risk of incident knee OA, medial and lateral JSN (Table). In both univariate and multivariate analyses diabetes had no effect on knee OA progression assessed by cartilage volume and BME volume changes over time.
Conclusions Diabetes is independently associated with decreased incidence of radiographic knee OA, while having no influence on its structural progression.
Disclosure of Interest : None declared