Background Viscosupplementation (VS) by intra-articular (IA) injections of hyaluronic acid (HA) is widely used worldwide to alleviate pain and improve function in patients with mild to moderate knee osteoarthritis (KOA). However, in most of the clinical trials, the rate of responders to treatment does not exceed 60 to 70%. One of the key points that remains to be clarified is the identification of predictive factors of success or failure to treatment.
Objectives To identify clinical and radiographic predictive factors of treatment's response or failure in patients treated with VS for KOA.
Methods We included in this post-hoc analysis, 166 patients, with fully available clinical and radiographic data, coming from the intent-to-treat population (n=205) of a controlled, multicentre, double-blind, randomized, clinical trial, which compared 2 HA viscosupplement in patients with symptomatic KOA. At baseline, demographic, anthropometric, clinical data (patient global assessment, WOMAC score, presence or lack of synovial effusion) radiologic data (Kellgren-Lawrence grade, OARSI score for joint space narrowing [JSN]) and all previous and present treatments for OA, previous IA steroids or HA injections were recorded. All patients received 3 weekly IA injections of HA. They were classified as Responders if they fulfilled the OMERACT-OARSI criteria, 6-month after injections. Mann-Whitney and chi-square tests were used to assess the association of quantitative or qualitative factors and OMERACT-OARSI response. Multivariate logistic regression analysis was used to identify predictors of response.
Results Clinical and demographic characteristics at baseline as well as treatment effectiveness did not statistically differ between the 2 HA groups, so that the data could be pooled. The studied population consisted in 101 women and 65 men, mean age 65.2, mean disease duration 48.7, mean BMI 27.7. 47 patients were obese. Seventy three patients had grade 3 JSN. At baseline, mean WOMAC pain and function scores were 9.8 and 27.5, respectively. Before study inclusion, 95 had at least one VS or corticosteroid injection. At month 6, WOMAC pain and total scores decreased by 43.4% and 41.9% respectively. One-hundred-thirteen patients (68.1%) were classified as Responders. In univariate analysis, increased BMI and OARSI grade were significantly associated with lack of response (p=0.002 and p=0.005, respectively). Similar associations, although not statistically significant, were found with previous IA HA or steroids injection (p=0.11) and age >65 years (p=0.13). Multivariate analysis confirmed the association of obesity and radiological severity with lack of response. Finally, we found a strong cumulative impact of potential predictive factors (age >65, previous VS or steroid IA injection, obesity and severe JSN) on VS response. Thus, all patients without predictive factors were Responders, versus only 69.1% and 28.6% of those with 2 or 4 risk factors, respectively (p<0.0001).
Conclusions VS should be chiefly considered in subjects with normal BMI and moderate tibio-femoral joint space narrowing. This study shows, for the first time, a strong cumulative impact of potential predictive factors of failure. This can be helpful for clinicians when they inform patients on the chance of success of VS.
Acknowledgement To LABRHA for authorization tu use its database
Disclosure of Interest None declared