253 PREDICTIVE FACTORS OF PRESCRIPTION IN THE TREATMENT OF KNEE OSTEOARTHRITIS. PART 2: FIRST TREATMENT CHANGES

Purpose: EULAR guidelines for the treatment of Knee osteoarthritis (KOA) state that optimal management requires a combination of nonpharmacological and pharmacological modalities and that paracetamol is to be tried first. Our aim was to evaluate the predictive factors of the first choice line therapeutic strategy in general practitioners (GP)’ practice in France and Spain. Methods: Observational, prospective, multicenter 1-yr cohort study including patients 50 yrs, suffering from ACR defined KOA, not treated within the previous 6 months except for non-pharmacological treatment and/or self-medication with analgesics or low-dose NSAIDs and for whom it was planned to initiate a first line treatment according to the clinical status of the patient and the EULAR recommendations. Statistics. To identify the main predictive factors, multivariate analyses (logistic regressions) were performed with 1st choice modalities as dependent variables and clinical data at baseline as explanatory measures. Results: 383 GPs (50% in each country) included 1038 patients (69% females; means±SD: age 68±10 yrs, BMI 28.9±4.8 kg/m2, KOA duration 5.2±6.1 yrs). Since baseline patients’ characteristics in France and Spain were similar, both populations were pooled. Paracetamol alone was prescribed in 828 patients (440 with/388 without non-pharmacological modalities, mainly diet and/or exercices) and paracetamol plus other drugs (NSAIDs, SYSADOAs) were prescribed in 210 patients (114 with/96 without non pharmacological modalities). 1. Predictive factors of the association paracetamol + other drugs versus paracetamol alone are presented in table 1. Global GPs’ assessment, pain at night and effusion were significant factors in univariate (p< 0.05), but not in the multivariate analyses, while sex, age, BMI, past or present comorbidities, uni/bilateral KOA, recent flare-up of the pain, patient’s pain and patient’s global assessment were not significant at all. 2. The predictive factors of the association paracetamol + nonpharmacological treatments versus paracetamol alone are presented in table 2. Sex (females), bilateral KOA, self-medication, sick leave were significant in univariate, but not in the multivariate analyses. Conclusions: The predictive factors leading to the associations of nonpharmacological or other pharmacological modalities to paracetamol in the first GPs’ prescription for treating KOA seem different. Previous pharmacological treatment and request for specialist advice are good predictors of ajunction of other pharmaceuticals to paracetamol, while previous non-pharmacological treatments, obesity, recent flare-up and pain at night are predictors for the adjunction of non-pharmacological modalities.