Background and aims
Patient-specific pain development over time can help in understanding osteoarthritis (OA) heterogeneity by defining groups with similar characteristics and likely different treatment needs. We aimed to identify pain trajectories, to describe patients within trajectory-group using baseline demographic and psychosocial characteristics, and to explore whether trajectories predict end-stage disease i.e. total knee replacement (TKR).
Methods
We used data from a three-year, double-blind randomised control trial of vitamin D supplementation on knee OA (N=474, placebo vs 800IU vitamin D, equal groups, mean age 64, 61% women per treatment-group). The pain was measured in 6-month intervals using the WOMAC questionnaire (0-100 scale). Latent class growth analysis was employed to identify pain trajectories, multinomial analysis to regress trajectories on baseline patients’ characteristics, and logistic regression to predict TKR by pain trajectories.
Results
We identified four distinct pain trajectories. The first trajectory (36.6%, baseline pain 17.7, slope -1.4); the second (38.6%, baseline pain 30.5, slope 0.8); the third (21.4%, baseline pain 48.0, slope 2.1) and the fourth trajectory-group (3.4%, baseline pain 71.5, slope 2.2). Baseline characteristics age, BMI, physical and psychological health were significantly different among groups, and of these, physical health distinguishing every trajectory from all others. The fourth trajectory had OR of 12.14 (95% CI 1.59-92.86) of having TKR.
Conclusions
There are distinct OA patient groups with similar pain experience described as different baseline pain and development over time. These groups differ in baseline characteristics and progression risk. Future clinical OA trials can benefit from identifying patients at risk for disease and symptom progression.

07 - PAIN TRAJECTORIES AS PHENOTYPES OF KNEE OSTEOARTHRITIS