Background: Knee pain is a common musculoskeletal complaint with an estimated annual population prevalence of 25% in people aged over 55 years. There are many causes of knee pain though osteoarthritis (OA) is one of the most frequent. Not all people with OA, however, have knee pain. There is discordance between pain intensity and disease severity, the reason for which is unknown. Variation in pain sensitivity may be one possible explanation. Quantitative sensory testing (QST) is a non-invasive technique using non-painful and painful stimuli to assess altered sensitivities in the skin and muscle. Little is known, however, about pain sensitivity in people with knee pain and the role of psychosocial factors in relation to pain sensitivity and pain intensity. Intra-articular steroids are a widely used and effective therapy for knee OA though response to treatment varies in both magnitude and duration of response. Pain sensitivity and/or psychosocial factors may explain some of the variation observed in response to treatment.Aims: To determine whether (i) greater sensitivity to stimuli is associated with higher levels of pain intensity in a population-based sample with knee pain, and whether those associations are mediated by psychosocial factors, (ii) there are changes in QST following intra-articular steroid injections in patients with symptomatic knee OA, and (iii) whether psychosocial factors and sensitivity to stimuli at baseline predict change in pain following intervention.Methods: 72 men and women with knee pain were recruited from a population-based cohort. All had QST assessments and completed a range of questionnaire instruments addressing pain intensity and psychosocial factors. QST assessments (including thermal, mechanical, vibration and pressure) were made at the most affected knee and contralateral forearm. Assessments of tender point count, wind-up ratio and diffuse noxious inhibitory control were also performed. Structural equation modelling was used to determine whether associations between QST measures and pain intensity were mediated by a latent psychosocial factor. In a separate open label trial of intra-articular steroid injections, 32 men and women with symptomatic knee OA underwent QST assessments and also completed questionnaires. The assessments were performed at both knees at the baseline visit (prior to injection) and at a post-injection visit 5-15 days later. Changes in QST were assessed using Wilcoxon matched pairs signed-rank with linear regression used to determine baseline QST predictors of change in pain.Results: In the observational study, mechanical hyperalgesia (tender point count, mechanical pain sensitivity, and allodynia), illness perceptions, catastrophizing and disability scores were positively associated with higher levels of pain intensity. Mediation analyses revealed stronger associations for the indirect effect including a latent psychosocial mediator between measures of mechanical hyperalgesia and global pain, and stronger associations for the direct effect between measures of mechanical hyperalgesia and knee pain. In the intervention study no changes in QST were observed between visits. However, lower baseline mechanical pain thresholds at the injected knee and illness perceptions predicted response to treatment.Conclusion: Illness perceptions and mechanical hyperalgesia can be used to identify subjects experiencing higher levels of global and knee pain intensity, and those who were more likely to respond to intra-articular steroid therapy. Changes in knee pain following intervention with steroid injection are not explained by changes in pain sensitivity.