Psychological factors have been consistently associated with osteoarthritis (OA)-related pain and disability. However, the relationship between psychologically-derived profiles and measures of sensory function in OA has not been previously reported. Individuals with knee OA (n=197) completed a battery of psychological, health and sensory assessments. The psychological variables were subjected to hierarchical cluster analysis hypothesizing that specific psychological profiles would emerge and that clusters would significantly differ in self-reported pain and disability. We also hypothesized that clusters would differ in their responses to laboratory tests of pain sensitivity and pain inhibition. The best solution yielded four clusters based on their profiles across multiple psychological measures. Cluster 1 was characterized by high positive affect/optimism with low pain vigilance and depression. This group had the lowest self-reported pain/disability and the least sensitivity to tactile, pressure and thermal pain (p’s<0.01). Cluster 2 had high optimism with low active/passive coping. They had also low self-reported pain/disability and demonstrated significant pain inhibition (p’s<0.05). Cluster 3 showed high active/passive coping with low optimism. These participants had high self-reported pain/disability and signs of central sensitization to tactile stimuli. Cluster 4 exhibited the lowest positive affect/optimism with the highest pain vigilance and negative affect. These individuals experienced the highest self-reported pain/disability including widespread pain (p’s<0.001). Cluster 4 was the most sensitive to tactile, pressure and thermal stimuli and showed significant central sensitization to tactile and thermal stimuli (p’s<0.001). Our findings suggest that knee OA represents a heterogeneous pain condition characterized by considerable variability in psychological profiles, which are associated with both clinical pain and somatosensory function. Multidisciplinary pain treatment approaches consistent with the biopsychosocial model of pain should provide significant advantages if these are targeted to profiles such as those present in our OA sample.
© 2013 Published by Elsevier Inc.