Pain perception in osteoarthritis (OA) is complex, with the intensity of pain reported in affected individuals often with little correlation to structural damage. Few studies have explored the neural networks responsible for processing pain in OA. Here, Kulkarni et al (Arthritis Rheum 2007; 56(4):1345) use functional positron emission tomography (PET) scanning to investigate the structures of the brain involved in pain processing in patients with knee OA.
Methods
Subjects with intermittently symptomatic knee OA treated with non-pharmacologic measures or simple analgesics only (i.e. no narcotic analgesics or antidepressants) underwent PET scanning under three conditions on three separate days:
1) During typical arthritic knee pain
2) During a period free of knee pain
3) Experimental pain in which heat pain stimuli was applied to the skin over the arthritic knee
Subjects discontinued analgesics at least 12 hours prior to testing and were asked to not engage in unaccustomed physical activities prior to undergoing pain-free and experimental pain testing.
Results
12 subjects (50% female, mean age 59 years, knee OA evenly split between right and left knee as the predominant symptomatic knee) underwent PET scanning. Group means for depression, anxiety, and catastrophizing scores did not change over the course of testing. Pain ratings (on a 100 mm scale) were low during pain-free testing (mean rating 4 ± 10) and elevated during both arthritic and experimental pain testing (mean rating 62.9 ± 11.4 and 62.4 ± 15.3, respectively), but did not significantly differ between arthritic and experimental pain testing.
Despite similar pain ratings, significant differences in location of brain metabolic activity were detected between the three testing conditions. Compared to pain-free testing, arthritic pain showed significant enhancement in all pain areas of the brain with activation bilaterally, including the area of the areas responsible for processing emotional and affective responses to pain (cingulate cortex), aversion conditioning and fear processing (amygdala), and attention (prefrontal and posterior parietal cortices). Both experimental pain and arthritic pain activated the same regions, although arthritic pain was associated with greater activation of the medial pain system (associated with emotion and fear) and the primary motor cortex, a finding possibly related to inhibition of the desire to move the leg with arthritic pain.
Conclusions
Pain from knee OA and experimental pain are both processed in brain areas associated with mood, emotions, and fear. However, OA pain was associated with greater activation in these areas.
Editorial Comment
The pain response in OA may be mediated as much by pain processing mechanisms in the brain as structural damage. These data suggest that the pain response in OA goes beyond the mere localization of sensations, but reaches deep into the emotional centers of the brain. Studies involving functional brain imaging may identify areas of the brain that may be targets for novel therapeutic agents, as treatment of both emotional and physical pain may be needed to alleviate the symptoms of painful knee OA in some patients.
