Osteoarthritis (OA) is the leading cause of disability in US adults and its prevalence is expected to double by 2020 [
1]. Historically, the "disease" of OA has been viewed primarily as damage to the cartilage and bone. As such, the magnitude of damage or inflammation of these structures is often associated with higher symptom levels. Population-based studies suggest otherwise; 30 to 50% of individuals with moderate to severe radiographic changes of OA are asymptomatic, and approximately 10% of individuals with moderate to severe knee pain have normal radiographs [
2,
3]. Psychosocial factors do account for some of this variance in pain and other symptoms, but only modestly [
4‐
6]. There also may be other bone, joint, or physical changes associated with symptom severity which are still not well-understood. The current failure of peripheral damage, inflammation, or other factors to explain the presence, absence, or severity of chronic pain suggests the need to identify additional salient factors that may be contributing to the experience of OA.
There is growing evidence in OA that there is a central component to pain. Recent studies of animal models provide support for central sensitization of nociceptive pathways [
7,
8]. In addition, focus groups identified a subset of patients with chronic, symptomatic knee OA who used pain quality descriptors that were suggestive of neuropathic pain [
9]. There are several studies suggesting that OA patients display diffuse hyperalgesia to mechanical or heat stimuli (that is, suggestive of central nervous system (CNS) mediation) [
10‐
12]. Kosek demonstrated that individuals with hip OA had reduced descending analgesic activity, which partially normalized following hip arthroplasty, suggesting the involvement of central factors influencing the activity of peripheral nociceptive input [
13]. Gwilym and colleagues used both experimental pain testing and more sophisticated functional neuroimaging procedures to show evidence of augmented CNS processing of pain in 20 OA patients [
14]. In a separate study, this same group showed that atrophy of the thalamus was seen at baseline on OA and improved following arthroplasty [
15], again underscoring the role of the CNS in OA pain. Finally, recent randomized controlled trials have demonstrated that compounds that alter pain neurotransmitters centrally, such as serotonin and norepinephrine, (for example, duloxetine, tricyclics), are efficacious in OA [
16,
17]. In aggregate, these studies do not imply that peripheral factors are unimportant in OA; rather, peripheral factors alone are insufficient to account for symptoms in some or many individuals with OA. Whether or not CNS augmentation plays a prominent role in OA pain is likely to be tied to genetic predisposition, environmental stressors, and the degree of illness burden a given person is experiencing at the time [
18‐
25]. Non-region-specific symptoms accompanying pain, such as fatigue, cognitive problems, sleep problems, and perturbations of mood, are systemically-mediated symptoms that may index more central involvement in the maintenance of illnesses, such as pain [
24,
26,
27]. These symptoms may be important to target treatment in addition to pain. For instance, in a previous study with an OA sample, fatigue was actually more related to functional disability than to pain [
28]. To optimize treatment, it may be necessary to better characterize people with OA and determine if there are subsets of people who have symptom presentations that may reflect different underlying pain mechanisms. Since OA treatment is still largely focused on alleviating pain at the peripheral site (for example, strength training to reduce knee joint stress), it is particularly important to examine if there is a subgroup of people with OA that has a symptom cluster supporting a clinical presentation of centrally-mediated pain. For those people, it may be important to focus on treatment that is more commonly used in centrally-mediated pain conditions, (such as behavioral treatments and lifestyle interventions). Classification of people into empirically-derived subgroups has been done in other chronic pain conditions, such as fibromyalgia and low back pain, based upon symptoms and psychosocial factors and has successfully identified subgroups of patients with distinct phenotypic characterization and with differential response to treatment [
29‐
31]. OA is similar to these other chronic pain conditions in that many patients experience multifocal pain, fatigue, sleep disturbances, and mood disorders in addition to joint pain; however, no studies have examined if there are distinct subgroups of people with OA who differ based on their symptomatology.
The purpose of this study was to identify specific subgroups of individuals who initially presented with symptomatic knee or hip OA. Subgroups would be based upon levels of comorbid centrally-mediated symptomatology, such as fatigue, sleep disturbance, mood, as well as other symptoms, which we refer to as illness burden, known to occur in other centrally-mediated pain states (for example, irritable bowel syndrome, temporomandibular joint disorder, interstitial cystitis, and so on) [
32]. We hypothesized that one subgroup would appear to represent OA patients who predominantly have joint pain, whereas other subgroups would have different symptom presentations that support differential manifestations of CNS-mediated symptomatology.