Background/Purpose: Pain in osteoarthritis (OA) may be nociceptive or neuropathic-like in nature. In knee OA, pain severity is known to be poorly correlated with joint damage. Identifying a patient’s pain profile may guide the provision of specific therapy. We sought to determine the prevalence of neuropathic-like pain (NP) and pain sensitisation (PS) as defined by self-report questionnaires in people with knee and hip OA.

Methods: MEDLINE, EMBASE, CINAHL were systematically searched in duplicate (1990-April 2020) for full text articles presenting the prevalence of NP and PS in knee and hip OA using self-report questionnaires, with ≥30 participants, age ≥18 years. Data were extracted into a prespecified form in duplicate. Risk of bias was assessed using the National Institute of Health Quality Assessment Tool for Observational Cohort and Cross Sectional Studies. Meta-analysis was performed using RevMan, using a random effects model. Heterogeneity between studies and sub-groups was assessed using Cochran’s Q and I² statistics.

Results: From 2706 non-duplicated references, 39 studies were included (2011-2020), from Europe (n=17), Asia (n=14), North America (n=4), Oceania (n=3) and Africa (n=1). Thirty-six studies reported on knee pain and 6 on hip pain. Study populations were recruited from hospital outpatients (n= 21), presurgical candidates (n=10), trials (n=5) and community (n=3). NP was defined using the following self-report questionnaire tools: PainDETECT (PDQ, n=30 studies), Douleur Neuropathique 4 (DN4, n=5), Self-Report Leeds Assessment of Neuropathic Symptoms and Signs (S-LANSS, n=2). PS was defined by Central Sensitisation Index (CSI, n=6) and Fibromyalgia Survey Questionnaire (FSQ n=1).

For knee OA, the prevalence of possible NP, defined by PDQ score ≥13 was 40% (95% CI 32-48%, p< 0.00001, I²=97%). The prevalence of probable NP, defined by PDQ score ≥19 was 20% (95% CI 15-24%, p< 0.00001, I²=94%). Using PDQ, there were no significant differences in prevalence between study population subgroups, for possible or probable NP (I² = 0%). The prevalence of NP using S-LANSS was 32% (95% CI 26-38%, I²=0%, p=0.43), and using DN4 was 41% (95% CI 24-59%, p< 0.001, I²=97%). The prevalence of PS using CSI was 36% (95% CI 12-59%, p< 0.001, I²=95%).

For hip OA, the prevalence of possible NP (PDQ score ≥13) was 28% (95% CI 10-45%, p< 0.001, I²=92%). The prevalence of probable NP (PDQ score ≥19) was 9% (95% CI 6-13%, I²=0%, p< 0.00001). Using PDQ, the overall prevalence of possible NP was similar across multiple study population sources (I² = 0%).The prevalence of NP using DN4 was 22% in 1 study. One study used FSQ to assess PS but did not report prevalence.

Conclusion: Using self-report questionnaire tools, NP was determined to be more prevalent in knee than hip OA. Further study is required into the prevalence of PS in knee and hip OA and differences between tools. The prevalence of NP in knee and hip OA were similar for each joint across study population sources. This suggests that NP pain is unrelated to OA severity. NP may be used to phenotype patients, enabling targeted analgesic and non-pharmacological therapy in OA and potential improvement to quality of life.

« Back to ACR Convergence 2020

ACR Meeting Abstracts - https://acrabstracts.org/abstract/systematic-review-and-meta-analysis-of-the-prevalence-of-neuropathic-like-pain-and-pain-sensitisation-in-people-with-knee-and-hip-osteoarthritis/