Neuropathic-like pain in knee osteoarthritis: exploring differences in knee loading and inflammation. A cross-sectional study.

Background: A subgroup of patients with knee osteoarthritis (OA) reports symptoms attributable to a neuropathic cause. Little to no attention has been invested on investigating differences in knee loading and inflammation in these patients.

Aim: To explore differences in inflammation and knee loading in patients with knee OA categorized based on the presence of neuropathic-like pain.

Design: Cross-sectional study.

Setting: Ghent University Hospital, Ghent, Belgium.

Population: Knee OA patients.

Methods: cross-sectional analysis of data from 96 patients (mean age 64.18±7.11 years) with primary knee OA participating in a randomized controlled trial. Participants were divided into three groups (unlikely, possible and indication of neuropathic-like pain) according to the modified painDETECT questionnaire (mPDQ). Data on demographics, symptoms and physical function were obtained by questionnaires. Effusion/synovitis and bone marrow lesions (BMLs) were measured using magnetic resonance imaging. Knee loading variables (knee adduction moment [KAM], KAM impulse, and knee flexion moment [KFM]) were assessed by 3D-motion analysis. One-way analysis of covariance (ANCOVA), Chi-square test and curve analyses were used to analyze continuous, categorical and loading variables respectively. Multinomial logistic regression was used to identify predictors for neuropathic-like pain.

Results: Patients with indication of neuropathic-like pain exhibited higher KAM impulse compared to those with no indication of neuropathic-like pain (standard mean difference (SMD): -0.036 Nm normalized to body weight and height per second, 95% CI: -0.071, -0.001) along with greater pain intensity (SMD: 3.87 units, 95% CI: 1.90, 5.84), stiffness (SMD: 1.34 units, 95% CI: 0.19, 2.48) and worse physical function (SMD: 13.98 units 95% CI: 7.52, 20.44). Curve analysis showed no significant differences in KFM and KAM between groups. Effusion/synovitis and BMLs did not differ significantly between groups. The best predictors for indication of neuropathic-like pain were KAM impulse, Hoffa and sex.

Conclusions: Knee OA patients with indication of neuropathic-like pain exhibited higher dynamic medial loading, greater pain severity and worse physical function, while inflammatory markers were not significantly different across mPDQ groups. Future longitudinal studies are warranted to strengthen the evidence and establish mechanisms to explain associations between neuropathic-like pain and knee loading.

Clinical rehabilitation impact: Knee loading is a modifiable factor and patients with neuropathic-like pain may benefit from offloading interventions.