Determinants of non-nociceptive pain in Rheumatoid Arthritis.

Introduction: Features suggestive of neuropathic pain (NP) have been described in RA in addition to nociceptive pain. We aimed to determine the clinical predictors of NP in RA patients and study its association with radiographic structural damage.

Methods: Cross-sectional study was performed with RA patients followed at our Rheumatology department. Patients with diagnosed neuropathy of other origin, non-RA related risk factors for NP (e.g. diabetes mellitus) or fibromyalgia according to expert opinion were excluded. Demographic and clinical data were collected and disease activity/functional measures were evaluated. Two questionnaires were applied to assess NP: the Leeds Assessment of Neuropathic Symptoms (LANSS) and the painDETECT questionnaire (PDQ). Radiographs performed in up to 12 months before/after the evaluation were classified according to the modified van der Heijde Sharp´s method. Univariate and multivariate logistic regression were performed to identify the predictors of NP.

Results: 112 patients were included. 86 (77%) were women, with a mean (SD) age of 55.1 (10.8) years and median disease duration of 13 [2-41] years. 45 (40%) patients had NP by the LANSS (≥12) and 28% had a possible/likely NP in the PDQ (≥13). Female sex was predictive of NP by both tests and disease duration was inversely associated with LANSS NP. After adjusting for those two variables, pain VAS and TJC were positive predictors of NP by both tests. The same was not true for SJC, ESR or CRP levels. DAS28-CRP was significantly associated with PDQ NP, losing its statistical significance after adjustment for TJC and pain VAS. The HAQ score increased the odds of NP for both tests, independently of DAS 28-CRP. Positivity for ACPA and previous/current hydroxychloroquine treatment had lower odds of NP. 90 patients performed radiographic evaluation. Joint narrowing score was a significant negative predictor of LANSS NP. After adjusting for global radiographic score, current methotrexate treatment had lower odds of LANSS NP and previous/current leflunomide was a positive predictor of NP by both tests.

Conclusion: NP was associated with disease activity/functional scores but not with objective inflammatory measures. Greater structural damage, increased disease duration and ACPA positivity did not seem to increase the odds of NP. Possible association of NP and underlying csDMARD treatment was uncovered.