The stability of pain phenotypes in people with hand osteoarthritis – results from the NOR-HAND study

Objective In a hand osteoarthritis (OA) cohort, we aimed to explore pain phenotypes and characterize their stabilities using the multidimensional framework developed by The Initiative on Methods, Measurement and Pain Assessment in Clinical Trials (IMMPACT). Methods We included participants attending the visits at baseline (2016–17) and follow-up (2019–21) in the Nor-Hand study. To identify pain phenotypes (i.e., classes), we conducted latent transition analyses modelling self-reported pain severity, neuropathic-like pain, fatigue, sleep, anxiodepressive symptoms, pain catastrophizing, and quantitative sensory testing (QST). We compared a model including QST with a model without. Longitudinal stability of classes was assessed. Changes in indicator variables across participants with vs. without between-class transitioning were compared. Results We analysed 213 participants (86.9 % women) with a baseline mean age (standard deviation) of 60.9 (6.0) years. Both models identified four classes exhibiting similar results on differences in pain severity and psychosocial burden (kappa 0.91). Although phenotype stability varied (probability range: 0.48–0.95), most participants ( ∼ 80 %) remained in the same class at both visits. Most participants transitioning between classes shifted to a less severe pain class (32/44 (72.7 %) in the QST model; 36/45 (80.0 %) in the clinical model), showing larger improvements in pain and psychosocial burden than non-transitioners. Conclusion Four distinct pain phenotypes were identified among persons with hand OA. Excluding QST from our model did not influence phenotype composition or characteristics. Various longitudinal phenotype stability was observed. Between-class transitions were often characterized by less pain and psychosocial burden, potentially due to regression to the mean or improved disease coping.