Purpose: Hand osteoarthritis (OA) is a polyarticular disease, affecting the interphalangeal (IP) joints and the thumb base, including the first carpometacarpal (CMC-1) and scaphotrapeziotrapezoid (STT) joints. Based on previous studies, thumb base OA can be considered a distinct hand OA subset with a high burden of disease. In a cross-sectional study it was shown that thumb base pain is more strongly associated with radiographic damage than with MRI-defined inflammation, yet it is unknown if and how thumb base pain changes over time, and whether this is related to changes seen on imaging. Our aim was to investigate the course of thumb base OA pain and its association with changes in MRI-defined inflammation and structural damage. Methods: Longitudinal data of the Hand OSTeoArthritis in Secondary care (HOSTAS) study, which included consecutive patients diagnosed with primary hand OA by their treating rheumatologist, were used. Patients who underwent hand radiography, MR imaging and clinical examination of the right thumb base at baseline and at two-year follow-up were studied. Pain on palpation of the thumb base was assessed by trained research nurses on a 0-3 scale. Baseline and follow-up MR images, T1 FSE and T2 FSE fatsat in two planes, were scored paired in known time-order by two readers (SvB; FPBK), blinded for clinical and radiographic data, with good reliability, following the OMERACT thumb base OA MRI scoring system (TOMS). The CMC-1 and STT joints were assessed for synovitis (0-3) and bone marrow lesions (BMLs; 0-6 and 0-9, respectively) with half-point increments allowed for delta scores. The smallest detectable change was used as threshold for increase and decrease. Standard posterior-anterior right-hand radiographs were assessed for the presence of osteophytes in CMC-1 and STT joints following the OARSI atlas by two readers in consensus (SvB; HMK), blinded for clinical and MR imaging data, with good reliability. Since pain was assessed for the thumb base as a whole, imaging scores of CMC-1 and STT joints were combined, comparing no change in both joints (i.e. ‘stable’) versus increase/decrease in at least one joint. A thumb base was also classified as ‘stable’, if one joint increased while the other decreased. In thumb bases without maximum pain or maximum MRI scores at baseline, associations between increase in MRI features and increase in thumb base pain were investigated using logistic regression, presented as odds ratios (ORs) with 95% confidence intervals (CIs). Similarly, in thumb bases with pain and presence of MRI features at baseline, associations between decrease in imaging features and decrease in thumb base pain were investigated. Associations were adjusted for the other investigated MRI features and presence of radiographic osteophytes at baseline. Results: Out of a total of 161 patients (82.6% women, mean age 60.8 years, 91.3% fulfilling ACR hand OA criteria) 64 had pain on palpation of the thumb base at baseline (of whom 11 with maximal score). At the two-year follow-up visit, pain had decreased/resolved in 31 patients and increased/developed in 33 patients. The majority of patients had stable synovitis (n=106) and BML (n=96) scores over two years, although decreased (n=22, n=26, resp.) and increased (n=29, n=36, resp.) scores were common. Increase in radiographic osteophytes was rarely (n=10) observed. Increase in synovitis or BML was associated with increased pain (Table 1). A decrease in BML was associated with a decrease in pain, although it did not reach statistical significance (Table 1). The longitudinal associations between MRI features and pain were not substantially influenced by correcting for baseline osteophyte status (Table 1). Presence of osteophytes on baseline radiographs was weakly associated with change in pain in univariate analyses and attenuated when adjusting for the change in MRI features (Table 1). Decrease of synovitis in patients with baseline pain was scarce (n=7), therefore ORs were not computed. Conclusions: In this cohort of hand OA patients, thumb base pain levels varied over the course of two years in approximately forty percent of patients, of whom approximately half experienced an increase and half a decrease in pain. Changes in MRI-defined synovitis and BMLs of the thumb base joints were positively associated with change in pain on palpation. Baseline osteophytes were not significantly associated with change in pain. While cross-sectionally MRI-defined inflammation may be a less important determinant of pain than radiographic damage, this study shows that a change in inflammatory features may indeed be relevant.