S122 Osteoarthritis and Cartilage Vol. 16 Supplement 4 3 distinct clinical features: presence of finger nodes (2 or more IPJs); thumb base involvement (1 or more CMCJ); and bony enlargement (2 or more IPJs or MCPJs). The resulting subgroups were then examined using ANOVA for differences in changes in self-reported pain and disability (AUSCAN) at 18 months. Results: 513 participants were eligible for the analysis. Presence of thumb base involvement and then finger nodes were the two most important clinical features in distinguishing subgroups. The overall mean change (SD) in AUSCAN scores in the total sample (baseline-18 months) was -0.27 (4.1) for pain and -0.43 (6.15) for disability (a negative score indicating deterioration). There was deterioration in both pain and disability scores in those participants with thumb involvement (with or without finger nodes) and this was statistically significant after adjusting for baseline AUSCAN score, age and gender (p = 0.02 for pain; p = 0.04 for function). There was a minimal improvement in the participants with finger nodes alone at baseline (Table 1). Differences did not reach statistical significance. Conclusions: Clinically different sub-groups can be identified in a com- munity dwelling population of older adults with hand problems. Presence of thumb base involvement at baseline contributed to poorer clinical outcome scores at 18 months. These findings are in line with EULAR recommendations for the diagnosis of hand OA. The extent to which clinical subgroups may predict poor long-term outcome in older adults with features of hand OA needs further evaluation. Table 1 Mean AUSCAN change at 18 months ANOVA for Thumb base not involved Thumb base involved group differences No finger nodes (n = 128) Has finger nodes (n = 176) No finger nodes (n = 56) Has finger nodes (n = 153) Pain -0.25 (3.8) 0.27 (4.4) -1.20 (4.2) -0.66 (4.0) P = 0.01; P = 0.11* Disability -0.61 (5.0) 0.36 (6.35) -0.96 (6.39) -0.96 (6.27) P = 0.02; P = 0.14* *Adjusting ANOVA models for baseline AUSCAN score, age and gender. 270 RELATING OBJECTIVE CT-BASED METRICS OF ACUTE FRACTURE SEVERITY TO THE INCIDENCE OF POST-TRAUMATIC OSTEOARTHRITIS T.P. Thomas, D.D. Anderson, T.V. Mosqueda, C.J. van Hofwegen, T.D. Brown, J.L. Marsh. University of Iowa, Iowa City, IA, USA Purpose: A new CT-based methodology was recently developed to objec- tively quantify fracture severity, working from the principle that mechanical energy is necessarily expended to create new free surface area in a brittle solid. The amount of energy thus expended is proportional to the amount of de novo interfragmentary surface area (Figure 1). We now measure other clinically relevant aspects of the comminuted fractures, including fragment displacement and articular fragmentation. To assess these additional measures, we compared them with experienced clinician opinion. In addition, for the first time we report on the ability of the metrics to predict post-traumatic osteoarthritis (PTOA) in ankles of patients with tibial plafond fractures at two years post-injury. We hypothesized that an injury severity metric that included objective measures of articular disruption, of fracture energy, and of fragment displacement/dispersal would be an accurate predictor of PTOA. Methods: Twenty tibial plafond fractures were chosen for study. These cases had previously undergone rank order analysis of overall injury severity by three experienced orthopaedic traumatologists. We now asked the clinicians to rank order these fractures looking at specific components of injury, including fracture displacement and articular comminution. A fragment segmentation algorithm was separately used to objectively measure articular comminution, fragment displacement, and lumped measures of overall fracture severity. Concordance rates were calculated to measure agreement between assessments made by the clinicians, and the CT-based computational measures. Kellgren-Lawrence (KL) scores were assessed at 2 years post-injury. Linear regression models were used to evaluate the ability of various combinations of the fracture severity measures to predict PTOA. Results: Concordance rates between clinican rank ordering and fragment displacement were high with rates from 82 to 89%. At two years post- injury, 13% of the patients had developed minimal PTOA (KL score = 2), and 31% had developed moderate to severe PTOA (KL scores 3). Linear regression indicated that fracture energy and articular comminu- tion, when combined, explain 73% of the variation in PTOA severity; however fragment displacement/dispersal did not correlate as strongly with degeneration (R 2 = 0.42). The combined fracture energy and articular comminution metric was a better predictor of KL scores than was clinician opinion (R 2 of 0.73 vs. 0.47, respectively). Conclusions: The findings of this study indicate that an objective CT- based fracture severity metric reasonably predicts subsequent PTOA, providing a strong impetus for further developing this novel paradigm. Financial support was provided by grants from the National Institutes of Health (AR46601, AR55533, and AR48939), The AO Research Fund and The Orthopaedic Trauma Association. Figure 1. A sample CT slice from the metaphysic shows the free bone surfaces that are identified by the edge detection algorithm (a). Interfrag- mentary surface area is the fractured surface area (red) not present in the pre-fractured template (green). Interfragmentary surface area is graphi- cally represented by the area between the intact and fractured curves (b). The dashed line represents the location along the tibia of the sample CT slice. (c) A 3D reconstruction illustrates the total interfragmentary surface area created during the fracture event. 271 CHANGE IN RADIOGRAPHIC SUBCHONDRAL BONE ATTRITION OVER TIME S. Reichenbach 1 , I. Watt 2 , S. Williams 3 , E. Nuesch 1 , P. Juni 1 , P.A. Dieppe 4 . 1 University of Bern, Bern, SWITZERLAND, 2 University of Leiden, Leiden, NETHERLANDS, 3 University of Bristol, Bristol, UNITED KINGDOM, 4 University of Oxford, Oxford, UNITED KINGDOM Purpose: Subchondral bone attrition (SBA) has become an increasingly important topic of investigation in research on pain and progression in knee osteoarthritis (OA), and is generally thought of as a late, irreversible finding in radiographic OA. SBA probably involves remodelling of the bone, resulting in flattening or depression of the articular surface. How- ever, we lack data on the natural course of radiographic SBA. Methods: Knee radiographs of participants in a population based, longi- tudinal observational study from the Somerset and Avon Survey of Health (SASH) were assessed for SBA both at baseline and 8 year follow- up. SBA was graded on posterioranterior weight bearing radiographs on a scale from 0-3 as follows: 0 = no SBA, 1 = mild SBA (<5 mm), 2 = moderate SBA (5-10 mm) and 3 = severe OA (>10 mm). One trained reader (SR) assessed both baseline and follow-up radiographs. We determined SBA to be present in any case with SBA > 0. Progression was considered to have occurred if the SBA grade had increased 8 years later. In addition, we recorded whether a total knee replacement (TKR) had been conducted. Each pair of radiographs was also qualitatively assessed by two independent readers (PD, IW) as follows: no change, progression of joint damage and bone loss, or bone remodeling with no further progression of joint damage. Baseline characteristics of subjects (day pain, night pain) and radiographic characteristics (K/L score, femoral and tibial attrition) were compared between the 4 groups (no change, reformation, progression, TKR). Results: We found 48 paired radiographs in 38 study participants where SBA had been read at baseline. Four paired radiographs were excluded in the quality assessment: one radiograph was judged not to have SBA (false positive reading), and three radiographs were judged not to have primary OA (secondary OA due to fractures and inherited malformation). In 19.2%, SBA was found in the femur only, in 37.5% it was only tibial, and in the remaining 33.3% it was found in both locations. K/L grade was high at baseline, indicating severe OA (median = 4, interquartile range (IQR) 3 to 4). At follow up after 8 years, more than half of the knees (54.2%) had been replaced (TKR). 22 paired radiographs without TKR were left for the qualitative assessment of the study: 12 knees were considered to be unchanged (54.5%), 6 knees (27.3%) showed progression, and 4 knees (18.2%) showed bone remodeling or improvement. Radiographs which