The prevalence of radiographic hip osteoarthritis is increased in high bone mass; a case-control study

Abstract

s / Osteoarthritis and Cartilage 22 (2014) S57–S489 S343 601 THE PREVALENCE OF RADIOGRAPHIC HIP OSTEOARTHRITIS IS INCREASED IN HIGH BONE MASS; A CASE-CONTROL STUDY S.A. Hardcastle y, P. Dieppe yz, C.L. Gregson y, D. Hunter x, G. Thomas x, N.K. Arden kx, T.D. Spector{, D.J. Hart{, M.J. Laugharne#, G.A. Clague yy, M.H. Edwards k, E. Dennison k, C. Cooper kx, M. Williams zz, G. Davey Smith xx, J.H. Tobias y. yMusculoskeletal Res. Unit, Univ. of Bristol, Bristol, UNITED KINGDOM; zUniv. of Exeter Med. Sch., Exeter, UNITED KINGDOM; xNIHR Musculoskeletal BioMed. Res. Unit, Univ. of Oxford, Oxford, United Kingdom; kMRC Lifecourse Epidemiology Unit, Univ. of Southampton, Southampton, United Kingdom; {Dept. of Twin Res. and Genetic Epidemiology, King’s Coll. London, London, United Kingdom; Dept. of Radiology, Royal United Hosp., Bath, United Kingdom; yyDept. of Radiology, Royal Glamorgan Hosp., Llantrisant, United Kingdom; zzDept. of Radiology, North Bristol NHS Trust, Bristol, United Kingdom; xxMRC Integrative Epidemiology Unit, Univ. of Bristol, Bristol, United Kingdom Purpose: Numerous epidemiological studies have reported an association between increased bone mineral density (BMD) and osteoarthritis (OA), but whether this represents cause or effect remains unclear. To establish whether higher BMD predisposes to OA, we aimed to determine whether individuals with High Bone Mass (HBM) have a higher prevalence of radiographic hip OA compared with controls. Methods: HBM cases were recruited from 15 UK centres by systematically screening DXA databases. HBM was defined in index cases as a total hip Z-score þ3.2 and L1 Z-score þ1.2, or vice-versa, and in firstdegree relatives of index cases as total hip Z-score plus L1 Z-score þ3.2; unaffected relatives were recruited as controls. AP pelvic X-rays were performed in participants aged 40 years. Age-stratified random sampling was used to select further population controls from the Chingford 1000-women and Hertfordshire cohort studies. All radiographs were assessed for features of OA (Croft score, osteophytes, joint space narrowing (JSN), cysts, sclerosis) by a single observer blinded to case-control status using an atlas. Minimum joint space width (JSW) was measured using a computer-aided method. Intra-observer repeatability for most features was good. Analyses used logistic regression, with generalised estimating equations to account for within-person clustering (right/left), adjusted a priori for age, gender and body mass index (BMI). Results: Analyses included 530 HBM hips in 272 cases (mean age 62.9 years, 74% female) and 1702 control hips in 863 controls (mean age 64.8 years, 84% female), after excluding poor quality films and hip replacements (n1⁄4109). The prevalence of radiographic hip OA, defined as Croft score 3, was increased in cases compared with controls (20.0% vs. 13.6%). Results of logistic regression analyses for the binary radiographic OA variables in HBM cases vs. controls are shown in the table below, adjusted for age, gender and BMI. Cases had an increased odds of hip OA compared with controls after adjustment (OR 1.52 [1.09, 2.11], p1⁄40.013). In analyses of individual radiographic features, osteophytes (both any osteophyte, and moderate ( grade 2) osteophytes) and subchondral sclerosis were also more prevalent in cases compared with controls. However, the prevalence of JSN was not increased in HBM cases, and measured JSW did not differ between the groups (mean difference 0.04mm [-0.05, 0.13], p1⁄40.39). Analyses were repeated in the different control groups separately, and stratified by gender, with broadly similar findings. Logistic regression analysis of radiographic hip OA variables in HBM cases vs. all