The Natural History of Bone Marrow Lesions in Community-Based Middle-Aged Women Without Clinical Knee Osteoarthritis Patricia A. Berry, BBioMedSci (Hons),* Miranda L. Davies-Tuck, BBioMedSci (Hons),* Anita E. Wluka, MBBS, FRACP, PhD,* ,† Fahad S. Hanna, PhD, BSc,* ,† Robin J. Bell, MBBS, FAFPHM, PhD, ‡,§ Susan R. Davis, MBBS, FRACP, PhD, ‡,¶ Jenny Adams, RNRM, and Flavia M. Cicuttini, MBBS, FRACP, PhD* Objective: Bone marrow lesions (BML) have been implicated in the pathogenesis of osteoarthritis, yet their exact role, etiology, and natural history remain unclear. The aim of this study was to examine the natural history of BML in a healthy population and identify risk factors associated with their persistence and incidence. Methods: One hundred forty-eight healthy middle-aged women had magnetic resonance imaging performed on their dominant knee at baseline and 2 years later to assess the presence, natural history, and risk factors for persistence and incidence of BML. Results: Approximately 46% of BML present at baseline completely resolved over 2 years. “Large” BML had the potential to improve, while the majority of “very large” remained stable. In those women with no BML at baseline, approximately 9% developed a BML over 2 years, the majority in the medial compartment. There was a trend toward weight being a risk factor for the develop- ment of “very large” BML (P = 0.08). Conclusions: The natural history of BML may be different in healthy persons compared with diseased states. The trend for weight as a risk factor for development of a “very-large” BML suggests there is potential to identify modifiable risk factors for BML in asymptomatic people and warrants further investigation. © 2009 Elsevier Inc. All rights reserved. Semin Arthritis Rheum 39:213-217 Keywords: knee, bone marrow lesions, osteoarthritis, magnetic resonance imaging B one marrow lesions (BML) have been implicated in the pathogenesis of osteoarthritis (OA) (1-3), yet their exact role and etiology remain unclear. They are present in both symptomatic (2-5) and healthy populations without knee pain or any history of signifi- cant knee injury (6-8). Previous studies on the natural history of BML have ex- amined symptomatic populations with established radio- graphic knee OA (4,5,9-11). In established knee OA, they are associated with pain (2), changes in cartilage metabolism as measured by type II collagen degradation marker CTX-II (4), and increased likelihood of cartilage loss (5,12). There is a paucity of studies examining the natural history of BML in asymptomatic, healthy populations. Moreover, factors asso- ciated with the development of incident BML in healthy populations have yet to be examined. The aim of this study was to examine the natural his- tory of BML in asymptomatic women with no clinical *Department of Epidemiology and Preventive Medicine, Monash University Medical School, Alfred Hospital, Prahran, Victoria, Australia. †Baker Heart Research Institute, Commercial Road, Melbourne, Victoria, Australia. ‡The Women’s Health Program, Department of Medicine, Monash University Medical School, Alfred Hospital, Prahran, Victoria, Australia. §Associate Professor. ¶Professor. Address reprint requests to: Patricia Berry, BBioMedSci (Hons), Department of Epidemiology and Preventive Medicine, Monash University Medical School, Alfred Hospital, Prahran, Victoria, Australia 3181. E-mail: Patricia.Berry@med.monash. edu.au. SJOGRENS SYNDROME 213 0049-0172/09/$-see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.semarthrit.2008.05.003