Objective: The objective of this study was to systematically review the literature to determine whether obesity and systemic factors, including age, gender, heritability, dietary factors, smoking, serum and urine biomarkers of cartilage or bone metabolism, bone-related factors, and medication, are associated with knee bone marrow lesions (BMLs) identified on magnetic resonance imaging in asymptomatic pre-osteoarthritis and osteoarthritis populations. Methods: Electronic searches of MEDLINE and EMBASE were performed from January 1, 1996 to September 30, 2012 using the following keywords: bone marrow lesion(s), bone marrow (o)edema, osteoarthritis, and knee. Studies examining obesity and non-biomechanical factors in relation to the presence, incidence, or change in BMLs were included. Two independent reviewers extracted data and assessed methodological quality of selected studies. Due to the heterogeneity of the studies, we performed a best evidence synthesis. Results: Among 30 studies, 17 were considered high quality. The study populations were heterogeneous in terms of symptoms and radiographic knee osteoarthritis. There was strong evidence for an association between serum lipids and BMLs and no association between age and BMLs. There was moderate evidence for a relationship between obesity and BMLs. There was limited evidence for gender, smoking, C-telopeptide of type I collagen, anti-bone-resorptive treatments, licofelone, and chondroitin sulfate. There was a paucity of evidence for heritability and conflicting evidence for dietary fatty acids. Conclusion: There is strong evidence for serum lipids and moderate evidence for obesity as risk factors for knee BMLs. Given the role of BMLs in the pathogenesis of knee osteoarthritis, identification of modifiable risk factors of BMLs and therapeutic interventions targeting BMLs has the potential to reduce the burden of knee osteoarthritis. ? 2013 Elsevier Inc. All rights reserved.