Introduction: High plasma homocysteine levels have been associated with incident osteoporotic fractures, but the mechanisms underlying this association are still unknown. It has been hypothesized that homocysteine might interfere with collagen cross-linking in bone, thereby weakening bone structure. Therefore, we wanted to investigate whether plasma homocysteine levels are associated with bone quality parameters, rather than with bone mineral density. Methods: Cross-sectional data of the B-PROOF study (n = 1227) and of two cohorts of the Rotterdam Study (RS-I (n = 2850) and RS-II (n = 2023)) were used. Data on bone mineral density of the femoral neck and lumbar spine were obtained in these participants using dual-energy X-ray assessment (DXA). In addition, participants of B-PROOF and RS-I underwent quantitative ultrasound measurement of the calcaneus, as a marker for bone quality. Multiple linear regression analysis was used to investigate the associations between natural-log transformed plasma levels of homocysteine and bone mineral density or ultrasound parameters. Results: Natural-log transformed homocysteine levels were inversely associated with femoral neck bone mineral density in the two cohorts of the Rotterdam Study (B = -0.025, p = 0.004 and B = -0.024, p = 0.024). In B-PROOF, no association was found. Pooled data analysis showed significant associations between homocysteine and bone mineral density at both femoral neck (B = -0.032, p = 0.010) and lumbar spine (B = -0.098, p = 0.021). Higher natural-log transformed homocysteine levels associated significantly with lower bone ultrasound attenuation in B-PROOF (B = -3.7, p = 0.009) and speed of sound in both B-PROOF (B = -8.9, p = 0.001) and RS-I (B = -14.5, p = 0.003), indicating lower bone quality. Pooled analysis confirmed the association between homocysteine and SOS (B = -13.1, p = 0.016). Results from ANCOVA-analysis indicate that differences in SOS and BUA between participants having a plasma homocysteine level above or below median correspond to 0.14 and 0.09 SD, respectively. Discussion: In this study, plasma levels of homocysteine were significantly inversely associated with both bone ultrasound parameters and with bone mineral density. However, the size of the associations seems to be of limited clinical relevance and may therefore not explain the previously observed association between plasma homocysteine and osteoporotic fracture incidence. (C) 2014 Elsevier Inc. All rights reserved.