Osteoporotic fractures have substantial clinical and public health impact. Bone quality is an important determinant of fracture risk. Quantitative ultrasound (QUS) of bone measured as broadband ultrasound attenuation (BUA) has been shown to predict fracture risk. However, there have been very few large population studies, particularly in men. We investigated the correlates of calcaneal BUA using a CUBA clinical machine in 15,668 middle and older aged men and women (42-82 years) from the UK, EPIC-Norfolk cohort. At all ages mean BUA was significantly greater in men than women (men, 90.1+/-17.6; women 72.1+/-16.5). The age-related decline in BUA was five times greater in women than men (-0.77 vs. -0.15 dB/MHz per year). Pre- and post-menopausal bone loss was 0.39 and 0.85 dB/MHz per year, respectively. In univariate regression BUA increased with weight and height by 0.45 dB/MHz per kg and 0.68 per cm in women and 0.24 dB/MHz per kg and 0.33 per cm in men. BUA increased with body mass index (BMI) by 0.84 dB/MHz per kg/m2 in women and 0.55 in men. However, weight was twice as influential as height in men and seven times as great in women. Age, weight and height explained 27% of the variance of BUA in women, but only 3% in men. Adjusted BUA was significantly lower in men and women with an existing history of any hip, wrist or spinal fracture both overall and when analysed for specific site. Figures were: all fractures 66.8 vs. 72.5 dB/MHz ( P<0.001), women; 84.1 vs. 90.5 ( P<0.001), men; hip fractures 61.9 vs. 72.2 dB/MHz ( P<0.001), women; 81.5 vs. 90.2 ( P<0.001), men; wrist fractures 66.6 vs. 72.5 dB/MHz ( P<0.001), women; 81.5 vs. 90.2 ( P<0.001), men; spinal fractures 68.1 vs. 72.1 dB/MHz ( P<0.01), women; 85.1 vs. 90.2 ( P<0.01), men. These differences equate to reductions of 14, 9 and 6% and 10, 7 and 6% for fractures of the hip, wrist and spine in the BUA of women and men, respectively. Thus, despite the overall gender difference in BUA the relative magnitude of a previous history of fracture was equally important in both men and women. Adjusted BUA was also lower in those with previous history of osteoporosis. In women currently taking hormone replacement therapy (HRT) the adjusted BUA was 5 dB/MHz or one-third of an SD greater than in those who did not. The BUA of those with a current smoking habit was 1.7% lower in women and 3.2% lower in men. Overall, there are substantial sex differences in the relationship of the physical and osteoporotic risk factors associated with BUA. A better understanding of these determinants of heel ultrasound may provide insights into how some of the sex differences in bone health can be explained and bone loss in later life minimised.