RESUMEN
In women, heel ultrasound (US) bone mineral density (BMD) has been shown to predict fracture risk, but the usefulness of this screening tool in men is not known. We measured the heel quantitative ultrasound index (QUI( in a convenience sample 185 of men (136 Caucasian, 1 Asian, and 48 African-American) with an average age of 63 yr (range of 25-85) undergoing BMD of the spine and hip by dual X-ray absorptiometry (DXA) to determine whether the heel measurement could predict central BMD. The average DXA T-score was -0.97, -1.20, and -1.61 for the spine, total hip, and femoral neck, respectively. The mean heel US BMD T-score (using the only available T-score, which was defined for Caucasian postmenopausal women) was -0.92. There were significant correlations among the various DXA measurements and the heel US BMD T-score (r = 0.373-0.483, p < 0.001). We defined arbitrarily osteopenia as a spine, total hip, or femoral neck T-score by DXA of < -1.5. We also made two different arbitrary definitions of osteoporosis by DXA: < -2.0 and < -2.5. Using these numbers as disease definitions, we determined the specificity, sensitivity, as well as positive and negative predictive values of using the heel US T-score to predict osteopenia or osteoporosis. Using various cutoffs for the heel T-score, we found that increasing the cutoff toward 0 increased the sensitivity but lowered the specificity. No cutoff was found that provided both good sensitivity and specificity. By analyzing the men by ethnic and age groups, we found that the best set of receiver operating characteristic (ROC) curves was derived from data using heel US to predict osteopenia and osteoporosis in men younger than age 65, although the areas under the ROC curve were approx 0.8. In conclusion, despite a strong correlation between the heel QUI and the spine and hip BMD by DXA, no heel T-score could predict osteopenia or osteoporosis with satisfactory sensitivity and specificity. It is possible that the use of risk factor assessment plus heel QUI might have better predictive value, and further studies are needed to determine whether heel QUI or other US determination is an independent risk factor for fracture in men.