A large proportion of patients with asymptomatic primary hyperparathyroidism (PHPT) have some degree of bone involvement which appears to be relatively greater for cortical than trabecular bone. However, the clinical meaning and pathophysiologic basis of this observation are unknown. In 77 postmenopausal women with asymptomatic PHPT, bone mineral density (BMD) was measured at the proximal and ultradistal forearm, the lumbar spine, the femoral neck, and Ward's triangle by dual-energy X-ray absorptiometry. The digitalized X-ray pictures of the nondominant hand were obtained from all patients and from 680 healthy postmenopausal women, to measure the outer (D) and inner (d) diameter of the second metacarpus. The cortical area per total area (CA/TA) and a bending breaking resistance index (D4-d4/D) were then calculated. In 29 of the patients not operated on and in 30 healthy pair-matched women, a second X-ray of the hand was obtained 5-12 years afterward. In patients with PHPT, the z score of CA/TA was significantly lower than zero [-0.97+/-0.99, standard deviation (SD)]. This is due to an enlargement of the inner diameter, despite a significant increase in the z score for the outer diameter. The z score of the DXA measurements was significantly lower than zero for the lumbar spine (-0.59+/-1.26), ultradistal radius (-1.03+/-0.91), proximal radius (-1.91+/-1.80), and Ward triangle (-1.81+/-1.07), but not for the femoral neck (-0.36+/-1.03). In subjects in whom two X-rays were obtained, per-decade endosteal resorption and periosteal apposition were statistically significant only in the PHPT patients. Both endosteal resorption and periosteal apposition were significantly greater in PHPT patients compared to healthy controls. The mean BBRI in PHPT patients was not different from that in controls, but the longitudinal changes were significantly greater than those observed in control subjects. Our radiogrammetry data may provide an original clue for understanding preferential cortical bone loss in PHPT patients. In cross-sectional and longitudinal studies, we have shown that in PHPT, both endosteal bone resorption and periosteal apposition are augmented. The former effect is predominant, which leads to significant diminution of cortical thickness. As a consequence of the enlargement of long bones, the areal BMD is somewhat underestimated, since the same amount of cortical bone is divided by a greater diameter. Furthermore, in term of mechanical properties, the increases in the cross-sectional area of appendicular bone segments might compensate in part for both the diminution of cortical thickness and a greater porosity of cortical bone.