Analgesic effects of etidronate, alendronate and risedronate were compared in patients with osteoporosis and/or osteoarthritis by measuring the fall of skin impedance along with conventional subjective pain-estimation by visual rating scale (VRS). One hundred ninety-nine postmenopausal women consulting the Osteoporosis and Osteoarthritis Clinic of Katsuragi Hospital complaining of back and/or knee pain were randomly divided into four groups; Group A (49 subjects) given 5 mg/day alendronate, Group E (50 subjects) 200 mg/day etidronate, Group R (50 subjects) 2.5 mg/day risedronate and Group P no bisphosphonate. None of the four groups showed significant deviation from others as to age and parameters of bone metabolism. Proportions of subjects with osteoporosis was 18-40%. Those with osteoarthritis of the spine and knee, higher than Grade II according to the Nathan and Lawrence-Kellgren scale, respectively, was 45 and 61%, respectively, without a significant difference among the four groups. Significant positive correlation was found between the fall of skin impedance and pain expressed in VRS. Attenuation of exercise-induced fall of skin impedance and also subjective pain expressed in VRS was greatest in Group E with a highly significant difference from Groups A (P = 0.0002 and P < 0.0001), R (P < 0.0001 and P = 0.0014) and P (P < 0.0001 and P < 0.0001). Neither A nor R showed significant difference from P as to the fall of skin impedance. Among the three bisphosphonates tested, etidronate appeared to be outstanding in analgesic effects.