BACKGROUND: We wanted to compare an ultrasound examination with the bone scintiscan to diagnose stress fractures. METHODS: Using the bone scintiscan as the reference standard, we conducted a prospective, double-blind study of 78 patients (87 percent were men, mean age 24 years) referred for bone scintiscan to rule out tibial stress fractures. After the participants were injected with radionuclide, we examined each tibia once using ultrasound adjusted for an active intensity of 2.0 W/cm2 and again with the wand turned off. The patient was blinded to the mode used. The patient's response to the ultrasound was considered positive if the patient reported pain as the wand passed over the tibia. A bone scintiscan was considered positive according to the criteria of Zwas. One sonography technician performed all examinations; both he and the nuclear medicine department were blinded to the other's findings. The final results were tabulated by a third, uninvolved party. A positive correlation between the scintiscan and ultrasound examination consisted of pain with active ultrasound and any degree of stress fracture in any part of the same tibia as found on the bone scintiscan. RESULTS: Thirty-five stress fractures were found on bone scintiscan, whereas only 15 were detected by ultrasound examination (sensitivity 43 percent). With ultrasound testing there were 22 false positives (specificity 49 percent) and 20 false negatives. These findings resulted in a positive predictive value of 41 percent and a negative predictive value of 51 percent. CONCLUSION: Ultrasound is not reliable in the diagnosis of tibial stress fractures. Bone scintiscan remains the test of choice.
BACKGROUND: We wanted to compare an ultrasound examination with the bone scintiscan to diagnose stress fractures. METHODS: Using the bone scintiscan as the reference standard, we conducted a prospective, double-blind study of 78 patients (87 percent were men, mean age 24 years) referred for bone scintiscan to rule out tibial stress fractures. After the participants were injected with radionuclide, we examined each tibia once using ultrasound adjusted for an active intensity of 2.0 W/cm2 and again with the wand turned off. The patient was blinded to the mode used. The patient's response to the ultrasound was considered positive if the patient reported pain as the wand passed over the tibia. A bone scintiscan was considered positive according to the criteria of Zwas. One sonography technician performed all examinations; both he and the nuclear medicine department were blinded to the other's findings. The final results were tabulated by a third, uninvolved party. A positive correlation between the scintiscan and ultrasound examination consisted of pain with active ultrasound and any degree of stress fracture in any part of the same tibia as found on the bone scintiscan. RESULTS: Thirty-five stress fractures were found on bone scintiscan, whereas only 15 were detected by ultrasound examination (sensitivity 43 percent). With ultrasound testing there were 22 false positives (specificity 49 percent) and 20 false negatives. These findings resulted in a positive predictive value of 41 percent and a negative predictive value of 51 percent. CONCLUSION: Ultrasound is not reliable in the diagnosis of tibial stress fractures. Bone scintiscan remains the test of choice.