Quantitative computed tomography (QCT) has been found useful in the evaluation of vertebral bone mineral density (BMD). It separates cortical from trabecular bone in the vertebral bodies. The accuracy of QCT, however, is limited because of the existence of unknown amounts of marrow fat in the trabecular bone. The purpose of this study is to investigate the precision and accuracy of QCT, with a particular emphasis on the advantage of dual energy technique over single energy technique, as well as to investigate age-related change of BMD in patients with no metabolic disorders. For evaluation of BMD, SOMATOM DR-H CT scanner was used, which provided dual energy scan with rapid kilovolt peak switching system. The lumbar vertebrae (L1-L3) were exposed at the middle portion together with the calibration phantom (Ca10(PO4)6(OH)2 200 mg/ml, 0 mg/ml). KV-separated images (KV-HI, KV-LO) and material-separated images (MAT-HI, MAT-LO) were calculated from dual-energy scan data. KV-separated data were considered as SEQCT data. In experiment, dipotassium hydrogen phosphate (K2HPO4) solution was used to simulate bone mineral, water to simulate lean soft tissue and ethanol (C2H5OH) to simulate fat. To investigate the variations originating from marrow fat, a series of phantoms (K2HPO4-C2H5OH-H2O mixtures) with a constant 10 gm% K2HPO4 concentration but with the various mixtures of C2H5OH ranging from 0-40% by volume was scanned with dual energy technique. The value of BMD obtained from KV-separated data was reduced about 12% per 10% fat by volume increase, while the value obtained by DEQCT technique was reduced 5.5% per 10% fat by volume increase. With the dual energy technique the error was greatly reduced. On the other hand, the higher precision of KV-separated images compared to MAT-HI images was obtained, as well as the relatively higher precision for determining high rather than low fraction of mineral. In conclusion, with the single energy technique, precision is high, and with the dual energy technique, accuracy is high. It was also noted that selection of the CT section was the most critical factor in clinical BMD analysis. Age-related change of BMD in vertebral trabecular and cortical bone was studied in 161 patients without bone metabolic disorders. There were two peaks of BMD in females in their 20s and 30s, and a single peak in males in their 20s.(ABSTRACT TRUNCATED AT 400 WORDS)