Peripheral quantitative computed tomography (pQCT) is increasingly being used to measure bone mineral density (BMD) in both research and clinical practice to monitor BMD changes. Repeated measurements in long-term follow-up study are an appropriate method to study the pattern of bone loss and the diagnostic value critically depends upon the precision (reproducibility). Positioning is one of the sources of imprecision. In this study, BMD at the locations around 4% length of the tibia and radius were measured by pQCT. The relationship between the change of BMD and the change of total cross-sectional-area (CSA) of the bone were analyzed in order to promote the follow-up-reproducibility of pQCT measurements. The results showed a decrease of CSA and increase of trabecular BMD from distal to proximal at the human distal radius, while a consistent decrease of CSA and apparent trabecular BMD from distal to proximal at the distal tibia was observed. It is suggested the follow-up location can be considered as the same location as the baseline measurement at the tibia if the CSA changed within -/+20 mm(2). As to the radius, the criteria are better to be -/+10 mm(2) of the CSA change. Otherwise, it is enough to judge the location only by checking the 4% location when both the 4% and shaft location of the bone are measured at one measurement. And some suggestions are also given to the machine manufacturer.