Normative Standards for HRpQCT Parameters in Chinese Men and Women.

Assessing bone architecture using high-resolution peripheral quantitative computed tomography (HRpQCT) has the potential to improve fracture risk assessment. The Normal Reference Study aimed to establish sex-specific reference centile curves for HRpQCT parameters. This was an age-stratified cross-sectional study and 1072 ambulatory Chinese men (n = 544) and women (n = 528) aged 20 to 79 years, who were free from conditions and medications that could affect bone metabolism and had no history of fragility fracture. They were recruited from local communities of Hong Kong. Reference centile curves for each HRpQCT parameter were constructed using generalized additive models for location, scale, and shape with age as the only explanatory variable. Patterns of reference centile curves reflected age-related changes of bone density, microarchitecture, and estimated bone strength. In both sexes, loss of cortical bone was only evident in mid-adulthood, particularly in women with a more rapid fashion probably concurrent with the onset of menopause. In contrast, loss of trabecular bone was subtle or gradual or occurred at an earlier age. Expected values of HRpQCT parameters for a defined sex and age and a defined percentile or Z-score were obtained from these curves. T-scores were calculated using the population with the peak values as the reference and reflected age- or menopause-related bone loss in an older individual or the room to reach the peak potential in a younger individual. These reference centile curves produced a standard describing a norm or desirable target that enables value clinical judgements. Percentiles, Z-scores, and T-scores would be helpful in detecting abnormalities in bone density and microarchitecture arising from various conditions and establishing entry criteria for clinical trials. They also hold the potential to refine the diagnosis of osteoporosis and assessment of fracture risk. © 2018 American Society for Bone and Mineral Research.