The absorptiometry T-score: influence of selection of the reference population and related considerations for everyday practice

• Published in: Joint, bone, spine : revue du rhumatisme bone, spine : revue du rhumatisme, 2003-08, Vol.70 (4), p.290-293
• Main Authors: Levasseur, Régis, Guaydier-Souquières, Geneviève, Marcelli, Christian, Sabatier, Jean-Pierre
• Format: Article
• Language: English
• Subjects: Absorptiometry; Absorptiometry, Photon; Adolescent; Adult; Aged; Aged, 80 and over; Bone Density; Bone mineral density; Data Interpretation, Statistical; Female; France - epidemiology; Humans; Lumbar Vertebrae - diagnostic imaging; Lumbar Vertebrae - metabolism; Middle Aged; Normative data; Osteoporosis; Osteoporosis - diagnosis; Osteoporosis - epidemiology; Premenopause; Reference data; Reference Values; Reproducibility of Results; Selection Bias; Sensitivity and Specificity; T-score
• ISSN: 1297-319X
• DOI: 10.1016/S1297-319X(03)00042-3

Objectives. – The WHO definition of osteoporosis is based on T-scores, which are calculated from mean bone mineral density values in reference populations composed of young adults. The definition of this reference population is a key-point. Methods. – We compared lumbar spine T-scores in three reference populations of young adults, one from the US and two from France. Results. – Reference values were higher in the American population than in the two French populations. As a result, the prevalence of a T-score indicating osteoporosis in a population of 2887 patients aged 20–87 years (mean age, 62.4 ± 9.5 years) was 32% with the American reference population and 22% and 24% with the two French populations. Thus, about 25% and 32% of patients classified as having osteoporosis with the American reference population did not have osteoporosis with the two French reference populations, respectively. There was no significant difference between the prevalences obtained with the two French populations. Conclusions. – Reference populations of young adults should be representative of the population of each country, for each measurement site, gender, and ethnic group. When evaluating individuals, in addition to the clinical setting and T-score, age should be taken into account. The Z-score remains useful.