Statistical Challenges in the Analyses of Bone Mineral Density in Pediatric Populations

• Published in: Therapeutic innovation & regulatory science Vol. 49; no. 4; pp. 569 - 575
• Main Author: Dinh, Phillip
• Format: Journal Article
• Language: English
• Published: Los Angeles, CA SAGE Publications 01.07.2015; Springer International Publishing; Springer Nature B.V
• Subjects: Adjustment; Age; Authorship; Bone density; Clinical medicine; Datasets; Drug Safety and Pharmacovigilance; Ethnicity; Nutrition research; Osteoporosis; Pediatrics; Pharmacotherapy; Pharmacy; Population; Special Populations: Analytical Report; Standard deviation; Values; Women; bone mineral density; T-score; Z-score; DXA; score
• ISSN: 2168-4790; 2168-4804; 2168-4804
• DOI: 10.1177/2168479015570334

Assessing bone mineral density (BMD) by dual-energy x-ray absorptiometry (DXA) is standard in clinical practice with good precision and reproducibility. One way to analyze BMD data is through the Z-score where a subject’s BMD is standardized against some reference population. This article highlights several potential problems in the calculation of the Z-score: (1) reference data are manufacturer dependent; (2) a transformation is recommended to reduce the skewness often seen in BMD data, however a transformation parameter is only available in some references using the Hologic manufacturer; (3) some reference data may be obsolete, as they were done under older scan models; (4) reference data are recommended to be interpolated to the subject’s age, and linear interpolation is questionable; (5) reference data do not exist for all age groups; (6) reference data depend on the race of the subjects, and not all races are representative; and (7) reference data are limited to certain body composition. Given the limitation in the BMD Z-scores, other clinical symptoms (eg, bone fractures) or laboratory findings (eg, bone biomarkers) should be assessed in conjunction with the BMD Z-score to evaluate a subject’s bone health.