ORIGINAL ARTICLE The impact of the Hologic vs the ICMR database in diagnosis of osteoporosis among South Indian subjects Sahana Shetty, Nitin Kapoor, Dukhabandhu Naik, Hesarghatta S Asha, Nihal Thomas and Thomas V Paul Department of Endocrinology, Diabetes and Metabolism, Christian Medical College, Vellore, India Summary Background and Objectives Recently, the Indian Council of Medical Research (ICMR) has published normative data for bone mineral density (BMD) measured by dual-energy X-ray absorptiometry (DXA) scanning. However, the impact this has had on the diagnosis of osteoporosis when compared to cur- rently used Caucasian databases has not been analysed. Hence, this study was undertaken to look at agreement between the Ho- logic Database (HD) based on BMD normative data in Cauca- sians and the ICMR database (ICMRD) in defining osteoporosis in subjects with or without hip fracture. Materials and Methods It is a cross -sectional study of 2976 subjects (men 341, women 2757) (mean age Æ SD = 62Á2 Æ 7Á2 years), including 316 subjects with low impact hip fracture: 2199 were from the hospital database, and 461 were healthy postmenopausal women from the community who underwent (DXA) scanning between January 2010 and March 2013. Recal- culated T scores from ICMRD were used for the diagnosis of osteoporosis and compared with HD. Results An almost perfect agreement existed between the two databases for the diagnosis of osteoporosis at the hip (j À0Á82, P < 0Á0001) in all subjects, and a moderate relationship existed in those with hip fracture (j À0Á65, P < 0Á0001). Seventy-three of 316 hip fracture subjects (23Á5%) defined as osteoporosis according to HD were classified as osteopenia according to IC- MRD. Conclusion The threshold of hip BMD T score for treating osteoporosis may have to be redefined if the ICMR reference database is used. Initiation of treatment in these subjects must be based on multiple fracture risk factor assessment in addition to looking at BMD. Further studies with a larger sample size of subjects with fracture are needed to validate our findings. (Received 8 January 2014; returned for revision 25 January 2014; finally revised 18 April 2014; accepted 6 May 2014) Introduction Osteoporosis is the common metabolic bone disease, which has a significant economic burden globally. The current population- based prevalence of people with low bone mineral density in the United States is 57 million, as demonstrated in the most recently released data by the National Osteoporosis Foundation, which also showed a huge variation in the prevalence among different ethnicities. 1 India is one of the largest affected countries in the world where one out of every eight men and one out of every third woman has osteoporosis. 2 An earlier peak incidence of osteoporo- sis has also been reported in Indians as compared to the western subjects who usually develop it one to two decades later. 1,2 Peak bone mass (PBM) is the maximum bone mass present at the end of skeletal maturation which is acquired by third decade of life. The PBM depends on genetic, nutritional and other envi- ronmental factors and thus differs in different ethnicities. 3 These factors may explain the variations seen in the bone mineral den- sity (BMD) across different adult populations. The dual-energy X-ray absorptiometry (DXA) remains the main diagnostic modality for diagnosing osteoporosis. 4 The current DXA machines have incorporated Caucasian-based BMD as normative data for the diagnosis of low bone mass. For example, the refer- ence database used in the Hologic machine is based on the data from 14 646 men and women, nationally representative U.S. adults collected during the National Health and Nutrition Exam- ination Survey III (NHANES-III, 1988–94). 5 The relevance of using Caucasian reference database for diag- nosing osteoporosis in other populations has been questioned in view of ethnic differences in attainment of peak bone mass. Also, most published Indian studies have shown a low bone mass across all age groups. This necessitated the Indian Council of Medical Research (an apex body in India for conducting medical research) to derive a reference data for bone mineral density. This was a multicentre study conducted between 2001– 2006 to arrive at the reference BMD at hip, forearm and spine in healthy Indian individuals (n = 808) aged 20–29 years. 6 However, the performance of ICMR database in comparison with Caucasian database in categorizing BMD into various groups has not been studied. So, we attempted to look at the agreement in classifying BMD by the T score derivations from Hologic Data- base (HD) and the Indian Council of Medical Research Database (ICMRD) in a south Indian semi-urban population. More Correspondence: Dr Thomas V Paul, Professor, Department of Endocri- nology, Diabetes and Metabolism, Christian Medical College, Ida Scud- der Road, Vellore 632 004, India. Tel.: +91 416 2283118; Fax: +91 416 4203570; E-mail: thomasvpaul@yahoo.com © 2014 John Wiley & Sons Ltd 519 Clinical Endocrinology (2014) 81, 519–522 doi: 10.1111/cen.12495