Histological age-at-death estimation in human bone: assessment of inter-population variation

Abstract

Age-at-death estimates in skeletal remains are hampered by observer bias and error, as well as individual and population variation. This study aimed at improving accuracy and reliability of age estimates obtained from applying methods involving the assessment of histomorphometric variables of bone and bone mineral density (BMD). The assessment of osteon population density (OPD) and size (length, surface area and volume) of Haversian systems were investigated in three populations by means of stereology. Bone slides prepared from South African black (n = 99), South African white (n = 94) and Danish white (n = 30) individuals were analysed using MicroBrightField’s StereoInvestigator software. ANCOVA results revealed a statistically significant difference (p < 0.001) between the three groups in terms of OPD. No statistically significant difference was seen in the size of secondary osteons between the groups. Linear regression analysis was used to construct population-specific formulae for age-atdeath estimation in South African white and black individuals. The Danish sample was used as a comparative group for white South African standards established. Age mimicry appeared to play a role in the over-estimation of age in the Danish individuals, after which the problem was corrected by combining the two samples in order to achieve normal age distribution within the larger sample. Secondly, DXA scans of the proximal femur were done for a subgroup of South African black (n = 64) and white (n = 59) individuals for estimation of age-at-death from bone mineral density. Results displayed a significant difference between white and black groups for total and neck BMD. White males and females differed in total and neck BMD, although black males and females differed only in terms of neck BMD. Age could be significantly correlated with BMD in the white population, but not in the black population. Regression analysis was also done for the complete sample (white and black individuals pooled), as ancestry is unknown in some cases. A significant correlation was seen for age versus neck BMD in the total sample and in sex-specific groups. The outcome of this study suggested that the three population samples are somewhat dissimilar in bone microstructure, with differences related to OPD and BMD present between the groups. Based on this outcome, it should be emphasized that age estimation standards should be adapted to accommodate combined groups and create to more generally applicable standards. Further research involving various disciplines is needed to better understand the underlying reason(s) for these differences between these populations.