1How and why can the analysis of anthropobiological remains be of interest in understanding the current state of health of human populations? The 2019 Revision of World Population Prospects foresees that the number of people aged 65 or over in the world will increase from 702.9 million in 2019 to 1,548.9 million in 2050, a percentage change of 120% (United Nations, 2019). Although specific regional circumstances could lead to significant differences, a corollary question inevitably arises: with the dramatic changes to our way of life over the past several decades, do humans still age the same way as before? More specifically, is it possible to observe secular trends in age-related biological parameters in Western European societies? Among these parameters, bone mineral density (BMD) of the total hip decreases with age. In particular, the contribution of post-menopausal bone loss to age-related bone loss renders females typically more susceptible to osteoporosis-related fractures in comparison to males (Stini, 2003; Brickley and Ives, 2008), a striking feature in northern European societies. Moreover, besides sex differences, not all populations show the same bone architecture, BMD or fragility (Nelson and Megyesi, 2004; Ruffing et al., 2006; Araujo et al., 2007; Looker et al., 2009; Nam et al., 2010; Wang and Seeman, 2012). Bone density can thus help to predict the risk of fractures, but age-related bone changes exhibit population-related variations in terms of incidence or degree, and this needs to be taken into account (Walker et al., 2012; Molleson and Orban, 2019).
2Among the many factors that contribute differently to BMD in ethnic groups, and in males and females are body size and body composition (body mass index), haematological parameters, hormonal status, use of oral contraceptives, number of pregnancies, genetic factors such as a family history of osteoporosis, diet (calcium intake), frequency and duration of exercise, occurrence of fracture or prior fracture, as well as lifestyle and socioeconomic status (Gur et al., 2003; Ruffing et al., 2006; 2007; Rautava et al., 2007; Morin et al., 2009; Scholtissen et al., 2009; Bow et al., 2011; Wei et al., 2011; Chantler et al., 2012; Borrè et al., 2015). Combinations of these factors could characterise populations located in different geographical areas, which may explain why indirect factors, such as climate (Rozman et al., 2003), have also been suggested as factors influencing BMD.
3To evaluate bone senescence, the most common method is X-ray absorptiometry (Laval-Jeantet et al., 1995), which measures bone mass and bone density. Measurements of anatomical variations related to aging can also be studied using morphometrics. For instance, the cross-sectional geometry of long bones appears to be a good indicator of the influential mechanical forces and a reasonable reflection of habitual activities (Pearson and Lieberman, 2004). Molleson and Orban (2019) therefore used this measure to assess the relationship between ageing and bone remodelling and thinning.
4This study examines the same collection of human skeletons as that studied by Molleson and Orban (2019) and provides an update of Gilissen and Orban (2022) by extending the comparison to non-human primates. This collection of known age individuals represents a well-documented skeletal sample derived from Schoten, a village in the outer suburbs of Antwerp, Belgium. The aim of the study is to illustrate the dynamics of BMD at the femoral region in males and females with reference to current standards and other samples.
5The limitations of the Schoten collection are detailed in Orban et al. (2011). The collection is small compared to other European osteological collections containing specimens of known (or estimated) age and sex and covering the period between the 11th and 16th centuries AD (Wharram Percy burials, North Yorkshire, England, Mays et al., 2007) or between the 18th and 19th centuries AD (Christ Church Spitalfields, London, England, Molleson and Cox, 1993). It should therefore be kept in mind that a study on this collection stands between a clinical case study and a population survey (Molleson and Orban, 2019). The analysis of this material is however of high interest for population health history (Walker, 2000). In particular, of the nine collections detailed in Orban et al (2011), only four (including Schoten) come from Northern Europe.
6Details of the socioeconomic context of Schoten in the 19th and 20th centuries, the age ranges, sex, health status, occupations and social situation of the inhabitants including the history of the acquisition of the specimens by the Royal Belgian Institute of Natural Sciences in Brussels, are provided by Orban et al. (2011), Molleson and Orban (2019), and Gilissen and Orban (2022). In the second half of the 19th century, the development of means of communication and the beginnings of industrialisation opened up the village of Schoten, and profound transformations marked its growth. In the space of a century, the village grew from a handful of houses and farms to a larger urban agglomeration. Rural Schoten has become a semi-industrial region. The socio-demographic evolution of the population is reflected in the increased number of inhabitants (from 1,408 in 1830 to 13,762 in 1931), the evolution of industry-related occupations and the proletarianisation of the population (Orban et al., 2011). The series consists of 48 adult skeletons (26 males and 22 females) and three juveniles (one 19 year old male, one 19 year old female and one 15 year old female). Orban et al. (2011) include a morphological analysis of the skeletal material and consider the series potential as a reference collection.
7Year of birth, sex and place of birth, but not the names, of each individual are documented. The registers do not include the causes of death. Briefly, the individuals, born between 1837 and 1916 within a radius of 50 km around Schoten, a suburb of Antwerp, Belgium, died in 1930 (N=5) or 1931 (N=46) and were buried in the old cemetery at Schoten. The last burial in the old cemetery took place on 24th June 1931, after which there were burials only in family plots. A new cemetery became operational from 29th June 1931. In 1946, the old cemetery was excavated (Orban et al., 2011; further details about Schoten’s cemeteries in Chabot and Camp, 2000) and the human remains were exhumed by the staff of the Anthropology section, under the supervision of F. Twiesselmann, Head of section, Anthropology and Prehistory, Royal Belgian Institute of Natural Sciences, Brussels, Belgium, under existing Belgian legislation (Orban et al., 2011; Molleson and Orban, 2019). Storage is by anatomical part in large wooden drawers, with fitted glass topped lids. The bones have been chemically treated, most probably with shellac flakes dissolved in alcohol (Orban et al., 2011). Compared with the important Spitalfields collection of identified skeletons (Molleson and Cox, 1993), where only 6% of the individuals are over 80 years old, the distribution of the individual ages at death is peculiar in the Schoten series, with an over-representation of older individuals. Twelve are older than 80 years (27% of the sample), sixteen are between 61 and 80 years old, thirteen are between 41 and 60 years old, seven are between 21 and 40 years old, and only three are under 20 years old.
8Together with radiogrammetry, the most commonly used technique for studying bone loss in past populations is Dual Energy X-ray Absorptiometry (DEXA) (Curate, 2014). In this study, BMD was measured through DEXA (or DXA). This technique measures the attenuation of an incident X-ray beam through the tissues, which is a function of the density of these tissues. More precisely, two energy beams of different intensity are sent through the body. The two measurements allow the assessment and elimination of the absorption component of soft tissues (muscle and fat). An image, in orthogonal projection with respect to the scanned surface, is reconstructed from the measured absorptions and is used to select the region of interest for measurement (target area, figure 1). Absorption through the radiographed bone is used to evaluate the bone mineral content, which affects the attenuation of the radioactive beam. More precisely, DEXA calculates the amount of hydroxyapatite in bone, expressing it in grams of mineral per area unit (table 1; figure 2). It therefore calculates the mass in a cross-sectional area of the bone rather than a mass per unit volume (Freitag and Barzel, 2002). The result of the measurement is the total bone mass of the bone region of interest, expressed in grams of hydroxyapatite or surface mass in grams per cm2 corresponding to the ratio of the bone mass to the surface of the projected image (Laval-Jeantet et al., 1995).
Figure 1
Image of the femoral proximal region, in orthogonal projection with respect to the scanned surface, used to select the measurement target areas. Target areas correspond to the yellow contours: T, trochanter; iT, intertrochanter; Fn, femoral neck (rectangle); Wt, Ward’s triangle (square). The total region covers the trochanter, the intertrochanter and the femoral neck |
Image de la région proximale du fémur en projection orthogonale par rapport au balayage de la surface et sélection des régions d’intérêt pour les mesures. Les régions d’intérêt correspondent aux contours jaunes : T, trochanter ; iT, intertrochanter ; Fn, col fémoral (rectangle) ; Wt, triangle de Ward (carré). La région totale couvre le trochanter, l’intertrochanter et le col du fémur
Table 1
DEXA report for a bone region (total proximal left femoral region) in an adult female from the Schoten series. The target areas are illustrated in figure 1. Basic results and the WHO classification (diagnosis of normality) for this individual are presented. Target area surface in cm2; BMC, bone mineral content (g); BMD, bone mineral density (g/cm2); T-score and Z-score (see methods). PR (Peak Reference) and AM (Age Matched) are not used in this study. See figure 2 for a plot of the total region value |
Rapport DEXA pour une région osseuse (ici, la région fémorale gauche proximale totale) chez une femme adulte de la série de Schoten. Les régions d’intérêt (ROI) sont illustrées sur la figure 1. Les résultats de base et la classification de l’OMS (diagnostic de normalité) pour cet individu sont présentés. Surface ROI en cm2 ; BMC, contenu minéral osseux (g) ; BMD, densité minérale osseuse (g/cm2) ; T-score et Z-score (voir méthodes). PR (Peak Reference) et AM (Age Matched) ne sont pas utilisés dans cette étude. Voir la figure 2 pour un graphique de la valeur totale de la région
Figure 2
Bone mineral density (BMD) values at total proximal femoral region (see figure 1) against the age-related normal range in an adult female. The reference database for the plot is from NHANES. See table 1 for the values |
Valeur de la densité minérale osseuse (BMD) dans la région fémorale proximale totale chez une femme adulte (voir figure 1) par rapport à la variabilité normale liée à l’âge. La base de données de référence pour le graphique est celle de NHANES. Voir le tableau 1 pour les valeurs
9In a clinical context, studies of BMD in the lumbar spine are probably the most common and have also provided significant results in anthropological studies (Bouchez et al., 2011). The proximal femur provides robust results that are independent of the DEXA analysis system (Fan et al., 2010) and preserves generally better than the lumbar spine in archaeological contexts. In addition, its positioning in the densitometer is much simpler. To determine BMD in the Schoten series, we analysed the proximal region of the left femur of each individual with a standard densitometer manufactured by Hologic (Hologic QDR-2000 series bone densitometer, Hologic, Bedford MA, USA). The bone was placed anterior-surface uppermost, so that the femoral neck lay flat, and the diaphysis was oriented parallel to the axis of the scanner following the protocol of Mays et al. (2006). Measurements were made at multiple skeletal sites including femoral neck, Ward’s triangle, trochanter and intertrochanter regions, and total hip (the latter corresponds to the femoral proximal metaphysis and does not include the femoral head) (figure 1). Trochanter and intertrochanter region measurements are not illustrated here, they are included in Gilissen and Orban (2022).
10Bone mineral density analysis of dry specimens requires the use of a soft tissue equivalent. For this purpose, researchers have used various products such as flour or ethanol gel. Most authors use water (Lees et al., 1993; Fulpin et al., 2001) or dry rice (Mays et al. 1998; 2006). Bones were placed in a plastic box containing dry rice as a soft-tissue substitute. We avoided additional product manipulation in our analysis, data acquisition has been made as consistent as possible. It must be highlighted here that clinical work confirms that the reliability of DEXA measurements is not influenced by whether or not the bone is covered by soft tissue (Dr. Eric Laurent, personal communication), although in most cases, according to our experience, the calibration of the device and hence the BMD measurements are not possible without a soft tissue equivalent.
11Bone mineral density is expressed as a T-score, i.e., the number of standard deviations (SD) from the young normal adult mean (a T-score of zero) or as a Z-score, i.e., the number of standard deviations from the age-matched mean. Bone mineral density is used to determine bone quantity, a correlate of the risk of osteoporosis and its precursor osteopenia (decrease in bone tissue without risk of fracture) (see Brickley and Ives, 2008, for a review). T-scores between -1 and -2.5 indicate low bone mass or osteopenia (up to 11% lower BMD). T-scores below -2.5 SD indicate excessive fragility or osteoporosis (BMD loss above 11%) (World Health Organization, http://courses.washington.edu/bone-phys/index.html). For the diagnosis of osteoporosis, Lewiecki et al. (2004) recommend considering the lowest T-score of the femoral neck or the total hip. Table 1 and figure 2 show an example of a summary of DEXA results for a bone region (total proximal femoral region) in an adult female from the Schoten series, as well as a diagnostic classification from the World Health Organization (WHO). The reference database is from the National Health and Nutrition Examination Survey (NHANES) available in 2003 and was provided by L. Gantois (Institut Jules Bordet, Brussels). It represents a population of individuals of European ancestry between 20 and 85 years old.
12Table 2 gives descriptive statistics (mean BMD values and standard deviations) at the different femoral regions for females and males and for age categories. Figures 3-8 show plots of BMD (g/cm2) for different femoral regions (see figure 1) as a function of age (years) in females and males from the Schoten collection (circles) together with data from a reference group of the same age and sex (squares, NHANES reference group, represented here by means and standard deviations). This paper presents the results pertaining to the total femoral proximal region, the femoral neck, and Ward’s triangle. Gilissen and Orban (2022) show additional plots of BMD (g/cm2) for the trochanter and the inter-trochanter. No femoral fractures were identified in any of the subjects. Raw data available under request to the authors.
Table 2
Femoral region mean bone mineral density (BMD in g/cm2) values and standard deviations (SD) for each sex and age group (in years) |
Moyennes des valeurs de densité minérale osseuse (BMD en g/cm2) avec l’écart type (SD) pour chaque sexe et chaque groupe d’âge exprimé en années
Figure 3
Bone mineral density (BMD in g/cm2) at the total femoral proximal region (see figure 1) as a function of age (yrs=years) in females from the Schoten collection (circles) together with data from a reference group of the same age and sex (squares, NHANES reference group, means and standard deviations). X-ray: L. Gantois and M. Lemort (Jules Bordet Institute, Brussels) |
Densité minérale osseuse (BMD en g/cm2) de la région fémorale proximale totale (cf. figure 1) en fonction de l’âge (yrs=ans) chez les sujets féminins de la collection de Schoten (cercles) représentée avec le groupe de référence pour le même sexe et catégorie d’âge (carrés, groupe de référence NHANES, moyennes et écarts-types). Radiographie : L. Gantois et M. Lemort (Institut Jules Bordet, Bruxelles)
Figure 4
Bone mineral density (BMD in g/cm2) at the total femoral proximal region (see figure 1) as a function of age (yrs=years) in males from the Schoten collection (circles) together with data from a reference group of the same age and sex (squares, NHANES reference group, means and standard deviations). X-ray: L. Gantois and M. Lemort (Jules Bordet Institute, Brussels) |
Densité minérale osseuse (BMD en g/cm2) de la région fémorale proximale totale (cf. figure 1) en fonction de l’âge (yrs=ans) chez les sujets masculins de la collection de Schoten (cercles) représentée avec le groupe de référence pour le même sexe et catégorie d’âge (carrés, groupe de référence NHANES, moyennes et écarts-types). Radiographie : L. Gantois et M. Lemort (Institut Jules Bordet, Bruxelles)
Figure 5
Bone mineral density (BMD in g/cm2) at the femoral neck (see figure 1) as a function of age (yrs=years) in females from the Schoten collection (circles) together with data from a reference group of the same age and sex (squares, NHANES reference group, means and standard deviations). X-ray: L. Gantois and M. Lemort (Jules Bordet Institute, Brussels) |
Densité minérale osseuse (BMD en g/cm2) de la région du col fémoral (cf. figure 1) en fonction de l’âge (yrs=ans) chez les sujets féminins de la collection de Schoten (cercles) représentée avec le groupe de référence pour le même sexe et catégorie d’âge (carrés, groupe de référence NHANES, moyennes et écarts-types). Radiographie : L. Gantois et M. Lemort (Institut Jules Bordet, Bruxelles)
Figure 6
Bone mineral density (BMD in g/cm2) at the femoral neck (see figure 1) as a function of age (yrs=years) in males from the Schoten collection (circles) together with data from a reference group of the same age and sex (squares, NHANES reference group, means and standard deviations). X-ray: L. Gantois and M. Lemort (Jules Bordet Institute, Brussels) |
Densité minérale osseuse (BMD en g/cm2) du col fémoral (cf. figure 1) en fonction de l’âge (yrs=ans) chez les sujets masculins de la collection de Schoten (cercles) représentée avec le groupe de référence pour le même sexe et catégorie d’âge (carrés, groupe de référence NHANES, moyennes et écarts-types). Radiographie : L. Gantois et M. Lemort (Institut Jules Bordet, Bruxelles)
Figure 7
Bone mineral density (BMD in g/cm2) at Ward’s triangle (see figure 1) as a function of age (yrs=years) in females from the Schoten collection (circles) together with data from a reference group of the same age and sex (squares, NHANES reference group, means and standard deviations). X-ray: L. Gantois and M. Lemort (Jules Bordet Institute, Brussels) |
Densité minérale osseuse (BMD en g/cm2) du triangle de Ward (cf. figure 1) en fonction de l’âge (yrs=ans) chez les sujets féminins de la collection de Schoten (cercles) représentée avec le groupe de référence pour le même sexe et catégorie d’âge (carrés, groupe de référence NHANES, moyennes et écarts-types). Radiographie : L. Gantois et M. Lemort (Institut Jules Bordet, Bruxelles)
Figure 8
Bone mineral density (BMD in g/cm2) at Ward’s triangle (see figure 1) as a function of age (yrs=years) in males from the Schoten collection (circles) together with data from a reference group of the same age and sex (squares, NHANES reference group, means and standard deviations). X-ray: L. Gantois and M. Lemort (Jules Bordet Institute, Brussels) |
Densité minérale osseuse (BMD en g/cm2) du triangle de Ward (cf. figure 1) en fonction de l’âge (yrs=ans) chez les sujets masculins de la collection de Schoten (cercles) représentée avec le groupe de référence pour le même sexe et catégorie d’âge (carrés, groupe de référence NHANES, moyennes et écarts-types). Radiographie : L. Gantois et M. Lemort (Institut Jules Bordet, Bruxelles)
13Table 3 presents a correlation matrix for the BMD values of the femoral target areas. The data comes from the females and males of the Schoten population. Table 4 suggests a diagnosis of osteopenia or osteoporosis based on femoral BMD values for the individuals of Schoten.
Table 3
Correlation matrix for the bone mineral density values of the femoral target areas. Data from the Schoten population, both sexes |
Matrice de corrélation des valeurs de densité minérale osseuse pour les différentes régions fémorales. Données de la population de Schoten, sujets féminins et masculins
Table 4
Diagnosis based on femoral bone mineral density values for the Schoten population, both sexes. No osteoporotic fractures were observed. Diagnosis according to WHO classification |
Diagnostic basé sur les valeurs de densité minérale osseuse fémorale pour la population de Schoten, sujets féminins et masculins. Aucune fracture ostéoporotique n’a été constatée. Diagnostic selon la classification de l’OMS
14It is important to remember that the individuals from the Schoten series were all buried in the old Schoten cemetery. It follows that this collection is rich in elderly individuals, more than octogenarians, and that the juvenile group is small (N=3). When investigating the health status of past populations, the sample size of such material will almost always be smaller than those in epidemiological studies on living subjects, with a consequent lack of statistical power. However, these studies have the advantage of determining whether some of the differences in BMD between populations observed today also existed in European populations in the past (Mays et al., 2006).
15Concerning the putative diagenesis due to the archaeological nature of the material, careful examination of comparable material from Medieval population samples of Trondheim (Norway) and Wharram Percy (England) led Mays et al. (2006) to the conclusion that the BMD values in these samples are likely representative of in vivo values, rather than being artefacts of diagenesis. This conclusion can certainly be applied to the Schoten material for which details on postmortem preservation are provided by Orban et al. (2011) and Molleson and Orban (2019). In addition, Curate (2014) reviewed both direct and indirect evidence suggesting that the diagenetical effects on bone mineral content are only marginal.
16The skewed age distribution of the Schoten sample provides an opportunity to address the emergence of new epidemiological problems such as osteoporosis, which is becoming increasingly common with age, especially in socioeconomically advantaged populations (Orban et al., 2011). More women experience fractures than the combined number of women who experience breast cancer, myocardial infarction and coronary death in one year, across ethnic and racial groups (Cauley, 2011). Prevention efforts could benefit from the study of secular trends in bone biology in various human populations. Moreover, BMD has been more investigated in the metacarpal bones than in the femoral region of ancient skeletal remains (Mays, 1996; 2000; 2001; 2006) probably because of the usual paucity of the latter material.
17The BMD values of Schoten’s males and females are within the range of the current reference population (NHANES). More precisely, individuals aged over 65 years appear to be well distributed around the mean (see also Gilissen et al., 2006; Orban et al., 2011; Molleson and Orban, 2019; Gilissen and Orban, 2022). In addition, although the juveniles in our sample are represented by only two females and one male, they could provide important values. Indeed, peak BMD appears to be fully attained at the time of late adolescence, as is the case in today’s populations. Peak BMD at skeletal maturity may be the single most important factor in the development of osteoporosis (Hernandez et al., 2003). On average, it is reached at 12 years and generally before 16 years at the femoral neck in men and women of European ancestry (Henry et al., 2004; Berenson et al., 2009).
18When considering possible differences between males and females, an asymmetry in data distribution can be observed in the Schoten females, whose BMD values up until the age of 50 appear higher (above the mean) than the average values of today’s NHANES reference population. Conversely, males are more evenly distributed around the mean (figures 3-8). These observations contrast with those of Srinivasan et al. (2012) on a cohort of young and old male and female individuals from Rochester, MN, USA, where femoral strength tended to be relatively similar across the sexes. Our observation is consistent with Molleson and Orban (2019), who scored Schoten femoral radiographs according to the criteria in Acsádi and Nemeskéri (1970), and who observed that thinning of the femoral cortex was detected only in a few individuals, mostly women, aged 65 or more (Molleson and Orban, 2019). This suggests that rates of bone loss were slower in this sample of women than in the modern reference sample. Interestingly, no osteoporotic fractures were observed in the Schoten sample.
19Lifestyle may have been a correlated factor contributing to the good health condition of younger Schoten individuals, especially females. Other results however suggest that when considering age-related bone loss, lifestyle factors may be less important in influencing the severity of osteoporosis. For instance, at medieval Wharram Percy, England (a homogeneous, rural peasant population, 11th-16th centuries AD), there was also significant age-related loss of bone substance in the femur, but the degree of loss was no less than, and perhaps even exceeded, that seen in recent subjects (Mays, 1996; Mays et al., 1998). It should be noted that in the Schoten sample, in contrast with the apparent good bone health of the younger female individuals, four females over 60 years old could be diagnosed with osteoporosis.
20Another difference between males and females lies in the variability of BMD values. Figures 3-8 show that the reference population’s standard deviations of BMD are larger in males than in females for every age range. Bone mineral density values for both male and female Schoten individuals fall within the range of the reference population and the variability of Schoten male BMD values even exceeds somewhat the range of variability in the reference sample. The Schoten female BMD values cover a narrower spectrum of variation at all sites of measurement. Mean BMD values are however higher in males than in females, which may compensate for the difference in variability between the sexes. However, table 3 shows better correlations between the BMD of the different femoral regions in females than in male individuals, in concordance with a higher BMD variability in males, and suggesting a more homogeneous bone structure in females than in males.
21A third difference between male and female individuals concerns the diagnosis of osteopenia and osteoporosis (table 4), where sex differences can clearly be observed. This concords with recent studies on sex differences in bone architecture, which indicate that several biomechanical variables lead to increased bone strength in males compared to females (Nelson and Megyesi, 2004). In the case of females, the number of pregnancies has an impact on BMD in postmenopausal women (Gur et al., 2003). This relationship varies with age and shows differences between skeletal sites. There is no additional data to further analyse the role of this parameter. For the diagnosis of osteoporosis, Aoki et al. (2000) consider that BMD measurements are significantly more sensitive at Ward’s triangle than at the femoral neck, trochanter, intertrochanteric region and total hip sites. This appears to be confirmed by the Schoten sample analysis. Interestingly, it is BMD at Ward’s triangle (and at the trochanter), but not at the femoral neck, that shows significant negative correlations with the number of pregnancies in postmenopausal women (Gur et al., 2003). At Ward’s triangle, BMD values decrease at a higher rate with age than at other target areas, as illustrated in figure 7. In fact, the NHANES reference database shows this trend and the Schoten female data, for which four cases of osteoporosis could be diagnosed, are within the range of these reference values. Each bone site is specific, because the conditions of biomechanical functioning of the bone govern remodelling. This singularity, due to their structure (spongious bone or cortical bone) or to their location in the skeleton as a whole, makes it impossible to make generalisations on overall mineral content based on local data (Laval-Jeantet et al., 1995). Correlations between the BMD of the different femoral regions (table 3) however suggest a more homogeneous bone structure in females than in males (see above).
22Rural life in the 19th and early 20th centuries was certainly characterised by the absence of mechanised means of transport as are generally available today, and Schoten’s inhabitants therefore remained active throughout their lifetimes. Considering cross-sectional cortical thickness in femora, the rate of aging appears to have been much slower for this late 19th to early 20th century sample than it is for modern samples (Molleson and Orban, 2019). The BMD data do not offer such a clear picture. Male individuals show a wide variation in age-related bone density loss, within the range of the reference sample. Whether this feature is linked to a wide range of professional activities is another research question. As far as the small size of our sample allows us to draw any conclusions, females show above average bone density values until the age of 50, they then show average bone densities and age-related bone density loss at all measurement sites.
23Femoral BMD measurements for a French (individuals of European ancestry) population are provided by Laval-Jeantet et al. (1995) and constitute a modern population reference sample comparable to the Schoten series. Data provided by Sansilbano-Collilieux et al. (1994) for the left/right femoral neck in a young (18-20 year old) and in an aged woman (over 50 years old) (necropolis of Cognac-Saint-Martin, 7th-15th century AD) are comparable to the Schoten data. Details pertaining to this material are provided by Gilissen and Orban (2022). Most interestingly, Curate et al. (2013) show on a Portuguese skeletal sample from the mid-19th - early 20th centuries that BMD was significantly higher in males when compared to females, as in the Schoten sample, and that the prevalence of age-related osteoporosis in the proximal femur was higher in women. Comparisons with modern Portuguese samples showed an equivalent pattern of BMD reduction. Curate et al. (2013) however observed that BMD was usually lower in the skeletal sample, which is not the case for Schoten. Overall, the decrease in bone mass observed in skeletal samples and its similarity to that observed in modern populations, despite huge lifestyle differences, suggests a profound ancestry of age-related bone loss and related diseases. In another geographical context, it is here interesting to note the much better bone maintenance among Neolithic females when compared to early modern times females from the same region in present-day Poland (Spinek et al., 2016).
24A striking contrast between female and male individuals is observable in the BMD values provided by Fulpin et al. (2001) and Mafart et al. (2008) for the femoral neck and Ward’s triangle. These data come from the skeletal remains buried in the cemetery of the crypt of Notre-Dame-du-Bourg’s cathedral in Digne, Alpes-de-Hautes-Provence, France. It was the burial place for the city’s inhabitants from the 11th to the 17th century AD without any known social discrimination. As can be seen in table 5, most of the elderly women (over 50 years old) had BMD values lower than the younger women (under 30 years old) and the difference between younger and older female individuals seems bigger than it is today. In comparison to both the NHANES reference database and the Schoten series, the female values appear to be in the lower range.
Table 5
Bone mineral densities (BMD in g/cm2) for Notre-Dame-du-Bourg (Mafart et al., 2008) transformed to Hologic values, using the following equations. For methods used to generate these conversion algorithms, see Mays et al. (2006). For the femoral neck: Hologic BMD=0.789126×Lunar BMD+0.081403. For Ward’s triangle: Hologic BMD 0.934769×Lunar BMD-0.140822 |
Les données de densité minérale osseuse (BMD en g/cm2) de Notre-Dame-du-Bourg d’après Mafart et al. (2008) ont été transformées en valeurs Hologic à l’aide des équations suivantes. Pour les méthodes utilisées pour générer ces algorithmes de conversion, voir Mays et al. (2006). Pour le col du fémur : BMD Hologic=0,789126×BMD Lunar+0,081403. Pour le triangle de Ward : BMD Hologic=0,934769×BMD Lunar-0,140822
25Mays et al. (2006) compared the Trondheim’s (12th-17th centuries AD) and Wharram Percy’s (11th-16th centuries AD, see above) skeletal collections’ BMD values (table 6, adapted from Mays et al., 2006). The Trondheim population resembled the Wharram Percy sample, in terms of both age-dependent loss of BMD for both sexes and peak BMD values. There is no evidence for differences between the two populations in peak bone density and no evidence for differences in patterns of age-related decline in BMD. There is however a slight difference if we consider the earlier loss of BMD at Ward’s triangle in the Trondheim males in the 30-49 years group, which is significantly lower than in the Wharram Percy males (Mays et al., 2006). This differs from the Schoten sample where a drop in BMD values can be observed at Ward’s triangle between 40-50 years in females but not in males. This is of interest because the Schoten sample is more recent than the Trondheim and Wharram Percy samples, and the current Norwegian and UK populations show differences in BMD, probably recent in origin (Mays et al., 2006). On this particular point, the overall results of Brødholt et al. (2021) highlight the significant variation in BMD during the prehistoric and historic periods in Norway.
Table 6
Trondheim bone mineral density values (BMD in g/cm2) from Mays et al. (2006), Wharram Percy values from Mays et al. (1998). Original values were transformed to Hologic values (see table 5 for the equations) |
Valeurs de densité minérale osseuse (BMD en g/cm2). Les valeurs des données de Trondheim proviennent de Mays et al. (2006). Les valeurs des données de Wharram Percy proviennent de Mays et al. (1998). Ces valeurs ont été transformées en valeurs Hologic (voir le tableau 5 pour les équations)
26This illustrates the importance of variation in BMD between human populations and its changes over time. As Molleson and Orban (2019) concluded for cross-sectional femoral cortical thickness, with regard to BMD, age-related trends in one sample will not necessarily be applicable to another, whether the two samples are contemporaneous or not.
27A more recent population, almost contemporary with Schoten, is the Christ Church Spitalfields collection. The remains were excavated during the restoration of the church and were dated from 1729-1852. They consist of 87 left femora from female individuals aged between 15 and 89 years, and 30 males between 25 and 88 years old. Lees et al. (1993) provided values for mean BMD at the femoral neck in 55 year old females (0.729) and males (0.883), and at Ward’s triangle (0.506 and 0.671, respectively; BMD values in g/cm2 transformed to Hologic values, see table 5 for the equations). They concluded that the rate of femoral bone loss is significantly higher in females today than two centuries ago, both before and after menopause. Notwithstanding, neither the NHANES nor the Schoten sample confirm this conclusion. The Christ Church Spitalfields data are within the range or close to the mean values of the current reference database (figures 3, 5 and 7). This observation is consistent with the aforementioned results of Curate et al. (2013). As indicated by Ruff et al. (2015), increasing mechanisation and urbanisation have relatively small effects on skeletal robustness.
28Bone loss is inevitable with age, as it appears to be the case for humans. It is also the case for other primates. Nonhuman primates are particularly useful for studying changes in the musculoskeletal system because they develop bone and muscle loss similar to humans during aging. Old World monkeys and great apes display haversian osteonal remodelling of cortical bone that occurs in humans but not in rodents, and they have a similar reproductive endocrine system that affects bone metabolism (Didier et al., 2016). In particular, chimpanzees are among the closest living relatives to humans and hence provide a crucial comparative model for investigating various aspects of human biology and evolution.
29We here present results of the recent study of Tayach (2022) on BMD analysis in a chimpanzee (Pan troglodytes) femoral bone series (adult males N=19, adult females N=12) housed in the Royal Museum for Central Africa. This series is housed in conditions similar to the Schoten series at the Royal Belgian Institute of Natural Sciences. Bone mineral density measurements were performed on the proximal region of the left femur of each specimen with a standard densitometer Lunar iDEXA enCORE Version 17 (GE Medical Systems, Chicago IL, USA). Details of the analysis and of the transformation to Hologic values are provided by Tayach (2022). Here, we present the results for the chimpanzee femoral neck region and for Ward’s triangle, and compare them to those from the Schoten series.
30Figures 9 and 10 illustrate BMD (g/cm2) at the femoral neck for humans (Homo sapiens) and chimpanzees (Pan troglodytes) between the quartiles of 25% to 75%. The horizontal line indicates the median, the cross represents the mean. Whiskers show the total range for females (F, figure 9) and males (M, figure 10). As in humans, chimpanzee BMD values are higher in males than in females. All means are significantly different (p<0.01) and chimpanzees have higher BMD values than humans (Tayach, 2022).
Figure 9
Bone mineral density (BMD in g/cm2) distribution at the femoral neck in female chimpanzees (Pan troglodytes) and in female human individuals from the Schoten series (Homo sapiens). X-ray: L. Gantois and M. Lemort (Jules Bordet Institute, Brussels). F: female |
Distribution de la densité minérale osseuse (BMD en g/cm2) au niveau du col du fémur dans une série de chimpanzés femelles (Pan troglodytes) et chez les sujets féminins de la série de Schoten (Homo sapiens). Radiographie : L. Gantois et M. Lemort (Institut Jules Bordet, Bruxelles). F : femelle
Figure 10
Bone mineral density (BMD in g/cm2) distribution at the femoral neck in male chimpanzees (Pan troglodytes) and in male human individuals from the Schoten series (Homo sapiens). X-ray: L. Gantois and M. Lemort (Jules Bordet Institute, Brussels). M: male |
Distribution de la densité minérale osseuse (BMD en g/cm2) au niveau du col du fémur dans une série de chimpanzés mâles (Pan troglodytes) et chez les sujets masculins de la série de Schoten (Homo sapiens). Radiographie : L. Gantois et M. Lemort (Institut Jules Bordet, Bruxelles). M : mâle
31For the femoral neck BMD, the interquartile range of the Schoten females is 0.49, compared with 0.23 for female chimpanzees. The interquartile range of the Schoten males is 0.23, compared with 0.31 for male chimpanzees. For BMD at Ward’s triangle, the interquartile range in Schoten females is 0.42, compared with 0.19 in female chimpanzees. The interquartile range in Schoten males is 0.22, compared with 0.20 for male chimpanzees (Tayach, 2022).
32Gunji et al. (2003) provided femoral neck BMD values for six young adult female chimpanzee skeletons measured by DEXA. These are two specimens from Mahale Mountains National Park, Tanzania, whose ages at death are estimated as 23 and 20 years old respectively (Hosaka et al., 2000), and four captive ones (18 years old and prime adults). These values (1.10-1.30 g/cm2) are in the upper range of the values provided by Tayach (2022) for the specimens stored in the Royal Museum for Central Africa.
33The modern human skeleton therefore appears to be more gracile than the one of a great ape. The ecological and phylogenetic components need to be explored. In their comparative study of the bone mechanical properties of the radius in chimpanzees, humans and Japanese macaques (Macaca fuscata) using peripheral quantitative computed tomography, Kikuchi et al. (2003) concluded that chimpanzees are more similar to macaques than to humans in terms of bone properties and that the metabolism of bone, including cortical bone density, might be conserved in these non-human primates. The high BMD values observed by Tayach (2022) in chimpanzees compared with humans could therefore provisionally be considered as a plesiomorphic (archaic) character. These observations are consistent with the fact that the human skeleton is unique in having a low fraction of trabecular bone volume compared to non-human primates. Trabecular volume fraction remained high throughout human evolution until it decreased significantly in recent modern humans (Chirchir et al., 2015). Chirchir et al. (2015) and Chirchir (2019) suggest a possible link between changes in the modern human skeleton and increased sedentism. In this framework, Ryan and Shaw (2015) observed a correspondence between human behaviour and trabecular bone structure in the proximal femur, indicating that more highly mobile human populations have trabecular bone structure similar to what would be expected for wild nonhuman primates of the same body mass. This appears to be somewhat at odds with the study by Madimenos et al. (2020). These authors recorded bone density values and observed a higher risk of osteoporosis among Bolivian Tsimane forager-horticulturalists, a rural population with high fertility and high levels of lifelong physical activity, than for populations in the United States. Moreover, Tsegai et al. (2018) found that trabecular bone volume fraction is, in most anatomical sites studied, significantly higher in chimpanzees than in humans, suggesting a systemic difference in trabecular structure unrelated to local loading regime. This systemic difference hypothesis is not ruled out by Chirchir (2019). Beyond the study of femoral BMD, these results highlight the importance of comparing multiple skeletal sites across a diverse set of primate species, as well as archaeological and recent samples from modern human populations, for a precise understanding of the dynamics of primate and human bone structure.
34In conclusion, BMD values in the homogeneous Schoten sample are within the range of the modern population reference sample (NHANES) but females show above-average bone density values up to the age of 50 years, followed by average bone densities until death, and this at all measurement sites. When considering rural female populations, high lifetime physical activity levels that increase mechanical loading, higher parity, short inter-birth interval and earlier age at first pregnancy are traits associated with reduced BMD and greater age-related decline in BMD when compared to reference populations (Stieglitz et al., 2015). However, it should be underlined that apparent similarities in lifestyle and environment may not necessarily translate into similar bone densities (Madimenos et al., 2020). We can at least here conclude that the trade-off between investing energy in reproduction and somatic repair among the Schoten females is comparable to that of females in contemporary Western societies. In addition, diagnosis of osteopenia and osteoporosis shows striking contrasts between males and females, and variability in BMD is systematically greater in males than in females for all femoral regions and for all age groups, a situation similar to the one observed in the NHANES reference sample. In a broader comparative context, human BMD values at the femoral neck are lower than for chimpanzees, but the variability appears to be more important in humans than among chimpanzees.
Acknowledgements: Access to the Schoten collection at the Royal Belgian Institute of Natural Sciences (RBINS) was supported by a Return Grant from the Belgian Federal Science Policy (Belspo) to EG. Sincere thanks to Prof. Marc Lemort and to Lionel Gantois at the Institut Bordet, Brussels, for so generously performing the DEXA analysis on both the human and chimpanzee series and for providing the NHANES reference dataset. Special thanks to Mathys Rotonda (Royal Museum for Central Africa) for his help with the figures, to Theya Molleson (Natural History Museum, London) for effective support in the earlier phases of this work and to Antoine Balzeau (Muséum National d’Histoire Naturelle, Paris), Caroline Polet (Royal Belgian Institute of Natural Sciences, Brussels), José Braga (Université Paul Sabatier, Toulouse), Yann Heuzé (CNRS, Université de Bordeaux), Stéphane Louryan (ULB) and two reviewers for their insightful comments and for providing key references.