AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 124:33 44 (2004)

Medieval Trabecular Bone Architecture: The Inuence of Age, Sex, and Lifestyle
S.C. Agarwal,1,2* M. Dumitriu,2 G.A. Tomlinson,3 and M.D. Grynpas2,4
1 2

Department of Anthropology, McMaster University, Hamilton, Ontario L8S 4L9, Canada Samuel Lunenfeld Research Institute, Mount Sinai Hospital, Toronto, Ontario M5G 1X5 3 Departments of Medicine, Toronto General Hospital and University of Toronto, Toronto, Ontario M5G 2C4, Canada 4 Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto, Ontario M5G 1L5, Canada KEY WORDS populations trabecular architecture; bone loss; bone quality; osteoporosis; past

ABSTRACT Osteoporosis has become a growing health concern in developed countries and an extensive area of research in skeletal biology. Despite numerous paleopathological studies of bone mass, few studies have measured bone quality in past populations. In order to examine age- and sex-related changes in one aspect of bone quality in the past, a study was made of trabecular bone architecture in a British medieval skeletal sample. X-ray images of 5-mm-thick coronal lumbar vertebral bone sections were taken from a total of 54 adult individuals divided into three age categories (18 29, 30 49, and 50 years), and examined using image analysis to evaluate parameters related to trabecular bone structure and connectivity. Signicant age-related changes in trabecular bone structure (trabecular bone volume (BV/TV), trabecular number (Tb.N), trabecular separation (Tb.Sp), and anisotropic ratio (Tb.An)) were observed to occur primarily by middle age with signicant differences between the youngest and two older age groups. Neither sex showed continuing change in trabecular structure between the middle and old age groups. Age-related changes in bone connectivity (num-

ber of nodes (N.Nd) and node-to-node strut length (Nd.Nd)) similarly indicated a change in bone connectivity only between the youngest and two older age groups. However, females showed no statistical differences among the age groups in bone connectivity. These patterns of trabecular bone loss and fragility contrast with those generally found in modern populations that typically report continuing loss of bone structure and connectivity between middle and old age, and suggest greater loss in females. The patterns of bone loss in the archaeological samples must be interpreted cautiously. We speculate that while nutritional factors may have initiated some bone loss in both sexes, physical activity could have conserved bone architecture in old age in both sexes, and reproductive factors such as high parity and extended periods of lactation could have played a key role in female bone maintenance in this historic population. The study of qualitative elements (such as trabecular architecture) is vital if we are to understand bone maintenance and fragility in the past. Am J Phys Anthropol 124:33 44, 2004. 2004 Wiley-Liss, Inc.

Osteoporosis has become a growing health concern in the aging populations of developed countries. Osteoporosis is a systemic skeletal disease characterized by a reduction in the amount of bone or bone mass, and a deterioration of the microstructure of bone tissue, caused by a signicant period in which bone resorption exceeds bone formation (Kanis, 1994; Manolagas and Jilka, 1995; Riggs and Melton, 1988). Osteopenia is the correct term for reduced bone, specically loss of mineral density or mineral in general (Kanis, 1990). It is perhaps appropriate to consider osteopenia as the risk factor for the clinical disorder osteoporosis (Melton and Wahner, 1989). The term osteoporosis is then reserved specically for bone fragility and increased susceptibility to fracture (Kanis, 1994; Riggs and Melton, 1988). However, more than loss of bone mass contributes to fragility. Extraskeletal factors such as falls and elements of bone quality, which refer to the structural and material properties of bone itself and its archi

tecture, contribute signicantly to bone fragility (Heaney, 1993; Kanis, 1990). Primary osteoporosis includes both postmenopausal (or type I) and agerelated (or type II) osteoporosis. While age-related osteoporosis is found in both sexes after about age 70 years, postmenopausal bone loss is accelerated in
Grant sponsor: Natural Sciences and Engineering Research Council of Canada; Grant sponsor: Social Sciences and Humanities Research Council of Canada. *Correspondence to: Dr. Sabrina C. Agarwal, Department of Anthropology, University of Toronto, 100 St. George Street, Toronto, Ontario M5G 3G3, Canada. E-mail: sabrina.agarwal@utoronto.ca Received 25 July 2000; accepted 29 April 2003. DOI 10.1002/ajpa.10335 Published online 11 August 2003 in Wiley InterScience (www. interscience.wiley.com).

2004 WILEY-LISS, INC.

34

S.C. AGARWAL ET AL.

females with the onset of menopause, and is found in females between 5575 years. Both show differing patterns of fracture and affected skeletal sites. However, osteoporosis is a heterogeneous disorder, associated with a number of risk factors including genetics, diet, nutrition, and mechanical usage. A number of physiologic and lifestyle factors, including physical activity, parity, and lactation, have also been suggested to affect both peak bone mass, and subsequent bone loss (e.g., Sowers and Galuska, 1993; Stevenson et al., 1989; Ward et al., 1995). In an attempt to better understand the patterns and prevalence of the disease, there has been steady interest in the paleopathology of osteoporosis. A number of paleopathological studies examined bone loss in past populations from diverse geographic regions with a variety of methodologies (Agarwal and Grynpas, 1996; Pfeiffer and Lazenby, 1994). However, the majority of studies concentrated on bone mass in past populations, with few addressing the important role of bone quality in bone fragility. A variety of biomedical investigations clearly established the role of bone quality in bone strength. For example, the importance of age-related changes in bone structure and material (Grynpas, 1993; Martin, 1993), fatigue and microdamage (Partt, 1993; Schafer et al., 1989), and trabecular architecture (Ding et al., 2002; Fazzalari, 1993; Mosekilde, 1993; Recker, 1993; Snyder et al., 1993; Thomsen et al., 2002) in bone fragility were all well-demonstrated. Bone quality would likely have played a similarly important role in bone fragility and fracture risk in past populations. Qualitative elements of bone strength such as material properties or microdamage cannot be examined in archaeological bone due to decomposition and burial. However, one aspect of bone quality that can be examined in archaeological bone is trabecular architecture. In order to examine age- and sex-related changes in bone quality in the past, this study documents and interprets vertebral trabecular architecture in a British medieval skeletal population. Overall patterns of age- and sexrelated changes in trabecular architecture were compared with those reported in modern populations, and factors that may have inuenced trabecular bone maintenance in the population are considered. MATERIALS Intact fourth lumbar vertebral samples were taken from 54 adult individuals (males 24, females 30) from a British archaeological population. Skeletons were excavated at the deserted medieval village of Wharram Percy, located in North Yorkshire, England. The burials are thought to primarily represent ordinary peasants who lived at the predominantly agricultural settlement of Wharram Percy or elsewhere in the parish, and are dated mainly to the medieval period between 1116th centuries AD (Beresford and Hurst, 1990; Mays, 1996; Mays et al., 1998; Sofaer Derevenski, 2000). Mays

(1996) and Mays et al. (1998) at English Heritage conducted initial sex determination and age-atdeath estimation of the skeletons. Sex was determined using standard osteological indicators (Brothwell, 1981), and age was determined using dental wear. Examination of all skeletons along with the assigned sex and age was made by the senior author to conrm age and sex assignments. Dental wear was specically recorded on the molar teeth using the system of Brothwell (1981), incorporating suggested modications (Mays, 1996; Mays et al., 1998). Wharram Percy is considered an ideal population in which to age with dental wear, as the sample has a signicant number of juveniles required to calibrate the rate of wear (Mays, 1998a; Mays et al., 1998). However, as aging of skeletal material is particularly problematic in the case of older individuals, individuals were conservatively assigned into three broad age categories: 18 29, 30 49, and 50 years. Previous work on Wharram Percy demonstrated age-related loss of bone quantity (bone mineral density and mass); however, the skeletons do not show a signicant number of typical fragility related fractures (Mays, 1996; Mays et al., 1998). The vertebra is ideal for the study of trabecular architecture, as it consists of a rich trabecular network constructed to withstand vertical compressive forces, its microstructure made up of vertical columns of bone connected by horizontal struts (Mosekilde, 1993; Partt, 1992). Age-related change and loss of both trabecular structure and connectivity are known (Mosekilde, 1993; Partt, 1992), and a number of studies conrmed that architecture does play a role, independent of mass, in bone strength (Compston, 1994; Garrahan et al., 1986; Kleerekoper et al., 1985; Mellish et al., 1991; Mosekilde, 1988, 1989, 1993; Partt, 1992; Vesterby et al., 1991). METHODS Image analysis Midcoronal vertebral sections (5 mm thick) were removed, using an Isomet low-speed saw (Buehler, Ltd.) (Fig. 1). The vertebral body was clamped with a custom-designed metal clamp and screw to keep the bone secure during sectioning. The Isomet micrometer was used to ensure removal of a measured parallel 5-mm-thick section. Sections were removed easily, with only very minute bone loss observed during cutting. Sections were cleaned and sonicated in 70% ethanol and distilled water to remove any loose dirt particles, and air-dried. The few sections that upon sectioning were found to contain excessive clay dirt that could not be easily dislodged, and that would obstruct measurement of trabecular architecture, were not included in the study. In order to obtain information on the qualitative elements of trabecular architecture, sections were examined with the use of image analysis conducted with X-rays of thick sections. X-rays of each section

MEDIEVAL TRABECULAR BONE ARCHITECTURE

35

Trabecular bone volume (BV/TV) trabecular bone area 100/total area measured . Average trabecular thickness (Tb.Th) bone area/12 perimeter. Trabecular number (Tb.N) Trabecular separation (Tb.Sp) Tb.Th. (BV/TV 1) . Anisotropic ratio (Tb.An)
Fig. 1. Five-millimeter-thick coronal section of fourth lumbar vertebra from Wharram Percy, showing rich anisotropic trabecular network.

(BV/TV) . Tb.Th

ratio of vertical to horizontal trabeculae (Grynpas et al., 1992) . The amount of bone (BV/TV) in a given section of bone is closely related to its organization, as described by the various structural and connectivity parameters. Several authors note that BV/TV is highly correlated to both strut-analysis parameters and histomorphological parameters such as trabecular thickness, separation, or number (Chappard et al., 1999; Compston et al., 1989; Croucher et al., 1996). However, as noted and demonstrated by Banse (2002), identical BV/TV in two given samples can still show entirely different structural organization, and as such, the study of qualitative parameters aside from BV/TV is required and justied. Data analysis All statistical analyses were conducted using SPSS statistical software (version 10.7). Descriptive statistics, including sample mean, range, standard deviation (SD), and standard error of the mean (SEM) were computed for each parameter. Normality of the data was examined with the Shapiro-Wilk test. The Levene test for homogeneity of variance was used to compare variability within each of the age groups. Within-group variances were sufciently close to permit use of standard ANOVA models. A comparison of means with a one-way analysis of variance (one-way ANOVA) was used to determine signicant differences in the image-analysis parameters across the three age groups. Tukeys test (a post hoc test that accounts for multiple comparisons) was used to determine which pairs of groups differed when the overall ANOVA P-value for a parameter was less than 0.05. These ANOVAs were performed rst on the sample as a whole, and then with each sex considered separately. A two-way ANOVA was used to determine if age-related changes in the image-analysis parameters were statistically different for each sex, allowing the consideration of age, sex, and their interaction as separate factors. A third set of analyses added BV/TV as a continuous covariate to the two-way ANOVA for each of the other image-analysis parameters. If differences across age groups in the mean value of a parameter

were taken on a Faxitron (Hewlett-Packard) machine at between 2527 kV for 10 sec, using Fuji medical mammography lm. X-ray lms were then scanned to obtain image les that could be accessed by a Quantimet 570 (Leica) image processing and analysis system. Mathematical morphology (Serra, 1982) and standard image analysis tools were used to analyze the images obtained. A masked area that included only the trabecular tissue of the entire section, excluding the surrounding cortical bone, was used. After thresholding the grey-level images and obtaining binary images, a pruned version of the binary trabeculae was created, which is termed a skeletonized image (Fig. 2), in order to obtain data on the orientation and connectivity of the individual trabecular branches (Grynpas et al., 1992; Lundon et al., 1997). Using multiple-point transformation, the connection pixels between the vertical and horizontal trabeculae are found (Lundon et al., 1997). The endpoint transformation is similarly used to mark the free ends of trabeculae on the skeletonized binary image (Lundon et al., 1997). These multiple points (termed nodes) and end points (termed terminus) on the image are then used to obtain the following strut analysis parameters, described by Mellish et al. (1991) (Fig. 3): number of nodes/mm2 (N.Nd), number of termini/mm2 (N.Tm), and terminus-to-terminus (Tm.Tm), node-to-terminus (Nd.Tm), and node-to-node (Nd.Nd) strut lengths (mm/mm2). These parameters reect the number of structural elements and the connectivity of the bone. For example, a greater number of nodes or greater node-to-node strut length would indicate a higher degree of connectedness, whereas a greater number of termini or terminus-to-terminus free struts would indicate a loss of connectivity (Mellish et al., 1991). Mathematical morphology was used to also obtain the following structural parameters. All calculations and nomenclature used are followed from the standards set forth by the ASBMR Histomorphometric Nomenclature Committee (Partt et al., 1987):

36

S.C. AGARWAL ET AL.

Fig. 2. Binary image (left) and skeletonized image (right), delineating trabecular branching of vertebral trabecular bone from Wharram Percy sample.

RESULTS Signicant results are presented in Tables 1 and 2 and expressed as mean SEM. The trabecular bone parameters are presented in two groups: those that reect trabecular structure, and those that are related to trabecular bone connectivity. Trabecular structure Age-related patterns are seen in the structural parameters of males. Both mean BV/TV and Tb.N show an age-related decrease. BV/TV shows signicant difference among the age groups with one-way ANOVA at P 0.008, while Tb.N shows signicant difference at P 0.011, with Tukeys post hoc tests showing signicant difference in both parameters again in groups 1 vs. 2 and 1 vs. 3 at P 0.05. Both Tb.Sp and Tb.An show an age-related increase with one-way ANOVA at P 0.014 and P 0.040, respectively, with Tukeys post hoc tests showing a signicant difference in both parameters in age groups 1 vs. 3 at P 0.05. Thus, males show significant loss of structure by middle age, with no statistically signicant change between middle and old age. Females also show age-related changes. Both mean BV/TV and Tb.N show differences with oneway ANOVA at P 0.028, and P 0.005, respectively, with Tukeys post hoc tests showing signicant difference in groups 1 vs. 2 in the case of BV/TV and groups 1 vs. 2 and 1 vs. 3 in the case of Tb.N. Both Tb.An and Tb.Sp show an increase by middle age, one-way ANOVA showing a signicant difference between the groups at P 0.016 for Tb.Sp and P 0.029 for Tb.An, with Tukeys post hoc tests showing differences to lie between age groups 1 and 2 at P 0.05. Thus, females seem to lose structure primarily in middle age, with again no loss of struc-

Fig. 3. Diagrammatic representation of skeletonized trabecula with node and termini with three connecting struts. Adapted from Compston et al. (1987).

were present in the two-way ANOVA, but were diminished in the ANOVA with BV/TV, then we inferred that these differences were a result of the parameters association with BV/TV. If the differences remained in the model with BV/TV, this was taken as evidence of a relationship beyond that stemming from the parameters association with BV/TV.

MEDIEVAL TRABECULAR BONE ARCHITECTURE

TABLE 1. Wharram Percy trabecular bone structure results1 N F1 F2 F3 P 0.05 M1 M2 M3 P 0.05
1

37

BV/TV (%) 37.4 2.4 27.4 2.9 29.8 2.0 1 vs. 2 38.3 3.1 28.9 2.0 26.4 2.4 1 vs. 2, 3

Tb.Th ( m) 189.8 170.7 184.7 ns 198.9 181.5 179.0 ns 6.8 5.7 11.4 5.5 9.9 8.3

Tb.N ( /mm2) 2.0 1.6 1.6 1 vs. 1.9 1.6 1.5 1 vs. 0.08 0.09 0.07 2, 3 1.2 0.08 0.08 2, 3

Tb.Sp ( m) 327.3 490.8 447.6 1 vs. 336.6 464.0 514.9 1 vs. 25.6 44.9 34.3 2 34.9 36.9 47.3 3

Tb.An 1.1 0.08 1.4 0.06 1.3 0.06 1 vs. 2 1.1 0.09 1.2 0.04 1.3 0.06 1 vs. 3

9 12 9 8 9 7

Trabecular structural (histomorphometric) parameters. 1, age group 18 29 years; 2, age group 30 49 years; 3, age group 50 years. F, females; M, males. BV/TV, trabecular bone volume, Tb.Th, average trabecular thickness, Tb.N, trabecular number, Tb.Sp, trabecular separation, Tb.An, anisotropic ratio. Statistical signicant differences between age groups are indicated at P 0.05 level with one-way ANOVA, whereby pairwise multiple comparison-procedures indicate which group was signicantly different. Results are expressed as mean SEM. For statistical procedures, see text. TABLE 2. Wharram Percy trabecular bone connectivity results1 N N.Nd 1.6 1.3 1.3 0.15 0.16 0.09 N.Tm 0.08 0.32 0.21 ns 2.2 0.18 2.7 0.24 2.7 0.33 ns 2.5 3.1 2.5 Tm.Tm 0.30 0.32 0.31 0.05 0.04 0.04 Nd.Nd 0.92 0.60 0.63 0.12 0.11 0.06 Nd.Tm 0.72 0.66 0.66 ns 0.67 0.67 0.60 ns 0.03 0.04 0.04 0.03 0.04 0.04

F1 F2 F3 P 0.05 M1 M2 M3 P 0.05
1

9 12 9 8 9 7

ns 1.8 0.30 1.2 0.14 0.96 0.10 1 vs. 3

ns 0.24 0.04 0.36 0.03 0.39 0.06 p 0.051

ns 1.0 0.17 0.54 0.08 0.45 0.08 1 vs. 2, 3

Trabecular connectivity parameters. 1, age group 18 29 years; 2, age group 30 49 years; 3, age group 50 years. F, females; M, males. N.Nd, number of nodes/mm2; N.Tm, number of termini/mm2; Tm.Tm, terminus-to-terminus; Nd.Nd, node-to-node; Nd.Tm, node-to-terminus strut lengths (mm/mm2). Statistically signicant differences between age groups are indicated at P 0.05 level with one-way ANOVA, whereby pairwise multiple comparison procedures indicate which group was signicantly different. Results are expressed as mean SEM. For statistical procedures, see text.

ture at all between middle and old age. Trabecular thickness (Tb.Th) shows no statistically signicant differences among age groups in either sex. Age effects are seen for TB.N and TB.Sp in a two-way ANOVA (P 0.0001 and P 0.0004), but they are no longer signicant when BV/TV is added as a covariate (P 0.11 and P 0.74, respectively). Tb.An shows signicant age effects in the two-way ANOVA with or without BV/TV (P 0.002 for both models). Trabecular connectivity Clear age-related patterns in connectivity are evident in males. A signicant difference in the age groups is seen with one-way ANOVA in N.Nd at P 0.033, with post hoc tests showing a difference between age groups 1 and 3 at P 0.05. Nd.Nd also shows an age-related decrease, the data requiring the appropriate transformation to stabilize variance. One-way ANOVA of the square root of mean Nd.Nd shows a signicant difference between groups at P 0.006, with signicant difference between age groups 1 vs. 2 and 1 vs. 3 with Tukeys tests at P 0.05. Terminus-to-terminus strut length (Tm.Tm) shows a statistical trend with one-way ANOVA for differences among age groups at P 0.050. Thus, males show signicant loss of connectivity by middle age. Interestingly, no statistically signicant differences are seen between the female age groups for any of the connectivity parameters.

None of the differences across age groups are significant in the two-way ANOVA with BV/TV as a covariate. No statistically signicant difference between the sexes is seen in any of the image-analysis parameters when age and sex are considered as separate factors in a two-way ANOVA. However, the small sample size of this data set means that the power of this analysis is low. DISCUSSION Assessment of methodology Only a small number of studies have examined trabecular structure in past populations (Brickley and Howell, 1999; Kneissel et al., 1994, 1997; Vogel et al., 1990). Examination of age- and sex-related changes in trabecular microstructure avoids the difculties associated with chemical diagenesis or the use of in vivo instruments to assess bone loss. Most importantly, it allows the examination of one aspect of bone quality. Aspects of bone quality, such as trabecular architecture, are known to play a vital role in bone fragility and fracture risk. It was demonstrated in this study that trabecular architecture in archaeological bone can be assessed with the use of image analysis. In the biomedical setting, image analysis has been used extensively by investigators to examine trabecular bone microstructure from samples that are processed as thin

38

S.C. AGARWAL ET AL.

sections (Amling et al., 1996; Compston et al., 1987; Croucher et al., 1996; Garrahan et al., 1986; Lundon et al., 1997; Thomsen et al., 2000; Vesterby, 1990), photographs of thick sections (Bergot et al., 1988; Twomey et al., 1983), computed tomography (CT) images (Gordon et al., 1998), and X-rays images (Geraets et al., 1990, 1993; Grynpas et al., 1992, 1994). Although the different methods of obtaining images are not directly comparable, the different images provide similar information on trabecular structure, and each offers different advantages. The choice of utilizing X-rays of thick sections to obtain images in this study was made for various reasons. The 5-mm large sections could not be easily processed to obtain thin sections, and three-dimensional (3-D) imaging instruments such as CT were not available. A preliminary study of the use of photographic images of the sections from the Wharram Percy sample was undertaken. Although the parameters measured from the direct photographs did correlate statistically with the measurements taken on X-ray images and demonstrated the same overall statistical sex- and age-related trends, there were several difculties with using photographic images. Photographs produced images of only the surface of the bone section, and it was often difcult to distinguish trabeculae due to shadowing. Much relied on subjective observer interpretation and the use of correct lighting of the section during photograph capture. Korstjens et al. (1998) discussed similar difculties with photographic images as compared to radiographic images. Digital image analysis of radiographic trabecular bone architecture was used in previous studies (Geraets et al., 1990; Grynpas et al., 1992, 1994), and the reliability of the technique is well-established (Korstjens et al., 1997). Radiographic images also capture trabecular structure throughout the entire section, and it was demonstrated that radiographic properties of trabecular bone architecture signicantly correlate to biomechanical characteristics (Korstjens et al., 1998). As such, radiographic images of archeological bone sections were used in this study to conduct image analysis. There are also concerns with microstructural analysis using two-dimensional (2-D) images. 2-D images are not an exact representation of the 3-D structure of the trabecular bone, but certainly are highly related to 3-D structure. In the case of strut analysis, nodes and node-to-node struts may be underestimated from 2-D images, but cannot be created artifactually (Compston, 1994; Compston et al., 1987). In contrast, termini and struts with one or more termini may be created by sectioning processes, and could be overestimated (Compston, 1994; Compston et al., 1987). All of the image analysis parameters in this study showed discernible patterns in trabecular structure and connectivity, and the measurements of bone structure and connectivity correlated well with another study that was made of bone mineral density (Agarwal, 2001). Only

trabecular thickness (Tb.Th) did not show a statistically signicant age-related trend. It is possible that radiographic images of sections were not ideal to accurately measure and differentiate minor ageor sex-related differences in trabecular thickness, and thus the observation of Tb.Th may be better suited for histological thin sections. Interpreting age-related patterns of change in trabecular architecture Comparison with previous studies. Age-related change in trabecular structure and connectivity in the Wharram Percy sample is seen to occur by middle age. Males appear to follow some gradual agerelated change in trabecular architecture, but in both sexes there appears to be no signicant change in bone structure or connectivity between middle and old age. Females show no signicant change connectivity at all. Given that present-day females are traditionally thought to show greater postmenopausal and age-related bone loss than males, the lack of sex differences in trabecular bone architecture, and in the age-related patterns of change, are also surprising. The small sample numbers in this study make interpretation of the results problematic. However, the few previous studies of trabecular architecture reported similar ndings to the present study. Vogel et al. (1990), in a study of European historical samples, found less pronounced age-related loss of bone volume and trabecular connectivity as compared to modern samples. A more recent study by Kneissel et al. (1997) of Nubian vertebrae reported signicant loss of trabecular structure and connectivity at an early age. However, they found no signicant loss in the later age groups encompassing individuals 30 60 years of age (Kneissel et al., 1997). A study by Brickley and Howell (1999) of mean trabecular length in a British postmedieval sample also noted age-related loss of structure to occur primarily in young age females, and loss to be less marked in the oldest female age group. However, signicant loss of structure was seen in the older male age group in this postmedieval sample, and sex differences were noted, with males showing greater intact structure in almost all age groups as compared to females (Brickley and Howell, 1999). In a study of cortical bone at Wharram Percy, Mays (1996) reported on age-related bone loss in the metacarpal, although, as with this study, he did not nd signicant loss between the middle and oldest age groups. However, subsequent study of femoral bone mineral density (BMD) by Mays (1998b) did nd a continuing decrease in BMD in females between middle and old age. The reason for the different patterns of age-related bone loss at Wharram Percy is unclear. Some of the discrepancies in previous studies may reect differences in tissue response (trabecular vs. cortical) and/or differing skeletal sites of bone loss. The femoral sites examined (neck and Wards triangle) by Mays (1998b) contained a mix of both cortical and trabecular tissue, as com-

MEDIEVAL TRABECULAR BONE ARCHITECTURE

39

pared to the present study that examined strictly trabecular tissue. Also, as with all absorptiometric measurements in archeological bone, it is unclear if chemical diagenesis may be affecting the bone mineral density results. Another important factor to consider is the parameter that the differing methodologies are measuring. Cortical thickness, bone mineral density, and trabecular bone architecture all give an indication of overall bone loss; however, they reect different aspects of bone maintenance. It is difcult to directly compare studies of trabecular bone loss in archeological samples with those of recent populations. Although similar parameters are examined, studies often use differing methodologies. Further, results in dry archeological bone are not always comparable to in vivo measurements of bone loss. However, it is possible to compare broad patterns and trends in bone loss and osteoporosis. For example, image-analysis studies of trabecular structure in modern vertebrae showed that both the total percentage of trabecular bone (also denoted as BV/TV) (Bergot et al., 1988, 1990) and the number of trabeculae (Twomey et al., 1983) show an age-related decrease. However, these studies reported sex differences in modern trabecular bone structure, with greater postmenopausal bone loss in females as compared to males (Bergot et al., 1988, 1990; Twomey et al., 1983). Strut-analysis of trabecular architecture in modern populations also showed greater loss of connectivity in females as compared to males (Compston et al., 1987). Compston et al. (1987) also noted that the age-related change in connectivity in females occurs primarily between the older age groups, whereas males show more linear age-related change. The patterns of change in trabecular architecture in the Wharram Percy sample are also unusual in comparison to modern patterns of bone loss, in that no signicant change is seen between the middle and oldest age groups. Further, females in this archaeological sample show change in trabecular architecture at a young age. This is not typically reported in any of the parameters in modern populations. Patterns of fragility. As mentioned earlier, the Wharram Percy skeletons do not show a signicant number of typical fragility-related fractures (Mays et al., 1996; Mays, 1998b). The vertebral bones used in this study showed no evidence of fracture, and none of the skeletons in the complete medieval Wharram Percy assemblage showed typical wrist or hip fractures (Mays, 1996, 1998b). Mays (1996, 1998b) did note some rib fractures and compression fractures of the vertebral bodies at Wharram Percy. However, while Mays (1996) correlated these fractures with reduced metacarpal bone mass, it is unclear if these fractures are truly fragility fractures aside from their association with bone mass at this distant skeletal site. It is possible that many of the observed rib fractures are trauma-related (Agarwal, 2001). Further, of the small number of compression

fractures at Wharram Percy, only a few are found in females, and it is unclear if additional pathology may be contributing to these vertebral change (Agarwal, 2001). Understanding the prevalence of bone loss and osteoporotic fracture in mortality samples is difcult. Studies of bone loss and fragility in populations past generally held the a priori expectation that age-related fragility fractures would have occurred in archaeological populations. As such, the low prevalence of fragility fractures in archaeological samples is often explained as a result of mortality bias, low life expectancy in the past, or inaccurate age-atdeath estimates. For example, while the low prevalence of age-related or fragility fractures in some past populations may mean that fracture was rare in the past as compared to modern populations, it could also reect heterogeneity in the oldest age groups, whereby the oldest individuals in skeletal samples may not be developing fragility fractures, as they represent an overall healthier stock that managed to survive into old age. This is particularly important to consider when comparing elderly individuals in the past and the present, as present-day old-aged individuals have beneted from modern medicine and may not be comparable to historic or prehistoric old-aged individuals. However, secular trends do not as easily explain the observed low prevalence in the past of postmenopausal fragility fractures. It was also suggested that perhaps few individuals in past populations would have reached extreme old age to suffer some types of fragility fracture, specifically age-related hip fractures (Mays, 1996, 2000). However, it should be noted that low life expectancy in the past is also related to high infant mortality. Surviving infancy, there was a good possibility of living to an old enough age to suffer fragility fractures (Brickley, 1997). In the case of females, life expectancy is related not only to infant mortality, but also to risks associated with childbirth. Further, there is no reason to believe that human longevity has changed over time, and there is evidence that people did indeed live into old age (Jackes, 2000). Using medieval demographic data from Russell (1948), Jackes (2000) showed that the average age at death of males over age 15 years is around 47 years during the plague, and about 54 years in the periods before and after the plague years. Jackes (2000) further stated that estimates of a 10% survival beyond age 60 would actually be conservative, highlighting the demographic data of Russell (1985) that noted a number of individuals who were expected to have lived beyond 60 across Europe and North Africa in the rst 1500 years AD, and the work of Sjovold (1978), who noted a signicant number of deaths between ages 70 80 in an Austrian village in the 250 years prior to 1852. Mays (1998a) estimated age at death in the Wharram Percy medieval skeletal sample. Using the same age groups as in this study in the total sample of 300 adults, Mays

40

S.C. AGARWAL ET AL.

(1998a) found a good number of older adults (40%), or a 4 in 10 chance of surviving to at least age 50. While the Wharram Percy assemblage represents a long period of time in which longevity may have varied or in which uctuations in birth and death rates may have inuenced the age structure of the assemblage, Mays (1998a) noted that the effects of both would have been minimal. Mays (1996, 1998a) also noted that the age distribution at Wharram Percy is rather similar to British documentary evidence. The study by Russell (1937) of age at death, using 13th and 14th century British postmortem Inquisition records from males 21 and over, found approximately 50% of individuals to have died at 50 years of age or older. Although these data exclude the plague years and are taken from only wealthy classes, it shows a roughly similar age at death structure to Wharram Percy (Mays, 1996, 1998a). Of course, as in Russell (1937), there certainly would have been only a small proportion of the population at Wharram Percy living into the seventies or eighties. While this would explain the lack of changes in trabecular architecture at Wharram Percy related to senile osteoporosis, it still does not fully explain the lack of observed changes expected in the females related to menopause. It should also be noted that in modern times, fracture risk is not tied exclusively to life expectancy. Today, there is a secular trend whereby the increment in the population over age 80 has and will continue to rise exponentially as compared to the overall population growth (Kanis, 1994). However, the change in demographics does not account entirely for the present increased incidence of several types of fragility fracture. For example, Kanis (1994) noted that hip fracture incidence has doubled in periods of time as short as a few decades in many parts of the world. Life expectancy is thus not the only factor involved in the increasing incidence of osteoporosis. Perhaps the most concerning problem with using archeological skeletal samples is age at death estimations (Jackes, 2000; Milner et al., 2000; Saunders and Hoppa, 1993). It is increasingly evident that while humans in the past did likely manage to live into old age, we cannot accurately age skeletons older than around 50 years of age. As in this study, the conservative approach in osteological studies has been to assign only broad age groups with a nal open-end age group of 45 or 50 to skeletons. However, it was suggested that if we cannot break down our age-at-death estimation after 50 into ner groups, we may not be able to adequately study the rates of degenerative or age-related conditions such osteopenia (Jackes, 2000). Again, while this may hold true when looking exclusively at age-related or senile bone loss and osteoporosis, certainly the use of broad age categories, such as those used in this study, are likely adequate to discern changes and patterns in bone structure and fragility in females that are related to menopause.

Mays (1996, 1998b) suggested that the lack of specically hip fractures at Wharram Percy may be related to the impact of energy during falls. He suggested that individuals at Wharram Percy may have been protected from bone fractures occurring during a fall as they would have occurred on soft, yielding surfaces such as grass or bare earth that would have reduced the force on impact. Alternatively, Pfeiffer (2000) suggested that the paucity of fragility fractures in the past may be related to the difculty in quantifying or recognizing fragility fractures in the archaeological record, particularly the variable degrees of vertebral compression fractures. However, the results of this study appear to support the paucity of visible vertebral compression fractures, and indicate no signicant loss of microscopic trabecular structure between middle and old age that would be expected to precede even a small degree of vertebral compression. Along with the low prevalence of fragility fractures in the Wharram Percy sample, the ndings in this study suggest that age-related patterns of trabecular architecture are different in the past, as expected from patterns of bone loss in modern populations. Influence of lifestyle factors. Lifestyle factors are known to play a signicant role in bone maintenance and subsequent bone loss. It is difcult to extrapolate the ndings on lifestyle factors in modern populations in order to hypothesize fracture risk in the past. However, certain factors seem more likely to have played a role in bone maintenance in the past than others. For example, while smoking, alcohol consumption, and caffeine consumption are cited risk factors of osteoporosis in modern populations, they are likely not as important in the interpretation of osteoporosis and bone loss in the past. Mays (1996) noted that smoking cannot be considered a risk factor at Wharram Percy, as tobacco was unknown in medieval England. Similarly, although alcohol consumption was likely popular historically, it is unclear if sufcient quantities were being consumed to be a risk factor for bone loss (Brickley, 1997; Mays, 1996). Factors such as nutrition, physical activity, and reproduction may have played more substantial roles in bone loss and osteoporosis in the past. Medieval peasant diets would have suffered during the winter months with diets short in protein and lacking in lipids, calcium, and other essential vitamins (Gies and Gies, 1990). However, it is not certain if this degree of poor nutrition would have translated into bone loss in the Wharram Percy population. Calcium deciency is often regarded as a potential factor in bone loss. It is unlikely that the Wharram Percy diet was decient in calcium, as many cattle bones and fragments of pottery vessels used for dairy products were found during excavation (Beresford and Hurst, 1990; Hurst, 1984; Mays, 1996; Ryder, 1974). Further, ndings in modern populations indicate that although calcium greatly inuences

MEDIEVAL TRABECULAR BONE ARCHITECTURE

41

the growing skeleton, it has only a weak inuence on adult bone loss (Kanis, 1994; Slemenda and Johnston, 1990). It is possible that vitamin D deciency may have had a more signicant effect on bone maintenance in the Wharram Percy population. However, no signs of osteomalacia were found in any of the medieval Wharram Percy skeletons, and Mays (1996) suggested that the outdoor lifestyle of the medieval peasants would have likely prevented vitamin D deciency. While nutritional factors could explain some of the early age loss and change in trabecular architecture at Wharram Percy, they do not easily explain the unusual ageand sex-related patterns of trabecular architecture observed in this study. It is possible that inuences of nutritional deciency on trabecular bone maintenance may have been counteracted to some degree by the competing inuence of physical activity. Physical activity is considered an important risk factor in osteoporosis, affecting both the achievement of peak bone mass and likely the subsequent rate of bone loss and deterioration of bone quality. The reduction in habitual physical activity in modern Western populations was suggested as a primary explanation for the varied and increasing geographic incidence in osteoporotic fractures (Kanis, 1994; Mosekilde, 1995). Studies of age-related change in bone mineral density in European archeological skeletons discussed the possible role of physical activity in preventing old age bone loss (Ekenman et al., 1995; Lees et al., 1993). Physical activity would have been undoubtedly greater, as compared to modern populations, in the Wharram Percy agricultural settlement. Physical labor would have been required in all areas of rural life. Men would have been responsible for tasks involving heavy lifting and the use of ox- or horsedrawn plows, while females would have been responsible for domestic work such as cleaning, gardening, caring for livestock, and weaving (Mays, 1996; Sofaer Derevesnki, 2000). However, the division of labor in medieval rural communities was likely uid (Bennett, 1997), and historical records emphasize equally active lifestyles in both sexes (Leyser, 1995; Mays, 1996; Razi, 1980). Women would have frequently joined in agricultural tasks and participated in manual work in the eld (Bennett, 1997). It is likely that bone maintenance in both sexes would have beneted from generally higher levels of physical activity as compared to modern populations. Although the role of physical activity in the distinct patterns of trabecular bone architecture observed in the Wharram Percy sample is unclear, it is possible that greater physical activity may have prevented loss of trabecular architecture in old age. The unusual patterns of trabecular architecture seen particularly in female skeletons from Wharram Percy may be more clearly explained by reproductive factors. Pregnancy and lactation are known to be high bone turnover states; however, the long-

term effect of pregnancy and lactation on bone loss and fragility is not clearly understood. Studies of the effects of pregnancy on bone found conicting results (Cross et al., 1995; Drinkwater and Chesnut, 1991; Kent et al., 1993; Naylor et al., 2000; Sowers et al., 1991); however, epidemiological evidence suggests that parity may decrease fracture risk and could increase bone density (Fox et al., 1993; Murphy et al., 1994; Sowers et al., 1992). While longitudinal studies indicate that bone loss (primarily trabecular bone loss) can occur during pregnancy and initial lactation (Afnito et al., 1996; Chan et al., 1982; Drinkwater and Chesnut, 1991; Hayslip et al., 1989; Kent et al., 1993; Lamke et al., 1977; Lopez et al., 1996; Sowers, 1996; Sowers et al., 1993, 1995), there is also convincing evidence that recovery of bone occurs with extended lactation and during weaning (Afnito et al., 1996; Kent et al., 1990; Lopez et al., 1996; Sowers, 1996; Sowers et al., 1993, 1995). It seems likely that reproductive factors would have played a role in bone maintenance in historical females, and previous studies of bone loss in archeological skeletons indeed discussed the possible role of pregnancy and lactation. For example, Poulsen et al. (2001) suggested that signicant decrease in bone mineral density in young Danish medieval skeletons could be the result of pregnancy and lactation stress. The authors in fact suggested that the physiological demands associated with pregnancy and breast-feeding may have increased mortality in young medieval women, although no evidence is given of the circumstances in which reproduction would have been so detrimental (Poulsen et al., 2001). In contrast, Vogel et al. (1990) suggested that parity may have played a role in better trabecular connectivity as compared to modern populations in female skeletons from European historical populations. Although we have no direct evidence of pregnancy and lactation practices at Wharram Percy, parity would likely have been higher in the rural medieval population, as compared to modern populations. Extended lactation was also practiced in medieval times. Rural medieval peasants, such as those at Wharram Percy, would likely have had to nurse their own children (Gies and Gies, 1990). The health benets of breastfeeding for both the mother and child (Gies and Gies, 1990), and the possible contraceptive function of breast-feeding, were also known by the medieval period, perhaps accounting for later weaning ages in the medieval period (Fildes, 1995). The patterns of age-related change in trabecular architecture in the female skeletons in this study may be related to the unique hormonal environment created by these reproductive practices. Increased parity and extended lactation would also have resulted in lower steroid exposure in the historic females, in comparison to modern Western females who give birth to fewer children and practice little or no breast-feeding (Pollard, 1999; Sperling and Beyene, 1997; Strassmann, 1997). Further, age of menarche is known to show a

42

S.C. AGARWAL ET AL.

secular change, specically decreasing in recent modern populations (Sperling and Beyene, 1997; Treloar, 1974), and although age at menopause has been regarded as generally stable through human evolution (Pavelka and Fedigan, 1991), there is evidence that many populations can show earlier onset of menopause in comparison to industrialized Western populations (Eaton et al., 1994; Sperling and Beyene, 1997). It was estimated that women in populations with life cycles that include high parity, extended lactation, late age of menarche, and early age at menopause could have as few as 50 100 menstrual cycles, while in comparison, modern Western females experience approximately as many as 420 menstrual cycles (Eaton et al., 1994; Sperling and Beyene, 1997; Strassmann, 1997; Weaver, 1998). Although it is uncertain where the reproductive patterns at Wharram Percy would fall in relation to these estimates, it is clear that the hormonal milieu would have been different in these historical females. The less dramatic change in bone after menopause observed in this study may simply reect the lower levels of steroid exposure in these historical females. Weaver (1998) specically suggested that the pattern of bone loss in modern females may be related to the sudden downregulation of bone forming osteoblast cells that are elevated with chronically high levels of estrogen. At the same time, reproductive patterns may additionally explain some of the observed poor trabecular bone architecture in the young-age females in this population. We may simply be observing skeletons of premenopausal women of reproductive age who were pregnant or breast-feeding at time of death. Their maternal skeletons may not have had the sufcient time to recover from the demands of pregnancy or the start of lactation, which it is capable of doing. Interpreting the patterns of bone loss and quality is speculative; however, reproductive patterns were undoubtedly different in this historic rural group in comparison to modern populations. In this light, it is not surprising that the patterns of change in trabecular bone architecture in the historical females differ from what is expected. CONCLUSIONS It was demonstrated in this study that trabecular architecture can be successfully examined in archeological bone with the use of image analysis. The study of trabecular architecture avoids the confounding issues of chemical diagenetic change, and provides a relatively simple assessment of age-related change in one aspect of bone quality. Although largely ignored in previous studies of archeological samples, it is known that bone quality plays a vital role in bone strength and fragility. Both sexes at Wharram Percy show a decrease in vertebral trabecular structure by middle age; however, this decrease appears more gradually in males. Only males show a signicant age-related change in connectivity, and both sexes show no change between middle and old

age in any parameter. While age-related changes in trabecular bone architecture can occur independent of changes in bone volume (Banse, 2002), in this study, BV/TV was highly correlated to the other structural histomorphometric and strut-analysis parameters, and the relationship between the various parameters with age could be explained by the close relationship with BV/TV. Along with the paucity of typical fragility fractures at Wharram Percy, these patterns of age-related change in trabecular architecture differ from the patterns generally observed in modern populations. While speculative, we suggest that greater levels of physical activity may explain the preservation of bone structure in both sexes, and that reproductive factors likely played an additional role in the case of female bone maintenance in the past. The patterns of change in trabecular architecture seen in historical females are not surprising, given the dramatically different reproductive behaviors and hormonal milieu that the Wharram Percy females would have been exposed to, as compared to modern females. While the interpretation of patterns of bone loss and fragility in the past is speculative and specically unique to each population, it is clear that a complete picture of osteoporosis in past populations can only be obtained with further paleopathological studies that emphasize the evolutionary and biocultural approach and that recognize the important role of bone quality. ACKNOWLEDGMENTS This study was funded in part by doctoral fellowships from the Natural Sciences and Engineering Research and Social Sciences and Humanities Research Councils of Canada (to S.C.A.). Access to the Wharram Percy skeletal material was given by the kind permission of Dr. Simon Mays and the English Heritage. LITERATURE CITED
Afnito P, Tommaselli GA, di Carlo C, Guida F, Nappi C. 1996. Changes in bone mineral density and calcium metabolism in breastfeeding women: a one year follow-up study. J Clin Endocrinol Metab 81:2314 2318. Agarwal SC. 2001. The inuence of age and sex on trabecular architecture and bone mineral density in three British historical populations. Ph.D. thesis, University of Toronto, Toronto. Agarwal SC, Grynpas MD. 1996. Bone quantity and quality in past populations. Anat Rec 246:423 432. Amling M, Posl M, Ritzel H, Hahn M, Vogel M, Wening VJ, Delling G. 1996. Architecture and distribution of cancellous bone yield vertebral fracture clues. A histomorphometric analysis of the complete spinal column from 40 autopsy specimens. Arch Orthop Trauma Surg 115:262269. Banse X. 2002. When density fails to predict bone strength. Acta Orthop Scand [Suppl] 73:157. Bennett JM. 1987. Women in the mediaeval English countryside. Oxford: Oxford University Press. Beresford M, Hurst J. 1990. Wharram Percy, deserted medieval village. London: Batsford/English Heritage. Bergot C, Laval-Jeantet AM, Preteux F, Meunier A. 1988. Measurement of anisotropic vertebral trabecular bone loss during aging by quantitative image analysis. Calcif Tissue Int 43:143 149.

MEDIEVAL TRABECULAR BONE ARCHITECTURE

Bergot C, Laval-Jeanet AM, Bouysse S, Laval-Jeanet M. 1990. Variations du volume osseux et de la densite minerale au cours du viellissement chez lhomme. Rev Rhum Mal Osteoartic 57: 791798. Brickley MB. 1997. Age-related bone loss and osteoporosis in archaeological bone: a study of two London collections, Redcross Way and Farringdon Street. Ph.D. dissertation, University College London, London. Brickley M, Howell PG. 1999. Measurement of changes in trabecular bone structure with age in an archaeological population. J Archaeol Sci 26:151157. Brothwell DR. 1981. Digging up bones. London: Oxford University Press. Chan GM, Slater P, Ronald N, Roberts CC, Thomas MR, Folland D, Jackson R. 1982. Bone mineral status of lactating mothers of different ages. Am J Obstet Gynecol 144:438 441. Chappard D, Legrand E, Pascaretti C, Basle MF, Audran M. 1999. Comparison of eight histomorphometric methods for measuring trabecular bone architecture by image analysis on histological sections. Microsc Res Tech 45:303312. Compston JE. 1994. Connectivity of cancellous bone: assessment and mechanical implications [editorial]. Bone 15:463 46. Compston JE, Mellish RW, Garrahan NJ. 1987. Age-related changes in iliac crest trabecular microanatomic bone structure in man. Bone 8:289 292. Compston JE, Mellish RW, Croucher P, Newcombe R, Garrahan NJ. 1989. Structural mechanisms of trabecular bone loss in man. Bone Miner 6:339 350. Cross NA, Hillman LS, Allen SH, Krause GF, Vieira NE. 1995. Calcium homeostasis and bone metabolism during pregnancy, lactation, and postweaning: a longitudinal study. Am J Clin Nutr 61:514 523. Croucher PI, Garrahan NJ, Compston JE. 1996. Assessment of cancellous bone structure: comparison of strut analysis, trabecular bone pattern factor, and marrow space star volume. J Bone Miner Res 11:955961. Ding M, Odgaard A, Linde F, Hvid I. 2002. Age-related variations in the microstructure of human tibial cancellous bone. J Orthop Res 20:615 621. Drinkwater BL, Chesnut CHD. 1991. Bone density changes during pregnancy and lactation in active women: a longitudinal study. Bone Miner 14:153160. Eaton SB, Pike MC, Short RV, Lee NC, Trussell J, Hatcher RA, Wood JW, Worthman CM, Jones NG, Konner MJ, et al. 1994. Womens reproductive cancers in evolutionary context. Q Rev Biol 69:353367. Ekenman I, Eriksson SA, Lindgren JU. 1995. Bone density in medieval skeletons. Calcif Tissue Int 56:355358. Fazzalari NL. 1993. Trabecular microfracture. Calcif Tissue Int 53:143147. Fildes V. 1995. The culture and biology of breastfeeding: an historical review of Western Europe. In: Stuart-Macadam P, Dettwyler KA , editors. Breastfeeding: biolcultural perspectives. New York: Aldine de Gruyter. p 101126. Fox KM, Magaziner J, Sherwin R, Scott JC, Plato CC, Nevitt M, Cummings S. 1993. Reproductive correlates of bone mass in elderly women. Study of Osteoporotic Fractures Research Group. J Bone Miner Res 8:901908. Garrahan NJ, Mellish RW, Compston JE. 1986. A new method for the two-dimensional analysis of bone structure in human iliac crest biopsies. J Microsc 142:341349. Geraets WG, Van der Stelt PF, Netelenbos CJ, Elders PJ. 1990. A new method for automatic recognition of the radiographic trabecular pattern. J Bone Miner Res 5:227233. Geraets WG, Van der Stelt PF, Elders PJ. 1993. The radiographic trabecular bone pattern during menopause. Bone 14:859 864. Gies F, Gies G. 1990. Life in a medieval village. London: Harper Row. Gordon CL, Lang TF, Augat P, Genant HK. 1998. Image-based assessment of spinal trabecular bone structure from high-resolution CT images. Osteoporos Int 8:317325. Grynpas M. 1993. Age and disease-related changes in the mineral of bone. Calcif Tissue Int [Suppl] 53:57 64.

43

Grynpas MD, Acito A, Dimitriu M, Mertz BP, Very JM. 1992. Changes in bone mineralization, architecture and mechanical properties due to long-term (1 year) administration of pamidronate (APD) to adult dogs. Osteoporos Int 2:74 81. Grynpas MD, Kasra M, Dumitriu M, Nespeca R, Very JM, Mertz BP. 1994. Recovery from pamidronate (APD): a two-year study in the dog. Calcif Tissue Int 55:288 294. Hayslip CC, Klein TA, Wray HL, Duncan WE. 1989. The effects of lactation on bone mineral content in healthy postpartum women. Obstet Gynecol 73:588 592. Heaney RP. 1993. Is there a role for bone quality in fragility fractures? Calcif Tissue Int [Suppl] 53:3 6. Hurst JG. 1984. The Wharram Percy project: results to 1983. Medieval Archaeol 28:77111. Jackes M. 2000. Building the bases for paleodemographic analyses: adult age determination. In: Katzenberg MA, Saunders SR, editors. Biological anthropology of the human skeleton. New York: Wiley Liss. p 417 466. Kanis JA. 1990. Osteoporosis and osteopenia. J Bone Miner Res 5:209 211. Kanis JA. 1994. Osteoporosis. Oxford: Blackwell Sience. Kent GN, Price RI, Gutteridge DH, Smith M, Allen JR, Bhagat CI, Barnes MP, Hickling CJ, Retallack RW, Wilson SG, et al. 1990. Human lactation: forearm trabecular bone loss, increased bone turnover, and renal conservation of calcium and inorganic phosphate with recovery of bone mass following weaning. J Bone Miner Res 5:361369. Kent GN, Price I, Gutteridge DH, Allen JR, Rosman KJ, Smith M, Bhagat CI, Wilson SG, Retallack RW. 1993. Effect of pregnancy and lactation on maternal bone mass and calcium metabolism. Osteoporos Int [Suppl] 3:44 47. Kleerekoper M, Villanueva AR, Stanciu J, Rao DS, Partt AM. 1985. The role of three-dimensional trabecular microstructure in the pathogenesis of vertebral compression fractures. Calcif Tissue Int 37:594 597. Kneissel M, Boyde A, Hahn M, Teschler-Nicola M, Kalchhauser G, Plenk H Jr. 1994. Age- and sex-dependent cancellous bone changes in a 4000y BP population. Bone 15:539 545. Kneissel M, Roschger P, Steiner W, Schamall D, Kalchhauser G, Boyde A, Teschler-Nicola M. 1997. Cancellous bone structure in the growing and aging lumbar spine in a historic Nubian population. Calcif Tissue Int 61:95100. Korstjens CM, Spruijt RJ, Geraets WG, Mosekilde L, van der Stelt PF. 1997. Reliability of an image analysis system for quantifying the radiographic trabecular pattern [letter]. IEEE Trans Med Imag 16:230 234. Korstjens CM, Spruijt RJ, Mosekilde L, Geraets WG, van der Stelt PF. 1998. Agreement between radiographic and photographic trabecular patterns. Acta Radiol (Stockh) 39:625 631. Lamke B, Brundin J, Moberg P. 1977. Changes of bone mineral content during pregnancy and lactation. Acta Obstet Gynecol Scand 56:217219. Lees B, Molleson T, Arnett TR, Stevenson JC. 1993. Differences in proximal femur bone density over two centuries. Lancet 341:673 675. Leyser H. 1995. Medieval women. London: St. Martins Press. Lopez JM, Gonzalez G, Reyes V, Campino C, Diaz S. 1996. Bone turnover and density in healthy women during breastfeeding and after weaning. Osteoporos Int 6:153159. Lundon K, Dumitriu M, Grynpas MD. 1997. Supraphysiologic levels of testosterone affect cancellous and cortical bone in the young female cynomolgus monkey. Calcif Tissue Int 60:54 62. Manolagas SC, Jilka RL. 1995. Bone marrow, cytokines, and bone remodeling. Emerging insights into the pathophysiology of osteoporosis. N Engl J Med 332:305311. Martin B. 1993. Aging and strength of bone as a structural material. Calcif Tissue Int [Suppl] 53:34 40. Mays SA. 1996. Age-dependent cortical bone loss in a medieval population. Int J Osteoarcheol 6:144 154. Mays S. 1998a. The archaeology of human bones. London: Routledge. Mays S. 1998b. Osteoporosis in earlier human populations. J Clin Densitometry 2:7178.

44

S.C. AGARWAL ET AL.

Serra J. 1982. Image analysis and mathematical morphology. London: Academic Press. Sjovold T. 1978. Inference concerning the age distribution of skeletal populations and some consequences for paleodemography. Anthropol Kozl 22:99 114. Slemenda CW, Johnston CC. 1990. Osteoporotic fractures. In: Simmons DJ, editor. Nutrition and bone development. New York: Oxford University Press. p 131147. Snyder BD, Piazza S, Edwards WT, Hayes WC. 1993. Role of trabecular morphology in the etiology of age-related vertebral fractures. Calcif Tissue Int [Suppl] 53:14 22. Sofaer Derevenski JR. 2000. Sex differences in activity-related osseous change in the spine and the gendered division of labor at Ensay and Wharram Percy, UK. Am J Phys Anthropol 111: 333354. Sowers M. 1996. Pregnancy and lactation as risk factors for subsequent bone loss and osteoporosis. J Bone Miner Res 11: 10521060. Sowers MR, Galuska DA. 1993. Epidemiology of bone mass in premenopausal women. Epidemiol Rev 15:374 398. Sowers M, Crutcheld M, Jannausch M, Updike S, Corton G. 1991. A prospective evaluation of bone mineral change in pregnancy. Obstet Gynecol 77:841 845. Sowers M, Clark MK, Hollis B, Wallace RB, Jannausch M. 1992. Radial bone mineral density in pre- and perimenopausal women: a prospective study of rates and risk factors for loss. J Bone Miner Res 7:647 657. Sowers M, Corton G, Shapiro B, Jannausch ML, Crutcheld M, Smith ML, Randolph JF, Hollis B. 1993. Changes in bone density with lactation. JAMA 269:3130 3135. Sowers M, Eyre D, Hollis BW, Randolph JF, Shapiro B, Jannausch ML, Crutcheld M. 1995. Biochemical markers of bone turnover in lactating and nonlactating postpartum women. J Clin Endocrinol Metab 80:2210 2216. Sperling S, Beyene Y. 1997. A pound of biology and a pinch of culture or a pinch of biology and a pound of culture? The necessity of integrating biology and culture in reproductive studies. In: Hager L, editor. Women in human evolution. New York: Routledge. p 137152. Stevenson JC, Lees B, Devenport M, Cust MP, Ganger KF. 1989. Determinants of bone density in normal women: risk factors for future osteoporosis? Br Med J [Clin Res] 298:924 928. Strassmann BI. 1997. The biology of menstruation in Homo sapiens: total lifetime menses, fecundity, and nonsynchrony in a natural-fertility population. Curr Anthropol 38:123129. Thomsen JS, Ebbesen EN, Mosekilde L. 2000. A new method of comprehensive static histomorphometry applied on human lumbar vertebral cancellous bone. Bone 27:129 138. Thomsen JS, Ebbesen EN, Mosekilde L. 2002. Predicting human vertebral bone strength by vertebral static histomorphometry. Bone 30:502508. Treloar AE. 1974. Menarche, menopause, and intervening fecundability. Hum Biol 46:89 107. Twomey L, Taylor J, Furniss B. 1983. Age changes in the bone density and structure of the lumbar vertebral column. J Anat 136:1525. Vesterby A. 1990. Star volume of marrow space and trabeculae in iliac crest: sampling procedure and correlation to star volume of rst lumbar vertebra. Bone 11:149 155. Vesterby A, Mosekilde L, Gundersen HJ, Melsen F, Holme K, Sorensen S. 1991. Biologically meaningful determinants of the in vitro strength of lumbar vertebrae. Bone 12:219 224. Vogel M, Hahn M, Caselitz P, Woggan J, Pompesius-Kempa M, Delling G. 1990. Comparison of trabecular bone structure in man today and an ancient population in Western Germany. In: Takahashi HE, editor. Bone morphometry. Tokyo, Japan: Nishimura Co. p 220 223. Ward JA, Lord SR, Williams P, Anstey K, Zivanovic E. 1995. Physiologic, health and lifestyle factors associated with femoral neck bone density in older women. Bone 16:373378. Weaver DS. 1998. Osteoporosis in the bioarchaeology of women. In: Grauer A, Stuart-Macadam P, editors. Sex and gender in paleopathological perspective. Cambridge: Cambridge University Press. p 27 46.

Mays S. 2000. Age-dependent cortical bone loss in women from 18th and early 19th century London. Am J Phys Anthropol 112:349 361. Mays SA, Lees B, Stevenson JC. 1998. Age-dependent bone loss in the femur in a medieval population. Int J Osteoarchaeol 8:97106. Mellish RW, Ferguson-Pell MW, Cochran GV, Lindsay R, Dempster DW. 1991. A new manual method for assessing two-dimensional cancellous bone structure: comparison between iliac crest and lumbar vertebra. J Bone Miner Res 6:689 696. Melton LJD, Wahner HW. 1989. Dening osteoporosis [editorial]. Calcif Tissue Int 45:263264. Milner GR, Wood JW, Boldsen JL. 2000. Paleodemography. In: Katzenberg MA, Saunders SR, editors. Biological anthropology of the human skeleton. New York: Wiley Liss. p 467 497. Mosekilde L. 1988. Age-related changes in vertebral trabecular bone architectureassessed by a new method. Bone 9:247250. Mosekilde L. 1989. Sex differences in age-related loss of vertebral trabecular bone mass and structure biomechanical consequences. Bone 10:425 432. Mosekilde L. 1993. Vertebral structure and strength in vivo and in vitro. Calcif Tissue Int [Suppl] 53:121126. Mosekilde L. 1995. Osteoporosis and exercise. Bone 17:193195. Murphy S, Khaw KT, May H, Compston JE. 1994. Parity and bone mineral density in middle-aged women. Osteoporos Int 4:162166. Naylor KE, Iqbal P, Fledelius C, Fraser RB, Eastell R. 2000. The effect of pregnancy on bone density and bone turnover. J Bone Miner Res 15:129 137. Partt AM. 1992. Implications of architecture for the pathogenesis and prevention of vertebral fracture. Bone [Suppl] 13:41 47. Partt AM. 1993. Bone age, mineral density, and fatigue damage. Calcif Tissue Int [Suppl] 53:82 86. Partt AM, Drezner MK, Glorieux FH, Kanis JA, Malluche H, Meunier PJ, Ott SM, Recker RR. 1987. Bone histomorphometry: standardization of nomenclature, symbols, and units. Report of the ASBMR Histomorphometry Nomenclature Committee. J Bone Miner Res 2:595 610. Pavelka MS, Fedigan LM. 1991. Menopause: a comparative life history perspective. Yrbk Phys Anthropol 34:1338. Pfeiffer S. 2000. Paleohistology: health and disease. In: Katzenberg MA, Saunders SA, editors. Biological anthropology of the human skeleton. New York: Wiley-Liss. p 287302. Pfeiffer SK, Lazenby RA. 1994. Low bone mass in past and present aboriginal populations. In: Draper HH, editor. Advances in nutritional research. New York: Plenum Press. p 3551. Pollard TM. 1999. Sex, gender and cardiovascular disease. In: Pollard TM, Hyatt SB, editors. Sex, gender, and health. Cambridge: Cambridge University Press. p 5374. Poulsen LW, Qvesel D, Brixen K, Vesterby A, Boldsen JL. 2001. Low bone mineral density in the femoral neck of medieval women: a result of multiparity? Bone 28:454 458. Razi Z. 1980. Life, marriage and death in a medieval parish. Cambridge: Cambridge University Press. Recker RR. 1993. Architecture and vertebral fracture. Calcif Tissue Int [Suppl] 53:139 142. Riggs RL, Melton LJ. 1988. Osteoporosis: etiology, diagnosis and management. New York: Raven Press. Russell JC. 1937. Length of life in England 1250 1348. Hum Biol 9:528 541. Russell JC. 1948. British medieval population. Albuquerque: University of New Mexico Press. Russell JC. 1985. The control of late ancient and and medieval population. Philadelphia: American Philosophical Society. Ryder ML. 1974. Animal remains from Wharram Percy. Yorkshire Archaeol J 46:4252. Saunders SR, Hoppa RD. 1993. Growth decit in survivors and non-survivors: biological mortality bias in subadult skeletal samples. Yrbk Phys Anthropol 36:127151. Schafer MB, Radin EL, Burr DB. 1989. Mechanical and morphological effects of strain rate on fatigue of compact bone. Bone 10:2072014.