Bone Microstructure and Changes in Tissue Mineralization Throughout Adulthood

Bone Microstructure and Changes in Tissue Mineralization throughout Adulthood

Amy Catherine Beresheim

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Department of Anthropology University of Toronto

Bone Microstructure and Changes in Tissue Mineralization throughout Adulthood

Amy Catherine Beresheim

Doctor of Philosophy

Department of Anthropology University of Toronto

2018 Abstract

This research aims to explore the global variability in the structural and material properties of mid-thoracic ribs within a large (n=213) sample of known age, sex, and government “race” designation from apartheid South Africa [DOD 1967-1998]. Most individuals in the study sample were non-white, and inferred to be of poor socioeconomic status. Using linearly polarized light microscopy (LPLM) on the full sample, and back-scattered scanning electron microscopy

(BSE-SEM) on a subset (n=143), histomorphometric parameters, and the average tissue mineralization of the cortical and trabecular bone compartments were assessed in photomontages of transverse rib cross-sections. The influence of body size on bone mass and histomorphometry was first considered. Tissue-level changes were then interpreted in a biosocial context, exploring variation associated with adverse apartheid living conditions and adult life history.

Body size does not appear to correlate with either bone mass or histomorphometry, suggesting that size-standardization may not be necessary in studies of rib bone microstructure. Compared to the women in the research sample, men exhibited delayed peak bone mass attainment, lower osteocyte densities, and lower average tissue mineralization. Women achieved peak bone mass when anticipated. Relative cortical area, osteon area, bone tissue mineralization, and cortical

ii porosity were the best histological indicators of menopause in women. Poorer bone health indices in males infer greater systematic marginalization under apartheid rule. Men in this sample may have been more susceptible to dietary deficiencies and substance abuse issues, leading to compromised bone mass and quality.

This research demonstrates that osteoporosis risk is not just a concern of the aged, white female population in South Africa. It provides novel data on an understudied population, and underscores the importance of skeletal research collections for the study of contemporary epidemiological issues.

iii

Acknowledgments

First, I would like to thank my doctoral supervisor, Susan Pfeiffer. She allowed me to expand my work beyond forensic research questions, and introduced me to two completely foreign places— Canada and South Africa. Throughout the dissertation process she has been firm, but also incredibly kind and supportive. I am very proud to be part of her legacy and deeply honored that she agreed to take me on as her final PhD student. I am also thankful for the wonderful network of former students and research collaborators that she has created and allowed me to join.

I am indebted to my own research collaborators, Susan Pfeiffer, Marc Grynpas, Amanda Alblas, Jarred Heinrich, Jacklynn Walters, and Michelle Cameron. Without them, none of this work could have been possible. I need to acknowledge Benedict Paige, Linda Greyling, support staff in the Division of Anatomy and Health Sciences at Stellenbosch University, and the Western Cape Government Inspectorate of Anatomy for granting and facilitating my access to such an important skeletal collection. I would also like to express my tremendous gratitude and appreciation for the undergraduate students who directly contributed to this research by helping me process my samples for imaging and analysis: Virginia Pritzler, Klara Komza, Meimei Fong, Matthew Gray and Liam Wadsworth.

I would like to thank my core committee, Susan Pfeiffer, Marc Grynpas and Mary Silcox, as well as the Evolutionary Anthropology faculty that have helped guide me through this process. I would be remiss if I did not explicitly acknowledge the Pfeiffer and Viola labs. More broadly, I would like to thank the administrative and custodial staff, as well as the graduate student body in the Department of Anthropology. In particular, I am grateful for Elizabeth Sawchuk, Lindi Masur, Emma Yasui, Andrew Holmes, Sarah Ranlett, Courtneay Hopper, Steven Dorland, Kathy Pitirri, Jarred Heinrich, Michelle Cameron, Thivviya Vairamuthu, Catherine Merritt, Bess Doyle, Eve Smeltzer, Walter Callaghan, Andrew Harris, and Amy Fox.

The real driving force behind this project has been my friends and family. I express my deep love and appreciation for lifelong pals Maxwell Losgar, Belinda Smith, Michelle Gilbert, Michael Simmons, Colleen Cheverko, James Scheuermann, Paul Meddaugh, Maria Darr, Michala Stock, Amy Johnson, and Christian Quintanilla. The Vesley, Dunlop, Herzman, MacNeil, and Conlin clans have all provided me with endless encouragement and support. I could not be more grateful

iv to belong to such an incredible family network. Above all others, I would like to thank my parents, Jane and the late Joseph Beresheim, my sister, Amanda, and my eternal rock, Cody MacNeil.

Table of Contents

Acknowledgments ...... iv Table of Contents ...... vi List of Tables ...... ix List of Figures ...... x List of Appendices ...... xiii List of Appendix Figures ...... xiv Chapter 1 Introduction ...... 1

1.1 Research Hypotheses ...... 5 1.2 Organization of the Thesis ...... 6 1.3 Literature Cited ...... 6

Chapter 2 Literature Review ...... 10

2.1 Modeling ...... 10 2.2 Remodeling and Bone Cells ...... 10 2.3 Measures of Bone Mass and Density ...... 12

2.3.1 Back-Scattered Scanning Electron Microscopy (BSE-SEM) ...... 13

2.4 Overview of Major Life History Events ...... 15

2.4.1 Acquisition of Peak Bone Mass ...... 15 2.4.2 Menopause and Primary Osteoporosis ...... 18

2.5 Rib Bone Histomorphometry ...... 19

2.5.1 Histomorphometric Changes in the Rib from Birth through Senescence ...... 21 2.5.2 Biomechanics of the Thorax and Body Size Considerations ...... 22

2.6 South African Apartheid and the Social Determinants of Health ...... 23

2.6.1 Peak Bone Mass Attainment in South Africa ...... 24 2.6.2 Osteoporosis and Fragility Fractures in South Africa ...... 25

2.7 Genetic Considerations ...... 26 2.8 Literature Cited ...... 27

Chapter 3 An Exploration of Body Size and Bone Mass on Cortical Bone Histomorphometry in Human Ribs...... 41

3.1 Abstract ...... 41 3.2 Introduction ...... 42 3.3 Materials and Methods ...... 44

3.3.1 Research Sample ...... 44 3.3.2 Histological Preparation, Imaging, and Analysis ...... 45 vi

3.3.3 Statistical Analysis ...... 49

3.4 Results ...... 49 3.5 Discussion ...... 52

3.5.1 Body Size ...... 52 3.5.2 Osteon Population Density (OPD) ...... 53 3.5.3 Osteon Area (On.Ar) ...... 54 3.5.4 Other Considerations ...... 55

3.6 Conclusion ...... 56 3.7 Acknowledgements ...... 56 3.8 Literature Cited ...... 57

Chapter 4 Sex-specific Patterns in Cortical and Trabecular Bone Microstructure in the Kirsten Skeletal Collection, South Africa ...... 62

4.1 Abstract ...... 62 4.2 Introduction ...... 63

4.2.1 Apartheid and the Social Determinants of Health ...... 64 4.2.2 Histomorphometry ...... 65 4.2.3 Study Objectives ...... 65

4.3 Materials and Methods ...... 66

4.3.1 Research Sample ...... 66 4.3.2 Sample Selection and Tissue Processing ...... 68 4.3.3 Ethics Statement ...... 69 4.3.4 Data Collection ...... 69 4.3.5 Statistical Analysis ...... 69

4.4 Results ...... 73

4.4.1 Cortical Bone Histomorphometry ...... 73 4.4.2 Trabecular Bone Histomorphometry ...... 73

4.5 Discussion ...... 74

4.5.1 Osteon Population Density (OPD) ...... 74 4.5.2 Osteon Area (On.Ar) ...... 74 4.5.3 Relative Cortical Area (Rt.Ct.Ar) ...... 75 4.5.4 Trabecular Bone Histomorphometry ...... 79 4.5.5 Future Directions ...... 82

4.6 Conclusion ...... 82 4.7 Acknowledgements ...... 82 4.8 Author Contributions ...... 83 4.9 Literature Cited ...... 83

vii

Chapter 5 Use of Backscattered Scanning Electron Microscopy to Quantify the Bone Tissue of Mid-Thoracic Human Ribs ...... 92

5.1 Abstract ...... 92 5.2 Introduction ...... 93

5.2.1 South African Apartheid and the Social Determinants of Health ...... 93 5.2.2 Back-Scattered Scanning Electron Microscopy (BSE-SEM) and Bone Tissue Mineralization ...... 94 5.2.3 Indicators of Bone Quality ...... 94

5.3 Materials and Methods ...... 95

5.3.1 Research Sample ...... 95 5.3.2 SEM Preparation ...... 96 5.3.3 Image Acquisition and Analysis ...... 97 5.3.4 Statistical Analysis ...... 98 5.3.5 Results ...... 104

5.4 Discussion ...... 105

5.4.1 Intracortical Porosity ...... 105 5.4.2 Osteocyte Lacunar Density ...... 109 5.4.3 Degree of Bone Mineralization ...... 112 5.4.4 South African Apartheid and Consequences to Bone Health ...... 115

5.5 Conclusion ...... 116 5.6 Literature Cited ...... 116

Chapter 6 Conclusion ...... 125

6.1 Revisiting Research Hypothesis ...... 125 6.2 Body Size ...... 126 6.3 Peak Bone Mass and Density ...... 126

6.3.1 South African Apartheid and Bone Health ...... 128

6.4 Histological Signatures of Menopause ...... 129 6.5 Final Remarks ...... 130 6.6 Literature Cited ...... 130

Appendix A: List of Abbreviations by Order of Appearance ...... 134

Appendix B: Full Research Sample ...... 136

Appendix C: BSE-SEM MATLAB Programs with Sample Images for Each Processing Step . 143

viii

List of Tables

Table 1: A Simplified Model of Bone Development, Maintenance, and Loss in Women ...... 1

Table 2: Descriptive statistics for the study sample and independent samples t-test between men and women. Significant values are in bold...... 47

Table 3: Partial correlations controlling for age in men, women, and the pooled sample. Significant values are in bold...... 50

Table 4 Hierarchical multiple regression output for OPD and On.Ar ...... 52

Table 5: Sample structure according to sex, age, and population group...... 66

Table 6: Descriptive statistics (mean ± SD or median (interquartile range)) for each histomorphometric variable are given by age cohort for men and women...... 71

Table 7: Observed outcomes at <35 years (mean ± SD or median (interquartile range)), and predicted outcomes at 80 years (estimated mean (95% prediction interval))...... 72

Table 8: Cortical and trabecular bone parameters (Descriptions modified from Vajda et al., 1995; Bloebaum et al., 1997; Dempster et al., 2013; Hunter and Agnew, 2016) ...... 100

Table 9: Descriptive statistics for cortical parameters for the study sample by sex and age cohort (mean ± SD if normally distributed, median (IQR) if non-normally distributed) ...... 101

Table 10: Descriptive statistics for trabecular parameters for the study sample by sex and age cohort (mean ± SD if normally distributed, median (IQR) if non-normally distributed) ...... 102

Table 11: Observed outcomes at 20-35 years (mean ± SD or median (interquartile range)), and predicted outcomes at 80 years (estimated mean (95% prediction interval))...... 103

List of Figures

Figure 1: (Left) Racial-demographic map of South Africa in 1970. Racial concentrations of 30% or more are given by magisterial district. The approximate location of the Kirsten Skeletal Collection is indicated by the arrow icon. (Right) Map of Africa, with South Africa colored in purple. (Figure modified from CIA, 1979)...... 4

Figure 2: (Left) Example of a BSE-SEM image showing a region of cortical bone with a few osteons of low mineralization, including an active BMU, surrounded by more highly mineralized lamellar bone. (Right) BSE-SEM image showing a region of high bone turnover in which there are many osteons of low mineralization...... 14

Figure 3: Light microscopy (top left) and back-scattered scanning electron microscopy (top right) image pair used to create a “relational” composite image (bottom), illustrating collagen fiber orientation within poorly mineralized bone. (Image from Goldman et al., 2005)...... 14

Figure 4: Cluster analysis from the Southern African Human Genome Programme (SAHGP) showing the average ancestral composition of various South African population groups. The populations included in the study are Sotho (SOT), Xhosa (XHS), Zulu from Soweto (ZUS), Coloured (COL), and the admixed Xhosa individual from South Africa (XHD). Additional populations used in this analysis are Baganda from Uganda (BAG), Bengali from Bangladesh (BEB), Utah Residents (CEPH) with Northern and Western European Ancestry (CEU), COL from Wellington (CWC), Northern and Central Khoesan including Ju/Õhoansi, G|ui, and G||ana and !Xuun (KSA), Southern Khoesan including Khwe, Karretjie, Nama and ≠Khomani (KSB), Luhya in Webuye, Kenya (LWK), South Eastern Bantu speakers from Schlebusch et al. 2012 (SEB2); Black South Africans from Soweto based on May et al. 2013 (SEB3), Malay from Singapore (SSMP); southwestern Bantu-speakers (SWB); Yoruba in Ibadan, Nigeria (YRI); Zulu from South Africa (ZUL) (Image from Choundhury et al., 2017)...... 27

Figure 5: (a) Complete rib cross-section under linearly polarized light (LPL) demonstrating Tt.Ar, Es.Ar, and On.Ar measurements in red. (b) Inset demonstrating data collection methods. Areas highlighted in blue indicate an intact osteon and areas highlighted in green indicate a fragmentary osteon (together forming OPD)...... 48

Figure 6: Scatterplot showing the relationship between OPD and On.Ar in men (orange circles) and women (red triangles) ...... 51

Figure 7: Scatterplot showing the relationship between Ct.Ar and On.Ar (Y-axis 1, blue x’s) and the relationship between Ct.Ar and OPD (Y-axis 2, green +’s) with smoothed curve fitted by LOESS (Jacoby, 2000)...... 55

Figure 8a-b: Sample structure according to sex, age, and population group. SAC in blue, SAB in green, and SAW in beige...... 67

x lines represent the 95% confidence interval of the prediction. SAC in blue circles, SAB in circles diamonds, and SAW in beige circles...... 76

Figure 10a-h: Trabecular bone parameters. Predicted age-related changes in bone volume to total volume ratio (BV/TV) (a-b), trabecular thickness (Tb.Th) (c-d), trabecular number (Tb.N) (e-f), and trabecular spacing (Tb.Sp) (g-h) in men and women. The solid line represents the fitted mean from the regression model, and the dashed lines represent the 95% confidence interval of the prediction. SAC in blue circles, SAB in green circles, and SAW in beige circles...... 80

Figure 11: Sample structure by age and sex...... 96

Figure 12: BSE-SEM photomontage of a transverse rib cross-section from a 50 year old South African Coloured (SAC) woman. The inferior rib surface is to the left in this image, while the superior rib surface is to the right. This density-dependent image renders bone in various grey levels, while non-bone spaces are assigned to black. The detailed image on the right shows several more highly mineralized Haversian systems (osteons), each with a central vascular canal. The largest pore represents a resorption space, bounded by scalloped reversal line. The longitudinal cracks are a product of tissue processing. The very small black spaces are osteocyte lacunae...... 98

Figure 13a-f: Cortical porosity parameters. Predicted age-related changes in cortical bone area (Ct.B.Ar) (a-b), mean pore area (Po.Ar) (c-d), mean pore diameter (Po.Dm) (e-f), for men and women. The solid line represents the fitted mean from the regression model, and the dashed lines represent the 95% confidence interval of the prediction. Men are represented by open circles, and women are represented by open triangles...... 106

Figure 14g-l: Cortical porosity parameters (cont.). Predicted age-related changes in number of pores (N.Po) (g-h), pore density (Po.Dn) (i-j), and cortical porosity (Ct.Po) (k-l) for men and women. The solid line represents the fitted mean from the regression model, and the dashed lines represent the 95% confidence interval of the prediction. Men are represented by open circles, and women are represented by open triangles...... 107

Figure 15a-f: Cortical osteocyte lacunar properties. Predicted age-related changes in cortical osteocyte lacunar number (Ct.Ot.Lc.N) (a-b), cortical osteocyte lacunar area (Ct.Ot.Lc.N) (c-d), and cortical osteocyte lacunar density (Ct.Ot.Lc.N) (e-f) for men and women. The solid line represents the fitted mean from the regression model, and the dashed lines represent the 95% confidence interval of the prediction. Men are represented by open circles, and women are represented by open triangles...... 108

Figure 16a-f: Trabecular osteocyte lacunar properties. Predicted age-related changes in trabecular osteocyte lacunar number (Tb.Ot.Lc.N) (a-b), trabecular osteocyte lacunar area (Tb.Ot.Lc.N) (c-d), and trabecular osteocyte lacunar density (Tb.Ot.Lc.N) (e-f) for men and women. The solid line represents the fitted mean from the regression model, and the dashed lines represent the 95% confidence interval of the prediction. Men are represented by open circles, and women are represented by open triangles...... 111

Figure 17a-d: Cortical and trabecular bone mineralization. Predicted age-related changes in cortical WMGL (a-b), and trabecular WMGL (c-d) for men and women. The solid line represents the fitted mean from the regression model, and the dashed lines represent the 95% confidence xi interval of the prediction. Men are represented by open circles, and women are represented by open triangles...... 114

xii

List of Appendices

Appendix A: List of Abbreviations by Order of Appearance ...... 134

Appendix B: Full Research Sample ...... 136

Appendix C: BSE-SEM MATLAB Programs with Sample Images for Each Processing Step . 143

xiii

List of Appendix Figures

Figure A1: Original BSE-SEM photomontage of transverse rib cross-section ...... 143

Figure A2: (Left) BSE-SEM photomontage of transverse rib cross-section with trabecular bone removed. (Right) Histogram of cortical bone grey levels...... 147

Figure A3: (Left) BSE-SEM photomontage of a transverse rib cross-section showing the manually traced trabecular bone compartment. (Right) Histogram of trabecular bone grey levels...... 148

Figure A4: Binary BSE-SEM photomontage of transverse rib cross-section with trabecular bone removed...... 149

Figure A5: Binary BSE-SEM photomontage of transverse rib cross-section with trabecular bone and osteocyte lacunae removed...... 150

Figure A6: Binary BSE-SEM photomontage of a transverse rib cross-section showing the manually traced trabecular bone compartment with osteocyte lacunae removed...... 151

Figure A7: Binary BSE-SEM photomontage of transverse rib cross-section with filled medullary cavity...... 151

Figure A8: Binary BSE-SEM photomontage of transverse rib cross-section with filled medullary cavity and cracks removed...... 153

Figure A9: Binary BSE-SEM photomontage of transverse rib cross-section with all voids completely filled for calculation of total bone area...... 154

xiv 1

Chapter 1 Introduction

The formative and resorptive activities of bone cells in modeling and remodeling determine bone mass and density. Throughout the human lifespan, bone tissues are modified by these processes in order to accommodate physiological and mechanical demands (Table 1). During major life history events such as puberty and menopause, bone responds to changes in gonadal sex steroid secretion by adjusting some of its structural (i.e. geometric size and shape) and material (i.e. microstructural organization and composition) properties (Balasch, 2003; Clarke and Khosla, 2010; Khosla and Pacifici, 2013; Walsh and Eastell, 2013). Accordingly, sex differences in the timing and extent of bone turnover are expected and should occur in predictable ways. Most of our knowledge of bone mass and density comes from clinical measures of contemporary western populations, using methods such as dual-energy x-ray absorptiometry (DXA) or quantitative computed tomography (QCT). A histological approach furthers our understanding of age- and sex-related variation because it provides vital information on bone quality: the suite of material properties that determines bone strength and fracture risk (Fritton and Schaffler, 2008). Studies of bone mass and quality are both critical for monitoring developmental and senescent changes throughout the human lifespan.

Table 1: A Simplified Model of Bone Development, Maintenance, and Loss in Women Table 1. A Simplified Model of Bone Development, Maintenance, and Loss Development Maintenance The Osteoporoses Age 0-30 yrs 31-54 yrs 55+ yrs Sequence BF→BR BR→BF BR→BF Activity BF>BR BF=BR BF

The identification of peak bone mass (i.e. the cessation of modeling and skeletal growth), and histological signatures of menopause, have important implications for reconstructing and interpreting the life histories of past peoples (Bogin and Smith, 1996; Roksandic and Armstrong, 2011; Agarwal, 2016). We can increase our knowledge of these universal conditions within and across populations, but more broadly, we can build an understanding of secular changes throughout human history (Smith, 1992; Bogin and Smith, 2000; Key, 2000; Nelson et al., 2003, 2015; Schwartz, 2012; Madimenos, 2015). Although teeth are more commonly used to infer life

2 history in archaeological or paleontological specimens (Robson and Wood, 2008; Guatelli- Steinberg, 2009; Kelley and Schwartz, 2012; Smith, 2013; De Castro et al., 2015), bone is remodeled throughout the entire lifespan, providing the most recent record of tissue modification. As dental tissues either lack or have limited reparative capabilities after they are fully formed, bone provides the only real means of assessing adult life phases and transitions in the past.

To assess how bone responds to changing physiological demands throughout the adult lifespan, this research employs both linearly polarized light (LPLM) and backscattered scanning electron microscopy (BSE-SEM). These methods can be used to generate high-resolution photomontages of full bone cross-sections, in which the cortical and trabecular compartments are easily partitioned to generate measures of bone mass. Coupled together, they allow individual remodeling events (i.e. resorption spaces and secondary osteons), osteocyte lacunar properties, and average tissue mineralization to be quantified in adjacent bone samples (Goldman et al., 2000, 2005). By exploring multiple aspects of bone tissue organization, this approach permits the detection of biologically meaningful variables. Based on previous work, it is expected that women will mature faster, but will generally display poorer indices of bone mass and quality throughout the entire adult lifespan (see Chapter 2).

In addition to appropriate imaging methods and biologically significant variables, the success of histological studies depends on a large study sample with sufficient intra-population variation and statistical power. Most skeletal research collections are heavily male-biased and largely consist of geriatric skeletons. The Kirsten Skeletal Collection (Stellenbosch University, South Africa) provides a unique opportunity to conduct this research because of the high proportion of women who died at different ages throughout the adult lifespan (Appendix A). However, population-level variation in the timing of major life events or unexpected sex-related differences may be driven by biosocial factors (Crews and Gerber, 2003). It is important to situate research samples into appropriate historical context in order to appreciate the human-environment interaction, and its impact on human skeletal biology. Population-level differences have often been inappropriately reduced to “racial” or genetic classifications that are completely divorced from their underlying etiology (Burgard, 2002). Multifaceted approaches to studying human health are needed.

The Kirsten Skeletal Collection is cadaver-derived, and largely comprised of non-white South Africans who lied and died during the apartheid era (1948-1994 AD). Major sources of body donations include unclaimed bodies from teaching hospitals and government mortuaries in the Cape Town Metropole, suggesting low socio-economic status and poor living conditions for most individuals in the study sample (Labuschagne and Mathey, 2000; Alblas et al., 2018). Under the Population Registration Act of 1950, individuals were officially classified as South African Black (SAB), South African White (SAW), or South African Coloured (SAC). Non- whites were forced to live in peripheral areas with poor infrastructure, inadequate water supply and waste removal services (Figure 1). Community and healthcare resources were allocated along racial lines, creating disparities in the distribution of health and disease among South African population groups (Andersson and Marks, 1988; Nightingale et al., 1990; Turton and Chalmers, 1990; van Rensburg and Benatar, 1993; Bertens et al., 2012). Further, racial discrimination and segregation rendered non-white groups more susceptible to dietary deficiencies and substance abuse problems (Wisner, 1989; London, 1999; Mager, 2004; Parry et al., 2005; Wechsberg et al., 2008; Naude et al., 2012; Myers et al., 2013; Gossage et al., 2014).

Environmental conditions during growth may influence the onset of developmental events, and contribute to structural and material differences manifest in adult bone (Pearson and Lieberman, 2004). However, epidemiological studies suggest general similarities in the timing of major life history events among the South African population groups. In 1977, average menarcheal age was 13.9 years for the SAB and 13.1 years for the SAW (Jones et al., 2009). While slight ethnic differences in the average age at menopause have been reported, it is around 48.9 years in SAB women, which does not differ significantly from values reported for SAW women (Walker et al., 1984). It is probable that adverse living conditions had less of an impact on life history, and more of an impact on other aspects of bone health. Historically, men have been the most vulnerable given their high representation in clinical osteoporosis cases under the apartheid regime (Grusin and Samuel, 1957; Seftel et al., 1966; Lynch et al., 1967, 1970; Wapnick et al., 1971).

The mid-thoracic rib is an ideal location for assessing human life history and systemic changes to cortical and trabecular bone tissues. The ribs are constrained by the biomechanics of respiration, and receive fewer weight-bearing loads compared to the long bones of the upper and lower limb (Pfeiffer et al., 2006; Cho and Stout, 2011). Microstructural differences between analogous sampling locations also appears to be minimal (Crowder and Rosella, 2007). However, the load complexity category of the rib is still largely unknown (Skedros, 2012). Before population-level trends can be fully appreciated, it is first necessary to consider underlying patterns of bone tissue organization. While hormonal states influence bone development, maintenance, and loss—it is the adaptive response to mechanical loads that results in normalized bone architecture (Martin et al., 1998). Bone mass, density, and quality may be influenced by body size as it relates to biomechanical stress and strain. By revealing any significant relationships between the structural and material properties of bone and measures of human body size, we are able to comment on the utility of these variables in biomechanical and life history/aging research.

1.1 Research Hypotheses

To reach these goals, the following hypotheses will be tested:

1. H0: Measures of human body size and bone mass will not be positively correlated with cortical bone microstructure.

HA: Measure of human body size and bone mass will be positively correlated with cortical bone microstructure.

2. H0: The structural and material properties of mid-thoracic ribs will not vary between males and females in early adulthood (i.e. during the development of peak bone mass).

HA1: Cross-sectional geometry, the composition of microstructural features, and average levels of bone tissue mineralization will differ in young adult men and women, reflecting ontogenetic factors. Women will mature faster, but generally display poorer indices of bone mass and quality.

HA2: The expected biological relationships defy expectation, suggesting that biosocial factors related to the South African apartheid political system likely influenced the variables of interest.

3. H0: The structural and material properties of mid-thoracic ribs will not vary between males and females following the predicted age of women’s menopause (50+ years).

HA1: The structural and material properties of mid-thoracic ribs will vary between males and females following the predicted age of women’s menopause (50+ years). Post- menopausal women will exhibit less bone, proportionately more intracortical remodeling, and lower mineralization densities than their male contemporaries.

HA2: The expected biological relationships defy expectation, suggesting that biosocial factors related to the South African apartheid political system likely influenced the variables of interest.

1.2 Organization of the Thesis

The next chapter provides a background on bone remodeling and a summary of reported changes to cortical and trabecular bone tissues throughout the human lifespan. I describe how bone mass and quality are measured and discuss the limitations of our current imaging technologies. I also review histological studies using the rib, elaborating on age- and sex-related patterns when possible. Chapter Three introduces the study sample and provides a brief introduction to life under South African apartheid. I discuss the current climate of bone health research in South Africa, and what is known about bone mass and quality in past South African population groups. Chapter Four tests my first hypothesis, and considers whether bone size proxies are correlated with histomorphometric variation. Chapter Five tests hypotheses two and three using transmitted light microscopy, while Chapter Six does the same using backscattered scanning electron microscopy. Chapter Seven contains my final conclusions, and suggested directions for future research.

1.3 Literature Cited

Agarwal SC. 2016. Bone morphologies and histories: Life course approaches in bioarchaeology. Am J Phys Anthropol 159:S130–S149. Alblas A, Greyling LM, Geldenhuys EM. 2018. Composition of the Kirsten Skeletal Collection at Stellenbosch University. S Afr J Sci 114:1–6. Andersson N, Marks S. 1988. Apartheid and Health in the 1980s. Soc Sci Med 27:667–681. Balasch J. 2003. Sex steroids and bone: Current perspectives. Hum Reprod Update 9:207–222. Bertens MGBC, Ulijaszek S, Kozieł S, Henneberg M. 2012. Late childhood and adolescence growth sensitivity to political transition: the case of South African Cape coloured schoolchildren during and post-apartheid. Anthropol Rev [Internet] 75:19–31. Available from: http://www.degruyter.com/view/j/anre.2012.75.issue-1/v10044-012-0002-6/v10044- 012-0002-6.xml Bogin B, Smith BH. 1996. Evolution of the human life cycle. Am J Hum Biol 8:703–716. Bogin B, Smith BH. 2000. Evolution of the Human Life Cyle. In: Stinson S, Bogin B, Huss- Ashmore O’Rourke R, editors. Human Biology: An Evolutionary and Biocultural Perspective. NY: Elsevier GmbH. p 377–424. Available from: http://onlinelibrary.wiley.com/doi/10.1002/evan.1360010406/abstract De Castro JMB, Modesto-Mata M, Martinón-Torres M. 2015. Brains, teeth and life histories in hominins: A review. J Anthropol Sci 93:21–42. Cho H, Stout SD. 2011. Age-associated bone loss and intraskeletal variability in the Imperial Romans. J Anthropol Sci [Internet] 89:109–25. Available from: http://www.ncbi.nlm.nih.gov/pubmed/21368344 Clarke BL, Khosla S. 2010. Female reproductive system and bone. Arch Biochem Biophys 503:118–28. Crews DE, Gerber LM. 2003. Reconstructing life history of hominids and humans. Coll

Antropol 27:7–22. Crowder C, Rosella L. 2007. Assessment of intra- and intercostal variation in rib histomorphometry: its impact on evidentiary examination. J Forensic Sci 52:271–6. Fritton JC, Schaffler MB. 2008. Bone Quality. In: Marcus R, Feldman D, Nelson D, Rosen C, editors. Osteoporosis. 3rd Editio. Burlington, MA: Elsevier Academic Press. p 625–641. Goldman HM, Blayvas A, Boyde A, Howell PGT, Clement JG, Bromage TG. 2000. Correlative Light and Backscattered Electron Microscopy of Bone — Part II: Automated Image Analysis. Scanning 22:337–344. Goldman HM, Thomas CDL, Clement JG, Bromage TG. 2005. Relationships among microstructural properties of bone at the human midshaft femur. J Anat 206:127–139. Gossage PJ, Snell CL, Parry CDH, Marais AS, Barnard R, de Vries M, Blankenship J, Seedat S, Hasken JM, May PA. 2014. Alcohol use, working conditions, job benefits, and the legacy of the “dop” system among farm workers in the Western Cape Province, South Africa: Hope despite high levels of risky drinking. Int J Environ Res Public Health 11:7406–7424. Grusin H, Samuel MD. 1957. A syndrome of osteoporosis in Africans and its relationship to scurvy. Am J Clin Nutr 5:644–650. Guatelli-Steinberg D. 2009. Recent studies of dental development in Neandertals: Implications for Neandertal life histories. Evol Anthropol 18:9–20. Jones LL, Griffiths PL, Norris S a., Pettifor JM, Cameron N. 2009. Age at menarche and the evidence for a positive secular trend in urban South Africa. Am J Hum Biol 21:130–132. Kelley J, Schwartz GT. 2012. Life-history inference in the early hominins Australopithecus and Paranthropus. Int J Primatol 33:1332–1363. Key CA. 2000. The evolution of human life history. World Archaeol 31:329–350. Khosla S, Pacifici R. 2013. Estrogen deficiency, postmenopausal osteoporosis, and age-related bone loss. In: Osteoporosis. Fourth Edi. Elsevier Inc. p 1113–1136. Labuschagne BCJ, Mathey B. 2000. Cadaver profile at University of Stellenbosch Medical School, South Africa, 1956 – 1996. Clin Anat 13:88–93. London L. 1999. The “dop” system, alcohol abuse and social control amongst farm workers in South Africa: a public health challenge. Soc Sci Med 48:1407–1414. Lynch SR, Berelowitz I, Seftel HC, Miller GB, Krawitz P, Bothwell TH. 1967. Osteoporosis in Johannesburg Bantu Males: Its Relationship to Siderosis and Ascorbic Acid Deficiency. Am J Clin Nutr 20:799–807. Lynch SR, Seftel HC, Wapnick AA, Charlton RW, Bothwell TH. 1970. Some aspects of calcium metabolism in normal and osteoporotic Bantu subjects with special reference to the effects of iron overload and ascorbic acid depletion. S Afr J Med Sci 35:45–56. Madimenos FC. 2015. An evolutionary and life-history perspective on osteoporosis. Annu Rev Anthropol 44:186–206. Mager A. 2004. “White liquor hits black livers”: Meanings of excessive liquor consumption in South Africa in the second half of the twentieth century. Soc Sci Med 59:735–751. Martin RB, Burr DB, Sharkley NA. 1998. Skeletal Tissue Mechanics. 1st Edition. New York: Springer Science and Business Media. Myers B, Kline TL, Browne FA, Carney T, Parry C, Johnson K, Wechsberg WM. 2013. Ethnic differences in alcohol and drug use and related sexual risks for HIV among vulnerable women in Cape Town, South Africa: implications for interventions. BMC Public Health [Internet] 13:174. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3598514&tool=pmcentrez&rend ertype=abstract%5Cnhttp://biomedcentral.com/1471-2458/13/174

Naude CE, Carey PD, Laubscher R, Fein G, Senekal M. 2012. Vitamin D and calcium status in South African adolescents with alcohol use disorders. Nutrients [Internet] 4:1076–1094. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3448088&tool=pmcentrez&ren dertype=abstract Nelson DA, Agarwal SC, Darga LL. 2015. Bone health from an evolutionary perspective: development in early human populations. In: Holick MF, Nieves JW, editors. Nutrition and Bone Health, Nutrition and Health. New York: Springer Science+Business Media, LLC. p 1–718. Nelson DA, Sauer NJ, Agarwal SC. 2003. Evolutionary Aspects of Bone Health. Clin Rev Bone Miner Metab 1:169–179. Nightingale EO, Hannibal K, Geiger J, Hartmann L, Lawerence R, Spurlock J. 1990. Apartheid medicine. JAMA J … [Internet] 264:2097–2102. Available from: http://jama.ama- assn.org/content/264/16/2097.short Parry CD, Pluddemann A, Steyn K, Bradshaw D, Norman R, Laubscher R. 2005. Alcohol Use in South Africa: Findings from the First Demographic and Health Survey (1998)*. J Stud Alcohol 66:91–97. Pearson OM, Lieberman DE. 2004. The aging of Wolff’s “law”: ontogeny and responses to mechanical loading in cortical bone. Am J Phys Anthropol Suppl 39:63–99. Pfeiffer S, Crowder C, Harrington L, Brown M. 2006. Secondary osteon and Haversian canal dimensions as behavioral indicators. Am J Phys Anthropol 131:460–468. van Rensburg HCJ, Benatar SR. 1993. The legacy of apartheid in health and health care. South African J Sociol 24:99–111. Robson SL, Wood B. 2008. Hominin life history: Reconstruction and evolution. J Anat 212:394– 425. Roksandic M, Armstrong SD. 2011. Using the life history model to set the stage(s) of growth and senescence in bioarchaeology and paleodemography. Am J Phys Anthropol 145:337– 347. Schwartz GT. 2012. Growth, development, and life history throughout the evolution of Homo. Curr Anthropol [Internet] 53:S395–S408. Available from: http://www.journals.uchicago.edu/doi/10.1086/667591 Seftel HC, Malkin C, Schmaman A, Abrahams C, Lynch SR, Charlton RW, Bothwell TH. 1966. Osteoporosis, scurvy, and siderosis in Johannesburg Bantu. Br Med J 1:642–646. Skedros JG. 2012. Interpreting load history in limb-bone diaphysis: important considerations and their biomechanical foundations. In: Crowder CM, Stout SD, editors. Bone Histology: an anthropological perspective. 1st Editio. Boca Raton: CRC Press. Smith BH. 1992. Life history and the evolution of human maturation. Evol Anthropol [Internet]:134–142. Available from: http://onlinelibrary.wiley.com/doi/10.1002/evan.1360010406/abstract Smith TM. 2013. Teeth and Human Life-History Evolution. Annu Rev Anthropol [Internet] 42:191–208. Available from: http://www.annualreviews.org/doi/10.1146/annurev-anthro- 092412-155550 Turton RW, Chalmers BE. 1990. Apartheid, stress and illness: The demographic context of distress reported by South African Africans. Soc Sci Med 31:1191–1200. Walker ARP, Walker BF, Ncongwane J, Human ENT. 1984. Age of menopause in black women in South Africa. Br J Obstet Gynaecol 91:797–801. Walsh JS, Eastell R. 2013. Role of estrogen in the age-related decline in bone microstructure. J

Clin Endocrinol Metab 98:519–21. Wapnick AA, Lynch SR, Seftel HC, Charlton RW, Jowsey J. 1971. The effect of siderosis and ascorbic acid depletion on bone metabolism, with special reference to osteoporosis in the Bantu. Br J Nutr 25:367–376. Wechsberg WM, Luseno WK, Karg RS, Young S, Rodman N, Myers B, Parry CD. 2008. Alcohol, cannabis, and methamphetamine use and other risk behaviours among Black and Coloured South African women: A small randomized trial in the Western Cape. Int J Drug Policy 19:130–139. Wisner B. 1989. Commodity relations and nutrition under apartheid: a note on South Africa. Soc Sci Med [Internet] 28:441–446. Available from: http://www.ncbi.nlm.nih.gov/pubmed/2648598

Chapter 2 Literature Review 2.1 Modeling

Bone modeling optimizes bone geometry and mass by conforming to the mechanical and physiological demands present throughout ontogeny (Maggiano, 2012). While its processes set the bone template, modeling essentially ceases upon skeletal maturation (Robling et al., 2006), and thus, it is not the focus of the proposed research.

2.2 Remodeling and Bone Cells

Remodeling is the process of bone maintenance and repair; it is continuous throughout life. Remodeling serves three major functions: (1) adjusting nutrient concentrations, (2) protecting against microfracture formation/propagation, and (3) replacing damaged tissue (Skedros, 2012). The process is executed by a complex of cells collectively known as a basic multicellular unit (BMU), which includes osteoclasts, osteoblasts, and osteocytes. These are the same cell types operational during modeling, but their activity is said to be coupled (or tethered) in remodeling. Remodeling is described in terms of phases, with an active BMU following a sequence of activation, resorption, and formation (ARF; Martin et al., 1998). Other authors have extended this process to include reversal and quiescent phases (Ortner, 2003; Parfitt, 2003). The entire remodeling process takes approximately 120 days to complete in cortical bone, culminating in the deposition of bone structural units (BSUs; Robling and Stout, 2008). In cortical bone these structures are called secondary osteons, or Haversian systems. In cancellous bone they are referred to as hemiosteons (Parfitt, 1994). Cortical bone remodeling is described below.

The BMU maintains a distinctive three-dimensional structure as it travels longitudinally through the diaphysis of long bones, roughly parallel to the main loading direction (Hert et al., 1994; Parfitt, 1994; Petrtyl et al., 1996). In ribs, they are oriented parallel to the length of the bone shaft. Osteoclasts are found along the leading edge(s) of this structure, known as the resorptive front or cutting cone. Their recruitment and activation is directed by cytokine receptor activator of nuclear factor kappa B ligand (RANKL; Boyce and Xing, 2007). Osteoclasts secrete collagenase, lysosomal hydrolases, and other enzymes, creating an acidic environment that degrades the organic portion of bone and mobilizes calcium and phosphorus minerals (Mescher,

2010). Osteoclasts excavate a tunnel that typically reaches 250-300µm in diameter, delimitating the cross-sectional size of the BSU that will eventually fill in its wake (Robling and Stout, 2008). The tunnel is quickly invaded by a blood capillaries and osteoprogenitor (bone-lining) cells, both originating from either the periosteum or endosteum (Mescher, 2010). These cells polish the scalloped edges of the resorption space and secrete a hyper-calcified layer of matrix that forms the cement or reversal line (Robling and Stout, 2008). Bone-lining cells subsequently differentiate into osteoblasts, which constitute the closing cone of the BMU.

Osteoblasts adhere to the reversal line and lay down layers of pre-osseous tissue in the resorption space until the diameter of the invading vessel is closely approximated (Robling et al., 2006). They deposit a matrix (osteoid) of type I collagen, glycoproteins, proteoglycans, and matrix vesicles (Mescher, 2010). Newly deposited bone undergoes a time-dependent increase in mineral density as inorganic material primarily comprised of calcium hydroxyapatite crystals

(Ca10(PO4)6(OH)2) fills the matrix volume. As a result, there is an inverse relationship between bone mineral density (BMD) and bone remodeling activity. During periods of active remodeling, there will be more “young” bone matrix with a lower mineral density than “old” bone matrix (Rauch and Schoenau, 2001). Osteoblasts either become locked in the solidifying matrix and transform into osteocytes (osteocytogenesis), flatten out on the new bone surface and differentiate into bone lining cells, or undergo apoptosis (Franz-Odendaal et al., 2006; Dallas and Bonewald, 2011; Xiong and Brien, 2012).

Osteocytes maintain bone tissue by sensing mechanical strain (mechanosensation), mediating mineral metabolism, and directing osteoclast and osteoblast activity (Bonewald, 2011; Klein- Nulend et al., 2013). Osteocytes rest in oval-shaped cavities, known as lacunae, and have radiating cytoplasmic processes called canaliculi. The canaliculi of adjacent osteocytes extend out into the matrix and make contact through gap junctions, enabling cellular communication, nutrient exchange, and waste removal (Mescher, 2010). It has been proposed that these dendritic extensions have the ability to adjust their position within the bone tissue through peri-lacunar remodeling (Bonewald and Johnson, 2008).

Secondary osteons are characterized by centripetally deposited lamellae that surround a centrally located Haversian canal, a conduit allowing the passage of blood, lymph, and nerve fibers (Robling and Stout, 2008). The canal is bordered by approximately 4 to 10 concentric lamellae,

12 each 3 to 7µm thick (Mescher, 2010). Secondary osteons can be differentiated from primary osteons by their scalloped reversal lines, the borders that mark the shift from osteoclastic to osteoblastic activity. Secondary osteons connect to each other via Volkmann’s canals, which are smaller than Haversian canals and run at oblique angles. Together, these vessels create a network for circulating blood and nutrients throughout cortical bone.

2.3 Measures of Bone Mass and Density

In clinical research, bone mass and density have been assessed using a variety of imaging techniques. Bone mineral content (BMC, g) and areal bone mineral density (aBMD, g/cm2) are most commonly measured by dual energy x-ray absorptiometry (DXA). Because DXA is highly influenced by bone size, results are normally described in terms of bone mass rather than tissue mineralization (Macdonald et al., 2013). Further, DXA cannot separate the more metabolically active trabecular bone from the more structurally important cortical bone. Peripheral or central quantitative computed tomography (QCT) can be used to calculate volumetric bone mineral density (vBMD, g/cm3), cross-sectional geometry, and the relative proportions of cortical and trabecular bone. While these two noninvasive techniques have provided important insights regarding bone mass attainment and loss, they cannot assess bone density at the material level because image resolution is insufficient to quantify cortical porosity (Petit et al., 2005).

High-resolution peripheral QCT (HR-pQCT) partially overcomes this limitation, and provides additional information regarding the organization of the trabecular bone compartment (e.g. number and thickness of trabeculae; Buie et al., 2007). HR-pQCT is sometimes used in conjunction with finite element analysis (FEA), to estimate bone material properties such as strength and stiffness (Allen and Krohn, 2013). However, the pixel size of 82µm is not sufficiently small to identify many of the Haversian canals within cortical bone, so this method underestimates true porosity (Walsh and Eastell, 2013). Micro-CT scanners typically produce scans with pixel sizes in the range of 1-30µm, and given that many have convenient desktop set- ups, have become a popular tool for ex vivo investigations of both cortical and trabecular bone microstructure. In vivo micro-CT imaging resolution is comparable to HR-pQCT, and applications are restricted to animal studies because of the small field of view and high radiation dose (Campbell and Sophocleous, 2014).

Less frequently, quantitative backscattered scanning electron microscopy (BSE-SEM) has been utilized to quantify the bone mineral density distribution of iliac crest biopsies (Boyde et al., 1995; Fratzl-Zelman et al., 2009; Koehne et al., 2014; Misof et al., 2014). BSE-SEM provides information on bone mineral content independent of any porosities in the sample, accounting for intracortical remodeling variability not captured by other methods. This method generates high resolution planar images, in which it is also possible to examine aspects of bone quality. In addition to assessing bone tissue mineralization, previous research has focused on enumerating and measuring osteocyte lacunae and intracortical porosities (Boyce and Bloebaum, 1993; Zebaze et al., 2010; Misof et al., 2014; Bromage et al., 2016).

2.3.1 Back-Scattered Scanning Electron Microscopy (BSE-SEM)

BSE-SEM is an imaging method that quantifies the intensity of electrons backscattered from an extremely thin surface layer (<1.5 µm in thickness). Since the z-plane is so small, false contrasts due to variation in section thickness are not of issue (Reid and Boyde, 1987). The signal obtained from a selected bone area is proportional to the weight fraction of calcium locally present in the bone tissue (Roschger et al., 1998). The degree of mineralization relates to the time since bone formation, and strongly relates to compressive bone strength (Follet et al., 2004). The most recently deposited bone is the least mineralized and appears darkest in image outputs (i.e. lowest grey levels). Intact osteons are less mineralized than surrounding interstitial or primary lamellar bone, which appear brighter in comparison (Reid and Boyde, 1987). Regions of high bone turnover will tend to have more osteons of lower mineralization, while regions with low bone turnover will tend to have concentrated areas with more highly mineralized bone (Goldman, 2001; Figure 2). The rate of bone mineralization is also dependent upon mineral availability, and can be decreased by nutritional deficiency or metabolic disease. The range of conditions that affect mineral homeostasis include inadequate dietary calcium, high dietary acid load, and increased physiological demand (e.g. pregnancy, lactation, intestinal malabsorption) (Roschger et al., 2008).

Given high machine availability and relatively low imaging costs, it is possible to generate full cross-sectional bone images for large study samples. Image resolution is sufficient to identify cortical and trabecular compartments such that their specific adaptations to targeted physiological conditions and mechanical loading can be appreciated. Further, light microscopy

(LM) images of the same fields of view can be directly compared, revealing significant correlations between the degree of mineralization and patterns of collagen fiber orientation (Goldman et al., 2005; Figure 3). Differences in grey level ranges may reflect absolute differences in the degree of mineralization associated with species (Zeininger et al., 2011), functional load history (Riggs et al., 1993; Goldman et al., 2005; Zeininger et al., 2011), age (Goldman et al., 2003), or disease states (Jones et al., 1999; Roschger et al., 2007, 2008).

2.4 Overview of Major Life History Events

The female reproductive system plays a major role in regulating the acquisition, maintenance, and loss of bone mass throughout the lifespan. Skeletal size and density are similar in males and females until early adolescence (Bonjour et al., 2012), wherein longitudinal and radial growth largely correspond to the growth hormone—insulin-like growth factor-1 (GH/IGF-1) system (Niu and Rosen, 2005). Sex differences in bone geometry and in mineralization density are mainly attributed to gonadal sex steroid secretions that coincide with the onset of puberty. Other physiological conditions such as pregnancy, lactation, and menopause can present significant challenges to the female skeleton and may contribute to sex-specific variation in remodeling throughout adulthood (Clarke and Khosla, 2010). At menopause, the withdrawal of ovarian sex steroids (i.e. estrogens, androgens, progesterone) prompts increased osteoclastic activity and leads to universal loss of bone mass and density in women. While both sexes demonstrate substantial age-related bone loss in advanced years, men do not experience a spike in bone resorption triggered by the end of the reproductive lifespan (Khosla et al., 2008).

2.4.1 Acquisition of Peak Bone Mass

Peak bone mass and density are important determinants of overall skeletal health, and serve as risk indicators for bone loss and fragility later in life (Heaney et al., 2000). While familial studies indicate that genetic factors account for between 60% and 80% of the variability in peak bone mass (Pocock et al., 1987; Ferrari et al., 1998), adult skeletal dimensions are largely shaped by environmental factors encountered throughout growth and development (Ruff, 2003a; b; Pearson and Lieberman, 2004). Bone metabolism is especially high from the perinatal period to the end of the first postnatal year, and gradually decreases throughout childhood and early adolescence (Bayer, 2014). Correspondingly, epidemiological studies propose that lasting ill-effects on bone health may arise during fetal life or in early infancy. Reduced maternal height, weight, and fat stores, as well as vitamin D deficiency, a history of smoking, and low socio-economic status have all been associated with reduced bone mass in children (Cooper et al., 2006).

During childhood and early adolescence, basal hormones from the GH-IGF-1 axis maintain slow but continuous bone growth until puberty. Skeletal size and mass increases are similar in pre- pubertal boys and girls (Bonjour et al., 2012). Although BSE-SEM research indicates that inter- individual variability in bone mineral density is high during the pre-pubertal growth period, there

16 are no statistically significant sex differences (Fratzl-Zelman et al., 2009). Hormonal differences expressed with the onset of puberty (11 to 13 years in girls and 14 to16 years in boys)1 are responsible for the sexual dimorphism of the adult human skeleton (Clarke and Khosla, 2010). Given that the pre-pubertal growth period lasts approximately 1 to 2 years longer in males than in females, and that GH and IGF-1 largely modulate linear growth (Wang and Seeman, 2008), greater sex differences in bone size and mass are anticipated in the lower limb than in the thorax.

In mid-to-late adolescence, the increased pulsatile secretion of gonadotropin-releasing hormone (GnHr) by the hypothalamus leads to increased serum GH, IGF-1, gonadotropins, and sex steroids (Giustina and Veldhuis, 1998). Luteinizing (LH) and follicle stimulating (FSH) hormones are secreted by gonadotropes in the anterior pituitary, and are responsible for regulating normal menstrual cycles in women (Clarke and Khosla, 2010). Androstenedione and testosterone are mainly derived from testicular secretions in men (Riggs et al., 2008). In women, they are secreted by the adrenal glands, but are also produced by ovarian interstitial and theca cells. These androgens are the most abundant circulating sex steroids in both sexes, but they play a more obvious role in male skeletal homeostasis (Balasch, 2003). Estrogen and progesterone are the main skeletal effectors in women, and are secreted by the ovaries. Estrone (E1) and estradiol (E2), the two main circulating forms of estrogen, can respectively be derived from androstenedione and testosterone (Riggs et al., 2008).

During puberty, increased longitudinal and transverse growth is supported by elevated levels of GH and IGF-1 and the onset of gonadal sex steroid secretion. High levels of GH and IGF-1 are maintained for 3 to 4 years before returning to pre-pubertal levels. Conversely, adult levels of serum sex steroids are achieved during puberty and are maintained until menopause or senility (Giustina and Veldhuis, 1998). Testosterone is believed to stimulate periosteal apposition; thus, most measures of bone size and mass markedly increase in pubertal males (Lobo, 2014). Estrogen inhibits periosteal expansion but promotes endosteal apposition, resulting in comparatively smaller subperiosteal and endosteal areas in females than in males (Wang et al. 2006). Early work by Garn and colleagues (Garn, 1970; Garn et al., 1972) proposed that

1 Due to large variation in the onset and end of the adolescent/pubertal growth spurt, this period can span the age range of 8 to 19 years in girls, and 10 to 22 years in boys (Malina et al., 2004).

17 endosteal apposition was associated with the pubertal estrogen surge, and served to create a calcium store for reproduction.

Peak bone mass attainment occurs approximately 1.4 years earlier in females than in males, with females demonstrating lower average BMC and aBMD values (Ferrari et al., 1998). Approximately 40% of adult skeletal bone mass is determined between 12 and 16 years of (Theintz et al., 1992; Mølgaard et al., 1999), and substantial sex differences are apparent by 14 years (Bailey et al., 1999). Cross-sectional and longitudinal DXA studies indicate that peak bone mass is generally established by early adulthood, but may be achieved as early as 14.8 years in females at weight-bearing sites such as the femoral neck (Baxter-Jones et al., 2011). At most skeletal locations, bone mass does not appear to increase between the third and fifth decade of (Bonjour et al., 1991; Katzman et al., 1991; Theintz et al., 1992; Matkovic et al., 1994; Fournier et al., 1997; Hopper et al., 1998). However, peak bone mass attainment may be delayed if adverse environmental conditions and sever biological stress are experienced throughout ontogeny (See section 2.7.1). There is also some evidence to suggest stronger buffering to extrinsic factors in the female skeleton (Stinson, 1985; Jantz and Jantz, 1999; Ross et al., 2003; Steckel, 2005; Cole et al., 2015; Heinrich, 2015).

Three-dimensional measures of BMD have greatly improved our understanding of the compositional changes that occur during puberty. QCT findings indicate that at some clinically relevant sites, vBMD values are actually higher in females during puberty. Cortical vBMD gradually increases during the 20 months following menarche, but remains relatively unchanged in males of matched maturational status (Loro et al., 2000; Kontulainen et al., 2005, 2006; Wang et al., 2005). More recently, HR-pQCT studies have revealed that males experience a transient period of fragility during the growth spurt of early puberty, and that gains in cortical vBMD are most pronounced during early adulthood in both sexes (Kirmani et al., 2009; Wang et al., 2010; Nishiyama et al., 2012). This research suggests that with advancing maturity, females develop denser and less porous cortices than males do. This may be the result of higher rates of intracortical remodeling associated with prolonged adolescent growth in boys compared with girls (Macdonald et al., 2013). Histological lines of evidence are needed to confirm this.

2.4.2 Menopause and Primary Osteoporosis

Menopause is defined as time subsequent to the last menstrual period, but menopause-related symptoms may occur before the cessation of menses (Clarke and Khosla, 2010). Perimenopausal changes are generally preceded by reduced fecundity, and are a consequence of fluctuating hormone levels. Ovarian aging and accelerated primordial follicle atresia are associated with small elevations in serum follicle stimulating hormone (FSH), as well as reductions in inhibin B and mullerian inhibiting substance (MIS or AMH). FSH levels rise several years before menopause, with substantial increase occurring two years prior (Sowers et al., 2013). FSH levels stabilize to a steady state during the menopause transition, and then decrease in the seventh decade of life (Perrien et al., 2006). Some evidence suggests that increased serum FSH may contribute to bone loss during the peri- and postmenopausal periods by expanding the osteoclast precursor pool (Tung and Iqbal, 2007; Imam et al., 2009).

Estrogen withdrawal leads to an increase in bone resorption that is not compensated by a concurrent increase in bone formation, resulting in net bone loss. Estrogen decline begins roughly two years before the last menstrual period (Randolf et al., 2011), as does androgen status (i.e. androstenedione and testosterone), although the latter cannot be adequately detected at the time of the perimenopause (Horstman et al., 2012). Marked reductions in serum estradiol (E2) and estrone (E1) are coincident with established menopause (Khosla et al., 1997). Accelerated rates of bone loss can first be documented when the menstrual cycle becomes irregular in the perimenopause (Greendale et al., 2012). Although rates of bone loss are site-specific, biochemical markers of bone metabolism indicate that bone resorption increases by 90% after the onset of menopause, whereas bone formation only increases by 45% (Garnero et al., 1996). In the four to eight years following menopause, 20 to 30% of trabecular bone and 5 to 10% of cortical bone is lost. Thereafter, the overall rate of bone loss slows to 1 to 2% per year and comparable amounts of trabecular and cortical bone are resorbed (Riggs et al., 2008).

Bone loss can largely be explained by the well-documented menopausal increase in osteoclastic activity, which produces extensive intracortical porosity through resorption spaces and Haversian canals (Khosla and Pacifici, 2013). Histomorphometric studies also suggest that reduced osteon infilling by osteoblasts increases intracortical porosity (Brockstedt et al., 1993; Parfitt et al., 1995; Power et al., 2012; Goldman et al., 2014). Brockstedt et al. (1993) found significant sex

19 differences in cortical porosity with age, which they attributed to decreased osteon wall thickness (W.Th) in females. Similar findings were reported by Parfitt and colleagues (1995).

While the molecular and cellular mechanisms through which estrogen deficiency leads to bone loss are not fully understood, potential models often consider osteocyte apoptosis and the subsequent decrease in inhibitory signals to osteoclasts (Emerton et al., 2011). Estrogen also appears to have direct effects on osteoclastic activity. For example, studies have demonstrated that osteoclast-specific deletion of ERα reduces osteoclast apoptosis and results in decreased bone mass (Nakamura et al., 2007; Martin-Millan et al., 2010). Estrogen deficiency may also have indirect effects on osteoblasts, which likely contribute to impaired bone formation in the years surrounding menopause (Khosla and Pacifici, 2013).

In western countries, the average age of menopause is 51 to 52 years, and 97% of women are postmenopausal by age 58 (Lobo, 2014). Age at menopause is heritable, with as much as 87% of the variance in timing attributable to genetic factors (de Bruin et al., 2001). While the age of menopause is tightly constrained, behavioral factors may produce some variability. Earlier natural menopause has been associated with less education, low socioeconomic status, nulliparity or having fewer children, never having used oral contraceptives, and smoking (Gold et al., 2001). The effects of smoking seem to be the most consistent, with menopause occurring 1 to 2 years earlier in smokers than in non-smokers (Van Noord et al., 1997; Gold et al., 2001). However, body mass and activity level have not been found to consistently influence the age at menopause (Baker et al., 2013).

2.5 Rib Bone Histomorphometry

Human ribs are particularly well-suited for microscopic analysis given their availability for destructive tissue processing. Rib bones are numerous, seldom used in routine osteological assessments, and easily removed from cadaver breast plates during autopsy. Compared to the long bones of the upper and lower limb, the ribs are subject to less complex biomechanical interactions and fewer of the weight-bearing loads, which produce variability in bone modeling and remodeling behavior (Pfeiffer et al., 2006). Histological research indicates that inter-element variability between analogous sampling locations within adjacent ribs tends to be minimal (Crowder and Rosella, 2007). The entire cortex can also be evaluated thanks to their relatively

20 small cross-sectional areas, thereby avoiding issues of research design for adequate tissue sampling.

Rib bone histology has found a number of important applications in biological anthropology. Analyses of cortical bone microstructure, have been used to infer diet (Brenton & Paine, 2007; Ericksen, 1976; Pfeiffer & King, 1983; Richman, Ortner, & Schulter-Ellis, 1979), study bone metabolism (Agnew and Stout, 2012; Skedros et al., 2013; Eleazer and Jankauskas, 2016), identify pathological conditions (Schultz, 2001; Schultz and Schmidt-Schultz, 2015; de Boer and Van der Merwe, 2016), and are popular in age-at-death estimation research (Kerley, 1965; Kerley and Ubelaker, 1978; Thompson, 1979; Ericksen, 1991; Stout and Paine, 1992; Thomas et al., 2000; Cho et al., 2002; Maat et al., 2006; Han et al., 2009; Pavón et al., 2010; Nor et al., 2013; Lee et al., 2014). The mid-thoracic ribs are favored assessment location because their development is constrained by the biomechanics of respiration, and they thus reflects systemic bone turnover (Cho & Stout, 2011; Pfeiffer et al., 2006). Ribs contain high proportions of medullary tissue, and undergo proportionately more modeling and remodeling during growth and development (Stein and Granik, 1976; Pfeiffer et al., 2006). Because bone turnover is high, the ribs are expected to capture comparatively more recent events in an individual’s metabolic history.

Unfortunately, almost all past histological studies are limited by small sample sizes that are heavily male biased. Most research on the ribs was conducted over a quarter of a century ago (Sedlin et al., 1963a; b; Sedlin, 1964; Epker and Frost, 1965a; b; Takahashi et al., 1965; Landeros and Frost, 1966; Frost and Wu, 1967; Frost, 1969; Wu et al., 1970; Burton et al., 1989), before digital imaging methods were introduced. Comparisons among more recent studies are also difficult because they often do not consider the same variables, and only some report values separated by sex. Although basic biology predicts that it should, histomorphometric research often does not reveal any sex-specific patterns (Stout and Paine, 1992; Stout and Lueck, 1995; Stout et al., 1996; Pratte and Pfeiffer, 1999; Mulhern, 2000). Studies utilizing improved imaging techniques and biologically relevant variables need to be conducted on larger samples that have a high proportion of females.

2.5.1 Histomorphometric Changes in the Rib from Birth through Senescence

The ribs protect vital organs, and provide attachments sites for the respiratory muscles, as well as for some muscles of the back and forelimb. The geometry of the rib cage is complex, and changes over time. During growth, rib bone dimensions increase as the entire thorax expands to achieve adult proportions (Sandoz et al., 2013). The most dramatic geometric changes occur during the first two years of life (Openshaw et al., 1984), and ultimate cross-sectional size is established by 20 to 30 years (Agnew et al., 2015). While absolute measures of bone mass (e.g. total subperiosteal, total cortical and trabecular area) increase until skeletal maturity (Sedlin et al., 1963a; b; Takahashi et al., 1965; Streeter and Stout, 2003; Pfeiffer, 2006; Agnew et al., 2013), bone remodeling and tissue mineralization vary considerably throughout the lifespan. The distribution of microstructural features likely reflects both the biomechanical loading environment and age-related changes.

Remodeling has been observed in utero and in neonates (Burton et al., 1989; Baltadjiev, 1995), but typically begins at about one year of age (Enlow, 1962). Cortical drift, coupled with a high activation frequency of remodeling events (Jowsey, 1960; Lacroix, 1971; Martin et al., 1998), results in medullary expansion and extensive intracortical porosity throughout growth and development (Sedlin et al., 1963a; Epker and Frost, 1965a; b; Frost, 1969; Agnew et al., 2013b). Relatively more remodeling events (i.e. resorption spaces, secondary osteons) occur in the pleural periosteocortex and the cutaneaous endocortex, and large areas of primary lamellar bone are observed in the pleural endocortex (Agnew et al., 2013). Reid and Boyde (1987) report that pleural cortex is more highly mineralized, and that BMDD is highly influenced by the inclusion of densely mineralized endochondral bone and trabeculae formed at the cartilage growth plate. It is believed that fragmentary osteons, which are numerous in adult, do not become a notable fraction of osteon population density (OPD) until the late teens (Streeter, 2005).

Most histological studies of the ribs have focused on bone remodeling that occurs after the cessation of longitudinal skeletal growth. Variability in mineralization density distributions decrease with age, indicating that the remodeling rates deaccelerate during the period of bone maintenance (Reid and Boyde, 1987). Lower density osteons occur more frequently in the cutaneous cortex of the adult rib, and probably reflect differential mechanical loading across the ribs (Reid and Boyde, 1987). As chronological age increases, the adult cortex becomes crowded

22 with complete and fragmentary secondary osteons, and the proportion of primary circumferential lamellar bone decreases (Reid and Boyde, 1987; Stout and Paine, 1992; Stout and Lueck, 1995; Stout et al., 1996; Pratte and Pfeiffer, 1999; Cho et al., 2002, 2006; Paine and Brenton, 2006; Cho and Stout, 2011). In healthy individuals, a positive linear relationship is evident until menopause or senescence, when bone resorption begins to outpace bone formation.

Intra-cortical porosities also increase with advancing age (Sedlin et al., 1963b; Landeros and Frost, 1966; Agnew and Stout, 2012; Agnew et al., 2013b), and are most numerous in the cutaneous cortex of the rib (Landeros and Frost, 1966; Dominguez and Agnew, 2014, 2016; Agnew et al., 2015). Large osteonal canals and resorption spaces indicate a negative remodeling imbalance between osteoclasts and osteoblasts (Rauch et al., 2007).

2.5.2 Biomechanics of the Thorax and Body Size Considerations

Traditionally, studies of load history tend to focus on a single mechanical property and have been limited to weight bearing bones of the appendicular skeleton. Researchers contend that non- weight bearing elements are less affected by differences in activity patterns, and are more indicative of physiological condition and/or nutritional stress. However, the biomechanics of the thorax have been poorly characterized (Skedros, 2012), and strain thresholds are known to determine a number of localized remodeling events in all bone tissues. Recently, there has been an increased interest in characterizing the loading environment on human ribs to better understand injury risk (Schoell et al., 2015).

Although in vivo and adequate in vitro strain data are lacking, the ribs are thought to be cyclically loaded in bending—experiencing tensile strains in the cutaneous cortex and compressive strains in the pleural cortex (Agnew and Stout, 2012). The rib cage is distorted with inspiration as the intermediate and distal ribs move anterolaterally to expand the chest cavity (Koulouris and Dimitroulis, 2001). As the thorax increases in anterior-posterior dimension, rib curvature is reduced and tensile and compressive strains at the mid-shaft decrease. Upon expiration, the thorax assumes its original shape and mid-shaft strains increase. When rib cage distortion was measured in resting and exercising subjects, studies demonstrated that deeper breathing was correlated with increased distortion (Kenyon et al., 1997; Sanna et al., 1999). Repetitive loads endured from muscle contractions and the force of gravity cause microscopic

23 bone fractures to form, initiating bone turnover as a remedial response. This line of reasoning infers that increased activity level may correlate with greater remodeling in the ribs2.

Body size is largely responsible for the mechanical stress experienced by skeletal tissue, and can be used as a surrogate for mechanical load (Wetzsteon et al., 2011). Generally, this phenomenon is surmised as bigger people, bigger muscles, bigger bones (Parfitt, 2004). The male thorax undergoes marked changes at puberty, largely attributed to muscle development of the upper body (Sinclair and Dangerfield, 1998). Unsurprisingly, adult men typically have larger rib bones and greater indices of bone mass than their female counterparts (Dupras & Pfeiffer, 1996; Sedlin, Frost, & Villanueva, 1963; Streeter & Stout, 2003; Takahashi & Frost, 1966). Similarly, ribcage size has been shown to positively correlate with height, weight, and body mass index (BMI) (Gayzik et al., 2008). It is probable that histomorphometric variation also relates to body size, as low body weight is a known risk factor for osteoporotic rib fracture (Margolis et al., 2000).

2.6 South African Apartheid and the Social Determinants of Health

In 1948, the National Party came into power and formalized the apartheid political system, arguing that the peaceful coexistence of South Africa necessitated racial segregation. Under apartheid legislation, individuals were officially classified as South African Black (SAB), South African White (SAW), South African Coloured (SAC), or South African Indian/Asian (SAI). The apartheid government fostered inequalities in South Africa by denying large sections of the non-white population access to land ownership and community level infrastructure. Territorial “homelands” of non-white population groups received inferior governmental health expenditures and resource allocations, resulting in large health disparities among the South African population groups (Price, 1986; Nightingale et al., 1990; van Rensburg and Benatar, 1993; Harris et al., 2011).

2 Alternatively, Tommerup et al. (1993) found no effect on rib histomorphology with increased breathing rate in exercised pigs.

2.6.1 Peak Bone Mass Attainment in South Africa

Biological stress can negatively modulate longitudinal and appositional skeletal growth, as well as the timing of epiphyseal fusion. Growth faltering is associated with food insecurity, limited access to clean water, poor sanitation and hygiene, and low quality healthcare (Matrins et al., 2011; Ngure et al., 2014). Although socioeconomic status (SES) is an imprecise marker for specific living conditions and risks, it can serve as a general indicator of biological stress. Given institutionalized racism and inequality under the apartheid political system, most non-whites were of low socioeconomic status (SES) (Williams et al., 2008).

Childhood hunger and undernutrition are major health concerns in South Africa, especially as they are deeply rooted in poverty and social inequality (Iversen et al., 2011). Even in the post- apartheid era (1994 AD – present), growth in SAW children continues to be superior to that of their non-white peers (Cameron, 2003). Growth stunting is substantially higher among SAB and SAC, compared to SAW children (Burgard, 2002). On average, SAB children consume less calcium, are shorter, and undergo significantly less formal education than their white counterparts (Nyati et al., 2006). Growth patterns also appear to differ between the sexes as South African boys are more commonly underweight (Kimani-Murage et al., 2010). Moreover, body weight was shown to be lower in boys than age-matched girls across six African countries, indicating either differential care practices or greater biological frailty (Madise et al., 1999). Similarly, Cole and colleagues (2015) found that SAB boys of low SES experienced delayed skeletal maturation in comparison to SAW boys, but did not observe any differences among girls from different income levels.

Research on bone mass in South African children is limited and inconclusive. Early metacarpal radiometry studies indicate lower bone mass indices in SAB children compared to SAW children (Walker et al., 1971). Norris and colleagues (2008) found that DXA-derived bone area and bone mineral content values directly correlated with SES in SAB children, while McVeigh and colleagues (2004) found no relationship between BMD and physical activity level in SAB children. In the latter study, SAB girls had higher measures of bone mass than their white counterparts, but there were no significant differences between SAB and SAW boys (McVeigh, 2004). Similarly, size-adjusted BMC, measured using a single-photon absorptiometer, was found to be significantly greater in SAC girls than in SAW girls, but there were no differences between

25 boys from different population groups (Patel et al., 1993). However, more recent work with DXA has revealed higher BMC in SAB versus SAW boys at the femoral neck, hip, and mid- radius, but no significant differences at the whole body, lumbar spine, and distal one-third of the radius (Vidulich et al., 2011). Using pQCT, Micklesfield and colleagues (2011) concluded that SAB children have wider diaphyses with greater measures of bone strength than SAW children. One interpretation is that post-apartheid living conditions have improved bone outcomes in the non-white population, but others have argued that genetic factors influence bone mass in SAB children (May et al., 2013).

2.6.2 Osteoporosis and Fragility Fractures in South Africa

Access to bone densitometry and care is limited in South Africa, with few DXA machines available, and mostly in urban centers (Fuleihan, Adib, & Nauroy, 2011). Available data are circumscribed to specific geographic areas, failing to capture the true variability in bone mass among South Africans. Costly BMD scans are not part of the paid screening benefits currently offered by most medical aid systems (Bateman, 2006; Adonis et al., 2013). According to a 2011 survey, less than 6% of insured South African women over 50 years of age underwent preventative DXA screenings, with a clear urban bias, and considerable regional variation in the utilization of screening services. It can be assumed that the general uninsured population screening rate is much lower than that of the health-insured population (Adonis et al., 2013).. While some research shows higher BMD values at the proximal femur location in SAB compared to SAW women, vertebral bone mass appears to be similar (Daniels et al., 1995; Conradie et al., 2014), and may even be higher in SAW women (Chantler et al., 2012). Likewise, early metacarpal radiometry studies demonstrated lower cortical densities in SAB compared to SAW population groups (Solomon, 1979; Walker, Walker, & Richardson, 1971). In terms of global comparisons, one study did not find significant any differences in mean BMD values, but greater outer diameters, cross-sectional areas, endosteal diameters, and section moduli in the femoral neck region of United States blacks compared to South African blacks (Nelson et al., 2004).

Osteoporotic fracture incidence Africa has been insufficiently surveyed to inform on health economic strategies for treatment. The most comprehensive research suggests that South Africa had among the lowest global incidences of hip fractures during the 1950s and 1960s (Cumming

26 et al., 1997; Cauley et al., 2014). Consistent with similar studies conducted in the United States, lower hip fracture rates are reported for the SAB than for the SAW population group (Dent et al., 1968; Solomon, 1968, 1979). More recent research suggests that vertebral fracture prevalence rates do not differ among South African population groups (Conradie et al., 2015), although vertebral fractures are often undiagnosed (Delmas et al., 2005). While fracture prevalence rates differ between boys and girls (Thandrayen et al., 2009, 2014), hip fractures seem to occur equally among elderly men and women (Solomon, 1968, 1979).

2.7 Genetic Considerations

Racial or ethnic differences in bone mass and microstructure are often inappropriately reduced to underlying genetic factors, leading to erroneous biological conclusions, especially if population admixture is not considered. This is of particular concern for South Africa, where there is an incredible amount of genetic diversity because of the country’s long and multifaceted history of migration and colonization. Indigenous Khoesan, classified as SACs under apartheid, demonstrate great genomic variation (Schlebusch et al., 2012), and probably represent the deepest historical population divergences among living people (Henn et al., 2010; Schuster et al., 2010; Schlebusch et al., 2013). Linguistic and genetic evidence indicates that the Bantu, whose descendants make up the SAB population, originated in the Nigeria-Cameroon grasslands, branching from the Niger-Congo language family (Vansina, 1995; de Filippo et al., 2011). They migrated to occupy much of the east and southern parts of Africa, reaching South Africa by as early as 500 C.E. (Gurdasani et al., 2015). Variants from this Bantu family included the Nguni group, who occupied the eastern coastal plains, and the Sotho-Tswana group, who lived on the interior (Pakendorf et al., 2017). The arrival of Europeans in the 1400s, the establishment of the Dutch-controlled Cape Colony in 1657, and the subsequent slave trade in the Western Cape further contributed to a complex pattern of genetic variation in South Africa (Petersen et al., 2013). Studies have identified up to five distinct ancestral populations in the SAC genome, with significantly different ancestry proportions among SAC individuals sampled from different regions of the country (de Wit et al., 2013; Choundry et al., 2017; Figure 4). In the Western Cape, SAC individuals show varying degrees of admixture with Khoesan, Bantu-speakers, Europeans, and populations from the Indian sub-continent (Choundry et al., 2017).

Separate migration paths of Bantu-speaking agro-pastoralists into South Africa, beginning roughly 2,000 years ago (Badenhorst, 2010), help to explain some of the genetic differences between SAB individuals from the southeastern and southwestern parts of country (Choundry et al., 2017). It has recently been proposed that East Africans are at much higher osteoporosis risk compared to West Africans because of the lactase persistence genotype (Hilliard, 2016). SABs are also genetically distinct from the West Africans that dominate the African-American gene pool (Patin et al., 2017; Silva et al., 2015; Tishkoff et al., 2009). Therefore, it may not be reasonable to assume greater bone mass in SABs compared to other population groups, as is typically observed for African Americans (Looker et al., 2009). Moreover, there is a tremendous amount of admixture within the United States African American population (Parra et al., 1998; Bryc et al., 2015), bringing into question the broader utility of these race-based characterizations as genetic proxies.

2.8 Literature Cited

Adonis L, An R, Luiz J, Mehrotra A, Patel D, Basu D, Sturm R. 2013. Provincial screening rates for chronic diseases of lifestyle, cancers and HIV in a health-insured population. South African Med J 103:309–312. Agnew AM, Kang YS, Moorhouse K, Herriott R, Bolte IV JH. 2013a. Age-related changes in

stiffness in human ribs. In: IRCOBI Conference. . p 257–269. Agnew AM, Moorhouse K, Kang YS, Donnelly BR, Pfefferle K, Manning AX, Litsky AS, Herriott R, Abdel-Rasoul M, Bolte IV JH. 2013b. The response of pediatric ribs to quasi- static loading: Mechanical properties and microstructure. Ann Biomed Eng 41:2501–2514. Agnew AM, Schafman M, Moorhouse K, White SE, Kang Y. 2015. The effect of age on the structural properties of human ribs. J Mech Behav Biomed Mater [Internet] 41:302–314. Available from: http://dx.doi.org/10.1016/j.jmbbm.2014.09.002 Agnew AM, Stout SD. 2012. Brief communication: Reevaluating osteoporosis in human ribs: The role of intracortical porosity. Am J Phys Anthropol 148:462–466. Allen MR, Krohn K. 2013. Skeletal Imaging. Basic Appl Bone Biol:93–113. Badenhorst S. 2010. Descent of iron age farmers in Southern Africa during the last 2000 years. African Archaeol Rev 27:87–106. Bailey DA, McKay HA, Mirwald RL, Crocker PR, Faulkner RA. 1999. A six-year longitudinal study of the relationship of physical activity to bone mineral accrual in growing children: the university of Saskatchewan bone mineral accrual study. J Bone Miner Res 14:1672– 1679. Baker JF, Davis M, Alexander R, Zemel BS, Mostou S, Shults J, Sulik M, Schiferl DJ, Leonard MB. 2013. Associations between body composition and bone density and structure in men and women across the adult age spectrum ☆. Bone 53:34–41. Balasch J. 2003. Sex steroids and bone: Current perspectives. Hum Reprod Update 9:207–222. Baltadjiev G. 1995. Micromorphometric Characteristics of Osteons in Compact Bone of Growing Tibiae of Human Fetuses. Acta Anat 154:181–185. Bateman C. 2006. South Africa under-prioritises osteoporosis. South African Med J 96:19–20. Baxter-Jones ADG, Faulkner R a, Forwood MR, Mirwald RL, Bailey D a. 2011. Bone mineral accrual from 8 to 30 years of age: an estimation of peak bone mass. J bone Miner Res Off J Am Soc Bone Miner Res 26:1729–39. Bayer M. 2014. Reference values of osteocalcin and procollagen type I N-propeptide plasma levels in a healthy Central European population aged 0-18 years. Osteoporos Int 25:729– 736. de Boer HHH, Van der Merwe AEL. 2016. Diagnostic dry bone histology in human paleopathology. Clin Anat 29:831–843. Bonewald LF. 2011. The amazing osteocyte. J Bone Miner Res 26:229–38. Bonewald LF, Johnson ML. 2008. Osteocytes, mechanosensing and Wnt signaling. Bone 42:606–615. Bonjour J, Chevalley T, Ferrari S, Rizzoli R. 2012. Peak bone mass and its regulation. In: Pediatric bone. Second Edi. Elsevier Inc. p 189–221. Bonjour J, Theintz G, Buchs B, Slosman D, Rizzoli R. 1991. Critical years and stages of puberty for spinal and femoral bone mass accumulation during adolescence. J Clin Endocrinol Metab 73:553–563. Boyce BF, Xing L. 2007. The RANKL/RANK/OPG pathway. Curr Osteoporos Rep 5:98–104. Boyce TM, Bloebaum RD. 1993. Cortical aging differences and fracture implications for the human femoral neck. Bone 14:769–778. Boyde A, Jones SJ, Aerssens J, Dequeker J. 1995. Mineral density quantitation of the human cortical iliac crest by backscattered electron image analysis: Variations with age, sex, and degree of osteoarthritis. Bone 16:619–627. Brenton BP, Paine RR. 2007. Reevaluating the Health and Nutritional Status of Maize- Dependent Populations: Evidence for the Impact of Pellagra on Human Skeletons from

South Africa. Ecol Food Nutr [Internet] 46:345–360. Available from: http://www.tandfonline.com/doi/abs/10.1080/03670240701486545 Brockstedt H, Kassem M, Eriksen EF, Mosekilde L, Melsen F. 1993. Age- and sex-related changes in iliac cortical bone mass and remodeling. Bone 14:681–691. Bromage TG, Juwayeyi YM, Katris JA, Gomez S, Ovsiy O, Goldstein J, Janal MN, Hu B, Schrenk F. 2016. The scaling of human osteocyte lacuna density with body size and metabolism. Comptes rendus - Palevol [Internet] 15:32–39. Available from: http://dx.doi.org/10.1016/j.crpv.2015.09.001 de Bruin JP, Bovenhuis H, van Noord P a, Pearson PL, van Arendonk J a, te Velde ER, Kuurman WW, Dorland M. 2001. The role of genetic factors in age at natural menopause. Hum Reprod 16:2014–2018. Bryc K, Durand EY, Macpherson JM, Reich D, Mountain JL. 2015. The genetic ancestry of african americans, latinos, and european Americans across the United States. Am J Hum Genet [Internet] 96:37–53. Available from: http://dx.doi.org/10.1016/j.ajhg.2014.11.010 Buie HR, Campbell GM, Klinck RJ, MacNeil J a., Boyd SK. 2007. Automatic segmentation of cortical and trabecular compartments based on a dual threshold technique for in vivo micro- CT bone analysis. Bone 41:505–515. Burgard S. 2002. Does Race Matter? Children’s Height in Brazil and South Africa. Demography 39:763–790. Burton P, Nyssen-Behets C, Dhem A. 1989. Haversian Bone Remodelling in Human Fetus. Acta Anat 135:171–175. Cameron N. 2003. Physical growth in a transitional economy: The aftermath of South African apartheid. Econ Hum Biol 1:29–42. Campbell GM, Sophocleous A. 2014. Quantitative analysis of bone and soft tissue by micro- computed tomography: applications to ex vivo and in vivo studies. Bonekey Rep [Internet] 3:1–12. Available from: http://www.portico.org/Portico/article?article=pgk2ph9dzmn Cauley JA, Chalhoub D, Kassem AM, Fuleihan GE-H. 2014. Geographic and ethnic disparities in osteoporotic fractures. Nat Rev Endocrinol [Internet] 10:338–351. Available from: http://www.nature.com/doifinder/10.1038/nrendo.2014.51 Chantler S, Dickie K, Goedecke JH, Levitt NS, Lambert E V., Evans J, Joffe Y, Micklesfield LK. 2012. Site-specific differences in bone mineral density in black and white premenopausal South African women. Osteoporos Int 23:533–542. Cho H, Stout SD. 2011. Age-associated bone loss and intraskeletal variability in the Imperial Romans. J Anthropol Sci [Internet] 89:109–25. Available from: http://www.ncbi.nlm.nih.gov/pubmed/21368344 Cho H, Stout SD, Bishop TA. 2006. Cortical bone remodeling rates in a sample of African American and European American descent groups from the American midwest: Comparisons of age and sex in ribs. Am J Phys Anthropol 130:214–226. Cho H, Stout SD, Madsen RW, Streeter MA. 2002. Population-specific histological age- estimating method: A model for known African-American and European-American skeletal remains. J Forensic Sci 47:12–18. Choudhury A, Ramsay M, Hazelhurst S, Aron S, Bardien S, Botha G, Chimusa ER, Christoffels A, Gamieldien J, Sefid-Dashti MJ, Joubert F, Meintjes A, Mulder N, Ramesar R, Rees J, Scholtz K, Sengupta D, Soodyall H, Venter P, Warnich L, Pepper MS. 2017. Whole- genome sequencing for an enhanced understanding of genetic variation among South Africans. Nat Commun [Internet] 8:2062. Available from: http://www.nature.com/articles/s41467-017-00663-9

Clarke BL, Khosla S. 2010. Female reproductive system and bone. Arch Biochem Biophys 503:118–28. Cole TJ, Rousham EK, Hawley NL, Cameron N, Norris SA, Pettifor JM. 2015. Ethnic and sex differences in skeletal maturation among the Birth to Twenty cohort in South Africa. Arch Dis Child [Internet] 100:138–143. Available from: http://adc.bmj.com/lookup/doi/10.1136/archdischild-2014-306399 Conradie M, Conradie MM, Kidd M, Hough S. 2014. Bone density in black and white South African women: contribution of ethnicity, body weight and lifestyle. Arch Osteoporos [Internet] 9:193. Available from: http://link.springer.com/10.1007/s11657-014-0193-0 Conradie M, Conradie MM, Scher AT, Kidd M, Hough S. 2015. Vertebral fracture prevalence in black and white South African women. Arch Osteoporos 10:203. Cooper C, Westlake S, Harvey N, Javaid K, Dennison E, Hanson M. 2006. Review: developmental origins of osteoporotic fracture. Osteoporos Int [Internet] 17:337–47. Available from: http://www.ncbi.nlm.nih.gov/pubmed/16331359 Crowder C, Rosella L. 2007. Assessment of intra- and intercostal variation in rib histomorphometry: its impact on evidentiary examination. J Forensic Sci 52:271–6. Cumming R, Nevitt M, Cummings SR. 1997. Epidemiology of hip fractures. Epidemiol Rev 19:244–257. Dallas SL, Bonewald LF. 2011. Dynamics of the Transition from Osteoblast to Osteocyte. Ann N Y Acad Sci 1192:437–443. Daniels ED, Pettifor JM, Schnitzler CM, Russell SW, Patel DN. 1995. Ethnic differences in bone density in female South African nurses. J Bone Miner Res 10:359–367. Delmas PD, van de Langerijt L, Watts NB, Eastell R, Genant H, Grauer A, Cahall DL. 2004. Underdiagnosis of Vertebral Fractures Is a Worldwide Problem: The IMPACT Study. J Bone Miner Res [Internet] 20:557–563. Available from: http://doi.wiley.com/10.1359/JBMR.041214 Dent CE, Engelbrecht HE, Godfrey RC. 1968. Osteoporosis of lumbar vertebrae and calcification of abdominal aorta in women living in Durban. Br Med J [Internet] 4:76–79. Available from: http://www.ncbi.nlm.nih.gov/pubmed/5696550 Dominguez VM, Agnew AM. 2014. Patterns in resorptive spaces in elderly rib cortices. Am J Phys Anthropol 153:107. Dominguez VM, Agnew AM. 2016. Examination of factors potentially influencing osteon size in the human rib. Anat Rec 299:313–324. Dupras TL, Pfeiffer SK. 1996. Determination of sex from adult human ribs. Can Soc Forensic Sci J 29:221–231. Eleazer CD, Jankauskas R. 2016. Mechanical and metabolic interactions in cortical bone development. Am J Phys Anthropol 160:317–333. Emerton KB, Hu B, Woo AA, Sinofsky A, Hernandez C, Majeska RJ, Jepsen KJ, Schaffler MB. 2011. Resorption Following Ovariectomy in Mice. Bone 46:577–583. Enlow DH. 1962. A Study o f the Post-Natal Growth and Remodeling of Bone. Am J Anat 110:79–101. Epker BN, Frost HM. 1965a. The Direction of Transverse Drift of Actively Forming Osteons in Human Rib Cortex. J Bone Joint Surg Am 47:1211–1215. Epker BN, Frost HM. 1965b. A histological study of remodeling at the periosteal, haversian canal, cortical endosteal, and trabecular endosteal surfaces in human rib. Anat Rec 152:129–135. Ericksen MF. 1976. Cortical bone loss with age in three native American populations. Am J Phys

Anthropol [Internet] 45:443–52. Available from: http://www.ncbi.nlm.nih.gov/pubmed/998766 Ericksen MF. 1991. Histologic Estimation of Age at Death Using the Anterior Cortex of the Femur. Am J Phys Anthropol 84:171–179. Ferrari S, Rizzoli R, Bonjour J. 1998. Heritable and nutritional influences on bone mineral mass. Aging (Milano) 10:205–213. Filippo C De, Barbieri C, Whitten M, Mpoloka SW, Gunnarsdóttir D, Bostoen K, Nyambe T, Beyer K, Schreiber H, Knijff P De, Luiselli D, Stoneking M, Pakendorf B. 2011. Y- chromosomal variation in Sub-Saharan Africa: insights into the history of Niger-Congo groups. Mol Biol Evol 28:1255–1269. Follet H, Boivin G, Rumelhart C, Meunier PJ. 2004. The degree of mineralization is a determinant of bone strength: A study on human calcanei. Bone 34:783–789. Fournier PE, Rizzoli R, Slosman DO, Theintz G, Bonjour JP. 1997. Asynchrony between the rates of standing height gain and bone mass accumulation during puberty. Osteoporos Int 7:525–532. Franz-Odendaal TA, Hall BK, Witten PE. 2006. Buried alive: how osteoblasts become osteocytes. Dev Dyn [Internet] 235:176–190. Available from: http://www.ncbi.nlm.nih.gov/pubmed/16258960 Fratzl-Zelman N, Roschger P, Misof BM, Pfeffer S, Glorieux FH, Klaushofer K, Rauch F. 2009. Normative data on mineralization density distribution in iliac bone biopsies of children, adolescents and young adults. Bone 44:1043–8. Frost HM. 1969. Tetracycline-based histological analysis of bone remodeling. Calcif Tissue Res [Internet] 3:211–237. Available from: http://link.springer.com/10.1007/BF02058664 Frost HM, Wu K. 1967. Histological measurement of bone formation rates in unlabeled contemporary, archaeological and paleontological compact bone. Fuleihan GE-H, Adib MG, Nauroy L. 2011. The Middle East & Africa Regional Audit: Epidemiology, Costs, & Burden of Osteoporosis in 2011. (Stenmark J, Misteli L, editors.). Nyon, Switzerland. Garn S. 1970. The earlier gain and later loss of cortical bone. Springfield: Charles C. Thomas. Garn SM, Nagy JM, Sandusky ST. 1972. Differential sexual dimorphism in bone diameters of subjects of European and African ancestry. Am J Phys Anthropol 37:127–129. Garnero P, Hausherr E, Chapuy MC, Marcelli C, Grandjean H, Muller C, Cormier C, Bréart G, Meunier PJ, Delmas PD. 1996. Markers of bone resorption predict hip fracture in elderly women: the EPIDOS Prospective Study. J Bone Miner Res 11:1531–1538. Gayzik FS, Yu MM, Danelson KA, Slice DE, Stitzel JD. 2008. Quantification of age-related shape change of the human rib cage through geometric morphometrics. J Biomech 41:1545–54. Giustina A, Veldhuis JD. 1998. Pathophysiology of the neuroregulation of growth hormone secretion in experimental animals and the human. Endocr Rev 19:717–797. Gold EB, Bromberger J, Crawford S, Samuels S, Greendale GA, Harlow SD. 2001. Factors Associated with Age at Natural Menopause in a Multiethnic Sample of Midlife Women. Am J Epidemiol 153:49–52. Goldman HM. 2001. Histocomposition and geometry at the human mid-shaft femur. Goldman HM, Bromage TG, Boyde A, Thomas CDL, Clement JG. 2003. Intrapopulation variability in mineralization density at the human femoral mid-shaft. J Anat 203:243–255. Goldman HM, Hampson NA, Guth JJ, Lin D, Jepsen KJ. 2014. Intracortical remodeling parameters are associated With measures of bone robustness. Anat Rec 297:1817–1828.

Goldman HM, Thomas CDL, Clement JG, Bromage TG. 2005. Relationships among microstructural properties of bone at the human midshaft femur. J Anat 206:127–139. Greendale G a., Sowers M, Han W, Huang MH, Finkelstein JS, Crandall CJ, Lee JS, Karlamangla AS. 2012. Bone mineral density loss in relation to the final menstrual period in a multiethnic cohort: Results from the Study of Women’s Health Across the Nation (SWAN). J Bone Miner Res 27:111–118. Gurdasani D, Carstensen T, Tekola-Ayele F, Pagani L, Tachmazidou I, Hatzikotoulas K, Karthikeyan S, Iles L, Pollard MO, Choudhury A, Ritchie GRS, Xue Y, Asimit J, Nsubuga RN, Young EH, Pomilla C, Kivinen K, Rockett K, Kamali A, Doumatey AP, Asiki G, Seeley J, Sisay-Joof F, Jallow M, Tollman S, Mekonnen E, Ekong R, Oljira T, Bradman N, Bojang K, Ramsay M, Adeyemo A, Bekele E, Motala A, Norris SA, Pirie F, Kaleebu P, Kwiatkowski D, Tyler-Smith C, Rotimi C, Zeggini E, Sandhu MS. 2015. The African Genome Variation Project shapes medical genetics in Africa. Nature [Internet] 517:327– 332. Available from: http://dx.doi.org/10.1038/nature13997 Han S-H, Kim S-H, Ahn Y-W, Huh G-Y, Kwak D-S, Park D-K, Lee U-Y, Kim Y-S. 2009. Microscopic age estimation from the anterior cortex of the femur in Korean adults. J Forensic Sci 54:519–522. Harris B, Goudge J, Ataguba JE, McIntyre D, Nxumalo N, Jikwana S, Chersich M. 2011. Inequities in access to health care in South Africa. J Public Health Policy [Internet] 32:S102–S123. Available from: http://dx.doi.org/10.1057/jphp.2011.35 Heaney RP, Abrams S, Dawson-Hughes B, Looker A, Marcus R, Matkovic V, Weaver C. 2000. Peak Bone Mass. Osteoporos Int 11:985–1009. Heinrich JT. 2015. Spatial Characterization of Rib Cortical Bone Microstructure and the Effect of Nutritional and Physiological Stresses. Henn BM, Gignoux CR, Jobin M, Granka JM, Macpherson JM, Kidd JM, Rodriguez-Botigue L, Ramachandran S, Hon L, Brisbin A, Lin a a, Underhill P a, Comas D, Kidd KK, Norman PJ, Parham P, Bustamante CD, Mountain JL, Feldman MW. 2011. Hunter-gatherer genomic diversity suggests a southern African origin for modern humans. Proc Natl Acad Sci U S A 108:5154–5162. Hert J, Fiala P, Petrttl M. 1994. Osteon Orientation of the Diaphysis of the Long Bones in Man. Bone 15:269–277. Hilliard CB. 2016. High osteoporosis risk among East Africans linked to lactase persistence genotype. Bonekey Rep [Internet] 5:1–9. Available from: http://www.portico.org/Portico/article?article=pgk3nd1r8j5 Hopper JL, Green RM, Nowson C a, Young D, Sherwin a J, Kaymakci B, Larkins RG, Wark JD. 1998. Genetic, common environment, and individual specific components of variance for bone mineral density in 10- to 26-year-old females: a twin study. Am J Epidemiol 147:17–29. Horstman AM, Dillon EL, Urban RJ, Sheffield-Moore M. 2012. The role of androgens and estrogens on healthy aging and longevity. Journals Gerontol A Biol Sci Med Sci 67:1140– 1152. Imam A, Iqbal J, Blair HC, Davies TF, Huang CL-H, Zallone A, Zaidi M, Sun L. 2009. Role of the pituitary-bone axis in skeletal pathophysiology. Curr Opin Endocrinol Diabetes Obes 16:423–429. Iversen PO, Plessis L du, Marais D, Morseth M, Hoisaether EA, Herselman M. 2011. Nutritional health of young children in South Africa over the first 16 years of democracy. South African J Child Heal 5:72–77.

Jantz LM, Jantz RL. 1999. Secular change in long bone length and proportion in the United States, 1800-1970. Am J Phys Anthropol 110:57–67. Jones SJ, Glorieux FH, Travers R, Boyde A. 1999. Clinical Investigations The Microscopic Structure of Bone in Normal Children and Patients with Osteogenesis Imperfecta : A Survey Using Backscattered Electron Imaging. Calcif Tissue Int 64:8–17. Jowsey J. 1960. Age Changes in Human Bone. Clin Orthop 17:210–218. Katzman DK, Bachrach LK, Carter DR, Marcus R. 1991. Clinical and Anthropometric Correlates of Bone Mineral Acquisition in Healthy Adolescent Girls. J Clin Endocrinol Metab 73:1332–1339. Kenyon CM, Cala SJ, Yan S, Aliverti a, Scano G, Duranti R, Pedotti a, Macklem PT. 1997. Rib cage mechanics during quiet breathing and exercise in humans. J Appl Physiol 83:1242– 1255. Kerley ER. 1965. The Microscopic Determination of Age in Human Bone. Am J Phys Anthropol 23:149–163. Kerley ER, Ubelaker DH. 1978. Revisions in the Microscopic Method of Estimating Age at Death. Am J Phys Anthropol 49:545–547. Khosla S, Amin S, Orwoll E. 2008. Osteoporosis in men. Endocr Rev 29:441–464. Khosla S, Atkinson EJ, Melton LJ, Riggs BL. 1997. Effects of age and estrogen status on serum parathyroid hormone levels and biochemical markers of bone turnover in women: A population-based study. J Clin Endocrinol Metab 82:1522–1527. Khosla S, Pacifici R. 2013. Estrogen deficiency, postmenopausal osteoporosis, and age-related bone loss. In: Osteoporosis. Fourth Edi. Elsevier Inc. p 1113–1136. Kimani-Murage EW, Kahn K, Pettifor JM. 2010. The prevalence of stunting, overweight and obesity, and metabolic disease risk in rural South African children. BMC public [Internet]:1–13. Available from: http://www.biomedcentral.com/1471-2458/10/158 Kirmani S, Christen D, van Lenthe GH, Fischer PR, Bouxsein ML, McCready LK, Melton LJ, Riggs BL, Amin S, Müller R, Khosla S. 2009. Bone structure at the distal radius during adolescent growth. J Bone Miner Res 24:1033–1042. Klein-Nulend J, Bakker AD, Bacabac RG, Vatsa A, Weinbaum S. 2013. Mechanosensation and transduction in osteocytes. Bone [Internet] 54:182–190. Available from: http://dx.doi.org/10.1016/j.bone.2012.10.013 Koehne T, Vettorazzi E, Küsters N, Lüneburg R, Kahl-nieke B, Püschel K, Amling M, Busse B. 2014. Trends in trabecular architecture and bone mineral density distribution in 152 individuals aged 30 – 90 years. Bone [Internet] 66:31–38. Available from: http://dx.doi.org/10.1016/j.bone.2014.05.010 Kontulainen S a, Macdonald HM, Khan KM, McKay H a. 2005. Examining bone surfaces across puberty: a 20-month pQCT trial. J Bone Miner Res 20:1202–1207. Kontulainen SA, Macdonald HM, Mckay HA. 2006. Change in cortical bone density and its distribution differs between boys and girls during puberty. J Clin Endocrinol Metab 91:2555–2561. Koulouris MG, Dimitroulis I. 2001. Structure and function of the respiratory muscles. Pheumon 14:185–217. Lacroix P. 1971. The internal remodeling of bones. In: Bourne GH, editor. The Biochemistry and Physiology of Bone, Vol. III. New York: Academic Press. p 119–14. Landeros O, Frost HM. 1966. Comparison of amounts of remodeling activity in opposite cortices of ribs in children and adults. J Dent Res 45:152–158. Lee U-Y, Jung G-U, Choi S-G, Kim Y-S. 2014. Anthropological Age Estimation with Bone

Histomorphometry from the Human Clavicle. Anthropologist 17:929–936. Lobo RA. 2014. Menopause and Aging. In: Strauss JF, Barbieri R, editors. Yen & Jaffe’s Reproductive Endocrinology: Physiology, Pathophysiology, and Clinical Management. Seventh Ed. Philadelphia, PA: Elsevier Saunders. Looker AC, Melton LJ, Harris T, Borrud L, Shepherd J, McGowan J. 2009. Age, gender, and race/ethnic differences in total body and subregional bone density. Osteoporos Int 20:1141– 1149. Loro ML, Sayre J, Roe TF, Goran MI, Kaufman FR, Gilsanz V. 2000. Early identification of children predisposed to low peak bone mass and osteoporosis later in life. J Clin Endocrinol Metab 85:3908–3918. Maat GJR, Maes A, Aarents MJ, Nagelkerke NJD. 2006. Histological age prediction from the femur in a contemporary Dutch sample. The decrease of nonremodeled bone in the anterior cortex. J Forensic Sci 51:230–237. Macdonald HM, Hoy CL, Mckay HA. 2013. Osteoporosis. In: Osteoporosis. Fourth Edi. Elsevier. p 1017–1036. Available from: http://dx.doi.org/10.1016/B978-0-12-415853- 5.00042-X Madise NJ, Matthews Z, Margetts B. 1999. Heterogeneity of child nutritional status between households: A comparison of six sub-Saharan African countries. Popul Stud (NY) 53:331– 343. Maggiano C. 2012. Making the mold: a microstructural perspective on bone modeling during growth and mechanical adaptation. In: Crowder CM, Stout SD, editors. Bone Histology: An Anthropological Perspective. 1st Editio. Boca Raton: CRC Press. Malina RM, Bouchard C, Bar-Or O. 2004. Growth maturation and physical activity. Champaign, IL: Human Kinetics. Margolis KL, Ensrud KE, Schreiner PJ, Tabor HK. 2000. Body size and risk for clinical fractures in older women. Ann Intern Med 133:123–127. Martin-Millan M, Almeida M, Ambrogini E, Han L, Zhao H, Weinstein RS, Jilka RL, O’Brien C a, Manolagas SC. 2010. The estrogen receptor-alpha in osteoclasts mediates the protective effects of estrogens on cancellous but not cortical bone. Mol Endocrinol 24:323–334. Martin RB, Burr DB, Sharkley NA. 1998. Skeletal Tissue Mechanics. 1st Editio. New York: Springer Science and Business Media. Matkovic V, Jelic T, Wardlaw GM, Llich JZ, Goel PK, Wright JK, Andon MB, Smith KT, Heaney RP. 1994. Timing of peak bone mass in Caucasian females and its implication for the prevention of osteoporosis: Inference from a cross-sectional model. J Clin Invest 93:799–808. Matrins VJB, Toledo Florêncio TMM, Grillo LP, Franco M do CP, Martins PA, Clemente APG, Santos CDL, Vieria M de FA, Sawaya AL. 2011. Long-lasting effects of undernutrition. Int J Environ Res Public Health 8:1817–1846. May A, Pettifor JM, Norris SA, Zane R. 2013. Genetic factors influencing bone mineral content in a black South African population. J Bone Res 31:708–716. McVeigh JA. 2004. Associations between physical activity and bone mass in black and white South African children at age 9 yr. J Appl Physiol [Internet] 97:1006–1012. Available from: http://jap.physiology.org/cgi/doi/10.1152/japplphysiol.00068.2004 Mescher A. 2010. Junqueria’s Basic Histology: Text and Atlas. 12th Editi. New York: McGraw- Hill Medical. Micklesfield LK, Norris SA, Pettifor JM. 2011. Determinants of bone size and strength in 13- year-old South African children: The influence of ethnicity, sex and pubertal maturation.

Bone 48:777–785. Misof BM, Dempster DW, Zhou H, Roschger P, Fratzl-Zelman N, Fratzl P, Silverberg SJ, Shane E, Cohen A, Stein E, Nickolas TL, Recker RR, Lappe J, Bilezikian JP, Klaushofer K. 2014. Relationship of Bone Mineralization Density Distribution (BMDD) in Cortical and Cancellous Bone Within the Iliac Crest of Healthy Premenopausal Women. Calcif Tissue Int 95:332–339. Mølgaard C, Thomsen BL, Prentice A, Cole TJ, Michaelsen KF. 1999. Whole body bone mineral content in healthy children and adolescents. Arch Dis Child 76:9–15. Mulhern DM. 2000. Rib remodeling dynamics in a skeletal population from Kulubnarti, Nubia. Am J Phys Anthropol 111:519–530. Nakamura T, Imai Y, Matsumoto T, Sato S, Takeuchi K, Igarashi K, Harada Y, Azuma Y, Krust A, Yamamoto Y, Nishina H, Takeda S, Takayanagi H, Metzger D, Kanno J, Takaoka K, Martin TJ, Chambon P, Kato S. 2007. Estrogen Prevents Bone Loss via Estrogen Receptor α and Induction of Fas Ligand in Osteoclasts. Cell 130:811–823. Nelson DA, Pettifor JM, Barondess DA, Cody DD, Uusi-Rasi K, Beck TJ. 2004. Comparison of Cross-Sectional Geometry of the Proximal Femur in White and Black Women from Detroit and Johannesburg. J Bone Miner Res [Internet] 19:560–565. Available from: http://doi.wiley.com/10.1359/JBMR.040104 Ngure FM, Reid BM, Humphrey JH, Mbuya MN, Pelto G, Stoltzfus RJ. 2014. Water, sanitation, and hygiene (WASH), environmental enteropathy, nutrition, and early child development: Making the links. Ann N Y Acad Sci 1308:118–128. Nightingale EO, Hannibal K, Geiger J, Hartmann L, Lawerence R, Spurlock J. 1990. Apartheid medicine. JAMA J … [Internet] 264:2097–2102. Available from: http://jama.ama- assn.org/content/264/16/2097.short Niu T, Rosen CJ. 2005. The insulin-like growth factor-I gene and osteoporosis: A critical appraisal. Gene 361:38–56. Van Noord P a H, Boersma H, Dubas JS, Te Velde E, Dorland M. 1997. Age at natural menopause in a population-based screening cohort: The role of menarche, fecundity, and lifestyle factors. Fertil Steril 68:95–102. Nor FM, Pastor RF, Schutkowski H. 2013. Age at death estimation from bone histology in Malaysian males. Med Sci Law [Internet] 54:203–208. Available from: http://www.ncbi.nlm.nih.gov/pubmed/24189643 Norris SA, Sheppard ZA, Griffiths PL, Cameron N, Pettifor JM. 2008. Current Socio-Economic Measures, and Not Those Measured During Infancy, Affect Bone Mass in Poor Urban South African Children. J Bone Miner Res 23. Nyati LH, Norris SA, Cameron N, Pettifor JM. 2006. Effect of ethnicity and sex on the growth of the axial and appendicular skeleton of children living in a developing country. Am J Phys Anthropol 130:135–141. Openshaw P, Edwards S, Helms P. 1984. Changes in rib cage geometry during childhood. Thorax 39:624–627. Paine RR, Brenton BP. 2006. Dietary health does affect histological age assessment: An evaluation of the Stout and Paine (1992) age estimation equation using secondary osteons from the rib. J Forensic Sci 51:489–492. Pakendorf B, Gunnink H, Sands B, Bostoen K. 2017. Prehistoric Bantu-Khoisan language contact: A cross-disciplinary approach. Lang Dyn Chang 7:1–46. Parfitt a M, Villanueva a R, Foldes J, Rao DS. 1995. Relations between histologic indices of bone formation: implications for the pathogenesis of spinal osteoporosis. J Bone Miner Res

10:466–473. Parfitt AM. 1984. The cellular basis of bone remodeling: The quantum concept reexamined in light of recent advances in the cell biology of bone. Calcif Tissue Int 36. Parfitt AM. 1994. Osteonal and Hemi-Osteonal Remodeling: The Spatial and Temporal Framework for Signal Traffic in Adult Human Bone. J Cell Biochem 55:273–286. Parfitt AM. 2004. The attainment of peak bone mass: What is the relationship between muscle growth and bone growth? Bone 34:767–770. Parra EJ, Marcini A, Akey J, Martinson J, Batzer MA, Cooper R, Forrester T, Allison DB, Deka R, Ferrell RE, Shriver MD. 1998. Estimating African American admixture proportions by use of population-specific alleles. Am J Hum Genet [Internet] 63:1839–51. Available from: http://www.sciencedirect.com/science/article/pii/S0002929707616280 Patel DN, Pettifor JM, Becker PJ. 1993. The effect of ethnicity on appendicular bone mass in white, coloured and Indian schoolchildren. South African Med J 83:847–853. Patin E, Lopez M, Grollemund R, Verdu P, Harmant C, Quach H, Laval G, Perry GH, Barreiro LB, Froment A, Heyer E, Massougbodji A, Fortes-Lima C, Migot-Nabias F, Bellis G, Dugoujon J-M, Pereira JB, Fernandes V, Pereira L, Van der Veen L, Mouguiama-Daouda P, Bustamante CD, Hombert J-M, Quintana-Murci L. 2017. Dispersals and genetic adaptation of Bantu-speaking populations in Africa and North America. Science (80- ) [Internet] 356:543–546. Available from: http://www.sciencemag.org/lookup/doi/10.1126/science.aal1988 Pavón MV, Cucina A, Tiesler V. 2010. New formulas to estimate age at death in Maya populations using histomorphological changes in the fourth human rib. J Forensic Sci 55:473–477. Pearson OM, Lieberman DE. 2004. The aging of Wolff’s “law”: ontogeny and responses to mechanical loading in cortical bone. Am J Phys Anthropol Suppl 39:63–99. Perrien DS, Achenbach SJ, Bledsoe SE, Walser B, Suva LJ, Khosla S, Gaddy D. 2006. Bone turnover across the menopause transition: Correlations with inhibins and follicle-stimulating hormone. J Clin Endocrinol Metab 91:1848–1854. Petersen DC, Libiger O, Tindall EA, Hardie RA, Hannick LI, Glashoff RH, Mukerji M, Fernandez P, Haacke W, Schork NJ, Hayes VM. 2013. Complex Patterns of Genomic Admixture within Southern Africa. PLoS Genet 9:10–13. Petit M a., Beck TJ, Kontulainen S a. 2005. Examining the developing bone: What do we measure and how do we do it? J Musculoskelet Neuronal Interact 5:213–224. Petrtyl M, Hert J, Fiala P. 1996. Spatial Organization of the Haversian Bone in Man.pdf. J Biomech 29:161–169. Pfeiffer S, Crowder C, Harrington L, Brown M. 2006. Secondary osteon and Haversian canal dimensions as behavioral indicators. Am J Phys Anthropol 131:460–468. Pfeiffer S, King P. 1983. Cortical bone formation and diet among protohistoric iroquoians. Am J Phys Anthropol 60:23–28. Pocock N a., Eisman J a., Hopper JL, Yeates MG, Sambrook PN, Eberl S. 1987. Genetic determinants of bone mass in adults: A twin study. J Clin Invest 80:706–710. Power J, Doube M, van Bezooijen RL, Loveridge N, Reeve J. 2012. Osteocyte recruitment declines as the osteon fills in: interacting effects of osteocytic sclerostin and previous hip fracture on the size of cortical canals in the femoral neck. Bone [Internet] 50:1107–14. Available from: http://www.ncbi.nlm.nih.gov/pubmed/22353552 Pratte DG, Pfeiffer S. 1999. Histological age estimation of a cadaveral sample of diverse origins. J Can Soc Forensic Sci 32:155–167.

Price M. 1986. Health care as an instrument of Apartheid policy in South Africa. Health Policy Plan [Internet] 1:158–170. Available from: https://academic.oup.com/heapol/article- lookup/doi/10.1093/heapol/1.2.158 Randolph JF, Zheng H, Sowers MFR, Crandall C, Crawford S, Gold EB, Vuga M. 2011. Change in follicle-stimulating hormone and estradiol across the menopausal transition: Effect of age at the final menstrual period. J Clin Endocrinol Metab 96:746–754. Rauch F, Schoenau E. 2001. Changes in bone density during childhood and adolescence: an approach based on bone’s biological organization. J Bone Miner Res 16:597–604. Rauch F, Travers R, Glorieux FH. 2007. Intracortical remodeling during human bone development--a histomorphometric study. Bone 40:274–80. Reid SA, Boyde A. 1987. Changes in the mineral density distribution in human bone with age: image analysis using backscattered electrons in the SEM. J Bone Miner Res 2:13–22. van Rensburg HCJ, Benatar SR. 1993. The legacy of apartheid in health and health care. South African J Sociol 24:99–111. Richman EA, Ortner DJ, Schulter-Ellis FP. 1979. Differences in intracortical bone remodeling in three aboriginal American populations: Possible dietary factors. Calcif Tissue Int 28:209– 214. Riggs BL, Khosla S, Iii LJM. 2008. Estrogen , Bone Homeostasis , and Osteoporosis. In: Osteoporosis. 3rd Editio. Elsevier Inc. p 1011–1039. Riggs CM, Lanyon LE, Boyde A. 1993. Functional associations between collagen fibre orientation and locomotor strain direction in cortical bone of the equine radius. Anat Embryol 187:231–238. Robling AG, Castillo AB, Turner CH. 2006. Biomechanical and molecular regulation of bone remodeling. Annu Rev Biomed Eng [Internet] 8:455–98. Available from: http://www.ncbi.nlm.nih.gov/pubmed/16834564 Robling AG, Stout SD. 2008. Histomorphometry of Human Cortical Bone. In: Katzenberg MA, Saunders SR, editors. Biological Anthropology of the Human Skeleton. Second Edi. Wiley- Liss. p 149–171. Roschger P, Dempster DW, Zhou H, Paschalis EP, Silverberg SJ, Shane E, Bilezikian JP, Klaushofer K. 2007. New observations on bone quality in mild primary hyperparathyroidism as determined by quantitative backscattered electron imaging. J Bone Miner Res 22:717–723. Roschger P, Fratzl P, Eschberger J, Klaushofer K. 1998. Validation of quantitative backscattered electron imaging for the measurement of mineral density distribution in human bone biopsies. Bone 23:319–326. Roschger P, Paschalis EP, Fratzl P, Klaushofer K. 2008. Bone mineralization density distribution in health and disease. Bone 42:456–66. Ross AH, Baker LE, Falsetti A. 2003. Sexual dimorphism a proxy for environmental sensitivity? A multitemporal view. J Washingt Acad Sci 89:1–12. Ruff C. 2003a. Growth in bone strength, body size, and muscle size in a juvenile longitudinal sample. Bone 33:317–329. Ruff C. 2003b. Ontogenetic adaptation to bipedalism: age changes in femoral to humeral length and strength proportions in humans, with a comparison to baboons. J Hum Evol 45:317– 349. Sandoz B, Badina A, Lambot K, Mitton D, Skalli W. 2013. Quantitative geometric analysis of rib, costal cartilage and sternum from childhood to teenagehood. Med Biol Eng Comput 51:971–979.

Sanna a, Bertoli F, Misuri G, Gigliotti F, Iandelli I, Mancini M, Duranti R, Ambrosino N, Scano G. 1999. Chest wall kinematics and respiratory muscle action in walking healthy humans. J Appl Physiol 87:938–946. Schlebusch CM, Lombard M, Soodyall H. 2013. MtDNA control region variation affirms diversity and deep sub-structure in populations from southern Africa. BMC Evol Biol [Internet] 13:56. Available from: http://bmcevolbiol.biomedcentral.com/articles/10.1186/1471-2148-13-56 Schlebusch CM, Skoglund P, Sjödin P, Gattepaille LM, Hernandez D, Jay F, Li S, Jongh M De, Singleton A, Blum MGB, Soodyall H, Jakobsson M. 2012. Genomic Variation in Seven Khoe-San Groups Reveals Adaptation and Complex African History. Science (80- ) 338:374–379. Schoell SL, Weaver AA, Vavalle NA, Stitzel JD. 2015. Age- and Sex-Specific Thorax Finite Element Model Development and Simulation. Traffic Inj Prev [Internet] 16:S57–S65. Available from: http://www.tandfonline.com/doi/full/10.1080/15389588.2015.1005208 Schultz M. 2001. Paleohistopathology of bone: A new approach to the study of ancient diseases. Am J Phys Anthropol [Internet] 116:106–147. Available from: http://doi.wiley.com/10.1002/ajpa.10024 Schultz M, Schmidt-Schultz TH. 2015. Is it possible to diagnose TB in ancient bone using microscopy? Tuberculosis [Internet] 95:S80–S86. Available from: http://dx.doi.org/10.1016/j.tube.2015.02.035 Schuster SC, Miller W, Ratan A, Tomsho LP, Giardine B, Kasson LR, Harris RS, Petersen DC, Zhao F, Qi J, Alkan C, Kidd JM, Sun Y, Drautz DI, Bouffard P, Muzny DM, Reid JG, Nazareth L V., Wang Q, Burhans R, Riemer C, Wittekindt NE, Moorjani P, Tindall EA, Danko CG, Teo WS, Buboltz AM, Zhang Z, Ma Q, Oosthuysen A, Steenkamp AW, Oostuisen H, Venter P, Gajewski J, Zhang Y, Pugh BF, Makova KD, Nekrutenko A, Mardis ER, Patterson N, Pringle TH, Chiaromonte F, Mullikin JC, Eichler EE, Hardison RC, Gibbs RA, Harkins TT, Hayes VM. 2010. Complete Khoisan and Bantu genomes from southern Africa. Nature 463:943–947. Sedlin ED. 1964. The ratio of cortical area to total cross-section area in rib diaphysis: a quantitative index of osteoporosis. Clin Orthop 36:161–168. Sedlin ED, Frost HM, Villanueva BS. 1963a. Variations in cross-section area of rib cortex with age. J Gerontol 18:9–13. Sedlin ED, Frost HM, Villanueva BS. 1963b. Age changes in resorption in the human rib cortex. J Gerontol 18:345–349. Silva M, Alshamali F, Silva P, Carrilho C, Mandlate F, Jesus Trovoada M, Černý V, Pereira L, Soares P. 2015. 60,000 years of interactions between Central and Eastern Africa documented by major African mitochondrial haplogroup L2. Sci Rep [Internet] 5:12526. Available from: http://www.nature.com/articles/srep12526 Skedros JG. 2012. Interpreting load history in limb-bone diaphysis: important considerations and their biomechanical foundations. In: Crowder CM, Stout SD, editors. Bone Histology: an anthropological perspective. 1st Editio. Boca Raton: CRC Press. Skedros JG, Keenan KE, Williams TJ, Kiser CJ. 2013. Secondary osteon size and collagen/lamellar organization (“ osteon morphotypes”) are not coupled, but potentially adapt independently for local strain mode or magnitude. J Struct Biol [Internet] 181:95–107. Available from: http://dx.doi.org/10.1016/j.jsb.2012.10.013 Solomon L. 1968. Osteoporosis and fracture of the femoral neck in the South African Bantu. J Bone Jt Surg 50B:2–13.

Solomon L. 1979. Bone Density in Ageing Caucasian and African Populations. Lancet 314:1326–1330. Sowers MR, Zheng H, Greendale G a, Neer RM, Cauley J a, Ellis J, Johnson S, Finkelstein JS. 2013. Changes in bone resorption across the menopause transition: effects of reproductive hormones, body size, and ethnicity. J Clin Endocrinol Metab 98:2854–63. Steckel RH. 2005. Young adult mortality following severe physiological stress in childhood: skeletal evidence. Econ Hum Biol 3:314–28. Stein ID, Granik G. 1976. Rib structure and bending strength: an autopsy study. Calcif Tiss Res 20:61–73. Stinson S. 1985. Sex Differences in Environmental Sensitivity During Growth and Development. Yearb Phys Anthropol 28:123–147. Stout SD, Lueck R. 1995. Bone remodeling rates and skeletal maturation in three archaeological skeletal populations. Am J Phys Anthropol 98:161–171. Stout SD, Paine RR. 1992. Histological age estimation using the rib and clavicle. Am J Phys Anthropol 87:111–115. Stout SD, Porro M a., Perotti B. 1996. Brief communication: A test and correction of the clavicle method of Stout and Paine for histological age estimation of skeletal remains. Am J Phys Anthropol 100:139–142. Streeter MA. 2005. Histomorphometric characteristics of the subadult rib cortex : Normal patterns of dynamic bone modedling and remodeling during growth and development. Streeter MA, Stout SD. 2003a. The Histomorphometry Of The Subadult Rib: Age-Associated Changes In Bone Mass And The Creation Of Peak Bone Mass. In: Agarwal SC, Stout SD, editors. Bone Loss and Osteoporosis: An Anthropological Perspective. New York: Klewer Academic/Plenum Publishers. p 91–101. Streeter MA, Stout SD. 2003b. The Histomorphometry Of The Subadult Rib: Age-Associated Changes In Bone Mass And The Creation Of Peak Bone Mass. In: Agarwal SC, Stout SD, editors. Bone Loss and Osteoporosis: An Anthropological Perspective. 1st Editio. New York: Klewer Academic/Plenum Publishers. p 91–101. Takahashi H, Frost HM. 1966. Age and Sex Related Changes in the Amount of Cortex of Normal Human Ribs. Acta Orthop Scand [Internet] 37:122–130. Available from: http://www.tandfonline.com/doi/full/10.3109/17453676608993272 Takahashi N, Epker BN, Frost HM. 1965. Relation between age and size of osteons in man. Henry Ford Hosp Med Bull 13:25–31. Thandrayen K, Norris S a, Micklesfield LK, Pettifor JM. 2014. Fracture patterns and bone mass in South African adolescent-mother pairs: the Birth to Twenty cohort. Osteoporos Int [Internet] 25:693–700. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3906554&tool=pmcentrez&ren dertype=abstract Thandrayen K, Norris SA, Pettifor JM. 2009. Fracture rates in urban South African children of different ethnic origins: The Birth to Twenty cohort. Osteoporos Int 20:47–52. Theintz G, Buchs B, Rizzoli R, Slosman D, Clavien H, Sizonenko PC, Bonjour JP. 1992. Longitudinal monitoring of bone mass accumulation in healthy adolescents: evidence for a marked reduction after 16 years of age at the levels of lumbar spine and femoral neck in female subjects. J Clin Endocrinol Metab 75:1060–1065. Thomas CDL, Stein MS, Feik SA, Wark JD, Clement JG. 2000. Determination of age at death using combined morphology and histology of the femur. J Anat 196:463–471. Thompson DD. 1979. The core technique in the determination of age at death of skeletons. J

Forensic Sci 24:902–915. Tishkoff SA, Reed FA, Friedlaender FR, Ranciaro A, Froment A, Hirbo JB, Awomoyi AA, Bodo J, Doumbo O, Ibrahim M, Juma AT, Maritha J, Lema G, Moore JH, Mortensen H, Nyambo TB, Omar SA, Powell K, Pretorius GS, Smith MW, Thera A, Wambebe C, Weber JL, Williams SM. 2009. The Genetic Structure and History of Africans and African Americans. Science (80- ) 324:1035–1044. Tommerup LJ, Raab DM, Crenshaw TD, Smith EL. 1993. Does weight-bearing exercise affect non-weight-bearing bone? J Bone Miner Res 8:1053–1058. Tung S, Iqbal J. 2007. Evolution, aging, and osteoporosis. Ann N Y Acad Sci [Internet] 1116:499–506. Available from: http://www.ncbi.nlm.nih.gov/pubmed/18083942 Vansina J. 1995. New Linguistic Evidence and “the Bantu Expansion.” J Afr Hist 36:173–195. Vidulich L, Norris SA, Cameron N, Pettifor JM. 2011. Bone mass and bone size in pre- or early pubertal 10-year-old black and white South African children and their parents. Calcif Tissue Int 88:281–293. Walker ARP, Walker BF, Richardson BD. 1971. Metacarpal bone dimensions in young and aged South African Bantu consuming a diet low in calcium. Postgrad Med J 47:320–325. Walsh JS, Eastell R. 2013. Role of estrogen in the age-related decline in bone microstructure. J Clin Endocrinol Metab 98:519–21. Wang Q, Alén M, Nicholson P, Lyytikäinen A, Suuriniemi M, Helkala E, Suominen H, Cheng S. 2005. Growth patterns at distal radius and tibial shaft in pubertal girls: a 2-year longitudinal study. J Bone Miner Res 20:954–961. Wang Q, Seeman E. 2008. Skeletal growth and peak bone strength. Best Pract Res Clin Endocrinol Metab 22:687–700. Wetzsteon RJ, Zemel BS, Shults J, Howard KM, Kibe LW, Leonard MB. 2011. Mechanical loads and cortical bone geometry in healthy children and young adults. Bone 78:1103–1108. Williams DR, Gonzalez HM, Williams S, Mohammed S a, Moomal H, Stein DJ. 2008. Perceived discrimination, race and health in South Africa. Soc Sci Med 67:441–452. De Wit E, Delport W, Rugamika CE, Meintjes A, M??ller M, Van Helden PD, Seoighe C, Hoal EG. 2010. Genome-wide analysis of the structure of the South African Coloured Population in the Western Cape. Hum Genet 128:145–153. Wu K, Schubeck K, Frost HM, Villanueva A. 1970. Haversian bone formation rates determined by a new method in a Mastodon, and in human diabetes mellitus and osteoporosis. Calcif Tissue Res 6:204–219. Xiong J, Brien CAO. 2012. Osteocyte RANKL: New Insights Into the Control of. J bone Miner Res 27:499–505. Zebaze RM, Ghasem-Zadeh A, Bohte A, Iuliano-Burns S, Mirams M, Price RI, Mackie EJ, Seeman E. 2010. Intracortical remodelling and porosity in the distal radius and post-mortem femurs of women: a cross-sectional study. Lancet [Internet] 375:1729–1736. Available from: http://dx.doi.org/10.1016/S0140-6736(10)60320-0 Zeininger A, Richmond BG, Hartman G. 2011. Metacarpal head biomechanics: A comparative backscattered electron image analysis of trabecular bone mineral density in Pan troglodytes, Pongo pygmaeus, and Homo sapiens. J Hum Evol [Internet] 60:703–710. Available from: http://dx.doi.org/10.1016/j.jhevol.2011.01.002

Chapter 3 An Exploration of Body Size and Bone Mass on Cortical Bone Histomorphometry in Human Ribs

AMY C. BERESHEIM1*, SUSAN K. PFEIFFER1,2, AMANDA ALBLAS3

Affiliations: 1Department of Anthropology, University of Toronto, 19 Russell Street, Toronto ON, M5S 2S2; 2Department of Archaeology, University of Cape Town, Private Bag X3, Rhondebosch 7701; 3Division of Anatomy and Histology, Department of Biomedical Sciences, Faculty of Medicine and Health Sciences, Stellenbosch University, Cape Town 8000

Correspondence to: Amy C. Beresheim, MA, Department of Anthropology, University of Toronto, 19 Russell Street, Toronto, ON M5S 2S2. Phone: 647-459-2284. Email: [email protected]

Accepted for publication in The Anatomical Record (January 2018) 3.1 Abstract

This study examines the influence of human adult body size and bone mass on cortical bone histomorphometry, and explores microstructural variation in mid-thoracic ribs. The sample consists of 213 individuals (nfemale=82, nmale=131, mean age-at-death=47.96±15.71 years) from the Kirsten Skeletal Collection, Stellenbosch University, South Africa. Maximum femur length (MFL) and femur maximum head diameter (FHD) are used as proxies for height and weight; total cross-sectional area (Tt.Ar), endosteal area (Es.Ar) and cortical area (Ct.Ar) are used to derive measures of bone mass. Histomorphometric variables include osteon population density (OPD) and osteon area (On.Ar). Partial correlations, controlling for age, test for significant relationships among variables. A hierarchical regression model is used to determine unique variable contributions to On.Ar and OPD. Body size measurements do not correlate with either bone mass or histomorphometric variables, suggesting that size-standardization may not be necessary in studies of rib bone microstructure. Age is the most significant factor affecting OPD, while OPD is the best predictor of On.Ar. These findings suggest that age-related secondary osteon crowding affects osteon geometry. Understanding the biological mechanisms that direct bone remodeling and determine microstructural variation is essential for interpreting histological data.

Key words: adult body size, bone mass, histomorphometry, bone remodeling, osteon population density (OPD)

3.2 Introduction

Adult body size affects various aspects of gross skeletal morphology, such that short individuals with low body mass tend to be under-aged and tall individuals with high body mass tend to be over-aged when standard age-at-death methods are applied (Merritt, 2015). Although osteon population density (OPD) and osteon area (On.Ar) are common variables in histological age-at- death estimation (Stout and Paine, 1992; Cho et al., 2002), and are sometimes used to infer diet and/or activity levels in past populations (Burr et al., 1990; Abbott et al., 1996; Paine and Brenton, 2006; Pfeiffer et al., 2006; Fahy et al., 2017), few studies have considered possible body size effects on cortical bone microstructure. The relationship between adult body size and OPD remains largely unexplored. Studies of weight-loss in older adults show increased bone turnover (Jensen et al., 2001; Shah et al., 2011), suggesting an inverse correlation between adult body size and OPD. Some investigators have identified significant relationships between body size and On.Ar, but with conflicting results. Burr (1992) found a positive correlation between body weight and On.Ar in immature macaque ribs and femora, but Britz and colleagues (2009) reported the opposite relationship at the femoral midshaft location among human adults. Maat and colleagues (2006) argued that body size does not significantly influence histological age-at- death prediction when using anterior wedges of the femoral midshaft.

The relationship between body size and histomorphometry may best be interpreted in terms of the biomechanical loading environment. Peak strain levels are remarkedly consistent across species (Rubin and Lanyon, 1984), largely because different bone morphologies produce different mechanical advantages (Biewener, 1991). However, weight gain can increase bone mass and alter bone microstructure by inducing an adaptive response to increased mechanical loads (Iwaniec and Turner, 2016). Although not always the case, osteon densities tend to be most concentrated in compression cortices where strain magnitudes are typically highest (Mason et al., 1995; Skedros et al., 2004, 2009). High strain is also thought to have an inhibitory effect on the resorptive activity of osteoclasts, thus limiting the size of forming osteons and On.Ar (van Oers et al., 2008). Studies of unloading in human bone are consistent with these trends. Prolonged weightlessness in space flight is associated with bone mineral loss and osteoporosis across the entire skeleton (Leblanc et al., 2007). In the absence of weight-bearing activity, such as in cases of amputation or paraplegia, long bones appear to demonstrate low osteon population densities and larger osteon areas (Lazenby and Pfeiffer, 1993; Schlecht et al., 2012).

Interactions among these variables are not well understood, nor is it clear how bone mass contributes to microstructural variability. At least at some assessment locations, OPD is negatively correlated with total cross-sectional area (Tt.Ar) and cortical area (Ct.Ar), while On.Ar demonstrates the inverse relationship with these measurements (Ural and Vashishth, 2006). Robust bones (i.e. wide relative to length) tend to have larger and more numerous osteons compared to more slender bones (Goldman et al., 2014). There is also a trend for larger osteons in larger bones (Frost et al., 1987; Pfeiffer, 1998; Pfeiffer et al., 2006; Goldman et al., 2014; Dominguez and Agnew, 2016; Goliath et al., 2016), but this may be explained by differences in their absolute cortical areas.

Histomorphometric studies of the axial and appendicular skeleton are often incongruous, reflecting different adaptations to metabolic and mechanical stimuli. The ribs are constrained by the mechanics of respiration and seldom experience large weight-bearing loads. They are believed to be more representative of systemic bone turnover, and have thus been a favored assessment site in anthropological studies of bone metabolism and aging (Robling and Stout, 2003; Agnew and Stout, 2012; Skedros et al., 2013; Eleazer and Jankauskas, 2016). The ribs have significantly smaller relative cortical areas compared to the long bones of the upper and lower limb, suggesting that they do not demonstrate the same allometric effect with body size (Stewart et al., 2015). Nonetheless, small body size is a known risk factor for osteoporotic rib fractures (Margolis et al., 2000; Compston et al., 2014). Ribcage size has also been positively correlated with height, weight, and body mass index (BMI) (Gayzik et al., 2008).

The objectives of this research are to explore the influence of adult body size and bone mass on bone histomorphometry, and to investigate histomorphometric relationships at the midshaft location of human ribs. Previous studies demonstrate an increase in OPD and a decrease in On.Ar with advancing age, but sex-based differences remain less straightforward (Stout and Paine, 1992; Cho et al., 2002, 2006; Cho and Stout, 2011; Goliath et al., 2016; Pfeiffer et al., 2016). We hypothesize that after controlling for age- and sex-related variation, there will be lower OPD values and smaller On.Ar measurements in small-bodied individuals with small bone areas, as compared to larger individuals.

3.3 Materials and Methods

3.3.1 Research Sample

The research sample is drawn from the Kirsten Skeletal Collection, curated at the Faculty of Medicine and Health Sciences, Stellenbosch University, South Africa. The collection is relatively representative of population demography of the Western Cape Province during the apartheid era (1948-1994 AD). Sex, age-at-death, cause-of-death, and government “race” designation are available from departmental records. Our sample, like the collection, is mainly comprised of South African Coloured adults. Our sampling protocol was designed to maximize the number of men and women in each 10-year age cohort. Targets were sometimes under met because of limited representation within the collection, or because of incomplete skeletal inventories. Priority for inclusion was given to individuals with relatively complete thoracic cages, such that ribs could be sided effectively and seriated (Dudar, 1993). One mid-thoracic rib (R5-R7) was selected from each individual based on cross-sectional completeness at the midshaft location. Ribs that exhibited antemortem trauma or pathological changes were excluded. Although most histological methods have been developed for the 6th rib, research indicates that cortical bone parameters are consistent across analogous sampling locations in adjacent ribs (Crowder and Rosella, 2007).

Associated body size measurements were collected following methods of Buikstra and Ubelaker (1994). Left side measurements were selected when both paired elements were present. In this study, maximum femur length (MFL) is used to represent height, while femur maximum head diameter (FHD) is used as a proxy for body weight (Merritt, 2015). While imperfect, these measures are often used in the study of past human populations. Limb bone length correlates well with stature, so that MFL can be used in predictive equations when living stature is unknown (Trotter and Gleser, 1952, 1958; Feldesman and Fountain, 1996). The femoral head grows to counteract the weight that is applied to it, such that its adult size should be proportional to body mass at the time of epiphyseal fusion (Ruff et al., 1997). FHD is a highly reproducible measurement; values from known-mass individuals have been used to derive body mass estimation formulae (Ruff et al., 1991; McHenry, 1992; Grine et al., 1995). It is particularly valuable in anthropological research because the femoral head is frequently preserved in ancient specimens.

Morphological characteristics of the sample are presented in Table 1. Data are given as mean ± standard deviation (SD) if normally distributed, and median ± interquartile range (IQR) if non- normally distributed. Compared to world-wide variation in body size, the Kirsten Skeletal Collection likely represents people who were of average height and weight (Auerbach and Ruff, 2004, 2006), but exhibited relatively little sexual size dimorphism given non-significant sex differences in MFL and FHD. Sexual size dimorphism [SSD = ln(male/female)] was calculated as 0.013 and 0.024, for MFL and FHD, respectively (Smith, 1999). Individuals in this sample are larger than archaeological populations from southern Africa, but have smaller indices of sexual dimorphism (SSD = 0.021 and SSD =0.083 for MFL and FHD, respectively) (Kurki, 2011). While we cannot fully explain these differences, we hypothesize that they may be related to differential growth stunting in modern South African children. Boys are more commonly underweight (Kimani-Murage et al., 2010), and more likely to engage in risky drug and alcohol use (Yach and Tollman, 1993; Flisher et al., 1996; Peltzer et al., 2011; Naude et al., 2012), possibly resulting in smaller adult body size in men. Although the sample is not sexually dimorphic in terms of body size, men have significantly larger cross-sectional rib area measurements than women.

3.3.2 Histological Preparation, Imaging, and Analysis

Histological preparation followed Crowder et al. (2012). Midshaft tissue samples were embedded in plastic under vacuum, after being washed in an ultrasonic bath. Thick-sections were cut and ground to a final thickness of 50 to100µm using a Buehler Isomet precision saw and a Buehler Ecomet grinding wheel. Sections were pressed until dry, and then mounted with glass coverslips using a toluene-based solution.

Virtual slides of transverse rib cross-sections were prepared as previously described (Pfeiffer et al., 2016). An Olympus BX-40 microscope with automated stage and Olympus Stream image- capture software were used to create linearly polarized light (LPL) image montages of each rib cross-section (Figure 5). Macroscopic geometrical properties include variables related to the bone surfaces (periosteal and endosteal), and microstructural parameters are limited to those most commonly employed in histological age-at-death estimation. Variables were either measured directly or calculated as follows (Cho et al., 2002, 2006):

1. Total area (Tt.Ar) = total cross-sectional/subperiosteal area in mm2 2. Endosteal area (Es.Ar) = area of the medullary cavity in mm2 3. Cortical area (Ct.Ar) = total bone area between the periosteal and endosteal borders of the rib. Ct.Ar is calculated as Tt.Ar – Es.Ar in mm2. 4. Relative cortical area (Rt.Ct.Ar) = the percentage of cortical bone in relation to total subperiosteal area. Rt.Ct.Ar is calculated as (Ct.Ar/Tt.Ar) 5. Intact osteon density (N.On.) = number of secondary osteons with an intact Haversian canal and bounded by a scalloped reversal line in #/mm2 6. Fragmentary osteon density (N.On.Fg) = number of secondary osteons with a partially visible Haversian canal that has been breached either by a neighboring osteon or a resorptive bay, and secondary osteons with no remnant of a Haversian canal present in #/mm2 7. Osteon population density (OPD) = (N.On + N.Fg.On/Ct.Ar) in #/mm2 8. Osteon Area (On.Ar) = area of bone contained within the cement line of a complete osteon in mm2. A complete osteon is defined as having an intact Haversian canal and a clear reversal line, thus excluding fragmentary osteons. A minimum of 25 intact osteons were measured, excluding variants such as drifting osteons3. The mean On.Ar for each histological cross-section was used in all analyses.

3 While osteon variants such as drifting osteons were avoided (Skedros et al., 2007), circular osteons with circular Haversian canals were not preferentially selected given recent research by Hennig and colleagues (2015). Using three-dimensional (3D) micro-CT renderings of the vascular canal network, they showed that osteons tend to have elliptical cross-sections, and that 3D osteon orientation was not strongly correlated with 2D osteon circularity.

Table 2: Descriptive statistics for the study sample and independent samples t-test between men and women. Significant values are in bold.

Males Females Pooled t-test

Mean ± SD / Mean ± SD / Mean ± SD / Variable n Median (IQR) Range n Median (IQR) Range n Median (IQR) Range t p

OPD (#/mm2) 131 20.01 ± 5.45 6.50 - 34.79 82 19.80 ± 5.06 8.21 - 31.40 213 19.21 ± 5.30 6.50 - 34.79 0.27 0.787 On.Ar (mm2) 131 0.034 (0.009) 0.018 - 0.043 82 0.034 (0.010) 0.020 - 0.049 213 0.034 (0.009) 0.18 - 0.043 0.628 0.531 Tt.Ar (mm2) 131 69.93 (20.16) 34.33 - 109.11 82 53.80 (16.08) 29.41 - 108.86 213 62.75 (23.11) 29.11 - 109.11 6.243 <0.001 Es.Ar (mm2) 131 44.31 (18.57) 10.47 - 88.72 82 30.36 (18.04) 8.62 - 86.92 213 39.73 (21.23) 8.62 - 88.72 5.988 <0.001 Ct.Ar (mm2) 131 24.33 (8.21) 11.39 - 48.80 82 22.24 (9.84) 10.60 - 47.08 213 23.98 (9.22) 11.39 - 48.80 1.751 0.081 Rt.Ct.Ar (mm2) 131 0.35 (0.12) 0.15 - 0.69 82 0.41 (0.20) 0.13 - 0.75 213 0.37 (0.16) 0.13 - 0.75 -3.696 <0.001 Age (years) 131 49.29 ± 15.84 17 - 79 82 45.83 ± 15 18 - 82 213 47.96 ± 15.71 17 - 82 1.570 0.118 MFL (mm) 90 438.00 ± 32.30 375 - 507 61 44.5.07 ± 28.67 380 - 505 152 440.65 ± 30.99 375 - 505 -1.444 0.151 FHD (mm) 90 43.78 ± 4.03 33.7 - 51.8 61 44.02 ± 4.00 33.0 - 52.0 151 43.89 ± 4.01 33.3 - 52.0 -0.841 0.402

3.3.3 Statistical Analysis

All statistical analyses were performed using SPSS 23. Model assumptions were graphically evaluated (histograms, scatter plots, and quantile-quantile plots). Non-normally distributed variables were transformed per the Box-Cox technique (Osborne, 2010), and the normalized data were used in all further analyses. Independent t-tests were used to detect sex-specific differences in the variable distributions. Partial correlations, controlling for age, determined significant relationships between variables in men, women, and the pooled sample. Multicollinearity assumptions were tested using the variance inflation factor (VIF) statistic. Hierarchical multiple regression (HMR) models were then created with OPD and On.Ar set as the dependent variables, evaluating all independent variables that had biologic plausibility and were significant in the correlation analysis. Cook’s distance was used to check for outliers in the data. In each HMR, age and sex were entered as covariates in the first analysis block. Because significant sex- differences for bone mass measurements were identified (Table 2), sex (0=male, 1=female) was included in the HMR models. The second block contained all the other variables of interest. Significance for all tests was set at p<0.05.

3.4 Results

Partial correlations are presented in Table 3. Although our body size measurements are all highly correlated with each other (data not shown), they do not demonstrate statistically significantly relationships with either bone mass or histomorphometric variables. In both sexes and in the pooled sample, OPD is positively correlated with Ct.Ar and Rt.Ct.Ar, but not Tt.Ar or Es.Ar. Similarly, On.Ar is negatively correlated with Ct.Ar. and Rt.Ct.Ar, but not Tt.At or Es.Ar. OPD and On.Ar are also highly negatively correlated with each other (Figure 6).

Results of the hierarchical multiple regressions are reported in Table 4. For the HMR examining OPD, age and sex were set as predictors in the first block of the analysis. This model is significant (F (2, 210)=43.398; p<0.001)), explaining 29.2% of the total variance in OPD. When the other variables were added into Block 2, the resulting model is also significant (F (5, 207)=30.890; p<0.001), explaining an additional 13.5% of the variance in OPD (F(3, 207)=16.249; p<0.001). Age, Ct.Ar, and On.Ar all demonstrate significant contributions to the model. Age is the most important predictor, uniquely explaining 15.4% of the variance in OPD.

Table 3: Partial correlations controlling for age in men, women, and the pooled sample. Significant values are in bold.

Variable Men Women Pooled OPD df r p df r p df r p On.Ar 128 -0.415 <0.001 79 -0.410 <0.001 210 -0.403 <0.001 Tt.Ar 128 -0.035 0.691 79 -0.126 0.262 210 -0.87 0.205 Es.Ar 128 0.091 0.304 79 0.026 0.821 210 0.038 0.205 Ct.Ar 128 -0.276 0.001 79 -0.376 0.001 210 -0.316 <0.001 Rt.Ct.Ar 128 -0.242 0.006 79 -0.247 0.026 210 -0.218 0.001 MFL 88 0.123 0.249 58 -0.226 0.083 149 0.004 0.960 FHD 88 -0.052 0.627 58 -0.090 0.494 149 -0.068 0.410 On.Ar Tt.Ar 128 0.096 0.277 79 -0.080 0.477 210 0.051 0.459 Es.Ar 128 -0.078 0.381 79 -0.214 0.055 210 -0.096 0.162 Ct.Ar 128 0.441 <0.001 79 0.291 0.008 210 0.391 <0.001 Rt.Ct.Ar 128 0.347 <0.001 79 0.324 0.003 210 0.315 <0.001 MFL 88 -0.046 0.670 58 0.207 0.112 149 0.026 0.75 FHD 88 0.011 0.919 58 0.034 0.799 149 0.007 0.932 Tt.Ar Es.Ar 128 0.917 <0.001 79 0.904 <0.001 210 0.923 <0.001 Ct.Ar 128 0.413 <0.001 79 0.408 <0.001 210 0.431 <0.001 Rt.Ct.Ar 128 -0.462 <0.001 79 -0.512 <0.001 210 -0.525 <0.001 MFL 88 0.047 0.663 58 -0.089 0.500 149 -0.042 0.606 FHD 88 0.149 0.162 58 -0.085 0.516 149 0.053 0.522 Es.Ar Ct.Ar 128 0.021 0.816 79 -0.012 0.913 210 0.059 0.396 Rt.Ct.Ar 128 -0.773 <0.001 79 -0.823 <0.001 210 0.535 <0.001 MFL 88 0.074 0.490 58 -0.101 0.441 149 -0.031 0.703 FHD 88 0.183 0.084 58 -0.101 0.442 149 -0.015 0.851 Ct.Ar Rt.Ct.Ar 128 0.612 <0.001 79 0.568 <0.001 210 0.535 <0.001 MFL 88 -0.046 0.664 58 0.041 0.758 149 -0.024 0.77 FHD 88 -0.040 0.711 58 0.054 0.684 149 -0.015 0.851 Rt.Ct. Ar MFL 88 -0.08 0.456 58 0.105 0.425 149 0.019 0.816 FHD 88 -0.178 0.093 58 0.102 0.438 149 -0.073 0.370

Figure 6: Scatterplot showing the relationship between OPD and On.Ar in men (orange circles) and women (red triangles)

Age and sex were also set as predictors in the first block of the HMR examining On.Ar. This model is significant (F(2, 210)=12.767; p<0.001)), explaining 10.8% of the total variance in On.Ar. When the other variables are added into Block 2, the resulting model is significant (F(5, 207)=20.571; p<0.001), explaining an additional 22.4% of the variance in On.Ar (F(3, 207)=23.089; p<0.001). Ct.Ar, Rt.Ct.Ar, and OPD all demonstrate significant contributions to the model. OPD is the most important predictor, uniquely explaining 7.3% of the variance in On.Ar.

Table 4 Hierarchical multiple regression output for OPD and On.Ar R R2 R2 change β t Sig sr2 OPD Block 1 0.541 0.292 Age 0.544 9.311 <0.001 Sex 0.040 0.682 0.496 Block 2 0.654 0.427 0.135 Age 0.392 6.698 <0.001 0.154 Sex -0.003 -0.057 0.954 0.000 On.Ar -0.292 -4.773 <0.001 0.063 Ct.Ar -0.156 -2.194 0.029 0.013 Rt.Ct.Ar -0.018 -0.292 0.809 0.000 On.Ar Block 1 0.329 0.108 Age -0.328 -5.009 <0.001 Sex -0.078 -1.198 0.232 Block 2 0.576 0.332 0.224 Age -0.020 -0.284 0.777 0.000 Sex -0.068 -1.091 0.276 0.004 OPD -0.34 -4.773 <0.001 0.073 Ct.Ar 0.199 2.594 0.010 0.022 Rt.Ct.Ar 0.163 2.056 0.041 0.014

3.5 Discussion

3.5.1 Body Size

Contrary to expectation, body size measurements do not covary with bone mass variables. These relationships tend to be more pronounced in weight-bearing than in non-weight-bearing bones (Edelstein and Barrett-Connor, 1993; McNeil et al., 2009), potentially explaining the lack of significant associations in our study. It is possible that these results are specific to our study sample, given the poor health status of most individuals in the Kirsten Skeletal Collection. Systemic effects on the variables of interest could be particularly marked on the rib since it is non-load-bearing. Cho and Stout (2011) report positive linear relationships between Tt.Ar, Es.Ar, Ct.Ar and estimated body mass in both the femur and rib, but do give descriptive statistics of their raw measurements to allow for study comparisons. Low sexual dimorphism could have also obfuscated any discernable patterns, as men with comparable body size to women have been shown to have larger skeletal size and bone mass (Nieves et al., 2004). Greater variation may be needed to appreciate any body size effects.

Our results do not support a causal relationship between body mass, OPD, and On.Ar in the ribs. It is possible that the bones of the appendicular and axial skeleton demonstrate differential responses to body size and inferred bone strain. The inverse relationship between body mass and On.Ar is likely to be bone specific, and even then it explains less than 10% of the overall variance in femoral On.Ar (Britz et al., 2009). Compared to weight-bearing bones, higher remodeling rates and significantly smaller relative cortical areas imply that these differences may be less apparent in the ribs (Epker et al., 1965; Stewart et al., 2015). In this light, our results suggest that it may not be necessary to employ size-standardization in studies of histological structures of the rib. This alleviates some research concerns because these data are not commonly available for existing histological collections, nor in many archaeological and forensic contexts.

One limitation of this exploration of body size on bone histomorphometry is that body mass and stature in life are not known; our measurements based on gross bone morphology can only serve as proxies. While informative, measurements of long bones cannot account for substantial changes to body mass that may have occurred between the cessation of skeletal growth and the death of the person. Our study also does not control for anticipated sexual dimorphism in body composition, such as differences in muscle mass and body fat distributions. Body composition is thought to contribute significantly to variability in bone structure and bone mineralization density (BMD). Skeletal muscle mass is positively correlated with BMD (Sun et al., 2003; Lee et al., 2016), but site-specific fat deposits are thought to exert differing effects on bone. Large amounts of subcutaneous fat have been associated with compromised bone strength, apparently because of increased bone remodeling, in adolescent females, but fat appears to have a protective effect on cortical bone mass in female adults (Dimitri et al., 2012). Visceral fat has been negatively associated with BMD and trabecular bone quality indicators across the entire lifespan (Cohen et al., 2013; Júnior et al., 2013). Given the socioeconomic circumstances of the persons who make up the Kirsten Skeletal Collection (Labuschagne and Mathey, 2000; Pfeiffer et al., 2016), obesity is likely to have been very rare (Puoane et al., 2002).

3.5.2 Osteon Population Density (OPD)

Consistent with previous research, age is the strongest predictor of OPD. The cortex becomes crowded with complete and fragmentary osteons with age. It has been observed that this

54 crowding eventually creates an OPD asymptote when evidence of earlier remodeling events is completely removed (Stout and Lueck, 1995). Older individuals with marked endosteal expansion have been shown to demonstrate greater episodic osteon clustering, suggesting that changes to cortical area also partially modulate osteon density (Heinrich, 2015). Our results show that OPD is highly correlated with Ct.Ar and Rt.Ct.Ar, but variation in Rt.Ct.Ar does not uniquely contribute to OPD prediction after controlling for age. Regardless of age, the absolute amount of cortical bone is an important factor for limiting the number of secondary osteons seen in a rib bone cross-section.

OPD is only weakly associated with habitual loading, but may be an important factor in fracture resistance within aging bone (Skedros, 2012). Bone micro-damage accumulation increases exponentially with age (Schaffler et al., 1995), faster than remodeling maintenance efforts can repair. While bone is poorly equipped to minimize micro-crack formation (Reilly and Currey, 1999), evidence suggests that it effectively abates the rate and length of micro-crack propagation. Increased OPD corresponds to a higher density of cement lines, which may serve as effective barriers to micro-crack growth (O’Brien et al., 2005; Gibson et al., 2006). Increased OPD may also enhance osteon de-bonding, bridging, and pullout by dissipating energy and preventing ultimate failure (Hiller et al., 2003; Dong and Guo, 2004; Bigley et al., 2006).

3.5.3 Osteon Area (On.Ar)

In our study, there is a significant negative correlation between OPD and On.Ar, consistent with observations by Schlecht et al. (2012) and Miszkiewicz (2016). OPD is the greatest predictor of On.Ar, suggesting that osteon crowding is an important factor for determining secondary osteon size. Yeni et al. (1997) found that osteon density positively correlates with fracture toughness, but that osteon size is inversely related to this property. Therefore, smaller osteon size may improve the fatigue properties of bone simply by facilitating increased osteon density (Britz et al., 2009).

OPD uniquely explains only a small proportion of the variation in On.Ar, indicating that other factors need to be considered. Our results support the assertion that large bones allow for larger osteons, but the local biomechanical environment likely has superseding influence. Britz et al. (2009) proposed that smaller osteons may be beneficial in aging cortices because they reduce the size of temporary defects (i.e. resorption spaces) that might contribute to catastrophic failure.

High intracortical porosity lowers the ultimate failure point in bone (Ural and Vashishth, 2007), but this explanation is insufficient, as On.Ar and cortical porosity do not appear to be directly related (Dominguez and Agnew, 2016). Regionally-specific collagen fiber orientations, mineral content, osteocyte lacunar properties, and other metric characteristics of secondary osteons all merit further investigation (Skedros, 2012).

3.5.4 Other Considerations

OPD does not vary between the pleural and cutaneous rib cortices (Gocha et al., 2016), but osteon clustering and geometric properties may be related to their proximity to the cortical bone surfaces. (Heinrich, 2015) observed a bimodal distribution of osteons along the bone cortex, with

56 less pronounced intracortical remodeling, and a maximum number of osteons closer to the periosteal than the endosteal envelope. Larger osteons are more likely to be located closer to the endosteal surface, but On.Ar is highly variable and a poor predictor of relative periosteal distance (Heinrich, 2015). While these phenomena may be explained by strain gradients, Skedros et al. (1997, 2001) have previously argued that large endosteal osteon size is more likely related to calcium mobilization near the bone marrow.

Our deduction that increased body mass is linearly related to increased bone strain may be too simple. The habitual load complexity of midshaft ribs is not known. Further, a causal relationship between strain and small osteon size is difficult to explain given empirical evidence of aging bone. While decreased On.Ar with age is well established (Currey, 1964; Takahashi et al., 1965; Jowsey, 1966; Evans, 1976; Yoshino et al., 1994; Han et al., 2009; Dominguez and Agnew, 2016; Pfeiffer et al., 2016), it seems unlikely that older bones are categorically subject to higher strain. It has been argued that the age-dependent decrease in On.Ar might be an artifact of our data collection methods (Takahashi et al., 1965). By virtue of their size, larger osteons are more likely to be remodeled and excluded from On.Ar measurements as OPD increases. Nonetheless, secondary osteons with complete cement lines still represent the most recent remodeling events in a given cortex, and the rarity of large osteonal variants in adult bone suggests that the smaller On.Ar values are real (Skedros et al., 2007).

3.6 Conclusion

Given that a clear relationship between adult body size and histomorphometric variation is not obvious, size standardization of histomorphometric data likely will not improve the interpretation of results, at least when these values are generated from the rib. This study supports the longstanding assumption that the rib is an ideal bone to study systemic effects independent of loading usually related to body weight and bone size. Given the relationships documented here, OPD and On.Ar continue to be useful for studies of age- and sex-related variation.

3.7 Acknowledgements

The authors thank Linda Greyling of Stellenbosch University, and Virginia Pichler, Klara Komza, Meimei Fong and Jarred Heinrich at University of Toronto.

3.8 Literature Cited Abbott S, Trinkaus E, Burr DB. 1996. Dynamic Bone Remodeling in Later Pleistocene Fossil Hominids. Am J Phys Anthropol 99:585–601. Agnew AM, Stout SD. 2012. Brief communication: Reevaluating osteoporosis in human ribs: The role of intracortical porosity. Am J Phys Anthropol 148:462–466. Auerbach BM, Ruff CB. 2004. Human body mass estimation: A comparison of “morphometric” and “mechanical” methods. Am J Phys Anthropol 342:331–342. Auerbach BM, Ruff CB. 2006. Limb bone bilateral asymmetry: variability and commonality among modern humans. J Hum Evol 50:203–18. Biewener AA. 1991. Musculoskeletal design in relation to body size. J Biomech 24:19–29. Bigley RF, Griffin L V, Christensen L, Vandenbosch R. 2006. Osteon interfacial strength and histomorphometry of equine cortical bone. J Biomech 39:1629–40. Britz HM, Thomas CDL, Clement JG, Cooper DML. 2009. The relation of femoral osteon geometry to age, sex, height and weight. Bone [Internet] 45:77–83. Available from: http://dx.doi.org/10.1016/j.bone.2009.03.654 Buikstra JE, Ubelaker DH. 1994. Standards for data collection from human skeletal remains. :1– 272. Burr DB. 1992. Estimated intracortical bone turnover in the femur of growing macaques: Implications for their use as models in skeletal pathology. Anat Rec 189:180–189. Burr DB, Ruff CB, Thompson DD. 1990. Patterns of skeletal histologic change through time: Comparison of an archaic Native American population with modern populations. Anat Rec 226:307–313. Cho H, Stout SD. 2011. Age-associated bone loss and intraskeletal variability in the Imperial Romans. J Anthropol Sci [Internet] 89:109–25. Available from: http://www.ncbi.nlm.nih.gov/pubmed/21368344 Cho H, Stout SD, Bishop TA. 2006. Cortical bone remodeling rates in a sample of African American and European American descent groups from the American midwest: Comparisons of age and sex in ribs. Am J Phys Anthropol 130:214–226. Cho H, Stout SD, Madsen RW, Streeter MA. 2002. Population-specific histological age- estimating method: A model for known African-American and European-American skeletal remains. J Forensic Sci 47:12–18. Cohen A, Dempster DW, Recker RR, Lappe JM, Zhou H, Zwahlen A, Müller R, Zhao B, Guo X, Lang T, Saeed I, Liu XS, Guo XE, Cremers S, Rosen CJ, Stein EM, Nickolas TL, Mcmahon DJ, Young P, Shane E. 2013. Abdominal fat is associated with lower bone formation and inferior bone quality in healthy premenopausal women: A transiliac bone biopsy study. J Clin Endocrinol Metab 98:2562–2572. Compston JE, Flahive J, Hosmer DW, Watts NB, Siris ES, Silverman S, Saag KG, Roux C, Rossini M, Pfeilschifter J, Nieves JW, Netelenbos JC, March L, Lacroix AZ, Hooven FH, Greenspan SL, Gehlbach SH, Díez-Pérez A, Cooper C, Chapurlat RD, Boonen S, Anderson F a., Adami S, Adachi JD. 2014. Relationship of weight, height, and body mass index with fracture risk at different sites in postmenopausal women: The global longitudinal study of osteoporosis in women (GLOW). J Bone Miner Res 29:487–493. Crowder C, Heinrich JT, Stout SD. 2012. Rib histomorphometry for adult age estimation. Methods Mol Biol 915:109–127. Crowder C, Rosella L. 2007. Assessment of intra- and intercostal variation in rib histomorphometry: its impact on evidentiary examination. J Forensic Sci 52:271–6. Currey JD. 1964. Some effects of ageing in human Haversian systems. J Anat 98:69–75. Dimitri P, Bishop N, Walsh JS, Eastell R. 2012. Obesity is a risk factor for fracture in children but is protective against fracture in adults : A paradox. Bone [Internet] 50:457–466. Available from: http://dx.doi.org/10.1016/j.bone.2011.05.011 Dominguez VM, Agnew AM. 2016. Examination of factors potentially influencing osteon size in the human rib. Anat Rec 299:313–324. Dong XN, Guo XE. 2004. Geometric determinants to cement line debonding and osteonal lamellae failure in osteon pushout tests. J Biomech Eng 126:387–390.

Dudar JC. 1993. Identification of rib number and assessment of intercostal variation at the sternal rib end. J Forensic Sci 38:788–797. Edelstein SL, Barrett-Connor E. 1993. Relation between body size and bone mineral density in elderly men and women. Am J Epidemiol 138:160–169. Eleazer CD, Jankauskas R. 2016. Mechanical and metabolic interactions in cortical bone development. Am J Phys Anthropol 160:317–333. Epker BN, Kelin M, Frost HM. 1965. Magnitude and location of cortical bone loss in human rib with aging. Clin Orthop Relat Res 41:198–203. Evans FG. 1976. Mechanical properties and histology of cortical bone from younger and older men. Anat Rec 185:1–11. Fahy GE, Deter C, Pitfield R, Miszkiewicz JJ, Mahoney P. 2017. Bone deep: Variation in stable isotope ratios and histomorphometric measurements of bone remodelling within adult humans. J Archaeol Sci 87:10–16. Feldesman MR, Fountain RL. 1996. “Race” specificity and the femur/stature ratio. Am J Phys Anthropol 100:207–224. Flisher AJ, Ziervogel CF, Chalton DO, Leger PH, Robertson BA. 1996. Risk-taking behaviour of Cape Peninsula high-school students. South African Med J 86:1090–1093. Frost HM, Colorado S, Pueblo C, Lafayette W. 1987. Secondary osteon populations: An algorithm for determining mean bone tissue age. Yearb Phys Anthropol 30:221–238. Gayzik FS, Yu MM, Danelson KA, Slice DE, Stitzel JD. 2008. Quantification of age-related shape change of the human rib cage through geometric morphometrics. J Biomech 41:1545–54. Gibson VA, Stover SM, Gibeling JC, Hazelwood SJ, Martin RB. 2006. Osteonal effects on elastic modulus and fatigue life in equine bone. J Biomech [Internet] 39:217–25. Available from: http://www.ncbi.nlm.nih.gov/pubmed/16321623 Gocha TP, Dominguez VM, Agnew AM. 2016. Spatial patterning (or lack thereof) in osteon population density in the human rib. Am J Phys Anthropol 159:S156. Goldman HM, Hampson NA, Guth JJ, Lin D, Jepsen KJ. 2014. Intracortical remodeling parameters are associated With measures of bone robustness. Anat Rec 297:1817–1828. Goliath JR, Stewart MC, Stout SD. 2016. Variation in osteon histomorphometrics and their impact on age-at-death estimation in older individuals. Forensic Sci Int [Internet] 262:282.e1-282.e6. Available from: http://dx.doi.org/10.1016/j.forsciint.2016.02.053 Grine FE, Jungers WL, Tobias P V, Pearson OM. 1995. Fossil Homo femur from Berg Aukas, northern Namibia. Am J Phys Anthropol 185:151–185. Han S-H, Kim S-H, Ahn Y-W, Huh G-Y, Kwak D-S, Park D-K, Lee U-Y, Kim Y-S. 2009. Microscopic age estimation from the anterior cortex of the femur in Korean adults. J Forensic Sci 54:519–522. Heinrich JT. 2015. Spatial Characterization of Rib Cortical Bone Microstructure and the Effect of Nutritional and Physiological Stresses. Hennig C, Thomas CDL, Clement JG, Cooper DML. 2015. Does 3D orientation account for variation in osteon morphology assessed by 2D histology? J Anat 227:497–505. Hiller LP, Hazelwood SJ, Yeh OC, Martin RB, Veterinury JDW, Duais D. 2003. Osteon pullout in the equine third metacarpal bone: effects of ex vivo fatigue. J Orthop Res 21:481–488. Iwaniec UT, Turner RT. 2016. Influence of body weight on bone mass, architecture, and turnover. J Endocrinol 230:R115–R130. Jacoby WG. 2000. Loess: A nonparametric, graphical tool for depicting relationships between variables. Elect Stud 19:577–613. Jensen LB, Kollerup G, Quaade F, SØRensen OH. 2001. Bone Mineral Changes in Obese Women During a Moderate Weight Loss With and Without Calcium Supplementation. J Bone Miner Res [Internet] 16:141–147. Available from: http://doi.wiley.com/10.1359/jbmr.2001.16.1.141 Jowsey J. 1966. Studies of Haversian systems in man and some animals. J Anat [Internet] 100:857–864. Available from: http://link.springer.com/10.1007/s00223-015-9957-9 Júnior IFF, Cardoso JR, Christofaro DGD, Codogno JS, César A, Moraes F De, Fernandes RA. 2013. The relationship between visceral fat thickness and bone mineral density in sedentary

obese children and adolescents. BMC Pediatr 13:37. Kimani-Murage EW, Kahn K, Pettifor JM. 2010. The prevalence of stunting, overweight and obesity, and metabolic disease risk in rural South African children. BMC public [Internet]:1–13. Available from: http://www.biomedcentral.com/1471-2458/10/158 Kurki HK. 2011. Pelvic dimorphism in relation to body size and body size dimorphism in humans. J Hum Evol [Internet] 61:631–643. Available from: http://dx.doi.org/10.1016/j.jhevol.2011.07.006 Labuschagne BCJ, Mathey B. 2000. Cadaver profile at University of Stellenbosch Medical School, South Africa, 1956 – 1996. Clin Anat 13:88–93. Lazenby RA, Pfeiffer SK. 1993. Effects of a nineteenth century below‐knee amputation and prosthesis on femoral morphology. Int J Osteoarchaeol 3:19–28. Leblanc AD, Spector ER, Evans HJ, Sibonga JD. 2007. Skeletal responses to space flight and the bed rest analog: A review. J Musculoskelet Neuronal Interact 7:33–47. Lee JE, Lee SR, Song H. 2016. Muscle mass is a strong correlation factor of total hip BMD among Korean premenopausal women. Osteoporos Sarcopenia [Internet] 2:99–102. Available from: http://dx.doi.org/10.1016/j.afos.2016.04.001 Maat GJR, Maes A, Aarents MJ, Nagelkerke NJD. 2006. Histological age prediction from the femur in a contemporary Dutch sample. The decrease of nonremodeled bone in the anterior cortex. J Forensic Sci 51:230–237. Margolis KL, Ensrud KE, Schreiner PJ, Tabor HK. 2000. Body size and risk for clinical fractures in older women. Ann Intern Med 133:123–127. Mason MW, Skedros JG, Bloebaum RD. 1995. Evidence of strain-mode-related cortical adaptation in the diaphysis of the horse radius. Bone 17:229–237. McHenry HM. 1992. Body size and proportions in early hominids. Am J Phys Anthropol 87:407–431. McNeil CJ, Raymer GH, Doherty TJ, Marsh GD, Rice CL. 2009. Geometry of a weight-bearing and non-weight-bearing bone in the legs of young, old, and very old men. Calcif Tissue Int 85:22–30. Merritt CE. 2015. The influence of body Size on adult skeletal age estimation methods. Am J Phys Anthropol 57:35–57. Miszkiewicz JJ. 2016. Investigating histomorphometric relationships at the human femoral midshaft in a biomechanical context. J Bone Miner Metab [Internet]:179–192. Available from: http://dx.doi.org/10.1007/s00774-015-0652-8 Naude CE, Carey PD, Laubscher R, Fein G, Senekal M. 2012. Vitamin D and calcium status in South African adolescents with alcohol use disorders. Nutrients [Internet] 4:1076–1094. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3448088&tool=pmcentrez&ren dertype=abstract Nieves JW, Formica C, Ruffing J, Zion M, Garrett P, Lindsay R, Cosman F. 2004. Males Have Larger Skeletal Size and Bone Mass Than Females, Despite Comparable Body Size. J Bone Miner Res [Internet] 20:529–535. Available from: http://doi.wiley.com/10.1359/JBMR.041005 O’Brien FJ, Taylor D, Lee TC. 2005. The effect of bone microstructure on the initiation and growth of cracks. J Orthop Res 23:475–480. van Oers RFM, Ruimerman R, van Rietbergen B, Hilbers PAJ, Huiskes R. 2008. Relating osteon diameter to strain. Bone 43:476–82. Osborne JW. 2010. Improving your data transformations : Applying the Box-Cox transformation. Pract Assessment, Res Eval 15:1–9. Paine RR, Brenton BP. 2006. Dietary health does affect histological age assessment: An evaluation of the Stout and Paine (1992) age estimation equation using secondary osteons from the rib. J Forensic Sci 51:489–492. Peltzer K, Davids A, Njuho P. 2011. Alcohol use and problem drinking in South Africa: findings from a national population-based survey. Afr J Psychiatry [Internet] 14:30–37. Available from: http://dx.doi.org/10.4314/ajpsy.v14i1.65466 Pfeiffer S. 1998. Variability in Osteon Size in Recent Human Populations. Am J Phys Anthropol

106:219–227. Pfeiffer S, Crowder C, Harrington L, Brown M. 2006. Secondary osteon and Haversian canal dimensions as behavioral indicators. Am J Phys Anthropol 131:460–468. Pfeiffer S, Heinrich J, Beresheim A, Alblas M. 2016. Cortical bone histomorphology of known- age skeletons from the Kirsten collection, Stellenbosch University, South Africa. Am J Phys Anthropol 160:137–147. Puoane T, Steyn K, Bradshaw D, Laubscher R, Fourie J, Lambert V, Mbananga N. 2002. Obesity in South Africa: the South African demographic and health survey. Obes Res 10:1038–1048. Reilly GC, Currey JD. 1999. The development of microcracking and failure in bone depends on the loading mode to which it is adapted. J Exp Biol 202:543–552. Robling AG, Stout SD. 2003. Histomorphology, geometry, and mechanical loading in past populations. In: Agarwal S, Stout SD, editors. Bone Loss and Osteoporosis: An Anthropological Perspective. 1st editio. New York: Klewer Academic/Plenum Publishers. p 207–228. Rubin CT, Lanyon LE. 1984. Regulation of bone formation by applied dynamic loads. J Bone Jt Surg [Internet] 66:397–402. Available from: http://www.ncbi.nlm.nih.gov/pubmed/6699056 Ruff CB, Scott WW, Liu AY-C. 1991. Articular and diaphyseal remodeling of the proximal femur with changes in body mass in adults. Am J Phys Anthropol 8:397–413. Ruff CB, Trinkaus E, Holliday TW. 1997. Body mass and encephalization in Pleistocene Homo. Nature 387:173–176. Schaffler MB, Choi K, Milgrom C. 1995. Aging and matrix microdamage accumulation in human compact bone. Bone 17:521–525. Schlecht SH, Pinto DC, Agnew AM, Stout SD. 2012. Brief communication: the effects of disuse on the mechanical properties of bone: what unloading tells us about the adaptive nature of skeletal tissue. Am J Phys Anthropol [Internet] 149:599–605. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23086658 Shah K, Armamento-villareal R, Parimi N, Chode S, Sinacore DR, Hilton TN, Napoli N, Qualls C, Villareal DT. 2011. Exercise training in obese older adults prevents increase in bone turnover and attenuates hormones. J Bone Miner Res 26:2851–2859. Skedros JG. 2012. Interpreting load history in limb-bone diaphysis: important considerations and their biomechanical foundations. In: Crowder CM, Stout SD, editors. Bone Histology: an anthropological perspective. 1st Editio. Boca Raton: CRC Press. Skedros JG, Hunt KJ, Bloebaum RD. 2004. Relationships of loading history and structural and material characteristics of bone: development of the mule deer calcaneus. J Morphol [Internet] 259:281–307. Available from: http://www.ncbi.nlm.nih.gov/pubmed/14994328 Skedros JG, Knight AN, Clark GC, Crowder CM, Dominguez VM, Qiu S, Mulhern DM, Donahue SW, Busse B, Hulsey BI, Zedda M, Sorenson SM. 2013. Scaling of Haversian canal surface area to secondary osteon bone volume in ribs and limb bones. Am J Phys Anthropol [Internet] 151:230–44. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23633395 Skedros JG, Mason MW, Bloebaum RD. 2001. Modeling and remodeling in a developing artiodactyl calcaneus: A model for evaluating Frost’s Mechanostat Hypothesis and its corollaries. Anat Rec 185:167–185. Skedros JG, Mendenhall SD, Kiser CJ, Winet H. 2009. Interpreting cortical bone adaptation and load history by quantifying osteon morphotypes in circularly polarized light images. Bone [Internet] 44:392–403. Available from: http://www.ncbi.nlm.nih.gov/pubmed/19049911 Skedros JG, Sorenson SM, Jenson NH. 2007a. Are distributions of secondary osteon variants useful for interpreting load history in mammalian bones? Cells Tissues Organs [Internet] 185:285–307. Available from: http://www.ncbi.nlm.nih.gov/pubmed/17587802 Skedros JG, Sorenson SM, Jenson NH. 2007. Are distributions of secondary osteon variants useful for interpreting load history in mammalian bones? Cells Tissues Organs 185:285– 307. Skedros JG, Su SC, Bloebaum RD. 1997. Biomechanical implications of mineral content and microstructural variations in cortical bone of horse, elk, and sheep calcanei. Anat Rec

316:297–316. Smith RJ. 1999. Statistics of sexual size dimorphism. J Hum Evol 36:423–459. Stewart MC, Goliath JR, Stout SD, Hubbe M. 2015. Intraskeletal variability of relative cortical area in humans. Anat Rec 298:1635–1643. Stout SD, Lueck R. 1995. Bone remodeling rates and skeletal maturation in three archaeological skeletal populations. Am J Phys Anthropol 98:161–171. Stout SD, Paine RR. 1992. Histological age estimation using the rib and clavicle. Am J Phys Anthropol 87:111–115. Sun AJ, Heshka S, Heymsfield SB, Wang J, Gallagher D. 2003. Is there an association between skeletal muscle mass and bone mineral density among African-American , Asian-American , and European-American women ? Acta Diabetol 40:S309–S313. Takahashi N, Epker BN, Frost HM. 1965. Relation between age and size of osteons in man. Henry Ford Hosp Med Bull 13:25–31. Trotter M, Gleser GC. 1952. Estimation of stature from long bones of American Whites and Negros. Am J Phys Anthropol 10:463–514. Trotter M, Gleser GC. 1958. A re-evaluation of estimation of stature based on measurements of stature taken during life and of long bones after death. Am J Phys Anthropol 16:79–123. Ural A, Vashishth D. 2006. Interactions between microstructural and geometrical adaptation in human cortical bone. J Orthop Res 24:1489–1498. Ural A, Vashishth D. 2007. Effects of intracortical porosity on fracture toughness in aging human bone: A µCT-based cohesive finite element study. J Biomech Eng 129:625–631. Yach D, Tollman SM. 1993. Public health initiatives in South Africa in the 1940s and 1950s: Lessons for a post-apartheid era. Am J Public Health 83:1043–1050. Yeni YN, Brown CU, Wang Z, Norman TL. 1997. The influence of bone morphology on fracture toughness of the human femur and tibia. Bone 21:453–459. Yoshino M, Imaizumi K, Miyasaka S, Seta S. 1994. Histological estimation of age at death using microradiographs of humeral compact bone. Forensic Sci Int 64:191–198.

Chapter 4 Sex-specific Patterns in Cortical and Trabecular Bone Microstructure in the Kirsten Skeletal Collection, South Africa

AMY C. BERESHEIM1*, SUSAN K. PFEIFFER1,2, MARC D. GRYNPAS3,4, AMANDA ALBLAS5

Institutions: 1Department of Anthropology, University of Toronto, 19 Russell Street, Toronto, Canada M5S 2S2; 2Department of Archaeology, University of Cape Town, Private Bag X3, Rondebosch, South Africa 7701 3Department of Laboratory Medicine and Pathobiology and Institute for Biomaterials and Biomedical Engineering, University of Toronto, 60 Murray Street, Box 42, Toronto, Canada M5T 3L9 4Lunenfeld-Tanenbaum Research Institute, Mount Sinai Hospital, 25 Orde Street, Suite 417, Toronto, Canada M5T 3H7 5Division of Anatomy and Histology, Department of Biomedical Sciences, Stellenbosch University, P.O. Box 241, Cape Town, South Africa 8000

*Correspondence to: Amy Beresheim, Department of Anthropology, University of Toronto, 19 Russell Street, Toronto, Canada M5S 2S2. Phone: 47-459-2284 E-mail: [email protected]

Grant Sponsors: Pilot Research Funding, University of Toronto, Department of Anthropology (AB); Social Sciences and Humanities Research Council of Canada (SP)

Accepted in the American Journal of Human Biology (January 2018) 4.1 Abstract

Objectives: This study provides bone histomorphometric reference data for South Africans of the Western Cape who likely dealt with health issues under the apartheid regime.

Methods: The 206 adult individuals (nfemale=75, nmale=131, mean=47.9±15.8 years) from the Kirsten Skeletal Collection, U. Stellenbosch, lived in the Cape Town metropole from the late 1960s to the mid-1990s. To study age-related changes in cortical and trabecular bone microstructure, photomontages of mid-thoracic rib cross-sections were quantitatively examined. Variables include relative cortical area (Rt.Ct.Ar), osteon population density (OPD), osteon area (On.Ar), bone volume fraction (BV/TV), trabecular number (Tb.N), trabecular thickness (Tb.Th), and trabecular spacing (Tb.Sp).

Results: All cortical variables demonstrate significant relationships with age in both sexes, with women showing stronger overall age associations. Peak bone mass was compromised in some men, possibly reflecting poor nutritional quality and/or substance abuse issues throughout adolescence and early adulthood. In women, greater predicted decrements in On.Ar and Rt.Ct.Ar suggest a structural disadvantage with age, consistent with post-menopausal bone loss. Age- related patterns in trabecular bone microarchitecture are variable and difficult to explain. Except for Tb.Th, there are no statistically significant relationships with age in women. Men demonstrate significant negative correlations between BV/TV, Tb.N, and age, and a significant positive correlation between Tb.Sp and age.

Conclusions: This research highlights sex-specific differences in patterns of age-related bone loss, and provides context for discussion of contemporary South African bone health. While the study sample demonstrates indicators of poor bone quality, osteoporosis research continues to be under-prioritized in South Africa.

Key words: Histomorphometry; bone loss and fragility; osteoporosis; South Africa; apartheid

4.2 Introduction

Overshadowed by HIV/AIDS, tuberculosis, malnutrition, and alcoholism, osteoporosis is not yet recognized as a major health problem in South Africa (Bateman, 2006). Although the National Osteoporosis Foundation of South Africa (NOFSA) has published guidelines on prevention and treatment (Hough et al., 2010), local data on the incidence of osteoporosis and associated fragility fractures is limited. Extrapolating from international statistics, an estimated 1.4 million females and 0.6 million males over 50 years currently suffer from osteoporosis (International Osteoporosis Foundation, 2011). Within the next two decades the oldest South African cohort is expected to exponentially increase, contributing to the burden of disease (International Osteoporosis Foundation, 2011).

Osteoporosis is typically diagnosed using dual energy X-ray absorptiometry (DXA), and is defined as a bone mineral density (BMD) that is 2.5 standard deviations or more below the young adult mean of a population. Established reference values from Europe and North America are of limited utility in South Africa because of compositional variation between populations, as well as substantial differences in climate, diet, and lifestyle (Micklesfield, Norris, & Pettifor,

2011). While there has been a recent push to generate BMD data specific to the South African population (Chantler et al., 2012; Conradie, Conradie, Kidd, & Hough, 2014), studies remain scarce, and are largely focused on postmenopausal bone loss in white women. However, compounded by structural violence and complex disease interactions, age-related decrements to bone quality are likely exacerbated in some subsets of the South African population. Men from historically disadvantaged groups within the apartheid political system [A.D. 1948-1994] have been (Grusin & Samuel, 1957; Lynch et al., 1967; Lynch, Seftel, Wapnick, Charlton, & Bothwell, 1970; Seftel et al., 1966; Wapnick, Lynch, Seftel, Charlton, & Jowsey, 1971), and may continue to be, the most vulnerable.

4.2.1 Apartheid and the Social Determinants of Health

Social, economic, and political institutions were structured along legally defined racial categories in apartheid South Africa. Under the Population Registration Act of 1950, people were arbitrarily classified into the South African Black (SAB), South African Coloured (SAC), South African White (SAW), and South African Asian/Indian (SAI) population groups. SABs, often referred to as Bantu, are genetically distinct from the West Africans that dominate the African-American gene pool (Patin et al., 2017; Silva et al., 2015; Tishkoff et al., 2009). The coloured designation is especially contentious as it was used as a catch-all for people who could not effectively be sorted into the classification system. The SAC represents a highly admixed population group with complex origins, and includes indigenous peoples of Khoesan and Griqua ancestry. A substantial part of this gene pool derives from Europeans that settled in the Cape after 1652, as well as from slaves imported from South East Asia and West Africa (Daya et al., 2013; de Wit et al., 2010). SAWs are typically described as having European ancestry, and SAIs are considered descendants of Indian indentured laborers (Inwood & Masakure, 2013).

The Group Areas Act of 1950 enforced the separation of these population groups. Although the story of racial inequality in South Africa is often framed in a black and white dichotomy, all non- white population groups (including SABs, SACs, and SAIs) were subject to race-based legislation (Inwood & Masakure, 2013). Non-whites were required to move to peripheral areas with inadequate housing, poor infrastructure, and limited access to education and healthcare. Institutionalized inequalities created disparities in the distribution of health and disease, as well as differential dietary and substance abuse patterns among population groups (Andersson &

Marks, 1988; Nightingale et al., 1990; van Rensburg & Benatar, 1993). Environmental conditions can have a significant impact on bone growth and degradation, affecting aspects such as bone size and density (Curate, 2014; Saunders, 2008). Despite the transition to democracy since 1994, racial residential segregation and socioeconomic differences persist (Ataguba, Day, & McIntyre, 2015; Harris et al., 2011). While the etiology of osteoporosis is multifactorial and complex, the legacy of apartheid undoubtedly continues to influence bone quality outcomes.

4.2.2 Histomorphometry

Bone mass and density have been clinically assessed using various imaging methods. While DXA is most commonly employed, its application to dry bone is controversial because of possible bone diagenesis (Agarwal, 2008; Agarwal & Grynpas, 1996; Kneissel et al., 1994; Mays, 1996). The DXA method cannot provide a true measure of bone geometry, distinguish cortical from trabecular bone compartments, or account for tissue-level heterogeneity (Beck, 2003). While destructive and not suitable to all historical collections, histomorphometry is considered the “gold standard” for bone metabolic and mineralization evaluation (Vidal et al., 2012). Although bone histomorphometry is most easily examined in the deceased, it can be applied to bone biopsies from the living. It is a useful alternative to DXA because the bone compartments can be analyzed separately and individual bone remodeling events (i.e. intracortical porosity, secondary osteons) are quantifiable. Rather than assessing BMD, histomorphometry is used to examine bone quality, specifically the contribution of bone microstructure to bone strength and fracture resistance (Fyhrie, 2005).

4.2.3 Study Objectives

A histomorphometric reference dataset will be helpful in the exploration of age-related variation in bone quality outcomes across adulthood in South African men and women. Cross-sectional data from the South African apartheid era can be used to investigate the extent of compromised bone health during this time, and to create an age trajectory supporting the prediction of age- related changes in cortical and trabecular bone microstructure. It is our hope that this research will ultimately be useful for understanding epidemiological transitions and bone health in post- apartheid South Africa, particularly for those living in the Western Cape Province.

4.3 Materials and Methods

4.3.1 Research Sample

The research sample consists of mid-thoracic ribs representing 206 adult skeletons (nfemale=75, nmale=131, mean age at death=47.9±15.8 years) from the Kirsten Skeletal Collection (Table 5; Figure 8a-b). Curated by the Division of Anatomy and Histology at Stellenbosch University Faculty of Medicine and Health Sciences in Parow, South Africa, it is a collection of skeletonized cadavers that were originally received for teaching dissections through a body donation program. Donor sources are mostly limited to the northern suburbs of the Cape Town and surrounding rural towns, and although a few were bequeathed, most cadavers were received as unclaimed bodies from local hospitals and government mortuaries. Roughly half of the cadavers used in dissections were skeletonized by technical staff for inclusion into the collection, which is currently used for both teaching and research purposes (Alblas, Greyling, & Geldenhuys, 2018; Labuschagne & Mathey, 2000). The collecting period spans from A.D.1956 to present day, with most of the skeletonized cadavers dating to the apartheid era. All individuals included in this study have information on government-designated “race”, sex, age-at-death, and cause-of-death. While we cannot ensure the accuracy of departmental records, the official focus on personal identification during the apartheid era suggests reliable documentation.

The composition of the skeletonized cadaver profile dramatically differs from that of the entire country, but more closely reflects population demographics of the Western Cape during the main period of collecting (Labuschagne & Mathey, 2000). Within the research sample, approximately 62% of the skeletons are classified as SAC, 22% as SAB, and 16 % as SAW. The percentage of females tends to be lower when unclaimed bodies are used (Gangata, Ntaba, Akol, & Louw, 2010), and the Kirsten Skeletal Collection is no exception with a male-to-female sex-ratio of roughly 3:1. The number of SAB females in this sample is especially low.

Table 5: Sample structure according to sex, age, and population group. Men Women Group n mean SD n mean SD SAC 71 47.8 16.6 56 43.6 15.0 SAB 38 47.5 15.6 7 36.7 11.4 SAW 22 57.3 11.4 12 58.7 12.1

Figure 8a-b: Sample structure according to sex, age, and population group. SAC in blue, SAB in green, and SAW in beige.

Dissection cadavers are disproportionally from underprivileged groups (Halperin, 2007). While racial inequalities are irrefutable, given the major sources of body donations, we presume that most whites in the sample were not fully shielded from the ill-effects of systematic marginalization and rapid industrialization. The socioeconomic status of most of the people represented in the Kirsten Skeletal Collection was low, implying marginal to poor housing, healthcare, and employment experiences. Townships often lacked adequate infrastructure and community facilities. Overcrowded households, absent streetlighting, unpaved roads, and sporadic waste removal services were typical. In the Cape Town Municipal area, coloured townships were initially developed without telephones, post or police offices, churches, community or sports centers, and in some cases, even without schools. The primary social gathering places were shebeens– the term for informal pubs (Theron, 1976).

One reflection of substandard living conditions is the high prevalence of pulmonary tuberculous (TB) in the SAB and SAC of the Western Cape (Labuschagne & Mathey, 2000). Individuals with a history of active TB have higher incidence rates of osteoporosis and fragility fractures (Chen et al., 2017). TB has been found to disproportionately affect the ribs (Steyn et al., 2013), such that indicators of poor bone quality may be especially marked in the study sample.

4.3.2 Sample Selection and Tissue Processing

Size and demographic features of the sample are detailed in Figure 8a-b and Table 5. Sampling procedures and tissue processing followed methods described elsewhere (Pfeiffer, Heinrich, Beresheim & Alblas, 2016). Briefly, mid-thoracic ribs were collected using a stratified sampling protocol based on population group, sex, and age. The target was to include 15 samples of each sex per 10-year age increment for each population group (i.e. 20–29 years, 30–39 years. . . 70–79 years). This target was sometimes unmet because of limited representation within the collection or insufficient rib tissue associated with a skeleton. Rib samples were only selected if they were R5, R6 or R7, had a complete cross-section at the mid-shaft, and did not exhibit any evidence of ante-mortem trauma or disease.

At the University of Toronto, mid-shaft samples were embedded in an epoxy resin under vacuum. Thick sections were cut from these prepared blocks using a Buehler Isomet precision saw, and ground to approximately 100 µm with a Buehler Ecomet grinding wheel. Thin sections

69 were polished using a diamond suspended paste, and then adhered to glass slides and coverslips using a toluene-based solution.

4.3.3 Ethics Statement

Written permission for tissue harvesting and transport to the University of Toronto was granted by Dr. Benedict J. Page, Head of the Division of Anatomy and Histology, as well as by the Western Cape Government Inspector of Anatomy. Research clearance was granted by the University of Toronto Research Ethics Board.

4.3.4 Data Collection

Photomontages of mid-thoracic rib cross-sections were generated using an Olympus BX-41 light microscope, an Olympus SC30 camera, a PriorOptiScan II automated stage, and Olympus cellSens software (v. 1.9). High-resolution images were captured at 100x magnification to create virtual slides of each sample under both bright field (BF) and linearly polarized light (LPL). To study age-related changes in cortical and trabecular bone microarchitecture, LPL virtual slides were quantitatively examined using two image-analysis software programs. Histomorphometric measurements for cortical bone were collected using Olympus cellSens software with a digitizing tablet, following definitions outlined by Cho, Stout, Madsen, and Streeter (2002). Osteon area (On.Ar), total area (Tt.Ar), and endosteal area (En.Ar) were manually traced. Relative cortical area (Rt.Ct.Ar) was derived from these measurements by subtracting En.Ar from Tt.Ar, and dividing by Tt.Ar (Tt.Ar - Es.Ar/Tt.Ar). Osteon population density (OPD) was calculated as the sum of intact and fragmentary osteons in each rib cross-section divided by the cortical area (Tt.Ar – Es.Ar = Ct.Ar). Histomorphometric parameters for trabecular bone were derived using the Bioquant Osteo II digitizing system (R&M Biometrics, Nashville, TN). Variables include bone volume density (BV/TV), trabecular thickness (Tb.Th), trabecular number (Tb.N), and trabecular spacing (Tb.Sp; Dempster et al., 2013). A plate model was used to estimate Tb.Th (Parfitt et al., 1983).

4.3.5 Statistical Analysis

70 apartheid race categories in biomedicine is contested (Erasmus, 2012), and we made the decision to omit them in our statistical analyses. These categorizations may obfuscate underlying biological factors that contribute to differential bone remodeling. Forensic validation studies performed on South African samples tend to underestimate known age-at-death (Paine & Brenton, 2006; Pratte & Pfeiffer, 1999), and the accuracy of these methods does not seem to improve when race-based prediction equations are employed (Pfeiffer, Heinrich, Beresheim, & Alblas, 2016). Studies of admixture in the SAC population also provide arguments against arbitrary genetic characterizations in clinical research (Choudhury et al., 2017; Daya et al., 2013; de Wit et al., 2010). Differences between young men and women (<35 years of age) were assessed using a t-test or Wilcoxon’s rank-sum test depending on variable distributions (Table 7). The overall relationship between age and histomorphometric parameters was investigated through either Pearson or Spearman correlations (Table 7).

Age-related changes in histomorphometric parameters and the differences between men and women were examined using linear regression. Model assumptions were checked graphically (histograms, scatter plots, and quantile-quantile plots), and variable transformations were performed per box-cox analysis where warranted (Osborne, 2010). Regression models tested for possible nonlinear associations by incorporating a quadratic age term (age2). To address issues of collinearity, “age-centered” (age – mean age) and “age-centered squared” were employed. “Sex”, “sex- and age-centered”, and “sex- and age-centered squared” interaction terms were included to assess sex-specific comparisons with age. Model fit was assessed using R2 coefficients, graphical analysis of residuals, and the Akaike information criterion (AIC). Predicted values were used to estimate age-related changes across the lifespan for both sexes. We report the estimated mean and 95% prediction interval at age 80 years. Similar statistical approaches have been utilized in population-based HR-pQCT studies (Dalzell et al., 2009; Hansen, Shanbhogue, Folkestad, Nielsen, & Brixen, 2014; Khosla et al., 2006; Macdonald, Nishiyama, Kang, Hanley, & Boyd, 2011).

While some researchers have attempted to size-standardize the structural properties of rib cross- sections (Eleazer & Jankauskas, 2016), we consider this step unwarranted because ribs do not demonstrate the same allomeric relationship to body size as the long bones of the upper and lower limbs (Stewart, Goliath, Stout, & Hubbe, 2015). Thus, only unadjusted values of bone area are reported. All statistical tests were performed in either Microsoft Excel or SPSS v. 21.

Table 6: Descriptive statistics (mean ± SD or median (interquartile range)) for each histomorphometric variable are given by age cohort for men and women.

Men Age 2 2 Cohort n OPD (#/mm ) On.Ar (mm ) Rt.Ct.Ar (%) BV/TV (%) Tb.Th (mm) Tb.N (#/mm) Tb.Sp (mm) ≤19 years 8 10.12 ± 3.58 0.037 (0.014) 0.47 (0.18) 0.17 (0.11) 0.09 (0.03) 1.73 ± 0.44 0.47 (0.34) 20-29 9 18.21 ± 5.10 0.034 (0.014) 0.39 (0.10) 0.19 (0.08) 0.10 (0.03) 1.64 ± 0.45 0.53 (0.23) 30-39 23 17.82 ± 4.20 0.034 (0.012) 0.41 (0.20) 0.14 (0.04) 0.11 (0.03) 1.41 ± 0.40 0.63 (0.25) 40-49 23 19.43 ± 3.98 0.036 (0.007) 0.37 (0.12) 0.15 (0.07) 0.09 (0.02) 1.51 ± 0.42 0.51 (0.29) 50-59 37 21.31 ± 4.85 0.034 (0.008) 0.32 (0.96) 0.12 (0.05) 0.09 (0.03) 1.29 ± 0.32 0.65 (0.22) 60-69 19 23.56 ± 4.87 0.033 (0.011) 0.31 (0.11) 0.14 (0.06) 0.10 (0.02) 1.44 ± 0.41 0.56 (0.32) 70+ 12 23.59 ± 3.96 0.034 (0.008) 0.33 (0.13) 0.15 (0.07) 0.10 (0.02) 1.34 ± 0.33 0.62 (0.18)

Women Age 2 2 Cohort n OPD (#/mm ) On.Ar (mm ) Rt.Ct.Ar (%) BV/TV (%) Tb.Th (mm) Tb.N (#/mm) Tb.Sp (mm) ≤19 years 1 8.41 0.042 0.49 0.16 0.15 1.06 0.80 20-29 17 16.10 ± 4.11 0.037 (0.005) 0.51 (0.16) 0.16 (0.08) 0.11 (0.27) 1.51 ± 0.48 0.59 (0.43) 30-39 12 19.12 ± 5.39 0.037 (0.009) 0.51 (0.16) 0.14 (0.07) 0.11 (0.01) 1.50 ± 0.57 0.65 (0.29) 40-49 16 19.49 ± 2.91 0.037 (0.009) 0.47 (0.19) 0.14 (0.09) 0.09 (0.02) 1.53 ± 0.40 0.55 (0.34) 50-59 12 22.03 ± 3.70 0.030 (0.004) 0.34 (0.13) 0.16 (0.04) 0.10 (0.01) 1.60 ± 0.33 0.51 (0.20) 60-69 14 23.93 ± 4.38 0.029 (0.005) 0.36 (0.13) 0.14 (0.11) 0.09 (0.03) 1.26 ± 0.47 0.66 (0.50) 70+ 3 24.89 ± 2.34 0.028 (N/A) 0.28 (N/A) 0.16 (N/A) 0.12 (N/A) 1.30 ± 0.16 0.64 (N/A)

Table 7: Observed outcomes at <35 years (mean ± SD or median (interquartile range)), and predicted outcomes at 80 years (estimated mean (95% prediction interval)).

Men Women

Predicted Change Between 20 and 80 Years Predicted Change Between 20 and 80 Years

Sex Observed Estimated Mean Abs. % Observed Estimated Mean Abs. % Diff in R2 c R2 c Sex x (<35 years) (PI) at 80 Years Diff Change (<35 years) (PI) at 80 Years Diff Change Young Age e Adultsd Cortical n = 30 n = 23

OPDb (#/mm2) 15.85 ± 5.69 25.87 (17.11-34.64) 10.02 63.2 0.557** 16.81 ± 5.12 26.51 (16.88-36.14) 12.23 72.8 0.587** 0.522 0.469

On.Ara (mm2) 0.036 (0.009) 0.031 (0.021-0.045) -0.01 -13.4 -0.200* 0.037 (0.006) 0.017 (0.010-0.274) -0.02 -54.3 -0.543** 0.221 0.025

Rt.Ct.Ara, b (%) 0.39 (0.14) 0.32 (0.16-0.53) -0.06 -15.5 -0.341** 0.51 (0.16) 0.14 (0.04-0.29) -0.31 -60.1 -0.527** 0.004 0.001

Trabecular

BV/TVa (%) 0.14 (0.08) 0.12 (0.05-0.23) -0.05 -34.7 -0.212* 0.16 (0.05) 0.13 (0.05-0.26) -0.08 -49.3 -0.143 0.866 0.853

Tb.Tha (mm) 0.10 (0.03) 0.10 (0.07-0.15) -0.004 -4.2 -0.061 0.11 (0.03) 0.08 (0.06-0.12) -0.03 -26.6 -0.228* 0.155 0.107

Tb.N (#/mm) 1.55 ± 0.45 1.24 (0.45-2.03) -0.31 -20.0 -0.250** 1.48 ± 0.45 1.61 (0.74-2.48) 0.134 -9.3 -0.107 0.533 0.361

Tb.Spa (mm) 0.53 (0.27) 0.71 (0.38-1.76) 0.18 34.5 0.211* 0.61 (0.40) 0.49 (0.28-0.49) -0.12 -19.4 0.107 0.612 0.417

4.4 Results

Summarized histomorphometric parameters in young men and women (<35 years) and predicted age-related changes are shown in Table 7. Regression plots are shown in Figures 9a-f and 10a-h.

4.4.1 Cortical Bone Histomorphometry

All cortical variables demonstrate significant relationships with age in both sexes, with women showing stronger overall age-associations. OPD increases with age, while On.Ar and Rt.Ct.Ar decrease with age. Correlation coefficients are low to moderate. OPD and On.Ar do not differ between young adult men and women, but Rt.Ct.Ar is significantly greater in young adult women. OPD and Rt.Ct.Ar show significant quadratic associations with age. Between 20 and 80 years, the predicted increase in OPD is similar in men and women. However, the sex x age interaction is significant for On.Ar and Rt.Ct.Ar. On.Ar decreases by 54.3% in women, compared to 13.4% in men. Similarly, Rt.Ct.Ar decreases by 60.1% in women, compared to 15.5% in men. Women exhibit greatest On.Ar and Rt.Ct.Ar during young adulthood (24 years and 31 years, respectively), and undergo accelerated decreases after 50 years of age. Men approach a maximum OPD value in their late 70s, but OPD values continue to increase in women. No asymptote was observed in women.

4.4.2 Trabecular Bone Histomorphometry

Linear models provide the best fit for all trabecular parameters. Trabecular bone indices are similar in young adult men and women, and all sex by age interactions are nonsignificant. While all trabecular parameters are significantly correlated with age in the total sample (p > 0.05), these relationships are maintained to varying degrees in the male and female subsamples. Most correlation coefficients are low, reflecting broadly scattered data. When men and women are analyzed separately, the positive correlations between BV/TV, Tb.N, and Tb.Sp with age are attenuated in the female sample. Although the correlation with age is non-significant, the regression model predicts that BV/TV decreases by 49.3% over the course of the female lifespan. Tb.Th demonstrates a significant 26.6% decline from young adulthood to 80 years. With the exception of a few outliers, Tb.Th appears remarkably consistent across the male lifespan. BV/TV and Tb.N both demonstrate a significant negative correlation with age in men. BV/TV

74 decreases by 34.7% and Tb.N decreases by 20.0% over the course of the male lifespan. Tb.Sp. is positively correlated with age in men, increasing by 34.5% from young adulthood to 80 years.

4.5 Discussion

This research provides novel data on bone health in understudied South African populations, for which there are few epidemiological studies available. As it is a cross-sectional study, age- related changes in bone quality outcomes are estimated rather than observed. Our data suggest that patterns of age-related bone loss differ between women and men, and between the cortical and trabecular compartments of mid-thoracic ribs. However, some caution is necessary in interpreting these results as samples were obtained through a body donation program. Full medical information was not available and unknown selection biases may limit their application to contemporary biomedical issues in South Africa.

4.5.1 Osteon Population Density (OPD)

OPD increases as chronological age increases in both men and women. New osteon formations often target preexisting systems (Maggiano et al., 2016), such that only a finite number of remodeling events can exist within a given cortex. While an OPD asymptote may occur as early as 50 years of age (Stout & Paine, 1994), our results show that males do not reach a plateau until the end of the seventh decade of life, with this phenomenon occurring even later in females. Chronically stressed males demonstrate lower OPD values compared to those experiencing minimal stress (Heinrich, 2015), potentially explaining the late OPD asymptote in this sample, and more broadly, why age-at-death estimation methods using this variable tend to underestimate age in South African skeletal collections.

4.5.2 Osteon Area (On.Ar)

Women demonstrate a significant decline in On.Ar with age, suggesting that this mechanism may be related to post-menopausal bone loss. Although there is a strong genetic component to On.Ar (Havill et al., 2013), osteon size is frequently considered in both aging and biomechanical research. Numerous studies have demonstrated a decrease in On.Ar with age (Currey, 1964; Dominguez & Agnew, 2016; Evans, 1976; Han et al., 2009; Jowsey, 1966; Landeros & Frost, 1964; Takahashi, Epker, & Frost, 1965; Yoshino, Imaizumi, Miyasaka, & Seta, 1994), with

75 several others also reporting significant sex-based differences (Britz, Thomas, Clement, & Cooper, 2009; Burr, Ruff, & Thompson, 1990; Goliath, Stewart, & Stout, 2016).

Many small osteons are thought to be more efficient at resisting damage propagation and accumulation than a smaller number of large osteons (Skedros, Keenan, Williams, & Kiser, 2013). As microcracks increase with age (Schaffler, Choi, & Milgrom, 1995), a high density of small osteons may improve the fatigue life of cortical bone. Reduced osteon size may also be an adaptive response to a change in the biomechanical environment. Endocortical bone resorption left uncompensated by new periosteal apposition will decrease the second moment of inertia, thereby increasing biomechanical stress and strain. Osteon size demonstrates an inverse relationship with strain (van Oers, Ruimerman, van Rietbergen, Hilbers, & Huiskes, 2008), such that smaller osteons are expected in women who experience substantial cortical bone loss in the decades following menopause. In agreement with this hypothesis, Dominguez and Agnew (2016) found a significant positive relationship between On.Ar. and percent cortical area (%Ct.Ar), even after controlling for age.

4.5.3 Relative Cortical Area (Rt.Ct.Ar)

Men and women in the Kirsten Skeletal Collection may be experiencing disparate levels of nutritional and/or physiological stress. While women demonstrate predicable age-related patterns in terms of their cortical bone mass, the expected biological relationships are obscured in men. Women achieve peak bone mass in their mid-20s, and consistent with the age of menopause, exhibit accelerated rates of bone loss after 50 years. Men have significantly smaller Rt.Ct.Ar in young adulthood, suggesting compromised or delayed peak bone mass attainment. This is highly unusual; other studies report average male bone mass in excess to females’ at every decade of adult life (Dupras & Pfeiffer, 1996; Sedlin, Frost, & Villanueva, 1963; Streeter & Stout, 2003; Takahashi & Frost, 1966). Malnutrition, alcohol abuse, and issues of comorbidity are presumably influencing this parameter in at least some of the men represented in this sample.

Figure 9a-f: Cortical bone parameters. Predicted age-related changes in osteon population density (OPD) (a-b), osteon area (On.Ar) (c-d), and relative cortical area (Rt.Ct.Ar) (e-f) in men and women. The solid line represents the fitted mean from the regression model, and the dashed lines represent the 95% confidence interval of the prediction. SAC in blue circles, SAB in circles diamonds, and SAW in beige circles.

4.5.3.1 Substance Abuse Issues

Historically, many individuals from the represented groups engaged in farm labor, as the Western Cape is the leading agricultural region in the country. Socioeconomic and health indicators among farm laborers in the Western Cape were especially poor. Adverse working conditions included poor sanitation, lack of electricity and running water, and low wages (London, 1999). Farm laborers were also not protected under minimum wage, statutory compensation, and unemployment insurance laws (Van der Horst, 1976). The practice of paying workers partially in alcohol, known as the “Dop” system, was introduced in the early years of colonial settlement in the Cape Colony. Despite its official prohibition in 1961, informal use was ongoing until at least the late 1990s, exacerbating issues of alcohol abuse in the Western Cape over the course of centuries (London, Nell, Thompson, & Myers, 1998; J. te W. Naude, London, Pitt, & Mahomed, 1998).

Chronic alcohol consumption directly inhibits osteoblastic bone formation and can led to secondary osteoporosis (Maurel, Boisseau, Benhamou, & Jaffre, 2012). Reduced bone volume and trabecular thickness has been documented in South African men with a history of alcohol abuse (Schnitzler et al., 1994). Men are more likely to be drinkers than women, with particularly high levels of use in the SAB, SAC, and non-urban populations (Gossage et al., 2014). In the 1998 South African Demographic and Health Survey (SADHS), 44.7% of men and 16.9% of women reported that they were regular alcohol consumers (Parry et al., 2005). However, heavy episodic drinking by pregnant women is associated with fetal alcohol syndrome (FAS) in infants, and the Western Cape has amongst the highest known FAS prevalence rates in the world (Croxford & Viljoen, 1999; Viljoen et al., 2005). There is also a strong link between alcohol consumption and other non-communicable diseases such as cancer, cardiovascular disease, liver disease, pancreatitis, and diabetes (Parry et al., 2005).

Early research suggested that amongst the SAB, osteoporosis, albeit rare, occurs at an uncharacteristically early age and is more common in men than women (Lynch et al., 1967). Dietary iron overload and vitamin C deficiency are purported causes, and may even be co- morbid factors as iron accelerates the catabolism of ascorbic acid (Schnitzler, Schnaid, MacPhail, Mesquita, & Robson, 2005). In a year-long study at the Baragwanath Hospital in the Soweto area of Johannesburg, 69% of patients admitted for osteoporosis treatment were

78 clinically diagnosed with scurvy (Grusin & Samuel, 1957). African hemosiderosis may be related to consuming large quantities of traditional beer prepared in iron pots or metal drums (Bothwell, Seftel, Jacobs, Torrance, & Baumslag, 1964).

Smoking and taking of snuff, which were ubiquitous at this time, have recently been linked to increased osteoporosis incidence in a study of South African women (Ayo-Yusuf & Olutola, 2014). Smoking prevalence rates were particularly high among SAB and SAC men, and low among SAB and SAI women at the end of the apartheid era (Yach et al. 1992). Sex differences in smoking behavior are established in adolescence, with apartheid era prevalence rates of 27.3% in SAB boys but only 0.8% in SAB girls (Strebel, Kuhn, & Yach, 1989). However, the negative effects of smoking on the attainment of peak bone mass are not yet well understood (Sahni & Kiel, 2015).

4.5.3.2 Dietary Deficiencies

Suboptimal skeletal development may have occurred in some individuals if they had dietary deficiencies in calcium and vitamin D like those observed among contemporary youths from the Western Cape (Naude, Carey, Laubscher, Fein, & Senekal, 2012). Dietary data for urban SAB on the Cape Peninsula indicate that dairy and vegetal consumption was especially poor (Bourne et al., 1994). An apartheid nutritional survey of the SAB revealed that 85-90% of adults receive less than 500 mg of daily calcium intake (Walker, 1965). Using metacarpal radiometry, follow-up research demonstrated that black adolescents consuming a low calcium diet had lower bone mass indices than their white contemporaries who received adequate calcium nutrition (Walker, Walker & Richardson, 1971). Information about nutritional deficiencies and associated low bone mass were seen as difficult to interpret since they did not correlate with increased fracture risk (Walker, 1972).

Adverse environmental circumstances may also affect the growth rates of males more than females, resulting in delayed skeletal maturation in boys but not in girls (Stinson, 1985). A recent study found the same pattern and timing of skeletal maturation among SAB and SAW girls from different socioeconomic groups, but significantly delayed skeletal maturation in SAB compared to SAW boys (Cole et al., 2015). While delayed puberty is a known risk factor for osteopenia in men (Finkelstein, Neer, Biller, Crawford, & Klibanski, 1992), the long term effects of these

79 factors remain uncertain as the rate and extent of change to Rt.Ct.Ar is much more pronounced in women across the entire adult lifespan.

4.5.4 Trabecular Bone Histomorphometry

Largely because of poor preservation of the medullary space in archaeologically derived bone, anthropological studies of bone loss and fragility seldom consider trabecular bone microstructure (Agarwal, Dumitriu, Tomlinson, & Grynpas, 2004; Brickley & Howell, 1999; Kneissel et al., 1994, 1997). Given that our sample was obtained through a body donation program, diagenesis was not an issue. To our knowledge, this is the first research to directly examine age-related changes to trabecular parameters in the ribs. We found unexpected patterns of trabecular bone loss in the Kirsten Skeletal Collection, including a lack of significant age-associations in women.

Trabecular indices do not differ significantly by sex. Trabecular parameters are highly variable among skeletal sites, such that our results are not directly comparable to those from other research (Amling et al., 1996). However, histomorphometric studies of trans-iliac bone biopsies in other South African samples have also failed to detect any sex-specific patterns with age (Schnitzler et al., 1990; Schnitzler, 1993). Age- and sex-based differences tend to be more pronounced at weight-bearing sites such as the proximal femur (Eckstein et al., 2007; Turunen, Prantner, Jurvelin, Kröger, & Isaksson, 2013). At least for the men in our study, trabecular bone networks seem to be progressively disconnected by plate perforation such that BV/TV and Th.N decrease and Tb.Sp increases with age. Consistent with observations by Parfitt and colleagues (1983), the change in BV/TV is much greater than changes to Tb.Th. Values for Tb.Th remain relatively constant across the adult lifespan. This differs from previous research that suggests this as a common mechanism of age-related bone loss in men (Macdonald et al., 2011).

This study’s large sample size and targeted sampling procedure increases the explanatory power of our results. Nevertheless, women were relatively underrepresented, and we had weak coverage at the age extrema. Given that trabecular parameters are especially variable in our female sample, as illustrated in the heterogeneous scatterplots, these limitations may have impeded our ability to deduce overarching patterns of age-related change. It is possible that the demographic makeup of the sample also influenced our results. SAW women are not represented in the younger age cohorts, nor are SAB women in the older age cohorts.

81 the prediction. SAC in blue circles, SAB in green circles, and SAW in beige circles.

4.5.4.1 Female Life History/Contraception/HRT

The extent to which racial inequalities and disparate living conditions affected women’s life history milestones are difficult to gauge. Available information suggests general similarities among the three population groups represented in our sample. In 1977, average menarcheal age was 13.9 years for SAB and 13.1 years for SAW (Jones, Griffiths, Norris, Pettifor, & Cameron, 2009). A more recent study suggests it may be later in SAB women (Beksinska, Smit, Kleinschmidt, & Farley, 2011). Slight ethnic differences in the timing of menopause have been described, but the reported average age of menopause in urban SAB women, 48.9 years, does not significantly differ from values reported for SAW women (Walker, Walker, Ncongwane, & Tshabalala, 1984). Based on these averages, roughly half of the Kirsten Collection female sample is estimated to have been post-menopausal at time of death.

While all South African population groups exhibited a decline in fertility across the apartheid era, this decline was consistently most pronounced among SAB women and least pronounced among SAW women. From 1970 to 1996, total fertility declined from 5.4 to 3.7 in SAB women, from 5.1 to 2.8 in SAC women, and from 3.1 to 1.9 in SAW women (Udjo, 2003). Compared to today, higher fertility rates may have provided protective effects on trabecular bone parameters in the aged female sample. In a meta-analysis conducted by Bayray and Enquselassie (2013), parity was generally shown to positively influence BMD in postmenopausal women.

Racial residential segregation influenced the availability and quality of family planning services and menopausal hormone therapies (Burgard, 2004). Despite being targets of strong state family planning programs in late-apartheid South Africa, SAB women were less likely to practice modern contraception than non-blacks. Among SAB women in the early-2000s, for which we have data, public sector hormone replacement therapy (HRT) sales were considerably less than those reported for the private sector (Podmore, Botha, & Gray, 2008). Interview and survey work suggests that many poor and uninsured women may not have known that treatment options were available (Mashiloane, Bagratee, & Moodley, 2001). Given this, we believe it is unlikely that contraception or HRT substantially influenced postmenopausal bone loss in our sample.

4.5.5 Future Directions

Age-related changes to cortical bone may be more pronounced than changes to trabecular bone (Chen, Zhou, Shoumura, Emura, & Bunai, 2010). In this study, cortical parameters appeared to be reasonably good indicators of major life history events, whereas trabecular parameters were highly variable and almost completely undiagnostic in women. While this may be related to high fertility, an improved female sample size and distribution might elucidate an underlying age- related pattern not observed in this study. Although 2-D and 3-D studies of trabecular microstructure are highly correlated (Cohen et al., 2010; Thomsen et al., 2005), the idealized plate model of trabecular struts may also have influenced our results. Newer 3-D methods might provide more accurate, more highly resolved data. Future research exploring BMD, intracortical porosity, and fragility fractures will be useful for retrospectively assessing osteoporosis risk in this population, thereby informing current healthcare delivery considerations.

4.6 Conclusion

Although bone loss is clearly an age-related phenomenon, osteoporosis has a multifactorial etiology, and the risk of developing the condition is also mediated by factors independent of estrogen withdrawal in post-menopausal women. When studying bone quality outcomes, it is important to situate the research sample into a broader biocultural framework. Racial differences in health outcomes are not inherent, but are strongly dictated by social inequalities that cannot adequately be controlled for in existing epidemiological models (Phelan & Link, 2015; Roberts, 2012). Systemic marginalization through the apartheid political system made males particularly vulnerable to malnutrition and substance abuse problems early in life, causing delayed peak bone mass attainment. While this deficit is recouped in mid to late adulthood, a late OPD asymptote indicates that most of the men in this sample were subject to chronic stress throughout their lives. Known causes of secondary osteoporosis and poor trabecular bone quality in older individuals suggest that non-white males were at considerable risk during this historical period/era.

4.7 Acknowledgements

The authors thank Stellenbosch University faculty member Linda Greyling, and University of Toronto research assistants Virginia Pichler, Klara Komza, and Meimei Fong. The work also benefited from input from Dr. Jarred Heinrich.

4.8 Author Contributions

AB analyzed the data and drafted the manuscript. AB, SP, and MG designed the study and directed implementation and data collection. SP, AA, and MG edited the manuscript for intellectual content and provided critical comments.

4.9 Literature Cited

Agarwal, S. C. (2008). Light and Broken Bones: Examining and Interpreting Bone Loss and Osteoporosis in Past Populations. In M. A. Katzenberg & S. R. Saunders (Eds.), Biological Anthropology of the Human Skeleton (Second Edi, pp. 387–410). Hoboken, NJ: John Wiley & Sons, Inc. https://doi.org/10.1002/9780470245842.ch12 Agarwal, S. C., Dumitriu, M., Tomlinson, G. A., & Grynpas, M. D. (2004). Medieval trabecular bone architecture: The influence of age, sex, and lifestyle. American Journal of Physical Anthropology, 124(1), 33–44. https://doi.org/10.1002/ajpa.10335 Agarwal, S. C., & Grynpas, M. D. (1996). Bone quantity and quality in past populations. The Anatomical Record, 246(4), 423–32. https://doi.org/10.1002/(SICI)1097- 0185(199612)246:4<423::AID-AR1>3.0.CO;2-W Agarwal, S. C., & Grynpas, M. D. (2009). Measuring and interpreting age-related loss of vertebral bone mineral density in a medieval population. American Journal of Physical Anthropology, 139(2), 244–252. https://doi.org/10.1002/ajpa.20977 Alblas, A., Greyling, L. M., & Geldenhuys, E. M. (2018). Composition of the Kirsten Skeletal Collection at Stellenbosch University. South African Journal of Science, 114(1), 1–6. Amling, M., Herden, S., Posl, M., Hahn, M., Ritzel, H., & Delling, G. (1996). Heterogeneity of the skeleton: comparison of the trabecular microarchitecture of the spine, the iliac crest, the femur, and the calcaneus. Journal of Bone and Mineral Research, 11(1), 36–45. https://doi.org/10.1002/jbmr.5650110107 Andersson, N., & Marks, S. (1988). Apartheid and Health in the 1980s. Social Science and Medicine, 27(7), 667–681. Ataguba, J. E.-O., Day, C., & McIntyre, D. (2015). Explaining the role of the social determinants of health on health inequality in South Africa. Global Health Action, 8(1), 28865. https://doi.org/10.3402/gha.v8.28865 Ayo-Yusuf, O. A., & Olutola, B. G. (2014). Epidemiological association between osteoporosis and combined smoking and use of snuff among South African women. Nigerian Journal of Clinical Practice, 17(2), 174–177. https://doi.org/10.4103/1119-3077.127542 Bateman, C. (2006). South Africa under-prioritises osteoporosis. South African Medical Journal, 96(1), 19–20. Bayray, A., & Enquselassie, F. (2013). The effect of parity on bone mineral density in postmenopausal women: A systematic review. Journal of Osteoporosis and Physical Activity, 1(2), 104. https://doi.org/10.4172/2329-9509.1000104 Beck, T. (2003). Measuring the structural strength of bones with dual-energy X-ray absorptiometry: principles, technical limitations, and future possibilities. Osteoporosis International, 14(Suppl 5), S81–S88. https://doi.org/10.1007/s00198-003-1478-0 Beksinska, M. E., Smit, J. A., Kleinschmidt, I., & Farley, T. M. M. (2011). Assessing menopausal status in women aged 40-49 using depot-medroxyprogesterone acetate, norethisterone enanthate or combined oral contraceptption. S Afr Med J, 101, 131–135.

Bothwell, T. H., Seftel, H. C., Jacobs, P., Torrance, J. D., & Baumslag, N. (1964). Iron overload in Bantu subjects: Studies on the availability of iron in Bantu beer. American Journal of Clinical Nutrition, 14, 47–51. Bourne, L. T., Langenhoven, M. L., Steyn, K., Jooste, P. L., Nesamvuni, A. E., & Laubscher, J. A. (1994). The food and meal pattern in the urban African population of the Cape Peninsula, South Africa: the BRISK Study. The Central African Journal of Medicine, 40(6), 140–148. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/7954728 Brickley, M., & Howell, P. G. T. (1999). Measurement of changes in trabecular bone structure with age in an archaeological population. Journal of Archaeological Science, 26(2), 151– 157. https://doi.org/10.1006/jasc.1998.0313 Britz, H. M., Thomas, C. D. L., Clement, J. G., & Cooper, D. M. L. (2009). The relation of femoral osteon geometry to age, sex, height and weight. Bone, 45(1), 77–83. https://doi.org/10.1016/j.bone.2009.03.654 Burgard, S. (2004). Factors associated with contraceptive use in late- and post-apartheid South Africa. Studies in Family Planning, 35(2), 91–104. Burr, D. B., Ruff, C. B., & Thompson, D. D. (1990). Patterns of Skeletal Histologic Change Through Time : Comparison of an Archaic Native American Population With Modern Populations. The Anatomical Record, 226, 307–313. Chantler, S., Dickie, K., Goedecke, J. H., Levitt, N. S., Lambert, E. V., Evans, J., … Micklesfield, L. K. (2012). Site-specific differences in bone mineral density in black and white premenopausal South African women. Osteoporosis International, 23(2), 533–542. https://doi.org/10.1007/s00198-011-1570-9 Chen, H., Zhou, X., Shoumura, S., Emura, S., & Bunai, Y. (2010). Age- And gender-dependent changes in three-dimensional microstructure of cortical and trabecular bone at the human femoral neck. Osteoporosis International, 21(4), 627–636. https://doi.org/10.1007/s00198- 009-0993-z Chen, Y. Y., Feng, J. Y., Ting, W. Y., Yen, Y. F., Chuang, P. H., Pan, S. W., … Su, W. J. (2017). Increased risk of incident osteoporosis and osteoporotic fracture in tuberculosis patients: a population-based study in a tuberculosis-endemic area. Osteoporosis International, 28(5), 1711–1721. Choudhury, A., Ramsay, M., Hazelhurst, S., Aron, S., Bardien, S., Botha, G., … Pepper, M. S. (2017). Whole-genome sequencing for an enhanced understanding of genetic variation among South Africans. Nature Communications, 8(1), 2062. https://doi.org/10.1038/s41467-017-00663-9 Cohen, A., Dempster, D. W., Müller, R., Guo, X. E., Nickolas, T. L., Liu, X. S., … Shane, E. (2010). Assessment of trabecular and cortical architecture and mechanical competence of bone by high-resolution peripheral computed tomography: Comparison with transiliac bone biopsy. Osteoporosis International, 21(2), 263–273. https://doi.org/10.1007/s00198-009- 0945-7 Cole, T. J., Rousham, E. K., Hawley, N. L., Cameron, N., Norris, S. A., & Pettifor, J. M. (2015). Ethnic and sex differences in skeletal maturation among the Birth to Twenty cohort in South Africa. Archives of Disease in Childhood, 100(2), 138–143. https://doi.org/10.1136/archdischild-2014-306399 Conradie, M., Conradie, M. M., Kidd, M., & Hough, S. (2014). Bone density in black and white South African women: contribution of ethnicity, body weight and lifestyle. Archives of Osteoporosis, 9(1), 193. https://doi.org/10.1007/s11657-014-0193-0 Croxford, J., & Viljoen, D. (1999). Alcohol Consumption by Pregnant Women in the Western

Cape. South African Medical Journal, 89(September), 962–965. Curate, F. (2014). Osteoporosis and paleopathology: A review. Journal of Anthropological Sciences, 92, 119–146. https://doi.org/10.4436/JASS.92003 Curate, J. F. T., Albuquerque, A., Correia, J., Ferreira, I., de Lima, J. P., & Cunha, E. M. (2013). A glimpse from the past: Osteoporosis and osteoporotic fractures in a portuguese identified skeletal sample. Acta Reumatologica Portuguesa, 38(1), 20–27. Currey, J. D. (1964). Some effects of ageing in human Haversian systems. Journal of Anatomy, 98, 69–75. Dalzell, N., Kaptoge, S., Morris, N., Berthier, A., Koller, B., Braak, L., … Reeve, J. (2009). Bone micro-architecture and determinants of strength in the radius and tibia: Age-related changes in a population-based study of normal adults measured with high-resolution pQCT. Osteoporosis International, 20(10), 1683–1694. https://doi.org/10.1007/s00198-008-0833-6 Daya, M., van der Merwe, L., Galal, U., Möller, M., Salie, M., Chimusa, E. R., … Hoal, E. (2013). A panel of ancestry informative markers for the complex five-way admixed South African coloured population. PloS One, 8(12), e82224. https://doi.org/10.1371/journal.pone.0082224 de Wit, E., Delport, W., Rugamika, C. E., Meintjes, A., Möller, M., van Helden, P. D., … Hoal, E. G. (2010). Genome-wide analysis of the structure of the South African Coloured Population in the Western Cape. Human Genetics, 128(2), 145–153. https://doi.org/10.1007/s00439-010-0836-1 Dempster, D. W., Compston, J. E., Drezner, M. K., Glorieux, F. H., Kanis, J. A., Malluche, H., … Parfitt, A. M. (2013). Standardized Nomenclature, Symbols, and Units for Bone Histomorphometry: A 2012 Update of the Report of the ASBMR Histomorphometry Nomenclature Committee. Journal of Bone and Mineral Research, 28(1), 1–16. https://doi.org/10.1002/jbmr.1805 Ding, M., & Hvid, I. (2000). Quantification of age-related changes in the structure model type and trabecular thickness of human tibial cancellous bone. Bone, 26(3), 291–295. https://doi.org/10.1016/S8756-3282(99)00281-1 Dominguez, V. M., & Agnew, A. M. (2016). Examination of factors potentially influencing osteon size in the human rib. Anatomical Record, 299(3), 313–324. https://doi.org/10.1002/ar.23305 Dupras, T. L., & Pfeiffer, S. K. (1996). Determination of sex from adult human ribs. Canadian Society of Forensic Science Journal, 29(4), 221–231. Eckstein, F., Matsuura, M., Kuhn, V., Priemel, M., Müller, R., Link, T. M., & Lochmüller, E.-M. (2007). Sex Differences of Human Trabecular Bone Microstructure in Aging Are Site- Dependent. Journal of Bone and Mineral Research, 22(6), 817–824. https://doi.org/10.1359/jbmr.070301 Eleazer, C. D., & Jankauskas, R. (2016). Mechanical and metabolic interactions in cortical bone development. American Journal of Physical Anthropology, 160(2), 317–333. https://doi.org/10.1002/ajpa.22967 Erasmus, Z. (2012). Apartheid race categories: daring to question their continued use. Transformation, 79, 1–11. Evans, F. G. (1976). Mechanical properties and histology of cortical bone from younger and older men. The Anatomical Record, 185, 1–11. https://doi.org/10.1002/ar.1091850102 Finkelstein, J. S., Neer, R. M., Biller, B. K. M., Crawford, J. D., & Klibanski, A. (1992). Osteopenia in men with a history of delayed puberty. The New England Journal of Medicine, 326(5), 600–604.

Fuleihan, G. E.-H., Adib, M. G., & Nauroy, L. (2011). The Middle East & Africa Regional Audit: Epidemiology, Costs, & Burden of Osteoporosis in 2011. (J. Stenmark & L. Misteli, Eds.). Nyon, Switzerland. Fyhrie, S. D. P. (2005). Summary - Measuring “Bone Quality.” Journal of Musculoskeletal and Neuronal Interactions, 5(4), 318–320. Gangata, H., Ntaba, P., Akol, P., & Louw, G. (2010). The reliance on unclaimed cadavers for anatomical teaching by medical schools in Africa. Anatomical Sciences Education, 3(4), 174–183. https://doi.org/10.1002/ase.157 Goliath, J. R., Stewart, M. C., & Stout, S. D. (2016). Variation in osteon histomorphometrics and their impact on age-at-death estimation in older individuals. Forensic Science International, 262, 282.e1-282.e6. https://doi.org/10.1016/j.forsciint.2016.02.053 Gossage, P. J., Snell, C. L., Parry, C. D. H., Marais, A. S., Barnard, R., de Vries, M., … May, P. A. (2014). Alcohol use, working conditions, job benefits, and the legacy of the “dop” system among farm workers in the Western Cape Province, South Africa: Hope despite high levels of risky drinking. International Journal of Environmental Research and Public Health, 11(7), 7406–7424. https://doi.org/10.3390/ijerph110707406 Grusin, H., & Samuel, M. D. (1957). A syndrome of osteoporosis in Africans and its relationship to scurvy. The American Journal of Clinical Nutrition, 5(6), 644–650. Halperin, E. C. (2007). The poor, the black, and the marginalized as the source of cadavers in United States anatomical education. Clinical Anatomy, 20(5), 489–495. https://doi.org/10.1002/ca.20445 Han, S.-H., Kim, S.-H., Ahn, Y.-W., Huh, G.-Y., Kwak, D.-S., Park, D.-K., … Kim, Y.-S. (2009). Microscopic age estimation from the anterior cortex of the femur in Korean adults. Journal of Forensic Sciences, 54(3), 519–522. https://doi.org/10.1111/j.1556- 4029.2009.01003.x Hansen, S., Shanbhogue, V., Folkestad, L., Nielsen, M. M. F., & Brixen, K. (2014). Bone microarchitecture and estimated strength in 499 adult Danish women and men: A cross- sectional, population-based high-resolution peripheral quantitative computed tomographic study on peak bone structure. Calcified Tissue International, 94(3), 269–281. https://doi.org/10.1007/s00223-013-9808-5 Harris, B., Goudge, J., Ataguba, J. E., McIntyre, D., Nxumalo, N., Jikwana, S., & Chersich, M. (2011). Inequities in access to health care in South Africa. Journal of Public Health Policy, 32(S1), S102–S123. https://doi.org/10.1057/jphp.2011.35 Havill, L. M., Allen, M. R., Harris, J. A. K., Levine, S. M., Coan, H. B., Mahaney, M. C., & Nicolella, D. P. (2013). Intracortical bone remodeling variation shows strong genetic effects. Calcified Tissue International, 93(5), 472–480. https://doi.org/10.1007/s00223-013- 9775-x Heinrich, J. T. (2015). Spatial Characterization of Rib Cortical Bone Microstructure and the Effect of Nutritional and Physiological Stresses. Retrieved from TSpace (http://hdl.hand http://hdl.handle.net/1807/69305 le.net/1807/69305). Hough, S., Ascott-Evans, B. H., Brown, S. L., Cassim, B., de Villiers, T. J., Lipschitz, S., … Sonnendecker, E. W. (2010). NOFSA Guideline for the Diagnosis and Management of Osteoporosis. Journal of Endocrinology, Metabolism and Diabetes of South Africa, 15(3), 1–200. https://doi.org/10.1080/22201009.2010.10872239 International Osteoporosis Foundation. (2011). The Middle East & Africa Regional Audit: Epidemiology, Costs, & Burden of Osteoporosis in 2011. Nyon, Switzerland: Fuleihan, G. E.-H., Adib, M. G., & Nauroy, L.

Inwood, K., & Masakure, O. (2013). Poverty and Physical Well-being among the Coloured Population in South Africa. Economic History of Developing Regions, 28(2), 56–82. https://doi.org/10.1080/20780389.2013.866382 Jones, L. L., Griffiths, P. L., Norris, S. a., Pettifor, J. M., & Cameron, N. (2009). Age at menarche and the evidence for a positive secular trend in urban South Africa. American Journal of Human Biology, 21(August 2008), 130–132. https://doi.org/10.1002/ajhb.20836 Jowsey, J. (1966). Studies of Haversian systems in man and some animals. Journal of Anatomy, 100(4), 857–864. https://doi.org/10.1002/ajpa Khosla, S., Riggs, B. L., Atkinson, E. J., Oberg, A. L., McDaniel, L. J., Holets, M., … Melton, L. J. (2006). Effects of sex and age on bone microstructure at the ultradistal radius: a population-based noninvasive in vivo assessment. Journal of Bone and Mineral Research : The Official Journal of the American Society for Bone and Mineral Research, 21(1), 124– 131. https://doi.org/10.1359/JBMR.050916 Kneissel, M., Boyde, A., Hahn, M., Teschler-Nicola, M., Kalchhauser, G., & Plenk, H. (1994). Age- and sex-dependent cancellous bone changes in a 4000y BP population. Bone, 15(5), 539–545. https://doi.org/10.1016/8756-3282(94)90278-X 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. Calcified Tissue International, 61(2), 95–100. https://doi.org/10.1007/s002239900302 Labuschagne, B. C. J., & Mathey, B. (2000). Cadaver profile at University of Stellenbosch Medical School, South Africa, 1956 – 1996. Clinical Anatomy, 13, 88–93. Landeros, O., & Frost, H. M. (1964). Comparison of amounts of remodeling activity in opposite cortices of ribs in children and adults. Journal of Dental Research, 45(1), 152–158. https://doi.org/10.1177/00220345660450010701 Lees, B., Stevenson, J. C., Molleson, T., & Arnett, T. R. (1993). Differences in proximal femur bone density over two centuries. The Lancet, 341(8846), 673–676. https://doi.org/10.1016/0140-6736(93)90433-H London, L. (1999). The “dop” system, alcohol abuse and social control amongst farm workers in South Africa: a public health challenge. Social Science & Medicine, 48(10), 1407–1414. https://doi.org/10.1016/s0277-9536(98)00445-6 London, L., Nell, V., Thompson, M.-L., & Myers, J. E. (1998). Western Cape - Collateral evidence from a study of occupational hazard. South African Medical Journal, 88(9), 1096– 1101. Lynch, S. R., Berelowitz, I., Seftel, H. C., Miller, G. B., Krawitz, P., & Bothwell, T. H. (1967). Osteoporosis in Johannesburg Bantu Males: Its Relationship to Siderosis and Ascorbic Acid Deficiency. The American Journal of Clinical Nutrition, 20(8), 799–807. Lynch, S. R., Seftel, H. C., Wapnick, A. A., Charlton, R. W., & Bothwell, T. H. (1970). Some aspects of calcium metabolism in normal and osteoporotic Bantu subjects with special reference to the effects of iron overload and ascorbic acid depletion. The South African Journal of Medical Sciences, 35, 45–56. Macdonald, H. M., Nishiyama, K. K., Kang, J., Hanley, D. a, & Boyd, S. K. (2011). Age-related patterns of trabecular and cortical bone loss differ between sexes and skeletal sites: a population-based HR-pQCT study. Journal of Bone and Mineral Research : The Official Journal of the American Society for Bone and Mineral Research, 26(1), 50–62. https://doi.org/10.1002/jbmr.171 Maggiano, I. S., Maggiano, C. M., Clement, J. G., Thomas, C. D. L., Carter, Y., & Cooper, D.

M. L. (2016). Three-dimensional reconstruction of Haversian systems in human cortical bone using synchrotron radiation-based micro-CT: Morphology and quantification of branching and transverse connections across age. Journal of Anatomy, 228(5), 719–732. https://doi.org/10.1111/joa.12430 Mashiloane, C. D., Bagratee, J., & Moodley, J. U. (2001). Awareness of and attitudes toward menopause and hormone replacement therapy in an African community. International Journal of Gynecology & Obstetrics, (76), 91–93. Maurel, D. B., Boisseau, N., Benhamou, C. L., & Jaffre, C. (2012). Alcohol and bone: Review of dose effects and mechanisms. Osteoporosis International, 23(1), 1–16. https://doi.org/10.1007/s00198-011-1787-7 Mays, S. (1996). Age dependent cortical bone loss in a medieval population. International Journal of Osteoarchaeology, 6, 144–154. https://doi.org/10.1002/(SICI)1099- 1212(199603)6:2<144::AID-OA261>3.0.CO;2-G Micklesfield, L. K., Norris, S. A., & Pettifor, J. M. (2011). Ethnicity and Bone: A South African Perspective. J Bone Miner Metab, 29, 257–267. Naude, C. E., Carey, P. D., Laubscher, R., Fein, G., & Senekal, M. (2012). Vitamin D and calcium status in South African adolescents with alcohol use disorders. Nutrients, 4(8), 1076–1094. https://doi.org/10.3390/nu4081076 Naude, J. te W., London, L., Pitt, B., & Mahomed, C. (1998). The “dop” system around Stellenbosch - Results of a farm survey. South African Medical Journal, 88(9), 1102–1105. Nightingale, E. O., Hannibal, K., Geiger, J., Hartmann, L., Lawerence, R., & Spurlock, J. (1990). Apartheid medicine. JAMA: The Journal of …, 264, 2097–2102. Retrieved from http://jama.ama-assn.org/content/264/16/2097.short Osborne, J. W. (2010). Improving your data transformations : Applying the Box-Cox transformation. Practical Assessment, Research & Evaluation, 15(12), 1–9. Paine, R. R., & Brenton, B. P. (2006). Dietary health does affect histological age assessment: An evaluation of the Stout and Paine (1992) age estimation equation using secondary osteons from the rib. Journal of Forensic Sciences, 51(3), 489–492. https://doi.org/10.1111/j.1556- 4029.2006.00118.x Parfitt, A. M., Mathews, C. H. E., Villanueva, A. B., Kleerekoper, M., Frame, B., & Rao, D. S. (1983). Relationships between surface, volume, and thickness of iliac trabecular bone in aging and in osteoporosis. Implications for the microanatomic and cellular mechanisms of bone loss. Journal of Clinical Investigation, 72(4), 1396–1409. https://doi.org/10.1172/JCI111096 Parry, C. D., Pluddemann, A., Steyn, K., Bradshaw, D., Norman, R., & Laubscher, R. (2005). Alcohol Use in South Africa: Findings from the First Demographic and Health Survey (1998)*. Journal of Studies on Alcohol, 66, 91–97. Patin, E., Lopez, M., Grollemund, R., Verdu, P., Harmant, C., Quach, H., … Quintana-Murci, L. (2017). Dispersals and genetic adaptation of Bantu-speaking populations in Africa and North America. Science, 356(6337), 543–546. https://doi.org/10.1126/science.aal1988 Pfeiffer, S., Crowder, C., Harrington, L., & Brown, M. (2006). Secondary osteon and Haversian canal dimensions as behavioral indicators. American Journal of Physical Anthropology, 131, 460–468. https://doi.org/10.1002/ajpa.20454 Pfeiffer, S., Heinrich, J., Beresheim, A., & Alblas, M. (2016). Cortical bone histomorphology of known-age skeletons from the Kirsten collection, Stellenbosch University, South Africa. American Journal of Physical Anthropology, 160, 137–147. https://doi.org/10.1002/ajpa.22951

Phelan, J. C., & Link, B. G. (2015). Is Racism a Fundamental Cause of Inequalities in Health? Annual Review of Sociology, 41(1), 311–330. https://doi.org/10.1146/annurev-soc-073014- 112305 Podmore, S. H., Botha, J. H., & Gray, A. L. (2008). Hormone therapy use in the South African public and private sectors 2001–2005. Maturitas, 60(2), 182–184. https://doi.org/10.1016/j.maturitas.2008.04.008 Pratte, D. G., & Pfeiffer, S. (1999). Histological age estimation of a cadaveral sample of diverse origins. Journal of the Canadian Society of Forensic Science, 32(March 2014), 155–167. https://doi.org/10.1080/00085030.1999.10757496 Roberts, D. (2012). Debating the Cause of Health Disparities. Cambridge Quarterly of Healthcare Ethics, 21(3), 332–341. https://doi.org/10.1017/S0963180112000059 Robling, A. G., & Stout, S. D. (2003). Histomorphology, geometry, and mechanical loading in past populations. In S. Agarwal & S. D. Stout (Eds.), Bone Loss and Osteoporosis: An Anthropological Perspective (1st editio, pp. 207–228). New York: Klewer Academic/Plenum Publishers. Sahni, S., & Kiel, D. P. (2015). Smoking, alcohol, and bone health. In M. F. Holick & J. W. Nieves (Eds.), Nutrition and Bone Health, Nutrition and Health (pp. 489–504). New York: Springer Science+Business Media, LLC. https://doi.org/10.1007/978-1-4939-2001-3 Saunders, S. R. (2008). Juvenile skeletons and growth-related studies. In A. M. Katzenberg & S. R. Saunders (Eds.), Biological Anthropology of the Human Skeleton (Second Edi, pp. 117– 147). Hoboken, NJ: John Wiley & Sons, Inc. Schaffler, M. B., Choi, K., & Milgrom, C. (1995). Aging and matrix microdamage accumulation in human compact bone. Bone, 17(6), 521–525. https://doi.org/10.1016/8756- 3282(95)00370-3 Schnitzler, C. M. (1993). Bone quality: A determinant for certain risk factors for bone fragility. Calcified Tissue International, 53(1 Supplement). https://doi.org/10.1007/BF01673398 Schnitzler, C. M., Macphail, A. P., Shires, R., Schnaid, E., Mesquita, J. M., & Robson, H. J. (1994). Osteoporosis in african hemosiderosis: Role of alcohol and iron. Journal of Bone and Mineral Research, 9(12), 1865–1873. https://doi.org/10.1002/jbmr.5650091205 Schnitzler, C. M., Pettifor, J. M., Mesquita, J. M., Bird, M. D. T., Schnaid, E., & Smyth, A. E. (1990). Histomorphometry of iliac crest bone in 346 normal black and white South African adults. Bone and Mineral, 10(3), 183–199. https://doi.org/10.1016/0169-6009(90)90261-D Schnitzler, C. M., Schnaid, E., MacPhail, A. P., Mesquita, J. M., & Robson, H. J. (2005). Ascorbic acid deficiency, iron overload and alcohol abuse underlie the severe osteoporosis in black african patients with hip fractures - A bone histomorphometric study. Calcified Tissue International, 76(2), 79–89. https://doi.org/10.1007/s00223-004-0053-9 Sedlin, E. D., Frost, H. M., & Villanueva, B. S. (1963). Age changes in resorption in the human rib cortex. J Gerontol, 18, 345–349. Seftel, H. C., Malkin, C., Schmaman, A., Abrahams, C., Lynch, S. R., Charlton, R. W., & Bothwell, T. H. (1966). Osteoporosis, scurvy, and siderosis in Johannesburg Bantu. British Medical Journal, 1, 642–646. Silva, M., Alshamali, F., Silva, P., Carrilho, C., Mandlate, F., Jesus Trovoada, M., … Soares, P. (2015). 60,000 years of interactions between Central and Eastern Africa documented by major African mitochondrial haplogroup L2. Scientific Reports, 5(1), 12526. https://doi.org/10.1038/srep12526 Skedros, J. G., Keenan, K. E., Williams, T. J., & Kiser, C. J. (2013). Secondary osteon size and collagen/lamellar organization (“ osteon morphotypes”) are not coupled, but potentially

adapt independently for local strain mode or magnitude. Journal of Structural Biology, 181(2), 95–107. https://doi.org/10.1016/j.jsb.2012.10.013 Stewart, M. C., Goliath, J. R., Stout, S. D., & Hubbe, M. (2015). Intraskeletal variability of relative cortical area in humans. Anatomical Record, 298(9), 1635–1643. https://doi.org/10.1002/ar.23181 Steyn, M., Scholtz, Y., Botha, D., & Pretorius, S. (2013). The changing face of tuberculosis: Trends in tuberculosis-associated skeletal changes. Tuberculosis, 93(4), 467–474. https://doi.org/10.1016/j.tube.2013.04.003 Stinson, S. (1985). Sex Differences in Environmental Sensitivity During Growth and Development. Yearbook of Physical Anthropology, 28, 123–147. https://doi.org/10.1002/ajpa.1330280507 Stout, S. D., & Paine, R. R. (1994). Bone remodeling rates: A test of an algorithm for estimating missing osteons. American Journal of Physical Anthropology, 93(1), 123–129. https://doi.org/10.1002/ajpa.1330930109 Strebel, P. M., Kuhn, L., & Yach, D. (1989). Smoking practices in the black township population of Cape Town. South African Medical Journal, 75(9), 428–431. https://doi.org/10.1136/jech.43.3.209 Streeter, M. A., & Stout, S. D. (2003). The Histomorphometry Of The Subadult Rib: Age- Associated Changes In Bone Mass And The Creation Of Peak Bone Mass. In S. C. Agarwal & S. D. Stout (Eds.), Bone Loss and Osteoporosis: An Anthropological Perspective (1st Editio, pp. 91–101). New York: Klewer Academic/Plenum Publishers. Takahashi, H., & Frost, H. M. (1966). Age and Sex Related Changes in the Amount of Cortex of Normal Human Ribs. Acta Orthopaedica Scandinavica, 37(2), 122–130. https://doi.org/10.3109/17453676608993272 Takahashi, N., Epker, B. N., & Frost, H. M. (1965). Relation between age and size of osteons in man. Henry Ford Hosp Med Bull, 13, 25–31. Thomsen, J. S., Laib, A., Koller, B., Prohaska, S., Mosekilde, L., & Gowin, W. (2005). Stereological measures of trabecular bone structure: Comparison of 3D micro computed tomography with 2D histological sections in human proximal tibial bone biopsies. Journal of Microscopy, 218(2), 171–179. https://doi.org/10.1111/j.1365-2818.2005.01469.x Theron, E. (976): Report of the Commission of Enquiry into Matters Relating to the Coloured Population Group (RP 38/1976). Pretoria: Government Printer. Tishkoff, S. A., Reed, F. A., Friedlaender, F. R., Ranciaro, A., Froment, A., Hirbo, J. B., … Williams, S. M. (2009). The Genetic Structure and History of Africans and African Americans. Science, 324(5930), 1035–1044. https://doi.org/10.1126/science.1172257.The Turunen, M. J., Prantner, V., Jurvelin, J. S., Kröger, H., & Isaksson, H. (2013). Composition and microarchitecture of human trabecular bone change with age and differ between anatomical locations. Bone, 54(1), 118–125. https://doi.org/10.1016/j.bone.2013.01.045 Udjo, E. O. (2003). A Re-Examination of Levels and Differential in Fertility in South Africa From Recent Evidence. Journal of Biosocial Science, 35(3), S0021932003004139. https://doi.org/10.1017/S0021932003004139 van Oers, R. F. M., Ruimerman, R., van Rietbergen, B., Hilbers, P. A. J., & Huiskes, R. (2008). Relating osteon diameter to strain. Bone, 43(3), 476–82. https://doi.org/10.1016/j.bone.2008.05.015 van Rensburg, H. C. J., & Benatar, S. R. (1993). The legacy of apartheid in health and health care. South African Journal of Sociology, 24(4), 99–111. https://doi.org/10.1080/02580144.1993.10431680

Vidal, B., Pinto, A., Galvão, M. J., Santos, A. R., Rodrigues, A., Cascão, R., … Canhao, H. (2012). Bone histomorphometry revisited. Acta Reumatologica Portuguesa, 37(4), 294– 300. https://doi.org/10.1177/002215549704500215 Viljoen, D. L., Gossage, J. P., Brooke, L., Adnams, C. M., Jones, K. L., Robinson, L. K., … May, P. A. (2005). Fetal alcohol syndrome epidemiology in a South African community: a second study of a very high prevalence area. Journal of Studies on Alcohol, 66(5), 593–604. https://doi.org/10.15288/jsa.2005.66.593 Walker, A. R. P. (1965). Osteoporosis and calcium deficiency. American Journal of Clinical Nutrition, 16, 327–336. Walker, A. R. P. (1972). The human requirement of calcium: should low intakes be supplemented? American Journal of Clinical Nutrition, 25(5), 518–530. Walker, A. R. P., Walker, B. F., Ncongwane, J., & Tshabalala, E. N. (1984). Age of menopause in black women in South Africa. BJOG: An International Journal of Obstetrics & Gynaecology, 91(8), 797–801. Walker, A. R. P., Walker, B. F., & Richardson, B. D. (1971). Metacarpal bone dimensions in young and aged South African Bantu consuming a diet low in calcium. Postgraduate Medical Journal, 47(548), 320–325. https://doi.org/10.1136/pgmj.47.548.320 Wapnick, A. A., Lynch, S. R., Seftel, H. C., Charlton, R. W., & Jowsey, J. (1971). The effect of siderosis and ascorbic acid depletion on bone metabolism, with special reference to osteoporosis in the Bantu. British Journal of Nutrition, 25, 367–376. Yach, D., McIntyre, D., & Saloojee, Y. (1992). Smoking in South Africa: the health and economic impact. Tobacco Control, 1(4), 272–280. https://doi.org/10.1182/blood-2015-04- 641225. Yoshino, M., Imaizumi, K., Miyasaka, S., & Seta, S. (1994). Histological estimation of age at death using microradiographs of humeral compact bone. Forensic Science International, 64(2–3), 191–198. https://doi.org/10.1016/0379-0738(94)90231-3

Chapter 5 Use of Backscattered Scanning Electron Microscopy to Quantify the Bone Tissue of Mid-Thoracic Human Ribs

AMY C. BERESHEIM1*, SUSAN K. PFEIFFER1,2, MARC D. GRYNPAS3,4, AMANDA ALBLAS5

*Correspondence to: Amy Beresheim, Department of Anthropology, University of Toronto, 19 Russell Street, Toronto, Canada M5S 2S2. Phone: 47-459-2284 E-mail: [email protected]

To be submitted to the American Journal of Physical Anthropology

5.1 Abstract

Introduction: Back-scattered scanning electron microscopy (BSE-SEM) is an underutilized imaging method in biological anthropology. It allows analysis of both the cortical and trabecular bone compartments through high-resolution images of large cross-sections. Machine availability, relatively low imaging costs, and automated data collection permit the study of bone quality and tissue mineralization.

Materials and methods: Using BSE-SEM imaging, osteocyte lacunar density (Ot.Lc.Dn), cortical porosity (Ct.Po), and bone tissue mineralization (WMGL) were quantified in mid- thoracic ribs from the Kirsten Skeletal Collection. Individuals (nfemale=75, nmale=68, mean age=46.3 years) were predominantly from the South African Coloured (SAC) population group

(nSAC=103, 72%), as classified under the apartheid government. Non-white males were particularly vulnerable to malnutrition and substance abuse issues during the apartheid era, the

93 period in which the individuals lived and died (DOD 1968-1998 AD). Age-related changes in histomorphometric parameters and sex differences were examined using correlation analysis, as well as linear and non-linear regressions.

Results: Young adult men have significantly less mineralized bone and fewer osteocyte lacunae in both cortical and trabecular bone compartments. Sex-related changes with age are mostly non- significant. Only women demonstrate a positive relationship between Ct.Po and age.

Discussion and Conclusion: Systemic disenfranchisement of non-white population groups affected bone health in South Africa, and may continue to do so today. Indicators of poor bone quality are evident in SAC males across the adult lifespan, indicating that osteoporosis and fracture risk are not just of concern to the aged white female population.

Key words: South African apartheid, bone histomorphometry, osteoporosis, BSE-SEM

5.2 Introduction

5.2.1 South African Apartheid and the Social Determinants of Health

Information on apartheid health indicators are needed for understanding epidemiological transitions in South Africa, and for determining whether conditions have improved since the implementation of a democratic government in 1994. Under the apartheid political system, individuals were officially classified as South African Black (SAB), South African White (SAW), or South African Coloured (SAC). Non-whites were required to live in peripheral areas with inadequate housing, poor infrastructure, and limited access to education, healthcare, and community services. Institutionalized racial inequalities created disparities in the distribution of health and disease, as well as differential dietary and substance use patterns among population groups (Andersson and Marks, 1988; Nightingale et al., 1990; van Rensburg and Benatar, 1993).

Epidemiological data on osteoporosis and fragility fractures in South Africa are inconsistent and sparse, especially for this historical period. Early metacarpal radiometry work demonstrated lower cortical densities in SAB compared to SAW population groups (Walker et al., 1971; Solomon, 1979). Using atomic absorption spectroscopy, Daniels et al. (1997) did not reveal any differences in bone mineral density (BMD) between SAB and SAW women at the distal radius or lumbar spine, but reported significantly lower values at the femoral neck location in SAW

94 women. The most comprehensive research suggests that South Africa had among the lowest global incidences of hip fractures during the 1950s and 1960s (Cumming et al., 1997; Cauley et al., 2014). Consistent with similar studies conducted in the United States, lower hip fracture rates are reported for SAB than for SAW (Dent et al., 1968; Solomon, 1968, 1979).

Novel information on apartheid health conditions may be obtained through the study of contemporary skeletal collections. Using a backscattered scanning electron microscopy (BSE- SEM) approach, this study aims to produce bone quality and tissue mineralization data for an understudied South African population from the Western Cape.

5.2.2 Back-Scattered Scanning Electron Microscopy (BSE-SEM) and Bone Tissue Mineralization

BSE-SEM allows the study of relative mineral content by using high energy electrons reflected from a sample (Boyde, 2012). While it is sometimes used to examine bone diagenesis in archaeological samples (Turner-Walker and Syversen, 2002; Morales et al., 2017), BSE-SEM has largely been eclipsed by newer 3-D imaging methods for the study of bone tissue mineralization and microstructural variation. However, 3-D techniques tend to be labor intensive, require extensive data storage, and are limited to small bone volumes. A major advantage of BSE-SEM is that full bone cross-sections can be analyzed at high resolution, improving statistics, and protecting against variability associated with stochastic remodeling and/or the local strain environment. Small intracortical porosities such as osteocyte lacunae and Haversian canals, which are difficult to capture using most micro-CT set-ups, can be enumerated and measured with relative ease. Further, relational light microscopy (LM) images of the same fields of view can be directly compared, revealing significant correlations between the degree of mineralization and patterns of collagen fiber orientation (Goldman et al., 2000, 2005).

5.2.3 Indicators of Bone Quality

Age-related studies of bone mass and density often focus on resorption at the endosteal surface, but intracortical remodeling is another important mechanism of bone loss (Zebaze et al., 2009, 2010). Small increases in cortical porosity can lead to a disproportionate decrease in bone strength, having important consequences for bone fragility and fracture susceptibility (McCalden et al., 1993; Yeni et al., 1997; Zioupos, 2001; Ural and Vashishth, 2007). It is also important to consider BMD relative to porosity, as the distribution of both relate to variations in the

95 remodeling rate. A recent study found that cortical porosity was higher when there was a lower average degree of mineralization in the iliac crest (Misof et al., 2014).

Osteocyte lacunar-canicular networks are critical determinants of bone health and function, given their role in mechanosensation and transduction (Klein-Nulend et al., 2013). Osteocyte lacunar properties have been proposed as an alternative index for assessing bone quality (Ma et al., 2008), with a growing interest the relationship between their morphology and their adaptive response (Vatsa et al., 2008; van Hove et al., 2009; Carter et al., 2013; van Oers et al., 2015; Hemmatian et al., 2017). Osteocyte apoptosis is also linked to increased intracortical resorption in response to bone fatigue (Cardoso et al., 2009; Bellido, 2015). Normative reference data on the osteocyte population in healthy bone and the effect(s) of age and sex are lacking, but osteocyte deficiency strongly correlates with low mechanical loading and bone microdamage (Qiu et al., 2005; Aguirre et al., 2006; Britz et al., 2012). Osteocyte lacunae are not always occupied by living cells, and the percent vacancy can increase with age (Skedros et al., 2016). Preservation and visualization of the cells themselves are difficult if not impossible in curated collections; therefore, the lacuna is used as a proxy in this research.

5.3 Materials and Methods

5.3.1 Research Sample

Mid-thoracic ribs derive from the Kirsten Skeletal Collection, Stellenbosch University, Tygerberg, South Africa. A subset of the original research sample used by Pfeiffer et al. (2016) was selected for this study. Mid-shaft thick-sections from each rib were previously embedded for histological analysis following Crowder et al. (2012). All female specimen blocks with complete infiltration were included. Males in each 10-year age cohort were randomly selected to match the female distribution. The present sample is comprised of 143 specimen blocks, each representing one individual (nmale=68, nfemale= 75, ages 12-90 years). Sample demographics are displayed in Figure 11. Age, sex, cause-of-death information, and “race” designation under the apartheid government are known. Most individuals included in this study were classified as South African

Coloureds (SAC, nSAC=103), followed by South African Blacks (SAB, nSAB=25), and South

African Whites (SAW, nSAW=15), respectively. Cause of death was not a determining factor for exclusion. These individuals likely experienced a myriad of conditions that could potentially influence variables of interest. However, individual outliers were not omitted as others have

96 previously demonstrated that variation in intracortical porosity and osteocyte lacunar parameters are normal (Agnew and Stout, 2012; Carter et al., 2013, 2014; Andronowski et al., 2017).

Figure 11: Sample structure by age and sex.

5.3.2 SEM Preparation

Prior to imaging, mid-shaft rib thick-sections were embedded under vacuum using Buehler Epothin epoxy resin and hardener. Using a precision saw, surplus material was removed from each specimen block to generate a cut surface for polishing. This surface was ground and polished using a series of increasingly fine sandpaper and diamond paste. Final polishing was performed with a cloth and 1 um diamond paste. Blocks were rinsed in an ultrasonic bath with distilled water to remove any debris and blotted dry with a clean tissue. Finish quality was visually inspected using a dissecting microscope (Boyde, 2012; Jones, 2012).

Polished specimen blocks were affixed to a 9x9 cm plexiglass plate using polymer clay, ensuring that the top surfaces of the blocks were level. Carbon tape was adhered to the blocks to make an electrically conductive connection to the base of the plate. Carbon evaporation was then performed using the Edwards Auto 306 vacuum coater system at Mount Sinai Hospital in Toronto, Ontario.

5.3.3 Image Acquisition and Analysis

BSE-SEM scans of transverse rib cross-sections were obtained using a FEI XL30 environmental scanning electron microscope (ESEM) at Mount Sinai Hospital in Toronto, Ontario. A Faraday cup electrometer was used to measure beam current. Images were captured at a working distance of 15 mm, 20 kV accelerating voltage, and approximately 0.95 nA beam current. Magnification was set at 100x, resulting in 4.65 um/pixel resolution.

MgF2 and SiO2 standards were used to calibrate gray levels during image capture, and to control for any errors caused by machine drift. Previous research has demonstrated that BSE-SEM imaging error is negligible, allowing comparisons between data from multiple imaging sessions (Vajda et al., 1995). After every sample was imaged, a histogram of standards was collected to ensure that peak grey levels remained at 125 for SiO2, and 150 for MgF2. The position of the two calibration peaks needed to be within 2 grey level values to be considered acceptable. Samples were reimaged if these criteria were not met. Brightness and contrast were adjusted when necessary to recreate the original histogram standards.

Images were manually stitched in Adobe Photoshop to create a montage of each rib cross-section (Figure 12). Photomontages were imported into Matlab (Mathworks, Natick, MA) for image processing and analysis. Using a series of image filters, photomontages were de-speckled and cleaned before being transformed to binary (grey levels 0 and 255). For each rib cross-section, the endosteal border was manually traced using a digitizing tablet and stylus to allow separate analyses of the cortical and trabecular bone compartments. Using customized scripts, lacunae and intracortical porosities (including Haversian canals and resorption spaces) were enumerated relative to the detected bone area in each compartment. Cracks created through tissue processing were excluded from these calculations. Given the characteristic size and shape of these microstructural features, it is possible to automate data collection. Cortical and trabecular bone parameters are listed in Table 7. All abbreviations follow Dempster et al. (2013).

Grey level histogram values were converted to weighed mean grey levels (WMGLs), an indicator of overall degree of mineralization for each individual (Vajda et al., 1995; Bloebaum et al., 1997). Each greyscale value was multiplied by its pixel count and divided by the total number of pixels examined, then the average grey level across the total number of observations was calculated. Black background values (grey levels 0-15), representing regions of non- mineralized bone (i.e., lacunae, Haversian canals, resorption spaces and cracks), were deleted before calculating the WMGLs.

5.3.4 Statistical Analysis

Following Shapiro-Wilks tests, data are presented as mean (± standard deviation) if normally distributed or median (interquartile range) if non-normally distributed. Descriptive statistics for each variable are given by sex and age cohort (Tables 8 and 9). Separate analyses were not performed for each population group given the small SAB and SAW sample sizes. Values before the third decade of life (<20 years) are presumed to represent modeling-related growth. Differences between young men and young women (20-35 years) were assessed using a t-test or Wilcoxon’s rank-sum test depending on each variable’s distribution. The overall relationship

99 between age and each parameter was investigated through either Pearson or Spearman correlations.

Age-related changes in histomorphometric parameters and the differences between men and women were examined using linear regression. Model assumptions were checked graphically (histograms, scatter plots, and quantile-quantile plots), and variable transformations were performed using the box-cox technique (Osborne, 2010). Regression models tested for possible nonlinear associations by incorporating a quadratic age term (age2). To address issues of collinearity, “age-centered” (age – mean age) and “age-centered squared” were employed. “Sex”, “sex- and age-centered”, and “sex- and age-centered squared” interaction terms were included to assess sex-specific comparisons with age. Model fit was assessed using R2 coefficients, graphical analysis of residuals, and the Akaike information criterion (AIC). Predicted values were used to estimate age-related changes across the lifespan for both sexes. We report the estimated mean and 95% prediction interval at age 80 years. Similar statistical approaches have been used in population-based HR-pQCT studies (Khosla et al., 2006; Dalzell et al., 2009; Macdonald et al., 2011; Hansen et al., 2014).

All statistical tests were performed in either SPSS v. 21 or Microsoft Excel.

100

Table 8: Cortical and trabecular bone parameters (Descriptions modified from Vajda et al., 1995; Bloebaum et al., 1997; Dempster et al., 2013; Hunter and Agnew, 2016)

101

Table 9: Descriptive statistics for cortical parameters for the study sample by sex and age cohort (mean ± SD if normally distributed, median (IQR) if non-normally distributed)

Men Ct.B.Ar Po.Dn Ct.Ot.Lc.N Ct.Ot.Lc.Ar Ct.Ot.Lc.Dn Age Cohort n (mm2) Po.Ar (um2) Po.Dm (um) Po.N (#) (#/mm2) Ct.Po (%) (#) (mm2) (#/mm2) Ct.WMGL <20 5 23.49 (7.63) 2302.46 (3437.31) 54.14 (34.97) 629 (258) 14.71 ± 3.50 0.0857 (0.07) 2305 (621) 0.0552 (0.02) 104.48 (26.43) 166.45 ± 9.09 20-29 9 23.37 (5.97) 1973.40 (1299.55) 50.13 (15.08) 721 (234) 13.26 ± 2.99 0.0615 (0.02) 2073 (1604) 0.0501 (0.04) 106.77 (48.95) 172.94 ± 10.87 30-39 11 20.83 (5.27) 1480.35 (2311.24) 43.41 (29.02) 571 (196) 11.28 ± 3.51 0.0442 (0.06) 1786 (869) 0.0432 (0.02) 76.50 (46.91) 170.56 ± 6.244 40-49 12 19.28 (9.19) 1978.49 (1597.08) 50.18 (9.77) 608 (284) 10.63 ± 3.42 0.0676 (0.03) 1528 (1305) 0.0406 (0.03) 70.24 (25.98) 175.45 ± 7.42 50-59 11 20.39 (6.74) 1665.02 (1215.10) 46.04 (15.58) 661 (103) 10.99 ± 2.56 0.0480 (0.04) 1361 (563) 0.0337 (0.02) 62.83 (18.17) 177.69 ± 8.25 60-69 10 19.75 (5.62) 2075.20 (1367.39) 51.40 (15.70) 674 (74) 11.52 ± 3.57 0.0747 (0.04) 1356 (658) 0.0346 (0.02) 68.07 (36.37) 170.29 ± 8.34 70+ 10 20.33 (10.21) 1801.56 (1200.97) 47.88 (14.89) 657 (339) 10.30 ± 1.75 0.0632 (0.04) 1500 (958) 0.0381 (0.03) 75.27 (22.11) 168.46 ± 4.67 Women Ct.B.Ar Po.Dn Ct.Ot.Lc.N Ct.Ot.Lc.Ar Ct.Ot.Lc.Dn Age Cohort n (mm2) Po.Ar (um^2) Po.Dm (um) Po.N (#) (#/mm2) Ct.Po (%) (#) (mm2) (#/mm2) Ct.WMGL <20 3 19.27 ± 3.31 4177.34 (N/A) 72.93 (N/A) 530 (N/A) 16.59 ± 4.41 0.1014 (N/A) 2541 (N/A) 0.0630 (N/A) 139.70 (N/A) 176.54 (N/A) 20-29 15 22.30 ± 4.06 1598.47 (991.29) 45.11 (13.98) 681 (393) 15.28 ± 3.78 0.0485 (0.01) 2715 (1486) 0.0663 (0.04) 120.21 (44.75) 183.79 (16.87) 30-39 11 22.65 ± 4.90 1489.81 (540.20) 43.55 (8.31) 654 (393) 15.91 ± 3.93 0.0475 (0.01) 2537 (966) 0.0641 (0.001) 107.37 (52.23) 184.22 (23.18) 40-49 16 19.94 ± 5.25 2086.74 (1492.49) 51.45 (17.94) 672 (171) 17.01 ± 3.90 0.0740 (0.05) 2143 (1112) 0.0538 (0.03) 94.21 (63.20) 180.31 (23.67) 50-59 11 21.33 ± 5.47 2286.19 (1498.60) 53.95 (17.24) 668 (285) 13.65 ± 2.29 0.0842 (0.05) 2216 (1395) 0.0554 (0.03) 103.12 (42.17) 173.44 (5.10) 60-69 14 16.92 ± 3.78 2401.10 (1766.56) 55.30 (20.56) 590 (103) 13.18 ± 4.07 0.0884 (0.06) 1344 (1153) 0.0351 (0.03) 88.11 (62.75) 172.72 (17.80) 70+ 5 12.99 ± 3.21 1986.12 (3925.06) 50.29 (42.20) 444 (139) 11.87 ± 1.74 0.0723 (0.11) 1436 (1282) 0.0355 (0.04) 125.29 (69.64) 169.92 (10.27)

102

Table 10: Descriptive statistics for trabecular parameters for the study sample by sex and age cohort (mean ± SD if normally distributed, median (IQR) if non-normally distributed)

Men Age Tb.Ot.Lc.N Tb.Ot.Lc.Ar Tb.B.Ar Tb.Ot.Lc.Dn Tb.WMGL Cohort n (#) (mm2) (mm2) (#/mm2) <20 5 193 (384) 0.0045 (0.011) 3.91 ± 2.00 56.53 (43.72) 160.87 ± 7.96 20-29 9 275 (273) 0.0064 (0.076) 4.05 ± 1.51 59.67 (28.97) 163.04 ± 10.21 30-39 11 228 (167) 0.0055 (0.005) 4.45 ± 2.05 52.42 (24.55) 158.21 ± 5.80 40-49 12 259 (319) 0.0065 (0.008) 5.17 ± 2.03 60.49 (28.16) 162.87 ± 9.12 50-59 11 142 (78) 0.0035 (0.002) 3.71 ± 1.82 33.82 (13.46) 163.57 ± 6.95 60-69 10 179 (145) 0.0043 (0.004) 4.67 ± 1.48 38.76 (30.06) 161.28 ± 9.45 70+ 10 186 (173) 0.0045 (0.004) 4.31 ± 1.43 42.39 (34.50) 156.94 ± 6.57

Women Age Tb.Ot.Lc.N Tb.Ot.Lc.Ar Tb.B.Ar Tb.Ot.Lc.Dn Tb.WMGL Cohort n (#) (mm2) (mm2) (#/mm2) <20 3 150 (N/A) 0.0042 (N/A) 1.52 (N/A) 73.81 (N/A) 169.39 ± 7.34 20-29 15 208 (184) 0.0052 (0.005) 2.58 (1.38) 81.74 (39.66) 167.80 ± 9.60 30-39 11 137 (140) 0.0035 (0.004) 2.02 (1.29) 62.20 (27.47) 171.54 ± 12.01 40-49 16 165 (122) 0.0045 (0.004) 2.05 (1.27) 75.26 (36.79) 163.17 ± 11.66 50-59 11 300 (338) 0.0073 (0.009) 3.71 (1.90) 67.87 (46.11) 159.01 ± 7.51 60-69 14 169 (177) 0.0047 (0.005) 2.60 (1.53) 64.63 (42.56) 160.73 ± 12.52 70+ 5 221 (196.67) 0.0060 (0.010) 2.73 (2.09) 99.73 (77.71) 157.63 ± 9.00

103

Table 11: Observed outcomes at 20-35 years (mean ± SD or median (interquartile range)), and predicted outcomes at 80 years (estimated mean (95% prediction interval)).

Men (n =68) Women (n =75) Predicted Change Between 20 and 80 Years Predicted Change Between 20 and 80 Years Sex Differences

Observed Estimated Mean Abs. Observed Estimated Mean Abs. In Young With Age % Change R2 c % Change R2 c (20-35 years) (PI) at 80 Years Diff (20-35 years) (PI) at 80 Years Diff Adultsd (Sex x Age)e Cortical n =16 n =20 Ct.B.Ar (mm2)a,b 23.36 (5.93) 19.743 (11.791-31.166) -3.6168 -15.48 -0.192 22.56 (6.88) 12.645 (7.026-21.105) -9.9151 -43.95 -0.405** 0.498 0.041 Po.Ar (um2)a 1946.36 (1765.14) 1871.791 (816.956-5333.283) -74.57 -3.83 -0.056 1552.16 (893.85) 2213.636 (931.853-6684.683) 1061.5 68.39 0.182 0.095 0.167 Po.Dm (um)a,b 49.78 (21.33) 50.772 (33.143-87.385) 0.99 1.99 -0.056 44.45 (12.49) 53.17 (34.245-93.522) 8.72 19.62 0.182 0.095 0.593 Po.N (#)a,b 694 (213) 625.615 (381.751-1025.26) -23.38 -3.37 -0.035 655 (297) 475.083 (288.33-782.798) -179.92 -27.47 -0.132 0.440 0.028 Po.Dn (#/mm2)a 12.86 (4.43) 9.78 (5.686-16.823) -3.08 -23.95 -0.082 15.71 (4.72) 12.22 (7.084-21.08) -3.49 -22.22 0.333** 0.026 0.722 Ct.Po (%)a 0.0610 (0.04) 0.062 (0.026-0.149) 0.0008 1.31 -0.001 0.0481 (0.01) 0.089 (0.037-0.216) 0.0409 85.03 0.370** 0.072 0.021 Ct.Ot.Lc.N (#)a,b 2029 (988) 1458.78 (636.967-3340.918) -570.22 -28.10 -0.376** 2626 (1760) 1443.15 (624.446-3335.244) -1182.9 -45.04 -0.413** 0.072 0.105 Ct.Ot.Lc.Ar (mm2)a,b 0.0495 (0.03) 0.038 (0.016-0.088) -0.012 -24.24 -0.356** 0.0642 (0.04) 0.037 (0.016-0.088) -0.027 -42.06 -0.381** 0.140 0.119 Ct.Ot.Lc.Dn (#/mm2)a,b 88.83 (42.21) 75.058 (33.837-137.208) -13.77 -15.50 -0.335** 112.14 (48.17) 118.323 (61.568-198.518) 6.18 5.51 -0.195 0.014 0.885 Ct.WMGLa,b 171.37 (10.40) 166.059 (153.16-184.284) -5.31 -3.10 -0.008 184.51 (17.36) 169.466 (155.528-189.655) -15.04 -8.15 -0.359** 0.002 0.019 Trabecular Tb.B.Ar (mm2)a 3.95 (2.68) 4.254 (1.482-9.792) 0.30 7.59 0.025 2.56 (1.40) 2.705 (0.807-6.836) 0.15 5.86 0.166 0.004 0.820 Tb.Lc.N (#)a 257 (187) 149.06 (36.35-611.247) -107.94 -42.00 -0.206 205 (216) 178.043 (43.098-735.513) -26.96 -13.15 0.046 0.336 0.203 Tb.Ot.Lc.Ar (mm2)a 0.0059 (0.0046) 0.004 (0.001-0.016) -0.0022 -37.29 -0.194 0.0051 (0.0054) 0.005 (0.001-0.02) -0.0005 -9.80 0.045 0.386 0.196 Tb.Ot.Lc.Dn (#/mm2)a,b 58.34 (37.13) 42.066 (15.736-92.563) -16.27 -27.89 -0.355** 67.42 (41.63) 82.073 (35.957-162.614) 14.65 21.73 -0.044 0.211 0.075 Tb.WMGL 160.10 ± 8.99 159.639 (140.179-140.179) -0.46 -0.29 -0.097 169.95 ± 11.66 157.025 (137.463-176.587) -12.92 -7.60 -0.342** 0.018 0.041

aVariable required box-cox transformation bChange with age is nonlinear cPearson correlation for normally distributed variables and Spearman rank correlation for non–normally distributed variables dp values from t-test or Wilcoxon’s rank sum test dependent on distribution of parameter (significant values are in bold) eComparison of slopes (significant values are in bold) ** Significant at the 0.01 level * Significant at the 0.05 level

104

5.3.5 Results

Except for Tb.WMGL, all histomorphometric parameters required transformation. Po.Ar, Po.Dn, Ct.Po, Tb.Ot.Lc.N, Tb.Ot.Lc.Ar, and Tb.WMGL were best characterized by linear relationships with age, whereas Po.Dn, Po.N, Ct.Ot.Lc.N, Ct.Ot.Lc.Ar, Ct.Ot.Lc.Dn, Tb.Ot.Lc.Dn, and Ct.WMGL were best characterized by quadratic relationships with age. Summarized data for young men and women (20-35 years) and predicted age-related changes are shown in Table 10. Regression plots for the cortical and trabecular bone compartments are shown in Figures 13a-f, 14g-l, 15a-f, 16a-f, and 17a-d.

Values for Ct.B.Ar are similar in young adult men and women, but only decrease significantly with age in women (-44%). The sex difference with age is also significant, with women demonstrating a more accelerated rate of decline beginning in middle adulthood. While men have significantly greater Tb.B.Ar in young adulthood (p=0.004), rates of age-related bone loss are similar in both sexes.

Po.Ar and Po.Dm do not demonstrate any notable relationships with age in either men or women. The age-related sex difference in Po.N is attenuated when the data are standardized by Ct.B.Ar. Po.Dn decreases (-22%) and Ct.Po increases (85%) significantly with age in women. There are no discernable trends for these variables in men. Young adult women have higher Po.Dn than their male counterparts (p=0.026), but rates of change do not significantly differ with age. Ct.Po markedly increases during the 5th decade of life in women, consistent with perimenopausal and menopausal bone loss.

Ct.Ot.Lc.N and Ct.Ot.Lc.Ar decrease with age in men (-24% and -16%, respectively) and women (-45%, and -42%, respectively), but only men exhibit a substantial reduction in Ct.Ot.Lc.Dn (- 16%) and Tb.Ot.Lc.Dn (-28%) with age. Men have lower Ct.Ot.Lc.Dn in young adulthood (p=0.014), but the sex difference with age is non-significant.

Ct.WMGL and Tb.WMGL decrease significantly with age in women (-8% and -8%, respectively). Women have much more highly mineralized bone in young adulthood (p=0.002 and p=0.018), and age-related changes to bone mineralization differ significantly between men and women. Men achieve peak WMGL values after women, and undergo a steeper decline in

105

Ct.WMGL only after the 6th decade of life. However, the pattern for Tb.WMGL remains relatively static throughout adult life in men.

5.4 Discussion

5.4.1 Intracortical Porosity

Early work suggested that the ratio of cortical porosity to bone area remains relatively constant across the adult lifespan in the human rib (Frost, 1963), but this assertion has been contested. Newer research has not identified any sex-based differences, but studies demonstrate a linear increase in Ct.Po with age (Dominguez and Agnew, 2016; Hunter and Agnew, 2016). This relationship has been described to varying degrees in men and women at other anatomical locations (Brockstedt et al., 1993; Feik et al., 1997; Bousson et al., 2001; Burghardt et al., 2010; Macdonald et al., 2011; Kazakia et al., 2013; Hansen et al., 2014; Bach-Gansmo et al., 2016; Vilayphiou et al., 2016), but direct comparisons to our results may not be appropriate given differences in mechanical loading between skeletal elements.

Only women demonstrate a significant increase in Ct.Po with age in our study sample. In men, the general pattern is similar but occurs later, and age-related fluctuations are much less extreme. Following expectations, Ct.Po is high during growth and development (Sedlin et al., 1963; Epker and Frost, 1965a; b; Frost, 1969; Agnew et al., 2013), when there is extensive modelling-related tissue drift. Values drop to a minimum in early adulthood, when peak bone mass is established (Agnew et al., 2015). Thereafter, Ct.Po slowly increases as the rib continues to remodel, until it tapers off in the decades following menopause. It is believed that the age-related increase in Ct.Po is mostly attributed to a coalescence of pores (i.e. increased Po.Ar), with only a small percentage explained by an increase in Po.Dn (Stein et al., 1999; Thomas et al., 2006; Cooper et al., 2007). In this study, Po.Dn actually decreased in women, and the correlation between age and Po.Ar was non-significant. These somewhat contradictory findings with increased Ct.Po can be reconciled by the fact that the oldest women in this sample had smaller average pore size compared to the three decades prior, but the greatest overall variation in pore size (Table 9).

106

107

108

109

5.4.2 Osteocyte Lacunar Density

Estrogen has anti-apoptotic effects on osteocytes (Tomkinson et al., 1997; Clarke and Khosla, 2010), leading to the expectation of higher densities of lacunae in young adult women. Similarly, osteoblast apoptosis triggered by sex steroid withdrawal may result in reduced osteocyte densities (i.e. fewer cell to differentiate into mature osteocytes) after menopause (Bradford et al., 2011). While the women in our study demonstrate greater lacunar abundance across the entire adult lifespan compared to men, the lack of significant relationships with age supports a paradigm shift away from estrogen-centric effects on this aspect of bone quality (Manolagas and Parfitt, 2010; Hunter and Agnew, 2016). In agreement, Qiu et al. (2002) also did not find a relationship between osteocyte lacunar density and menopause.

The number of osteocyte lacunae per unit of bone area only decreases significantly with age in men, possibly indicating a higher likelihood of damageability (Qiu et al., 2005; Ma et al., 2008). A decreased osteocyte lacunar density may disrupt the canalicular fluid flow, reducing the detection of microdamage and subsequent reparative efforts (Busse et al., 2010a). Osteoporotic bone has been found to contain fewer osteocyte lacunae than controls (Qiu et al., 2003b; Mullender et al., 2005; Zarrinkalam et al., 2012), although this relationship is somewhat controversial (Mullender et al., 1996; Oliveira et al., 2016). Numerous studies report age-related declines in osteocyte lacunar density to varying degrees in both sexes (Mullender et al., 1996, 2005, Qiu et al., 2002a; b, 2006; Vashishth et al., 2002; Busse et al., 2010a; Bernhard et al., 2013; Bach-Gansmo et al., 2016; Hunter and Agnew, 2016). However, others have also demonstrated the inverse relationship (Vashishth et al., 2005), or no relationship at all (Carter et al., 2013, 2014). The predictive ability of an individual's chronological age on osteocyte lacunar density is clearly limited. It may be that a decline in the lacunar-canalicular areal fraction correlates much more strongly with age (Ashique et al., 2017).

Substantial variation in osteocyte lacunar density has been reported for human bone. Lower osteocyte lacunar densities in this study, as compared to Bromage et al. (2016), may relate to the higher resolution of these colleagues’ captured BSE-SEM images. While lacunae diameters average 8 um, they can vary between 3 um to 20 um (Hannah et al., 2010). Our methods permitted the detection of lacunae greater than 4 um in diameter. Failure to capture the smallest osteocytes has an unknown impact on our results. As osteopenic osteocytes are thought to be

110 relatively large (van Hove et al., 2009), the error may not be substantial. However, it is still possible that men have higher densities of small osteocyte lacunae, which may help to explain the unexpected sex differences across the adult lifespan. Study differences may also relate to the younger average age of the sample used by Bromage et al. (2016), or to real differences in osteocyte lacunar densities between the femur and rib. Although Hunter and Agnew (2016) did not find any significant differences between osteocyte parameters in these bones, relatively fewer osteocytes are expected in the rib as they experience categorically less mechanical loading (Tommerup et al., 1993; Robling and Stout, 2003; Pfeiffer et al., 2006). Higher osteocyte densities are observed in the periosteal cortex relative to the endosteal cortex (Qiu et al., 2002b; Busse et al., 2010b), as torsion and bending stresses decrease with increased proximity to the center of a cross-section (Carter and Beaupré, 2007). Others have also demonstrated decreased osteocyte lacunar density in weightlessness and disuse (Aguirre et al., 2006; Britz et al., 2012).

Using traditional light microscopy, both Hunter and Agnew (2016) and Qiu et al. (2003a) report average osteocyte lacunar densities upwards of 848/mm2 in the human rib. The extremely thin imaging plane in this study, compared to a Z-range of 50-120 um in a typical histological thin- section, largely explains discrepancies among studies utilizing 2-D imaging methods. The latter method records osteocyte lacunar densities throughout the entire section thickness, such that it overestimates true osteocyte lacunar values. Standardization of imaging methods, osteocyte parameters, and measurement protocols is warranted.

111

112

5.4.3 Degree of Bone Mineralization

Existing data on rib tissue mineralization are scarce, and sample sizes are insufficient to access sex- and age-related variation. Calculating the ratio of ash weight to dry weight, Kemper et al. (2005, 2007) did not find any significant changes to percent mineralization with age, but both of these studies only had a sample of six individuals. Using BSE-SEM, Reid and Boyde (1987) found a positive correlation between age and bone mineralization density, but their sample size was limited to thirteen individuals and did not cover the more advanced decades. These data do not necessarily conflict with our results as the average degree of mineralization increases until approximately 40 years in women, and about 45 years in men. Age-related patterns are similar in the cortical and trabecular bone compartments, but the former is more highly mineralized. Our data on peak bone mineralization are consistent with metacarpal radiometry work, which gives peak bone mass values between 45 and 50 years for the SAB population group (Solomon, 1979).

Low bone mineralization is associated with osteoporosis in post-menopausal women (Roschger et al., 2008), but may relate to calcium and vitamin D deficiencies or high dietary acid load (Brickley et al., 2007; Heaney and Layman, 2008). Co-morbid factors such as HIV can further compromise BMD in both men and women (Yin et al., 2005; Kruger and Nell, 2017). Low bone mineralization has also been associated with increased microcrack density and diffuse damage area (Norman et al., 2008).

In our study, lower average densities in males across the entire adult lifespan suggest that they may have been at higher risk of osteoporosis and fragility fracture. However, BSE-SEM work conducted on the femur and ilium reveal similar findings, complicating the interpretation of these results. At the femoral mid-shaft location, Goldman et al. (2003) found that age-related decrements to bone mineralization were most apparent in males, particularly between the middle and older age cohorts. In the iliac crest, normal males appear to have higher concentrations of low and medium density bone, while normal females have proportionately more highly mineralized bone (Boyde et al., 1995). Although Koehne et al. (2014) report a positive correlation between age and tissue mineralization in the trans-iliac crest, they also identified higher values in women compared to men. High tissue mineralization in women alternatively may suggest a lower rate of bone renewal and less efficient remodeling (Goldman et al., 2003).

113

However, our porosity data, coupled with our previous work on osteon population density (OPD; Chapter 4), suggests that this is not the case.

Average cortical WMGL levels are higher in our study compared to Goldman et al. (2003) , but are similar to values reported in Zeininger et al. (2011), which respectively analyzed the femur and third metacarpal. These results are surprising, given the comparatively thin rib cortex, but could be explained by differential BSE-SEM imaging conditions. The relevance of these studies to rib tissue mineralization is also uncertain given differences in load history. However, systemic remodeling likely has similar effects across the entire skeleton. For example, elite female rowers with rib pain are more likely to have low BMD Z-scores at the lumbar spine and femoral neck (Dimitriou et al., 2014), indicating heavy upper body exercise left uncompensated by adequate nutrition can have measurable full body effects. The standardization of BSE-SEM imaging protocols and additional research exploring intra-individual variation between skeletal elements, is needed.

114

Figure 17a-d: Cortical and trabecular bone mineralization. Predicted age-related changes in cortical WMGL (a-b), and trabecular WMGL (c-d) for men and women. The solid line represents the fitted mean from the regression model, and the dashed lines represent the 95% confidence interval of the prediction. Men are represented by open circles, and women are represented by open triangles.

115

5.4.4 South African Apartheid and Consequences to Bone Health

This study examined bone quality in a sample of twentieth century South Africans who had experienced difficult lives. Although our rib data are non-normative, they serve to define the range of bone histomorphometric variation that developed under adverse apartheid living conditions. Health status is assumed to be relatively poor for most individuals in this sample, given available epidemiological data for non-whites living on the Western Cape at this time (Williams et al., 2008; Inwood and Masakure, 2013; Mpeta and Inwood, 2017). This presumption is also supported by cadaver donation sources, cause of death information, and previous research conducted on the broader Kirsten Skeletal Collection (Labuschagne and Mathey, 2000; Geldenhuys et al., 2016; Alblas et al., 2018).

Sex-based differences may be exacerbated by the systematic marginalization of non-white male population groups. Heavy alcohol consumption is a longstanding endemic in South Africa, wherein men are more likely to have a history of chronic alcohol abuse (Parry et al., 2005; Schneider et al., 2007; Peltzer et al., 2011; Gossage et al., 2014). Gender bias in alcohol consumption is not unique to South Africa (Wilsnack et al., 2009). Globally, it is one of the major causes of osteoporosis in men (Kim et al., 2003; Giusti and Bianchi, 2015; Misiorowski, 2017). Higher use has been shown to decrease BMD (Malik et al., 2009; Maurel et al., 2012), and increase the number of empty osteocyte lacunae (Maurel et al., 2014). With continued high consumption, men may be more susceptible to BMD decrements than their female counterparts (Malik et al., 2009).

South Africa’s continued issues with problem drinking can partially be attributed to the legacy of the “Dop” system, in which farm laborers were given alcohol as a condition or benefit of employment (London 1999, 2000). During the apartheid era, middle-aged SAB males constituted a high proportion of clinical osteoporosis cases; researchers scapegoated the cultural practice of drinking sorghum beer brewed in iron drums (Grusin and Samuel, 1957; Seftel et al., 1966; Lynch et al., 1967, 1970; Wapnick et al., 1971; Schnitzler et al., 1994, 2005). Heavy weekend binge drinking was reportedly highest among the SAC and SAB groups. Other risk factors included low levels of education and living in rural areas (Parry et al., 2005). Alcohol use and cigarette smoking occurred at high prevalence rates in trauma patients (Peden et al., 2000), and

116 alcohol-related violence was most common among the SAC group (Butchart and Brown, 1991; Geldenhuys et al., 2016).

5.5 Conclusion

The objective of this research was to explore age- and sex-related patterns in the structural and material properties of mid-thoracic ribs to help build an understanding of apartheid health conditions in predominately non-white South Africans. Results echo our previous work on bone mass and microstructure in a similar subset of the Kirsten Skeletal Collection (Pfeiffer et al., 2016; Chapter 4). While age associations are generally stronger in women, the men in our study also demonstrate indicators of poor bone quality. Compared to their female counterparts, they exhibit lower bone mineralization and osteocyte lacunar densities throughout the entire adult lifespan.

This study also demonstrates the relevance of skeletal biology research on curated and forensic samples for contemporary biomedical issues. Increased global prevalence, particularly in urbanized areas of developing countries, has made osteoporosis an emerging public health concern (Handa et al., 2008). While a sizeable proportion of the population is potentially at risk, osteoporosis research continues to be under-prioritized in South Africa (Bateman, 2006).

5.6 Literature Cited

Agnew AM, Moorhouse K, Kang YS, Donnelly BR, Pfefferle K, Manning AX, Litsky AS, Herriott R, Abdel-Rasoul M, Bolte IV JH. 2013. The response of pediatric ribs to quasi- static loading: Mechanical properties and microstructure. Ann Biomed Eng 41:2501–2514. Agnew AM, Schafman M, Moorhouse K, White SE, Kang Y-S, Balasch J, Bala Y, Seeman E, Abràmoff MD, Magalhães PJ, Ram SJ, Matsuo K, Irie N, Boyce BF, Xing L, Agnew AM, Stout SD, Bala Y, Depalle B, Farlay D, Douillard T, Meille S, Follet H, Chapurlat R, Chevalier J, Boivin G, Ascenzi M-GG, Lomovtsev A, Nakamura I, Takahashi N, Jimi E, Udagawa N, Suda T, Bajaj D, Geissler JR, Allen MR, Burr DB, Fritton JC, Augat P, Schorlemmer S, Ascenzi M-GG, Drife JO, Atkins GJ, Findlay DM, Ascenzi M-GG, Liao VP, Lee BM, Billi F, Zhou H, Lindsay R, Cosman F, Nieves J, Bilezikian JP, Dempster DW. 2015. The effect of age on the structural properties of human ribs. J Mech Behav Biomed Mater [Internet] 41:302–314. Available from: http://link.springer.com/10.1007/s00223-015-9971-y Agnew AM, Stout SD. 2012. Brief communication: Reevaluating osteoporosis in human ribs: The role of intracortical porosity. Am J Phys Anthropol 148:462–466. Aguirre JI, Plotkin LI, Stewart SA, Weinstein RS, Parfitt AM, Manolagas SC, Bellido T. 2006. Osteocyte Apoptosis Is Induced by Weightlessness in Mice and Precedes Osteoclast Recruitment and Bone Loss. J Bone Miner Res [Internet] 21:605–615. Available from:

117

http://doi.wiley.com/10.1359/jbmr.060107 Andersson N, Marks S. 1988. Apartheid and Health in the 1980s. Soc Sci Med 27:667–681. Andronowski JM, Mundorff AZ, Pratt I V., Davoren JM, Cooper DML. 2017. Evaluating differential nuclear DNA yield rates and osteocyte numbers among human bone tissue types: A synchrotron radiation micro-CT approach. Forensic Sci Int Genet [Internet] 28:211–218. Available from: http://dx.doi.org/10.1016/j.fsigen.2017.03.002 Ashique AM, Hart LS, Thomas CDL, Clement JG, Pivonka P, Carter Y, Mousseau DD, Cooper DML. 2017. Lacunar-canalicular network in femoral cortical bone is reduced in aged women and is predominantly due to a loss of canalicular porosity. Bone Reports [Internet] 7:9–16. Available from: http://dx.doi.org/10.1016/j.bonr.2017.06.002 Bach-Gansmo FL, Brüel A, Jensen MV, Ebbesen EN, Birkedal H, Thomsen JS. 2016. Osteocyte lacunar properties and cortical microstructure in human iliac crest as a function of age and sex. Bone [Internet] 91:11–19. Available from: http://dx.doi.org/10.1016/j.bone.2016.07.003 Bateman C. 2006. South Africa under-prioritises osteoporosis. South African Med J 96:19–20. Bellido T. 2015. Osteocyte-driven bone remodeling. Calcif Tissue Int 94:25–34. Bernhard A, Milovanovic P, Zimmermann EA, Hahn M, Djonic D, Krause M, Breer S, Püschel K, Djuric M, Amling M, Busse B. 2013. Micro-morphological properties of osteons reveal changes in cortical bone stability during aging, osteoporosis, and bisphosphonate treatment in women. Osteoporos Int [Internet] 24:2671–80. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23632826 Bloebaum RD, Skedros JG, Vajda EG, Bachus KN, Constantz BR. 1997. Determining mineral content variations in bone using backscattered electron imaging. Bone 20:485–490. Bousson V, Meunier A, Bergot C, Vicaut E, Rocha M a, Morais MH, Laval-Jeantet a M, Laredo JD. 2001. Distribution of intracortical porosity in human midfemoral cortex by age and gender. J Bone Miner Res 16:1308–1317. Boyde A. 2012. Scanning Electron Microscopy of Bone. In: Helfrich MH, Ralston SH, editors. Bone Research Protocols, Methods in Molecular Biology. Vol. 816. Springer Science and Business Media. p 365–400. Boyde A, Jones SJ, Aerssens J, Dequeker J. 1995. Mineral density quantitation of the human cortical iliac crest by backscattered electron image analysis: Variations with age, sex, and degree of osteoarthritis. Bone 16:619–627. Bradford PG, Gerace K V, Roland RL, Chrzan BG. 2011. Estrogen regulation of apoptosis in osteoblasts. Physiol Behav 99:181–185. Brickley MB, Mays S, Ives R. 2007. An investigation of skeletal indicators of vitamin D deficiency in adults: Effective markers for interpreting past living conditions and pollution levels in 18th and 19th century Birmingham, England. Am J Phys Anthropol 132:67–79. Britz HM, Carter Y, Jokihaara J, Leppänen O V., Järvinen TLN, Belev G, Cooper DML. 2012. Prolonged unloading in growing rats reduces cortical osteocyte lacunar density and volume in the distal tibia. Bone [Internet] 51:913–919. Available from: http://dx.doi.org/10.1016/j.bone.2012.08.112 Brockstedt H, Kassem M, Eriksen EF, Mosekilde L, Melsen F. 1993. Age- and sex-related changes in iliac cortical bone mass and remodeling. Bone 14:681–691. Bromage TG, Juwayeyi YM, Katris JA, Gomez S, Ovsiy O, Goldstein J, Janal MN, Hu B, Schrenk F. 2016. The scaling of human osteocyte lacuna density with body size and metabolism. Comptes rendus - Palevol [Internet] 15:32–39. Available from: http://dx.doi.org/10.1016/j.crpv.2015.09.001

118

Burghardt AJ, Kazakia GJ, Ramachandran S, Link TM, Majumdar S. 2010. Age- and Gender- Related Differences in the Geometric Properties and Biomechanical Significance of Intracortical Porosity in the Distal Radius and Tibia. J Bone Miner Res 25:983–993. Busse B, Djonic D, Milocanovic P, Hahn M, Puschel K, Ritchie RO, Djuric M, Amling M. 2010a. Decrease in the osteocyte lacunar density accompanied by hypermineralized lacunar occlusion reveals failure and delay of remodeling in aged human bone. Aging Cell 9:1065– 1075. Busse B, Djonic D, Milovanovic P, Hahn M, Püschel K, Ritchie RO, Djuric M, Amling M. 2010b. Decrease in the osteocyte lacunar density accompanied by hypermineralized lacunar occlusion reveals failure and delay of remodeling in aged human bone. Aging Cell 9:1065– 1075. Butchart A, Brown DSO. 1991. Non-fatal injuries due to interpersonal violence in Johannesburg - Soweto: Incidence, determinants and consequences. Forensic Sci Int 52:35–51. Cardoso L, Herman BC, Verborgt O, Laudier D, Majeska RJ, Schaffler MB. 2009. Osteocyte apoptosis controls activation of intracortical resorption in response to bone fatigue. J Bone Miner Res [Internet] 24:597–605. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2659511&tool=pmcentrez&ren dertype=abstract Carter D, Beaupré G. 2007. Skeletal Function and Form: Mechanobiology of Skeletal Development, Aging, and Generation. New York, NY: Cambridge University Press. Carter Y, Suchorab JL, Thomas CDL, Clement JG, Cooper DML. 2014. Normal variation in cortical osteocyte lacunar parameters in healthy young males. J Anat 225:328–336. Carter Y, Thomas CDL, Clement JG, Cooper DML. 2013. Femoral osteocyte lacunar density, volume and morphology in women across the lifespan. J Struct Biol 183:519–526. Cauley JA, Chalhoub D, Kassem AM, Fuleihan GE-H. 2014. Geographic and ethnic disparities in osteoporotic fractures. Nat Rev Endocrinol [Internet] 10:338–351. Available from: http://www.nature.com/doifinder/10.1038/nrendo.2014.51 Clarke BL, Khosla S. 2010. Female reproductive system and bone. Arch Biochem Biophys 503:118–28. Cooper DML, Thomas CDL, Clement JG, Turinsky AL, Sensen CW, Hallgrímsson B. 2007. Age-dependent change in the 3D structure of cortical porosity at the human femoral midshaft. Bone 40:957–965. Crowder C, Heinrich J, Stout SD. 2012. Rib histomorphometry for adult age estimation. In: Bell LS, editor. Forensic Microscopy for Skeletal Tissues: Methods and Protocols, Methods in Molecular Biology. Vol. 915. Springer Science+Business Media. p 109–127. Available from: http://link.springer.com/10.1007/978-1-61779-977-8 Cumming R, Nevitt M, Cummings SR. 1997. Epidemiology of hip fractures. Epidemiol Rev 19:244–257. Dalzell N, Kaptoge S, Morris N, Berthier A, Koller B, Braak L, Van Rietbergen B, Reeve J. 2009. Bone micro-architecture and determinants of strength in the radius and tibia: Age- related changes in a population-based study of normal adults measured with high-resolution pQCT. Osteoporos Int 20:1683–1694. Daniels ED, Pettifor JM, Schnitzler CM, Moodley GP, Zachen D. 1997. Differences in mineral homeostasis, volumetric bone mass and femoral neck axis length in black and white South African women. Osteoporos Int 7:105–112. Dempster DW, Compston JE, Drezner MK, Glorieux FH, Kanis JA, Malluche H, Meunier PJ, Ott SM, Recker RR, Parfitt AM. 2013. Standardized Nomenclature, Symbols, and Units for

119

Bone Histomorphometry: A 2012 Update of the Report of the ASBMR Histomorphometry Nomenclature Committee. J Bone Miner Res 28:1–16. Dent CE, Engelbrecht HE, Godfrey RC. 1968. Osteoporosis of lumbar vertebrae and calcification of abdominal aorta in women living in Durban. Br Med J [Internet] 4:76–79. Available from: http://www.ncbi.nlm.nih.gov/pubmed/5696550 Dimitriou L, Weiler R, Lloyd-Smith R, Turner A, Heath L, James N, Reid A. 2014. Bone mineral density, rib pain and other features of the female athlete triad in elite lightweight rowers. BMJ Open [Internet] 4:e004369. Available from: http://bmjopen.bmj.com/lookup/doi/10.1136/bmjopen-2013-004369 Dominguez VM, Agnew AM. 2016. Examination of factors potentially influencing osteon size in the human rib. Anat Rec 299:313–324. Epker BN, Frost HM. 1965a. The Direction of Transverse Drift of Actively Forming Osteons in Human Rib Cortex. J Bone Joint Surg Am 47:1211–1215. Epker BN, Frost HM. 1965b. A histological study of remodeling at the periosteal, haversian canal, cortical endosteal, and trabecular endosteal surfaces in human rib. Anat Rec 152:129–135. Feik SA, Thomas CD, Clement JG. 1997. Age-related changes in cortical porosity of the midshaft of the human femur. J Anat 191:407–16. Frost HM. 1969. Tetracycline-based histological analysis of bone remodeling. Calcif Tissue Res [Internet] 3:211–237. Available from: http://link.springer.com/10.1007/BF02058664 Geldenhuys E-M, Burger EH, Alblas A, Greyling LM, Kotze SH. 2016. The association between healed skeletal fractures indicative of interpersonal violence and alcoholic liver disease in a cadaver cohort from the Western Cape, South Africa. Alcohol 52:41–48. Giusti A, Bianchi G. 2015. Treatment of primary osteoporosis in men. Clin Interv Aging [Internet] 10:105–15. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=4283986&tool=pmcentrez&ren dertype=abstract Goldman HM, Blayvas A, Boyde A, Howell PGT, Clement JG, Bromage TG. 2000. Correlative Light and Backscattered Electron Microscopy of Bone — Part II: Automated Image Analysis. Scanning 22:337–344. Goldman HM, Bromage TG, Boyde A, Thomas CDL, Clement JG. 2003. Intrapopulation variability in mineralization density at the human femoral mid-shaft. J Anat 203:243–255. Goldman HM, Thomas CDL, Clement JG, Bromage TG. 2005. Relationships among microstructural properties of bone at the human midshaft femur. J Anat 206:127–139. Gossage PJ, Snell CL, Parry CDH, Marais AS, Barnard R, de Vries M, Blankenship J, Seedat S, Hasken JM, May PA. 2014. Alcohol use, working conditions, job benefits, and the legacy of the “dop” system among farm workers in the Western Cape Province, South Africa: Hope despite high levels of risky drinking. Int J Environ Res Public Health 11:7406–7424. Grusin H, Samuel MD. 1957. A syndrome of osteoporosis in Africans and its relationship to scurvy. Am J Clin Nutr 5:644–650. Handa R, Ali Kalla A, Maalouf G. 2008. Osteoporosis in developing countries. Best Pract Res Clin Rheumatol 22:693–708. Hannah KM, Thomas CDL, Clement JG, Carlo F De, Peele AG. 2010. Bimodal distribution of osteocyte lacunar size in the human femoral cortex as revealed by micro-CT. Bone [Internet] 47:866–871. Available from: http://dx.doi.org/10.1016/j.bone.2010.07.025 Hansen S, Shanbhogue V, Folkestad L, Nielsen MMF, Brixen K. 2014. Bone microarchitecture and estimated strength in 499 adult Danish women and men: A cross-sectional, population-

120

based high-resolution peripheral quantitative computed tomographic study on peak bone structure. Calcif Tissue Int 94:269–281. Heaney RP, Layman DK. 2008. Amount and type of protein influences bone health. Am J Clin Nutr 87:1567–1570. Hemmatian H, Bakker AD, Klein-Nulend J, van Lenthe GH. 2017. Aging, Osteocytes, and Mechanotransduction. Curr Osteoporos Rep 15:401–411. van Hove RP, Nolte PA, Vatsa A, Semeins CM, Salmon PL, Smit TH, Klein-Nulend J. 2009. Osteocyte morphology in human tibiae of different bone pathologies with different bone mineral density - Is there a role for mechanosensing? Bone [Internet] 45:321–329. Available from: http://dx.doi.org/10.1016/j.bone.2009.04.238 Hunter RL, Agnew AM. 2016. Intraskeletal variation in human cortical osteocyte lacunar density: Implications for bone quality assessment. Bone Reports [Internet] 5:252–261. Available from: http://dx.doi.org/10.1016/j.bonr.2016.09.002 Inwood K, Masakure O. 2013. Poverty and Physical Well-being among the Coloured Population in South Africa. Econ Hist Dev Reg [Internet] 28:56–82. Available from: http://www.tandfonline.com/doi/abs/10.1080/20780389.2013.866382 Jones CG. 2012. Scanning electron microscopy: Preparation and imaging for SEM. In: Bell LS, editor. Forensic Microscopy for Skeletal Tissues. Vol. 915. Springer Science+Business Media, LLC. p 1–20. Available from: http://link.springer.com/10.1007/978-1-61779-977-8 Kazakia GJ, Nirody JA, Bernstein G, Sode M, Burghardt AJ, Majumdar S. 2013. Age-and gender-related differences in cortical geometry and microstructure: Improved sensitivity by regional analysis. Bone 52:623–631. Kemper AR, McNally C, Kennedy EA, Manoogian SJ, Rath AL, Ng TP, Stitzel JD, Smith EP, Duma SM. 2005. Material Properties of Human Rib Cortical Bone from Dynamic Tension Coupon Testing. Stapp Car Crash J 49:199–230. Kemper AR, Mcnally C, Pullins CA, Freeman LJ, Duma SM, Rouhana SW. 2007. The Biomechanics of Human Ribs : Material and Structural Properties from Dynamic Tension and Bending Tests. Stapp Car Crash J 51:235–273. Khosla S, Riggs BL, Atkinson EJ, Oberg AL, McDaniel LJ, Holets M, Peterson JM, Melton LJ. 2006. Effects of sex and age on bone microstructure at the ultradistal radius: a population- based noninvasive in vivo assessment. J Bone Miner Res 21:124–131. Kim MJ, Shim MS, Kim MK, Lee Y, Shin YG, Chung CH, Kwon SO. 2003. Effect of chronic alcohol ingestion on bone mineral density in males without liver cirrhosis. Korean J Intern Med [Internet] 18:174–80. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=4531624&tool=pmcentrez&ren dertype=abstract Klein-Nulend J, Bakker AD, Bacabac RG, Vatsa A, Weinbaum S. 2013. Mechanosensation and transduction in osteocytes. Bone [Internet] 54:182–190. Available from: http://dx.doi.org/10.1016/j.bone.2012.10.013 Koehne T, Vettorazzi E, Küsters N, Lüneburg R, Kahl-nieke B, Püschel K, Amling M, Busse B. 2014. Trends in trabecular architecture and bone mineral density distribution in 152 individuals aged 30 – 90 years. Bone [Internet] 66:31–38. Available from: http://dx.doi.org/10.1016/j.bone.2014.05.010 Kruger MJ, Nell TA. 2017. Bone mineral density in people living with HIV: a narrative review of the literature. AIDS Res Ther [Internet] 14:35. Available from: http://aidsrestherapy.biomedcentral.com/articles/10.1186/s12981-017-0162-y Labuschagne BCJ, Mathey B. 2000. Cadaver profile at University of Stellenbosch Medical

121

School, South Africa, 1956 – 1996. Clin Anat 13:88–93. Lynch SR, Berelowitz I, Seftel HC, Miller GB, Krawitz P, Bothwell TH. 1967. Osteoporosis in Johannesburg Bantu Males: Its Relationship to Siderosis and Ascorbic Acid Deficiency. Am J Clin Nutr 20:799–807. Lynch SR, Seftel HC, Wapnick AA, Charlton RW, Bothwell TH. 1970. Some aspects of calcium metabolism in normal and osteoporotic Bantu subjects with special reference to the effects of iron overload and ascorbic acid depletion. S Afr J Med Sci 35:45–56. Ma YL, Dai RC, Sheng ZF, Jin Y, Zhang YH, Fang LN, Fan HJ, Liao EY. 2008. Quantitative associations between osteocyte density and biomechanics, microcrack and microstructure in OVX rats vertebral trabeculae. J Biomech 41:1324–1332. Macdonald HM, Nishiyama KK, Kang J, Hanley D a, Boyd SK. 2011. Age-related patterns of trabecular and cortical bone loss differ between sexes and skeletal sites: a population-based HR-pQCT study. J Bone Miner Res [Internet] 26:50–62. Available from: http://www.ncbi.nlm.nih.gov/pubmed/20593413 Malik P, Gasser RW, Kemmler G, Moncayo R, Finkenstedt G, Kurz M, Fleischhacker WW. 2009. Low bone mineral density and impaired bone metabolism in young alcoholic patients without liver cirrhosis: A cross-sectional study. Alcohol Clin Exp Res 33:375–381. Manolagas SC, Parfitt a M. 2010. What old means to bone. Trends Endocrinol Metab [Internet] 21:369–74. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2880220&tool=pmcentrez&ren dertype=abstract Maurel DB, Benaitreau D, Jaffré C, Toumi H, Portier H, Uzbekov R, Pichon C, Benhamou CL, Lespessailles E, Pallu S. 2014. Effect of the alcohol consumption on osteocyte cell processes: A molecular imaging study. J Cell Mol Med 18:1680–1693. Maurel DB, Boisseau N, Benhamou CL, Jaffre C. 2012. Alcohol and bone: Review of dose effects and mechanisms. Osteoporos Int 23:1–16. McCalden RW, McGeough JA, Baker MB, Court-Brown CM. 1993. Age-related changes in the tensile properties of cortical bone. J Bone Jt Surg 75:1193–1205. Misiorowski W. 2017. Osteoporosis in men. Menopause Rev 16:70–73. Misof BM, Dempster DW, Zhou H, Roschger P, Fratzl-Zelman N, Fratzl P, Silverberg SJ, Shane E, Cohen A, Stein E, Nickolas TL, Recker RR, Lappe J, Bilezikian JP, Klaushofer K. 2014. Relationship of Bone Mineralization Density Distribution (BMDD) in Cortical and Cancellous Bone Within the Iliac Crest of Healthy Premenopausal Women. Calcif Tissue Int 95:332–339. Morales NS, Catella L, Oliva F, Sarmiento PL, Barrientos G. 2017. A SEM-based assessment of bioerosion in Late Holocene faunal bone assemblages from the southern Pampas of Argentina. J Archaeol Sci Reports [Internet]:1–10. Available from: http://dx.doi.org/10.1016/j.jasrep.2017.07.012 Mpeta B, Inwood K. 2017. Black living standards in South Africa before democracy: New evidence from heights. Mullender MG, Van Der Meer DD, Huiskes R, Lips P. 1996. Osteocyte density changes in aging and osteoporosis. Bone 18:109–113. Mullender MG, Tan SD, Vico L, Alexandre C, Klein-Nulend J. 2005. Differences in osteocyte density and bone histomorphometry between men and women and between healthy and osteoporotic subjects. Calcif Tissue Int 77:291–296. Nightingale EO, Hannibal K, Geiger J, Hartmann L, Lawerence R, Spurlock J. 1990. Apartheid medicine. JAMA J … [Internet] 264:2097–2102. Available from: http://jama.ama-

122

assn.org/content/264/16/2097.short Norman TL, Little TM, Yeni YN. 2008. Age-related changes in porosity and mineralization and in-service damage accumulation. J Biomech 41:2868–2873. van Oers RFM, Wang H, Bacabac RG. 2015. Osteocyte Shape and Mechanical Loading. Curr Osteoporos Rep 13:61–66. Oliveira PS, Rodrigues JA, Shibli JA, Piattelli A, Iezzi G, Perrotti V. 2016. Influence of osteoporosis on the osteocyte density of human mandibular bone samples: A controlled histological human study. Clin Oral Implants Res 27:325–328. Osborne JW. 2010. Improving your data transformations : Applying the Box-Cox transformation. Pract Assessment, Res Eval 15:1–9. Parry CD, Pluddemann A, Steyn K, Bradshaw D, Norman R, Laubscher R. 2005. Alcohol Use in South Africa: Findings from the First Demographic and Health Survey (1998)*. J Stud Alcohol 66:91–97. Peden M, van der Spuy J, Smith P, Bautz P. 2000. Substance abuse and trauma in Cape Town. South African Med J [Internet] 90:251–255. Available from: http://www.tandfonline.com/doi/abs/10.1080/0953732032000199061 Peltzer K, Davids A, Njuho P. 2011. Alcohol use and problem drinking in South Africa: findings from a national population-based survey. Afr J Psychiatry [Internet] 14:30–37. Available from: http://dx.doi.org/10.4314/ajpsy.v14i1.65466 Pfeiffer S, Crowder C, Harrington L, Brown M. 2006. Secondary osteon and Haversian canal dimensions as behavioral indicators. Am J Phys Anthropol 131:460–468. Pfeiffer S, Heinrich J, Beresheim A, Alblas M. 2016. Cortical bone histomorphology of known- age skeletons from the Kirsten collection, Stellenbosch University, South Africa. Am J Phys Anthropol 160:137–147. Qiu S, Fyhrie DP, Palnitkar S, Rao DS. 2003a. Histomorphometric assessment of Haversian canal and osteocyte lacunae in different-sized osteons in human rib. Anat Rec [Internet] 272A:520–525. Available from: http://doi.wiley.com/10.1002/ar.a.10058 Qiu S, Rao D., Palnitkar S, Parfitt A. 2002a. Relationships between osteocyte density and bone formation rate in human cancellous bone. Bone 31:709–711. Qiu S, Rao DS, Fyhrie DP, Palnitkar S, Parfitt AM. 2005. The morphological association between microcracks and osteocyte lacunae in human cortical bone. Bone 37:10–15. Qiu S, Rao DS, Palnitkar S, Parfitt AM. 2002b. Age and distance from the surface but not menopause reduce osteocyte density in human cancellous bone. Bone 31:313–318. Qiu S, Rao DS, Palnitkar S, Parfitt AM. 2003b. Reduced iliac cancellous osteocyte density in patients with osteoporotic vertebral fracture. J Bone Miner Res 18:1657–1663. Qiu S, Rao DS, Palnitkar S, Parfitt AM. 2006. Differences in osteocyte and lacunar density between Black and White American women. Bone 38:130–135. Reid SA, Boyde A. 1987. Changes in the mineral density distribution in human bone with age: image analysis using backscattered electrons in the SEM. J Bone Miner Res [Internet] 2:13– 22. Available from: http://www.ncbi.nlm.nih.gov/pubmed/3455153 van Rensburg HCJ, Benatar SR. 1993. The legacy of apartheid in health and health care. South African J Sociol 24:99–111. Robling AG, Stout SD. 2003. Histomorphology, geometry, and mechanical loading in past populations. In: Agarwal S, Stout SD, editors. Bone Loss and Osteoporosis: An Anthropological Perspective. 1st editio. New York: Klewer Academic/Plenum Publishers. p 207–228. Roschger P, Paschalis EP, Fratzl P, Klaushofer K. 2008. Bone mineralization density distribution

123

in health and disease. Bone 42:456–66. Schneider M, Norman R, Parry C, Bradshaw D, Pluddemann A, South African Comparative Risk Assessment Collaborating G. 2007. Estimating the burden of disease attributable to alcohol use in South Africa in 2000. South African Med J [Internet] 97:664–672. Available from: http://ovidsp.ovid.com/ovidweb.cgi?T=JS&CSC=Y&NEWS=N&PAGE=fulltext&D=med5 &AN=17952223%5Cnhttp://nhs4315967.on.worldcat.org/atoztitles/link?sid=OVID:medlin e&id=pmid:17952223&id=doi:&issn=0256- 9574&isbn=&volume=97&issue=8&spage=664&pages=664-72&date=2007&ti Schnitzler CM, Macphail AP, Shires R, Schnaid E, Mesquita JM, Robson HJ. 1994. Osteoporosis in african hemosiderosis: Role of alcohol and iron. J Bone Miner Res 9:1865– 1873. Schnitzler CM, Schnaid E, MacPhail AP, Mesquita JM, Robson HJ. 2005. Ascorbic acid deficiency, iron overload and alcohol abuse underlie the severe osteoporosis in black african patients with hip fractures - A bone histomorphometric study. Calcif Tissue Int 76:79–89. Sedlin ED, Frost HM, Villanueva BS. 1963. Variations in cross-section area of rib cortex with age. J Gerontol 18:9–13. Seftel HC, Malkin C, Schmaman A, Abrahams C, Lynch SR, Charlton RW, Bothwell TH. 1966. Osteoporosis, scurvy, and siderosis in Johannesburg Bantu. Br Med J 1:642–646. Skedros JG, Weaver DJ, Doutre MS. 2016. Osteocyte size, shape, orientation, and population density. Osteologie 2. Solomon L. 1968. Osteoporosis and fracture of the femoral neck in the South African Bantu. J Bone Jt Surg 50B:2–13. Solomon L. 1979. Bone Density in Ageing Caucasian and African Populations. Lancet 314:1326–1330. Stein MS, Feik SA, Clement JG, Wark JD. 1999. An automated analysis of intracortical porosity in human femoral bone across age. J Bone Miner Res 14:624–632. Thomas CDL, Feik SA, Clement JG. 2006. Increase in pore area, and not pore density, is the main determinant in the development of porosity in human cortical bone. J Anat 209:219– 230. Tomkinson A, Reeve J, Shaw RW, Noble BS, Medical G. 1997. The Death of Osteocytes via Apoptosis Accompanies Estrogen Withdrawal in Human Bone. J Clin Endocrinol Metab 82:3128–3135. Tommerup LJ, Raab DM, Crenshaw TD, Smith EL. 1993. Does weight-bearing exercise affect non-weight-bearing bone? J Bone Miner Res 8:1053–1058. Turner-Walker G, Syversen U. 2002. Quantifying histological changes in archaeological bones using BSE-SEM image analysis. Archaeometry 44:461–468. Ural A, Vashishth D. 2007. Effects of intracortical porosity on fracture toughness in aging human bone: A µCT-based cohesive finite element study. J Biomech Eng 129:625–631. Vajda E, Skedros JG, Bloebaum RD. 1995. Consistency in calibrated backscattered electron images of calcified tissues and mineral analyzed in multiple imaging sessions. Scanning Microsc 9:741–755. Vashishth D, Gibson G, Kimura J, Schaffler MB, Fyhrie DP. 2002. Determination of bone volume by osteocyte population. Anat Rec 267:292–295. Vashishth D, Gibson GJ, Fyhrie DP. 2005. Sexual dimorphism and age dependence of osteocyte lacunar density for human vertebral cancellous bone. Anat Rec - Part A Discov Mol Cell Evol Biol 282:157–162. Vatsa A, Breuls RG, Semeins CM, Salmon PL, Smit TH, Klein-Nulend J. 2008. Osteocyte

124

morphology in fibula and calvaria - Is there a role for mechanosensing? Bone 43:452–458. Vilayphiou N, Boutroy S, Sornay-Rendu E, Van Rietbergen B, Chapurlat R. 2016. Age-related changes in bone strength from HR-pQCT derived microarchitectural parameters with an emphasis on the role of cortical porosity. Bone [Internet] 83:233–240. Available from: http://dx.doi.org/10.1016/j.bone.2015.10.012 Walker ARP, Walker BF, Richardson BD. 1971. Metacarpal bone dimensions in young and aged South African Bantu consuming a diet low in calcium. Postgrad Med J 47:320–325. Wapnick AA, Lynch SR, Seftel HC, Charlton RW, Jowsey J. 1971. The effect of siderosis and ascorbic acid depletion on bone metabolism, with special reference to osteoporosis in the Bantu. Br J Nutr 25:367–376. Williams DR, Gonzalez HM, Williams S, Mohammed S a, Moomal H, Stein DJ. 2008. Perceived discrimination, race and health in South Africa. Soc Sci Med 67:441–452. Wilsnack RW, Sharon C. Wilsnack, Kristjanson AF, Vogeltanz-Holm ND, Gmel G. 2009. Gender and alcohol consumption: patterns from the multinational GENACIS project. Addiction 104:1487–1500. Yeni YN, Brown CU, Wang Z, Norman TL. 1997. The influence of bone morphology on fracture toughness of the human femur and tibia. Bone 21:453–459. Yin M, Dobkin J, Street W, York N. 2005. Bone mass and mineral metabolism in HIV+ postmenopausal women. Osteoporos Int 16:1345–1352. Zarrinkalam MR, Mulaibrahimovic A, Atkins GJ, Moore RJ. 2012. Changes in osteocyte density correspond with changes in osteoblast and osteoclast activity in an osteoporotic sheep model. Osteoporos Int 23:1329–1336. Zebaze R, Bohte A, Mackie E, Seeman E. 2009. Age-related bone loss: The effect of neglecting intracortical porosity. Bone [Internet] 44:S117–S118. Available from: http://dx.doi.org/10.1016/j.bone.2009.01.261 Zebaze RM, Ghasem-Zadeh A, Bohte A, Iuliano-Burns S, Mirams M, Price RI, Mackie EJ, Seeman E. 2010. Intracortical remodelling and porosity in the distal radius and post-mortem femurs of women: a cross-sectional study. Lancet [Internet] 375:1729–1736. Available from: http://dx.doi.org/10.1016/S0140-6736(10)60320-0 Zeininger A, Richmond BG, Hartman G. 2011. Metacarpal head biomechanics: A comparative backscattered electron image analysis of trabecular bone mineral density in Pan troglodytes, Pongo pygmaeus, and Homo sapiens. J Hum Evol [Internet] 60:703–710. Available from: http://dx.doi.org/10.1016/j.jhevol.2011.01.002 Zioupos P. 2001. Accumulation of in-vivo fatigue microdamage and its relation to biomechanical properties in ageing human cortical bone. J Microsc 201:270–278.

125

Chapter 6 Conclusion 6.1 Revisiting Research Hypothesis

This chapter evaluates the results of this work against the research hypotheses presented in Chapter 1. Supported hypotheses are in bold. 1. Body Size

H0: Measures of human body size and bone mass will not be correlated with cortical bone microstructure.

HA: Measures of human body size and bone mass will be correlated with cortical bone microstructure.

2. Peak Bone Mass and Density

H0: The structural and material properties of mid-thoracic ribs will not differ between males and females in early adulthood.

HA2: The expected biological relationships defy expectation, suggesting that biosocial factors related to the South African apartheid political system likely influenced the variables of interest.

3. Histological Signatures of Menopause

H0: The structural and material properties of mid-thoracic ribs will not differ between males and females following the predicted age of women’s menopause (50+ years).

HA1: The structural and material properties of mid-thoracic ribs will differ between males and females following the predicted age of women’s menopause (50+ years). Post-

126

menopausal women will exhibit less bone, proportionately more intracortical remodeling, and lower mineralization densities than their male contemporaries.

HA2: The expected biological relationships will either be absent or defy expectation, suggesting that biosocial factors related to the South African apartheid political system likely influenced the variables of interest.

I conclude by discussing some of the limitations of this study, as well as the implication of these results on future bone health research in South Africa.

6.2 Body Size

The first objective of this research was to examine the influence of adult body size and cross- sectional bone mass on histomorphometric variation in the human rib cortex. Available body size proxies (maximum femur length and femur head diameter) do not correlate with either bone mass measurements or the histological properties of secondary osteons, suggesting that size- standardization is unnecessary in studies of rib bone microstructure. This is of relevance to anthropologists because well-recognized variables reflecting body size are not available for many existing histological collections, nor can we anticipate their accessibility in bioarcheological and forensic research contexts.

This study did not identify any functional relationships between body size and cortical bone histomorphometry, suggesting that osteon geometry may be of limited utility in future allometric research (see Felder et al., 2017 for an alternative perspective). Biomechanically induced variation appears to be more heavily influenced by the local strain environment, as larger bones allow for larger osteons. This research also reiterates the importance of using the rib in studies of bone metabolism and aging. Age is the most significant factor affecting osteon population density (OPD), while OPD is the best predictor of osteon area (On.Ar). These findings indicate that age-related secondary osteon crowding has a significant influence on osteon geometry.

6.3 Peak Bone Mass and Density

This research utilized linearly polarized light microscopy (LPLM) and backscattered scanning electron microscopy (BSE-SEM) to examine bone mass and tissue mineralization in the Kirsten Skeletal Collection. Because this study utilized a non-weight bearing bone, results are believed

127 to be less influenced by body size and activity level variation, and more representative of metabolic bone turnover (Tommerup et al., 1993; Robling and Stout, 2003; Pfeiffer et al., 2006; Eleazer and Jankauskas, 2016).

Systemic racism imposed by the apartheid political system (1948-1994) may have rendered non- white South Africans particularly vulnerable to early life stresses, resulting in delayed peak bone mass attainment in the research sample. Relative cortical area (Rt.Ct.Ar) values are highest in men during young adulthood, but do not exceed those of age-matched women until after the fifth decade of life. This is highly unusual. Other studies report average bone mass of men to be in excess to that of women at every decade of adult life (Sedlin et al., 1963; Takahashi and Frost, 1966; Dupras and Pfeiffer, 1996; Streeter and Stout, 2003). There also appears to be a significant age lag in the achievement of peak bone density, with men demonstrating lower average tissue mineralization across the entire adult lifespan in both the cortical and trabecular bone compartments. The average degree of mineralization increases until approximately 40 years in women, and about 50 years in men. This is about a decade later than expected. In contemporary western populations, maximum BMD values derived using DXA are typically reached by the third or fourth decade of life at most skeletal locations (Bonjour et al., 1991; Katzman et al., 1991; Theintz et al., 1992; Matkovic et al., 1994; Fournier et al., 1997; Hopper et al., 1998; Baxter-Jones et al., 2011; Jackowski et al., 2011).

While it is possible that men and women experienced similar levels of systematic oppression under apartheid rule, the female skeleton may be less sensitive to environmental perturbation (Stinson, 1985; Ross et al., 2003; Heinrich, 2015). Secular changes to long bone length are less pronounced in women, suggesting greater buffering to extrinsic factors (Jantz and Jantz, 1999). Women also have the capacity to recover bone losses associated with pregnancy and lactation in the post-weaning period (Bayray and Enquselassie, 2013), indicating high metabolic plasticity. Cumulatively, these data suggest that males may be more susceptible to allostatic load, malnutrition, and/or substance abuse problems during growth and development. In agreement, Cole et al. (2015) found that SAB boys of low socioeconomic status have delayed maturation rates compared to boys from more privileged groups, but did not identify any income-based patterns among girls.

128

Biosocial factors may also explain the observed patterns in bone mass and tissue mineralization. There is evidence to suggest that males are more likely to engage in risky binge drinking and drug use, starting in adolescence and continuing into adulthood (Flisher et al., 1996; Peden et al., 2000; Parry et al., 2005; Wechsberg et al., 2008; Peltzer et al., 2011; Harker Burnhams et al., 2014). These behaviors may have a negative impact on skeletal maturation. South African boys are more likely to be underweight than South African girls (Kimani-Murage et al., 2010), indicating either differential care practices or greater biological frailty (Madise et al., 1999). Low sexual dimorphism in the study sample may also suggest differential growth stunting in boys and girls (Stulp and Barrett, 2016). Population-based studies that fail to account for possible underlying differences in growth and development are highly susceptible to drawing incorrect conclusions about the biological processes directing bone turnover. These behaviors likely have lasting impacts on bone health, and may lead to differential osteoporosis risk in men and women later in life. It has been demonstrated that sex-based variation in bone quality is more likely the result of age-related bone gain than bone loss (Duan et al., 2001).

6.3.1 South African Apartheid and Bone Health

There is a historical precedent for poor bone quality in non-white males in South Africa, as they constituted a high proportion of clinical osteoporosis cases under the apartheid regime (Grusin and Samuel, 1957; Seftel et al., 1966; Lynch et al., 1967, 1970; Wapnick et al., 1971). These earlier studies were mainly limited to SAB men from Johannesburg, located in the Gauteng Province, a region of South Africa where population demographics markedly differ from the present research. The identification of poor skeletal health in male contemporaries from three different population groups in the Western Cape implicate biosocial rather than genetic causative factors.

Considering that the bulk of cadavers are unclmained bodies from local hospitals and government morturaries (Labuschagne and Mathey, 2000; Alblas et al., 2018), SAWs in the Kirsten Skeleton Collection likely represent people from a lower socioeconmic eschelon. It is important to note that the inclusion of indigent SAW males in the research sample does not diminish the differential risks posed to non-whites under apartheid rule. While conditions of abject poverty are used to explain the observed microstructural patterns, non-whites were much more likely to be impoverished during this time. Our results suggest that racial differences in

129 skeletal health outcomes are not inherent, but are strongly dictated by social inequalities that cannot adequately be controlled for in existing epidemiological models (Roberts, 2012; Phelan and Link, 2015).

6.4 Histological Signatures of Menopause

By combining LPLM and BSE-SEM, multiple aspects of cortical and trabecular bone remodeling could be evaluated in order to better understand adult bone health and the menopause transition in marginalized South Africans. While it is possible that non-white men and women experienced similar levels of systematic oppression under apartheid rule, the female skeleton may be less sensitive to environmental perturbation (Stinson, 1985; Ross et al., 2003; Heinrich, 2015). Secular changes to long bone length are less pronounced in women, suggesting greater buffering to extrinsic factors (Jantz and Jantz, 1999). Women also have the capacity to recover bone losses associated with pregnancy and lactation in the post-weaning period (Kovacs, 2005), indicating high metabolic plasticity.

Contrary to expectations, trabecular bone parameters and osteocyte lacunar properties do not appear to be good indicators of menopause in the older women. Cortical bone parameters may prove to be the most useful for detecting menopause in other past populations. Rt.Ct.Ar, osteon area (On.Ar), cortical bone mineralization (Ct.WMGL), and intracortical porosity variables appear to be the most diagnostic, as they closely follow life history expectations for this population (Walker et al., 1984; Jones et al., 2009). Endosteal expansion left uncompensated by new periosteal apposition will decrease the second moment of inertia, thereby increasing biomechanical stress and strain. Osteon size demonstrates an inverse relationship with strain (van Oers et al., 2008), such that smaller osteons are expected in women who experience substantial endocortical bone loss in the decades following menopause. Accordingly, the women in this sample demonstrate a significant decline in On.Ar with age. While pore area (Po.Ar) does not vary with age, pore density (Po.Dn) is highest in the peri-menopause, when accelerated rates of bone turnover are first anticipated. Cortical porosity (Ct.Po) is highest in the years following the predicted age of menopause, but eventually levels off in the final decades of life.

130

6.5 Final Remarks

While this is fundamentally a cross-sectional study, results were interpreted longitudinally, following widespread practice. It is important to note that independent variables were assumed to be temporally consistent throughout the period in which the individuals represented by the Kirsten Skeletal Collection lived. This study is also subject to other limitations of the osteological paradox (Wood et al., 1992). Unknown selection bias tempers the strength of research conclusions. Compared to longitudinal results, cross-sectional data can overestimates age- and sex-related patterns (Burt et al., 2017). Observation and data collection was performed in two-dimensions, which may exaggerate actual changes to bone tissues (Hennig et al., 2015). As imaging technologies continue to advance, nondestructive three-dimensional analyses of microstructural variation will become available for both skeletal and living populations (Genant et al., 2008; Cooper et al., 2011; Ito, 2011).

There is a paucity of bone health information available for South Africans, both past and present. This research provides novel data on an understudied population, evaluating cortical and trabecular bone histomorphology in predominately non-white individuals. Patterns of microstructural variation may be attributed to political and socioeconomic factors that influence disease exposure, access to quality healthcare, and the ability to obtain nutritious food. The dissertation was designed to capture the effect of adverse apartheid living conditions on well- established variables of bone mass and quality. It demonstrated the importance of incorporating environmental stress throughout ontogeny and major life history events into existing models for interpreting bone microstructure. Further, the project highlights the utility of curated skeletal collections for the study of contemporary biomedical issues. By recognizing vulnerabilities in men and non-white population groups, current public health endeavors can more effectively address osteoporosis and fracture risk in the post-apartheid era.

6.6 Literature Cited

Alblas A, Greyling LM, Geldenhuys EM. 2018. Composition of the Kirsten Skeletal Collection at Stellenbosch University. S Afr J Sci 114:1–6. Baxter-Jones ADG, Faulkner R a, Forwood MR, Mirwald RL, Bailey D a. 2011. Bone mineral accrual from 8 to 30 years of age: an estimation of peak bone mass. J bone Miner Res Off J Am Soc Bone Miner Res 26:1729–39. Bayray A, Enquselassie F. 2013. The effect of parity on bone mineral density in postmenopausal women: A systematic review. J Osteoporos Phys Act [Internet] 1:104. Available from:

131

http://www.esciencecentral.org/journals/the-effect-of-parity-on-bone-mineral-density-in- postmenopausal-women-a-systematic-review-2329-9509.1000104.php?aid=19379 Bonjour J, Theintz G, Buchs B, Slosman D, Rizzoli R. 1991. Critical years and stages of puberty for spinal and femoral bone mass accumulation during adolescence. J Clin Endocrinol Metab 73:553–563. Burt LA, Hanley DA, Boyd SK. 2017. Cross-sectional Versus Longitudinal Change in a Prospective HR-pQCT Study. J Bone Miner Res 32:1505–1513. Cole TJ, Rousham EK, Hawley NL, Cameron N, Norris SA, Pettifor JM. 2015. Ethnic and sex differences in skeletal maturation among the Birth to Twenty cohort in South Africa. Arch Dis Child [Internet] 100:138–143. Available from: http://adc.bmj.com/lookup/doi/10.1136/archdischild-2014-306399 Cooper DML, Erickson B, Peele a. G, Hannah K, Thomas CDL, Clement JG. 2011. Visualization of 3D osteon morphology by synchrotron radiation micro-CT. J Anat 219:481–489. Duan Y, Turner CH, Kim BT, Seeman E. 2001. Sexual dimorphism in vertebral fragility is more the result of gender differences in age-related bone gain than bone loss. J Bone Miner Res 16:2267–2275. Dupras TL, Pfeiffer SK. 1996. Determination of sex from adult human ribs. Can Soc Forensic Sci J 29:221–231. Eleazer CD, Jankauskas R. 2016. Mechanical and metabolic interactions in cortical bone development. Am J Phys Anthropol 160:317–333. Felder AA, Phillips C, Cornish H, Cooke M, Hutchinson JR, Doube M. 2017. Secondary Osteons Scale Allometrically In Mammalian Humerus And Femur. R Soc open sci [Internet] 4:170431. Available from: https://www.biorxiv.org/content/early/2017/04/30/131300 Flisher AJ, Ziervogel CF, Chalton DO, Leger PH, Robertson BA. 1996. Risk-taking behaviour of Cape Peninsula high-school students. South African Med J 86:1090–1093. Fournier PE, Rizzoli R, Slosman DO, Theintz G, Bonjour JP. 1997. Asynchrony between the rates of standing height gain and bone mass accumulation during puberty. Osteoporos Int 7:525–532. Genant HK, Engelke K, Prevrhal S. 2008. Advanced CT bone imaging in osteoporosis. Rheumatology 47. Grusin H, Samuel MD. 1957. A syndrome of osteoporosis in Africans and its relationship to scurvy. Am J Clin Nutr 5:644–650. Harker Burnhams N, Parry C, Laubscher R, London L. 2014. Prevalence and predictors of problematic alcohol use, risky sexual practices and other negative consequences associated with alcohol use among safety and security employees in the Western Cape, South Africa. Subst Abuse Treat Prev Policy [Internet] 9:14. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3944609&tool=pmcentrez&ren dertype=abstract Hennig C, Thomas CDL, Clement JG, Cooper DML. 2015. Does 3D orientation account for variation in osteon morphology assessed by 2D histology? J Anat 227:497–505. Hopper JL, Green RM, Nowson C a, Young D, Sherwin a J, Kaymakci B, Larkins RG, Wark JD. 1998. Genetic, common environment, and individual specific components of variance for bone mineral density in 10- to 26-year-old females: a twin study. Am J Epidemiol 147:17–29. Ito M. 2011. Recent progress in bone imaging for osteoporosis research. J Bone Miner Metab

132

29:131–140. Jackowski SA, Kontulainen SA, Cooper DML, Lanovaz JL, Baxter-Jones ADG. 2011. The timing of BMD and geometric adaptation at the proximal femur from childhood to early adulthood in males and females: a longitudinal study. J bone Miner Res Off J Am Soc Bone Miner Res 26:2753–61. Jantz LM, Jantz RL. 1999. Secular change in long bone length and proportion in the United States, 1800-1970. Am J Phys Anthropol 110:57–67. Jones LL, Griffiths PL, Norris S a., Pettifor JM, Cameron N. 2009. Age at menarche and the evidence for a positive secular trend in urban South Africa. Am J Hum Biol 21:130–132. Katzman DK, Bachrach LK, Carter DR, Marcus R. 1991. Clinical and Anthropometric Correlates of Bone Mineral Acquisition in Healthy Adolescent Girls. J Clin Endocrinol Metab 73:1332–1339. Kimani-Murage EW, Kahn K, Pettifor JM. 2010. The prevalence of stunting, overweight and obesity, and metabolic disease risk in rural South African children. BMC public [Internet]:1–13. Available from: http://www.biomedcentral.com/1471-2458/10/158 Kovacs CS. 2005. Calcium and Bone Metabolism During Pregnancy and Lactation. J Mammary Gland Biol Neoplasia 10:105–118. Labuschagne BCJ, Mathey B. 2000. Cadaver profile at University of Stellenbosch Medical School, South Africa, 1956 – 1996. Clin Anat 13:88–93. Lynch SR, Berelowitz I, Seftel HC, Miller GB, Krawitz P, Bothwell TH. 1967. Osteoporosis in Johannesburg Bantu Males: Its Relationship to Siderosis and Ascorbic Acid Deficiency. Am J Clin Nutr 20:799–807. Lynch SR, Seftel HC, Wapnick AA, Charlton RW, Bothwell TH. 1970. Some aspects of calcium metabolism in normal and osteoporotic Bantu subjects with special reference to the effects of iron overload and ascorbic acid depletion. S Afr J Med Sci 35:45–56. Madise NJ, Matthews Z, Margetts B. 1999. Heterogeneity of child nutritional status between households: A comparison of six sub-Saharan African countries. Popul Stud (NY) 53:331– 343. Matkovic V, Jelic T, Wardlaw GM, Llich JZ, Goel PK, Wright JK, Andon MB, Smith KT, Heaney RP. 1994. Timing of peak bone mass in Caucasian females and its implication for the prevention of osteoporosis: Inference from a cross-sectional model. J Clin Invest 93:799–808. van Oers RFM, Ruimerman R, van Rietbergen B, Hilbers PAJ, Huiskes R. 2008. Relating osteon diameter to strain. Bone 43:476–82. Parry CD, Pluddemann A, Steyn K, Bradshaw D, Norman R, Laubscher R. 2005. Alcohol Use in South Africa: Findings from the First Demographic and Health Survey (1998)*. J Stud Alcohol 66:91–97. Peden M, van der Spuy J, Smith P, Bautz P. 2000. Substance abuse and trauma in Cape Town. South African Med J [Internet] 90:251–255. Available from: http://www.tandfonline.com/doi/abs/10.1080/0953732032000199061 Peltzer K, Davids A, Njuho P. 2011. Alcohol use and problem drinking in South Africa: findings from a national population-based survey. Afr J Psychiatry [Internet] 14:30–37. Available from: http://dx.doi.org/10.4314/ajpsy.v14i1.65466 Pfeiffer S, Crowder C, Harrington L, Brown M. 2006. Secondary osteon and Haversian canal dimensions as behavioral indicators. Am J Phys Anthropol 131:460–468. Phelan JC, Link BG. 2015. Is Racism a Fundamental Cause of Inequalities in Health? Annu Rev Sociol [Internet] 41:311–330. Available from:

133

http://www.annualreviews.org/doi/10.1146/annurev-soc-073014-112305 Roberts D. 2012. Debating the Cause of Health Disparities. Cambridge Q Healthc Ethics [Internet] 21:332–341. Available from: http://www.journals.cambridge.org/abstract_S0963180112000059 Robling AG, Stout SD. 2003. Histomorphology, geometry, and mechanical loading in past populations. In: Agarwal S, Stout SD, editors. Bone Loss and Osteoporosis: An Anthropological Perspective. 1st editio. New York: Klewer Academic/Plenum Publishers. p 207–228. Ross AH, Baker LE, Falsetti A. 2003. Sexual dimorphism a proxy for environmental sensitivity? A multitemporal view. J Washingt Acad Sci 89:1–12. Sedlin ED, Frost HM, Villanueva BS. 1963. Variations in cross-section area of rib cortex with age. J Gerontol 18:9–13. Seftel HC, Malkin C, Schmaman A, Abrahams C, Lynch SR, Charlton RW, Bothwell TH. 1966. Osteoporosis, scurvy, and siderosis in Johannesburg Bantu. Br Med J 1:642–646. Stinson S. 1985. Sex Differences in Environmental Sensitivity During Growth and Development. Yearb Phys Anthropol 28:123–147. Streeter MA, Stout SD. 2003. The Histomorphometry Of The Subadult Rib: Age-Associated Changes In Bone Mass And The Creation Of Peak Bone Mass. In: Agarwal SC, Stout SD, editors. Bone Loss and Osteoporosis: An Anthropological Perspective. 1st Editio. New York: Klewer Academic/Plenum Publishers. p 91–101. Stulp G, Barrett L. 2016. Evolutionary perspectives on human height variation. Biol Rev 91:206–234. Takahashi H, Frost HM. 1966. Age and Sex Related Changes in the Amount of Cortex of Normal Human Ribs. Acta Orthop Scand [Internet] 37:122–130. Available from: http://www.tandfonline.com/doi/full/10.3109/17453676608993272 Theintz G, Buchs B, Rizzoli R, Slosman D, Clavien H, Sizonenko PC, Bonjour JP. 1992. Longitudinal monitoring of bone mass accumulation in healthy adolescents: evidence for a marked reduction after 16 years of age at the levels of lumbar spine and femoral neck in female subjects. J Clin Endocrinol Metab 75:1060–1065. Tommerup LJ, Raab DM, Crenshaw TD, Smith EL. 1993. Does weight-bearing exercise affect non-weight-bearing bone? J Bone Miner Res 8:1053–1058. Walker ARP, Walker BF, Ncongwane J, Human ENT. 1984. Age of menopause in black women in South Africa. Br J Obstet Gynaecol 91:797–801. Wapnick AA, Lynch SR, Seftel HC, Charlton RW, Jowsey J. 1971. The effect of siderosis and ascorbic acid depletion on bone metabolism, with special reference to osteoporosis in the Bantu. Br J Nutr 25:367–376. Wechsberg WM, Luseno WK, Karg RS, Young S, Rodman N, Myers B, Parry CD. 2008. Alcohol, cannabis, and methamphetamine use and other risk behaviours among Black and Coloured South African women: A small randomized trial in the Western Cape. Int J Drug Policy 19:130–139. Wood JW, Milner GR, Harpending HC, Weiss KM, Cohen MN, Eisenberg LE, Hutchinson DL, Jankauskas R, Česnys G, Katzenberg MA, Lukacs JR, Mcgrath JW, Roth EA, Ubelaker DH, Wilkinson RG. 1992. The Osteological Paradox: Problems of Inferring Prehistoric Health from Skeletal Samples. Curr Anthropol 33:343–370.

134

Appendix A: List of Abbreviations by Order of Appearance

LPLM: linearly polarized light microscopy

BSE-SEM: backscattered scanning electron microscopy

SAB: South African Black

SAC: South African Coloured

SAI: South African Indian/Asian

SAW: South African White

DXA: dual energy x-ray absorptiometry

BMD: bone mineral density

MFL: maximum femur length

FHD: femur maximum diameter

Tt.Ar: total cross-sectional area

Es.Ar: endosteal area

Ct.Ar: cortical area

Rt.Ct.Ar: relative cortical area

N.On: intact osteon density

N.On.Fg: fragmentary osteon density

OPD: osteon population density

On.Ar: osteon area

HMR: hierarchical multiple regression

BV/TV: trabecular bone volume fraction

Tb.N: trabecular number

Tb.Th: trabecular thickness

Tb.Sp: Trabecular spacing

135

Ct.B.Ar: cortical bone area

Po.Ar: mean pore area

Po.Dm: mean pore diameter

Po.N: pore number

Po.Dn: pore density

Ct.Po: cortical porosity

Ct.Ot.Lc.N: cortical osteocyte lacunar number

Ct.Ot.Lc.Ar: cortical osteocyte lacunar area

Ct.Ot.Lc.Dn: cortical osteocyte lacunar density

Ct.WMGL: cortical weighted mean grey level

Tb.B.Ar: trabecular bone area

Tb.Ot.Lc.N: trabecular osteocyte lacunar number

Tb.Ot.Lc.Ar: trabecular osteocyte lacunar area

Tb.Ot.Lc.Dn: trabecular osteocyte lacunar density

Tb.WMGL: trabecular weighted mean grey level

136 Appendix B: Full Research Sample

KSC U of T Group Accession Accession Rib # YOB YOD/DOD COD Sex Age ID # # An 202 S-67-202 R Mid-T 1917 May 3, 1967 Unknown M SAC 50 An 203 S-68-203 L Mid-T 1932 October 9, 1968 Tuberculosis Pulmonary M SAB 36 An 219 S-69-219 R Mid-T 1930 January 29, 1969 Tuberculosis M SAC 39 An 221 S-69-221 R Mid-T 1924 March 2, 1969 Tuberculosis Pulmonary F SAC 45 An 222 S-69-222 R6 1911 February 9, 1969 Aneurysm, Aortic haemorrhage M SAB 58 An 225 S-69-225 L5 1901 January 12, 1969 Carcinoma of Stomach M SAC 68 An 226 S-70-226 R6 1948 August 9, 1970 Hodgekin's Disease F SAB 22 An 231 S-70-231 R Mid-T 1940 September 28, 1970 Tuberculosis M SAC 30 An 232 S-70-232 R Mid-T 1900 September 28, 1970 Tuberculosis M SAC 70 An 235 S-70-235 L Mid-T 1930 September 6, 1970 Pneumonia (lobar) M SAC 40 An 236 S-70-236 R Mid-T 1910 September 5, 1970 Carcinoma of R Lung M SAB 60 An 239 S-70-239 L Mid-T 1910 June 30, 1970 Pneumonia M SAB 60 An 242 S-70-242 R6 1908 September 22, 1970 Carcinoma of Oesophagus M SAC 62 An 265 S-70-265 R Mid-T 1931 July 17, 1970 Cardiac Failure, Tuberculosis, Pulmonary F SAC 39 An 508 S-70-508 R6 1942 July 31, 1970 Renal and Cardiac Failure M SAC 28 An 247 S-71-247 L6 1928 January 9, 1971 Carcinoma of Breast with Metastases F SAW 43 An 261 S-71-261 R7 1895 March 7, 1971 Coronary Thrombosis M SAC 76 An 275 S-72-275 L7 1912 June 3, 1972 Intracerebral Bleeding F SAC 60 An 279 S-72-279 R Mid-T 1934 August 6, 1972 Cardiac Failure, Emphyzema M SAB 38 An 282 S-72-282 R Mid-T 1950 July 11, 1972 Carcinoma of Cervix F SAC 22 An 286 S-72-286 L Mid-T 1943 July 26, 1972 Tuberculosis Pulmonary F SAC 29 An 291 S-72-291 R6 1894 May 30, 1972 Emphysemia M SAC 78 An 295 S-72-295 R6 1900 August 18, 1972 Myocardial Infarction F SAC 72 An 298 S-72-298 L Mid-T 1938 August 9, 1972 Carcinoma of Cervix F SAC 34 An 305 S-73-305 L5 1899 December 9, 1973 Gangreen both feet, Gen. Toxaemia M SAC 74 An 306 S-73-306 R Mid-T 1928 September 5, 1973 Tuberculosis Pulmonary F SAC 45 An 308 S-73-308 R7 1934 December 27, 1973 Unknown F SAC 39 An 309 S-73-309 L7 1935 October 10, 1973 Carcinoma of Breast with Metastases F SAC 38 An 310 S-73-310 R7 1907 August 12, 1973 COPD (Chronic Obstructive Airways Disease), Pneumonia M SAC 66 An 312 S-73-312 R Mid-T 1935 August 24, 1973 Tuberculosis Pulmonary M SAC 38

137 An 315 S-73-315 R Mid-T 1928 May 16, 1973 Carcinoma of Cervix F SAB 45 An 316 S-73-316 R Mid-T 1905 July 21, 1976 Cardiac Failure M SAB 68 An 320 S-73-320 R Mid-T 1926 August 23, 1973 Hypertensive Cardiac Failure M SAB 47 An 322 S-73-322 R7 1908 February 8, 1973 Coronary Thrombosis F SAW 65 An 318 S-74-318 R6 1912 August 10, 1974 Cardiac Failure M SAW 62 An 339 S-74-339 R Mid-T 1946 May 12, 1974 Hypertensive Cardiac Failure F SAC 28 An 340 S-74-340 R6 1901 April 2, 1974 Cardiopulmonary Failure M SAC 73 An 343 S-74-343 R Mid-T 1914 July 15, 1974 Natural Causes M SAB 60 An 377 S-74-377 R Mid-T 1929 April 13, 1974 Cardiac Insufficiency F SAC 45 An 383 S-74-383 L Mid-T 1924 February 14, 1974 Hepatic Failure, Severe Malnutrition F SAC 53 An 394 S-74-394 R Mid-T 1925 August 5, 1974 Unknown M SAC 49 An 396 S-74-396 R Mid-T 1939 September 18, 1974 Chronic Myeloid Leukaemia M SAC 35 An 363 S-75-363 R Mid-T 1896 October 5, 1975 Carcinoma of Bronchus M SAB 79 An 366 S-75-366 R6 1925 December 3, 1975 Cardiomiopathy M SAB 50 An 367 S-75-367 R Mid-T 1914 November 6, 1975 Carcinoma of Tonsillar Palate Fossa M SAC 61 An 374 S-75-374 R Mid-T 1958 June 15, 1975 Status Epilepticus M SAC 17 An 378 S-75-378 R Mid-T 1925 May 22, 1975 Carcinoma of Stomach M SAC 50 An 387 S-75-387 R Mid-T 1917 May 16, 1975 Carcinoma of Lung M SAW 58 An 388 S-75-388 R Mid-T 1944 February 22, 1975 Tuberculosis Pulmonary M SAB 31 An 389 S-75-389 R6 1904 May 15, 1975 Respiratory Failure M SAC 71 An 390 S-75-390 R Mid-T 1916 June 5, 1975 Carcinoma of Oropharynx M SAC 59 An 392 S-75-392 R6 1896 September 3, 1975 Tuberculosis, Bronchopneumonia M SAB 79 An 395 S-75-395 R Mid-T 1955 May 4, 1975 Tuberculosis Pulmonary M SAC 20 An 398 S-75-398 R Mid-T 1949 December 19, 1975 Carcinoma of Tongue M SAC 26 An 400 S-75-400 R Mid-T 1923 April 12, 1975 Cardiac Failure, Mediastinal Tumour M SAB 52 An 412 S-76-412 R Mid-T 1958 December 1, 1976 Cardiac and Renal failure M SAB 18 An 429 S-76-429 R6 1943 December 30, 1976 Tuberculous Pulmonary, Fibrosis and Bulla Form M SAB 33 An 439 S-76-439 R Mid-T 1945 December 13, 1976 Respiratory Failure M SAB 31 An 448 S-76-448 L6 1944 February 21, 1976 Cardiac Failure F SAC 32 An 417 S-77-417 R Mid-T 1922 June 20, 1977 Carcinoma of Cervix, Stage 4 F SAC 55 An 424 S-77-424 R6 1907 March 23, 1977 Brain Tumor M SAC 70 An 427 S-77-427 R Mid-T 1923 February 12, 1977 Alcoholic Liver Cirrhosis, Cancer of L Lung M SAW 54 An 428 S-77-428 R Mid-T 1951 January 24, 1977 Cardiac Failure F SAC 26 An 441 S-77-441 R6 1911 December 13, 1977 Cerebrovascular Accident F SAC 66 An 445 S-77-445 R Mid-T 1945 October 11, 1977 Lymphoma M SAC 32

138

An 450 S-77-450 R Mid-T 1924 January 27, 1977 Amoebic Dysentery F SAB 53 An 451 S-77-451 R6 1917 August 25, 1977 Carcinoma of Oesophagus and Larynx F SAC 60 An 456 S-77-456 R Mid-T 1944 February 23, 1977 Rheumatic Cardiac Disease Chronic M SAB 33 An 459 S-77-459 R6 1926 June 12, 1977 Carcinoma of Oesophagus M SAB 51 An 461 S-77-461 R Mid-T 1926 September 26, 1977 Carcinoma of Bronchus M SAC 51 An 760 S-77-760 L Mid-T 1920 July 28, 1977 Tumour (brain metastases), Anaemia M SAB 57 An 833 S-77-833 R7 1954 December 27, 1977 Cardiopulmonary Arrest F SAC 23 An 435 S-78-435 R Mid-T 1923 February 6, 1978 Subarachnoid Haemorrhage M SAB 55 An 467 S-78-467 R Mid-T 1958 September 6, 1978 Endocarditis M SAB 20 An 470 S-78-470 L Mid-T 1959 November 30, 1978 Tuberculosis Pulmonary, Cardiac Failure M SAC 19 An 474 S-78-474 R Mid-T 1926 July 28, 1978 Cerebrovascular Accident M SAC 52 An 486 S-78-486 L6 1910 November 17, 1978 Carcinoma of Pharynx F SAC 65 An 487 S-78-487 L6 1952 March 2, 1978 Large Cell Lymphoma of Small Bowel M SAB 26 An 490 S-78-490 L6 1960 October 30, 1978 Cardiac Failure F SAC 18 An 494 S-78-492 R6 1920 January 27, 1978 Hepatic Failure M SAB 58 An 493 S-78-493 L6 1927 January 29, 1978 Pulmonary Embolism M SAB 51 An 494 S-78-494 R7 1920 January 27, 1978 Hepatic Failure M SAB 58 An 497 S-78-497 R6 1900 January 3, 1978 Cerebrovascular Accident, Lung Abscess M SAB 78 An 498 S-78-498 R Mid-T 1935 March 10, 1978 Bronchiectasis, Lung abscess M SAB 43 An 499 S-78-499 R Mid-T 1921 November 21, 1978 Carcinoma of Tongue M SAB 57 An 505 S-78-505 R Mid-T 1941 April 17, 1978 Cardiac Valve Lesions, Gastroenteritis M SAB 37 An 513 S-78-513 L6 1906 May 7, 1978 Cerebrovascular Accident, Small intestine obstruction M SAW 75 An 517 S-78-517 R Mid-T 1928 December 29, 1978 Meningitis M SAC 50 An 506 S-79-506 R6 1913 September 1, 1979 Myocardial Infarction F SAW 66 An 519 S-79-519 L7 1924 March 6, 1979 Carcinoma of Lung to Brain M SAW 55 An 520 S-79-520 R6 1935 February 6, 1979 Carcinoma of Cervix and Renal Failure F SAB 44 An 536 S-79-536 L Mid-T 1921 August 24, 1979 Asthma F SAC 58 An 538 S-79-538 R6 1933 June 1, 1979 Bronchopneumonia M SAW 46 An 548 S-79-548 L7 1901 January 11, 1979 Carcinoma of Stomach M SAC 78 An 554 S-79-554 L7 1913 August 16, 1979 COPD (Chronic Obstructive Airways Disease) M SAW 66 An 572 S-79-572 L7 1915 November 21, 1979 Pneumonia, Carcinoma of Cervix F SAW 64 An 565 S-80-565 R Mid-T 1956 August 27, 1980 Carcinoma of Ovaries F SAC 24 An 591 S-80-591 R7 1946 October 26, 1980 Renal Failure M SAB 34 An 607 S-80-607 L7 1918 October 4, 1980 Myocardial Infarction F SAW 62 An 588 S-81-588 R5 1956 June 19, 1981 Tuberculosis Pulmonary, Cardiac Arrest F SAC 22

139

An 590 S-81-590 R7 1925 November 29, 1981 Epilepsy and Brain Haemorrhage F SAC 56 An 596 S-81-596 R7 1938 July 21, 1981 Lymphoma (possible) M SAB 43 An 599 S-81-599 L6 1920 September 23, 1981 Carcinoma of Oesophagus M SAB 61 An 605 S-81-605 L7 1953 August 1, 1981 Hypothermia M SAC 28 An 610 S-81-610 R7 1921 August 5, 1981 Tuberculosis Pulmonary M SAB 60 An 612 S-81-612 R Mid-T 1921 April 28, 1981 Cerebrovascular Accident M SAC 60 An 615 S-81-615 L Mid-T 1915 November 9, 1981 Lupus Erythematosus, Carcinoma of Pancreas F SAW 66 An 623 S-81-623 R7 1918 August 7, 1981 Myocardial Infarction M SAC 63 An 624 S-81-624 L6 1903 February 19, 1981 Coronary Thrombosis M SAW 78 An 637 S-81-637 R6 1929 December 17, 1981 Artrial Fibrulation M SAW 52 An 648 S-81-648 R7 UNK UNK Chronic Alcoholism with Delirium Aspiration, Pneumonia M SAW 35 An 682 S-81-682 L7 1905 August 29, 1981 Cerebrovascular Accident M SAC 76 An 632 S-82-632 L7 1928 February 17, 1982 Myocardial Infarction M SAW 54 An 635 S-82-635 R7 1940 October 29, 1982 Myocardial Infarction F SAW 42 An 643 S-82-643 R6 1932 January 16, 1982 Ascites and Hepatomegaly, Carcinoma M SAB 50 An 652 S-82-652 R Mid-T 1924 November 26, 1982 Carcinoma of Mouth Floor M SAC 58 An 654 S-82-654 R7 1916 December 19, 1982 Myocardial Infarction M SAW 66 An 668 S-82-668 R Mid-T 1940 December 22, 1982 Malignant Hypertension and Renal Failure M SAB 42 An 651 S-83-651 R6 1941 July 31, 1983 Septicaemia with Diffusion, Intravascular Stollings Defek F SAW 42 An 656 S-83-656 R6 1922 July 16, 1983 Respiratory Failure and DOPS F SAC 61 An 658 S-83-658 R7 1938 March 23, 1983 Myocardial Infarction M SAW 45 An 660 S-83-660 R7 1937 May 1, 1983 Natural Causes M SAW 46 An 661 S-83-661 L7 1945 July 14, 1983 Carcinoma of Bronchus with Brain Metastases F SAB 38 An 665 S-83-665 R7 1949 December 1, 1983 Tuberculosis Pulmonary F SAC 34 An 667 S-83-667 R7 1945 August 16, 1983 Myocardial Infarction M SAC 38 An 669 S-83-669 L7 1950 April 15, 1983 Carcinoma of Cervix F SAC 33 An 671 S-83-671 R7 1920 October 15, 1983 Carcinoma of Stomach F SAC 63 An 672 S-83-672 R Mid-T 1931 March 5, 1983 CCF (Congestive Cardiac Failure) M SAC 52 An 683 S-83-683 L7 1905 February 16, 1983 Coronary Infarction, Myocardial Ischemia M SAW 78 An 684 S-83-684 R Mid-T 1921 August 4, 1983 Bronchopneumonia F SAC 62 An 702 S-83-702 R7 1945 December 7, 1983 Cardiopulmonary Arrest F SAC 38 An 714 S-83-714 L Mid-T 1918 December 27, 1983 Hypertension, Gastritis, Pharingitis F SAC 65 An 736 S-83-736 R Mid-T 1936 December 20, 1983 Carcinoma of Bronchus M SAC 47 An 690 S-84-690 L7 1934 August 15, 1984 Pneumonia (Lobar) F SAC 50 An 691 S-84-691 R7 1925 September 6, 1984 Myocardial Infarction M SAW 59

140

An 694 S-84-694 R Mid-T 1961 March 12, 1984 Hodgekin's Lymphoma M SAC 23 An 695 S-84-695 R Mid-T 1940 October 18, 1984 Lung Abscess M SAB 44 An 700 S-84-700 L Mid-T 1928 February 15, 1984 Pneumonia, Abdomen Sepsis M SAC 56 An 704 S-84-704 R7 1932 September 2, 1984 Myocardial Infarction M SAW 52 An 729 S-84-729 R7 1924 May 25, 1984 Cardiac Arrest F SAC 60 An 713 S-85-713 L7 1935 February 24, 1985 Cardiac Failure F SAW 50 An 725 S-85-725 R6 1937 March 21, 1985 Congestive Cardiac Failure, Pericardial Effusion F SAC 48 An 732 S-85-732 R Mid-T 1926 January 9, 1985 Myocardial Infarction M SAW 59 An 741 S-85-741 R Mid-T 1929 September 25, 1985 Cerebrovascular Accident M SAC 56 An 745 S-85-745 1965 July 15, 1985 Cardiac Arrest, Rheumatic Myocardial Infarction M SAB 20

An 752 S-85-752 R Mid-T 1942 December 13, 1985 Renal and Biventricular Cardiac Failure F SAC 43 An 754 S-85-754 R7 1946 December 2, 1985 Metastatic Cervix Carcinoma F SAC 39 An 755 S-85-755 R7 1930 September 19, 1985 Metastatic Cervix Carcinoma F SAC 55 An 759 S-85-759 R7 1957 December 12, 1985 Bronchopneumonia F SAC 28 An 761 S-85-761 R Mid-T 1952 November 25, 1985 Carcinoma of Maxillary Antrum with Lymphnode Metastases M SAC 33 An 765 S-85-765 L Mid-T 1944 June 20, 1987 Carcinoma of Left Antrum M SAC 42 An 766 S-85-766 R Mid-T 1932 November 26, 1985 Carcinoma of Mouth Floor and Tongue M SAC 53 An 767 S-85-767 R Mid-T 1920 December 25, 1985 Natural Causes M SAC 65 An 751 S-86-751 L7 1946 July 3, 1986 Subarachnoid Haemorrhage Branched into Ventricles F SAC 40 An 753 S-86-753 R7 1944 February 26, 1986 Carcinoma of Cervix F SAC 42 An 757 S-86-757 R Mid-T 1930 January 12, 1986 Hepatic Failure, Chronic Pancreatitis M SAW 56 An 758 S-86-758 L6 1953 February 22, 1986 Tuberculosis M SAB 33 An 771 S-86-771 R Mid-T 1924 May 14, 1986 Coronary Thrombosis M SAC 62 An 772 S-86-772 R Mid-T 1945 May 22, 1986 Septicaemia M SAC 41 An 774 S-86-774 R Mid-T 1945 June 17, 1986 Tuberculosis Pulmonary F SAC 41 An 846 S-86-846 R Mid-T 1950 July 26, 1986 Tuberculosis Pulmonary, Keto Acidosis F SAB 28 An 777 S-87-777 L Mid-T 1935 February 5, 1987 Tuberculosis Pulmonary M SAC 52 An 781 S-87-781 R Mid-T 1932 March 8, 1987 Carcinoma of Oesophagus M SAC 55 An 783 S-87-783 R Mid-T 1960 March 30, 1987 Tuberculosis Pulmonary F SAC 27 An 784 S-87-784 R Mid-T 1933 April 15, 1987 Carcinoma of Oesophagus M SAB 54 An 788 S-87-788 R Mid-T 1959 May 17, 1987 Natural Causes M SAC 28 An 790 S-87-790 R Mid-T 1952 June 16, 1987 Cardiac Arrest M SAC 35 An 791 S-87-791 R Mid-T 1929 June 23, 1987 Carcinoma of Pancreas M SAC 58 An 792 S-87-792 R Mid-T 1946 June 29, 1987 Bronchiectasis with Respiratory Failure M SAC 41 An 796 S-87-796 R Mid-T 1933 October 29, 1987 Intracerebral Bleeding with Respiratory Arrest F SAC 54

141

An 797 S-87-797 R Mid-T 1959 November 17, 1987 Cardiopulmonary Arrest F SAC 28 An 801 S-87-801 R Mid-T 1939 July 15, 1987 Alcoholic Liver Cirrhosis M SAC 48 An 802 S-87-802 R6 1920 August 4, 1987 COPD (Chronic Obstructive Airways Disease) M SAC 67 An 808 S-87-808 L Mid-T 1940 October 11, 1987 Tuberculosis Pulmonary M SAC 47 An 809 S-87-809 R Mid-T 1933 December 22, 1987 Cardiac Failure and Arrhythmia M SAC 54 An 827 S-87-827 R Mid-T 1946 August 24, 1987 Carcinoma of Liver M SAC 41 An 811 S-88-811 R Mid-T 1950 January 11, 1988 Tuberculosis Disseminated M SAC 38 An 813 S-88-813 R6 1927 January 12, 1988 Cerebrovascular Accident M SAC 61 An 814 S-88-814 R7 1919 January 30, 1988 Small Cell Bronchus Carcinoma M SAW 69 An 817 S-88-817 R7 1919 April 18, 1988 Cardiogenic Shock M SAW 69 An 819 S-88-819 R7 1929 April 6, 1988 Tuberculosis, Pulmonary Embolism M SAC 59 An 820 S-88-820 R Mid-T 1927 April 22, 1988 Carcinoma of Cervix F SAW 61 An 824 S-88-824 R6 1927 March 25, 1988 Carcinoma of Oesophagus F SAW 61 An 843 S-88-843 R Mid-T 1952 December 4, 1988 Carcinoma of Oesophagus M SAB 36 An 844 S-89-844 R Mid-T 1962 January 26, 1989 Upper Cervical Dislocation and Spinal Chord Compression F SAB 27 An 847 S-89-847 R6 1959 February 17, 1989 Renal Failure F SAC 30 An 849 S-89-849 R7 1962 April 27, 1989 Tuberculosis Pulmonary M SAC 27 An 851 S-89-851 R7 1954 June 2, 1989 Tuberculosis Pulmonary M SAC 35 An 853 S-89-853 R Mid-T 1945 August 5, 1989 Tuberculosis M SAC 44 An 854 S-89-854 R Mid-T 1963 August 11, 1989 Tuberculosis Pulmonary F SAC 26 An 855 S-89-855 R Mid-T 1945 August 28, 1989 Diabetes Mellitus M SAC 44 An 857 S-89-857 L Mid-T 1955 September 23, 1989 Epilepsy, Cardiorespiratory Failure, Epilepsy M SAC 34 An 859 S-90-859 R Mid-T 1956 March 16, 1990 Tuberculosis Pulmonary F SAC 34 An 862 S-90-862 R6 1949 May 25, 1990 Cardiopulmonary Arrest, Hepatic Coma M SAB 41 An 863 S-90-863 R Mid-T 1950 June 12, 1990 Pancreatitis Acute M SAC 40 An 864 S-90-864 R7 1952 July 11, 1990 Respiratory Failure F SAC 38 An 865 S-90-865 R Mid-T 1949 August 14, 1990 Hepatic Coma, Hepatic Failure F SAC 41 An 866 S-90-866 R Mid-T 1935 August 19, 1990 Pneumonia M SAC 55 An 867 S-90-867 L7 1970 September 1, 1990 Cardiopulmonary Arrest M SAC 20 An 868 S-90-868 L7 1941 November 8, 1990 Tuberculosis Pulmonary M SAB 49 An 869 S-90-869 R Mid-T 1950 September 23, 1990 Cardiac Pathology Complicated by Infective Endocarditis M SAC 40 An 871 S-91-871 R Mid-T 1966 February 2, 1991 Haemoptysis Massive F SAC 25 An 872 S-91-872 R7 1944 April 1, 1991 Natural Causes F SAC 47 An 873 S-91-873 R Mid-T 1952 May 2, 1991 Pneumonia M SAC 39 An 874 S-91-874 R Mid-T 1951 July 16, 1991 Cerebral Haemorrhage M SAC 40

142

An 875 S-91-875 R Mid-T 1972 October 23, 1991 Hepatoma with Gastrointestinal Bleeding M SAC 19 An 880 S-91-880 R7 1936 November 12, 1991 Carcinoma of Stomach F SAC 55 An 884 S-92-884 R Mid-T 1969 August 11, 1992 Tuberculosis Pulmonary F SAC 23 An 889 S-94-889 R Mid-T 1922 May 26, 1994 Cardiopulmonary Failure F SAC 72 An 888 S-95-888 R6 1913 May 27, 1995 Coronary Thrombosis F SAW 82 An 891 S-95-891 L6 1931 October 17, 1995 Myocardial Infarction M SAW 64 An 895 S-95-895 L6 1935 July 4, 1995 Cardiopulmonary Failure F SAC 60 An 935 S-98-935 R Mid-T 1947 April 19, 1998 Respiratory Arrest M SAC 51

143

Appendix C: BSE-SEM MATLAB Programs with Sample Images for Each Processing Step

% imageprogram.m

% Portions modified from Freehand_masking_demo.m by ‘Image Analyst’

https://www.mathworks.com/matlabcentral/profile/authors/1343420-image-analyst

% This program reads in all .tif files in the current directory and presents them to the user to identify a Region-of-Interest (ROI). The ROI is manually traced by the user to separate the cortical and trabecular bone compartments. The program then cleans the image, and extracts several bone parameters: pore area(um^2),mean pore diameter(um^2),pore density(/mm^2),cortical porosity (%), endosteal area(mm^2), cortical area(mm^2), total area(mm^2), cort lacunae size 1(#), cort lacunae size 2(#), cort lacunae size 3(#), cort lacunae area(mm^2), cort lacunae density(#/mm^2), trab lacunae size 1(#), trab lacunae size 2(#), trab lacunae size 3(#), trab lacunae area(mm^2), and trab lacunae density(#/mm^2).

Figure A1: Original BSE-SEM photomontage of transverse rib cross-section if(~isdeployed) cd(fileparts(which(mfilename))); end clc; % Clear command window.

144 clear; % Delete all variables. close all; % Close all figure windows except those created by imtool. imtool close all; % Close all figure windows created by imtool. workspace; % Make sure the workspace panel is showing. fontSize = 16; Conn_Limit = 7; % Limit which defines how lacunae are eliminated m = 3; % maximum Lacunae size Apx = (3000/645)^2; % Area per pixel um^2/px dirName = cd; %# folder path files = dir( fullfile(dirName,'*tif') ); %# list all *.tif files files = {files.name}'; %'# file names txtpath = strcat(dirName,'\','export.csv'); fileID = fopen(txtpath,'w'); fprintf(fileID,'Filename, mean pore area(um^2),mean pore diameter(um^2),pore density(/mm^2),cortical porosity, endosteal area(mm^2), cortical area(mm^2), total area(mm^2), cort lacunae size 1(#), cort lacunae size 2(#), cort lacunae size 3(#), cort lacunae area(mm^2), cortical lacunae density(#/mm^2), trab lacunae size 1(#), trab lacunae size 2 (#), trab lacunae size 3(#), trab lacunae area(mm^2), trab lacunae denisty(#/mm^2)\n'); % Create .csv file with headers

Base = linspace(0,255,256); Base = Base.'; for i=1:numel(files) ask = 'N'; fname = fullfile(dirName,files{i}); %# full path to file folder = fullfile(matlabroot, '\toolbox\images\imdemos'); fullFileName = fname; mkdir ([strrep(files{i},'.tif','')]) cd ([strrep(files{i},'.tif','')])

%Load in Image for Processing orgImage = imread(fullFileName); orgImage = orgImage(:,:,1);

%Despeckale Process

145 A = im2bw(orgImage,0.0625); %Convert to Black White - Default Settings BW2 = bwareaopen(A,8); %Eliminate all connected regions that are < 8 pixels in size) BW3 = im2uint8(BW2); % Convert Image to 8-bit BW Matrix BW4 = (1/255).*BW3; % Normalize non-zero elements grayImage = BW4.*orgImage;

% Save Image imwrite(orgImage,strrep(files{i},'.tif','^Original.tif')); % Save Image imwrite(grayImage,strrep(files{i},'.tif','_.tif')); while 1 imshow(grayImage, []); axis on; title('Original Grayscale Image', 'FontSize', fontSize); set(gcf, 'Position', get(0,'Screensize')); % Maximize figure

hFH = impoly(); accepted_pos = wait(hFH); prompt = 'Is this ROI Acceptable? Y/N [Y]: '; ask = input(prompt,'s'); if ask ~= 'Y'; continue end break end

%hFH = imfreehand(); % Create a binary image ("mask") from the ROI object. binaryImage = hFH.createMask(); xy = hFH.getPosition;

% Calculate the area, in pixels. numberOfPixels1 = sum(binaryImage(:)) endosteal_area(i) = numberOfPixels1*Apx/10^6; % Can also be calculated by taking the fractional pixels into account. numberOfPixels2 = bwarea(binaryImage)

% Get coordinates of the boundary of the freehand drawn region. structBoundaries = bwboundaries(binaryImage);

146 xy=structBoundaries{1}; % Get n by 2 array of x,y coordinates. x = xy(:, 2); % Columns. y = xy(:, 1); % Rows.

% Burn line into image by setting it to 0 wherever the mask is true. burnedImage = grayImage; burnedImage(binaryImage) = 0;

% Mask the image and display it. % Will keep only the part of the image that's inside the mask, zero outside mask. blackMaskedImage = grayImage; blackMaskedImage(~binaryImage) = 0;

% Save Image imwrite(burnedImage,strrep(files{i},'.tif','_Cortical.tif'));

% Create Histogram of Cortical Section A = imhist(burnedImage); A(1) = 0; figure bar (A,'histc') saveas (gcf,strrep(files{i},'.tif','_Cortical_Histogram.tif')); close

% Create Array for Export to .csv A(:,2)=A; A(:,1)=Base(:); % Export to .csv csvwrite(strrep(files{i},'.tif','_Cortical_Histogram.csv'),A); axis on; title('Masked Inside Region', 'FontSize', fontSize);

% Crop the image. leftColumn = min(x); rightColumn = max(x); topLine = min(y); bottomLine = max(y); width = rightColumn - leftColumn + 1;

147 height = bottomLine - topLine + 1; croppedImage = imcrop(blackMaskedImage, [leftColumn, topLine, width, height]);

Figure A2: (Left) BSE-SEM photomontage of transverse rib cross-section with trabecular bone removed. (Right) Histogram of cortical bone grey levels.

% Save Image imwrite(croppedImage,strrep(files{i},'.tif','_Trabecular.tif')); % Create Histogram of Trabecular Section A = imhist(croppedImage); A(1) = 0; figure bar (A,'histc') saveas (gcf,strrep(files{i},'.tif','_Trabecular_Histogram.tif')); close

% Create Array for Export to .csv A(:,2)=A; A(:,1)=Base(:); % Export to .csv csvwrite(strrep(files{i},'.tif','_Trabecular_Histogram.csv'),A);

148 axis on; title('Cropped Image', 'FontSize', fontSize);

Figure A3: (Left) BSE-SEM photomontage of a transverse rib cross-section showing the manually traced trabecular bone compartment. (Right) Histogram of trabecular bone grey levels.

% Eliminate Lacunae in Cortical Bone B = zeros(size(burnedImage,1),size(burnedImage,2),4); B(:,:,1) = im2bw (burnedImage,0.0625); imwrite(B(:,:,1), strrep(files{i},'.tif','_Cortical_BW.tif'));

149

Figure A4: Binary BSE-SEM photomontage of transverse rib cross-section with trabecular bone removed.

Lc_Area = 0; Calculate Lacunae in Cortical Bone for j = 2:m+1 Binv = ~B(:,:,j-1); Binv2 = bwareaopen(Binv,j); B(:,:,j) = ~Binv2; imwrite(B(:,:,j), strrep(files{i},'.tif',['_Cortical_BW_Clean_',num2str(j-1),'.tif'])); Lc(i,j-1) = (sum(sum(B(:,:,j)))-sum(sum(B(:,:,j-1))))/(j-1); Lc_Area = Lc_Area + Lc(i,j-1)*(j-1)*Apx/10^6; end

150

Figure A5: Binary BSE-SEM photomontage of transverse rib cross-section with trabecular bone and osteocyte lacunae removed.

% Eliminate Lacunae in Trabecular Bone B = zeros(size(croppedImage,1),size(croppedImage,2),4); B(:,:,1) = im2bw (croppedImage,0.0625); imwrite(B(:,:,1), strrep(files{i},'.tif','_Trabecular_BW.tif'));

Lm_Area = 0; % Calculate Lacunae in Trabecular Bone for j = 2:m+1 Binv = ~B(:,:,j-1); Binv2 = bwareaopen(Binv,j); B(:,:,j) = ~Binv2; imwrite(B(:,:,j), strrep(files{i},'.tif',['_Trabecular_BW_Clean_',num2str(j-1),'.tif'])); Lt(i,j-1) = (sum(sum(B(:,:,j)))-sum(sum(B(:,:,j-1))))/(j-1); Lt_Area = Lt_Area + Lt(i,j-1)*(j-1)*Apx/10^6; End

151

Figure A6: Binary BSE-SEM photomontage of a transverse rib cross-section showing the manually traced trabecular bone compartment with osteocyte lacunae removed. blankImage = imread(strrep(files{i},'.tif',['_Cortical_BW_Clean_' num2str(m) '.tif'])); blankImage(binaryImage) = 255; imwrite(blankImage,strrep(files{i},'.tif',['_Cortical_BW_Clean_Blank.tif'])); cd ..

Figure A7: Binary BSE-SEM photomontage of transverse rib cross-section with filled medullary cavity.

%Calculate Area and Pore metrics using analyzeimage.m function

152 [mean_pore_area(i),mean_pore_diameter(i),pore_density(i),cortical_porosity(i),total_area(i), cortical_area(i)] = analyzeimage(files{i}, Apx,Trabecular_Area(i)); %# load file

%Calculate Lacunae Densities Lc_Density = sum(Lc(:))/cortical_area(i); Lt_Density = sum(Lt(:))/Trabecular_Area(i);

% Print data to text file for further analysis fprintf(fileID,'%s,%f,%f,%f,%f,%f,%f,%f,%f,%f,%f,%f,%f,%f,%f,%f,%f,%f\n',files{i},mean_pore_area(i),mean_pore_ diameter(i),pore_density(i),cortical_porosity(i),Trabecular_Area(i),cortical_area(i),total_area(i), Lc(i,1), Lc(i,2), Lc(i,3), Lc_Area, Lc_Density, Lt(i,1), Lt(i,2), Lt(i,3), Lt_Area, Lt_Density); close end %Program End fclose('all');

153 % analyzeimage.m function [mean_pore_area,mean_pore_diameter,pore_density,cortical_porosity,total_area, cortical_area] = analyzeimage(imgpath, Apx, Trabecular_Area) se = strel('disk',1); % Create structural element to close voids cd ([strrep(imgpath,'.tif','')])

%// Read in image as binary im = im2bw(imread(strrep(imgpath,'.tif','_Cortical_BW_Clean_Blank.tif'))); figure, imshow(im), title('Original image'); iminvers = ~im; % Create image inverse [L,num] = bwlabel(iminvers); % Count number of pores number = num - 1;

IM2 = imclose(im,se); % Close small voids with structural element and store image filled_image = imfill(IM2,'holes'); % Close all pores and store image

% Save Image imwrite(IM2,strrep(imgpath,'.tif','_Cortical_BW_Clean_Blank_Closed.tif'));

Figure A8: Binary BSE-SEM photomontage of transverse rib cross-section with filled medullary cavity and cracks removed.

% Save Image

154 imwrite(filled_image,strrep(imgpath,'.tif','_Cortical_BW_Clean_Blank_Closed_Filled.tif'));

Figure A9: Binary BSE-SEM photomontage of transverse rib cross-section with all voids completely filled for calculation of total bone area. cd .. imarea = sum(im(:))*Apx; % um^2 Calculated by adding up all white pixels fillarea = sum(filled_image(:))*Apx; % um^2 Calculated by adding up all white pixels in filled_image holesarea = fillarea - imarea; % um^2

% Calculate parameters mean_pore_area = holesarea / number; mean_pore_diameter = sqrt(mean_pore_area/pi)*2; pore_density = number / (fillarea)*10^6; total_area = fillarea/10^6; %mm^2 cortical_area = total_area- Trabecular_Area; %mm^2 cortical_porosity = holesarea / cortical_area / 10^6; end