Skeletal biology over the life span: A view from the surfaces


  • James H. Gosman,

    Corresponding author
    1. Department of Anthropology, 4034 Smith Laboratory, The Ohio State University, Columbus, OH 43210-1106
    • Department of Anthropology, 4034 Smith Laboratory, 174 W. 18th Avenue, The Ohio State University, Columbus, OH 43210-1106, USA
    Search for more papers by this author
  • Samuel D. Stout,

    1. Department of Anthropology, 4034 Smith Laboratory, The Ohio State University, Columbus, OH 43210-1106
    Search for more papers by this author
  • Clark Spencer Larsen

    1. Department of Anthropology, 4034 Smith Laboratory, The Ohio State University, Columbus, OH 43210-1106
    Search for more papers by this author


The biocultural interpretation of skeletal remains is based upon the foundation of skeletal biology. In this review we examine the current state of skeletal biology research outside of the mainstream anthropology literature. The focus is on the structural changes of bone development and growth, and modeling and repair in the four bone surfaces: periosteal, Haversian, endosteal, and trabecular. The pattern of skeletal changes is placed within the framework of the human life span. New perspectives and direction of research on the environmental, biological, and genetic influences on modeling and remodeling processes are discussed chronologically at each bone surface. Implications for biological anthropologists are considered. This approach emphasizes variation in skeletal biology as a dynamic record of development, maturity, and aging. Yrbk Phys Anthropol 54:86–98, 2011. © 2011 Wiley Periodicals, Inc.

Skeletal biology has taken on a central role in a range of research areas that pertain to living and past primates, including humans. In this review we aim to examine the current state of understanding of skeletal biology, linking the morphology and structural changes of bone development, growth, remodeling, and repair within, on, and between the four surfaces of bone: periosteal, cortical/haversian, endosteal, and trabecular. The pattern of changes is placed within the framework of the human life span. Of specific interest are the age-associated and bone surface-specific influences of hormonal changes, the molecular and biochemical microenvironment of bone, and the mechanical environment on bone density, geometry, and microarchitecture. These types of biological data intersect with skeletal growth, development, and bone maintenance. The variations and perturbations of these processes provide physical anthropologists the opportunities and tools to interpret biocultural and behavioral aspects of ancient and recent human groups, such as bone mass accrual during subadult years and bone loss in later years. This contribution to the anthropological skeletal biology literature underscores the highly dynamic nature of skeletal tissues. It references the growing recognition that the mechanical and nutrient environment strongly influences bone beginning in utero and continues throughout life. Our goal is not to provide an exhaustive review of the bone biology or related bioanthropology literature, but rather to provide research perspectives currently being explored by bone science outside of anthropology-specific literature, that contribute to the growing discussion of the such issues as development and aging.


Bone is composed of two major compartments or envelopes in which modeling and remodeling occur, namely periosteal and endocortical. These, in turn, have the following available surfaces: periosteal, intracortical, endosteal, and trabecular (Martin et al., 1998; see Fig. 1). The modeling process involves the independent actions of bone formation and bone resorption on different surfaces, whereas the remodeling process involves the sequentially synchronized and coupled actions of bone resorption and bone formation on the same surface. Each surface, over the human life span, may have an age-specific (re)modeling response to various environmental, hormonal, and mechanical influences. The complex interactions within and among the nominal bone surfaces and their microenvironments are topics of intense research scrutiny. The discussion to follow, which treats them more or less as discrete entities, is organized that way for heuristic purposes. To this end, we present a selective review of the morphology and biological processes of these bone surfaces based on recent bone science research. The importance of this type of investigative framework for physical anthropology, bioarchaeology, and other osteologically related research is to establish an age-related context for the interpretation of skeletal data.

Figure 1.

High resolution microCT slice demonstrating the bone surfaces. PS, periosteal surface; IC, intracortical surface; ES, endosteal surface; TB, trabecular surface.


Periosteal morphology

The term “periosteum” combines the Greek peri “around, about” with osteon “bone,” indicating the outermost layer of long bone structure (Harper, 2008). This specialized tissue surface provides essential cellular processes and functions, which are important to changes in bone shape and biomechanical competence. The complexity of periosteal morphology is becoming apparent from recent research; a short digression into its functional morphology may be useful as background information. The periosteum is traditionally described as having two layers: an outer “fibrous” and an inner “cambium” layer. Researchers, however, have shown that this is an over simplification. The morphology and functional regions of the periosteum vary with both age and location (Anderson and Danylchuk, 1978; Raisz, 1984; Brommage et al., 1999). Ellender and coworkers (1988) in a study on rat caudal vertebra describe three layers which change during growth and development. The inner cambium layer contains osteoblasts which change from their initial elongated shape, to activated cuboidal cells, and then ultimately to a flattened quiescent state with age. Additional components of this layer include various progenitor cells, microvascular connections, and sympathetic neural elements (Allen et al., 2004). The middle layer described by these authors contains blood vessels, progenitor cells, and mononuclear cells, all gradually involuting with age. The outer fibrous layer containing collagen, fibroblasts, a neural and vascular network (Chanavaz, 1995), and elastin fibers (Taylor, 1992) becomes more dense, but less thick during development and skeletal maturity (Ellender et al., 1988). Periosteal processes, especially periosteal bone formation, are reduced in both rate and sensitivity to environmental changes (e.g., mechanical, hormonal sex steroid, growth hormone, and PTH), cytokine (BMP) influences, and age (see Allen et al., 2004 for references).

The response of the periosteum to changes in the hormonal and mechanical microenvironment is thought to reside in the highly cellular components of the cambium layer, and possibly in the pluripotential endothelial pericytes located in this highly vascularized tissue. Unfortunately, the true nature of these cellular elements remains ambiguous. Much of this research is based on cell cultures from calvarial periosteal cells. The relationship of these cells to periosteal osteoblasts and the validity of calvarial periosteum as a general model for the postcranial skeleton have not been established (Allen etal., 2004). These data based on calvarial cells gain additional uncertainty when one considers the anatomic and site specificity of periosteal morphology and process throughout the skeleton (Anderson and Danylchuk, 1978; Bromage et al., 1999).

Periosteal bone formation: Local and systemic influences

Periosteal cells respond to physiological levels of mechanical strain by increased cell proliferation, angiogenic response, and release of nitric oxide (NO) and PGE2 (prostaglandin E2) production, which exert a bone anabolic (formation) effect through various molecular pathways (Keila et al., 2001; McKensie and Silva, 2011; Robling et al., 2006; Turner and Robling, 2004). The signaling pathways by which the periosteal surface “senses” the need for and “effects” bone formation are largely unknown. Molecular dissection-related bone research is slowly providing insight. For example, the recently identified molecule periostin may be instrumental in the recruitment and differentiation of osteoblast precursors cells in the periosteum (Horiuchi et al., 1999; Oshima et al., 2002). It is expressed by preosteoblasts in the periosteal cambium layer during cell proliferation and secreted into the extracellular matrix (Oshima et al., 2002). Periostin, however, remains relatively poorly defined in its role as a secreted matrix protein related to periosteal morphology and function. Recent research has also associated it with mechanically stressed connective tissues such as a tendons, heart valves, and periodontal ligaments. These findings suggest that periostin has a biomechanical role in connective tissues by way of the regulation of Collagen 1 fibrillogenesis (Norris et al., 2007; Hoersch and Andrade-Navarro, 2010). Periostin is likely only one of many molecules to be discovered which are important in periosteal bone apposition and relevant to the understanding of bone's response to mechanical forces.

Parathyroid hormone (PTH) appears to have surface-specific and multimodal-metabolic effects in human bone. Intermittent daily exposure to PTH induces an increase in cortical wall thickness, which is a result of both periosteal and endocortical bone formation (Burr et al., 2001; Dempster et al., 2001; Lindsay et al., 2007). Paradoxically, continuous exposure to PTH in humans may stimulate bone formation on the periosteal surface, but not on the endocortical surface (Parfitt, 2002b). It increases bone matrix protein production in periosteal cells, but inhibits matrix production of endosteal cells in culture (Midura et al., 2003). Intermittent exposure to PTH is thought to enhance osteoblast differentiation (anabolic), whereas continuous exposure enhances osteoclastogenesis (catabolic; Locklin et al., 2003).

The influence of insulin-like growth factor (IGF-I) and growth hormone (GH) on human skeletal growth and development and periosteal apposition in particular has been intensely studied over the past two decades (see review Yakar et al., 2010). The IGF-I/GH interaction has a well-established positive effect on osteoblast differentiation and function in both humans and animal models. Animal models have indicated serum IGF regulation of the periosteal mechanism in modeling and remodeling, transverse and longitudinal bone growth, the resultant bone geometry, and catch up growth (Yakar et al., 2002, 2009; Fisher et al., 2005; Ellis et al., 2010). These investigations demonstrate that the complete loss of both serum and tissue IGF-1 has a deleterious effect on bone growth and or gain at all ages, whereas tissue IGF-1 appears to affect early growth and development, namely the early postnatal and prepubertal periods (Yakar et al., 2010).

Sex steroids have a profound age-related influence on the metabolism, structure, and function of periosteal cells, which express estrogen receptors (alpha and beta), androgen receptors, and enzymes instrumental in sex steroid interconversion. A comprehensive review of these influences and mechanisms is beyond the scope of this article (see Veldhuis et al., 2006 for a recent review). Generally, estrogen inhibits periosteal apposition, while androgens enhance periosteal apposition. Estrogen receptor-alpha (ERα) in cortical bone is a major regulator of mechanical load-induced periosteal bone apposition. Androgen receptors in the periosteum are associated with testosterone-induced periosteal bone formation. In addition, Venken and collaborators (2008) propose that both androgens and estrogens are necessary to maximize male periosteal expansion. The sex steroid balances in puberty and in aging are essential factors in the early development of sexual dimorphism of bone size, shape, and geometry and the later loss of bone architecture (Seeman, 2003). These age- and sex-modulated findings can be explained by a dosage-dependent effect in which low dosages of estrogen (e.g., prepubertal, postmenopausal, and in males) stimulates periosteal expansion, and high dosages of estrogen as occurs in pubertal and premenopausal females inhibits periosteal expansion.

The interactions between periosteal and endocortical bone surfaces are complex and incompletely understood at present. The patterns of periosteal age-related changes are perhaps best considered within the broader framework of the relative balance of modeling and remodeling processes involving both periosteal apposition/resorption and endocortical formation/resorption (Balena, 1992; Parfitt, 2002a; Bliziotes et al., 2006). Genetic data in animal models suggest that there are genetic components to the responsiveness of this system and the resulting periosteal/skeletal dimensions (Masinde et al., 2003; Wergedal et al., 2005; Kesavan et al., 2006). Duren and co-workers (2011) have used genotyped children from the Fels Longitudinal Study to demonstrate changes in the genomic regions influencing pediatric bone structure during growth and development. These type of data support the concept that the interaction of genetic and biological factors contribute to the variation in the temporal sequences of change in anatomical region-, site-, and sex-specific skeletal surfaces.


Trabecular bone ontogeny

The highly conserved and regulated process of endochondral ossification is responsible for the basic trabecular bone framework in which subsequent biologically and mechanically driven modeling/remodeling occurs (Gosman and Ketcham, 2009). Trabecular bone morphogenesis from growth plate cartilage to the secondary spongiosa is thought to be qualitatively predictable and very similar among mammalian species (Salle et al., 2002). Descriptive qualitative and quantitative histomorphometric data on the growth plate and associated metaphyseal region during human growth and development are well established in the literature (Felts, 1954; Fazzalari et al., 1997; Kneissel et al., 1997; Byers et al., 2000; Glorieux et al., 2000; Parfitt et al., 2000; Hall, 2005). Trabecular bone mass increases with age via an increase in trabecular thickness until skeletal maturity at approximately age 20.

Quantitative study of human fetal bone development by histomorphometric and microCT methods suggests that trabecular bone microstructure is influenced by maternal–fetal perturbations (Salle et al., 2002; Nuzzo et al., 2003). Results demonstrate an extremely rapid rate of trabecular bone metabolism, cell division, and modeling, especially in the last trimester of fetal development. This is manifested by increasing bone volume fraction, trabecular thickness, brisk matrix mineralization, and increasing hydroxyapatite crystal size. Research into the fetal origins of reduced bone mass in utero commonly use bone mineral density (BMD) derived from dual energy X-ray absorptiometry or ultrasound scanning (Mahon et al., 2010). Although BMD is not the equivalent of 3D microstructural measurements, the general trends are correlative (Nuzzo et al., 2003). Evidence suggests that environmental influences during early life (intrauterine) interact with the genome in establishing the functional level of a variety of metabolic processes involved in skeletal growth, including neonatal bone mass (Cooper et al., 2002; Javaid and Cooper, 2002). Adjusting for gestational age, neonatal bone density is positively associated with birth weight, birth length, and placental weight. Maternal factors negatively associated with neonatal bone density are maternal smoking, maternal nutrition at 18-weeks gestation, and high maternal physical activity (Godfrey et al., 2001). Neonatal bone density has been demonstrated to be lower among winter births than among summer birth, which is associated with winter month maternal vitamin D deficiency (Javaid and Cooper, 2002).

The emerging model for fetal programming links environmental conditions during embryonic and early development with risks of disease later in life (Godfrey, 2002). Mechanisms for the induction of fetal programming in regards to skeletal development are thought to include the nutritional environment (in the most general sense). This may permanently alter gene expression within the growth plate affecting endocrine systems such as growth hormone/insulin-like growth factor I, the hypothalamic-pituitary-adrenal-gonadal axis, and vitamin D (Cooper et al., 2002 and references cited within; Mohan et al., 2010). In addition, the nutrient environment may permanently reduce cell numbers in the growth plate. The high growth rates of the fetus are mostly the result of cell replication; fewer cells equal a reduced capacity for growth. The slowing of growth that is occurring is considered an adaptation to undernutrition (Cooper et al., 2002). The importance of this from a potential bioanthropological research perspective, is that undernutrition and other adverse influences arising in fetal life can have a permanent effect on skeletal size, volume, and structure throughout the life span (Tobias et al., 2005). The consequences of undernutrition and other intrauterine environmental factors on skeletal mineralization, neonatal bone density, and later attainment of peak bone mass is evident from recent human studies (Yin et al., 2009; Harvey et al., 2010). These are key variables in the study of human skeletal variation and the human condition. Perturbations are thought to have the possibility of beginning during maternal development as a fetus (grandmother/mother). This brings into play a transgenerational approach to interpretations of skeletal morphology (Gluckman et al., 2007).

Byers and coworkers (2000) used quantitative histomorphometric analysis of the costochondral junctions to suggest that proliferative, hypertrophic, and primary spongiosa regions of the growth plate/metaphysis are the “active growth unit.” These zones are responsible for producing an increase in bone volume fraction (BV/TV) by an increase in cartilage septa and trabecular thickness. The bone volume fraction and trabecular thickness were found to increase with age while trabecular number decreased. The secondary spongiosa was characterized by a more stable consolidation of trabecular structure. These researchers noted that trabecular bone structural parameters changed most rapidly during the first year of life. It is during this period that rapid growth and a marked change in biomechanical influences on trabecular bone coincide. This ontogenetic trabecular bone pattern can be described by the mechanical usage endochondral ossification model outlined by Frost and Jee (1994a,b) and Wong and Carter (1990). This model proposes that mechanical strain is the controlling mechanism for endochondral ossification and that mechanical usage effects on trabecular bone result in a mechanically adapted state (Frost and Jee, 1994a). The researchers account for the early loss of relatively underloaded, small trabeculae (decreasing BV/TV), and the recovery of mass (increasing BV/TV) by increased trabecular thickness. Empirical data support the importance of mechanical usage in the early age-related loss of primary spongiosa, and subsequent reworking of the trabecular structure (Frost and Jee, 1994b; Gosman and Ketcham, 2009).

In spite of the importance of ontogenetic patterning for understanding the functional adaptation of trabecular bone structure, very little quantitative data are available for trabecular structure in skeletally immature individuals. Until recently, work on trabecular bone ontogeny has focused on nonprimate mammals, including pigs, sheep and dogs (Nafei et al., 2000a,b; Tanck et al., 2001; Wolschrijn and Weijs, 2004). Tanck and coworkerss (2001) analyzed bone microarchitecture and mechanical behavior in pig tibiae and vertebrae, finding that trabecular bone increases in volume and anisotropy with age and body mass. They found that the initial response to increased loads is to add bone mass and then refine that mass later into a more mechanically efficient, anisotropic structure. Wolschrijn and Weijs (2004) analyzed the ulnar coronoid process of dogs, revealing that the bone volume fraction and trabecular thickness increase with age while trabecular number displays the opposite trend, decreasing in older individuals. Wolschrijn and Weijs (2004) note that dogs do not stand until 1.5 weeks and don't walk with steady gait until about 3 weeks. Therefore, they may retain primary bone from the fetal growth stage until several weeks after birth.

The reorganization of trabecular structure in pigs and dogs matches the results from ontogenetic studies in humans by Ryan and Krovitz (2006) and Gosman and Ketcham (2009), which further suggest that mechanical loading from locomotion is one of the essential driving forces in the adaptation of bone structure. In separate works, these researchers have analyzed bone development in the proximal humerus and proximal femur (Ryan and Krovitz, 2006; Ryan et al., 2007) and proximal tibia (Gosman, 2007; Gosman and Ketcham, 2009) in a growth series from two prehistoric North American human populations. Ryan and coworkers (2006, 2007) documented the developmental patterns in a sample of proximal femora and humeri from the Norris Farms #36 skeletal series, ranging in age from fetal through ∼ 11-years old. Gosman (2007) and Gosman and Ketcham (2009) completed a study of trabecular bone development in the proximal tibia from neonatal to 24 years of age in a sample from the SunWatch Village site. Both Norris Farms #36 and SunWatch Village were late prehistoric village maize agriculturalists. These studies (Ryan and Krovitz, 2006; Ryan et al., 2007; Gosman and Ketcham, 2009) quantify the ontogenetic patterns of structural change in trabecular bone in the postcranial skeleton of humans during the development and maturation of unassisted bipedalism. The trabecular architecture across these three anatomical locations (proximal femur, humerus, and tibia) is remarkably similar over the early human life span (see Fig. 2). The bone volume fraction (BV/TV) values for all three bones start out very high in the youngest individuals, decrease to reach a minimum around 1 year of age, and then progressively increase to adult levels through development. The higher BV/TV in the lower limb suggests the influence of weight-bearing mechanical forces. The trabecular thickness results also show a divergence between the humerus on the one hand and the femur and tibia on the other, with relatively increased thickness in the latter two elements compared to the former (cf. mechanically relevant upper and lower limb divergent structure in adults: Lochmüller et al., 2002; Nikander et al., 2006). Trabecular thickness does increase progressively with age in all three anatomical locations. The results for trabecular number and anisotropy (trabecular directionality) remain relatively similar across all three bones over the subadult portion of the life span (Ryan et al., 2010).

Figure 2.

Results of analyses of human trabecular bone structural ontogeny in the proximal humeral (red hexagons), femoral (blue triangles), and tibial (grey squares) metaphyses from the Norris Farms #36 site (humerus and femur) and the SunWatch Village site (tibia). Courtesy of T. Ryan. [Color figurecan be viewed in the online issue, which is available at]

The quantified trabecular bone structural parameters clearly document the process of endochondral ossification at the metaphyseal margin. The dense, uniformly oriented bony columns representing the primary spongiosa are laid down during the initial ossification of the cartilage model. They are then remodeled into divergent morphologies reflecting the locomotor and mechanical demands of each bone as well as the underlying genetic framework for skeletal patterning (Judex et al., 2004). While the results suggest a broad similarity in developmental processes across the three anatomical sites, they also provide evidence of a mechanical influence from locomotion. The distinct divergence in structure between the upper and lower limb elements in these analyses strongly suggests a mechanical component driving the remodeling of trabecular bone.

The remodeling of trabecular architecture includes an increase in the bone volume fraction (BV/TV) and trabecular thickness with a concurrent decrease in the number of trabeculae (Gosman and Ketcham, 2009). Researchers have also demonstrated differentiation of age- and gait-related structural zones within the femoral head, femoral neck, and proximal tibia, suggesting that the trabecular bone is adapting to the normal loads during walking (Ryan and Krovitz, 2006; Gosman, 2007; Gosman and Ketcham, 2009). A link can be made between 3D microstructure of tibial trabecular bone and specific events during gait maturation, including changes in the relative BV/TV between the medial and lateral tibial condyles corresponding to the development of the femoral bicondylar angle, as well as an increase in BV/TV in the posterior aspect of the tibial condyles associated with a greater functional flexion arc in the knee (Gosman, 2007; Gosman and Ketcham, 2009; Gosman et al., 2010). Using finite element models to predict elastic properties (the relationship between stress and strain) of trabecular bone during development, Ryan et al. (2007) demonstrated the mechanical significance of the structural differences between the femur and humerus. The elastic properties including stiffness (the resistance to deformation by an applied force) of femoral trabecular bone increased at a much faster rate than those of the humerus after the acquisition of unassisted bipedal walking. These results match those of Ruff on the relative chronological development of strength characteristics in the cortical bone of the humerus and femur in humans (Ruff, 2003). The pattern of ontogenetic changes in trabecular bone architecture suggests a different developmental trajectory and chronology than cortical bone. Human long bone trabecular microarchitecture has been demonstrated to reach its essential adult configuration by the time an individual reaches late childhood (Gosman and Ketcham, 2009), whereas the geometric properties of cortical bone continue to change past adolescence (Hubbell et al., 2011).

Locomotor ontogeny

Experimental studies of human locomotor development may explain many of the ontogenetic changes in trabecular bone described above. Several comprehensive studies of locomotor kinematics during ontogeny show major differences in joint kinematics in young children compared to adults that may have an important impact on bony development (Haywood, 1986; Sutherland et al., 1988; Sutherland, 1997; Grimshaw et al., 1998). In particular, children younger than ∼ 7-years old use a significantly more flexed knee and hip during bipedal walking compared to adults. As children mature, their joints become more extended, but do not reach adult patterns of extension until age seven (Sutherland et al., 1988). Relatively extended hips and knees are hallmarks of modern human bipedalism, and the evolution of these kinematic characteristics is considered a key component of hominin evolution (Sockol et al., 2007). Additionally, the femur remains relatively vertical (rather than abducted or adducted) during early ontogeny. This may have important implications for the development of the bicondylar angle (Shefelbine et al., 2002). Finally, a prominent heel strike, which is associated with large forces in adults, is not present until ∼ 18 months (Sutherland et al., 1988).

Other researchers have examined the kinetics of human locomotor ontogeny, and their results also show gradual changes during development (Chester et al., 2006; Chester and Wrigley, 2008). Ground reaction forces are higher at younger ages (as a percentage of body weight) and become lower as locomotor maturity is reached (Beck et al., 1981). Increased forces at young ages are due to relatively large vertical, fore-aft and medio-lateral forces (Beck et al., 1981). Combined with the kinematic changes, these studies suggest that forces that must be resisted by lower limb bones are both higher, and are experienced at different angles relative to limb segments in younger children compared to adults.

The amount of stride-to-stride variation in locomotor biomechanics may also play a major role in trabecular development. During human locomotor development, there is a general shift from high stride-to-stride variability in kinematic and kinetic parameters to lower variability with age (Adolph et al., 2003; Lasko-McCarthey et al., 1990). This variation is likely an essential element in the pattern of changes seen during development of trabecular architecture and organization in humans. The results from these types of studies strongly suggest an age-specific mechanical influence on trabecular bone development. Additional research is needed to understand the relative importance of the various factors including mechanical loading, general developmental processes, biomolecular microenvironmental changes, and genetic patterning. A current collaborative research project is examining the spatial and temporal patterns of change in trabecular bone development as related to loading regime changes in both magnitude and direction during locomotor ontogeny (Ryan et al., 2010). Trabecular bone research has been emboldened by the continued development of ever more powerful investigative tools (e.g., ex vitro and in vivo microCT imaging, quantitative 3D visualization, molecular dissection techniques, and finite element modeling), all of which can be applied to unravel the mysteries and ambiguities in skeletal data and contribute to the bioarchaeological research agenda.

Age- and sex-associated changes in trabecularbone

Seeman and Delmas (2006) have outlined four processes in bone which contribute to the age- and sex-associated loss of trabecular and cortical bone mass and structure: 1) reduction in bone formation at the cellular/basic metabolic unit level (BMU), with a reduction in the amount of bone formed in each BMU; 2) reduction in bone formation from attenuation of periosteal bone modeling; 3) increased bone resorption in each BMU, accentuated for a time by sex hormone deficiency (primarily estrogen); and 4) an overall increase in the bone remodeling rate after menopause. Age-related bone loss is now thought to begin as early as the third decade of life, continuing at varying rates until the end of life (Riggs et al., 2008).

The differential effects of remodeling on trabecular bone compared to cortical bone are due to the greater surface area in the former. This allows for more remodeling sites per unit volume (Parfitt et al., 1983). With age, each remodeling event results in a net loss of bone as more bone is resorbed than is reformed. This, in turn, establishes the age-associated pattern of trabecular thinning, loss of trabecular plates, and loss of connectivity, all producing a progressive loss in bone strength (Ding, 2000; van der Linden et al., 2001; Ammann and Rizzoli, 2003; van der Linden and Weinans, 2007; Djuric et al., 2010). Recent research indicates that bone loss in men is characterized by trabecular thinning, whereas bone loss in women by trabecular loss, perforation and loss of connectivity (Khosla et al., 2006; Riggs et al., 2008). The deficit in mechanical integrity is greater for the latter. Finite element modeling indicates that the effect of reductions in trabecular number on bone strength is two to five times greater than that of reductions in trabecular thickness (Silva and Gibson, 1997). Numerous populational-based studies using 3-D high-resolution CT imaging and the quantification of trabecular bone structural parameters demonstrate that men begin adult life with better trabecular microarchitecture and have less age-related microstructural deterioration. That is in men, trabeculae are thicker at the attainment of peak bone mass and lose less connectivity and fewer trabeculae with aging (Dalzell et al., 2009; Melton et al., 2010; Macdonald et al., 2011).


Growth and development

A fundamental aspect of long bone cortical bone growth is the orderly and progressive process of shape readjustment by modeling and remodeling on all the surfaces. The ontogenetic patterns of change in length and diameter during growth results in the repositioning of local regions by way of a reciprocal or complementary process of bone apposition on selected surfaces and bone resorption on others (Ohman et al., 1997; Zebaze et al., 2005). The end result is cortical bone which can be characterized as a morphologically complex and heterogeneous structure consisting of components of periosteal, Haversion, endosteal-derived, and compacted trabecular bone (see Enlow, 1963 for the classic discussion of this; see Fig. 3). Longitudinal bone growth by endochondral ossification creates new primary trabecular bone, but is also instrumental in the formation of metaphyseal cortical bone. Cadet and coworkers (2003) have demonstrated that the longitudinal and appositional growth of the metaphyseal cortex occurs by coalescence (compaction) of trabecular bone formed at the periphery of the growth plate. This coalescence is associated with an increased osteoblast surface on the peripheral endochondral trabecular bone, likely related to the inductive effects of the nearby periosteum, and modulated by mechanical forces (Sundaramurthy and Mao, 2006; Tanck et al., 2006). MicroCT scan images suggest that this process is age-related with the merging of peripheral trabeculae beginning to occur between 6 months and 1 year of age (Gosman, 2007). The cortical bone growth process can be further classified into four developmental phases based on the histological study of modern subadult ribs by Streeter (2010). These phases describe the presence and distribution of woven and lamellar bone as well as primary, secondary and drifting osteons, reflecting the results of age-specific bone modeling and remodeling.

Figure 3.

Long bone inward growth and reshaping are accomplished by the combination of periosteal resorption/formation, trabecular bone compaction, and endosteal deposition. Side A depicts inward reshaping and compaction into a preexisting trabecular region (1) and new endosteal apposition (2). Side B demonstrates the resulting heterogeneous cortical structure including both a whorled arrangement from trabecular compaction (y) and regular circumferential lamellae (x). Adapted from Enlow (1963).

The population variance of cortical bone phenotypic traits such as cross-sectional area and density are established early in life, generally by 1–2 years of age (Maresh, 1961). Growth-related tracking of these bone traits may be modulated by the effect of the environment, life style, and disease, thus contributing to patterns of human skeletal variation (Cooper et al., 1995). Ontogenetically related changes in cortical bone strength become site-specifically adapted to local and more distant mechanical forces by the selective and strategic distribution of bone mass and shape (Seeman, 2008; Wang et al., 2009). Wang and coworkers (2005) have demonstrated this pattern in tibial cross-sectional geometry in a 2-year longitudinal study of prepubertal and pubertal girls. Periosteal formation was found to favor the anterior and posterior cortical surfaces, increasing the bending strength preferentially in the anteroposterior compared to the mediolateral axes. Hubbell and coworkers have presented data from a developmentally seriated archaeological skeletal sample of tibiae supporting this pattern (Hubbell et al., 2011). Carpenter and Carter (2008) have described a cortical bone mechanobiological computational model incorporating the influences of periosteal tensile and pressure strains on periosteal and endosteal bone formation and resorption. Simulations of the model on rat tibia development suggest that cortical cross-sectional shape is a result of the combination and interaction of far-field loads (body mass) and local periosteal surface loads from adjacent muscle activity. This indicates the importance of bone-muscle interaction in addition to body shape and mass to the development of diaphyseal cross-sectional morphology (Schoenau and Frost, 2002).

Ethnic variations in phenotypic aspects of cortical bone developmental change in children have been observed. Three recent studies using peripheral quantitative computed tomography (pQCT) to examine bone density and cortical area are noteworthy. In their study of African-American, European-American, and Hispanic children, Wetzsteon et al. (2009) reported that ethnic differences in bone strength are manifest in children. On the basis of the measurements taken at the radius and tibia, African-American and Hispanic children exhibit significantly higher bone strength than European-American children due to greater volumetric density and cortical area. Gilsanz and coworkers (1998) report that, while black children had greater total areas than white children, cortical area did not differ. They conclude that bone mass distributed farther from the center of the bone provides greater strength to the bones of black children. These data are consistent with evidence that the apparent ethnic differences in cortical bone density during growth are due to bone size differentials and are developmental stage-specific (Seeman, 1997; Leonard et al., 2010).

Peak growth velocity and cortical bone strength

The progressive increase in cortical strength during growth and development takes a “break” during the period of peak growth velocity (PGV) of height in adolescence in both sexes. PVG is one of the maturity indicators related to the pubertal growth spurt. It typically takes place at a mean of 11.5 years in girls and 13.5 years in boys (Tanner and Whitehouse, 1976). The incidence of fractures in the distal forearm is highest at this time (Bailey et al., 1989). Cortical geometry and bone mass are major determinant of fracture risk at this site (Muller et al., 2003). Recent research has explored whether age-specific, relative cortical bone thinning, and microarchitectural changes could explain this increase in forearm fracture risk during puberty (Kirmani et al., 2009; Wang et al., 2010).

Parfitt (1994) took note of this phenomenon, suggesting that cortical thinning during the adolescent growth spurt was associated with a degree of secondary hyperparathyroidism resulting from the strong demand for calcium. Kirmani and coworkers (2009) compared trabecular and cortical bone high-resolution pQCT-derived structural parameters in 6- to 21-year olds. Trabecular bone parameters remained unchanged, except for males in late puberty and maturity. The interesting finding was the transient increase in cortical porosity during the period of peak growth velocity in both sexes; this was associated with an increase in PTH levels. This chronology coincided with the spike in forearm fracture risk in both sexes. After this period, cortical thickness and density underwent consolidation and a linear increase through skeletal maturity.

Functional adaptation

Cortical skeletal surfaces under the influence of sex hormones during the prepubertal to young–adult portion of the life span are highly responsive to mechanical loads generated by physical activity. Recent research has investigated and highlighted the importance of sport-specific activities, body shape, and underlying genetic influences on age-associated human skeletal structure and geometry (Kodoma et al., 2000; Ruff et al., 2006; Lagerholm et al., 2010; Wallace et al., 2010; Shaw and Stock, 2009a,b, 2011). These types of responses have been studied effectively using the mechanically loaded rat forelimb animal model. This research has demonstrated the extent of local response, investigated signaling pathways, and characterized the time course of bone response (formation) after the initiation of load (see Robling et al., 2006 for a review on these topics).

The chronology of bone formation/resorption changes are ultimately related to time-dependent gene expression patterns (Liedert et al., 2006). Recent research has investigated the timing of activation of these gene sequences (Mantila Roosa et al., 2011). This is life span consideration at an entirely different scale (e.g., biomolecular). It contributes to this discussion by adding mechanism and chronology to the process of bone functional adaptation: shining light into the “black box.” Data indicate that new osteoblasts are present on the bone surface within 48 hours after mechanical loading. This is followed by matrix production and bone formation lasting for 5 weeks (Turner et al., 1998). Six weeks after loading, bone formation has returned to baseline levels (Schriefer et al., 2005). Mantila Roosa and coworkers (2011) have investigated this pattern of bone formation in regards to a time sequence of loading-induced gene expression, identifying six clusters of multiple genes. These time-related patterns characterized essential periods of bone formation including osteoblast proliferation and differentiation, matrix production, and bone cell accommodation (desensitization).

Cortical aging

Changes in cortical (and trabecular) bone composition and structure over an individuals life span have significant consequences in its functional properties. Age-related changes in the bone collagen fibril network and mineralization have been relatively underserved in the anthropological bone science literature. Factors contributing to this status include the preservation-based limitations in studying collagen in archaeological skeletons as well as a general lack of focus on these issues. During growth and maturation, bone strength and stiffness increases, partly due to higher bone mineral density, changes in the relative chemical composition, and reworking of the structure and orientation of the collagen fiber and mineral crystals (Tanck et al., 2004; Skedros et al., 2006; Isaksson et al., 2009). This changing composition and structural modification of bone tissue becomes especially important with aging (Wang et al., 2002). Alterations in the collagen component of bone, specifically an increase in cross-linkages and impairment in the isomerization (molecular rearrangement) of collagen, contribute to increasing brittleness with age (Viguet-Carrin et al., 2006; Isaksson et al., 2010).

The contribution of cortical bone to bone quality and bone loss with age, is often underestimated. A significant proportion of skeletal bone mass is cortical. For example, the amount of cortical bone in vertebrae has been estimated at 30–60%, and the cortical bone of the vertebral body bears 45–75% of the axial load (Roux et al., 2010). In the femoral neck, the cortical shell contributes 40–90% of the bending rigidity (Burr, 2010). Cortical thinning and increased intracortical porosity are features associated with progressive bone fragility later in life. Periosteal apposition during adulthood is widely thought to be an adaptive response to the decrease in cortical bone strength produced by endosteal bone loss (Alhborg et al., 2003). Cross-sectional data supporting this concept have a number of confounding factors, such as the precision of methods used to detect very small changes, sex differences, and secular changes in bone size (Lazenby, 1990; Balena et al., 1992). Szulc and coworkers (2006) provide interesting insight in a prospective study of perimenopausal women, demonstrating that in premenopausal women endocortical bone loss was associated with periosteal apposition necessary to maintain cortical area and bending strength by distributing bone mass further from the neutral axis. After menopause, however when endocortical bone loss is increased and periosteal apposition is decreased, this balance is no longer maintained. Cortical thinning and bone fragility emerges (Seeman, 2008).

Accelerated endosteal remodeling in later ages is associated with physiologic sex hormone deficiency in women, sex hormone deficiency in men, and secondary hyperparathyroidism in both women and men (Seeman, 2008b). In addition to cortical thinning, this detrimental effect on cortical bone is also characterized by the combination intracortical porosity and cortical-endosteal resorption (“trabecularization” of cortical bone; Brancaccio et al., 2003; Zebaze et al., 2010). The increased rate of endosteal and intracortical remodeling combined with a decrease in bone formation in each BMU increases the endosteal surface area and porosity of the intracortical bone (Foldes et al., 1991). The total area of cortical porosity increases as a result of void coalescence. Bone loss becomes more cortical than trabecular, and structural degradation progresses (Yeni et al., 1997). Age-related intraskeletal variability in these processes underscores the importance of developing bone-site-specific standards. This is critical if normal and pathological states are to be identified and the results of different studies of skeletal series are to be compared (Eckstein et al., 2007; Seeman et al., 2001).


Ontogeny through aging

The cellular machinery which both constructs (modeling) and maintains (remodeling) bone size, shape, and structure is the key toward understanding the deposition and/or removal of bone from the various surfaces (see Datta et al., 2008 for a review of bone cellular biology). The growth-related processes in establishing the essential bone size, contours, and architecture, are directed toward obtaining peak bone mass and strength (see Fig. 4). This trajectory is established early in life, well before puberty (Loro et al., 2000). The skeletal maturity- and aging-related processes contribute to repairing fatigue damage from repetitve loading with reconstruction of the structure. It is the imbalance in the remodeling process, the rate of remodeling, and the constraints to periosteal bone formation, which lead to progressive bone loss and structural degradation during aging (Abright et al., 1941; Seeman and Delmas, 2006). This degradation has profound implications for increased risk of fracture due to reduced skeletal mass in later adulthood (Glencross and Agarwal, 2011). Influences on bone adaptation and morphology during the latter portion of the human lifespan includes polygenetic interactions, senescence and accumulating perturbations in the cellular processes, both hormonal deficiency (sex hormones) and excess (secondary hyperparathyroidism), changes in the bone microenvironment in regards to various growth factors and signaling molecules, loss of muscle mass and physical activity, nutritional imbalances, and as yet unkown biomolecular factors. Events over a person's life span interact with these factors, and contribute to the variability and complexities for the interpretation of skeletal data.

Figure 4.

On the surfaces. During growth periosteal apposition and the combination of relatively less endocortical resorption prior to puberty and relatively more peripubertal endocortical apposition thickens the cortex. Trabecular bone surfaces thicken and become more plate-like. Endocortical, intracortical, and trabecular resorption during aging result in cortical thinning, porosity and trabecular loss or thinning. Age-related periosteal apposition (bold outline) can partially offset these changes. (From Seeman, 2007; Springer, reproduced by permission).


Implications for biological anthropologists

Deconstructing bone geometry and mass into surfaces with age-specific patterns of change in structure has important implications for bioanthropological skeletal research of past (bioarchaeology, paleoanthropology) and living (human biology) groups. Three interrelated examples are discussed: methodological issues, biomechanical applications, and bone accural/loss patterns. Three methods commonly used in the study of bone loss in archaeological samples are metacarpal radiogrammetry (Ives and Brickley, 2005), cortical rib histomorphometry (Stout and Lueck, 1995), and vertebral trabecular architecture (Agarwal and Grynpas, 2009). Each of these methods focuses on a different bone element and surface. Beauchesne and coworkers (2011) have recently demonstrated that each method (and surface) may have a different age-related pattern of bone loss, thus offering alternative interpretations. Their results emphasize the importance of consideration of surface-to-surface comparisons when choosing analytical methods to analyze bone fragility in cross comparative research.

Biomechanical analysis of bone geometry is an integral component of the anthropologist's toolkit, useful in assessing general aspects of past activity patterns and body shape (Shackelford, 2007; Ruff, 2008). The diversity of long bone mass distribution is attributable to surface-specific bone formation/resorption in response to growth and aging, mechanical demand, and genetic patterning (Seeman, 2008a). This offers a portal to examine bone's heterogeneic adaptations surface-by-surface, sector-by-sector, and age-by-age. Wang and coauthors (2011, p 935) have published data testing this concept, framed in the hypothesis that “bone loss is the reversal of accural.” The authors report that the distribution of bone loss during older age in weight-bearing long bones reverses the preferential distribution of bone formation during early life. We argue that age- and bone surface-specific data may contribute to a more finely grained bioanthropological interpretation of bone structure, such as when and where a surface responds to demand (or lack thereof).

The future

This review has presented perspectives of time and history across the human life span in shaping bone mass and structure. It is the challenge to anthropology to place these age-related changes in skeletal morphology into the broad framework of human variation. Advanced technologies for quantifying and visualizing skeletal data are proving important to the advancement of integrated, cross-disciplinary research. Exciting possibilities exist for the expansion of anthropological research into newly developing disciplines concerned with topics such as the transgenerational influences (maternal–fetal environment), the interaction between bone and the immune system (osteoimmunology), and the influences of nutrition on differential gene expression and skeletal morphology (nutrigenomics). This article has developed the case that a view of skeletal surfaces across the life span may contribute to the understanding of the pathways of change and circumstances reflected in the human skeleton.


The authors thank Robert Sussman for his support toward the contribution of this article to the Yearbook of Physical Anthropology. The ideas for this article began in their individual and collaborative research programs. Many of the ideas expressed here were engendered by the ongoing discussions at and support from The Ohio State University. The authors thank the editors and anonymous reviewers for their constructive comments and suggestions which have resulted in a greatly improved manuscript.