THE RECOGNITION THAT bone size is an important determinant of bone strength has led inexorably to a consideration of factors that affect periosteal events. Similarly, there has been a growth in awareness of sex differences in fracture risk, and in turn, of sex-specific effects on bone size and geometry.(1) Even more recently, treatment with parathyroid hormone has been speculated to exert its antifracture effect in part through an increase in periosteal bone apposition (in addition to its major effects on trabecular bone mass), thus serving to identify the outer bone envelope as a target of therapeutic intervention.(2–4) Although imperfect, techniques have been developed that make the evaluation of bone size and geometry more feasible in clinical investigations,(5, 6) portending further interest in this area of investigation. Nevertheless, periosteal biology itself remains little investigated and poorly understood. Thus, it may be useful to consider several basic biological issues that could profoundly affect the consideration of the periosteum. The nature of the periosteum at the femoral neck is of special importance. In this discussion, the periosteum is considered a unique tissue compartment at the outer bone surface that, at least conceptually, could be the site of cellular events important for overall skeletal competence. The purpose of this perspective is to stimulate an expansion of the ways in which we consider periosteal physiology, pathophysiology, and therapeutics, specifically by re-examining common perceptions about periosteal cellular activity and their implications. Purposefully, somewhat provocative suggestions are intended to stimulate the development and testing of new hypotheses.
FRACTURE RESISTANCE AND BONE SIZE
Cellular events at the periosteal surface are responsible for bone diameter. Before considering periosteal biology, it is important to appreciate the role of bone size in the determination of strength. Certainly, the biomechanical competence of bone is affected by a host of factors, including mass, material properties, microarchitecture, and overall size/geometry. The size of bone is a particularly important contributor to bone strength because the resistance of bone to bending or torsional forces is related exponentially to its diameter. So, small contributions in the circumference of a bone add considerably to its fracture resistance. For instance, in the long bone schematically depicted in Fig. 1, the section modulus (z, an expression of bending or torsional resistance) is expressed as z = π/4r where ro is the periosteal radius and ri is the endosteal radius.(7) This is a geometrical parameter that does not take into account any material property, such as elastic modulus. In a typical example of the importance of geometry, a 65-year-old woman may have a femoral shaft ro = 1.6 cm and ri = 1.2 cm, with the section modulus being approximately 2.2 cm3. If ro increased by only 1 mm, the section modulus increases 32%, to 2.9 cm3. This represents an increase of approximately 2 population SDs(8) and is a meaningful increase in strength. Thus, relatively small increments in periosteal radius, particularly if they continue to accumulate over long periods of time, conceptually have the potential to considerably alter fracture resistance. Bone size has been linked to fracture risk.(9–11)
PERIOSTEAL BONE FORMATION DURING GROWTH
Periosteal bone apposition is a cardinal feature of skeletal development. Long bones grow wider as they grow taller, and it is commonly recognized that there is wide individual variation in this process (“big-boned” vs. “small-boned”). In fact, after adjustment for height or weight, there is a large range in bone size, indicating that periosteal apposition is affected by a distinct set of determinants.(12) In humans, some of the most obvious are gender (males > females) and race (blacks > whites > Asians).(13–15) Geographical differences in bone size are also marked, even within racial boundaries.(16) These differences in bone size parallel, and could explain, some of the gender, racial, and geographical differences in fracture rates. Disorders of bone size expansion, such as childhood illness at critical periods of development, have been proposed to contribute to the variation in adult bone strength and fracture likelihood.(17) In light of all this, it's clear that a better recognition of the factors that contribute to the maturation of bone size should be sought.
The cell biology that underlies bone growth is complex and not well understood. In long bones, the periosteum arises from mesenchymal tissue adjacent to the epiphyses where chondrocytic cells form a cellular collar outside the growth plate. This perichondrium ossifies in concert with the primary spongiosum of the epiphysis, and osteoblasts subsequently appear on the bone surface to form additional new bone and to result in progressive circumferential enlargement during growth. Gradual periosteal expansion occurs during childhood with a marked acceleration at puberty (in parallel with accelerated growth in bone length).(18) The molecular control of these events is uncertain. In light of the gender differences in bone size, sex steroids have been proposed as important regulators of periosteal apposition. Animal studies support a positive effect of androgens and a negative effect of estrogens on periosteal bone formation rates.(19) A role for insulin-like growth factor 1 (IGF-1) in the regulation of periosteal apposition has long been postulated, especially in concert with sex steroids during puberty.(20) Many other factors are probably involved as well. For instance, mechanical force applied in vivo induces the expression of a variety of genes in the periosteum(21) and a rapid transformation of quiescent periosteal surfaces to those on which bone formation occurs.(22) In fact, it has been suggested that the mechanical loading environment is a primary modulator of periosteal apposition.(23) Also, genetic analyses have implicated a variety of chromosomal regions (and presumably genes) in the control of bone size in humans and mice.(24, 25) In light of their effects on bone formation in other skeletal compartments, other lifestyle and environmental factors (e.g., nutrition, alcohol, and tobacco use)(23, 26) may modulate periosteal bone formation, but their effects have not been well examined. The importance of bone size in the determination of fracture resistance compels additional research to understand periosteal growth physiology and the factors that influence it.
PERIOSTEAL BONE FORMATION IN ADULTHOOD
Biomechanical analyses suggest that changes in bone size (or lack thereof) have important implications for the determination of fracture risk in older adults.(27, 28) Animal studies, from rodents to primates, document the persistence of periosteal bone formation throughout life, albeit at a slower rate than during growth, and there is the strong suggestion that bone size may continue to increase during adulthood. At present, most evaluations of change in bone size in humans are small and cross-sectional and are thus subject to limited power and cohort effects,(29–32) but some longitudinal studies support the increase in bone size with age.(8, 32, 33) As Lazenby has pointed out,(32) there are very few data concerning the rate at which bones may expand, how much individual or population difference exists, whether it occurs at some skeletal sites or some areas of bone more than others, or if it is intermittent or continuous. Clearly, the issue has been inadequately investigated despite its probable importance.
Mechanical events have usually been assumed to underlie the observation that bone size can increase in adults.(34) One attractive model posits that gradual endosteal bone loss with aging leads to cortical thinning and thus more bending stress on the outer surface of bone, in turn leading to the stimulation of periosteal bone apposition as a biomechanical compensation(8, 32) On the other hand, periosteal expansion also seems to occur in early adulthood, at a time when endosteal resorption has not begun, suggesting that events at the periosteum don't only reflect mechanical influences.(32) Moreover, less loaded bones (metacarpal, skull) also experience periosteal expansion in adults. Although likely important, the relative role of mechanical forces in the determination of periosteal responses in unknown.
Some studies suggest that men experience a greater rate of periosteal growth with aging.(1, 32, 35, 36) If there is a sex difference leading to a greater ability to maintain bone strength and fracture resistance in men, it may in part explain the relative protection against age-related fracture risk enjoyed by men.(28, 37) On the other hand, several reports note similar amounts of periosteal expansion in men and women,(32, 38) and it's not clear why there are such discrepant results. Certainly there are methodological differences in measurement techniques, anatomical sites considered, and clinical populations that could all contribute to the variation among reports.
As during growth, other factors may also influence bone size during aging (nutrition, endocrine factors, lifestyle variation, etc.). For instance, Beck et al. reported that inactivity and estrogen use were associated with reduced rates of proximal femoral size expansion in postmenopausal women.(8) Although estrogen therapy has been associated with an overall fracture risk reduction(39) (at least in part because of protection against endosteal and trabecular bone loss), a reduction in periosteal bone formation would exert a negative effect on biomechanical competence. If estrogen suppresses periosteal bone formation, perhaps postmenopausal women who are not treated with estrogen have more periosteal bone formation than do premenopausal women—a potentially positive result of low estrogen levels. The periosteal effects of selective estrogen receptor modulators or nongenotropic estrogens are unclear. Might other factors known to adversely affect osteoblast viability or bone formation (e.g., glucocorticoids, alcohol, renal dysfunction, vitamin D deficiency, etc.) contribute to a failure of periosteal expansion and increased fracture propensity? Conversely, stimulators of periosteal bone formation should offer new opportunities to improve bone strength. For instance, parathyroid hormone therapy (and even mild hyperparathyroidism) may increase bone size and strength through complex effects on bone forming elements on the periosteal surface.(4) If the postulated sex difference in bone size is a result of androgen action, as some animal studies suggest,(40) it lends support to the potential use of androgenic compounds, acting through an effect on bone size, in the prevention of age-related fracture. The emergence of the periosteum as a target for pharmacotherapeutics, for instance with parathyroid hormone or androgenic agents, promises to alter approaches to fracture risk reduction.
PERIOSTEAL BONE RESORPTION
Despite much recent attention to the potential importance of periosteal bone formation, there has been very little consideration of the occurrence or importance of osteoclastic resorption on the surface of bone. Periosteal resorption is a somewhat heretical concept. It is frequently assumed that there is an inexorable expansion of the periosteum through isolated new bone formation, or modeling, and that resorption is rare on the periosteal surface. However, it is unequivocal that periosteal resorption occurs in some situations. For instance, Parfitt et al. have pointed out the drift in bone surfaces that accompanies growth, including the dramatic resorption that must occur on the medial ileal surfaces during pelvic enlargement (Fig. 2).(41) Analogous events occur in other flat bones (mandible, skull, scapulae). Similarly, longitudinal growth of appendicular bones is accompanied by rapid periosteal resorption of the metaphysis (“waisting”) to create the more slender diaphysis (Fig. 3). Essentially, the periosteal radius (and size) of the bone shrinks during that process,(42) and strength is maintained by simultaneous endocortical bone apposition to form a thickened cortex. While there is simply very little information concerning the presence or absence of resorption on most adult periosteal surfaces, Epker and Frost actually described periosteal resorption (and remodeling) in adults on the surface of ribs almost 40 years ago(43) and Balena et al. examined periosteal remodeling on the surface of the ileum in women.(44) In the later studies, the extent of periosteal eroded surface equaled that on the endocortical surface (although there were fewer osteoclasts present on the periosteal surface and in general the remodeling rate was considered much slower than on the endosteal surfaces). It was estimated that the bone formation present on the periosteum all occurred on previously eroded surfaces—in other words, bone formation occurred only as part of remodeling and did not result from modeling. Virtually no other information exists concerning the nature of periosteal remodeling events or their impact on bone health. Nevertheless, there are clear illustrations of the phenomenon. For instance, the alveolar ridge of the mandible can be rapidly lost after tooth loss reduces the mechanical forces on it.(45) In an example of how disease can affect the periosteum, hyperparathyroidism has been classically associated with “subperiosteal” bone resorption. In severe forms, a reduction in mineralized bone size (classically of the phalanges) can be observed radiologically. Whether some or all of this osteoclastic activity originates on the periosteal surface or occurs as a result of exuberant Haversian remodeling (tunneling) within the subperiosteal cortex is unclear. However, the result is a reduction in the effective circumference of bone and arguably its resistance to fracture. To what extent losses of periosteal bone contribute to the increased fracture risk of advanced hyperparathyroidism is unexplored. In summary, the circumference and to some extent the biomechanical strength, of bone should be considered a function of the balance between periosteal bone formation and resorption. However, the rate of periosteal remodeling and the factors that influence it at critical skeletal sites (vertebrae, proximal femur) are unknown.
The changes in bone size during development and aging take on a new complexion if one considers the periosteum a remodeled surface. For instance, it is reasonable to consider that individual, gender, and racial differences in bone size during growth are not only the result of differences in bone formation but also may be determined by differences in the extent of periosteal resorption and the balance between the two processes. If postmenopausal women expand the periosteal circumference to a lesser extent than men with aging, is it because they form less bone on that surface (potentially caused by lower androgen levels) or do they have more osteoclastic activity on the periosteal surface, as they do on other surfaces? Does the secondary hyperparathyroidism that occurs with aging contribute to a higher rate of periosteal resorption and a relatively negative periosteal bone balance? In fact, is it possible that some individuals lose periosteal bone with aging and experience a reduction in bone size with adverse biomechanical consequences? In the unique longitudinal studies reported by Heaney et al.,(46) in which change in bone size in postmenopausal women was assessed over decades from radiographs, there was a net decrease in radial and metacarpal diameter with age. In the studies of Beck et al., there was, on average, an increase in femoral bone size with aging in older women, but the confidence limits of the rate of change clearly overlap 0 (a pattern also observed by Heaney et al. in the femur) and thus are at least consistent with a net loss of bone size in a subset of the population.(8, 46) This conclusion must be considered speculative, at least in part, because methods available for measuring bone size suffer from edge threshold effects and limited resolution.
The presence of periosteal resorption may not only influence the pathophysiology of bone fragility but also the understanding of the effects of commonly used therapies. Obviously, antiresorptive drugs affect endosteal and trabecular bone balances, but they could also modulate events at the periosteal surface. A reduction in surface-based osteoclastic activity would be expected to have beneficial effects on bone size, but a simultaneous reduction in overall remodeling might prevent gains in bone formation. The reduction in periosteal apposition noted by Beck et al. in older women receiving estrogen replacement therapy is consistent with this model (as it is with a simple reduction in bone formation by estrogen). The effects of antiresorptive drugs on periosteal osteoclastic activity should be evaluated, and the actions of other anti-osteoporosis therapies should be considered in this context.
THE FEMORAL NECK
The femoral neck is of unique importance in the field of osteoporosis. It has long been axiomatic that the periosteum of the femoral neck is disrupted by the attachment of the capsule of the hip joint and that periosteal bone apposition doesn't occur in that area.(47) That assertion is more and more difficult to support. Not only is it incompatible with the obvious expansion in femoral neck periosteal circumference that happens during growth but also with the now numerous reports of growth in femoral neck size with aging.(8, 46) Moreover, we recently documented classical periosteal bone formation in the femoral neck in a series of rhesus macaques by using histomorphometric analyses after double tetracycline labeling.(48) Bone formation was routinely observed, and the tissue level bone formation rate was sufficient to add substantively to femoral size over time. The histological presence of bone formation, and clinical evidence of increases in bone size, at the femoral neck provides a potentially important target for therapies that stimulate periosteal bone growth and femoral neck strength.
These studies in primates also revealed strong evidence of periosteal bone resorption at the femoral neck, with surface-based resorption pits containing typical multinucleated cells that stained positively for acid phosphatase (Fig. 4). This osteoclastic activity was apparently not the result of the emergence of intracortical tunneling at the bone surface. Osteoclastic activity, and the modulation of bone size through periosteal bone resorption, has the potential to influence femoral neck fracture risk. An anti-osteoclastic effect could also contribute to the proximal femoral antifracture effect of antiresorptive therapy.
Although the presence of the capsule apparently does not abrogate the presence of periosteal bone remodeling at the femoral neck, it might present other difficulties in the assessment of periosteal function in older individuals in vivo. Shea et al.(49) recently described calcification that occurs at the bone-capsular interface at the femoral neck in the elderly. This mineralization was capsular, yet quite dense. It is unknown how, or if, this extra skeletal mineral affects femoral neck fracture propensity. While the demarcation between bone and capsular mineral was clear in the ex vivo preparations used by Shea et al.,(49) the distinction between increases in bone size through periosteal formation and an increase in apparent femoral neck diameter through capsular mineralization adjacent to bone would be very difficult with current in vivo measurement techniques. Thus, a portion of the age-related increase in femoral neck size described by radiographic methods (including DXA) in the elderly could be artifactual. Obviously, this important issue should be further resolved, and methods must be sought to allow accurate measures of biomechanically effective femoral neck size in vivo.
In conclusion, bone size is a critical element of bone strength and possibly of fracture resistance. Cellular events at the periosteum modulate bone size throughout an individual's lifespan. It is probable that change in bone size is the result of the balance between periosteal osteoblastic and osteoclastic activities (as at other bone surfaces). Currently too little is known of the control of periosteal osteoblastic activity or of the clinical importance of variations in periosteal bone formation. Moreover, the nature and impact of periosteal bone resorption is virtually unexplored. Under some circumstances, bone resorption conceivably could exceed formation with the result being a fall in bone circumference and strength. By adding to bone mass at the perimeter, periosteal bone formation has the potential to increase bone strength, and it should be an attractive target for the development of new therapies. Whereas antiresorptive agents could prevent periosteal bone loss, they could also prevent periosteal formation by virtue of their reduction in overall bone remodeling. Periosteal remodeling events apparently occur at the femoral neck, a site of particular clinical relevance, but the evaluation of changes in femoral neck size may be obscured by calcification of adjacent capsular structures. Periosteal bone biology is insufficiently understood and should be the focus of concerted research effort. The development of noninvasive techniques to accurately assess changes induced by periosteal events is particularly important.
Michael Bliziotes, Michael Parfitt, Robert Klein, and Gary Beaupre provided great help in the review of the manuscript. Glenda Evans and Russ Turner provided the photomicrograph of the periosteum. This work was supported by NIAMS AR45647 and VA Merit Review funding.