Pathophysiology of Osteoporosis
ROBERT P. HEANEY, MD, FACP, FAIN
ABSTRACT: As with many chronic diseases
that express themselves late in life, osteoporo-
sis is distinctly multifactorial both in etiology
and in pathophysiology. Osteoporotic frac-
tures occur because of a combination of injury
and intrinsic bony fragility. The injury comes
most often from a combination of falls, poor
postural reflexes that fail to protect bony parts
from impact, and reduced soft tissue padding
over bony prominences. The bony fragility it-
self is a composite of geometry, low mass den-
sity, severance of microarchitectural connec-
tions in trabecular structures, and accumu-
lated fatigue damage. Reduced bone mass, in
turn, is caused by varying combinations of go-
nadal hormone deficiency, inadequate intakes
of calcium and vitamin D, decreased physical
activity, comorbidity, and the effects of drugs
used to treat various unrelated medical condi-
tions. Finally, the often poor outcome from hip
fracture in the elderly is partly caused by asso-
ciated protein-calorie malnutrition. An ade-
quate preventive program for osteoporotic
fracture must address as many of these factors
as possible, ie, it must be as multifaceted as
the disease is multifactorial. KEY INDEXING
TERMS: Osteoporosis; Fracture; Pathophysiol-
ogy; Bone fragility. [Am J Med Sci 1996;312(6):
251-256.]
O steoporosis is defined elsewhere in this sympo-
sium as a condition of skeletal fragility charac-
terized by reduced bone mass and microarchitect-
ural deterioration of bone tissue, with a consequent
increase in risk of fracture . Low bone mass is, there-
fore , visualized as a risk factor for fracture. The term
osteoporosis is also used to designate a bone mass
value more than 2.5 standard deviations below the
young adult mean. The term osteoporosis, therefore,
designates both one of the risk factors for fragility
From Creighton University, Omaha, Nebraska.
Submitted June 25, 1996; accepted in revised form July 16,
1996.
Correspondence: Robert P. Heaney, MD, Creighton University,
2500 California Plaza, Omaha, NE 68178.
THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES
and the condition offragility itself. It is, in this way,
analogous to hypertension, in which the term desig-
nates both the blood pressure level and the condition
of increased risk of untoward vascular events.
It is useful to see both sides of this ambivalent
definition because the pathophysiology of osteoporo-
sis ultimately involves the development not just of
low bone mass, but of both the other skeletal compo-
nents of fragility and the extraskeletal factors that
lead to fracture . The interplay of the multiplicity
of these pathogenetic mechanisms is illustrated in
Figure 1, which situates the more important groups
of causal factors relative to the undesired outcome:
fracture. A consequence of this multifactorial patho-
physiology is the realization that a comprehensive
approach to prevention will have to be multifaceted.
Before beginning a systematic survey of pathoge-
netic factors, it will be useful to examine the relative
importance of mass and nonmass factors . It is often
said that mass is the most important determinant
of fragility, but this is probably not correct. Simula-
tions suggest that mass explains less than half of
the observed fracture risk, 1 and clinical studies have
demonstrated that expressed fragility (for example,
in the form of a prior vertebral fracture) is a stronger
predictor of future fracture than is low bone mass by
itself.2 The reasons for this nonmass-related fragility
will be discussed later, but an example will illustrate
the aggregate importance. Figure 2 presents age-
specific fracture risk gradients for various values of
forearm bone mineral density (BMD). Three features
stand out: 1) risk rises with declining bone mass, as
expected; 2) risk also rises with age, even when bone
mass is held constant; and 3) the gradients are
steeper with advancing age, that is, a given defi-
ciency of bone makes a bigger difference in an older
person than in a younger one. As Figure 2 indicates,
the age effect is actually larger than the bone mass
effect.
Age, of course, is only a surrogate for a multiplicity
of factors. These include the following: falling more
often, slow postural reflexes that cause a person to
fall in such a way as to strike vulnerable bony parts,
loss of soft tissue protection over bony prominences,
accumulation of unremodeled fatigue damage in the
bony material, loss of critical trabecular connectiv-
ity, and possibly other factors predisposing to frac-
ture.
251
Pathophysiology
Architecture
Falls and
/ometry
St~~~~h -
Bone mass ..,.._ Nutrition
~erclse
'
Padding Bone material
strength
Postural
reflexes
Figure 1. Interplay of the major groups of pathogenetic factors
leading to osteoporotic fractures. (Reprinted with permission from
Robert P . Heaney, 1991.)
Six of the 11 articles in this symposium deal with
treatment of low bone mass or prevention of bone
loss. Bone mass, as we have seen, is but one of the
factors contributing to osteoporosis. Accordingly, al-
though my primary concern is with pathophysiology,
I shall also briefly indicate preventive approaches
for certain nonmass pathogenetic factors.
Extraskeletal Factors
Falls. Almost all osteoporotic fractures result from
injury-the application of more force to the bone
than it can sustain given its fragility. Such injury is
usually incurred in a fall. Although only 2% to 5%
of falls in the elderly result in fracture, both fre-
quency of falling and mechanics of falling change
with advancing age. These alterations are caused by
deterioration in gait and postural stability, de-
creased muscle strength, malnutrition, and comor-
bidity and medications. (Drugs associated with pos-
tural hypotension-and psychotropic drugs gener-
ally-have been especially implicated.) Probably
even more important than falls is the way the elderly
person falls. Falls to the side, striking the lateral
surface of the hip, will often cause a hip fracture
even in healthy younger individuals. 4 But younger
individuals generally shift their bodies to land on
softer parts or break the force of the fall with their
arms. Elderly individuals have slower response
times and more often fall to the side, so they more
often sustain force directly on the hip.
Energy Absorption. The foregoing discussion makes
clear why energy absorption is of critical importance
in determining whether fracture occurs when force
is suddenly applied to a bony structure. Both soft
tissue padding over the hip and padding of the sur-
face on which the fall occurs distribute the energy
of impact over a larger area of bone, and thereby
lessen point loads on the bone. That is partly why
hip fracture risk varies inversely with body weight
(even after adjusting for height). It also explains the
protective benefits achieved by artificial hip pads in-
corporated into undergarments for the elderly.5
252
Clearly, for those elderly individuals who already
Hormones
have fragile skeletons, attention to this pathogenetic
component of the osteoporotic fracture problem will
/
be needed to reduce numbers of fractures.
Bone Geometry. Three examples of the importance
of bone geometry have been elucidated. One is hip
axis length, the distance from the lateral surface of
the trochanter to the inner surface of the pelvis,
Lifestyle
along the axis of the femoral neck. For any given
bone density, short hip axis length results in an ar-
chitecturally stronger structure. 6 Available evidence
indicates that, holding bone mass constant, hip frac-
ture risk approximately doubles for every increase
of 1 standard deviation in hip axis length. This geo-
metric difference is probably the major reason why,
after adjusting to the same bone density values,
Asian people have about half the hip fracture rate
of white people.
Likewise, large vertebral body end-plate areas re-
sult in lower spine pressure values under bending
and compression for individuals of the same body
size. 7 Those with small vertebral bodies are there-
fore more likely to sustain wedge or compression
fractures. Finally, the force sustained by a bone
when struck in falling is a function of body height.
Other factors being equal, a tall person striking the
lateral surface of the hip in falling is more likely to
sustain a fracture than is a short person.
Such geometric factors contribute to individual
fracture risk and explain a substantial portion of the
population level variance in fracture rate. In each
situation, however, the ultimate pathogenesis of the
fracture is the fall and the force sustained by the
bone on impact.
160
80+
140
::.::~ 120
- ......
11'1"'
llii::~ 100 ] 75-79
UJc.
~8
f-0
80 70-74
u- 60 65-69
~~
u..- 40
60-64
20
0
>1.0 0.90-.99 0.80-.89 0.70-.79 0.60-.69 <0.60
BONE MASS (g/cm)
Figure 2. Age-specific fracture risk gradients for forearm bone
mineral density, redrawn from Hui et al. 3 Note that bone density
declines to the right. For any given bone density (eg, the dashed,
vertical line at a bone mass density of 0.7) fracture risk rises with
age. Note, also, that the gradient of risk is steeper with advancing
age. (Reprinted with permission from Robert P . Heaney, 1991.)
December 1996 Volume 312 Number 6

Skeletal Factors
Decreased Bone Mass. Bone mass and density are
the most obvious of the bone strength factors, and
certainly the most well studied. For a given material
density and architectural configuration, bone
strength rises as approximately the square of struc-
tural density. Therefore, any factor decreasing bone
mass will, of necessity, weaken bone: a 30% reduc-
tion, such as is common in osteoporosis, will decrease
strength by about 50%. Extensive epidemiologic data
indicate that fracture risk increases 2 to 3 time for
every drop of 1 standard deviation in bone mass at
any given site. (As noted in Figure 2, this effect var-
ies in magnitude at different ages and, therefore, the
cited gradient is only an approximation.)
Low bone mass (or density) is itself multifactorial
in etiology involving: 1) genetic predisposition; 2)
failure to achieve genetic potential during growth;
3) excessive leanness; 4) disuse; 5) gonadal hormone
deficiency; 6) inadequate calcium and vitamin Din-
take; 7) lifestyle and medical factors (eg, smoking,
alcohol abuse, corticosteroids); 8) remodeling errors;
9) medical diseases such as malabsorption syn-
dromes, biliary cirrhosis, and posttransplant dis-
ease; and 10) miscellaneous factors. Several of these
contributing causes are themselves related. For ex-
ample, excessive leanness is frequently associated
with both ovarian dysfunction and low calcium in-
take. Nevertheless, each of these factors appears to
make a contribution on its own. Some are covered
in greater detail in other articles.
Before describing these effects, it will be useful to
review the process by which bone tissue continually
renews itself. Remodeling consists of osteoclastic re-
sorption of old, dead, damaged, or underused bone,
followed by osteoblastic replacement with fresh, new
bone. It serves not only to replace damaged regions,
but to adjust bone shape and density to current pat-
terns of loading. All changes in bone mass, whether
they are losses that weaken bone or gains produced
by treatment regimens, occur through adjustments
in the balance between the resorption and formation
components of remodeling.
Bony renewal involves not just replacing tissue,
but seeking always to restore optimal mass density.
This process is controlled by a classic negative feed-
back loop in which some cellular apparatus, most
likely resident in the osteocytes, senses the degree
of bending (technically called "strain") produced by
routine loading of a part. During typical use, this
strain amounts to a deformation of0.1% to 0.15% in
any given dimension. The signals from the sensing
apparatus in bone adjust the local balance between
bone formation and bone resorption to increase or
decrease local bone mass (and hence stiffness) and,
thereby, to achieve the desired reference level of
bending.
Failure to reach the genetic potential for bone den-
THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES
Heaney
sity is most often a result of inadequate exercise and
insufficient calcium intake (alone or in combination).
In other words, suboptimal loading results in subop-
timal stimulation of bone deposition. The result is
that bone mass falls short of the genetic limit. Insuf-
ficient calcium intake during growth operates in a
similar way: it limits how much bone can be amassed
within the envelope produced by growth and exercise
(for additional information, see the article by Reid
on calcium, vitamin D, and exercise).
Disuse is a well established cause of reduced bone
mass. Bedrest and paralytic states regularly result
in bone loss. 8 The mechanism is the normal opera-
tion of the feedback loop controlling density: un-
derused bone is thinned until its bending on use
matches the strain setpoint. With age, most individ-
uals decrease the extent and intensity of physical
activity. If weight stays the same, bone mass falls
along with exercise level. Even modest programs of
physical activity in middle-aged and elderly individ-
uals can halt or reverse this loss, 9 •10 demonstrating
that it is not inexorable.
The foregoing discussion of exercise presumes
maintenance of normal body weight. However, if de-
creased physical activity is coupled with weight gain,
the system operates differently. Weight is the single
largest determinant of bone density in mature
adults, explaining approximately half the population
level variance. The reasons behind this fact are not
known. Almost certainly the effect involves more
than just increased mechanical loading from car-
rying the extra pounds, as lean mass is better corre-
lated with bone density than is either fat mass or
total body weight. Also, weight effects are evident
in the hand and forearm bones, 11 which, for bipedal
humans, are not weight bearing. The bone structure
of overweight individuals behaves as if it had a lower
setpoint for the reference level of bending under
load. The importance of body weight is shown by the
fact that fracture risk in women who are obese is
about one-third that of women of normal weight.
This is partly because of the increased bone density
and partly because of the extra padding. In the oppo-
site direction, excessively lean women have low bone
mass, excessive menopausal loss, and high fracture
risk.
By contrast, gonadal hormone loss (occurring
abruptly at menopause, and gradually in women
with anorexia nervosa or athletic amenorrhea, and
in aging men), appears to raise the strain setpoint
of the bone mass regulatory system. In other words,
after gonadal hormone loss, prior bone mass is
sensed as being excessive. The relatively sharp drop
in hormone levels across menopause results in re-
modeling-induced loss of bone amounting to about
15% over 5 or more years. (The quantity varies from
region to region; 15% loss is what is typically found
in the spine.) This loss is not specifically related to
253
Pathophysiology
exercise or calcium intake and cannot, therefore, be
appreciably altered by diet or exercise regimens.
However, gonadal deficiency bone loss would stop at
the 15% quantum if diet and exercise were adequate.
But often they are not. Calcium requirement rises
with age and food intake falls with age. As a result,
calcium deficiency increases with age. This factor in
combination with decreased exercise causes bone
loss to accelerate with advancing age. Women in
their 80s typically lose. bone at about the same rate
as in the early postmenopause. 12 In contrast to the
early postmenopausal years, adequate calcium and
vitamin D will greatly reduce bone loss at this
age.12,13
Smoking and alcohol abuse affect bone in uncer-
tain ways. Alcohol in moderation is associated with
decreased fracture risk, but excess alcohol is directly
toxic to osteoblasts and is often associated with re-
duced bone mass, with decreased calcium and vita-
min D intakes, and with excessive urinary loss of
calcium.
Remodeling errors are of two main types. The first
is excavation of over-large Haversian spaces in corti-
cal bone. Radial infilling is regulated by signals from
the outermost osteocytes, and generally advances in-
ward no more than 90 11M from the outer rim. Hence,
the excavation of large external diameters, which
may simply occur randomly, leads to large central
Haversian canals. Inevitably, these accumulate with
age, resulting in increased cortical porosity. In a sim-
ilar way, osteoclast penetration of trabecular plates,
or severing of trabecular beams, removes the scaf-
folding needed for osteoblastic replacement of re-
sorbed bone. In both ways random remodeling errors
(in addition to the other factors cited) tend to reduce
both cancellous and cortical bone density and struc-
tural integrity. These remodeling errors represent a
kind of entropy.
It is sometimes said that high remodeling rates
accelerate bone loss, and there is empirical evidence
of correlation between remodeling levels and rates
of bone loss in postmenopausal women. 14 It has even
been suggested that high levels of bone biochemical
markers of remodeling can be used to diagnose osteo-
porosis or to predict fracture risk. Such claims would
seem to go well beyond the available data. Moreover,
it is not clear from studies to date whether the con-
nection between the remodeling and the bone loss is
causal. It is often presumed that remodeling is al-
ways associated with some bone loss and that, there-
fore, the higher the remodeling rate, the greater the
loss. But that, plainly, is not always true. Investiga-
tional use ofhuman parathyroid hormone in osteopo-
rosis has shown that, as long as vitamin D status is
adequate, spine bone mass increases impressively
despite greatly increased bone remodeling activity. 15
Among other things, a high remodeling rate is a
marker for inadequate calcium and vitamin D sta-
254
--
Cl
1100
~ 1000 ········-----------------------------·-··r
0...1 Estrogen deficiency 15%
5 900
t
~-
c>- Disuse/entrop; ···fj%
~--;Q~~ r
800
0
co
...1 ~ deficiency 16%
~
0
1-
700
600
~ L
0 5 10 15 20
TIME MENOPAUSAL (yrs)
Figure 3. Diagrammatic partition of age-related bone loss in a
typical postmenopausal woman. The estrogen-deficiency loss as-
sumes no postmenopausal hormone replacement. The magnitude
of the calcium-deficiency loss is typical; it may be nonexistent in
some individuals, or substantially greater in others. (Reprinted
with permission from Robert P . Heaney, 1991.)
tus, 13·16 and it may be that, in the samples studied
to date, varying degrees of calcium deficiency were
the underlying cause of both the bone loss and the
high remodeling. Remodeling is typically elevated
with advancing age but can be restored to normal
young adult values with adequate calcium.13
An illustrative partition of postmenopausal loss in
women is shown in Figure 3, which superimposes
entropic and nutritional losses on estrogen defi-
ciency loss. The estrogen- and calcium-related losses
are preventable. Note that, in the early menopausal
years, estrogen deficiency dominates. Note also that
later, if calcium and vitamin D intakes are not ade-
quate, nutritional loss can account, in the last analy-
sis, for more bone loss than did menopausal estrogen
loss. The entropic loss is not preventable, at least
intrinsically. However, because it is a function of the
ongoing remodeling rate, it will likely be more severe
in individuals with high remodeling activity.
One of the least understood, and yet most fascinat-
ing, aspects of the bone mass regulatory system is
the strain setpoint. Setpoints exist in all physiologic
regulatory systems. Familiar examples include regu-
lation of blood pressure and respiratory rate. In no
system, however, is the molecular basis of the set-
point known, and bone is no exception. Although
there is remarkable constancy of strain under rou-
tine load across all bones in all mammalian skele-
tons studied to date (ie, bending on the order of0.1-
0.15%), there still is variability around that level.
The setpoint is probably the ultimate basis for ge-
netic variations in bone density. (For example, black
people have, presumably, a lower setpoint than
white people.) I have already suggested that estro-
December 1996 Volume 312 Number 6

gen alters the setpoint, but it is not known whether
other hormones or drugs might have similar effects,
nor whether the setpoint itself might be affected in
(and thus be the ultimate pathophysiologic basis for)
some forms of osteoporosis. (Such an alteration
seems to be present, for example, in osteogenesis
imperfecta. Scientific attention in that disorder has
focused mainly on the collagen defects and their ge-
netic basis. However, these defects explain neither
the low bone mass nor the fact that the exuberant
fracture calluses produced in this disease are remod-
eled down to the deficient bone mass that is a hall-
mark of osteogenesis imperfecta.) Clearly, the ideal
therapy for the low bone mass component of osteopo-
rosis would be an agent that lowered the strain set-
point. This would cause the skeleton's own remodel-
ing apparatus to increase bone density, in a sense
naturally.
Defects in Bone Quality. I noted above that the
strength of a bony structure varies roughly as the
square of density, assuming maintenance of archi-
tectural integrity and constant intrinsic strength of
the bony material. Here is where fatigue damage,
trabecular connectedness, and other architectural
factors enter into consideration. Fatigue damage
weakens the material, changing its intrinsic
strength, and trabecular disconnection weakens the
structure.
Fatigue Damage. Fatigue damage consists of ul-
tramicroscopic rents in the basic bony material, re-
sulting from the inevitable bending that occurs when
a structural member is loaded. Fatigue damage is
the principal cause of failure in all mechanical engi-
neering structures, and designers usually compen-
sate by increasing the massiveness of a structure so
that it bends little and therefore has a greater mar-
gin of safety. Bones, by contrast, are designed both
to be more resilient and to take care of fatigue dam-
age by repairing it. Therefore, they have a relatively
narrow margin of safety (about 2 X over peak physio-
logic loads).
Fatigue-damaged bone, being inherently weaker,
deforms more under loading and thereby produces
a signal greater than the reference level of bending.
The remodeling apparatus detects and replaces the
damaged regions. Fractures related to fatigue dam-
age occur whenever the damage occurs faster than
remodeling can repair it, or whenever the remodel-
ing apparatus is defective. March fractures and the
fractures of radiation necrosis are well-recognized
examples of fatigue fractures caused by these two
mechanisms, respectively.
Fatigue damage definitely occurs in normal bone
under ordinary use, 17 although it is less certain what
its precise role is in predisposing to osteoporotic frac-
ture. It would, on the other hand, be surprising if
bone were the only structural material known in
which fatigue damage were not an important factor
THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES
Heaney
in failure. In addition, there is suggestive evidence
in certain osteoporotic fractures (notably hip), that
remodeling repair may be specifically defective at
the site that ultimately fractures. 18 Why remodeling
surveillance or effectiveness might fail locally is not
known. Nevertheless, it is clear that such failure
would lead to accumulation of fatigue damage and,
therefore, to local weakening of bone.
Trabecular Disconnection. Bone structures loaded
vertically, such as the vertebral bodies and femoral
and tibial metaphyses, derive a substantial portion
of their structural strength from a system of hori-
zontal trabeculae, which brace the load-bearing ver-
tical trabeculae and limit lateral bowing and conse-
quent snapping under vertical loading. Severance of
such trabecular connections is known to occur pref-
erentially in postmenopausal women 19 and is consid-
ered to be a major reason for the large female:male
preponderance of vertebral osteoporosis. Long, un-
supported vertical trabeculae clearly make a struc-
ture flimsy. This is shown by the extraordinarily
high prevalence of microscopic trabecular fracture
calluses in vertebral bodies from women examined
at autopsy-typically 200 to 450 healing or healed
fractures per vertebral body. Whereas at any given
time many of these will be healed well enough to
be structurally competent, others will be fresh and
structurally weak. Such trabecular fractures are
asymptomatic and their accumulation silently weak-
ens the cancellous structure of the vertebral body.
The ability to predict future fracture on the basis of
prior vertebral fractures 2 is probably due in part to
the presence of such otherwise undetected trabecu-
lar defects. In other words, that may be why prior
fracture seems to predict future fracture even when
bone density is relatively high. The reason for prefer-
ential osteoclastic severance ofhorizontal trabecular
is not known. It is sometimes attributed to over-
aggressive osteoclastic resorption, but that answer
seems more descriptive than explanatory.
Outcome After Fracture
In addition to the disability and pain of osteopo-
rotic fractures, at least two typical fractures (spine
and hip) are known to be associated with excessive
mortality. 20 Also, hip fracture carries with it a sub-
stantial risk ofloss of independence. A growing body
of work makes plain that these bad outcomes are
not inevitable, and that they can be at least partially
prevented. 21 - 23 Hip fractures are concentrated in the
malnourished elderly. The best predictor of mortal-
ity in the group as a whole is low serum albumin on
admission. 22 Although hospital meals are nutri-
tionally adequate, they are seldom consumed by the
elderly individual hospitalized with hip fracture and
recovering from surgical repair. As a result, they do
nothing to alleviate the malnutrition, and the pa-
tients' situation worsens while under care. Aggres-
255
Pathophysiology
sive nutritional supplementation (particularly with
protein) dramatically improves hip fracture out-
comes-fewer deaths, shorter hospital stays, less in-
stitutionalization, and better protection of (or even
some restoration of) bone mass. 23 •24 Attention to the
nutritional problems of hip fracture patients is as
essential as is orthopedic repair of the broken bone.
References
1. Heaney RP. Is there a role for bone quality in fragility frac-
tures? Calcif Tissue Int. 1993; 53:S3-S5.
2. Ross PD, Davis JW, Epstein RS, Wasnich RD. Preex-
isting fractures and bone mass predict vertebral fracture inci-
dence in women. Ann Intern Med. 1991; 114:919-23.
3. Hui SL, Slemenda CW, Johnston CC Jr. Age and bone
mass as predictors of fracture in a prospective study. J Clin
Invest. 1988;81:1804-9.
4. Robinovitch SN, Hayes WC, McMahon TA. Prediction of
femoral impact forces in falls on the hip. J Biomech Eng.
1991; 113:366-74.
5. Lauritzen JB, Petersen MM, Lund B. Effect of external
hip protectors on hip fractures. Lancet. 1993;341:11-3.
6. Faulkner KG, Cummings SR, Black D, Palermo L,
Gliier C-C, Genant HK. Simple measurement of femoral
geometry predicts hip fracture: the study of osteoporotic frac-
tures. J Bone Miner Res. 1993;8:1211-7.
7. Gilsanz V, Loro ML, Roe TF, Sayre J, Gilsanz R, Schulz
EE. Vertebral size in elderly women with osteoporosis. J Clin
Invest. 1995;95:2332-7.
8. LeBlanc AD, Schneider VS, Evans HJ, Engelbretson
DA, Krebs JM. Bone mineral loss and recovery after 17
weeks of bed rest. J Bone Miner Res. 1990;5:843-50.
9. Dalsky GP, Stocke KS, Ehsani AA, Slatopolsky E, Lee
WC, Birge SJ Jr. Weight-bearing exercise training and lum-
bar bone mineral content in postmenopausal women. Ann
Intern Med. 1988; 108:824-8.
10. Smith EL, Gilligan C, McAdam M, Ensign CP, Smith
PE. Deterring bone loss by exercise intervention in premeno-
pausal and postmenopausal women. Calcif Tissue Int.
1989;44:312-21.
11. Heaney RP, Barger-Lux MJ, Davies KM, Ryan RA,
Johnson ML, Gong G. Bone dimensional change with age:
Interactions of genetic, hormonal, and body size variables.
(unpublished) 1996.
12. Chapuy MC, Arlot ME, Duboeuf F, Brun J, Crouzet B,
Arnaud S, et al. Vitamin D 3 and calcium to prevent hip
256
fractures in elderly women. N Engl J Med. 1992;327:1637-
42.
13. McKane WR, Khosla S, O'Fallon WM, Robins SP, Bur-
ritt MF, Riggs BL. A high calcium intake reverses the sec-
ondary hyperparathyroidism and increased bone resorption
of elderly women. J Clin Endocrinol Metab. 1996;81:1699-
1703.
14. Christiansen C, Riis BJ, Rodbro P. Prediction of rapid
bone loss in postmenopausal women. Lancet. 1987; 1:1105-
8.
15. Slovik DM, Rosenthal DI, Doppelt SH, Potts JT Jr, Daly
MA, Campbell JA, et al. Restoration of spinal bone in osteo-
porotic men by treatment with human parathyroid hormone
(1-34) and 1,25-dihydroxyvitamin D. J Bone Miner Res.
1986; 1:377-81.
16. Phang JM, Berman M, Finerman GA, Neer RM, Rosen-
berg LE, Hahn TJ. Dietary perturbation of calcium metabo-
lism in normal man: Compartmental analysis. J Clin Invest.
1969; 48:67-77.
17. Burr DB, Martin RB, Schaffler MB, Radin EL. Bone re-
modeling in response to in vivo fatigue microdamage. J Bio-
mech. 1985; 18:189-200.
18. Eventov I, Frisch B, Cohen Z, Hammel I. Osteopenia,
hematopoiesis and osseous remodeling in iliac crest and fem-
oral neck biopsies: A prospective study of 102 cases offemoral
neck fractures . Bone. 1991; 12:1-6.
19. Kleerekoper M, Villanueva AR, Stanciu J, Sudhaker R,
Parfitt AM. The role of three-dimensional trabecular micro-
structure in the pathogenesis of vertebral compression frac-
tures. CalcifTissue Int. 1985;37:594-7.
20. Cooper C, Atkinson EJ, Jacobsen SJ, O'Fallon WM,
Melton LJ III. Population-based study of survival after os-
teoporotic fractures. Am J Epidemiol. 1993;137:1001-5.
21. Bastow MD, Rawlings J, Allison SP. Benefits of supple-
mentary tube feeding after fractured neck of femur. Br Med
J. 1983;287:1589-92.
22. Rico H, Revilla M, Villa LF, Hernandez ER, Fernandez
JP. Crush fracture syndrome in senile osteoporosis: A nutri-
tional consequence. J Bone Miner Res. 1992;7:317-9.
23. Delmi M, Rapin C-H, Bengoa J-M, Delmas PD, Vasey
H, Bonjour J-P. Dietary supplementation in elderly pa-
tients with fractured neck of the femur. Lancet. 1990;
335:1013-6.
24. Rizzoli R, Schuerch MA, Slosman D, Bonjour J-P. Pro-
tein supplements increase serum IGF-1 and prevent proxi-
mal femur bone loss in elderly with a recent hip fracture . In:
Proceedings of 1996 World Congress on Osteoporosis, Am-
sterdam. May 18-23, 1996; Amsterdam: Elsevier Science
Publishers; 1996 (In press).
December 1996 Volume 312 Number 6