THE ANATOMICAL RECORD 219:l-9 (1987)

Bone “Mass” and the “Mechanostat”:

A Proposal
H.M. FROST
Southern Colorado Clinic, Pueblo, CO 81004; Department of Anatomy, Purdue University,
West Lafayette, IN 47907

ABSTRACT The observed fit of bone mass to a healthy animal’s typical mechan-
ical usage indicates some mechanism or mechanisms monitor that usage and control
the three longitudinal growth, bone modeling, and BMU-based remodeling activities
that directly determine bone mass. That mechanism could be named a mechanostat.
Accumulated evidence suggests it includes the bone itself, plus mechanisms that
transform its mechanical usage into appropriate signals, plus other mechanisms
that detect those signals and then direct the above three biologic activities. In viuo
studies have shown that bone strains in or above the 1500-3000 microstrain range
cause bone modelling to increase cortical bone mass, while strains below the 100-
300 microstrain range release BMU-based remodeling which then removes existing
cortical-endosteal and trabecular bone. That arrangement provides a dual system in
which bone modeling would adapt bone mass to gross overloading, while BMU-based
remodeling would adapt bone mass to gross underloading, and the above strain
ranges would be the approximate “setpoints” of those responses.
The anatomical distribution of those mechanical usage effects are well known. If
circulating agents or disease changed the effective setpoints of those responses their
bone mass effects should copy the anatomical distribution of the mechanical usage
effects. That seems to be the case for many agents and diseases, and several exam-
ples are discussed, including postmenopausal osteoporosis, fluoride effects, bone loss
in orbit, and osteogenesis imperfecta.
The mechanostat proposal is a seminal idea which fits diverse evidence but it
* -

requires critique and experimental study

Manuscripts circulated privately by the author in July could call it a mechanostat. As a relation that idea could
and October 1986’ proposed a novel mechanism for con- be written thus:
trolling bone “mass” that could be involved in the patho-
genesis of osteopenias and osteoporoses and that could MU bone -+ mechanostat -+ bone mass effect
tL--,-,,-,-------,--,,-1
-+

mediate the bone mass effects of some circulating agents, Relation (1)
genetics, drugs, and diseases. The mechanism can ex-
plain some otherwise enigmatic bone-biologic and clini- The dashed arrow shows the feedback loop between
cal phenomena and it fits abundant evidence from many the biologic mechanisms and bone architecture and mass
skeletal science subspecialties. The text will present the originally proposed by the author and others (p. 46,
concept, cite behavioral and anatomical features that Frost, 1964)and now accepted by most biomechanicians
underlie it, and offer a few examples of how it might (Cowin, 1986; Currey, 1984; Rubin, 1984). How such a
work in man and other mammals. The text focuses on mechanostat may be constructed and work are described
the narrow matter of bone mass in the sense of how next. It should contain three tissue-level biological
much bone exists after subtracting its marrow and vas- mechanisms: growth, modeling, and remodeling. Those
cular spaces. To adhere to that theme the text must mechanisms can indeed affect bone mass and do respond
ignore bone quality, architecture, and many other to mechanical usage in special ways, apparently under
matters.
THE MECHANOSTAT: ITS PARTS Received October 2, 1986; accepted February 24,1987.
Address reprint requests to H.M. Frost, M.D., Southern Colorado
It is rarely remarked but true that bone mass fits the Clinic, 2002 Lake Ave., Pueblo, CO 81004.
typical mechanical usage of healthy skeletons in a spe- ‘The original informal manuscripts were distributed to 71 partici-
cial sense: the mass can be overadequate but never inad- pants in the 1986 Hard Tissue Workshop sponsored by the University
equate. That implies some mechanism(s) monitors the of Utah, organized by Prof. W.S.S. Jee, and held in August at Sun
mechanical usage of bone and makes its biologic mech- Valley, ID, and also to 105participants in the 1986 Gordon Conference
anisms correct serious misfits between that mass and its on Biomechanics and Biomechanical Engineering held in August at
Andover, NH.
mechanical usage (MU). That mechanism would behave 2A “minimodeling” mechanism can occur on trabecular surfaces
like a home thermostat in that it would go “ON” in but it is not a major concern in the control of bone mass, even though
response to an error and “OFF” in its absence, so one it does affect trabecular orientation and architecture.
0 1987 ALAN R. LISS, INC.
2 H.M. FROST

Fig. 1. Modeling and modeling drifts. A,B: As a n infant’s long bone the middle of the shaft to the right. The three-quarter view in the
(solid line) grows to the length and overall size of the adolescent right-hand drawing (C) diagrams a resorption drift (Ft) on the left
(dashed line), highly patterned motions in tissue space of its periosteal surface of diaphyseal segment and a formation drift (F) on its right
surfaces (shown) and cortical-endosteal surfaces (not shown) maintain surface. The cross section to the left of C shows the resulting motions
the bone’s shape. Those surface motions are called modeling drifts. or drifting of the periosteal and cortical-endosteal surface (reproduced
Similar kinds of drifts but in a different pattern can correct (dashed by permission, H.M Frost. “Osteogenesis Imperfecta. The Setpoint
line) a child’s fracture that healed with a deformity (solid line), as in Proposal.” Clin Orth. 216:200, 1987).
the second drawing. The cross sections below show how the drifts move
the control and coordination of two other mechanisms, existing ones. In man and other large mammals model-
the MES mechanisms defined later, which mediate the ing drifts become ineffective around and after skeletal
message traffic between a bone’s mechanical usage and maturity, so they primarily control bone accumulation
its cells. Here mechanical usage includes all physical during growth (Enlow, 1963; Frost, 1985).
loads and motions imposed on the bony skeleton by a In the new bone lexicon, bone remodeling turns bone
subject’s normal weekly physical activities. over in small packets named BMU (basic multicellular
units) that couple a n initial resorption to a final forma-
tion process (Frost, 1986a; Jee, 1983; Parfitt, 1984a,b;
The Tissue-Level Biologic Mechanisms Recker, 1983). It occurs on all four skeletal “enve-
Briefly, longitudinal growth of bone adds new primary lopes”-the periosteal, haversian, corticalendosteal, and
spongiosa and cortical length to the bone bank (Enlow, trabecular bone surfaces-and it does so throughout life,
1963; Jee, 1983). Increasingly vigorous MU tends to whereas macromodeling drifts only affect compacta and
increase those additions, while decreasing MU retards primarily during growth.’ The amounts of resorption
the additions but cannot reverse them. The proof for this and formation in completed remodeling packets can vary
and some subsequent statements lies in differences in according to anatomical location and the influence of
the bones of congenitally paralyzed and normal limbs drugs, disease, and mechanics, and in ways that can
(Frost, 1986a, 1987). It was only recently realized that cause net gains or losses of bone in small increments or
the bone mass effects of the biological mechanisms re- decrements, respectively, that are signified by A B.BMU
spond more to the magnitudes of the forces on bones and shown in Figure 2 (Frost, 198613). Averaged over the
than to their frequency (Frost, 1985, 1986a). As exam- whole skeleton, the typical remodeling packet, or BMU,
ples, marathon runners do not usually have massive resorbs about 20 parts of bone but replaces about 19
bones but weight lifters do. (Frost, 1985, 1986b). Because of the negative value of
The resorption and formation drifts that provide most the A B.BMU where bone touches marrow, remodeling
bone macromodeling move periosteal and cortical-endos- usually removes that bone in discrete small decrements
teal bone surfaces in tissue space to determine diaphy- or “bites” (Coupron, 1981; Parfitt, 198413).As one result,
seal outside diameter and cortical cross-section area, as increased remodeling usually increases losses of trabec-
shown in Figure 1(Albright and Brand, 1987; Jaworski, ular and cortical-endosteal bone while decreased remod-
1984; Jee, 1983). Increasingly vigorous mechanical usage eling reduces those losses and could be said to conserve
makes modeling increase cortical bone bank deposits, that bone (Anderson, 1986; Jee et al., 1983; Martin,
while disuse retards future deposits but seldom reduces 1986; Parfitt, 1986).That arrangement also allows adults
 THE MECHANOSTAT 3

It

Fig. 2. The remodeling packet or basic multicellular unit (BMU). Top That increment is the meaning of A B.BMU. The “stair graph” in the
row: When on a quiescent bone surface seen in cross section at A, a drawing below shows that a series of such BMUs over several years
new remodeling packet or BMU is activated or recruited; then new would build up the affected bone surface. Here A B.BMU = +. At H
osteoclasts are produced locally to create a resorption bay (B). They are equal amounts of bone are resorbed and replaced per completed BMU
followed by osteoblasts that fill the bay back up with new bone (C). so no net gain or loss occurs even after repeated cycles of local remod-
That ARF (activation-resorption-formation) sequence normally takes eling as in the drawing below. Here A B.BMU = 0, and that is what
some 3 months to pass in healthy humans, and it turns over about 0.1- usually happens on intracortical haversian surfaces. I shows a net
.05 mm3 of bone per completed packet. Similar cellular and temporal decrement per completed BMU so A B.BMU = -, and as in the draw-
sequences construct secondary osteons, but with a different geometry. ing below, that causes a progressive net loss of bone in quantized small
Second row: To illustrate a n important property of the above mecha- decrements or “bites.” That normally happens to bone adjoining mar-
nism, the horizontal and vertical scales are switched and idealized, so row, e.g., to trabeculae and corticalendosteal bone, and the amount of
each drawing corresponds to the one above it and the one on the right bone removed by a typical such “bite” would be on the order of 0.003
is a “BMU graph” (after the author). Third row: At G an excess of bone mm3. (This and the next figure reproduced with permission. H.M.
was made per completed BMU so a permanent increment of new bone Frost. Intermediary Organization of the Skeleton. C.R.C. Press, Boca
has been added, which is the usual case on periosteal bone surfaces. Raton, 1986.)
to turn over some 10-30 x more bone annually than they are shown by voluminous clinical, experimental, and
lose. histomorphometric evidence, some cited by Currey
Increasingly vigorous mechanical usage tends to de- (19841, DeLuca et al. (19811, Frame and Potts (1983),
press the number of new remodeling packets that are Heaney et al. (1978),Lanyon (19841, Lips (19821, Minaire
being recruited, and throughout an affected bone-as (19731, Uhthoff (1986),Recker (19831, Smith et al. (19841,
shown in the bottom left compared to the bottom middle Whedon (1984), Wronski and Morey (1982, 1983a,b),
drawing in Figure 3. Increasing mechanical usage may Wronski et al. (19861, and Young et al. (1986).
also make the A B.BMU less negative, as in the second
row of drawings in the figure, or even positive in se- Importance of the recruitment processes
lected locations. Those effects would tend to conserve Osteoclasts and osteoblasts are short-lived dependent
trabecular and cortical-endosteal bone. Acute disuse re- -
parts of the also-short-lived(e.g., 3 months) multicel-
moves that depression, allowing remodeling to increase lular drifts and BMU, which include precursor and sup-
as at the bottom right in Figure 3, so it takes more porting cells, capillaries, and special organizations at
numerous and often larger A B.BMU “bites” from bone the microscopic level (Jee, 1983). If recruitment of those
in touch with marrow (Jaworski et al., 1980; Uhthoff mechanisms stops, then human growth, modeling, re-
and Jaworski, 1978;Jee et al., 1983; Wronski and Morey, modeling, and net gainsllosses of bone also stop within
1983a). Accordingly, in children and adults increasing some 3 months, for as Jaworski (1984), Parfitt (1984a,b),
mechanical usage tends to conserve existing spongiosa and the author (Frost, 1983, 1986a,b) have stated, con-
and corticalendosteal bone, while decreased usage ac- tinued modeling and remodeling absolutely require the
celerates losses of those bone fractions, so spongiosa continued recruitment of the biologic mechanisms which
becomes osteopenic and the marrow cavity enlarges. determine where, when, and if they act, and how long
Table 1 summarizes those effects of mechanics which and how much.
4 H.M. FROST

Fig. 3.Top row: These “stair graphs” (after P.J.Meunier) show the the author (Frost, 1986a-c). Third row: During mechanical disuse the
normal A B’BMU signs and the associated local bone mass effects on net bone decrement per completed BMU can and usually does increase,
the periosteal (P),haversian (H),and cortical-endosteal and trabecular usually as suggested in the left drawing. But the middle and right
(C-E, T) bone surfaces or envelopes. With continuing remodeling the situations can also occur, the former under treatment with adrenal-
P envelope progressively expands, the H envelope remains nearly cortical steroids. Bottom row: Changing the number of BMU created
unchanged, but bone is progressively lost on the C-E, T envelopes. annually will proportionally change the rate of bone loss on the C-E
Those phenomena account for most of the normal human adult’s slight and T envelopes, given a constant net decrement per completed BMU.
increase in bone diameter with aging and the accompanying reduction If the middle drawing represents normal, then increasing mechanical
of spongiosa, marrow cavity expansion, and cortical thinning. The usage or a lowered remodeling MES setpoint would change the situa-
latter occurs because the marrow cavity expands faster than the per- tion to that on the left. Decreasing usage or an elevated setpoint would
iosteal envelope and diameter. Second row: These BMU graphs follow change the situation to that on the right. The immediate and cumula-
a given locus on a bone surface over time toward the right, say for 2 tive effects of such setpoint changes should be the same whether due
years. In that time three remodeling packets arise on it and complete to hormones, other circulating agents, disease, age, drugs, genetics,
their ARF sequences. This system has the property that given net toxic agents, nutrition, local biochemical agents, or combinations
decrements of bone per typical completed BMU can accompany normal thereof. However, they could also accompany separate concurrent ef-
(on the left), greatly reduced (in the middle), or greatly increased (on fects of such agents on the responding modeling and remodeling sys-
the right) total amounts of bone formed, resorbed, and turned over. tems, andor on the subject’s vigor of daily activities. The author
Thus the BMU mechanism can cause local bone gainsflosses indepen- believes such combined effects do occur and the “noise” they add to
dently of global resorption or formation. Influences on that phenome- strictly mechanical effects partly explains why it took so long to per-
non of this system’s transients and steady states and of changes in the ceive the mechanostat’s footprints and properties.
remodeling space have been discussed in detail by Jaworski (1984)and

The above facts mean a mechanostat should certainly that causes peak bone strains smaller than about 1,500
include the growth, modeling, and remodeling mecha- microstrain, but they are enabled when strains rise to
nisms. Attention turns next to what controls their re- the 1,500-2,500 range, whereupon they increase cortical
sponses to mechanical usage. bone mass. That tends to lower peak strains back toward
that threshold range or minimum effectivestrain (hence,
The MES Mechanisms MES; Frost, 1983), which in effect switches modeling
Setpoint effects ON and OFF. That behavioral threshold persists be-
Different mechanisms appear to control the responses tween birth and maturity while typical peak mechanical
of the modeling and remodeling mechanisms to bone loads on many bones increase more than 50-fold (Al-
mechanics. As for modeling, after 1974, studies of in bright and Brand, 1987; Cochran, 1982; Currey, 1984;
vivo bone strains reviewed by Bouvier (1985), Currey Frost, 1986a).
(19841, and Lanyon (1984) showed that special strain That 1,500-2,500-microstrain range can be considered
ranges apply to the bone mass effects of modeling and the normal setpoint of the MES for mechanically con-
remodeling. Normal lamellar bone fractures when its trolled bone modeling, and at least one of its functions
mechanical strains attain some 25,000 microstrain seems to be making modeling adjust bone mass to keep
(Cochran, 1982). Yet mechanically controlled bone mod- typical peak strains below the threshold range (Frost,
eling drifts remain dormant during mechanical usage 1986b,c; Rubin, 1984). Too small or large a setpoint
 THE MECHANOSTAT 5

should have the predictable bone mass effects listed in done or cited by Lanyon (1984), Jaworski et al. (19801,
Table 2. During growth a high setpoint would retard Jaworski (1984). Jee (1983), Minaire (19731, Uhthoff
further additions of compacta, while a low setpoint would (1986), Wronski and Morey (1982, 1983a,b),and Wron-
increase them. An increased setpoint would have the ski et al. (1986) indicate that under mechanical usage
same bone mass effects as decreased vigor of mechanical that causes smaller peak bone strains than some 100-
usage, and conversely. Such behavior means a mecha- 300 microstrain, remodeling can proceed rapidly due to
nism or mechanisnls, called here the modeling MES active recruitment of new BMU, but as strains exceed
mechanism, should monitor bone's mechanical usage that range recruitment of new remodeling packets de-
and tell the modeling drifts what to do about it. Biome- clines or is depressed, and to one-half to one-twentieth
chanicians recognize that, and some candidates for the of the derepressed rate. Too high an MES setpoint (set
signals that do it are under study at present by Lanyon range?) for mechanically controlled remodeling would
(19841, Otter et al. (19851, and others. let the latter increase, which would accelerate removal
Remodeling responses to mechanics differ. Studies of bone adjoining marrow as in Figure 3 bottom right.
Too low an MES setpoint would depress remodeling and
conserve that bone as in Figure 3 bottom left. Here too
TABLE 1. Mechanical usage (MU) effects on bone growth, an increased setpoint would have the same bone mass
modeling, global remodeling, and mass* effects as decreased mechanical usage, and conversely.
Also listed in Table 2, those effects appear to persist
Cumulative activity from birth through adult life, which would mean a
MU increased: growth and modeling increase; remodeling
declines mechanism, called the remodeling MES mechanism
MU decreased: growth and modeling decrease; remodeling here, should manage them. Note here that mechanical
increases bone loads always cause corresponding bone strains
Compact bone mass (Cochran, 1982; Currey, 1984)and that mechanical usage
MU increased: mass increases in children, is conserved in could influence modeling and remodeling by effects in
adults addition to bone strains-for example, by changing bone
MU decreased: gains decrease in children, mass decreases blood flow or intraosseous pressure (Frost, 1986a; Lan-
in adults due to marrow cavity expansion yon, 19841.
Trabecular bone mass Consequently for both modeling and remodeling, in-
MU increased: existing spongiosa is conserved at all ages;
additions of new spongiosa increase in children creasing mechanical usage would have the same bone
MU decreased: loss of existing spongiosa increases at all mass effects as increased setpoints. It is intriguing that
ages; additions of new spongiosa decrease in children the modeling and remodeling activity responses to
Bone architecture changing mechanics are opposite, yet their bone mass
MU increased: Children: thicker cortex, greater outside responses are in the same sense, - for example, when
bone diameter, denser spongiosa, smaller marrow cavity, one increases gains the other decreases losses, while if
slightly longer bone one reduces gains the other increases losses.
Adults: conserved spongiosa and cortical-endosteal bone
MU decreased: Children: smaller outside diameter of bone,
osteopenic spongiosa, slightly shorter bone Construction of the MES mechanisms
Adults: larger marrow cavity, osteopenic spongiosa
As proposed elsewhere (Frost, 1985, 1986a-c,1987),for
The total amount of spongiosa in the bone bank increases during
growth because the growth-related additions of new primary spongiosa mechanics to control biology those mechanisms should
exceed the removal by continuing remodeling of the existing have mechanical and biological parts, such that a me-
spongiosa. When growth stops a t maturity, its additions of new chanical load (L) on a bone (B) would generate a primary
spongiosa do likewise, whereupon the net losses due to continuing mechanical signal (Sl), which is then detected (D) by
remodeling begin to become apparent (table adapted from material in
(Frost, 1987). cells which would generate secondary signals (Sz) ad-
Growth and modeling drifts are not normally effective in human dressed to the responding modeling %) and remodeling
adults. &.) systems. That message traffic in both MES mecha-

TABLE 2. Predicted steady-state effects of changed MES setpoints for modeling and remodeling compared to MU
effects in adults and children'
Cortical
cross-
Setpoint Cortical Spongiosa Outside section Trabecular Remodeling
change mass mass diameter area thickness rate
MU-N N2 N N N N N
MES-N N N N N N N
MU-D (adult) D D N D D N
MES-I (adult) D D N D D N
MU-I (adult) I I N I I N
MES-D (adult) I I N I I N
MU-D (child) D D D D D N
MES-I (child) D D D D D N
MU-I (child) I I I I I N
MES-D (child) I I I I I N
'The early and late effects of changes in vigor of mechanical usage and in the MES setpoints or set ranges differ in some significant and
meaningful respects. The text explains some of those differences.
2Code: N = normal; D = decreased; I = increased. MES = minimal effective strain.
6 H.M. FROST

TABLE 3. Comparison of early and late bone mass effects in adult life of MU, disease, drugs, and hormones'
Early effects Late (steady-state)effects
Annual Annual New
net net steady-
bone bone state
loss, Annual Annual loss, Annual Annual bone
global modeling remodeling global modeling remodeling mass
MU-N N 0 N N
MES-N N 0 N N
MU-D I 0 I N
MES-I I 0 I N
Oophorectomy I 0 I N
Paralysis I 0 I N
Adrenal steroids I 0 I N
1"hyperparathyroidism I 0 I N
Calcium deficiency I 0 I N 0 N D
MU-I D 0 D N 0 N I
MES-D D 0 D N 0 N I
Estrogen RX D 0 D N 0 N I
Thyroidectomy D 0 D N 0 N I
Caicitonin D 0 D N 0 N I
Weightlifting D 0 D N 0 N I
'Code: N = normal; D = decreased; I = increased; 0 = no change.

nisms and their general construction should look like ing the MES mechanisms would raise their effective
this, even if some of their cells and other details differ: setpoints, while making them overreactive would lower
the setpoints. That mode of action should cause two
MES mechanism special and otherwise enigmatic phenomena:
1)The bone mass effects of such agents should dupli-
L -+B + S1 -+ D + S2 -+(R,,R,) Relation"(2) cate those of changing mechanical usage and MES set-
Doints with respect to kind and anatomical distribution.
mechanical -+ + biological 2) Those bone mass effects should go through an early
phase of relatively rapid change, followed by relative
Clearly any mechanostat should include those MES stabilization at a new bone mass level.
mechanisms too, to control the observed and differing
activity and bone mass responses of modeling and re-
modeling to mechanical usage. Accordingly relation (1) Comparison of Fact to Hypothesis
can be expanded thus: Table 3 compares the bone mass effects of altered
mechanical usage and MES setpoints, and of some drugs,
diseases, activities, and nutritional states, taken from
---
-------Mechanostat - ------------- various sources (including Albright and Brand, 1987;
MU -+ boie -+ MES mechs. modeling/remodeling Brianqon and Meunier, 1979; Charles et al., 1985;
-+

systems bone mass Cruess, 1982; DeLuca et al., 1981; Frame et al., 1973;
t
-+

Frame and Potts, 1983; Frost, 1986a; Heaney et al.,

??? Relation (3) 1978; Kuhlencordt and Bartelheimer, 1980; Lips, 1982;
Orthopaedic Research Society Abstracts [ORSA], 1975-
1987; Peck 1984; Recker, 1983; Smith et al., 1984; Urist,
The question marks under the MES mechanisms ask 1980). They do show striking similarities and the above
how nonmechanical factors such as hormones might af- two phases, which strongly suggest they act via common
fect them. That novel question is discussed next, observ- controlling mechanisms, for which the MES mecha-
ing first that agents such as hormones would not likely nisms seem to be the most visible candidates. Relation
promptly change the mechanical nature and behavior of (1)could be expanded further to account for those fea-
existing bone, and thus the mechanical-side contribu- tures. Thus, including the feedback loop in that relation:
tion to the MES setpoints, but they could promptly affect r-'---'"'----''----'---------------------- I
the biological-side cells in ways that changed the appar- M U ! I-- ------______ Mechanostat -----------
I I
ent MES setpoints. -11 I

? cii I

i
!
SYNTHESIS bone-+ MES mechanisms modelinghemodeling
-+

An Intriguing Possibility I
1 4
The above material suggests that some circulating
and other agents and diseases might affect bone mass
by making the MES mechanisms either somewhat deaf, hormones. nutrition
or somewhat overreactive, to the skeleton's normal me- biochemical messengers
chanical usage, rather than or as well as by acting on disease genetic
existing osteoclasts and osteoblasts, or on their intact
parent modeling drifts and remodeling packets. Deafen-
toxic agents "x'
Relation (4)
 THE MECHANOSTAT 7

“ X ’ in that relation means this matter could have Fluoride Treatment

more to it; the relation encodes a seminal concept, not If fluoride lowered the effective MES setpoints by some
the closing argument to a jury. The dashed arrows indi- 15%, the now-oversensitive mechanostat would “per-
cate that nonmechanical factors such a s disease could ceive” a spurious overloading of the subject’s bone, so it
also change mechanical usage to affect bone mass with- would try to conserve what was there and even try to
out necessarily acting directly on the MES or modeling thicken trabeculae and add woven bone to existing la-
or remodeling mechanisms, or that they could have com- mellar bone. Those things do happen in many humans
bined effects on mechanics, the MES, and the respond- treated with the agent (Charles et al., 1985) and in those
ing biologic system. Much clinical, pathological, and animal species in which bone mass responds to fluoride
experimental evidence does suggest that such combined (e.g., rabbit, dog, and cattle but not rats) (Braincon and
effects occur. Meunier, 1979).
EXAMPLES Chronic Illness
The novelty of the mechanostat proposal does not lie It is also rarely remarked but true that the osteopen-
in the underlying bone-biological and biomechanical ias and osteoporoses that accompany many chronic ill-
facts, which are sound if perhaps arcane to some readers nesses accompany chronically and typically reduced
because of inefficient interdisciplinary communication. physical activities, for in the author’s experience such
The novelty lies in how the proposal relates them and patients are usually feeble and relatively inactive. Their
their effects on each other and on bone mass. The pro- diseases would include chronic hepatic, pulmonary, and
posal would merit consideration to the degree it could renal failure, malnutrition, anemias, polyarticular ar-
explain otherwise enigmatic facts. For that reason some thritis, motor dystrophies, and neurologic pareses. The
familiar clinical situations will illustrate how the au- mechanostat should sense that decreased activity and
thor believes the mechanostat could and probably does reduce the bone masses in adults or retard further ac-
work. cumulations in children until the remaining masses fit
the reduced physical activity, whereupon further losses
should decline to minimal levels. Again that would be
Postmenopausal Osteoporosis
the effect of the feedback loop in relations (1, 4). Those
If the endocrine and other changes associated with changes do happen in such patients, so direct effects of
menopause raised the effective MES setpoints some 30% circulating biochemical and endocrine agents on osteo-
by acting on the biological sides of the MES mecha- clasts and osteoblasts need not cause or be the sole cause
nisms, the mechanostat should “perceive” a spurious of their bone mass reductions.
30% excess of bone, so as Table 1 suggests, it should The same events should occur after paralysis or other
derepress remodeling, which should increase and begin causes of acute disuse, and do, as shown or reviewed, for
to remove that excess bone, but only bone in touch with example, by Frost (1986a1, Lips (19821, Minaire (1973),
marrow, because only there is the A B-BMU negative. Jaworski et al. (19801, Jee et al. (19831, and Wronski and
That would make spongiosa osteopenic and cortices Morey (1982, 1983a, b).
thinner and enlarge the marrow cavity. Adults lack
effective modeling drifts, so no great change in outside
Genetically Determined Bone Mass Features
bone diameter should occur, although i t might increase
slightly due to the positive ABeBMU on that envelope. If a genetic factor set the MES setpoints somewhat
But because the A B-BMU is zero on the haversian lower than the norm for most people and races, the
envelope and positive on the periosteal, no permanent mechanostat should sense that a normal amount of bone
bone losses should occur on either. When the bone mass is somewhat inadequate, so it would make affected in-
had reduced some 30%(e.g., to - 70%of normal), then dividuals accumulate somewhat more bone than the
under continued normal physical activities the now- norm during growth, and then retain more bone than
somewhat-overloaded remaining bone should generate the norm throughout life. As R.R. Recker has also sug-
- 30% larger-“louder”- mechanical signals than be- gested (in private correspondence), that could explain
fore, which the somewhat deaf mechanostat should hear why U.S. Negroes tend to have more bone than Cauca-
again, whereupon it would depress remodeling to nor- sians and Orientals, and also the rare asymptomatic
mal or even subnormal rates. That effect of the feedback Caucasian who has unusually dense bones.
loop in relations (1, 4) would also reduce further bone Contrariwise, if a genetic factor made the setpoints
losses. Thus a phase of increased bone turnover and loss too high from birth the mechanostat would sense a spu-
would occur, followed by reduced turnover and bone rious excess of bone mass, so it would reduce accumula-
mass stabilization at a lower level. tion of bone during growth. That would cause such
Precisely those events and sequences do occur during individuals to have osteopenic spongiosa, reduced corti-
and after menopause, as reviewed recently by, among cal thickness and outside bone diameter, and bones that
others, Johnston (1986), Melton (19861, Mosekilde and fracture easily. Due to the resulting impairment in mod-
Mosekilde (1986), Mosekilde et al. (1985), and Parfitt eling, malunions in children should tend to persist and
(1986, 1984 a, b). They also follow oophorectomy in dogs, accumulate rather than correct with further growth,
rats, and baboons as shown andor reviewed by Ander- while the higher the setpoint the earlier the condition
son (1986), Martin (1986), and Wronski et al. (1986). should become apparent clinically, and the more severe
Adrenal-corticosteroid hormones would have similar ef- it should be. The anatomical distributions of the bone
fects, plus certain nuances due to added direct effects of mass deficits should be those predicted by the properties
those agents on the responding modeling and remodel- of the mechanostat described in the section of this re-
ing mechanisms via the dashed arrow in relation (4). view called The Mechanostat: Its Parts. Those phenom-
8 H.M. FROST

ena do occur in osteogenesis imperfecta (Frost, 1986a,b; ias, and the bone mass effects of hormones, nutrition,
Sillence, 1981). biochemical agents, genetics, and physical activity. It
suggests biomechanical factors are more intimately in-
Life in Orbit
volved in some metabolic bone diseases and bone mass
The bone losses of astronauts have often been assumed features of other diseases than was realized in the past.
to reflect this scheme: lack of gravity --t putative circulat- It is proposed here, not as the truth of bone mass mat-
ing agent -+ osteoclasts + bone loss. That concept over- ters, but rather as another likely one of the many truths
looks some implications of a basic biomechanical fact: that explain the health and diseases of a sophisticated,
body weight is not the major mechanical load on the complex and highly integrated system, the living skele-
spine and lower-extremity bones. Rather, that weight is ton. An old adage among anatomists holds that function
a resistance muscle forces must overcome to move the dictates form. The mechanostat could be one of the many
body’s mass against the pull of earth’s gravity, and as mechanisms that, collectively, control that process and
Cochran (1982), Currey (19841, and many other biome- thereby define its properties, potentials, and limits.
chanicians have reported, those muscles work against
such poor lever arms that peak muscle forces on spine ACKNOWLEDGMENTS
and lower-extremity bones can exceed body weight alone The author would like to thank Mrs. Betty Uhernik,
by two to ten times. Normal bone mass should and does who typed the many drafts of this work, Mr. David
fit the demands of those larger loads. Gavin, who supplied the drawings, and Drs. C. Ander-
In orbit that weight and its resistance to muscle con- son, D. Burr, Z.F.G. Jaworski, W.S.S. Jee, C.C. Johnston,
traction is zero, so the remaining resistance to muscles Jr., and R.R. Recker for helpful critique and suggestions
is the inertial resistance of the body’s limb masses to concerning the concept and how to present it more
being moved. That resistance is about 10% of the resis- clearly. Thanks are also rendered to colleagues of the
tance on terra firma. Accordingly, in orbit a n astro- Southern Colorado Clinic who made available the facil-
naut’s skeleton enters a state of profound disuse ities for preparing the manuscript.
compared to life on earth. The mechanostat should per-
ceive that, derepress remodeling, and begin to remove LITERATURE CITED
what it “perceives” as a huge excess of bone relative to Albright, J.A., and R.A. Brand (1987) The Scientific Basis of Ortho-
the needs of ongoing mechanical usage. The mechano- paedics. 2nd ed. Appleton-Century-Crofts,New York.
stat’s properties, e.g., the value of the A B-BMU on the Anderson, C. (1986) Three studies of oophorectomy and estrogen and
progestone replacement in Beagle dogs. Reviewed at the 1986 Uni-
four bone envelopes, require the bone removed to be versity of Utah-sponsored Hard Tissue Workshop. Calcif. Tissue
primarily that adjoining marrow. The terra firma facts Int., (in press).
apply with less force to the upper extremities than to Bouvier, M. (1985) Application of in vivo bone strain measurement
techniques to problems of skeletal adaptations. Yrbk. Phys. An-
the spine and lower extremities because humans are thropol., 28:237-247.
bipedal, not quadripedal, so the above bone losses should Brianqon, D., and Meunier P.J. (1979) Le Traitement de I’Osteoporose
be proportionally smaller in the uppers. In growing sub- par l’hssociation Fluorure de Sodium, Calcium, Vitamine D. Univ.
jects such as rats or puppies the mechanostat’s proper- Claude Bernard, Lyon.
Charles, P., L. Mosekilde, and F.T. Jensen (1985)The effects of sodium
ties require those load reductions also to retard modestly fluoride, calcium phosphate and vitamin D2 for one to two years on
both longitudinal growth and expansile bone model calcium and phosphate metabolism in postmenopausal women with
drifts, a s suggested in Table 1. spinal crush fracture osteoporosis. Bone, 6t201-206.
All of those things do happen (Jee et al. 1983; ORSA Cochran, G . Van B. (1982) A Primer of Orthopaedic Biomechanics.
Churchill-Livingstone, Edinburgh.
1975-1987; Smith et al., 1984; Urist, 1980; Wronski and Courpron, P. (1981) Bone tissue mechanisms underlying osteoporoses.
Morey, 1982, 1983a,b). That suggests NASA investiga- Orthop. Clin. N. Am., 121513446,
tors might turn from seeking a mysterious circulating Cowin, S.C. (1986) Wolffs Law of trabecular architecture at remodel-
agent&) released by loss of gravity that makes osteo- ing equilibrium. J. Biomech. Eng., 108:83-88.
clasts remove bone in orbit to seeking how to lower MES Cruess, R.L. (1982) The Musculoskeletal System. Embryology, Bio-
chemistry and Physiology. Churchill Livingstone, Edinburgh.
setpoints andor how to suppress the increased recruit- Currey, J.D. (1984) The Mechanical Adaptations of Bones. Princeton
ment of remodeling packets that occurs in orbit andor University Press, Princeton.
devise a stratagem that duplicates in orbit the large DeLuca, H.F., H.M. Frost, W.S.S. Jee, C.C. Johnston, Jr., and A.M.
Parfitt (1981) Osteoporosis. Recent Advances in Pathogenesis and
resistance to muscle forces encountered on terra firma. Treatment. University Park Press, Baltimore.
More examples of plausible workings of the proposed Enlow, D.H. (1963)Principles of Bone Remodeling. Charles C. Thomas,
mechanostat could be given and further details of the Springfield.
above ones, too, but the above make the point: the mech- Frame, B., A.M. Parfitt, and H. Duncan (1973) Clinical Aspects of
anostat’s properties can explain in considerable detail Metabolic Bone Disease. Excerpta Medica, Amsterdam.
known and varied but otherwise enigmatic features of Frame, B., and J.T. Potts, Jr. (1983) Clinical Disorders of Bone and
Mineral Metabolism. Excerpta Medica, Oxford.
some familiar clinical situations. Frost, H.M. (1964) The Laws of Bone Structure. Charles C. Thomas,
Springfield.
Frost, H.M. (1983) A determinant of bone architecture. The minimum
CONCLUSION effective strain. Clin. Orthop., 175:286-292.
That concludes this text and its argument, which pre- Frost, H.M. (1985)The pathomechanics of osteoporoses. Clin. Orthop.,
200: 198-225.
sents succinctly a seminal idea that invites critique, Frost, H.M. (1986a) Intermediary Organization of the Skeleton. C.R.C.
testing, and refinement by others. If it has the merit Press, Boca Raton.
some of us believe (including D. Burr, R.V. Evans, W.S.S. Frost, H.M. (198613) Pathogenic mechanisms of the osteoporoses. In:
Jee, C.C. Johnston, Jr., B. Martin, A.M. Parfitt, and R.R. Current Concepts of Bone Fragility. H. Uhthoff, ed. Springer Ver-
lag, Berlin, pp. 329-361.
Recker) it suggests some changes in the strategy and Frost, H.M. (1986~)Setpoint proposal for osteogenesis imperfecta. In:
focus of the research and treatment that apply to some Current Concepts of Bone Fragility. H. Uhthoff, ed. Springer-Ver-
clinical human problems such as osteoporoses, osteopen- lag, Berlin, pp. 215-219.
 THE MECHANOSTAT 9

Frost, H.M. (1987) Vital biomechanics. Proposed general concepts for Parfitt, A.M. (1986) Review of bone mass and turnover in pre-, intra-
skeletal adaptations to mechanical usage. Calcif. Tissue Int., (in and postmenopausal women. Given at the 1986 University of Utah-
press). Sponsored Hard Tissue Workshop.
Heaney, R.P., R.R. Recker, and P.D. Saville (1978)Menopausal changes Parfitt, A.M. (1984a) The cellular basis of hone remodeling: The quan-
in bone remodeling. J. Lab. Clin. Med., 92964-970. tum concept reviewed in the light of recent advances in the cell
Jaworski, Z.F.G. (1984)Lamellar bone turnover system and its effector biology of bone. Calcif. Tissue Int. [Suppl], 36:37-45.
organ. Calcif. Tissue Int. [Suppl.], 36:46-55. Parfitt, A.M. (198413) Age-related structural changes in trabecular and
Jaworski, Z.F.G., M. Liskova-Kiar, and H. Uhthoff (1980)Effect of long cortical bone: Cellular mechanisms and biochemical consequences.
term immobilization on the pattern of bone loss in older dogs. J. Calcif. Tissue Int. [Supl], 36t123-128.
Bone Joint Surg., 6ZB:104-110. Peck, W. (1984) Bone and Mineral ResearcM3 (Ed). Excerpta Medica,
Jee. W.S.S. (1983) The Skeletal Tissues. In: Histology, 5th ed. W.L. Amsterdam.
Weiss, ed. Elsevier-North Holland, New York, pp. 200-255. Recker. R.R. (1983)Bone Histomorphometry. Techniques and Interpre-
Jee, W.S.S., T.J. Wronski, E.R. Morey, and D.B. Kimmel(1983) Effects tation. C.R.C. Press, Boca Raton.
of spaceflight on trabecular bone in rats. Am. J. Physiol. 13t310- Recker, R.R., P.D. Saville, and R.P. Heaney (1978) Effect of estrogens
314. and calcium carbonate on bone loss in postmenopausal women.
Johnston, C.C., Jr. (1986) A review of pre-, intra- and postmenopausal Metab. Bone Dis. Relat. Res., 1:7-13.
bone loss in women. Given at the 1986 University of Utah-spon- Rubin, C.T. (1984) Skeletal strain and the functional significance of
sored Hard Tissue Workshop. bone architecture. Calcif. Tissue Int. L%~ppl.],36:ll-18.
Kuhlencordt, F., and H. Bartelheimer (1980) Klinische Osteologie. Sillence, D. (1981) Osteogenesis imperfecta. An expanding panorama
Springer Verlag, Heidelberg. of variants. Clin. Orthop., 159:ll-20.
Lanyon, L.E. (1984) Functional strain as a determinant for bone re- Smith, E.L., Jr., P.E. Smith, C.J. Ensign, and M.M. Shea (1984) Bone
modeling. Calcif. Tissue Int. [Supl.], 3656-61. involution decrease in exercising middle-aged women. Calcif. Tis-
Lips, P. (1982) Metabolic Causes and Prevention of Femoral Neck sue Int. [Suppl.], 36:129-138.
Fractures. PhD Thesis. University of Amsterdam, Amsterdam. Uhthoff, H. (1986) Current Concepts of Bone Fragility. Springer-Ver-
Martin, B. (1986) Comparison of bone turnover and bone mass findings lag, Berlin.
in oophorectomized and estrogen treated oophorectomized dogs. Uhthoff, H., and Z.F.G. Jaworski (1978) Bone loss in response to long
Reviewed at the 1986 University of Utah-sponsored Hard Tissue term immobilization. J. Bone Joint Surg. 60B:420429.
Workshop. Calcif. Tissue Int., (in press). Urist, M.R. (1980) Fundamental and Clinical Bone Physiology. J.B.
Melton, L.J. (1986) Epidemiology of pre- and postmenopausal bone loss Limincott. Philadelohia.
in women. A review of recent studies in the U.S. and Europe. Given Whedbh, G.D.'(1984) Disuse osteoporosis: Physiological aspects. Calcif.
at the 1986 University of Utah-sponsored Hard Tissue Workshop. Tissue Int. [Suppl.], 36r146-150.
Minnaire, P. (1973) L'Osteoporose d'Immobilization. Donnbes Biolo- Wronski, T.J., and E.R. Morev (1982) Skeletal abnormalities in rats
giques et Histologiques Quantitatives. PhD Thesis. Ediprim, Lyon. induced by simulated weightlessness. J. Metab. Bone Dis. Relat.
Mosekilde, L., A. Viidik, and L. Mosekilde (1985)Correlation between Res., 4:69-76.
the compressive strength of iliac and vertebral trabecular bone in Wronski, T.J., and E.R. Morey (1983a) Inhibition of cortical and trabec-
normal individuals. Bone, 6:291-296. ular bone formation in the long bones of immobilized monkeys.
Mosekilde, L., and L. Mosekilde (1986)Normal vertebral body size and Clin. Ortho., 181:269-276.
compressive strength: Relations to age and to vertebral and iliac Wronski, T.J., and E.R. Morey (198313) Effect of spaceflight on perios-
trabecular bone compressive strength. Bone, 7:207-212. teal bone formation in rats. Am. J. Physiol., 13t305-309.
Orthopaedic Research Society Abstracts (1975-1987). American Acad- Wronski, T.J., C.C. Walsh, and L.A. Ignaszewski (1986) Histologic
emy of Orthopaedic Surgeons, Chicago. evidence for osteopenia and increased bone turnover in ovariecto-
Otter, M., J. Shoenung, and J.D. Williams (1985)Evidence for different mized rats. Bone, 7r119-124.
sources of stress-generated potentials in wet and dry bone. J. Or- Young, D.R., W.J. Niklowitz, R.J. Brown, and W.S.S. Jee (1986) Immo-
thop. Res., 3:321-324. bilization-associated osteoporosis in primates. Bone, 7: 109-118.