Introduction
Anorexia nervosa (AN) is a condition characterized by chronic, severe, self-induced
malnutrition, resulting in hypogonadotropic hypogonadism, an acquired state of
growth hormone resistance, relative hypercortisolemia, and dysregulation of various
appetite-regulating hormones and adipocytokines, all of which have critical negative
effects on skeletal maturation, growth and bone mass accrual. The lifetime
prevalence of AN, as defined by the DSM-VI, is reported to be as high as 2.2%
in women (). It occurs in 0.2-1% of adolescent girls and
1-4% of college-age young women (). Although AN affects
females predominantly, up to 10% of cases are seen in males, and the
incidence of males with this condition appears to be increasing in recent years
(). While many of the clinical consequences of AN appear
to be reversible with recovery, this may not hold true for its impact on the
skeleton. In this Perspective, we will review the skeletal consequences
resulting from hormonal changes of AN in both girls and boys.
Hormonal determinants of adverse skeletal outcomes in anorexia
nervosa
Growth hormone resistance and low insulin-like growth factor I
AN results in a nutritionally-acquired resistance to GH with impressively low
levels of IGF-1 () (Fig. 1).
GH has direct stimulatory effects on the differentiation and proliferation of
osteoblast precursors, and also acts indirectly through IGF-1 to stimulate the
differentiation of these precursors (). Since both are bone
anabolic factors that rise in early puberty and peak in mid-puberty, low IGF-1
levels coupled with GH resistance appear to contribute to the low bone formation
rates observed in AN, as indicated by low levels of bone formation markers. GH
resistance occurs both at the level of the liver and bone, and the latter is
suggested by a lack of association between GH concentrations and levels of bone
formation markers in girls with AN, whereas strong positive associations are
observed in healthy controls (). Some studies (), but not all (), have reported an adverse impact on stature
in adolescents with AN, consistent with low IGF-1 levels and resistance to GH.
However, an effect on stature is also dependent on the duration and severity of
AN. Therefore, timing of AN onset within the adolescent period and in relation
to pubertal onset, progression and the growth spurt has specific and
differential consequences on both immediate and long-term skeletal outcomes.
Hypogonadotropic hypogonadism
AN is characterized by hypothalamic amenorrhea and consequent hypoestrogenic
() and hypoandrogenic states ().
Most studies indicate an inverse association between bone density measures and
duration of amenorrhea (), and low estradiol levels
in AN are predicted by the severity of nutritional status ().
In adolescence, estrogen increases bone mass accrual by two core mechanisms:
first, progressive elevation of estrogen in early puberty stimulates increases
in GH and IGF-1, which have longitudinal bone growth effects and increase
periosteal bone apposition (), and second, surges of estrogen
levels in late adolescence have an osteoclast antiresorptive effect, especially
at endosteal surfaces. The antiresorptive effects of estrogen on bone are
mediated by the osteoprotegerin (OPG)-RANK-RANKL system and by inhibition of
various proinflammatory cytokines that otherwise stimulate osteoclast
differentiation and activation (). Estrogen stimulates
secretion of OPG, the soluble decoy receptor of RANK, which competes with RANKL
for binding to RANK and inhibits osteoclastic activity. Interestingly, although
estrogen levels are low in AN, OPG levels are high (), likely
an adaptive mechanism given the low bone density state. However, proinflammatory
cytokines are elevated in adolescent girls with AN (), and may
contribute to the impaired state of bone metabolism.
In addition, androgen levels are depressed in girls () and boys with AN (), and
likely contribute to low BMD. Studies in adolescents () and
adults () with AN have demonstrated that low testosterone
levels predict low BMD and low lean mass, and the latter is a particularly
important determinant of bone density measures in this population. Data
regarding adrenal androgens, however, are contradictory, with some studies
reporting reduced dehydroepiandrosterone sulfate (DHEAS) levels in young women
with AN, which correlate inversely with markers of bone turnover (), whereas other studies have not been able to demonstrate a
difference in DHEAS levels in adult women () and adolescents
() with AN compared with controls. It should be
acknowledged that hypogonadism likely acts synergistically with other deficient
elements to impact BMD, as administration of oral estrogen alone, for example,
does not improve BMD in adult women or in adolescents with AN (). Of note, the issue of adherence to oral estrogen and
compliance with the recommended regime has not been addressed in some of these
studies.
Hypercortisolemia
A relative state of hypercortisolemia (Fig. 2), as
evidenced by elevated urinary and serum cortisol levels compared with controls,
has been described repeatedly in both adult women () and girls with AN (),
likely a consequence of both increased cortisol secretory burst frequency and
decreased clearance (). Cortisol suppresses bone formation
by inhibiting the replication, differentiation and function of osteoblasts
through various mechanisms, and by inducing apoptosis of mature osteoblasts and
osteocytes (). Cortisol potentiates osteoclast
function, and also inhibits both renal tubular calcium reabsorption and calcium
absorption from the GI tract (). Furthermore, glucocorticoids
may directly decrease the secretion of GH () and hence lower
IGF-1, thus having further detrimental effects. Excess cortisol, both from
exogenous and endogenous sources, has been consistently linked with low BMD
(), and an inverse association
between BMD and cortisol levels has been reported amongst adults with AN (). Furthermore, an inverse association has clearly been
demonstrated between markers of bone formation, such as osteocalcin, and
cortisol in adults () and adolescents () with
AN.
Neuroendocrine gastrointestinal-derived peptides and
adipocytokines
Several peptide hormones have recently been linked to bone health in AN. These
include neuroendocrine gastrointestinal-derived peptides regulating food intake
and certain adipocytokines.
Ghrelin
AN is associated with increased secretion of ghrelin, a gastric-derived
orexigenic peptide, which is a GH secretagogue (),
stimulates osteoblast proliferation in vitro (), and
is expressed in cartilage () and osteoblasts. Elevated
ghrelin levels in girls with AN predict high GH and cortisol burst frequency
(). Importantly, as described above, GH resistance and
hypercortisolemia both negatively affect bone. Interestingly, in a follow-up
study, ghrelin was demonstrated to strongly positively predict BMD in
healthy girls, but not in girls with AN (), suggesting that
AN may confer a state of ghrelin resistance.
Leptin and peptide YY (PYY)
The role of the leptin network and other hypothalamic hormones in regulating
eating behaviors has received significant attention over the last decade.
However, the potential additional function of this system in the central
control of BMD is a relatively new area of focus. Leptin is an anorexigenic
adipocytokine that is depressed in AN ().
Leptin-deficient and leptin-resistant mice are obese and hypogonadal and yet
have high trabecular (but low cortical) BMD ().
Furthermore, administration of leptin to leptin-deficient mice leads to a
decrease in trabecular, but an increase in cortical, BMD, suggesting a
negative impact of leptin on trabecular bone, and a positive impact on
cortical bone. Takeda et al. demonstrated that leptin regulates bone
formation via the sympathetic nervous system (), whereby a
β-adrenergic agonist decreases bone mass in leptin-deficient and
wild-type mice, while a β-adrenergic antagonist increases bone mass in
wild-type and ovariectomized mice without affecting body weight. In contrast
to animal studies, a positive association between leptin and BMD has been
reported in humans even at sites of trabecular bone (). The
relationship between leptin and bone metabolism in AN remains to be
clarified.
Both leptin and peptide YY (PYY) are peripherally-released molecules that
serve as long-term and short-term signals, respectively, of the
energy-replete state (). Their effects on the central
nervous system are to decrease subsequent caloric intake. PYY is an
intestinally-derived anorexigen that acts via the Y2 neuropeptide Y (NPY)
receptor to decrease NPY secretion and inhibit caloric intake (). The downstream mechanism connecting Y receptor activation
to decreased BMD is yet to be well-elucidated. However, it has been shown
that Y2 receptor knockout mice have a two-fold increase in trabecular bone
volume at the distal femoral metaphysis with increased trabecular number and
thickness. Furthermore, osteoclast surface was not affected in these
animals, but osteoclast number was reduced. Osteoblast surface and number,
osteoid surface, and mineralizing surface were all unaffected but rates of
mineral apposition and bone formation were increased. As there is no
detectable Y receptor expression in bone, this was interpreted to represent
an effect mediated within the central nervous system. These data suggest
that signaling through the Y2 receptor suppresses bone formation, and more
specifically trabecular bone development (). In humans, PYY
has been demonstrated to negatively correlate with markers of bone formation
and resorption () and an association between elevated
levels of PYY in AN and low BMD has also been reported (),
suggesting that PYY may act as a catabolic signal to bone. With the
presumption that leptin and PYY both relay a catabolic signal to bone, the
results of their skeletal effects have been examined in human models
studying phenotypes at opposite extremes, obesity and AN. Obese individuals
have elevated leptin and low PYY levels, whereas individuals with AN have
low leptin and elevated PYY levels. Obesity is associated with increased BMD
() and AN with decreased BMD (),
therefore it appears that if these peptides exert independent effects on
bone, PYY may have a dominant effect compared to leptin. Conversely, Wortley
et al. recently reported a PYY knock-out mouse model with an
osteopenic phenotype, including a reduction in trabecular bone mass and a
decrease in bone strength (). These results suggest that
elevated PYY levels in AN may reflect resistance at the level of bone to the
positive effects of PYY.
Adiponectin
Adiponectin is an adipokine that is known to affect bone metabolism (), and consequently has been examined in the
pathogenesis of low BMD in AN. Although one would expect this fat-dependent
hormone to be disturbed in this state of under-nutrition, the specific
direction of derangement has been somewhat baffling such that adiponectin
levels have been reported as normal in adolescents with AN (), and both high () and low () in
adults with AN. Of note, high adiponectin levels are associated with low BMD
in healthy adults (). Adiponectin receptors are
expressed both on osteoblasts and osteoclasts () and a role for adiponectin in suppressing OPG and increasing expression
of RANKL has been described, suggesting that high adiponectin levels may
cause increased osteoclastic activity resulting in low BMD (). Adiponectin, however, is also reported to increase osteoblastic
activity, which theoretically should be associated with increased bone
formation (). Again, this puzzling inconsistency
led to a study that aimed to integrate human phenotypes into the equation
and determine associations of BMD with adiponectin ().
Interestingly in this study, adiponectin did not differ in AN subjects
versus controls, although adiponectin levels 60 minutes after oral glucose
did trend higher in the AN group. Despite the lack of difference in
adiponectin levels between AN subjects and controls, an inverse and
independent association between adiponectin and BMD in AN was
established.
Bone mineral density in anorexia nervosa
AN is complicated by severe bone loss. Osteopenia is present in 92% and
osteoporosis in 38% of young women with AN, with less than 15% of
women having normal bone density, despite an average age only in the early twenties
(). Low bone density in adults with AN is associated with
an uncoupling of bone turnover, with low levels of bone formation markers and high
levels of bone resorption markers. In women with active AN, bone loss occurs
rapidly, at an average annual rate of 2.5% (), highlighting
the importance of early intervention for women with AN. Data regarding skeletal
effects of weight recovery in women with AN are inconsistent and not particularly
reassuring; some suggest that weight recovery results in increases in BMD, whereas
others are not able to demonstrate significant improvement (). Of concern, most studies do agree that some degree
of residual bone loss is common even several years after recovery from AN (). Additionally, body weight history appears to be a crucial
predictor of both the presence of low bone density as well as of recovery;
specifically, patients with a history of a critical body mass index less than 16.4
± 0.3 kg/m2 appear to remain at high-risk for osteoporosis
even several years after their recovery ().
Bone mineral density in adolescent girls with anorexia nervosa
In teenagers with AN, low BMD is associated with a reduced bone turnover state,
evidenced by a decrease in both bone formation and bone resorption markers
(). This is in sharp contrast to the vigorous bone turnover
state characteristic of early puberty (). Furthermore, compared
to a 1-year follow-up period of continued spine bone mass accrual observed in
healthy adolescents, girls with AN have a plateauing of bone density and bone
mineral content (). Teenage girls with AN have been reported to
have Z-scores of < –1 at the spine and hip in as many as
50% and 30%, and Z-scores of < –2 at the spine and
the hip in 9% and 10% (). For whole body bone
mineral content adjusted for height, 30% of girls with AN had Z-scores of
< –1 and 4% had Z-scores of < –2
(). In this particular study, investigators took note of
the fact that areal bone density measurements by dual-energy X-ray
absorptiometry (DXA) are affected by stature, leading to a potential
underestimation of bone density in very short children. This is of concern given
that some studies have reported short stature in AN, raising the possibility
that reports of low BMD in AN are at least in part a consequence of shorter
stature and impaired radial bone growth (). However, these
investigators observed that girls with AN also have lower bone mineral apparent
density (BMAD), a surrogate measure for volumetric bone density that adjusts for
stature (), and were not short, indicating that lower bone
density observed in their cohort was not a consequence of short stature and
consequently short bones ().
With weight gain, there is a significant increase in markers of both bone
formation and bone resorption in adolescent girls with AN, suggesting the
potential to revert to a more physiological and pubescent state of increased
bone turnover. Additionally, an increase in bone formation markers in the first
6 months after weight gain is predictive of increases in bone mineral content in
the subsequent 6 months (). In girls with AN who gained
10% of their BMI and resumed menses, modest increases in bone density and
BMAD at the spine and the whole body were observed over a 1-year follow-up
period (), and these data suggest that sustained recovery of
menses and weight should lead to a significant increase in bone mass and BMD. In
contrast, girls not gaining weight are expected to have continued BMD and BMAD
loss with progressive and significant decreases in their Z-scores over time
(Fig. 3). Therefore, weight gain and resumption of
menses are to be strongly encouraged since even modest improvements are clearly
preferable to a continued detriment of bone health with persistence in
underweight.
Long-term bone mineral density deficits in women with a history of
anorexia nervosa
Although bone mass accrual may improve with weight gain and menstrual recovery,
significant residual deficits certainly persist into adulthood. In a study of
adult women with AN, Biller et al. reported that women with an onset of
amenorrhea before the age of 18 years had lower bone density than those who
developed amenorrhea after 18 years, even after controlling for duration of
amenorrhea (). In another study of 19 women with teenage-onset
AN, bone density at the femur, although not at the spine, was significantly
lower than in controls even after full recovery from AN 14-23 years later (mean
21 years) (). These investigations further emphasize that
adolescence serves as a critical period for establishing life-long bone
health.
Most studies indicate an inverse association between BMD measures and duration of
amenorrhea () as well as duration of illness
() with few obvious or hopeful prognostic indicators of BMD
recovery. For example, in a 25-month follow-up study of women aged 19-37 years
(mean age 26 years) with a 1-17-year history of amenorrhea (mean 5.8 years),
some of whom developed AN in the adolescent years, there was no significant
difference in the mean change in bone density between women who attained greater
than 80% of ideal weight, resumed menses, took estrogen or calcium, or
who exercised vigorously ().
The primary import of BMD is to predict future fracture risk. However, there are
only a few studies that have examined risk of fracture in the follow-up of women
with AN. Interestingly, Wentz et al. () reported
similar rates of fracture in women with AN versus controls after 11 years of
onset of AN (4/36 AN and 5/43 controls). These authors, however, did not
describe the nature of fractures observed, and their AN group was
distinguishable from most reported in the literature in that the women in this
study did not differ in BMD from controls (). In contrast,
Rigotti et al. reported a 7.1-fold greater risk of non-spine fractures in
women with AN during follow-up compared with normal women in the same age range
(). Analogously, Lucas et al. reported a 57%
cumulative incidence of fractures at the hip, spine, and radius in women with AN
40 years after diagnosis of their eating disorder and a standardized incidence
ratio of fractures of 2.9 compared to a population of healthy women (95%
confidence interval, 2.0-3.9) (). These later data suggest
significant fracture risk resulting from AN.
Bone mineral density in adolescent boys with anorexia nervosa
Although AN is primarily a disease of females, it is increasingly being
recognized in males as well. Regardless, there are only a few controlled studies
investigating BMD, bone turnover markers, or their predictors in adolescent boys
with AN. In one uncontrolled study (), which did not examine
bone turnover markers, boys with the lowest BMD had the longest duration of
illness, and the lowest physical activity and calcium intake. A recent study
examined absolute and height-adjusted measures of BMD and levels of bone
turnover markers in adolescent boys with AN; this cohort was found to have lower
BMD and corresponding Z-scores at the spine, hip, femoral neck, trochanter,
intertrochanteric region, and whole body, compared with controls (Fig. 4) (). Height-adjusted measures (lumbar BMAD
and whole body bone mineral content/height) were lower and bone formation
and resorption markers were reduced as well, indicating decreased bone turnover.
IGF-I was shown to be an important predictor of bone turnover markers, and, as
expected, testosterone levels and lean mass predicted BMD ().
Long-term bone mineral density deficits in men with a history of anorexia
nervosa
Analogous to AN, age-onset osteoporosis has traditionally been considered a
disease of women, however, men also incur substantial bone loss with aging and
experience vertebral fractures at approximately one-third the rate of women
(). The pattern of bone loss in men is distinctly
different from that seen in females. In females, BMD begins to decline slowly
after the age of 40 until the age of 55, when the loss of bone density
accelerates. The pattern among males is a slower decline after the age of 40
such that only after age 60 does their BMD decrease to levels almost equivalent
to females of equal age (). The lower prevalence of
osteoporosis in men is a consequence of greater accumulation of skeletal bone
mass during the adolescent and young adult years, greater bone size, and the
absence of a distinct equivalent of menopause. However, recognition of the
significant vulnerability older men still have to low BMD is an important
impetus to investigate long-term affects of AN in adult men.
There are few data reflecting long-term changes in bone density in men with
teenage-onset AN. One study reported lower bone density in three men with AN
compared with controls after 11 years of AN recovery; this difference was
significant for the whole body, but not for the spine or hip (). However, the small sample size makes globalization of these findings
challenging. In reaction to scant investigations addressing long term
osteoporosis risk in men with AN, Mehler et al. ()
studied severity of bone loss in male patients with eating disorders, compared
results to females with AN, and identified factors that contribute to low BMD in
these men. They demonstrated a strikingly high prevalence of low BMD in male
patients with a history of AN. Moreover, the severity of deficiency was greater
than in females with the same eating disorder, especially for those with a very
low BMI and longer duration of illness. Their report strongly implies that
osteoporosis in adult men with AN is likely under-reported and/or
under-investigated, and serves to heighten awareness of the risk for
osteoporosis in male patients with eating disorders.
Alterations in bone microarchitecture in anorexia nervosa
Alterations in bone microarchitecture in AN may explain fracture risk independent
of BMD. Advances in CT imaging allow for noninvasive evaluation of trabecular
microstructure at peripheral sites in vivo. Lawson et al. recently
performed a cross-sectional study of 23 women (12 with AN and 11 healthy
controls) to determine hormonal predictors of trabecular bone microarchitecture
(). Bone microarchitectural measures, including apparent
(app.) bone volume fraction, app. trabecular thickness, and app. trabecular
number, were reduced and app. trabecular spacing was increased in AN versus
controls, consistent with other studies (). Moreover,
decreased structural integrity at the ultradistal radius was associated with
decreased BMD at all sites except the total hip. IGF-I, leptin, testosterone,
and free testosterone levels predicted greater structural integrity based on all
microarchitectural parameters measured. The investigators concluded that bone
microarchitecture is abnormal in women with AN and that endogenous IGF-I,
leptin, and androgen levels predict disrupted bone microarchitecture independent
of BMI (). While techniques used in this study require further
validation, they may prove crucial in revealing additional information about
bone fragility and fracture risk in AN. A similar study has been performed in
adolescents with AN. As in adults, adolescents with AN had lower bone trabecular
volume and trabecular thickness, and greater trabecular separation than controls
(). Of concern, these changes were observed despite the
fact that DXA measures of bone density did not differ between girls with AN and
normal weight controls. These data suggest that changes in bone
microarchitecture may occur even before changes in bone density become
evident.
Therapeutic strategies to increase bone density in anorexia
nervosa
Weight gain and resumption of menses
Several studies show that BMD is not significantly improved by weight gain in
AN (). However, Soyka et al. showed that low
levels of bone turnover markers, present in girls with AN at baseline, did
significantly increase with improvement in nutritional status (). Furthermore, increases in surrogate markers of bone
turnover correlated with an improvement in lumbar and total BMC and BMD in
their AN group. Mika et al. reported corroborating data and
demonstrated that weight rehabilitation in a 2-year period following an
inpatient feeding program led to restoration of bone formation activity in
adolescents with AN, despite a lack of increase in BMD ().
Interestingly, Iketani et al. demonstrated specific benefit to spine
BMD with weight gain, although not to the level of that in controls (). Not surprisingly, BMD further increased with resumption
of menses. A recent study () demonstrated that even
short-term weight gain with menstrual recovery is associated with
stabilization of BMD measures in adolescent girls with AN. Moreover, even in
the absence of menstrual recovery, weight gain seemed to have some positive
effect on whole body parameters, albeit to a lesser extent than that noted
associated with menstrual recovery (Fig. 3).
Importantly, in this study there was no improvement in bone parameters at
the spine solely with weight gain, restating the significance of gonadal
steroids in optimizing trabecular BMD ().
Estrogen replacement
Hypothalamic hypogonadism is a synergistic rather than sole element in the
etiology of low BMD in AN, and multiple studies have shown that estrogen
replacement alone does not significantly improve BMD in this condition,
despite its anti-resorptive effects (). In
fact, a study by Miller et al. demonstrated that despite findings
indicating that resumption of menstrual function was important for
improvement of spine BMD and weight gain critical for improvement in hip
BMD, no increase in BMD was observed in women with AN receiving oral
contraceptive pills (OCPs) even with a mean weight gain of 11.7%
(). This result is speculatively attributed to the
IGF-1- and androgen-suppressive effects of high doses of estrogen in OCPs
(). Conversely, transdermal estrogen
preparations appear to be less IGF-1-suppressive (), and
their role in treating low BMD in AN remains to be determined. These results
call into question the standard clinical practice of prescribing OCPs to
adolescents and women with AN.
Recombinant human growth hormone (rhGH) and insulin-like growth
factor-1 (rhIGF-1)
Although, hypoestrogenemia is an important contributing factor to low BMD in
AN (), the lack of improvement in BMD with
estrogen-containing OCPs () suggests that correction
of other nutritionally dependent factors, independent of, or in addition to,
estrogen are necessary to improve BMD in AN. As discussed earlier, AN is
associated with an acquired resistance to GH and consequent low levels of
IGF-1 (), a nutritionally dependent bone trophic hormone
that stimulates osteoblast function and collagen formation (). In other conditions of GH resistance, such as liver disease or renal
failure, supraphysiologic doses of (rh)GH have been utilized therapeutically
to overcome this resistance (); it would thus follow
that similar interventions may be effective in increasing BMD in AN. In a
recent study, Hashizume et al. reported that administration of high
doses of (rh)GH to AN subjects caused an increase in IGF-1 levels (). A confounder of this study was that an increase in BMI
was also observed, which could independently result in increased IGF-1
levels. In contrast, IGF-1 acts downstream of GH, and investigators have
also examined the impact of (rh)IGF-1 administration on bone formation
markers in adult women with AN. In this study, (rh)IGF-1 administration led
to a normalization of the low levels of IGF-1 and caused a significant
increase in markers of bone turnover ().
The impact of (rh)IGF-1 administration on bone formation in children with AN
had not been studied until most recently. Bone accretion in children is a
physiologically different state from the maintenance of bone mass in adults,
with the former being a high formation and resorption state () leading to a net increase in bone mass, while in the latter, similar
rates of bone formation and resorption result in no net changes in bone
mass. IGF-1 is an important determinant of the pubertal increase in bone
mass (), therefore, effects of (rh)IGF-1 may differ in
adolescent AN girls from those observed in adults with this disorder.
Additionally, the safety of rhIGF-1 administration in adolescents with this
disorder was unknown. To investigate the hypothesis that subcutaneous
(rh)IGF-1 administration in adolescents with AN would stimulate bone
formation, investigators examined responses of surrogate markers of bone
formation, N-terminal propeptide of type 1 procollagen (PINP), and of bone
resorption, C-telopeptide (CTX), to short-term (rh)IGF-I administration
(7-9-day period) in 10 consecutive adolescent girls with AN. They compared
results to those in 10 age-matched girls with AN who did not receive
(rh)IGF-1 and found that this intervention, when given in a dose of 30-40
mcg/kg twice daily, successfully increased IGF-1 levels to a high normal
range and was associated with significant selective increases in levels of
PINP without a concomitant increase in CTX (). This study
revealed that the effects of (rh)IGF-1 were immediate and well-tolerated,
and the associated increase in bone formation markers suggests that there
may be a role for this medication as adjunctive therapy for low BMD in
adolescents with AN.
It is notable that when (rh)IGF-1 is given in conjunction with OCPs, the
result is a small but significant increase in BMD in adult women with AN
(). This is likely because of the combination of the
anabolic effects of (rh)IGF-1 and the anti-resorptive effects of estrogen.
This combination has not been studied in adolescent girls with AN.
Bisphosphonates
Thus far, only two studies have been published examining the role of
bisphosphonates in treating low BMD in AN. Bisphosphonate administration
results in decreased osteoclast bone resorption and is an effective and
commonly implemented treatment for post-menopausal osteoporosis. In a
double-blind, randomized, placebo-controlled study in adolescents with AN,
Golden et al. demonstrated that alendronate treatment resulted in a
modest within-group increase in both spine and femoral neck BMD. However,
BMD improvement in this group was primarily determined by weight
restoration, and increases were not statistically significant when compared
with their placebo group (). In contrast, in an
uncontrolled study of adults with AN receiving risedronate, Miller et
al. showed a decrease in bone resorption with an increase in BMD,
even without significant weight gain (). It is crucial to
note that at this time there is a lack of safety data regarding use of
bisphosphonates in woman of reproductive age, and bisphosphonates are not
FDA-approved in the U.S. for premenopausal women (other than for those
receiving glucocorticoids). Therefore, these medications should remain
confined to the research arena until more safety and efficacy data are
available, particularly in adolescents.
Effects of anorexia nervosa on adult stature in boys and girls
The onset and duration of AN in relation to the adolescent pubertal growth spurt,
achievement of peak height velocity and epiphyseal closure, determine adult height
and also may, at least partially, explain the discriminate outcome in girls versus
boys. Diminished stature in comparison with target height, both in boys and girls,
has been reported in earlier studies (), however,
more recent studies do not fully corroborate these data (). In
fact, one study even indicated greater than expected height in girls with AN ().
The hypogonadal state, associated with AN, causes a delay in bone age (BA), allowing
for a longer duration available for growth, and may offset deficits incurred by low
IGF-I levels, as long as duration of illness is not prolonged ().
Therefore, a delay in BA may contribute to preservation of height potential in some
adolescents with AN. Recent data also indicate that in girls with AN whose BA is
<15 years, an inverse association exists between change in height standard
deviation scores (SDS) over a 1-year period and the delay in BA in relation to
chronological age. These data suggest that girls with AN who have a delay in BA are
more likely to catch up for height SDS than those without delay. It should also be
noted that other important determinants of height measures in girls with AN in this
study were duration of illness and severity of growth deficits before weight
rehabilitation (). Despite the robust power of this study,
Lantzouni et al. reported contradictory data, and showed that following
nutritional rehabilitation in their cohort of girls with AN, an acceleration in
growth velocity was not sufficient to prevent statural deficits (). This discrepancy in outcome may reflect greater severity of illness in subjects
in the latter study.
Because the pubertal growth spurt occurs later in boys than in girls (), boys with onset of AN in the teenage years may be at greater
risk for short stature than girls with this disorder. The pubertal growth spurt
begins approximately 2 years later in boys than in girls, and peak height velocity
occurs at Tanner stage 4 in boys versus Tanner stage 3 in girls (). Furthermore, growth is almost complete at a BA of 15 years in
girls versus 17 years in boys (). Girls may thus be close to their
target height at the time of onset of AN, whereas significant growth potential may
exist in boys who develop AN at the same time. Correspondingly, Modan-Moses et
al. recently reported significant short stature in boys with onset of AN in
the adolescent years, and although weight restoration was associated with some
catch-up growth in this study, complete catch-up did not occur ().