0021-972X/01/$03.00/0 Vol. 86, No.

4
The Journal of Clinical Endocrinology & Metabolism Printed in U.S.A.
Copyright © 2001 by The Endocrine Society

LETTERS TO THE EDITOR

Fractures after Long-Term Alendronate Therapy Safety of Long-Term Alendronate
To the editor: To the editor:
In the September 2000 issue of your journal, a study by Tonino et al. We agree with Dr. Ott that the long-term safety of alendronate is
(1) described the skeletal effects of 7 yr of treatment with alendronate. important because patients with osteoporosis may be treated for ex-
The study concluded that 7 yr of continuous treatment resulted in better tended periods of time. However, we believe that the data support the
skeletal effects than shorter therapy. I am concerned about the vertebral use of alendronate through at least 7 yr. There are several reasons to be
fracture rates that were presented in this paper. These fractures were circumspect about Dr. Ott’s comments.

Downloaded from https://academic.oup.com/jcem/article/86/4/1842/2848945 by guest on 24 April 2024

confirmed by radiographs, but routine radiographs to check for verte- Our study was too small to detect meaningful differences in fracture
bral fractures were not done. During the last 2 yr 3.3%/yr of women incidence in the 2-yr extension. At least 2000 women must be followed
taking alendronate (10 mg/day) had a clinical vertebral fracture. This is for 3 yr to have adequate statistical power for detecting a 40% reduction
higher than reported during the first 3 yr of this study, when the rate in fracture incidence (1). It is also not appropriate to attempt compar-
of morphometric vertebral fractures (defined by measurements from isons of fracture rates to other studies, or to the experiences of the same
radiographs) was 2.1%/yr in the placebo group and 1.1%/yr in the group at an earlier time, because incidence rates may vary between
alendronate groups (2). Direct statistical comparisons can’t be made populations for reasons that are not apparent, even if characteristics such
because there was no control group at yr 6 and 7. as bone mineral density seem to be similar, and it is well known that
The method of ascertaining fractures was different in the two papers. fracture rate increases with age (2). Although it was not mentioned in
Other studies suggest that clinical vertebral represent only about a third our paper, spinal radiographs were scheduled for all participants at the
of total vertebral fractures. For example, in the Fracture Intervention end of yr 7; new vertebral fractures (including asymptomatic fractures
Trial (FIT), among women with a baseline vertebral fracture, the rate of identified on radiographs) were identified by the local investigators and
new clinical vertebral fractures was 1.7%/yr in women taking placebo reported as adverse experiences. This difference in methodology, from
and 0.8%/yr in women taking alendronate. The rate of morphometric the standardized morphometric assessment in yr 1–3, makes it impos-
fractures was 5%/yr in the placebo group and 2.7%/yr in the alendro- sible to perform valid comparisons of incidence rates and may explain,
nate group (3). In the other arm of the FIT study, where women did not in part, why the incidence seems to be higher in yr 6 and 7. In fact, if one
have a fracture at baseline, the morphometric vertebral fracture rate was looks at vertebral fractures reported as adverse experiences by inves-
0.9%/yr in the placebo group and 0.5%/yr in the alendronate group. The tigators during yr 1–3, the incidence of vertebral fractures in the patients
clinical vertebral fracture rate was only 0.2%/yr in placebo group and who received alendronate was 3.8%/yr, which is quite comparable with
0.1%/yr in the alendronate group (4). the incidence of 3.3%/yr seen during yr 6 and 7. This is consistent with
In the women taking long-term alendronate, the rate of vertebral the similar rate of height decrease observed during yr 1–3 (1.0 mm/yr)
fractures was at least three times higher during yr 6 and 7 than during and yr 6 and 7 (1.2 mm/yr).
yr 1–3, despite the fact that the bone density of the spine was increasing. We disagree with the hypothesis that “after (the first 4 years of use) the
The rates were also higher than predicted from the data in the FIT study. profound suppression of the bone formation rate may begin to have a
The discrepancy cannot be explained by selection of more severe cases negative effect.” Both bone resorption and formation markers are substan-
for the long-term study, because the baseline characteristics were sim- tially higher among postmenopausal osteoporotic women compared with
ilar. It also cannot be explained by aging of the population, because the premenopausal levels. As noted in our paper, bone turnover makers return
FIT participants were older. to within the premenopausal range during treatment with alendronate and
There is no doubt that alendronate increases bone strength and de- are not excessively suppressed. Treatment with alendronate (10 mg daily)
creases fracture rate during the first 4 yr of use, but after that the produced a stable (⬃55%) decrease in bone-specific alkaline phosphatase,
profound suppression of the bone formation rate may begin to have a a marker of bone formation, such that mean bone-specific alkaline phos-
negative effect. I hope these findings will lead to further studies of the phatase was 8.3 U/L, well within the normal range for premenopausal
long-term effect of alendronate on bone strength. women [mean (sd), 8.2 U/L (2.8); Ref. 3] after 7 yr of treatment. The
continued increase in spine bone mineral density through 7 yr is further
Susan M. Ott evidence that bone formation is not excessively suppressed.
Division of Metabolism Importantly, the annual incidence of nonvertebral fractures during yr
University of Washington 6 and 7 was not substantially different than during yr 1–3; this does not
Seattle, Washington 98195-6426 support any increase in fracture incidence with longer use of alendro-
nate. In conclusion, we do not believe there is any evidence of an increase
in fracture risk with continued use of alendronate up to 7 yr.
References
1. Tonino RP, Meunier PJ, Emkey R, et al. 2000 Skeletal benefits of alendronate:
Richard P. Tonino, Arthur Santora, and Philip D. Ross
7-year treatment of postmenopausal osteoporotic women. J Clin Endocrinol Good Health, P.C. (R.P.T.), South Burlington, Vermont 05403;
Metab. 85:3109 –3115. and Merck and Company, Inc. (A.S., P.D.R.), Rahway, New
2. Liberman UA, Weiss SR, Broll J, et al. 1995 Effect of oral alendronate on bone Jersey
mineral density and the incidence of fractures in postmenopausal osteoporosis.
N Engl J Med. 333:1437–1443. References
3. Black DM, Cummings SR, Karpf DB, et al. 1996 Randomised trial of effect of
alendronate on risk of fracture in women with existing vertebral fractures. 1. Black DM, Reiss TF, Nevitt MC, Cauley J, Karpf D, Cummings SR. 1993
Lancet. 348:1535–1542. Design of the Fracture Intervention Trial. Osteoporos Int. 3(Suppl 3):S29 –S39.
4. Cummings SR, Black DM, Thompson DE, et al. 1998 Effect of alendronate on 2. Hochberg M, Thompson D, Black D, Quandt S, Cauley J, Geusens P, for the
risk of fracture in women with low bone density but without vertebral fractures. FIT Research Group. 2000 The effect of alendronate on age-specific incidence
J Am Med Assoc. of key osteoporotic fractures. J Bone Miner Res. 5(Suppl 1):S552.
3. Garnero P, Shih WJ, Gineyts E, Karpf DB, Delmas PD. 1994 Comparison of

Received November 27, 2000. Address correspondence to: Susan M.

Ott, M.D., Division of Metabolism, Box 356426, University of Washing- Received January 22, 2001. Address correspondence to: Richard P.
ton, Health Science Building, Room BB545, 1959 NE Pacific Street, Se- Tonino, M.D., 368 Dorset Street, Suite 1, South Burlington, Vermont
attle, Washington 98195-6426. 05403.

1835
1836 LETTERS TO THE EDITOR JCE & M • 2001
Vol. 86 • No. 4

new biochemical markers of bone turnover in late postmenopausal osteoporotic Before analyzing the hypothetical disadvantages of AIs, estrogen
women in response to alendronate treatment. J Clin Endocrinol Metab. 79: actions at puberty in the male have to be underlined. First, estrogens act
1693–1700. on bone tissue in men. Epiphyseal closure and skeletal maturation do not
happen without estrogens (2, 3). The finding of eunuchoid proportions
Aromatase Inhibitors in Pubertal Boys: of the skeleton in estrogen-deficient men (2) suggests that estrogens are
Clinical Implications involved in the establishment of the proportions of the growing skeleton,
leading to eunuchoid body habitus (3). It seems that growth spurt also
To the editor:
is an estrogen-dependent process (2), although a conclusive consent on
I enjoyed the manuscript by Mauras et al. (1) because it opens this issue has not been reached, because a detailed growth chart of
interesting perspectives in the possible clinical use of aromatase estrogen-deficient men has never been provided (2, 3). Estrogens are
inhibitors (AIs) in the treatment of growth disorders, as previously needed to achieve the peak of bone mass (2, 3) during early to late
suggested by Grumbach and Auchus (2). AI constitutes a selective adolescence.
tool useful to obtain a delay in epiphyseal closure during skeletal Second, estrogens could modulate reproductive function in men.
development. The advantage of using AIs consists in a selective block Aromatization of androgens is required for a normal male sexual be-
of estrogen action without affecting biological effects due to andro- havior in mice, but much less is known about estrogens and human male

Downloaded from https://academic.oup.com/jcem/article/86/4/1842/2848945 by guest on 24 April 2024

gens per se (1). Thus, a pharmacological “decoupling” of androgen sexuality (4). Estrogens are necessary for a normal fertility in male
action from that of its estrogen metabolite is reached. In this view, AIs rodents. Presently, more research is required to prove whether the same
permit to extend longer the time of growth in boys with growth mechanisms operate in the human testis, although recent findings (e.g.
disorders (e.g. GH deficiency, idiopathic short stature), thus leading the presence of estrogen receptors and aromatase activity in the male
to an increased powerful of GH therapy and to a better final height. reproductive system) strongly support an estrogen role on human sper-
Presently, skeletal maturation could be restrained only by using matogenesis. In any case, estradiol is the major modulator of gonado-
GnRH analogs. AIs, however, do not affect the progression of pu- tropin secretion in normal men, acting both at pituitary (1, 2, 3, 5) and
bertal development because a block of estrogen action could occur hypothalamic levels (6), and estrogens are consequently involved in the
together with a normal advancement of virilization. Thus, AIs could control of testicular function.
prevent the psychological problems related to delayed puberty that Third, abnormalities of lipid and carbohydrate metabolism in men
often occur when GnRH analogs are used. with congenital estrogen deficiency (2, 3) suggest that estrogens could
The authors state that AIs are safe in young adolescents (1), and modulate some metabolic pathways, having a protective role also on the
this is the case if a short-term period of treatment is considered. By male cardiovascular system (2).
contrast, the issue becomes very complex if the treatment is carried Stated both well-documented and supposed effects of estrogens in
out for long time, as it is necessary in the case of short stature, being men, it should be considered that AIs could interfere with each of these
a great prolongation of the period available for growth desirable to physiological processes at puberty (Table 1). A delay in skeletal matu-
fill the gap of height. Several issues need to be clarified before AIs will ration could induce a disproportional growth, accounting for final eu-
be considered safe at puberty. Puberty, in fact, is the time for the nuchoid body proportions. A peak of bone mass lower than normal
children to mature sexual organs, to reach both adult proportions of could result from AI administration, predisposing the occurrence of
the skeleton and a complete adult psychosexual behavior. If unde- osteoporosis later during adulthood. Accordingly, spermatogenesis, the
sirable effects due to pharmacological treatments occurred in this psychosexual behavior, the cardiovascular system, and some metabolic
fragile phase of development some physiological functions could be pathways also could be impaired with an increased risk for health
compromised forever. during adulthood.
AIs directly decrease serum estradiol with an increase of the tes-
Received October 31, 2000. Address correspondence and requests for tosterone to estradiol ratio (1). The imbalance between circulating sex
reprints to: Vincenzo Rochira, M.D., Cattedra di Endocrinologia, Uni- steroids is more severe if we consider also the effects of AIs on
versità di Modena e Reggio Emilia, Policlinico di Modena, Via del Pozzo, gonadotropin feedback (6): higher LH levels cause a testosterone
71, 41100 Modena, Italy. E-mail: [email protected]. overproduction without a concomitant increase of estradiol (1), a

TABLE 1. Role of estrogen at puberty in the male and possible advantages and disadvantages of using AIs in young adolescent males

Estrogen actions at puberty in the male Advantages of using AIs Disadvantages of using AIs
Skeletal development
Epiphyseal closure and growth arresta Slowing down epiphyseal maturation delays A delay in epiphyseal closure and growth
growth arrest and prolongates growth. arrest could lead to disproportional
Potential benefits in growth disorders growth of some skeletal districts:
alone or in combination with r-hGH eunuchoid body proportions.
therapy.
Achievement of peak bone massa None Decreased peak of bone mass. Increased risk
of osteoporosis during adulthood.
Growth spurtb Absence of growth spurt? Continuing linear Reduced height velocity? Eunuchoid body
growth. proportions.
Establishment of final skeletal None Eunuchoid body proportions.
proportionsa
Feedback of gonadotrophinsa Decreased circulating estrogens Increased serum LH and Testosterone.
Severe imbalance of testosterone to
estradiol ratio. Macroorchidism?
Fertilityb None Risk of disrupted spermatogenesis.
Macroorchidism?
Behavioral changesb Continuing skeletal growth together with an Risk of an impairment of both the
unaffected virilization with a positive development and the achievement of a
psychological effect (especially when normal adult male psychosexual behavior.
compared with GnRH analogs).
Cardiovascular system and metabolic Uncertain Uncertain
pathwaysb
a
Certain.
b
Uncertain.
 LETTERS TO THE EDITOR 1837

hormonal pattern accounting for macro-orchidism in congenital es- 6. Hayes FJ, Seminara SB, Decruz S, Boepple PA, Crowley Jr WF. 2000 Aro-
trogen-deficient men. What are the consequences due to such a severe matase inhibition in the human male reveals a hypothalamic sit of estrogen
imbalance in circulating sex steroids during an important phase of feedback. J Clin Endocrinol Metab. 85:3027–3035.
7. Lee PA. 1999 Central precocious puberty. An overview of diagnosis, treatment,
development, which is puberty? Awaiting an answer to this question
and outcome. Endocrinol Metab Clin North Am. 28:901–918.
from future studies, some ethical perspectives should be considered.
The aim of this letter is to consider clinical implications related to the
pioneering use of these drugs in boys to promote a successful debate Author’s Response: Aromatase Inhibitors in Pubertal
among researchers on the possible inherent risks and benefits. The Boys—Clinical Implications
real impact of AIs on several physiological processes, which take
place at puberty, remains to be well established because a well- To the editor:
conducted, controlled clinical trial has never been performed. Pres- I appreciate the opportunity to reply to the letter of Dr. Rochira
ently, although the unique work performed by Mauras et al. (1) opens regarding the use of aromatase inhibitors in pubertal boys.
this novel form of therapy to study, it also leaves several questions In our work published in JCEM this summer (1), we report our results
unanswered. In that short-term treatment, the effects of AIs on skel- on the investigation of the intermediate metabolism of substrates, bone
etal maturation, fertility, and sexual behavior were not studied in calcium metabolism, body composition and strength in young males

Downloaded from https://academic.oup.com/jcem/article/86/4/1842/2848945 by guest on 24 April 2024

detail (1). Besides, caution is needed when data are extrapolated from given an aromatase inhibitor (Anastrozole) for 10 weeks and compared
a short-term treatment because a drug affecting physiological pro- the results with those of identical experiments in young men given a
cesses related to the developing body reveals its potential negative GnRH analog. In it, we clearly restrict our conclusions to the effects
effects when a long-term treatment is performed. What’s more, the observed over 10 weeks and only speculate to the potential use of this
issue is complicated by the fact that some adverse events could be compound in delaying epiphyseal fusion in short pubertal boys. I per-
detectable only later during adult life (e.g. infertility). sonally agree with the substance of the concerns expressed by Dr.
Again, what is the impact of AIs on local estrogen production? A block Rochira, and I echo the warning that the use of this and other similar
of estrogenic functions involves both autocrine and paracrine mechanisms aromatase inhibitors cannot be advocated until careful, methodical stud-
in each tissue in which aromatase is expressed. One of the tissue-specific ies, like the one reported, can be carried out in a large enough group of
implications related to the iatrogenic “decoupling” of androgen from es- patients and the data analyzed. Much more work needs to be done, and
trogen action in adolescent boys is represented by environmental changes such studies are presently underway.
in the testis (testosterone is higher with estradiol lower than normal). Are That estrogen is important in bone health is not questioned here.
the consequences on spermatogenesis occurring later during adulthood Actually, we used the model of aromatase inhibition with concomitant
predictable? The answer to this latter question is not so easy. Presently, in GH treatment in pubertal boys, and preliminary results show no det-
fact, we still do not know in detail the effects on male fertility of the rimental effects on bone mineralization after 1 yr of combined treatment
treatment with GnRH analogs during adolescence (7). (unpublished observations). More data are also needed on the issue of
To establish both the true rates of possible adverse events and the the impact of estrogen deficiency on sperm function and fertility. More
efficacy of AIs in male growth disorders, results from a long-term is written about the negative effects of estrogen exposure to the testis
treatment are anxiously awaited. Obviously, the following parame- than the effects of estrogen deficiency. Animal data clearly offer mixed
ters should be carefully monitored: sperm count, testicular size and results with both apparent disruption of spermatogenesis in mice lack-
morphology, height velocity, anthropometric measurements of sev- ing functional aromatase (2), and no difference in the number of Sertoli
eral skeletal districts, bone mineral density, and psychosexual be- cells, germ cells, the volume of seminiferous epithelium, tubule lumens,
havior. Changes in these parameters during long-term treatment of and interstitium between controls, and Anastrozole-treated rats after 1
adolescent boys will permit to establish in detail the side effects of yr (3). Extrapolating the effect of complete, lifelong aromatase blockade
these drugs. Presently, it is only possible to speculate on some adverse on sperm function from the published report of the man with this
events only by starting from our knowledge on the physiological aromatase deficiency (whose brother also had decreased sperm counts
estrogen role in males. Thus, also progress in the field of estrogen even though he had no aromatase gene mutation; Ref. 4), to the effects
physiology will probably permit us to forecast side effects due to the of timed pharmacological aromatase blockade may not be directly com-
experimental use of AIs in boys. parable. The latter achieves a 50% reduction in circulating estrogen
In conclusion, AI has to be regarded as an experimental regimen at concentrations and not a complete suppression. A critical question yet
puberty, being a therapeutic approach not miming physiological pu- to be answered, however, is not only if this intervention, like treatment
bertal changes of sex steroids. To transfer the use of AIs in pubertal boys with a GnRH analog, affects sperm function, but more importantly, is
into clinical practice, results from clinical trials on long-term treatment it reversible after discontinuation of treatment. This family of com-
need to be available and researchers need to clarify all of the androgen pounds, and the new class of estrogen receptor blockers, will require
actions that are really to be ascribed to estrogen. Accordingly, until much thorough and analytical study in the years to come. Only when proven
larger studies on these two issues have been done, physicians should efficacious and safe in long-term studies, can their potential usefulness
continue to exercise caution with regard to the potential for adverse in the pharmacological arsenal available to the clinican be fully assessed.
effects of AI administration to male adolescents.
Nelly Mauras
Vincenzo Rochira Division of Endocrinology
Department of Internal Medicine Nemours Children’s Clinic
University of Modena and Reggio Emilia Jacksonville, Florida 32207
41100 Modena, Italy

References References
1. Mauras N, O’Brien KO, Klein KO, Hayes V. 2000 Estrogen suppression in 1. Mauras N, O’Brian KO, Derter Klein K, Hayes V. 2000 Estrogen suppression
males: metabolic effects. J Clin Endocrinol Metab. 85:2370 –2377. in males: metabolic effects. J Clin Endocrinol Metab. 85:2370 –2377.
2. Grumbach MM, Auchus RJ. 1999 Estrogen: consequences and implications of 2. Robertson KM, O’Donnell L, Jones MC, et al. 1999 Impairment of spermat-
human mutations in synthesis and action. J Clin Endocrinol Metab. ogenesis in mice lacking a functional aromatase (Cyp19) gene. Proc Natl Acad
84:4677– 4694. Sci USA. 96:7986.
3. Faustini-Fustini M, Rochira V, Carani C. 1999 Oestrogen deficiency in men: 3. Turner KJ, Morley M, Atanassaua N, et al. 2000 Effect of chronic administration
where are we today? Eur J Endocrinol. 140:111–129. of an aromatase inhibitor to adult male rats on pituitary and testicular function
4. Carani C, Rochira V, Faustini-Fustini M, Balestrieri A, Granata A. R.M., 1999 and fertility. J Endocrinol. 164:225.
Role of estrogen in male sexual behaviour: insights from the natural model of
aromatase deficiency. Clin Endocrinol (Oxf). 51:517–525.
5. Rochira V, Faustini-Fustini M, Balestrieri A, Carani C. 2000 Estrogen re-
placement therapy in a man with congenital aromatase deficiency: effects of Received December 4, 2000. Address correspondence to: Nelly Mau-
different doses of transdermal estradiol on bone mineral density and hormonal ras, M.D., Division of Endocrinology, Nemours Children’s Clinic, 807
parameters. J Clin Endocrinol Metab. 85:1841–1845. Nira Street, Jacksonville, Florida 32207.
1838 LETTERS TO THE EDITOR JCE & M • 2001
Vol. 86 • No. 4

4. Carani C, Qin K, Simoni M, et al. 1997 Effect of testosterone and estradiol in TABLE 1. Classification of prolactinomas according to an
a man with aromatase deficiency. N Engl J Med. 337:95. ascending degree of aggressive behavior

Practical Classification of Prolactinomas for 1. Microprolactinoma (pituitary tumor of ⬍10 mm in diameter

Clinical Use associated with symptomatic hyperprolactinemia)
2. Macroprolactinoma (pituitary tumor of at least 10 mm in
To the editor: diameter associated with PRL levels usually over 200 ng/mL)
Notable discrepancies appear when analyzing the results from var- 3. Invasive prolactinoma (restricted to invasion of the carvernous
ious studies assessing treatment outcomes of patients with prolactino- sinuses space)
mas. Recently, based on a retrospective review of 46 male patients with 4. Aggressive prolactinoma (defined as invasive and DA resistant)
hyperprolactinemia managed with dopamine agonist (DA) therapy 5. Metastatic prolactinoma
alone, Pinzone et al. (1) conclude that PRL normalization is comparable
in male patients with macroprolactinomas vs. those with microprolacti-
nomas. This study also shows that the response rates to bromocriptine shown to exhibit a more severe clinical course both in human and in
(BRC) and cabergoline (CAB) were similar in both groups. This is rather animal models; and combined invasiveness and DA resistance usually

Downloaded from https://academic.oup.com/jcem/article/86/4/1842/2848945 by guest on 24 April 2024

surprising because CAB has been previously reported to be more ef- does not allow normalization of PRL levels, whatever the choice of
fective than BRC in normalizing PRL levels in patients with prolacti- therapy. The proposed classification may also help the clinician to plan
nomas (2) and because macroprolactinomas in men have been shown to the treatment and inform about the prognosis.
be more aggressive tumors with a higher prevalence rate of invasion Etienne Delgrange and Julian E. Donckier
toward cavernous sinus and resistance to BRC (3). Moreover, this is not Department of Internal Medicine and Endocrinology
in keeping with the results from the large-scale study by Verhelst et al. Université Catholique de Louvain at Mont-Godinne University
(2), who found that the probability of reaching normal PRL levels during Hospital
CAB treatment was significantly higher in patients of both sexes with a B-5530 Yvoir, Belgium
microprolactinoma than in patients with a macroprolactinoma. Thus,
what is wrong? Probably the absence of uniform classification of patients References
with presumed prolactinomas.
In the study by Pinzone et al. (1), inclusion criteria for “micropro- 1. Pinzone JJ, Katznelson L, Danila DC, Pauler DK, Miller CS, Klibanski A. 2000
lactinomas” (n ⫽ 12) were a serum PRL level more than 15 ng/mL and Primary medical therapy of micro- and macroprolactinomas in men. J Clin
Endocrinol Metab. 85:3053–3057.
either a normal computed tomography or magnetic resonance imaging
2. Verhelst J, Abs R, Maiter D, et al. 1999 Cabergoline in the treatment of hy-
scan of the pituitary or a tumor less than 1 cm in diameter. Obviously, perprolactinoma: a study in 455 patients. J Clin Endocrinol Metab. 84:2518 –2522.
this definition includes patients with idiopathic hyperprolactinemia. 3. Delgrange E, Trouillas J, Maiter D, Donckier J, Tourniaire J. 1997 Sex-related
Although sexual dysfunction was present in all cases, testosterone levels difference in the growth of prolactinomas: a clinical and proliferation marker
were normal in 3 of 11 patients. In this setting, it has been recommended study. J Clin Endocrinol Metab. 82:2102–2107.
to rule out the presence of a biologically inactive macroprolactinemia (4). 4. Guay AT, Sabharwal P, Varma S, Malarkey WB. 1996 Delayed diagnosis of
On the other hand, 4 of 11 patients remained testosterone deficient at psychological erectile dysfunction because of the presence of macroprolactine-
follow-up despite normalization of PRL levels in at least 2 of them. The mia. J Clin Endocrinol Metab. 81:2512–2514.
5. Delgrange E, Donckier J. 1997 Gonadal dysfunction in males with prolacti-
absence of restoration of gonadal function in men with small prolacti-
noma: from functional modification to irreversible damage? Eur J Endocrinol.
noma and mild testosterone deficiency is unusual (5), which suggests 136:630 (Letter).
that PRL was not the cause of hypogonadism and should prompt the 6. Feigenbaum SL, Downey DE, Wilson CB, Jaffe RB. 1996 Transsphenoidal
search for alternative diagnosis. Finally, it must be remembered that the pituitary resection for preoperative diagnosis of prolactin-secreting pituitary
finding of a “microadenoma” on scan in a patient with moderately adenoma in women: long term follow-up. J Clin Endocrinol Metab. 81:1711–
elevated PRL levels may not always indicate the presence of a prolacti- 1719.
noma (6). Thus, the small group studied probably represents a highly
heterogenous population. Nevertheless, normalization of PRL levels Medical Therapy of Prolactinomas in Men
(⬍15 ng/mL) was achieved in 83% (10 of 12 patients). Moreover, it must
be stressed that the two patients considered “DA resistant” had also To the editor:
nearly normal PRL values (between 15 and 17.5 ng/mL). Regarding Drs. Delgrange and Donckier have raised several issues concerning
macroprolactinomas, the study did not consider patients with tumors our study regarding primary dopamine agonist treatment in men with
invading the cavernous sinuses separately, although a preliminary re- prolactinomas. As the authors correctly state, our study shows that the
port indicates that invasive macroprolactinomas may be more fre- rate of PRL normalization in response to dopamine agonists is compa-
quently resistant to BRC than noninvasive ones (3). Highly aggressive rable in men with micro- or macroprolactinomas. These male subjects
macroprolactinomas often require multiple treatment modalities, but had not undergone surgery or radiation therapy and, thus, were treated
the authors excluded patients treated by surgery and/or adjunctive with dopamine agonists alone. The men presented with a spectrum of
radiotherapy. Thus, the more aggressive macroprolactinomas should radiographic and clinical findings, and the heterogeneity of prolacti-
logically be searched for in the large group of patients (n ⫽ 77) excluded noma presentation is well established. The diagnosis of “idiopathic
from the study. hyperprolactinemia” is somewhat problematic because such patients are
In conclusion, the study by Pinzone et al. (1) shows that PRL nor- often found to have lesions either at surgery or on magnetic resonance
malization is comparable in a group of male patients with idiopathic imaging scans as the sensitivity of such scans improves over time. As
hyperprolactinemia or presumed microprolactinoma and in a selected Drs. Delgrange and Donckier suggest, invasive macroprolactinomas
group of patients with macroprolactinomas not requiring surgery or may respond differently to dopamine agonists than noninvasive ones.
radiation therapy. To facilitate comparison of treatment results between However, the intent of the study was to present the results of men with
centers, additional studies should refer to a precise classification of prolactinomas treated with medical management alone without prese-
prolactinomas. We propose such a classification in Table 1, which mainly lecting a defined group. Prolactinoma responsiveness to therapy is often
relies on the following evidences: microprolactinomas have a benign difficult to judge based on proven biologic, endocrinologic, or pathologic
course and rarely progress in size over time; invasion of the cavernous classification. Invasive prolactinomas may respond to dopamine agonist
sinuses can be determined by radiological criteria and is associated to therapy with marked decreases in tumor size and normalization of PRL
higher proliferation activity; DA-resistant prolactinomas have been levels. In contrast, patients with microprolactinomas may show resis-
tance to dopamine agonist administration presumable due to differences
Received November 20, 2000. Address correspondence to: Etienne
Delgrange, M.D., Department of Internal Medicine and Endocrinology, Received December 1, 2000. Address correspondence to: Anne Kli-
Université Catholique de Louvain at Mont-Godinne University Hospi- banski, M.D., Neuroendocrine Unit, Massachusetts General Hospital, 55
tal, Solidarité Mutualiste Chrétienne, B-5530 Yvoir, Belgium. Fruit Street, Bulfinch 457B, 5-194 Joseph, Boston, Massachusetts 02114.
 LETTERS TO THE EDITOR 1839

in dopamine receptor concentration or binding characteristics. There- locomotion (walking and running) are concerned (6), we have clearly
fore, we categorized patients based on magnetic resonance imaging scan demonstrated that patients and healthy controls moved with the same
size. metabolic cost and efficiency of locomotion, provided that walking and
Drs. Delgrange and Donckier question our finding of similar efficacy running velocities are expressed as Froude number (which takes into
between dopamine agonists. The only way to determine whether one account the scale differences between the two groups). Nevertheless, it
dopamine agonist is more effective than another is through a prospec- is noteworthy that our patients with childhood-onset GHD were actually
tive, randomized, controlled study of both agents. Given the inherent unable to run at speeds higher than 8 km/h⫺1 for the time needed to
limitations of a retrospective chart review, one can only make conclu- reach a metabolic steady state (6). At this maximally attained speed their
sions on PRL normalization in treated patients. Clearly, prospective specific VO2 was about 25 mL/min⫺1 䡠 kg⫺1 and their average heart rate
studies are warranted to address this specific question. was about 180 beats/min⫺1, which, from their measured resting and
Finally, the natural history and therapeutic response to dopamine age-estimated maximum heart rate, would correspond to about 90% of
agonists of microprolactinomas has been well characterized in previous the maximum, strongly suggesting a a remarkable reduction in VO2 max
prospective as well as retrospective series in women. However, the also in childhood-onset patients with GHD.
natural history of untreated microprolactinomas in men has not been The considerable differences between patients with childhood-onset
well characterized due to the necessity of treating the low testosterone and adult-onset GHD (related to the duration, age of appearance, degree

Downloaded from https://academic.oup.com/jcem/article/86/4/1842/2848945 by guest on 24 April 2024

levels or sexual dysfunction that typically leads to presentation. Fur- of the disease) make mandatory for the future to analyze separately the
thermore, there are few data regarding the therapeutic responses to data recorded in so different clinical conditions.
dopamine agonist therapy in men. Certainly a better understanding of Moreover, the apparent discrepancy between the effects of rec-GH on
this disease in men would be of benefit to this increasing population of muscle strength (which remained unchanged) and submaximal aerobic
patients with this diagnosis. performance (significantly improved), as reported by Woodhouse et al.
(1), warrant further attention and should be the focus of future studies
Anne Klibanski and Laurence Katznelson investigating the effects of rec-GH on muscle.
Neuroendocrine Unit
Massachusetts General Hospital Alessandro Sartorio, Claudio Lafortuna, and Marco V. Narici
Boston, Massachusetts 02114 Istituto Auxologico Italiano (A.S.), IRCCS,
20145 Milano, Italy; Istituto Tecnologie Biomediche
Physical Performance in Growth Hormone- Avanzate (C.L.), CNR, 20090 Milano, Italy; and
Deficient Adults Department of Exercise and Sport Science (M.V.N.),
Metropolitan University of Manchester, Alsager, United Kingdom
To the editor: ST7 2HL
We have read with interest the paper by Woodhouse et al. (1), dealing
with the evaluation of submaximal aerobic performance in adults with References
adult-onset GH deficiency (GHD) before and after recombinant GH
1. Woodhouse LJ, Asa SL, Thomas SG, Ezzat S. 1999 Measures of submaximal
(rec-GH) treatment. performance evaluate and predict functional response to growth hormone (GH)
The authors point out how the perception of increased fatigue and treatment in GH-deficient adults. J Clin Endocrinol Metab. 84:4570 – 4577.
impaired physical performance experienced by these patients derives 2. Cuneo RC, Salomon F, Wiles CM, Hesp R, Sonksen PH. 1991 Growth hormone
from a markedly reduced maximal aerobic capacity (VO2max). Thus, treatment in growth hormone deficient adults. II. Effects on exercise perfor-
even the execution of ordinary daily activities requires a greater fraction mance. J Appl Physiol. 70:695–700.
of VO2max, imposing to the patients a discomfort out of proportion to 3. Nass R, Huber RM, Klauss V, Muller OA, Schopohl J, Strasburger CJ. 1995
their exercise task. Effect of growth hormone (hGH) replacement therapy on physical work ca-
The study by Woodhouse et al. (1) suggests that these changes may pacity and cardiac and pulmonary function in patients with hGH deficiency
acquired in the adulthood. J Clin Endocrinol Metab. 80:552–557.
be, in part, related to the lack of GH-insulin-like growth factor I actions 4. Bottinelli R, Narici M, Pellegrino MA, et al. 1997 Contractile properties and
on muscle mass, because the rec-GH treatment was associated with a fiber type distribution of quadriceps muscles in adults with childhood-onset
marked increase in skeletal muscle insulin-like growth factor I messen- growth hormone deficiency. J Clin Endocrinol Metab. 82:4133– 4138.
ger RNA, improving aerobic capacity indices and self-reported fatigue 5. Narici M, Ferretti G, Susta D, Faglia G, Sartorio A. 1999 Maximum anaerobic
during low-intensity exercise. performance of childhood-onset GH-deficient adults. Growth Horm IGF Res.
A picture thus emerges, in line with a common view (2, 3), of the 9:228 –235.
adult-onset GH-deficient patient as a poor performer, provided with a 6. Minetti AE, Ardigò LP, Saibene F, Ferrero S, Sartorio A. 2000 Mechanical and
below normal amount of contractile elements devoted to locomotor metabolic profile of locomotion in adults with childhood-onset GH deficiency.
Eur J Endocrinol. 142:35– 41.
tasks and to cardiorespiratory support of exercise.
Our group reached quite different conclusions in studies of short-
statured childhood-onset GH-deficient adults, which seem to have a
normally “proportioned” contracting machinery (4 – 6). In fact, al- Growth Hormone Deficiency and Physical Function
though absolute values of quadriceps strength and fiber cross-sec-
tional area (CSA) of patients were significantly lower than controls To the editor:
(4), differences disappeared once the absolute values were normal- We agree with Dr. Sartorio and his colleagues that the effects of rhGH
ized for quadriceps CSA and the subjects’ height. Normal muscle treatment on muscle size, strength, and anaerobic capacity in adult
twitch kinetics and fatiguability in the quadriceps of GH-deficient GH-deficient (GHD) patients are limited. We did not see marked im-
patients were fully consistent with the lack of a significant shift in provement in muscle strength despite an increase in mean fiber area
fiber type proportion (4). (both type I and type II fibers) following rhGH treatment as compared
In another recent study from our group (5), analyzing the anaerobic with placebo or baseline (1). However, the significant improvement in
performance of adults with GH deficiency, we concluded that both aerobic function that we have described (1) suggests that the capacity of
lactacid and alactacid maximum anaerobic power were similar in con- skeletal muscle to utilize oxygen for energy expenditure maybe en-
trols and patients when absolute differences (35%) were adequately hanced and/or there is an enhanced ability to deliver oxygen in response
normalized for body mass. These findings indicate that the GH-deficient to rhGH treatment. The improvement in cardiac function associated
patients’ ability to sustain anaerobic power and their rate of fatigue were with rhGH treatment (2) is certainly compatible with the latter possi-
comparable, in relative terms, with those of healthy controls. bility. Nevertheless, it may well be that the interface of oxygen delivery
Furthermore, as far as the aspects of mechanics and energetics of

Received March 2, 2000. Address correspondence to: Alessandro Received December 2, 2000. Address correspondence to: Shereen
Sartorio, M.D., Research Center for Growth Disorders, Italian Insti- Ezzat, M.D., Division of Endocrinology, Mount Sinai Hospital, Univer-
tute for Auxology, Via Ariosto 13, 20145 Milan, Italy. E-mail: sartorio@ sity of Toronto, 600 University Avenue, Suite 429, Toronto, Ontario,
auxologico.it. Canada M5G 1X5.
1840 LETTERS TO THE EDITOR JCE & M • 2001
Vol. 86 • No. 4

and utilization is critical in GHD subjects and that examination of muscle range associated with normal BFR could define its optimal level in
function alone may be less revealing in the GH deficiency state. hemodialysis patients, we proposed that their Biosource B-IRMA PTH
Dr. Sartario’s studies comparing childhood-onset GHD adults to age-, should be between 40 and 60 pg/mL (i.e. between 1 and 1.5 the upper
sex-, and activity-matched control subjects suggest that the reduced limit of normal to prevent adynamic bone disease). This proposed op-
muscle size and strength in GHD adults is a function of reduced body timal range is quite lower than that proposed with N-IRMA not only
size. Although this explanation is plausible for short-statured adults when expressed in absolute concentrations but also when expressed in
who have been GHD since childhood, this does not seem to be the folds of the upper limit of normal (ULN) since Quarles et al. (5) in 1992
situation for GHD adults who have a normally developed body size. and Wang et al. (6) in 1995 proposed in hemodialysis patients, 1.5–2.5 and
Moreover, while correction for reduced body size in adults with child- 2– 4 times the ULN, respectively, the upper threshold being a little higher
hood-onset GHD may negate the physiological differences seen in ab- in CAPD patients according to Wang et al. (Ref. 6; 5 times the ULN).
solute terms, the functional consequences cannot be so easily dismissed. Four hypotheses may be discussed to explain the discrepancy between
The fact remains that to perform activities of daily living (such as walk- our proposition for optimal PTH range and those of our American col-
ing, climbing stairs, rising from a chair, etc.) these GHD adults must leagues: 1) the role of a possible 7– 84 PTH fragment accumulation in renal
move a larger body mass, carrying a heavier load of fat with less met- failure; 2) the role of the antagonistic effect of 7– 84 PTH fragments on the
abolic machinery available to perform the work than their non-GHD bone remodeling effect of endogenous 1– 84 PTH; 3) the role of a difference

Downloaded from https://academic.oup.com/jcem/article/86/4/1842/2848945 by guest on 24 April 2024

counterparts. in aluminum overload leading to a higher bone resistance to endogenous
We agree with Dr. Sartario and his colleagues, who suggest that PTH remodeling effect; and 4) the role of calcitriol treatment.
future studies need to consider the onset of GHD. Such differences may 1. Accumulation of 7– 84 in renal failure is a possibility to consider
be a potential source contributing to the variable responses to GH treat- since Brossard et al. (7) showed with high-performance liquid-chroma-
ment on muscle and functional performance in GHD adults. tography determination that the proportion of non-1– 84 PTH fragments
in the intact PTH detected by N-IRMA was greater in hemodialyzed
Linda J. Woodhouse, Sylvia L. Asa, Scott G. Thomas, and patients than in healthy controls whether the patients are in stable
Shereen Ezzat condition (50 vs. 21%), have PTH stimulation by hypocalcemia (36 vs.
University of Toronto 13%), or PTH suppression by hypercalcemia (65 vs. 25%). However, in
Toronto, Ontario, Canada M5G 1X5 the article by John et al. (1) the proportion of non-1– 84 PTH recognized
by N-IRMA is not much greater in dialysis patients than in healthy
References
controls. Indeed, we calculated it by using the ratio of the ULN range
1. Woodhouse LJ, Asa SL, Thomas SG, Ezzat S. 1999 Measures of submaximal to be 52% in controls (65–31/65) and 59% in the eight dialysis patients
performance evaluate and predict functional response to growth hormone (GH) (688 –277/688) in basal condition and 50% after stimulation by citrate-
treatment in GH-deficient adults. J Clin Endocrinol Metab. 84:4570 – 4577. induced hypocalcemia. Only when PTH secretion was suppressed by
2. Lombardi G, Colao A, Ferone D, et al. 1997 Effect of growth hormone on cardiac calcium infusion was the proportion much higher (78%), this latter
function. Horm Res. 48:38 – 42.
phenomenon being easily explained by the well known increased ca-
tabolism of intact 1– 84 PTH to C-terminal fragments within the para-
Optimal Range of Plasma Concentration of True thyroid glands (8). Furthermore, Slatopolsky et al. (9) recently showed
1– 84 Parathyroid Hormone in Patients on that the proportion of non-1– 84 PTH fragment recognized with N-IRMA
Maintenance Dialysis was not lower in transplanted patients with better renal function than
in dialysis patients since the contrary is rather observed: 44% vs. 34%.
To the editor: Thus, the 1.5- to 5-fold higher PTH levels found necessary to maintain
We have read with interest the rapid communication of John et al. (1) normal BFR in the dialysis populations of Quarles et al. (5) and Wang
showing that the new Scandibodies laboratory immunoradiometric as- et al. (6) cannot be totally explained by a disproportionate retention or
say (IRMA) for measurement of intact PTH detects full-length human secretion of non-1– 84 fragments, since the proportion of these latter has
PTH but not amino-terminally truncated fragments, in contrast to most been documented only marginally higher in stable dialysis patients than
commercially available IRMAs for so-called intact PTH, which measure in nonuremic patients in the study of John et al. (Ref. 1; 59% vs. 52%).
also 7– 84 PTH as shown by Lepage et al. (2). 2. The hypothesis of Slatopolsky et al. (9) that the 7– 84 PTH fragments
In contrast to the Nichols Institute Allegro IRMA for intact PTH could, furthermore, have an antagonistic action vs. endogenous 1– 84
(N-IRMA), our PTH IRMA (B-IRMA) (Ref. 3; later commercialized by PTH on bone remodeling in uremic patients is, therefore, quite attractive.
Biosource for the last 10 yr) had the particularity to have a much lower This hypothesis is based on the observation that human 7– 84 PTH
normal range of 3– 40 pg/mL vs. 10 – 65 pg/mL for N-IRMA. The reason (hPTH 7– 84) did not increase cAMP production in osteoblast-like cells
for this discrepancy was not clear until 1992 when we compared the re- and antagonized in nonuremic rats the hypercalcemic and phosphaturic
covery of 7– 84 PTH fragments at a wide concentration range in our assay effects of hPTH 1– 84. However, to support the conclusion that this
and in that of Nichols Institute. The recovery of fragment (7– 84) was 100% antagonistic effect proven in nonuremic rats is the explanation for the
with N-IRMA and less than 1% in B-IRMA. Therefore, the nondetection of development of ABD in dialysis patients, one would like to document
7– 84 was a satisfactory explanation for the lower normal range. a higher proportion of non-1– 84 PTH measured by N-IRMA in patients
The clinical relevance of such difference is quite obvious for the with ABD than in those with osteitis fibrosa. Since according to John et
indirect evaluation of bone remodeling in hemodialysis patients. In 1991, al. (1) this proportion is rather lower (52% and 55%) in the two patients
Cohen Solal et al. (4) published in this journal the (1– 84) PTH levels with with ABD than in the six patients with osteitis fibrosa (61% and 59%),
the B-IRMA of 23 hemodialysis patients for whom a bone biopsy had we do not think that this explanation is quite satisfactory.
been performed. This allowed us to classify six of them in the adynamic 3. Therefore, we think that the main reason for the discrepancy be-
bone disease (ABD) group [low bone formation rate (BFR) with low tween our proposition regarding the optimal range for intact PTH is
osteoid thickness and low osteoblastic and osteoclastic surfaces], eight related to a greater bone resistance to the PTH remodeling effect in the
in the mild lesion group (with normal BFR and mild increase in osteo- American dialysis patients than in the Amiens dialysis patients, in re-
blastic and osteoclastic surface), and nine in the osteitis fibrosa group lation to their higher exposition to aluminum. Indeed, none of the
(with high BFR and increased osteoblastic and osteoclastic surfaces). All Amiens patients had traces of aluminum staining on their bone biopsy
the nine patients with osteitis fibrosa had PTH levels greater than or because they had never been exposed to aluminum either by the dia-
equal to 69 pg/mL (i.e. 1.7 the upper limit of normal). On the contrary, lysate or by the phosphate-binders, these latter having been definitively
all the six patients with ABD had PTH levels within the normal range. excluded since 1980. On the contrary, even though Quarles et al. (5)
Four of eight patients with mild lesion had PTH levels within the normal excluded patients with stainable aluminum, aluminum phosphate-bind-
range, three being above and one below. Considering that an intact PTH ers were still prescribed in their center. As regards the patients of Wang
et al. (6), only those with more than 25% of their osteoid/calcified bone
interfaces positive for aluminum were eliminated, so that most of them
Received October 9, 2000. Address correspondence to: Dr. A. had positive aluminum staining from 1–24% of their bone interfaces
Fournier, Hôpital sud Service de Néphrologie Médecine Interne du CHU because of previous aluminum-phosphate-binder ingestion. Since it is
d’Amiens, Avenue René Laënnec, 80054 Amiens, France. well documented that aluminum has a direct inhibitory effect on os-
 LETTERS TO THE EDITOR 1841

teoblasts, even when no stainable aluminum is present (10), it is very concentrations as determined by any particular assay method and the
likely that the difference in aluminum overload is the main explanation histological features of bone in patients with renal osteodystrophy. Such
for the differences in optimal range of PTH between our center and the information is essential to properly apply the results of PTH measure-
American ones. ments to the clinical management of patients with renal bone disease.
4. The role of treatment by calcitriol has to be eliminated since cal- Technical differences among currently available PTH assays probably
citriol treatment even without aluminum phosphate binder can further account, at least in part, for discrepancies among published results from
down-regulate PTH-PTHrP receptor (11) and decrease high bone re- different research laboratories. A number of other factors may also
modeling without a significant decrease of plasma intact PTH (12, 13). contribute, some of which have been cited by our European colleagues
This factor can, however, be eliminated as an explanation for our dis- in their correspondence. It seems unwarranted, however, to speculate
crepancy with the proposal of Wang et al. (6) since none of our patients retrospectively and without additional supporting information about
and of those of Wang et al. (6) were receiving calcitriol. the potential role of bone aluminum toxicity as a factor that contributes
The fact that in our unique population of patients without exposition to the skeletal resistance of PTH that may occur in renal failure.
to aluminum, or calcitriol, we documented a subpopulation with truly Most of the published work that describes the relationship between
normal secretion of PTH and low BFR, formally establishes for the first PTH levels in serum or plasma and bone histology in patients with renal
time that the uremic bone of these patients is not only resistant to the disease has been done using the immunoradiometric (IRMA) PTH assay

Downloaded from https://academic.oup.com/jcem/article/86/4/1842/2848945 by guest on 24 April 2024

hypercalcemic effect of exogenous PTH [shown by Massry et al. (14) in originally developed by Nussbaum et al. (1). This assay has been widely
1974] but also to the remodeling effect of endogenous PTH. used for more than a decade in clinical research and by commercial
medical diagnostic laboratories worldwide (Allegro PTH; Nichols In-
A. Fournier, M. E. Cohen Solal, R. Oprisiu, H. Mazouz, P.
stitute Diagnostics, San Juan Capistrano, CA). It remains the reference
Morinire, G. Choukroun, and R. Bouillon
standard for the noninvasive assessment of renal osteodystrophy be-
Nephrologie—CHU (A.F., M.E.C.S., R.O., H.M., P.M., G.C.),
Amiens, France; and LEGENDO University Hospital (R.B.) cause abundant supportive bone histopathological data are available
Leuven, Belgium (2–11). A similar abundance of clinical and laboratory information has
generally not been obtained in studies that have used other PTH-IRMAs
to assess patients with renal bone disease.
References A new, third-generation PTH-IRMA has recently been introduced
1. John M, Goodman W, Gao P, Cantor T, Salusky I, Jueppner M. 1999 A novel (12, 13). Although the relevant characterization data have not been
immunoradiometric assay detects full-length human PTH but not aminoter- published until now, the detection system used by Fournier et al. seems
minally truncated fragment: implication for PTH measurements in renal fail- to be very similar to that used in the studies by us (12) and by Slatopolsky
ure. J Clin Endocrinol Metab. 84:4287– 4290. et al. (13). This kind of PTH assay specifically detects full-length, bio-
2. Lepage R, Roy L, Brossard J, et al. 1998 A non-(1– 84) circulating parathyroid logically active 1-84 PTH and fails to cross-react with large amino-
hormone (PTH) fragment interferes significantly with intact PTH commercial
assay measurements in uremic samples. Clin Chem. 44:805– 809.
terminally truncated PTH-derived peptides such as 7-84 PTH that are
3. Bouillon R, Coopmans W, Degroote D, Radoux D, Eliard P. 1990 Immuno- retained in the plasma of patients with renal failure and are detected by
radiometric assay of parathyrin with polyclonal and monoclonal region- the Nichols’ PTH-IRMA (14, 15). The new IRMAs have the potential,
specific antibodies. Clin Chem. 36:271–276. therefore, to improve both precision and accuracy in the laboratory
4. Cohen Solal M, Sebert J, Boudailliez B, et al. 1991 Comparison of intact, mi- diagnosis of renal osteodystrophy.
dregion, and carboxy terminal assays of parathyroid hormone for the diagnosis Despite such considerations, the relationship between serum or
of bone disease in hemodialysed patients. J Clin Endocrinol Metab. 73:516 –524. plasma PTH levels as determined by third-generation PTH-IRMAs and
5. Quarles L, Lobaugh B, Murphy G. 1992 Intact parathyroid hormone overes-
the underlying histological subtype of renal osteodystrophy has yet to
timates the presence and severity of parathyroid-mediated osseous abnor-
malities in uremia. J Clin Endocrinol Metab. 75:145–150. be examined. Such information is essential to critically evaluate the
6. Wang M, Hercz G, Sherrard D, Segre G, Pei Y. 1995 Relationship between intact potential use of these new methods for diagnosing and monitoring the
PTH (1– 84) parathyroid hormone and bone histomorphometry parameters in evolution of renal bone disease. Until the data become available, guide-
dialysis patients without aluminum toxicity. Am J Kidney Dis. 26:836 – 844. lines about the concentrations of PTH in serum or plasma that corre-
7. Brossard JH, Cloutier M, Roy L, Lepage R, Gascon-Barr ZM, D’Amour P. 1996 spond to specific histological lesions of bone in patients with renal
Accumulation of a non-(1– 84) molecular form of parathyroid hormone (PTH) osteodystrophy must rely on available published information. Extrap-
detected by intact PTH assay in renal failure: importance in the interpretation
olation of results obtained using new PTH-IRMAs by comparing them
of PTH values. J Clin Endocrinol Metab. 81:3923–3929.
8. Kronenberg HM, Bringhurst FR, Segre GV, Potts JT. 1994 Parathyroid hor- to data obtained using previously developed PTH-IRMAs is insufficient
mone biosynthesis and metabolism. In: Bilezikian JP, Levine MA, Marcus R, to establish their value as a clinical diagnostic tool.
eds. Parathyroids basic and clinical concepts. New York: Raven Press; 125–137. The development of more precise and accurate methods for measur-
9. Slatopolsky E, Finch J, Clay P, et al. 2000 A novel mechanism for skeletal ing PTH and PTH-derived peptide fragments in serum or plasma is
resistance in uremia. Kidney Int. 58:753–761. likely to provide additional insight into the physiology and pathophys-
10. Fournier A, Moriniére P, Cohen Solal M. 1994 Adynamic bone disease in iology of parathyroid gland function and bone metabolism in patients
patients with uremia. Curr Opin Nephrol Hypertens. 3:396 – 410. with chronic renal failure. Such work may ultimately lead to improve-
11. Gonzalez E, Martin K. 1996 Bone cell response in uremia. Semin Dialysis.
9:339 –346.
ments in diagnostic precision and more refined guidelines for clinical
12. Hamdy NAT, Kanis J, Beneton M, et al. 1995 Effect of alfacalcidol on natural management. Future recommendations should be founded, however, on
course of bone disease in mild to moderate renal failure. Br Med J. 310:358 –363. abundant biochemical and histopathological data obtained from pa-
13. Goodman W, Ramirez J, Beun T, et al. 1994 Development of adynamic bone tients who have been well-characterized demographically. Only then
in patients with secondary hyperparathyroidism after intermittent calcitriol can the value of these technical advances in assay methodology be fully
therapy. Kidney Int. 46:1160 –1166. determined.
14. Massry S, Coburn J, Lee D, et al. 1973 Skeletal resistance to parathyroid hormone
in renal failure: study in 105 human subjects. Ann Intern Med. 78:357–364. William G. Goodman, Markus R. John, Harald Jüppner, and
Isidro B. Salusky
Parathyroid Hormone-Immunoradiometric Assays as Departments of Pediatrics and Medicine (W.G.G., I.B.S.),
Noninvasive Predictors of Renal Osteodystrophy: University of California at Los Angeles School of Medicine, Los
Angeles, California 90095; and Endocrine Unit, Massachusetts
The Need for Bone Histomorphometric Validation
General Hospital and Harvard Medical School (M.R.J., H.J.),
To the editor: Boston, Massachusetts 02114
The comments of Fournier et al. emphasize the importance of ade-
quately documenting the relationship between serum or plasma PTH References
1. Nussbaum SR, Zahradnik RJ, Lavigne JR, et al. 1987 Highly sensitive two-site
Received January 22, 2001. Address correspondence to: Harald Jüp- immunoradiometric assay of parathyrin, and its clinical utility in evaluating
pner, M.D., Endocrine Unit, Massachusetts General Hospital, Wellman patients with hypercalcemia. Clin Chem. 33:1364 –1367.
501, Boston, Massachusetts 02114. 2. Salusky IB, Ramirez JA, Oppenheim WL, Gales B, Segre GV, Goodman WG.
1842 LETTERS TO THE EDITOR JCE & M • 2001
Vol. 86 • No. 4

1994 Biochemical markers of renal osteodystrophy in pediatric patients un- a difference seen for bioavailable testosterone not adjusted for BMI? Non
dergoing CAPD/CCPD. Kidney Int. 45:253–258. sex hormone-binding globulin-bound testosterone includes albumin-
3. Pletka PG, Strom TB, Hampers CL., et al. 1976 Secondary hyperparathyroid- bound testosterone, which may account for up to 20% of total testos-
ism in human kidney transplant recipients. Nephron. 1976, 17:371–381.
terone (4). In view of the age of the subjects, if one is to adjust for BMI,
4. Sherrard DJ, Hercz G, Pei Y, et al. 1993 The spectrum of bone disease in
end-stage renal failure—an evolving disorder. Kidney Int. 43:436 – 442. surely one should also adjust for variations in serum albumin. Caution
5. Pei Y, Hercz G, Greenwood C, et al. 1993 A plastic osteodystrophy in diabetic in adjusting for covariates in reporting clinical findings has recently been
patients. Kidney Int. 44:159 –164. highlighted (5). Presentation of the unadjusted data from this study
6. Hercz G, Pei Y, Greenwood C, et al. 1993 A plastic osteodystrophy without would be very informative.
aluminum: the role of “suppressed” parathyroid function. Kidney Int. The authors report an assay sensitivity of 0.07 nmol/L for testoster-
44:860 – 866. one, however, standard RIAs for testosterone are notoriously inaccurate
7. Wang M, Hercz G, Sherrard DJ, Maloney NA, Serge GV, Pei Y. 1995 Rela- at the lower end of the female range. We are surprised that such distinct
tionship between intact 1-84 parathyroid hormone and bone histomorpho-
differences were observable in values below 0.5 nmol/L. Furthermore,
metric parameters in dialysis patients without aluminum toxicity. Am J Kidney
Dis. 26:836 – 844. the actual RIA used for measuring total testosterone is not stated. The
8. Pei Y, Hercz G, Greenwood C, et al. 1995 Risk factors for renal osteodystrophy: most pronounced difference in total testosterone increase was reported
a multivariant analysis. J Bone Miner Res. 10:149 –156. to occur between the 6th and 7th decades, yet for the 6th decade there

Downloaded from https://academic.oup.com/jcem/article/86/4/1842/2848945 by guest on 24 April 2024

9. Mathias RS, Salusky IB, Harmon WH, et al. 1993 Renal bone disease in were only 29 women with intact ovaries and seven after oophorectomy.
pediatric patients and young adults treated by hemodialysis in a childrens Hence, we caution a 35% increase on a mean value of less than 0.5
hospital. J Am Soc Nephrol. 12:1938 –1946. nmol/L in so few women becoming a generalizable fact. Furthermore,
10. Goodman WG, Ramirez JA, Belin TR, et al. 1994 Development of adynamic reporting of values adjusted for BMI gives us little feel for what the
bone in patients with secondary hyperparathyroidism after intermittent cal-
actual normal values were, and this is exacerbated in Table 2 in which
citriol therapy. Kidney Int. 46:1160 –1166.
11. Salusky IB, Kuizon BD, Belin T, et al. 1998 Intermittent calcitriol therapy in values were adjusted for both age and BMI.
secondary hyperparathyroidism: a comparison between oral and intraperito- Although ovarian stromal hypertrophy and hyperplasia may some-
neal administration. Kidney Int. 54:907–914. times persist or develop after menopause, probably secondary to ele-
12. John MR, Goodman WG, Gao P, Cantor TL, Salusky IB, Jüppner H. 1999 A vated LH levels and individual sensitivity, resulting in increased tes-
novel immunoradiometric assay detects full-length human PTH but not ami- tosterone production (6), other researchers have not found this to be the
no-terminally truncated fragments: implications for PTH measurements in general rule (7). That “. . . levels in women more than 70 yr of age or 20
renal failure. J Clin Endocrinol Metab. 84:4287– 4290. yr post menopause were comparable to premenopausal levels” is also
13. Slatopolsky E, Finch J, Clay P, et al. 2000 A novel mechanism for skeletal
questionable. In Fig. 3 in the article, the dotted lines indicate the mean
resistance in uremia. Kidney Int. 58:753–761.
14. Brossard JH, Cloutier M, Roy L, Lepage R, Gascon-Barre M, D’Amour P. 1996 testosterone level for premenopausal women as being between 0.6 – 0.7
Accumulation of a non-(1– 84) molecular form of parathyroid hormone (PTH) nmol/L. However, according to Sinha-Hikim et al. (8), who defined the
detected by intact PTH assay in renal failure: importance in the interpretation range of total and free testosterone levels during the normal menstrual
of PTH values. J Clin Endocrinol Metab. 81:3923–3929. cycle in healthy women, the mean testosterone level across the cycle was
15. Lepage R, Roy L, Brossard JH, et al. 1998 A non-(1– 84) circulating parathyroid 1.20 ⫾ 0.69 nmol/L and the mean free testosterone was 12.8 ⫾ 5.59
hormone (PTH) fragment interferes significantly with intact PTH commercial pmol/L. These levels were measured using an equilibrium dialysis
assay measurements in uremic samples. Clin Chem. 44:805– 809. method, and the sensitivity of the total testosterone assay was increased
to 0.008 nmol/L. The results from that study are sound and have been
What Are “Normal” Testosterone Levels for Women? used as references in many other studies. Furthermore, these testoster-
one levels are in agreement with those reported in other literature (9, 10).
To the editor: If we accept the data of Sinha-Hikim et al. (8) as reference levels, then
The cross-sectional study by Laughlin et al. (1), comparing endoge- the testosterone levels in the “intact postmenopausal women” in the
nous sex hormones levels with hysterectomy and oophorectomy status, study by Laughlin et al. (1) are, in fact, very low and do not actually
chronological age, and years since menopause in older women, rein- increase back to premenopausal levels (1.20 ⫾ 0.69 nmol/L). Although
forces the possible adverse endocrine sequelae of oophorectomy in older in Results the authors state (referencing a 1997 paper) these premeno-
women. However, we question some aspects of the data analysis and, pausal values as being from the same laboratory as used for this study,
hence, the clinical significance and interpretation of some of the said they are inconsistent with the widely accepted range. Were these also
findings. BMI adjusted? Either way, the comparison is not meaningful. Of note,
The inference from the manuscript is that, in women with intact the actual citation for the testosterone levels for premenopausal women
ovaries, there is an increase in total testosterone levels after the time of in Fig. 3 is that of an article by S. E. Bulun and E. R. Simpson, a study
menopause and with increasing age reaching premenopausal levels of levels of aromatase cytochrome P450 transcripts in adipose tissue of
concomitant with falling androstenedione levels. That testosterone women. The bioavailable values reported are ⬃10-fold greater than the
should rise while androstenedione falls is incongruous because adrenal normal range for free testosterone. Reporting of free testosterone would
androgen precursor levels fall linearly with age (2) and, following the have been far more meaningful.
menopause, peripheral conversion of androstenedione becomes a major The role of androgens in women is becoming increasingly more
source of circulating testosterone (3). We would appreciate the authors recognized and established. Certainly, the use of androgens, particularly
proposing a hypothesis for this incongruity. testosterone, has been shown to influence life aspects, such as mood,
The reason for adjusting total testosterone and bioavailable testos- women’s general well being and restoration of sexual desire. However,
terone for body mass index (BMI) in postmenopausal women is not there is limited data establishing normal androgen values for women of
justified. Because weight may vary significantly with increasing age and differing ages, to enable us to define those with “androgen deficiency.”
years since menopause, and the authors suggest that sex hormone- It is, therefore, necessary to highlight the incongruencies and short-
binding globulin concentrations were inversely related to BMI, one is left comings of the paper by Laughlin et al. (1), and the need for larger
to wonder if without adjustment for BMI whether any variation oc- prospective studies to establish the variations in testosterone levels in
curred. The authors do not report a relationship between either testos- women with age.
terone or androstenedione and BMI, making the need for the adjustment
most curious. Adjustment for BMI is not clinically informative, because Susan Davis and Jane Tran
the absolute bioavailable circulating levels are the values of physiolog- The Jean Hailes Foundation
ical significance in terms of direct androgen action and as precursors for Clayton, Victoria 3168, Australia
extragonadal estrogen biosynthesis in tissues such as bone. Indeed, was

References
Received April 27, 2000. Address correspondence to: Susan Davis, 1. Laughlin G, Barrett-Connor E, Kritz-Silverstein D, Von Muhlen D. 2000
M.D., Ph.D., Department of Epidemiology and Preventive Medicine, The Hysterectomy, oophorectomy, and endogenous sex hormone levels in older
Jean Hailes Foundation, Melbourne, 291 Clayton Road, Clayton, Victoria women: the Rancho Bernado Study. J Clin Endocrinol Metab. 85:645– 651.
3168, Australia. 2. Zumoff B, Rosenfeld RS, Strain GW. 1980 Sex differences in the 24 hour mean
 LETTERS TO THE EDITOR 1843

plasma concentrations of dehydroisoandrosterone (DHA) and dehydroisoan- assay.) Although the sensitivity of the Sinha-Hikim total testosterone
drosterone sulfate (DHAS) and the DHA to DHAS ratio in normal adults. J Clin assay is very high, total testosterone levels were apparently measured
Endocrinol Metab. 51:330 –334. directly in the dialysate. Equilibrium dialysis would not eliminate cross-
3. Judd HL, Lucas WE, Yen SSC. 1994 Effect of oophorectomy on circulating
reacting substances, which may account for higher (1.04 ⫾ 0.76 nmol/L)
testosterone and androstenedione levels in patients with endometrial cancer.
Am J Obstet Gynecol. 118:793–798. early follicular phase testosterone levels with this assay compared with
4. Luthold WW, et al. 1993 Serum testosterone fractions in women: normal and the Yen assay. In our view, steroid hormone assays may yield differing
abnormal clinical states. Metabolism. 42:638 – 643. results for a variety of reasons. Direct comparisons are valid only when
5. Assmann S, Pocock S, Enos L, Kasten L. 2000 Subgroup analysis and other- all samples are measured with the same assay techniques, as was the case
(mis)uses of baseline data in clinical trials. Lancet. 355:1064 –1069. in our paper.
6. Procope B. 1968 Studies on the urinary excretion, biological effects and origin Drs. Davis and Tran correctly point out that the finding of an increase
of estrogens in postmenopausal women. Acta Endocrinol (Copenh). 135:1– 86. in testosterone levels with postmenopausal aging was, in large part,
7. Labrie F, Belanger A, Cusan L, Gomez J-L, Candas B. 1997 Marked decline
based on lower levels among only 29 intact women in the 50- to 59-yr
in serum concentrations of adrenal C19 sex steroid precursors and conjugated
andrgoen metaboliltes during aging. J Clin Endocrinol Metab. 82:2396 –2402. age range. As stated in the article, “although age-specific levels were
8. Sinha-Hikim I, et al. 1998 The use of a sensitive equilibrium dialysis method based on relatively few women for the youngest decade, the similarity
for the measurement of free testosterone levels in healthy, cycling women and of hormone patterns after stratification by decade and by years since

Downloaded from https://academic.oup.com/jcem/article/86/4/1842/2848945 by guest on 24 April 2024

in human immunodeficiency virus-infected women. J Clin Endocrinol Metab. menopause supports the validity of the age-related associations.”
83:1312–1318. Among the intact women, 125 were less than 20, 136 were 20 –30, and
9. Judd HL, Yen SSC. 1973 Serum and androstenedione and testosterone levels 119 were more than 30 yr menopausal. Testosterone levels were 25%
during the menstrual cycle. J Clin Endocrinol Metab. 36:475– 481. higher among intact women more than 20 yr postmenopause compared
10. Abraham GE. 1974 Ovarian and adrenal contribution to peripheral androgens
with intact women less than 20 yr postmenopause.
during the menstrual cycle, J Clin Endocrinol Metab. 39:340 –346.

Postmenopausal Testosterone 3) A plausible biological explanation for the findings.

To the editor: The increase in circulating testosterone levels with postmenopausal
aging is likely to be ovarian in origin. Testosterone levels did not increase
Drs. Davis and Tran raise important issues that can be summarized with age among the oophorectomized women, and the metabolism and
to three main areas: 1) the effect of adjusting results for body mass index interconversion of androgens in women is not altered after menopause
(BMI); 2) the validity of the testosterone measurements and their relation (6). In recent studies, ovarian vein testosterone levels were related to the
to premenopausal levels; and 3) a plausible biological explanation for the degree of stromal hyperplasia in 18 postmenopausal women (7) and
findings. correlated positively with age in 52 women aged 42– 69 yr (8). Post-
menopausal ovarian androgen production is thought to be at least par-
1) The effect of adjusting results for BMI. tially gonadotropin driven. Gonadotropin receptors have been identi-
fied in the stroma of menopausal ovaries (9), and peripheral steroid
The relation between hormone levels and chronological and meno-
levels in postmenopausal women increase after hCG stimulation (10)
pausal age was analyzed 1) without adjusting for covariates, 2) adjusting
and decline after administration of long-acting GnRH agonists (11, 12).
for BMI, and 3) adjusting for BMI, smoking, exercise, and alcohol con-
Thus, stimulation by elevated gonadotropins could account for in-
sumption. In the interest of conserving space, only the BMI-adjusted
creased ovarian testosterone synthesis in older women. Although post-
results were presented. The association (or absent association) with age
menopausal gonadotropin levels tend to decline with advancing age
did not differ in unadjusted analyses, and the total and bioavailable
(13), they remain well above premenopausal levels. It is possible that
testosterone levels were virtually identical to the BMI-adjusted values
prolonged exposure to high gonadotropin stimulation may up-regulate
shown in the figures. The minimal impact of BMI on bioavailable tes-
ovarian responses to LH. To our knowledge, no published studies ad-
tosterone in this study may reflect the narrow range of BMI among the
dress this possibility.
Rancho Bernardo cohort (95% confidence interval, 23.8 –24.5 kg/m2), in
The absence of a postmenopausal increase in androstenedione levels
whom BMI did not vary across age groups.
with age similar to the increase in testosterone is not surprising. As
shown in previous studies and confirmed in this one, the postmeno-
2) The validity of the testosterone measurements and their pausal ovary contributes 40% of circulating testosterone, but only 10 –
relation to premenopausal levels. 15% of circulating androstenedione. Thus, an increase in ovarian an-
drostenedione production would be masked by the many-fold greater
Total testosterone levels were measured by a RIA developed in the adrenal contribution.
steroid research laboratory of Samuel Yen, M.D., during the 1970s using Our conclusion that testosterone levels return to the premenopausal
an antibody produced at the University of California–San Diego. An range in older, intact postmenopausal women needs clarification. The in-
important feature of steroid hormone measurements in this laboratory ternal (unpublished) normal laboratory values for this assay are 0.72 ⫾ 0.34,
is sample purification by solvent extraction and celite column chroma- 1.02 ⫾ 0.48, and 0.93 ⫾ 0.38 nmol/L for the early follicular, mid-cycle, and
tography before RIA. The testosterone assay sensitivity of 0.07 nmol/L luteal phases, respectively. Because testosterone levels do not undergo
is comparable with published values (0.03– 0.05 nmol/L) for assays cyclic changes in postmenopausal women, our results suggest that older,
using similar sample purification steps (1–3). Because of the high sen- as well as younger, postmenopausal women are relatively testosterone
sitivity and specificity of this method, measured values are lower than deficient when compared with younger cycling women, with levels similar
direct serum measurements with commercial kits. However, the testos- to the early follicular phase. We appreciate this opportunity to expand and
terone levels in this study were similar to levels in other large studies clarify our findings. Clearly, interpretation of cross-sectional data are sub-
of postmenopausal women using comparable assay methods (1–3). ject to error. Prospective studies are needed to define with certainty the
The authors also question the validity of the testosterone level cited pattern of changing testosterone levels in older women.
for premenopausal women. Levels for postmenopausal women were
compared with a recently published value (0.63 ⫾ 0.23 nmol/L) from the
Yen laboratory for 32 women sampled in the early follicular phase of the Gail Laughlin and Elizabeth Barrett-Connor
menstrual cycle (4). A 1979 study (5) using the same assay reported a Department of Family and Preventive Medicine
24-h mean level of 0.69 ⫾ 0.04 nmol/L for early follicular phase women, University of California–San Diego
therefore, the assay/laboratory has good internal consistency. (The 1973 La Jolla, California 92093-0607
Judd and Yen article cited by Davis and Tran used an earlier, less specific
References
Received December 18, 2000. Address correspondence to: Elizabeth 1. Hankinson SE, Willett WC, Manson JE, et al. 1998 Plasma sex steroid hor-
Barrett-Connor, M.D., Department of Family and Preventive Medicine, mone levels and risk of breast cancer in postmenopausal women. J Natl Cancer
University of California–San Diego, Stein Clinical Research Sciences Inst. 90:1292–1299.
Building, 9500 Gilman Drive, La Jolla, California 92093-0607. 2. Cauley JA, Lucas FL, Kuller LH, et al. 1999 Elevated serum estradiol and
1844 LETTERS TO THE EDITOR JCE & M • 2001
Vol. 86 • No. 4

testosterone concentrations are associated with a high risk for breast cancer. features we would associated with isolated spontaneous hypogo-
Ann Intern Med. 130:270 –277. nadism. The sole object was, of course, the preservation of the un-
3. Meldrum DR, Davidson BJ, Tataryn IV, Judd HL. 1981 Changes in circulating broken voice, and in the only recorded postmortem examination of
steroids with aging in postmenopausal women. Obstet Gynecol. 57:624 – 628.
a castrato the dimensions of the larynx were strikingly small with the
4. Arroyo A, Laughlin GA, Morales AJ, Yen SCC. 1997 Inappropriate gonad-
otropin secretion in polycystic ovary syndrome: influence of adiposity. J Clin vocal cords the length of a female high soprano (4).
Endocrinol Metab. 82:3728 –3733. With reference to the effect of castration on longevity, a study I carried
5. Lachelin GCL, Barnett M, Hopper BR, Brink G, Yen SSC. 1979 Adrenal out on a group of castrated compared with intact singers born at similar
function in normal women and women with the polycystic ovary syndrome. times showed that life-span was unchanged (2).
J Clin Endocrinol Metab. 49:892– 898.
6. Longcope C, Baker S. 1993 Androgen and estrogen dynamics: relationships John S. Jenkins
with age, weight, and menopausal status. J Clin Endocrinol Metab. 76:601– 604. Emeritus Professor of Endocrinology
7. Sluijmer AV, Heineman MJ, Koudstaal J, et al. 1998 Relationship between St. George’s Hospital Medical School
ovarian production of estrone, estradiol, testosterone, and androstenedione SW17 ORE London, United Kingdom
and the ovarian degree of stromal hyperplasia in postmenopausal women.
Menopause. 5:207–210. References
8. Ala-Fossi S-L, Maenpaa J, Aine R, Punnonen R. 1998 Ovarian testosterone

Downloaded from https://academic.oup.com/jcem/article/86/4/1842/2848945 by guest on 24 April 2024

secretion during perimenopause. Maturitas. 29:239 –245. 1. Wilson JD, Roehrborn C. 1999 Long-term consequences of castration in men:
9. Kobayashi M, Nakano R, Shima K. 1989 Immunohistochemical localization lessons from the Skoptzy and the eunuchs of the Chinese and Ottoman courts.
of pituitary gonadotropins and estrogen in human postmenopausal ovaries. J Clin Endocrinol Metab. 84:4324 – 4331.
Acta Obstet Gynecol Scand. 72:76 – 80. 2. Jenkins J. 1998 The voice of the castrato. Lancet. 351:1877–1880.
10. Greenblatt RB, Colle ML, Mahesh VB. 1976 Ovarian and adrenal steroid 3. Haböck F. 1927 Die Kastraten und ihre Gesangkunst. Stuttgart: Deutsche Ver-
production in the postmenopausal woman. Obstet Gynecol. 47:383–387. lags-Anstalt; 238.
11. Dowsett M, Cantwell B, Lala A, Jeffcoate SL, Harris AL. 1988 Suppression 4. Tandler J, Grosz S. 1909 Über den Einfluss der Kastraten auf den Organismus.
of postmenopausal ovarian steroidogenesis with the luteinizing hor- Arch entw Mech Org. 27:35.
mone-releasing hormone agonist goserelin. J Clin Endocrinol Metab.
66:672– 677.
12. Sluijmer AV, Heineman MJ, De Jong FH, Evers JLH. 1995 Endocrine activity Citation of “Validation” References for Sphygmocor-
of the postmenopausal ovary: the effects of pituitary down-regulation and Based Estimates of Central Aortic Blood Pressure
oophorectomy. J Clin Endocrinol Metab. 80:2163–2167.
13. Kwekkeboom DJ, de Jong FH, van Hemert AM, Vandenbroucke JP, Valk- To the editor:
enburg HA, Lamberts SWJ. 1990 Serum gonadotropins and ␣-subunit
Elsheikh et al. (1) describe how they attempted to assess central
decline in aging normal postmenopausal women. J Clin Endocrinol Metab.
70:944 –950. arterial hemodynamics in women with Turner’s syndrome.
The whole thrust of their paper suggests that the noninvasive central
arterial assessments of “aortic” blood pressure that they adopted have
Long-Term Consequences of Castration in Men been validated— but no data to support this assertion are actually of-
To the editor: fered or cited.
Indeed, in the measurements section of their paper, the authors write:
I read with great interest the paper “Long-term consequences of “The aortic pulse waveform was derived from the radial pulse wave-
castration in men: lessons from the Skoptzy and the eunuchs of the form by using a validated transfer function (4) implemented in the
Chinese and Ottoman courts,” published in December 1999 (1). How- Sphygmocor software. Measurements using this method have been
ever, the authors are at fault when, in the last paragraph, they state that shown to correspond closely to intraarterially recorded waves (2– 4)” (1).
the “so-called” castrati singers were a heterogenous group containing However, this is not correct.
only a few singers who had their testes removed or crushed. As I have Rather, examination of the three references cited from 1989 (2, 3) and
previously indicated (2), castration was, indeed, performed on a very 1996 (4) make no mention of the Sphygmocor/BPAS device (PWV Med-
large scale in Italy specifically to satisfy the ever increasing demand for ical, Sydney, Australia) or its radial artery generalized transfer function
the unique castrato voices, reaching its peak in the mid 18th century (GTF). Indeed, none of them show that radial artery pulse waveforms
when, according to the great authority on the castrato, Franz Haböck, as can be used to derive central aortic pulse waveforms noninvasively. One
many as 4000 castrations were carried out annually (3). It is, unfortu- article overviews the basic technique of applanation tonometry (2), and
nately, not true that only those with extraordinary singing ability were the other two references show how carotid artery tonometry [a technique
chosen for the procedure. Only a few reached the top opera houses, not used by Elsheikh et al. (1)] can provide noninvasive assessments of
many sang in ordinary church choirs, and there were many whose ascending aortic blood pressure (3, 4).
mutilation was to no purpose. The castrati nearly all came from poor The literature sometimes cited as having “validated” the radial artery
families in Italy, and fathers permitted the operation in the hope (only GTF used within the Sphygmocor/BPAS system has recently been over-
too often misplaced) that fame and fortune achieved by being a great viewed elsewhere (5). None of the above three articles even feature,
castrato would come to them and their poverty-stricken families. because they have nothing to do with the radial artery GTF.
The castrati singers flourished over a century and a half before the This issue is of importance because in the 7⫹ yr since the report
medical studies on the Skoptzy or the Chinese and Ottoman eunuchs describing the original radial artery GTF was published (6), and the U.S.
were carried out. Moreover, despite its popularity in the 18th century, Patent for the technique was granted (7), there has been an absolute
the practice of producing castrati was strictly illegal, according to the paucity of validation work with the approach reported in the scientific
canon law of the Roman Church, and every effort was made to keep literature. This lack of validation studies is particularly noticeable when
the identity of the operators and their origin vague. Euphemisms the measurements have to be calibrated noninvasively, as in the study
were used to account for the existence of a particular castrato such by Elsheikh et al. (1). Indeed, what little “validation” data for noninva-
as disease of the testes or accidental injury— being gored by a wild sive use of the Sphygmocor are available in the literature, based on 20
boar was a favorite reason! It is clear, however, that unlike the groups patients (8), suggests the errors associated with the method for the
discussed in the present paper, only the testes were removed; total estimation of central aortic blood pressure to be substantial. In fact, these
removal of all the genitalia was never carried out. Only a handful of errors are much greater than those permitted by the American Associ-
castrati lived on into the early years of the 20th century, in the Vatican, ation for Medical Instrumentation under guidelines endorsed by the U.S.
but 18th century descriptions of the appearances of the castrati in- Food and Drugs Administration for noninvasive blood pressure mea-
clude, in addition to the powerful high-pitched singing voice of a suring devices (5, 9). This has led to the conclusion that, at this time, there
strange quality, tallness of stature, smooth beardless face, rounding
of the hips, enlargement of the breasts, and a tendency to obesity—all
Received October 18, 2000. Accepted November 14, 2000.
Address correspondence and requests for reprints to: Eldon D. Leh-
Received February 29, 2000. Address correspondence to: John S. Jen- mann, Department of Imaging (MR Unit), National Heart and Lung Insti-
kins, M.D., 40 Hampstead Way, NW11 7JL London, United Kingdom. tute, Imperial College of Science, Technology and Medicine, Sydney Street,
E-mail: [email protected]. London SW3 6NP, United Kingdom. E-mail: [email protected].
 LETTERS TO THE EDITOR 1845

are no accurate or well validated methods for measuring blood pressure 13. Cameron J. 1999 Estimation of arterial mechanics in clinical practice and as a
noninvasively in vivo in the aorta (10). research technique. Clin Exp Pharmacol Physiol. 26:285–294.
Notwithstanding this, driven it would seem by commercial claims (7) 14. Cameron JD, McGrath BP, Dart AM. 1998 Use of radial artery applanation
tonometry and a generalized transfer function to determine aortic pressure
rather than evidence-based medicine, the Sphygmocor GTF approach now
augmentation in subjects with treated hypertension. J Am Coll Cardiol.
is being tried in all manner of disease states completely unrelated to the 14 32:1214 –1220.
subjects who provided data at cardiac catheterization for the original gen- 15. Cameron JD, McGrath BP, Dart AM. 1999 Use of radial artery applanation
eration of the radial artery GTF (6). In this respect, even if one believes the tonometry. Reply [Letter]. J Am Coll Cardiol. 34:952.
claims that a single radial artery GTF can be used accurately and robustly 16. Bulpitt CJ, Rajkumar C, Cameron JD. 2000 Central aortic blood pressure
to measure central aortic blood pressure noninvasively in all subjects of all measurements [Letter]. J Hum Hypertens. 14:531.
ages, heights, weights, and blood pressures, both on and off medication
(claims that are disputed because no evidence has, in fact, been published Homocysteine, Folate, Vitamin B12, and
in the peer-reviewed literature), it seems completely implausible that this Transcobalamins in Patients Undergoing Successive
will work for all disease states as well.
Connected with this, a recent query about where is the evidence that
Hypo- and Hyperthyroid States
the Sphygmocor system can be used to accurately and noninvasively To the editor:

Downloaded from https://academic.oup.com/jcem/article/86/4/1842/2848945 by guest on 24 April 2024

predict central aortic blood pressure in patients with another endocrine
disorder (diabetes; Ref. 11) was not able to offer any evidence or data to In the March 2000 issue of this journal, Lien et al. (1) reported a transient
support the claims that were being made (12). Furthermore, a search on increase in both plasma homocysteine and serum cholesterol during short-
Medline up to September 2000 using “Sphygmocor” and “sphygmo- term iatrogenic hypothyroidism, which may confer increased cardiovas-
cardiography” as search terms confirms the absolute paucity of valida- cular risk. The observations of Lien et al. (1) are in accordance with the
tion data that are available in the literature for the technique. results of our previous preliminary study (2).
Other researchers have noticed this as well and have independently We now repeat plasma homocysteine, serum folate, and vitamin B12 in
raised separate fundamental concerns about the use of the method, patients with successive hypo- and hyperthyroid states. Cobalamin-bind-
especially for the determination of the augmentation index (13–15) and ing proteins, transcobalamins, were also determined to explain changes in
central aortic blood pressure (16). vitamin B12 concentrations. Forty-five patients [age, 44.5 ⫾ 12.3 yr (23–78);
Unfortunately, citing papers from elsewhere in the field, which have not sex ratio M/F, 11/34] who had undergone total thyroidectomy for well-
involved the Sphygmocor’s radial artery GTF, as “validation” references, differentiated thyroid carcinoma were studied 4 weeks after the withdrawal
may make it appear on a cursory inspection that there have, in fact, been of thyroı̈dal hormone therapy and then 14 weeks after the resumption of
more validation studies using the device than are really the case. treatment to suppress the thyrotropin concentration. Hypo- and thyrotoxic
Given all the above, Elsheikh et al. (1) may wish to exercise caution states were evidenced by TSH concentrations (Table 1). Total EDTA-plasma
in making claims about the “validity” of the noninvasive approach they homocysteine, serum folate, cobalamin, serum total B12 binding capacity,
have adopted, which at present remains completely unproven, espe- apo-haptocorrin, and apo-transcobalamin II were measured by methods
cially in endocrine disease. described previously (4).
Homocysteine concentrations were significantly higher in hypo- than in
hyperthyroid state (mean increase, 5.3 ⫾ 4.6 ␮mol/L; Table 1). Furthermore,
Eldon D. Lehmann moderate hyperhomocysteinemia was observed for 10 of 45 patients (22%)
Department of Imaging (MR Unit) with hypothyroidism [range, 17.5–27.2 ␮mol/L; reference range, 6.0 –16.0
National Heart and Lung Institute ␮mol/L (5)]. Homocysteinemia was normal in all patients in the hyper-
Imperial College of Science, Technology and Medicine thyroid state. On univariate analysis, homocysteine was inversely related
London SW3 6NP, United Kingdom to folate (Rho, ⫺0.33; P ⫽ 0.02). Our results suggested that folate levels may
account in determining homocysteine as we confirmed the observation (1)
References of a moderate decline in serum folate in hypothyroid state, with a P value
1. Elsheikh M, Bird R, Casadel B, Conway GS, Wass JAH. 2000 The effect of at the limit of significance (P ⫽ 0.07). Vitamin B12 was significantly higher
hormone replacement therapy on cardiovascular hemodynamics in women in the hypothyroid state than in the hyperthyroid state, as found by Lien
with Turner’s syndrome. J Clin Endocrinol Metab. 85:614 – 618. et al. (1). Transcobalamin levels were determined attempting to explain
2. Kelly R, Hayward C, Ganis J, Daley J, Avolio A, O’Rourke M. 1989 Non- changes in vitamin B12 concentrations. Except in two cases, apo-hapto-
invasive registration of the arterial pressure pulse waveform using high- corrin was low in all patients in the two states whereas apo-transcobalamin
fidelity applanation tonometry. J Vasc Med Biol. 1:142–149. II was not significantly decreased in hypothyroid patients. Vitamin B12 was
3. Kelly R, Karamanoglu M, Gibbs H, Avolio A, O’Rourke M. 1989 Noninvasive not correlated to transcobalamins. We concluded that the increased vitamin
carotid pressure wave registration as an indicator of ascending aortic pressure.
B12 observed in hypothyroid state could not be explained by changes in
J Vasc Med Biol. 1:241–247.
4. Chen CH, Ting CT, Nussbacher A, et al. 1996 Validation of carotid artery transcobalamins.
tonometry as a means of estimating augmentation index of ascending aortic Lien et al. (1) have compared the results at 2-week intervals during two
pressure. Hypertension. 27:168 –175. phases. We compared the results between two states, hypo- and hyper-
5. Lehmann ED. 2000 Regarding the accuracy of generalized transfer functions thyroidism, in a larger cohort of patients. The increase in homocysteine
for estimating central aortic blood pressure when calibrated non-invasively concentrations during hypothyroidism may be explained by changes in
[Letter]. J Hypertens. 18:347–349. folate status and also by modifications in enzymes involved in homocys-
6. Karamanoglu M, O’Rourke MF, Avolio AP, Kelly RP. 1993 An analysis of the teine metabolism, distribution or clearance (1, 6), and/or by concurrent
relationship between central aortic and peripheral upper limb pressure waves
changes in renal function (1). Changes in activities of 5,10-methylenetet-
in man. Eur Heart J. 14:160 –167.
7. O’Rourke MF. 1993 Method for ascertaining the pressure pulse and related rahydrofolate reductase and methionine synthase have been reported dur-
parameters in the ascending aorta from the contour of the pressure pulse in ing both hyper- and hypothyroid states in an animal model (7). Data about
the peripheral arteries. United States Patent No. 5,265,011. normalization of hyperhomocysteinemia with levothyroxine are conflict-
8. Takazawa K, O’Rourke MF, Fujita M, et al. 1996 Estimation of ascending ing. Normalization was obtained after 3–9 months in a study of 14 patients
aortic pressure from radial arterial pressure using a generalised transfer func- (8), but failed after 2 months in a cohort of 14 patients (9).
tion. Z Kardiol. 85(Suppl 3):137–139. In conclusion, homocysteine was increased in 22% of our patients in the
9. Association for the Advancement of Medical Instrumentation. 1992 Elec- hypothyroidism stage. This mild hyperhomocysteinemia was rather ex-
tronic or automated sphygmomanometers. Arlington, VA: AAMI.
plained by a modification of folates status, with a mild decrease of blood
10. Lehmann ED. 2000 Aortic pulse-wave velocity versus pulse pressure and
pulse-wave analysis [Letter]. Lancet. 355:412. concentration and a negative correlation between folates and homocys-
11. Lehmann ED. 2000 Where is the evidence that radial artery tonometry can be teinemia, than by a modification of vitamin B12 status and transport.
used to accurately and non-invasively predict central aortic blood pressure in
patients with diabetes? [Letter] Diabetes Care. 23:869 – 870.
12. Brooks B, Molyneaux L, Yue DK. 2000 Where is the evidence that radial artery Received November 12, 2000. Address correspondence to: Françoise
tonometry can be used to accurately and noninvasively predict central aortic Barbé, Laboratoire de Biochimie A, Centre Hospitalier Universitaire de
blood pressure in patients with diabetes? Response to Lehmann [Letter]. Di- Nancy, Hôpital de Brabois—Hôpital d’Adultes, Rue du Morvan, 54511
abetes Care. 23:870 – 871. Vandoeuvre-les-Nancy, France. E-mail: [email protected].
1846 LETTERS TO THE EDITOR JCE & M • 2001
Vol. 86 • No. 4

TABLE 1. Homocysteine, folate, vitamin B12, and transcobalamin results in patients undergoing successive hypo- and hyperthyroid
states

n Hypothyroid state Hyperthyroid state P Reference range

Serum TSH (mIU/L) 45 83.9 ⫾ 52.5 0.03 ⫾ 0.04 ⬍0.001 0.25– 4.0
Serum T4 (pmol/L) 26 3.3 ⫾ 1.5 19.5 ⫾ 4.6 ⬍0.001 11.0 –25.0
Plasma homocysteine (␮mol/L) 45 12.4 ⫾ 5.8 7.1 ⫾ 3.5 ⬍0.001 6.0 –16.0
Serum folate (nmol/L) 24 13.4 ⫾ 6.0 15.5 ⫾ 6.4 0.075 6.0 –36.0
Serum vitamin B12 (pmol/L) 24 385 ⫾ 102 313 ⫾ 71 ⬍0.001 180 – 810
Serum total B12 binding capacity (pmol B12/L) 27 464 ⫾ 70 477 ⫾ 67 0.35
Serum apo-haptocorrin (pmol B12/L) 27 102 ⫾ 24 102 ⫾ 33 0.98 148 – 627
Serum apo-transcobalamin II(pmol B12/L) 27 362 ⫾ 57 374 ⫾ 50 0.069 295–1328
Results are expressed as the mean ⫾ SD. Paired t test was used to evaluate plasma and serum data between hypothyroid and hyperthyroid
states. Two-tailed at P ⬍ 0.05 was reported as statistical significance.

Downloaded from https://academic.oup.com/jcem/article/86/4/1842/2848945 by guest on 24 April 2024

Françoise Barbé, Marc Klein, Abalo Chango, Sophie Frémont, concentration during the hypothyroid phase were more strongly
Philippe Gérard, Georges Weryha, Jean-Louis Guéant, and associated with changes in serum creatinine than in folate, suggesting
Jean-Pierre Nicolas a renal mechanism (5).
Departments of Biochemistry A, INSERM XR308 (F.B., A.C., The covariation between tHcy and serum cholesterol was equally
S.F., P.G., J.-P.N.) and INSERM 0014 (P.G., J.-L.G.). and strong, but a mechanistic link between homocysteine and cholesterol
Department of Endocrinology (M.K., G.W.), Centre Hospitalier metabolism is not readily apparent. The strong relation between tHcy
Universitaire de Nancy, Hôpital de Brabois–Hôpital d’Adultes and cholesterol should remind us that covariations certainly do not
54511 Vandoeuvre-les-Nancy, France prove causality.
Thyroid status has a profound influence on a variety of biochem-
References ical processes (6 – 8), of which some may have secondary effects on
homocysteine metabolism. For example, thyroid hormones markedly
1. Lien EA, Nedrebo BG, Varhaug JE, Nygard O, Aakvaag A, Ueland PM. 2000 affect riboflavin metabolism, mainly by stimulating flavokinase
Plasma total homocysteine levels during short-term iatrogenic hypothyroidism.
and thereby the synthesis of flavin mononucleotide and flavin adenine
J Clin Endocrinol Metab. 85:1049 –1053.
2. Barbé F, Klein M, Chango A, Weryha G, Nicolas JP, Leclère J. 1999 Hypothy- dinucleotide (FAD; Refs. 7 and 8). Conceivably, these metabolic changes
roidism increases plasma homocysteine concentrations. J Endocrinol Invest. 22:28. may affect homocysteine metabolism, because flavin mononucleotide and
3. Ubbink JB, Vermaak WJH, Bissbort S. 1991 Rapid high-performance liquid FAD serve as cofactors for enzymes involved in the metabolism of vitamin
chromatographic assay for total homocysteine levels in human serum. J Chro- B6, cobalamin, and folate (9). Among these enzymes, the FAD-dependent
matogr. 565:441– 446. methylenetetrahydrofolate reductase should be considered, because this
4. Namour F, Olivier JL, Abdelmouttaleb I, et al. Transcobalamin codon 259 enzyme is recognized as a possible mediator of changes in tHcy level
polymorphism in HT29 and Caco-2 cells and in Caucasians: relation to transco-
according to riboflavin status (9).
balamin and homocysteine concentration in blood. Blood. In press.
5. Nicolas JP, Chango A. 1997 Dérégulation du métabolisme de l’homocystéine In conclusion, the nice study of Barbé et al. provides longitudinal data
et conséquences pour le système vasculaire. Bull Acad Natle Méd. 181:313–331. in a large number of patients and brings strong support that thyroid status
6. Nedrebo BG, Ericsson UB, Nygard O, et al. 1998 Plasma total homocysteine is an important determinant of plasma tHcy and affects serum folate levels.
levels in hyperthyroid and hypothyroid patients. Metabolism. 47:89 –93. However, the question of the mechanism(s) behind hyperhomocysteinemia
7. Nair CPP, Viswanathan G, Noronha J. 1994 Folate-mediated incorporation of in hypothyroid patients remains to be elucidated.
ring-2-carbon of histidine into nucleic acids: influence of thyroid hormone.
Metabolism. 43:1575–1578. Ernst A. Lien, Steinar Hustad, Bjørn G. Nedrebø, Ottar Nygård,
8. Hussein WI, Green R, Jacobsen DW, Faiman C. 1999 Normalization of hy- and Per M. Ueland
perhomocysteinemia with l-thyroxine in hypothyroidism. Ann Intern Med.
LOCUS for Homocysteine and Related Vitamins
131:348 –351.
9. Catargi B, Parrot-Roulaud F, Cochet C, Ducassou D, Roger P, Tabarin A. 1999 Armauer Hansens hus
Homocysteine, hypothyroidism, and effects of thyroid hormone replacement. University of Bergen
Thyroid. 9:1163–1166. N-5021 Bergen, Norway

Total Plasma Homocysteine in Hypo- and References

Hyperthyroidism: Covariations and Causality
1. Nedrebø B, Ericsson U-B, Ueland PM, Refsum H, Lien EA. 1994 Plasma levels
To the editor: of the atherogenic amino acid homocysteine in hyper- and hypothyroid pa-
tients. Eur J Clin Endocrinol. 130(Suppl 2):47.
There are consistent reports that patients with hypothyroidism have 2. Nedrebø BG, Ericsson U-B, Nygård O, et al. 1998 Plasma levels of the athero-
elevated total homocysteine (tHcy) in plasma and that tHcy is reduced genic amino acid homocysteine in hyper- and hypothyroid patients. Metab-
following therapy with T4 (1–5). The withdrawal of thyroid hormone olism. 47:89 –93.
replacement therapy before radioscintigraphy of thyroidectomized pa- 3. Catargi B, Parrot-Roulaud F, Cochet C, Ducassou D, Roger P, Tabarin A. 1999
tients provides controlled conditions for the study of changes in tHcy under Homocysteine, hypothyroidism, and effect of thyroid hormone replacement.
variable thyroid status. This strategy was adopted in a recent study from Thyroid. 9:1163–1166.
our group (5) and in the study on 45 patients reported by Barbé et al. 4. Hussein WI, Green R, Jacobsen DW, Faiman C. 1999 Normalization of hy-
perhomocysteinemia with L-thyroxine in hypothyroidism. Ann Intern Med.
Barbé et al. found lower serum folate in the hypothyroid compared
131:348 –351.
with the hyperthyroid state, and they observed the usual inverse 5. Lien EA, Nedrebø BG, Varhaug JE, Nygård O, Aakvaag A, Ueland PM. 2000
relationship between folate and tHcy. From this observation they Plasma total homocysteine levels during short-term iatrogenic hypothyroid-
inferred that the changes in tHcy may be explained by altered folate ism. J Clin Endocrinol Metab. 85:1049 –1053.
status or by a modification of the activity of folate-metabolizing 6. Danforth Jr E. 1983 The role of thyroid hormones and insulin in the regulation
enzymes, as has previously been suggested (2). Like Barbé et al., we of energy metabolism. Am J Clin Nutr. 38:1006 –1017.
found lower levels of serum folate during short-term iatrogenic hy- 7. Lee SS, McCormick DB. 1985 Thyroid hormone regulation of flavocoenzyme
pothyroidism (5). We observed, however, that the changes in tHcy biosynthesis. Arch Biochem Biophys. 237:197–201.
8. Rivlin RS. 1979 Hormones, drugs and riboflavin. Nutr Rev. 37:241–245.
9. Hustad S, Ueland PM, Vollset SE, Zhang Y, Bjørke-Monsen AL, Schneede
Received January 22, 2001. Address correspondence to: Ernst A. Lien, J. 2000 Riboflavin as a determinant of plasma total homocysteine: effect mod-
M.D., Division of Pharmacology, University Hospital of Bergen, N-5021 ification by the methylenetetrahydrofolate reductase C677T polymorphism.
Bergen, Norway. Clin Chem. 46:1065–1071.