The Journal of Clinical Endocrinology & Metabolism Vol. 83, No. 10 3436-3438
Copyright © 1998 by The Endocrine Society
III. Which Testosterone Assay Should Be Used In Older Men?1
S. R. Plymate
Veterans Affairs HCSPS and
Geriatric Research Education and Clinical Center
University of Washington School of Medicine
Seattle, Washington 98195
 |
Introduction
|
|---|
IN WOMEN, the determination of
age-associated gonadal failure is usually signaled as ovulation ceases
with the cessation of the menstruation. A decrease in estrogen
production after menopause is also evident by the onset of physical
symptoms such as hot flushes, dry vagina, etc. that are easily
discerned by the woman. However, measurement of estrogen levels during
the follicular phase of the menstrual cycle preceding menopause and
again after menopause does not necessarily reflect the magnitude of
menopausal symptoms or marked increase in gonadotropins.
The determination of age-associated hypogonadism (so-called andropause)
is not as clear in men as it is in women. It is possible that
andropause may not even occur in the majority of men, if it exists at
all, as an independent entity outside of a disease-associated decline
in plasma concentrations of testosterone. As men age, there is a
decline in serum total testosterone levels that begins about the age of
40 yr. In cross-sectional studies the annual decline in total and
"free" testosterone is 0.4% and 1.2%, respectively (1). However,
it should be noted that these studies are cross-sectional, and although
the serum testosterone level declines with age, the mean value for a
group of disease-free men is rarely below the lower limits of the
normal range for most laboratories, even into the ninth decade of life
(2). The diagnosis of an andropause in an individual man is not
necessarily further substantiated by the measurement of serum
gonadotropins. In men, careful measurements of luteinizing hormone (LH)
and FSH made in pooled blood samples reveals a small but significant
increase in serum gonadotropin levels (3, 4, 5). However, unlike in women,
the increase in blood gonadotropin levels in older, otherwise healthy
men is not above the normal range for healthy adult men.
Testosterone circulates in the blood 98% bound to protein. In men,
approximately 40% of the binding is to the high affinity sex hormone
binding globulin (SHBG), association constant (Ka) 1 x
109 L/mol, and approximately 60% bound weakly to albumin
(Ka = 3 x 104 L/mol) (6, 7). A number of
measurements for testosterone are available from clinical laboratories,
and it is important to understand the differences among them (7, 8, 9, 10, 11).
The measurement of serum testosterone or "total" testosterone is
usually performed by a radioimmunoassay and measures free plus protein
bound testosterone. "Free" or dialyzable testosterone measurements
are estimates of the fraction of testosterone in blood that is not
bound to protein. These assays require determination of the percentage
of unbound testosterone by a dialysis procedure, estimation of total
testosterone, and the calculation of free testosterone. Free
testosterone can also be calculated if total testosterone, SHBG, and
albumin concentrations are known (8). Kits are available for
determination of free testosterone without dialysis and are used to
provide a free testosterone measurement by many laboratories.
Unfortunately, these measurements are often inaccurate, especially when
the testosterone levels are low and SHBG levels are elevated (9, 12). A
third measurement of testosterone commonly made is that of
"bioavailable" or non-SHBG bound testosterone (10). This
measurement determines the amount of testosterone not bound to SHBG and
includes that which is nonprotein bound and weakly bound to albumin. It
is this fraction that supposedly is "bioavailable" to tissues. This
measurement takes into account that SHBG is precipitated by a lower
concentration of ammonium sulfate, 50%, than albumin. Thus by
precipitating a serum sample with 50% ammonium sulfate and measuring
the testosterone value in the supernate, non-SHBG bound or bioavailable
testosterone is measured. This fraction of testosterone can also be
calculated if total testosterone, SHBG, and albumin levels are
known.
To evaluate the clinical utility of serum testosterone measurements, it
is important to identify whether or not the normal laboratory range for
testosterone is valid in the older man, or if there are changes in
testosterone dynamics with age that may affect interpretation of
testosterone levels. Serum testosterone levels begin to decline in
normal healthy men on no medications, in the mid- to late-thirties.
This decline is linear into the nineties, at a rate of 0.4%/year. If
men with chronic medical illnesses are evaluated (illness defined as
hypertension, heart disease, diabetes, but not a hospitalized patient),
it appears that the same age-associated decline in serum testosterone
occurs, but that at any age the level for a man with a chronic illness
is 10–15% below that of healthy age-matched men. In addition to this
decline with age in total testosterone, there is an associated increase
in sex hormone binding globulin (SHBG), the circulating high affinity
binding protein for testosterone (2, 5, 13). Therefore, as man ages,
the total testosterone level decreases, but the serum binding of
testosterone increases. This increase in testosterone binding results
in a "free" or bioavailable testosterone level that decreases to a
greater extent than total testosterone. As a result, the availability
of the free active form of testosterone in the serum is further reduced
compared with total testosterone. Clinically, free T is decreased with
age to a greater extent than total T. However, it is unclear whether
free T is the only bioavailable fraction of testosterone for most of
the tissues. Using various model systems and depending on the tissue
and rate of blood flow through the tissue, the total concentration of
circulating testosterone available to the organs can be calculated (6).
The problem arises when studies attempt to determine whether certain
age- and gender-associated problems, such as impotence, are better
correlated with measurements of free or bioavailable testosterone than
total testosterone (10). However, when men with impotence and low
levels of bioavailable testosterone receive testosterone replacement
therapy, there is no greater improvement in their impotence compared
with similarly treated impotent men with normal levels of testosterone
(14). Thus, although bioavailable and free testosterone measurements
are lower in men with possible androgen related symptoms, they
apparently do not respond to androgen replacement. These data point to
two problems, 1) symptoms that we may feel are androgen related are
probably multifactorial and not reliable measures of androgen action.
Secondly, the measurement of testosterone that relies on an assessment
of protein-bound T may not be a reflection of a deficiency of T, but
may be the result of a change in the binding protein by a problem
independent of the androgen state of the man. For example, a decrease
in SHBG may occur in type II diabetes mellitus as a result of an
increase in insulin or insulin-like growth factor I levels (15, 16).
This reduction in SHBG level, unrelated to androgen deficiency and
associated with decreased total androgen in diabetic patients, may be
coincidentally related to impotence, when in fact, the impotence is due
to diabetes mellitus. Clinical situations such as these, point out the
difficulty in assessing androgen status when there is no good
independent marker of androgen action that can be used in
vivo. In addition, there are no well designed clinical trials that
have indicated that one method of testosterone measurement is better
than any another at defining a group of men who are androgen deficient
and will respond to testosterone replacement therapy. Therefore, either
until appropriate clinical studies are published or until a good marker
of testosterone action becomes available, a total serum testosterone is
as good a measurement, and less expensive, than a more complex and
labor intensive measure of free or bioavailable testosterone. In
addition, one must be aware that assays that purport to measure free
testosterone without using a dialysis method or without calculating
free testosterone levels based on separate measurements of testosterone
and SHBG, can markedly underestimate by as much as 100% the true free
hormone level (9). This sort of assay error can result in an incorrect
overestimation of the degree or prevalence of androgen deficiency.
An additional factor that needs to be taken into account when looking
at the differences in testosterone levels between young and elderly
men, is the difference in circadian variation. Young men clearly have a
circadian rhythm of testosterone, with the zenith in the morning
between 0600 h and 0800 h and the nadir in the late afternoon
between 1700 h and 1800 h (5, 17). In elderly men, the
circadian rhythm, if detected, is much flatter and tends not to be
consistent between individuals. Clinically, this means that the
difference in blood testosterone levels between young and elderly men
can be readily detected when measurements are made in the morning
during the zenith of the testosterone rhythm, but the difference may be
lost if samples are collected in the late afternoon. However, even
though the difference in testosterone secretory patterns is clearly
altered between young and elderly men, the clinical significance of
this altered secretory pattern has not yet been determined. With the
development of testosterone delivery systems that can mimic the
circadian variation in testosterone secretion, clinical studies may be
designed that can address the biological significance of the circadian
variation in testosterone.
In summary, based on the data currently available, the measurement of
total blood testosterone is the most appropriate test to determine
whether an older individual is hypogonadal or not. If the total
testosterone level is less than 7.0 nm/L (2 ng/mL), the individual
should be considered hypogonadal and, regardless of age, further
evaluation should be the same as for any other hypogonadal individual.
The question that cannot be answered with our present data base,
however, is whether or not otherwise healthy men or those with chronic
illnesses with testosterone levels of less than 10.5 nm/L (3 ng/mL) but
greater than 7.0 nm/L (2 ng/mL) are hypogonadal, and whether these
hypogonadal men would benefit from replacement, especially if their
gonadotropin levels are normal. Even though older men have reduced
muscle mass, decrease in body hair, and decrease in bone mass, there
have been no conclusive data demonstrating that these findings are the
result of age-associated reduction in testosterone levels. In fact, the
argument can just as well be made that even if an elderly man has a
testosterone level in what would be considered the normal range, his
current level is below what it would have been when he was younger and,
therefore, he is "hypogonadal." This concept is supported by the
evidence of a steady decline in testosterone levels, which reportedly
start to fall by the beginning at the fourth decade of life.
Measurements of free or bioavailable testosterone should be considered
experimental, until they are clearly shown to be a better marker of
hypogonadism in the elderly than total testosterone levels. Therefore,
measurement of free or bioavailable testosterone is not currently
justified.
When a testosterone level is found to be below 7.0 nm/L (2 ng/mL),
evaluation should proceed with measurement of gonadotropins, prolactin,
etc., regardless of age. However, the issue that is not clear is what
should be done for the healthy individual who does present with a
testosterone level between 7.0 and 10.5 nm/L (2–3 ng/mL). Should these
individuals have gonadotropin and prolactin levels measured? This
decision should be made based on the primary purpose for which the
testosterone measurement was initially made. If the testosterone level
was obtained because of the clinical suspicion of hypogonadism and the
value obtained was below the lower limits of normal for the reference
laboratory, then gonadotropin levels should be determined. If the
gonadotropin levels obtained are elevated, then appropriate evaluation
and treatment for primary testicular failure should be performed. If
the gonadotropin levels are low, evaluation for secondary hypogonadism
should be considered, including measurement of prolactin and
consideration of a pituitary imaging procedure. In these men with
borderline low levels of testosterone, if prolactin and gonadotropin
measurements do not provide an indication of the reason for the
marginal testosterone levels, consideration can be given to some
measurement of free or bioavailable testosterone. For the reasons
discussed earlier in this section, a free testosterone level measured
by a kit should be used cautiously. A more appropriate choice would be
bioavailable testosterone measured by ammonium sulphate precipitation,
free testosterone by a dialysis method, or separate measurements of
SHBG and testosterone with calculation of free testosterone. If the
free or bioavailable measurement is decreased and the man is clinically
hypogonadal but the cause is not clear, consideration to testosterone
replacement could be given, although the clinical response measurements
have not yet been defined. Note that the reason for the testosterone
measurement in these cases was a clinical suspicion of hypogonadism.
Until the results of appropriate clinical studies (several
NIH-sponsored studies addressing this issue are nearing completion)
become available, measurement of testosterone in older men without a
clinical suspicion of hypogonadism is inappropriate.
|
Footnotes
|
|---|
The author wishes to thank Dr. Sanjay Asthana
for review of the manuscript.
|
References
|
|---|
-
Gray A, Feldman HA, McKinley JB, Longcope C. 1991 Age, disease, and changing sex hormone levels in middle-aged men:
results of the Massachusetts Male Aging Study. J Clin Endocrinol
Metab. 73:1016–1025.[Abstract/Free Full Text]
-
Vermeulen A, Deslypere J. 1985 Testicular
endocrine function in the ageing male. Maturitas. 7:273–279.[CrossRef][Medline]
-
Tenover J, Dahl K, Hsueh A, Lim P, Matsumoto A, Bremner
W. 1987 Serum bioactive and immunoreactive follicle-stimulating
hormone levels and the response to clomiphene in healthy young and
elderly men. J Clin Endocrinol Metabol. 64:1103–1108.[Abstract/Free Full Text]
-
Tenover J, Matsumoto A, Clifton D, Bremner W. 1988 Age-related alterations in the circadian rhythms of pulsatile
luteinizing hormone and testosterone secretion in healthy men. J
Gerontol. 43:163–169.
-
Plymate SR, Tenover JS, Bremner WJ. 1989 Circadian
variation in testosterone, sex hormone binding globulin, and calculated
non-sex hormone binding globulin bound testosterone in healthy young
and elderly men. J Androl. 10:366–371.[Abstract/Free Full Text]
-
Pardridge W. 1988 Selective delivery of sex
steroid hormones to tissues in vivo by albumin and by sex
hormone-binding globulin. Ann NY Acad Sci. 538:173–192.[Medline]
-
Hammond G, Nisker J, Jones L, Siiteri P. 1980 Estimation of the percentage of free steroid in undiluted serum by
centrifugal ultrafiltration dialysis. J Biol Chem. 255:5023–5026.[Abstract/Free Full Text]
-
Sødergard R, Backström T, Shanbhag V, Carstensen
H. 1982 Calculation of free and bound fractions of testosterone
and estradiol-17ß to human plasma proteins at body temperature. J
Steroid Biochem. 16:801–810.[CrossRef][Medline]
-
Rosner W. 1997 Errors in measurement of plasma
free testosterone. J Clin Endocrinol Metabol. 82:2014–2015.[Free Full Text]
-
Nankin H, Calkins J. 1986 Decreased bioavailable
testosterone in aging normal and impotent men. J Clin Endocrinol
Metab. 63:1418–1423.[Abstract/Free Full Text]
-
Umstot E, Baxter J, Anderson R. 1985 A
theoretically sound and practicable equilibrium dialysis method for
measuring percentage of free testosterone. J Steroid Biochem. 22:639–648.[CrossRef][Medline]
-
Giraudi G, Cenderelli G, Migliardi M. 1988 Effect
of tracer binding to serum proteins on the reliability of a direct free
testosterone assay. Steroids. 52:423–424.[CrossRef][Medline]
-
Tenover J, Matsumoto A, Plymate S, Bremner W. 1987 The effects of aging in normal men on bioavailable testosterone and
luteinizing hormone secretion: Response to clomiphene citrate. J
Clin Endocrinol Metab. 65:1118–1126.[Abstract/Free Full Text]
-
Morley J, Perry H, Kaiser F, et al. 1993 Effects of
testosterone replacement therapy in old hypogonadal males: a
preliminary study. J Am Geriatr Soc. 41:149–152.[Medline]
-
Peiris A, Stagner J, Plymate S, Vogel R, Heck M, Samols
E. 1993 Sex hormone binding globulin levels in normal men: role of
pulsatile insulin secretion. J Clin Endocrinol Metab. 76:279–282.[Abstract]
-
Plymate S, Matej L, Jones R, Friedl K. 1988 Inhibition of sex hormone binding globulin (SHBG) production in the
human hepatoma (Hep G2) cell line by insulin and prolactin. J Clin
Endocrinol Metab. 67:460–464.[Abstract/Free Full Text]
-
Bremner W, Vitiello M, Prinz P. 1983 Loss of
circadian rhythmicity in blood testosterone levels with aging in normal
men. J Clin Endocrinol Metab. 56:1278–1285.[Abstract/Free Full Text]
Top
Introduction
References
