Mild Hypogonadotropic Gladys W. Strain,
Robert S. Rosenfeld,
Hypogonadism Jacob Kream,
in Obese Men
James J. Strain, Richard Deucher,
Joseph Levin, and David K. Fukushima
To evaluate the pituitary-gonadal axis of obese men, we compared the 24-hour mean plasma concentrations of total and free testosterone and of dihydrotestosterone, FSH, and LH in 21 healthy obese men, aged 18-50, and 24 age-matched healthy nonobese men. In the obese men, we also measured the volume of ejaculate and the number and motility of sperm, and potency by history and by measurement of nocturnal penile and investigated libido by psychiatric interview, tumescence. As a group, the obese men had less than two-thirds the normal mean plasma levels of total testosterone. free testosterone, and FSH; the difference from normal was highly significant for all three. 24 hr LH levels were normal, which is inappropriately low in view of the subnormal testosterone levels. 24 hr mean levels of dihydrotestosterone and spermatogenesis, libido, and potency were essentially normal. Taken together, the findings represent a state of mild hypogonadotropic hypogonadism. which thus appears to be characteristic of obese men. This abnormality probably results from partial suppression of the pituitary by the elevated plasma estrogen levels we and others find in these men.
HERE HAVE BEEN many studies of the pituitary-gonadal axis in obese women, stimulated by the readily observable menstrual irregularities and fertility problems that many of these women have, but the pituitary-gonadal axis of obese men has received less attention, possibly because these men lack obvious clinical stigmata that might point to reproductive dysfunction. Obese men have been reported by two groups to have elevated plasma estrogen levels,’ ’ and we have confirmed this finding.” There has been one report” of subnormal testosterone production, and four groups’.’ ‘” have reported subnormal plasma total testosterone levels in obese men; three of the latter groups’.‘,“’ report free testosterone levels: one’ reports normal levels; one’ reports an inverse correlation between the levels and the degree of obesity, with subnormal levels in 2 out of 10 men; and the third’” reports significantly subnormal levels in men more than I OO’Xabove desirable weight. The plasma gonadotropin levels of obese men have been reported in six studies:‘,’ ” one of these’ reports subnormal FSH levels in some of the men and the others’,‘.’ ” report normal levels; however, one of the latter’ points out that the fact that gonadotropin levels are not elevated in the face of subnormal testosterone levels means that obese men in general have depressed pituitary function. So far as we can determine, plasma dihydrotestosterone levels and sperm counts have not been reported in obese men. WC investigated the pituitary-gonadal axis of 21 otherwise healthy obese men by measuring their 24 hr mean plasma concentration of total and free testosterone. amd of dihydrotestosterone, follicle-stimulating hormone (FSH), and luteinizing hormone (LH), and by determining the number and motility of sperm in the ejaculate. We also investigated the libido of these men. by structured psychiatric interviews, and their potency, by history and by measurement of nocturnal penile tumescence. Our findings indicate that obese men characteristically manifest a state of mild hypoMetabolrsm,
Vol. 3 1, No. 9 (September), 1982
gonadotropic and potency.
but have normal
IO) and ranging 127%
obese men, aged 18-50
+ SD: 33 + 14). were studied.
+ SD: 33 L weight
consent was obtained
from all subjects.
These were measured
pooling aliquots determining
24 hr blood-sampling
by sampling from
blood every 20 mitt for 24 hr. each
in each pool. Details
have been previously
and of the
of’ Plasma Dih?ldrotestosterone
These steroids same antibody
from this laboratory.”
for both steriods
the testosterone procedural
used is generated
to a plasma
Front the Clinical and Medical
try. and the Institute
had been separated
losses. is equivalent
at 100. 200. 400, and
Keceivrd Jtir publication. Supported AM-27292
in part hy Grants RR-53
the L1.S. .-lir Force
Research. Address Medical
Israel N I’.
10003. 1~ 1982 by Grune & Stratton, 0026 -04U5/82/3
Inc. I .00/o
STRAIN ET AL.
800 pg dose levels yielded mean recoveries of lO6%, 95%, 97% and 95% respectively after correcting for procedural losses, The intraassay and interassay coefficients of variation were 7.2% (n = 20) and 13.1% (n = 14) respectively. The least detectable dose in the DHT assay is equivalent to a plasma level of 6 ng/dl. Addition of unlabeled DHT in amounts of IO pg. 20 pg. 40 pg, and 80 pg to 100 ml portions of plasma with an assayed value of 10.4 pg yielded recoveries of 8 1%. 100%. IO1 %, and 103% respectively after correction for procedural losses and subtraction of the “0” value. The intraassay and interassay coefficients of variation for a male plasma pool were 5.1% (n = 13) and 10.1% (n = 37) respectively. Determination of Plasma Free Testosterone Free testosterone is calculated using a computer program based upon the equations of Yates and Urquhart14 utilizing plasma albumin concentration, plasma SHBG determined by the method of Rosner,15 the total tesosterone concentrations, and the affinity constants for the binding of testosterone to SHBG (8.4 x IO’M ‘) and to albumin (3.7 x IO’ - Mm’) respectively. The sensitivity and variability of the determination of free testosterone are dependent upon the corresponding parameters of the assays of total testosterone, SHBG, and albumin. In our laboratory, the intra-assay and interassay coefficients of variation of the SHBG assay are 5.7% (n = 20) and 9.8% respectively. Free testosterone values calculated in this way have been shown to be essentially identical to those measured by equilibrium dialysis2’ Plasma albumin is determined by an automated procedure in the central laboratories of Montefiore Hospital, and the intra-assay and interassay coefficients of variation are both reported to be less than 5% at a range of values from 2.555.5 g/dl.
of Plasma FSH and LH
The methods used are the standard double-antibody procedures reported from Midgely’s laboratory.‘6,‘7 Reagents are obtained from the NPA-NIAMD as described, except for the second antibody, goat anti-rabbit-globulin, which is purchased from Pantex, Santa Monica, California. In particular, antiserum to hLH Batch #2 has been used in the studies described in this report. The relative potency (RP) of LER (in terms of the 2nd IRP HMG) using this antiserum is 2.9 ng/mlU. For the FSH assay, antiserum Batch #S was used; the corresponding RP of LER 907 is 17.5 ng/mlU. The LH antiserum cross-reacts completely with hCG. The minimum detectable dose for the LH and FSH assays are 0.4 mlU/ml and 2.9 mlU/ml respectively. In ten different assays the mean percentage of the individual tracer specifically bound to the LH antibody (final dilution 1:800,000) and to the FSH antibody (final dilution 1:250,000) was 43.6 (range 40.0-48.6) and 35.5 (range 3 I. l-44.3) respectively. The mean non-specific binding to the LH and FSH antibodies was 2.1 percent (range 1.9-2.8) and I.9 percent (range 1.5-2.0) respectively. The intra-assay variability for the LH and FSH assays was 4.1% and 6.1% respectively. The interassay variability for the LH and FSH assays was 6.7% and 10.0% respectively.
Determination of the Volume of Ejaculate and the Number and Motility of Sperm
these men by history during the interview, and was additionally evaluated in I I of them by measurement of nocturnal penile tumescence according to the method of Fisher.‘”
Statistical Methods Plasma FSH, LH dihydrotestosterone, total testosterone. and free testosterone values were distributed more nearly log-normally than normally in both obese and nonobese men. Accordingly, intergroup comparisons were made by applying Student’s t test, 2-tailed, to the logarithms of the values. The volume of ejaculate and sperm density were plotted linearly on probability paper to determine the presence of a normal distribution of values within populations.
24-Hr Mean Plasma Total and Free Testosterone The total testosterone values of the obese men were markedly subnormal (Fig. I); their geometric mean level was 268 ng/dl (95% C.L. I30 to 552). 42% lower than the geometric mean level of 461 ng/dl (95% C.L. 259 to 822) in the nonobese controls (p < 0.0001). The plasma free testosterone values showed a very close linear correlation with the total testosterone values in both the obese men and the nonobese men, and the regression lines for these two correlations did not differ significantly (/ 2.6; p = 0. I), so that the free testosterone levels of the obese men were likewise markedly subnormal (Fig. 2); their geometric mean level was 7.0 ng/dl (95% C.L. 2.9-17.0). 44% lower than the geometric mean level of 12.6 ng/dl (95% C.L. 6.9-23.1) in the nonobese controls (p i 0.001).
24-Hr Mean Plasma Dihydrotestosterone The plasma dihydrotestosterone levels of the obese men did not differ significantly from those of the nonobese controls: the geometric means were 94 and 98 ng/dl. respectively.
24-Hr Mean Plasma FSH and LH The FSH levels of the obese men were markedly subnormal (Fig. 3); their geometric mean level was 5.2 mlU/ml (95% C.L. 2.31 I .8), 36% lower than the geometric mean of 8. I mlU/ml(95% C.L. 3.8-l 7.4) in the nonobese controls (p = 0.001). The geometric mean LH level of the obese men. excluding the two clear-cut outliers designated by the symbol “x” in Fig. 4, was 10.2 mlU/ml. not significantly different from that of the nonobese controls, 10.5 mlU/ml. The two outliers had definitely subnormal values; these two men also had two of the lowest total testosterone values ( I73 and 207 ng/dl) and two of the lowest FSH values (2.9 and 3.4 mlU/ ml).
lp< : & I-, a’ ZEP 3%
Ejaculates were obtained from I6 of the obese men by masturbation after 3 days of continence. Clinical urologists blind to the study design measured the volume of the ejaculate and the number of sperm (in a counting chamber); the percentage of motile sperm was estimated in all samples.
Evaluation of Libido and Potency Libido structured
was evaluated in all the obese men in the course of a interview by a psychiatrist. Potency was evaluated in all
Twenty-four-hour mean plasma total testosterone in Fig. 1. age-matched nonobese and obese men.
Fig. 2. Twenty-four-hour mean plasma age-matched nonobese and obese men.
Volume of‘Ejaculate Sperm
of our subjects
of sperm per ml. of
and Bunge plotted
show a unimodal from
Fig. 6 shows the sperm concentrations
in the ejaculates. log-normal
had a geometric
mal. The which
those of Nelson (95%
and those tested also had normal
exclusively was bisexual.
two were exclusively
332% above desirable
body weight. was
not sexually active; his size made even masturbation
(95% C.L. had a
do not differ forms
the obese men had essentially
the values for the obese men were not subnor-
1~S.D. of percentage
and Bunge C.L.
mean value of 46,000.000/ml while
Fig. 5 shows the ejaculate
and the two are indistinguishable
and the concentration
Libido and Poterq
and those of Nelson
and Number and Motility of
in our obese population
The volume of ejaculate ejaculate
was 69 -r IX.
in our laboratory.
As a group, the obese men we studied had less than two thirds the normal mean plasm levels of total testosterone, free testosterone, and FSH. Taken together, these findings represent a state of mild hypogonadotropic hypogonadism, which thus appears to be characteristic of obese men. Spermatogenesis,
- --t .
/ o Nonokw (Nellon t Bunga)
. . . .,... 5
Twenty-four-hour and obese
Ok80 Men (Thir Study)
, . , 60
OF POPULATION WITH AN EQUAL LOWER EJACULATE VOLUME
STRAIN ET AL.
Obese MenfThis Study) /
loo: J /
/ ./ /
PERCENTAGE OF POPULATION WITH AN EQUAL OR LOWER SPERM COUNT Fig. 6. men.
of sperm counts in nonobese
libido, and potency were normal, which also qualifies the hypogonadism as relatively mild. The persistence of normal dihydrotestosterone levels in the face of decreased plasma levels of free and total testosterone is interesting, and implies that the relationship between these levels and the rate of DHT formation is more complex than a simple first-order reaction. Several previous authors have presented data suggestive of the presence of hypogonadotropic hypogonadism in obese men. Glass et a!.19 found normal levels of FSH and LH in these men. but commented that ‘6 . . . the absence of elevated gonadotropins . . . in these subjects with low free testosterone suggests . . . inappropriate suppression of their hypothalamic-pituitary axis.” Amatruda et a! reported in an abstract’ that “ . . . serum FSH was slightly reduced in two out of six morbidly obese patients” and concluded that “. . defect these data . . . suggest a possible hypothalamic to explain the depressed plasma testosterone”; in their subsequent full paper,” these workers reported normal FSH and LH levels but commented that “. . normal
LH concentration in the presence of significant decreases in total T and free T index . . . suggests . the hypothalamic-pituitary-gonada! axis was not responding appropriately to a low free T index.” Our data with respect to LH are similar, in that we found “inappropriately low” (though literally normal) values, but our findings with respect to FSH were much more definite: the values of the obese men as a group were displaced downwards, and the mean of the obese men was less than ?/3 the normal mean. It was this consistency of the subnormal FSH (we attribute the consistency to our measuring 24 hr mean plasma concentrations instead of “spot” concentrations) that led us to tie a!! the observations together into a picture of subclinical hypogonadotropic hypogonadism. The pathogenesis of this abnormality in obese men requires further investigation. We hypothesize that it results from partial suppression of the pituitary by the elevated plasma estrogen levels we’ and otherslm4 find in these men. Since these elevated levels, in turn, probably result from increased conversion of S4androstenedione to estrone,’ 4 it follows that suppression of adrenocortica! secretion of A4-androstenedione should lower them and thereby eliminate the hypogonadotropic hypogonadism. Preliminary studies in four patients in our laboratory” indicate that a 1 wk course of corticosteroids (which depressed plasma A’-androstenedione levels by about 55%) produced substantial normalization of the plasma estrone, estradio!, testosterone, and FSH levels. Though the results of these pilot studies require confirmation in a larger group of patients, they appear to support our hypothesis. ACKNOWLEDGMENT The authors gratefully
tance of Anne Thalassinos and Nathan noassay) and William Terri
Katz (hormone radioimmu-
David Schessel. and
Marc Schessel (24 hr blood-sampling and NPT studies).
REFERENCES I. Schneider
R. et al: Increased
estrogen production in obese men. J Clin Endocrinol Metab 48:633-
Normalwerten. 3. Kley
Dtsch med Wschr 104: 1487-I Edelmann
plasma sex hormones to different subjects. Metabolism
of obesity in male
sion of androstenedione
to estrogens in obese males. J Clin Endo-
crinol Metab 5 1:I 128-l
Kream J, et al: Obese young men have women do
deficiency of morbid
et al: Testosterone Clin
9. Glass AR, Swerdloff
RS, Bray GA, et al: Low serum testoste-
Clin Endocrinol Metab45:121 IO. Amatruda
globulin in massively obese men. J l-1219,
in obese males. J Clin End&in01 Il.
gonadal axis during prolonged total starvation
3O:lOl I, 1981 Raith
Depressed plasma testosterone and fractional binding of testosterone
elevated plasma estrogen levels but obese premenopausal not. Metabolism
obesity. Clin Res 22:647A,
rone and sex-hormone-binding
4. KIey HK, Keselaers T, Peerenboom H, et al: Enhanced conver-
5. Zumoff B, Strain GW,
deficiency of morbid obesity: Possible role of the hypothalamus.
482, 1979 HL:
in obese males. Acta
1971 7. Barbato
638, 1979 2. Kley
Clin Nutr 22:767-770,
L, et al: Pituitaryin obese men. Am J
IN OBESE MEN
12. Zumoff B, Rosenfeld RS, Strain GW, et al: Sex differences in the 24-hour mean plasma concentrations of dehydroisoandrosterone (DHA) and dehydroisoandrosterone sulfate (DHAS) and the DHA to DHAS ratio in normal adults. J Clin Endocrinol Metab 51:330333.1980 13. Boyar RM, Rosenfeld RS, Kapen S, et al: Human puberty: Simultaneous augmented secretion of luteinizing hormone and testosterone during sleep. J Clin Invest 54:609-618, 1974 14. Yates FE, Urquhart J: Control of plasma concentrations of adrenocortical hormones. Physiol Rev 42:359-443, 1962 IS. Rosner W: A simplified method for the quantitative determination of testosterone-estradiol-binding globulin activity in human plasma. J Clin Endocrinol Metab 34:983-988, 1972 16. Midgely AR: Radioimmunoassay for human follicle-stimulating hormone. J Clin Endocrinol Metab 27:295-299, 1967 17. Midgely AR: Radioimmunoassay: A method for human
chronic gonadotropin 79:10-18, 1966
18. Fisher C, Schiavi RC, Edwards A, et al: Evaluation of nocturnal penile tumescence in the differential diagnosis of sexual impotence. Arch Gen Psychiatry 36:431-437, 1979 19. Nelson CMK, changing parameters 255033507. 1974
Bunge RG: Semen of male fertility
Analysis for Fertil Steril
20. Strain GW, Zumoff B, Levin J, et al: Reversal of hyperestrogenemia and hypogonadotropic hypogonadism in obese men by corticoid administration. Am J Clin Nutr 34:618. 1981 (Abstr) 21. Mall GW. Rosenfield RL: Testosterone binding and free plasma androgen concentrations under physiological conditions: Characterization by flow dialysis technique. J Clin Endocrinol Metab 49:730, 1979