The Relationship Of Serum Testosterone To Erectile Function In Normal Aging Men
We evaluated the variation in serum testosterone in normal aging men and its relationship with erectile function. In a study that was not community based and during a free screening program for prostate cancer 1,071 men were invited to complete a sexual activity questionnaire, that is the abridged 5-item version of the International Index of Erectile Function (IIEF-5), as a diagnostic tool for erectile dysfunction. Possible scores on the IIEF-5 are 1 to 25 and erectile dysfunction was classified into 5 categories based on the scores, namely severe-1 to 7, moderate-8 to 11, mild to moderate-12 to 16, mild-17 to 21 and none-22 to 25. Serum total testosterone was measured between 8:00 and 10:00 a.m. in all men. Of the 1,071 men 965 (90.1%) were included in this study, of whom 88% were white and 12% were black. Mean age was 60.7 years. In this sample the prevalence of all degrees of erectile dysfunction was estimated to be 53.9%. The degree of erectile dysfunction was mild in 21.5% of cases, mild to moderate in 14.1%, moderate in 6.3% and severe in 11.9%. According to age the erectile dysfunction rate was 36.4% in the 40 to 49, 42.5% in the 50 to 59, 58.1% in the 60 to 69, 79.4% in the 70 to 79 and 100% in the 80 years and older groups (p <0.05). The variation in mean serum total testosterone in the age groups was not statistically significantly different (p >0.05). Pearson coefficients of age and total testosterone did not reveal any significant correlation (r = 0.00376, p = 0.907), similar to IIEF-5 score and total testosterone (r = 0.0163, p = 0.612). However, analysis of the variables IIEF-5 and age showed a statistically significant inverse or negative relationship (r = -0.3449, p <0.05). Erectile dysfunction showed a clear association with aging but no consistent correlation of total testosterone with erectile condition was identified.
THE RELATIONSHIP OF SERUM TESTOSTERONE TO ERECTILE
FUNCTION IN NORMAL AGING MEN
ERNANI LUIS RHODEN, CLAUDIO TELO
KEN, PAULO ROBERTO SOGARI AND
CARLOS ARY VARGAS SOUTO
From the Andrology Division, Department of Urology, Santa Casa Hospital and Fundac¸a˜ o Faculdade Federal de Cieˆncias Me´dicas, Porto
Purpose: We evaluated the variation in serum testosterone in normal aging men and its
relationship with erectile function.
Materials and Methods: In a study that was not community based and during a free screening
program for prostate cancer 1,071 men were invited to complete a sexual activity questionnaire,
that is the abridged 5-item version of the International Index of Erectile Function (IIEF-5), as a
diagnostic tool for erectile dysfunction. Possible scores on the IIEF-5 are 1 to 25 and erectile
dysfunction was classified into 5 categories based on the scores, namely severe—1 to 7, moder-
ate— 8 to 11, mild to moderate—12 to 16, mild—17 to 21 and none—22 to 25. Serum total
testosterone was measured between 8:00 and 10:00 a.m. in all men.
Results: Of the 1,071 men 965 (90.1%) were included in this study, of whom 88% were white
and 12% were black. Mean age was 60.7 years. In this sample the prevalence of all degrees of
erectile dysfunction was estimated to be 53.9%. The degree of erectile dysfunction was mild in
21.5% of cases, mild to moderate in 14.1%, moderate in 6.3% and severe in 11.9%. According to
age the erectile dysfunction rate was 36.4% in the 40 to 49, 42.5% in the 50 to 59, 58.1% in the
60 to 69, 79.4% in the 70 to 79 and 100% in the 80 years and older groups (p ⬍0.05). The variation
in mean serum total testosterone in the age groups was not statistically significantly different (p
⬎0.05). Pearson coefficients of age and total testosterone did not reveal any significant correla-
tion (r ⫽ 0.00376, p ⫽ 0.907), similar to IIEF-5 score and total testosterone (r ⫽ 0.0163, p ⫽
0.612). However, analysis of the variables IIEF-5 and age showed a statistically significant
inverse or negative relationship (r ⫽⫺0.3449, p ⬍0.05).
Conclusions: Erectile dysfunction showed a clear association with aging but no consistent
correlation of total testosterone with erectile condition was identified.
KEY WORDS: penis, impotence, testosterone, questionnaires, aging
Accompanying the aging process are certain critical condi-
tions that are often associated with decreased testosterone,
such as decreased muscle tissue mass, increased body fat
mass, decreased bone mass, osteoporosis, decreased sense of
well-being, depression, decreased libido and increased erec-
The decrease in serum androgen associ-
ated with aging in normal males is accompanied by a decline
in testicular function, including lower serum testosterone
and bioavailable testosterone, and increased androgen bind-
ing to the sex hormone globulin.
3, 5, 6
Because of these aspects,
the decrease in non sex hormone binding globulin bound
testosterone, called bioavailable testosterone, is often much
greater than the decrease in total testosterone. Another as-
pect is the fact that is unknown whether the age related
decline in serum testosterone in men is universal.
Normal male sexual function depends on a complex inter-
play of psychological, neurological, vascular and endocrine
There is considerable controversy on the relative
importance of each factors in the initiation and maintenance
of erection, especially the role of serum testosterone.
studies show that aging in men is associated with decreased
sexual interest and activity, particularly with an increased
prevalence of erectile dysfunction.
Generally androgens enhance libido and the frequency of
sexual acts but a causal relationship between altered andro-
gen levels and erectile dysfunction has not been proved.
2, 4, 10
However, recent data support the theory that androgens,
mainly free testosterone, have a beneficial effect on sexual
On the other hand, the real relationship of total
testosterone and sexual function, more specifically erectile
function, is less clear.
We assessed serum total testosterone
in normal aging men and determined its association with
self-reported erectile function.
MATERIALS AND METHODS
The ethics committee at our hospital approved this study.
Patients were previously informed of the research details and
they agreed to participate in the study. This noncommunity
based series included 1,071 men who were invited to partic-
ipate during a free screening program for prostate cancer,
that is Prostate Cancer Awareness Week at Santa Casa
Hospital-Porto Alegre, Brazil, from July 26 to 30, 1998. This
number of men was previously established without any type
of selection criteria by the program. This aspect needs careful
consideration regarding its possible influence on the results,
mainly due to the extremes of age. In addition to questions on
urinary symptoms and physical examination, all men who
presented completed a sexual activity questionnaire, that is
the abridged 5-item version of the International Index of
Erectile Function (IIEF-5),
as a diagnostic tool for erec-
They were not asked about sexual function,
risk factors or co-morbidities. As previously described by
Rosen et al, this questionnaire consists only of 5 questions
and each IIEF-5 item is scored on a 5-point ordinal scale, on
which lower values represent poorer sexual function.
Accepted for publication November 21, 2001.
HE JOURNAL OF UROLOGY
Vol. 167, 1745–1748, April 2002
Copyright © 2002 by A
MERICAN UROLOGICAL ASSOCIATION,INC.
Printed in U.S.A.
Thus, a response of 1 on a question was considered least
functional, whereas a response of 5 was considered most
functional. Possible scores on the IIEF-5 are 1 to 25. Scores
above 21 were considered normal erectile function and scores
at or below this cutoff were considered erectile dysfunction.
Erectile dysfunction was classified into 4 categories based on
IIEF-5 scores, namely severe—1 to 7, moderate— 8to11,
mild to moderate—12 to 16, mild—17 to 21 and none—22 to
Total testosterone determined by a commercially avail-
able radioimmunoassay kit was the only serum hormone
measured. Blood samples were obtained between 8:00 and
10:00 a.m. Normal total testosterone was 241 to 827 ng/dl.
The screening was advertised in print and electronic me-
dia, and participants were self-selected after responding to
this publicity. The media were only directed toward prostate
evaluation. All men who were patients at the division of
urology or andrology, those on intracavernous pharmacolog-
ical therapy or oral drugs for erectile dysfunction, those with
a penile prothesis and those with a major psychiatric or
penile anatomical disorder, or who reported no sexual activ-
ity during the last 6 months were excluded from study. Anal-
ysis of variance (ANOVA) followed by the Bonferroni and
chi-square tests was used for statistical analysis as well as
the Pearson correlation with p ⬍0.05 considered statistically
Of the 1,071 men 965 (90.1%) were included in the study. A
total of 106 men (9.9%) were excluded from analysis because
of failure to complete all protocol criteria. Of the respondents
850 (88%) were white and 115 (12%) were black. Mean age
was 60.7 years (range 45 to 90).
In this sample the prevalence of all degrees of erectile
dysfunction was estimated to be 53.9%. Erectile dysfunction
was mild in 21.5% of cases, mild to moderate in 14.1%,
moderate in 6.3% and severe in 11.9% (table 1). According to
age the erectile dysfunction rate was 36.4% in the 40 to 49,
42.5% in the 50 to 59, 58.1% in the 60 to 69, 79.4% in the 70
to 79 and 100% in the 80 years and older group. This rate was
statistically different among all age groups (p ⬍0.05) except
in the 40 to 49 and 50 to 59-year-old groups (p ⬎0.05, table 2).
Table 3 lists mean total testosterone per individual in the
various decades of life. The variation in mean serum total
testosterone in the age groups did not show any statistically
significant difference (p ⬎0.05). However, we observed a
greater percent of men with subnormal total testosterone in
the 80-year-old and older group (p ⬍0.05) compared with the
other age groups. Men 70 years old or older more often
presented with subnormal total testosterone compared with
those 40 to 49 years old (p ⬍0.05, table 4).
When considering only the 144 men with the maximum
score of 25 points on the IIEF-5, we observed that 143 (99.3%)
had normal total testosterone and only 0.7% showed a sub-
normal level of the androgen. The Pearson coefficients of the
variables age and total testosterone did not reveal any sig-
nificant correlation (r ⫽ 0.00376, p ⫽ 0.907). Furthermore, no
correlation was noted of IIEF-5 score with total testosterone
(r ⫽ 0.0163, p ⫽ 0.612, see figure). However, analysis of the
variables IIEF-5 score and age showed a statistically signif-
icant inverse or negative relationship (r ⫽⫺0,3449, p ⬍0.05).
Data in this study do not represent a randomly selected
population from within a community. Study participants
sought medical attention in a free screening program. There-
fore, it is possible that these data may not represent national
or even regional status. Individuals seeking medical atten-
tion in a screening program may be more concerned with
health than the general population. On the other hand, pa-
tients with co-morbidity and low quality of life may have no
interest in participating in this type of program.
The National Institutes of Health recommends routine
measurement of serum testosterone in men with erectile
However, studies show that this recommenda-
tion is irrational because of the overall low yield of clinically
significant endocrine abnormalities.
The current study
confirm this finding.
In men there is a gradual decrease in androgen with ag-
4, 5, 16–24
There is great variability among individuals in
the degree of this reduction and not all aging men have
hypogonadism to a clinically significant degree.
less, little is known about the true incidence of hormonal
abnormalities in an unselected group of men with erectile
failure and, in addition, there is confusion concerning the
effects that hormonal abnormalities have on erectile func-
10, 21, 23
With aging free testosterone decreases by 1.2% yearly and
albumin bound testosterone decreases by 1% yearly. Sex
hormone binding globulin, the carrier protein that binds 60%
of circulating testosterone, increases by 1.2% yearly with the
net effect that total serum testosterone decreases more
slowly at 0.4% yearly than the free or albumin bound testos-
terone pools alone.
4, 5, 7,11, 24
TABLE 1. Prevalence of normal erectile function and erectile
dysfunction in the study population
Erectile Function (IIEF-5) No. Pts. (%)
Normal 445 (46.1)
Dysfunction: 520 (53.9)
Mild 208 (21.5)
Mild–moderate 136 (14.1)
Moderate 61 (6.3)
Severe 115 (11.9)
TABLE 2. Prevalence of erectile dysfunction according to age group
Age Group No. Pts.
40–49 11 4 (36.4)
50–59 470 200 (42.5)
60–69 334 194 (58.1)
70–79 131 103 (79.4)
80 or older 19 19 (100)
Statistically significant difference among all age groups, except 40 to 49 and
50 to 59 years (ANOVA and Bonferroni test p ⬍0.05).
TABLE 3. Serum total testosterone according to age group
Age Group No. Pts.
Mean Serum Total
Testosterone ⫾ SD (ng./dl.)
40–49 11 503.18 ⫾ 194.90
50–59 470 527.68 ⫾ 186.66
60–69 334 536.43 ⫾ 188.08
70–79 131 554.52 ⫾ 201.43
80 or older 19 424.84 ⫾ 195.43
ANOVA (F ⫽ 2.15, p ⫽ 0.072) p ⬍0.05.
TABLE 4. Number of patients per serum total testosterone level
according to age group
Serum Total Testosterone (ng./dl.)
Than 241 (%)
Than 827 (%)
40–49 0 10 (91) 1 (9)
50–59 15 (3.2) 422 (89.8) 33 (7)
60–69 13 (3.9) 298 (89.2) 23 (6.9)
70–79 8 (6.1)* 112 (85.5) 11 (8.4)
80 or older 5 (26.3)† 14 (73.7) 0†
* Versus other age groups chi-square test p ⬍0.05.
† Versus 40 to 49-year-old age group chi-square test p ⬍0.05.
RELATIONSHIP OF TESTOSTERONE TO ERECTILE FUNCION1746
Testosterone production is regulated by the hypothalamic-
pituitary-testis axis and there is no clear consensus on the
endocrine mechanism of the rather modest decline in andro-
gen with aging.
2, 7, 10,17–20, 25
Several studies indicate that
aging is associated with changes at all 3 levels of the axis but
probably most predominantly at the testicular level.
4, 6, 18
Furthermore, there is increased binding of testosterone to its
carrier proteins, resulting in a lower level of free, biologically
5, 7, 17
Nickel et al reported a 17.5% inci-
dence of hypothalamic-pituitary-gonadal axis abnormalities
in patients with an initial diagnosis of erectile failure, al-
though in only 12.1% did abnormalities clearly contribute to
However, plasma testosterone below the lower normal
limit occurs only in a minority of elderly men, including 7% in
the 40 to 60, 20% in the 60 to 80 and 35% in the older than 80
years old group.
When compared with the rate of erectile
dysfunction in the various decades of life (table 2), these data
show much lower values. Erectile dysfunction data in other
such as the Massachusetts Male Aging Study
indicate a higher rate of this condition in the
various decades of life than the rate of hormonal abnormal-
ities. These aspects show that other significant conditions
have a significant influence on sexual condition.
It seems safe to say that when androgen values are within
the reference range, they are not a significant factor in erec-
tile dysfunction in aging men.
Sexual function decreases
with age but hormonal changes are not the major determin-
ing factor in this decline.
Korenman et al stated that hor-
monal hypogonadism without compensatory gonadotropin el-
evation is common in older men.
which is also common in aging men, is independent of go-
Because there is usually no cause and effect
relationship of low testosterone and erectile dysfunction, it
was suggested that the 2 conditions should be evaluated
separately because they are separate entities or it may indi-
cate that the process is more complex and factors other than
testosterone alone are involved.
Others, such as Nieschlag et al, did not observe any corre-
lation of testosterone with sexual activity in vigorous elderly
However, Davidson et al clearly noted a significant
but weak correlation of sexual activity with free testoster-
Similarly Tsitouras et al reported higher mean testos-
terone in men with high sexual activity.
Another interesting aspect is that the effects of androgen
replacement on sexual activity and erectile function are
rather disappointing, which is not surprising in view of the
complexity of the factors determining sexual activity,
whereas erectile function in elderly men rarely has an endo-
crine but more often has a vascular or neurological ori-
2, 6, 10,32, 33
Study results show a more consistent impact of
testosterone on behavior related directly to libido and not on
erectile function which, on the other hand can be retained
when serum androgen decreases below the normal
32, 34, 35
In absolute terms testosterone in aging men is almost
always within the normal reference values despite the pos-
sibility of a significant decrease during an individual life-
time. In the MMAS median free testosterone at ages 40 and
70 years was 0.23 and 0.16 nmol./l., respectively, with a
within-group standard deviation of 74%. Thus, a substantial
number of elderly men continue to have serum testosterone
within the wide range present in much younger men.
tably normal total testosterone assay results are widely vari-
able at 241 to 827 ng./dl. This aspect and the relatively small
numbers in the youngest and oldest age groups may have
skewed the results and may explain the lack of a positive
relationship in the current study. Another aspect is the pos-
sible lack of a relationship of total testosterone in the aging
process with sexual function, as described by others, as well
as the possible influence on this type of androgen by another
factor, such as sex hormone binding globulin.
scribed previously, sex hormone binding globulin increases
with aging and this fact results in a lower decrease in serum
total testosterone than in nonsex hormone binding globulin
bound testosterone. Further longitudinal studies with longer
followup would probably explain many doubts regarding this
Another aspect that must be considered is the widely rec-
ognized major influences of androgen in the sexual desire
domain. A possible skewing effects of this aspect on our
results was possible because the IIEF-5 determines only the
erectile condition of sexual function. However, the actions of
libido over the other sexual functions is well known. Unfor-
tunately we did not evaluate this aspect in the current study.
In the MMAS, Feldman et al did not note any consistent
correlation of erectile dysfunction with testosterone
there was a significant association of this condition with
the relationship of androgens to sexual function. The most
common conclusions indicate that testosterone correlates
positively with sexual desire and sleep related erection but
not with erectile dysfunction or the frequency of coitus.
The latter finding confirms evidence in androgen replace-
ment studies in hypogonadal men that androgen is more
important for sustaining sexual desire and sleep related erec-
tion than for maintaining the erectile response to external
7, 34, 38
Pearson correlation of total testosterone with IIEF-5 scores (r ⫽ 0.01623, p ⫽ 0.615)
RELATIONSHIP OF TESTOSTERONE TO ERECTILE FUNCION
The age associated decrease in androgen varies highly
among individuals. Many elderly men have an androgen
level that is in the normal range of young men, and so there
is no reason to treat them with androgens whatever their
complaints may be.
16, 23, 38
Moreover, few controlled studies
have assessed the effects of androgen substitution on clinical
signs generally attributed to androgen deficiency, such as
asthenia, lack of energy, decreased bone mass and osteopo-
rosis, decreased libido and sexual activity, and erectile dys-
11, 17, 38
There are enormous variations in testosterone values
among men of all ages. We did not detect any significant
variations in these levels in the current study. No significant
correlation with erectile function was identified but there
was a consistent relationship of the latter with age. Another
aspects is that the free testosterone fraction and not total
testosterone, which was considered in this study, or the pos-
sible relationship of various levels of androgen may possibly
be more useful in erectile function.
As noted by Sternbach, it would be erroneous to attribute
erectile dysfunction to decreasing testosterone when there
are other variables that must be considered as the etiology of
complaints related to diminished libido, erectile and ejacula-
tory dysfunction, and reduced sexual activity.
ables include availability of a partner, fear of performance
failure, impaired penile perfusion, chronic illness, depres-
sion, medications, neuropathy, smoking, and alcohol and
This study does not confirm a positive relationship of erec-
tile dysfunction and low serum total testosterone. However,
it shows a direct relationship of an increased prevalence of
erectile dysfunction and aging. Further studies are needed to
determine the possible relationships of specific fractions of
androgens (free testosterone) with erectile functions.
1. Jain, P., Rademaker, A. W. and Mcvary, K. T.: Testosterone
supplementation for erectile dysfunction: results of a meta-
analysis. J Urol, 164: 371, 2000
2. Fahmy, A. K., Mitra, S., Blacklock, A. R. E. et al: Is the meas-
urement of serum testosterone routinely indicated in men with
erectile dysfunction? BJU Int, 84: 482, 1999
3. Tenover, J. S.: Effects of testosterone supplementation in the
aging male. J Clin Endocr Metab, 75: 1092, 1992
4. Tenover, J. L.: Male hormone replacement therapy including
“andropause”. Endocr Metabol Clin North Am, 27: 969, 1998
5. Tserotas, K. and Merino, G.: Andropause and the aging male.
Arch Androl, 40: 87, 1998
6. Vermeulen, A. and Kaufman, J. M.: Ageing of the hypothalamo-
pituitary-testicular axis in men. Horm Res, 43: 25, 1995
7. Gooren, L. J. G.: The age-related decline of androgen levels in
men: clinically significant? Br J Urol, 78: 763, 1996
8. Feldman, H. A., Goldstein, I., Hatzichristou, D. G. et al: Impo-
tence and its medical and psychosocial correlates: results of
the Massachusetts Male Aging Study. J Urol, 151: 54, 1994.
9. Jonler, M., Moon, T., Brannan, W. et al: The effect of age,
ethnicity and geographical location impotence and quality of
life. Br J Urol, 75: 651, 1995.
10. Nickel, J. C., Morales, A., Condra, M. et al: Endocrine dysfunc-
tion in impotence: incidence and cost-effective screening.
J Urol, 132: 40, 1984.
11. Kim, Y. C.: Testosterone supplementation in the aging male. Int.
J Impot Res, 11: 343, 1999.
12. Rosen, R. C., Cappelleri, J. C., Smith, M. D. et al: Development
and evaluation of (an) abridged, 5-item version of the Interna-
tional Index of Erectile Function (IIEF-5) as a diagnostic tool
for erectile dysfunction. Int J Impot Res, 11: 319, 1999.
13. Rosen, R. C., Riley, A., Wagner, G. et al: The International Index
of Erectile Function (IIEF): a multidimensional scale for as-
sessment of erectile dysfunction. Urology, 49: 822, 1997.
14. Lehmann, K., Eichlisberger, R. and Gasser, T. C. Lack of diag-
nostic tools to prove erectile dysfunction: consequences for
reimbursement? J Urol, 163: 91, 2000.
15. NIH Consensus Development Panel on Impotence. JAMA, 270:
16. Werner, A. A.: The male climacteric. Report of two hundred and
seventy-three cases. JAMA, 132: 188, 1946
17. Vermeulen, A.: The male climacterium. Ann. Med., 25: 531, 1993
18. Urban, R. J.: Neuroendocrinology of aging in the male and fe-
male. Endocr Metab Clin North Am, 21: 921, 1992
19. Lamberts, S. W., Van den Beld, A. and Van der Lely, A.-J.: The
endocrinology of aging. Science, 278: 419, 1997
20. Burns-Cox, N.: The andropause: fact or fiction? Postgrad. Med J,
73: 553, 1997
21. Hafez, B.: Recent advances in clinical/molecular andrology. Arch
Androl, 40: 187, 1998
22. Kim, Y. C. Testosterone supplementation in the aging male. Int
J Impot Res, 11: 343, 1999
23. Morales, A., Heaton, J. P. W. and Carson, C. C. Andropause: a
misnomer for a true clinical entity. J Urol, 163: 705, 2000
24. Longcope, C., Goldfield, S. R. W., Brambilla, D. J. et al: Andro-
gens, estrogens, and sex hormobne-binding globulin in middle-
aged men. J Clin Endocr Metab, 71: 1442, 1990
25. Lue, T.: Erectile dysfunction. N Engl J Med, 342: 1802, 2000
26. Wagner, G. and Saenz de Tejada, I. Update on male erectile
dysfunction. BMJ, 316: 678, 1998
27. Joahannes, C. B., Araujo, A. B., Feldman, H. A. et al: Incidence
of erectile dysfunction in men 40 to 69 years old: longitudinal
results from the Massachusetts Male Aging Study. J Urol,
163: 460, 2000
28. Korenman, S. G., Morley, J. E., Mooradian, A. D. et al: Second-
ary hypogonadism in older men: its relation to impotence. J
Clin Endocr Metab, 71: 963, 1990
29. Nieschlag, E., Lammers, U., Freischem, C. W. et al: Reproduc-
tive function in young fathers and grandfathers. J Clin Endocr
Metab, 55: 676, 1982
30. Davidson, J. M., Chen, J. J., Crapo, L. et al: Hormonal changes
and sexual function in aging men. J Clin Endocr Metab, 37: 71,
31. Tsitouras, R. D., Martin, F. E. and Harman, S. M. Relationship
of serum testosterone to sexual activity in eldery men. J
Gerontol, 37: 288, 1982
32. Sternbach, H. Age-associated testosterone decline in men: clini-
cal issues for psychiatry. Am J Psychiatry, 155: 1310, 1998
33. Morales, A., Johnston, B., Heaton, J. P. W. et al: Testosterone
supplementation for hypogonadal impotence:assessment of
biochemical measures and therapeutic outcomes. J Urol, 157:
34. Kwan, M., Greenleaf, W. J., Mann, J. et al: The nature of andro-
gen action on male sexuality: a combined laboratory-self-
report study on hypogonadal men. J Clin Endocr Metab, 57:
35. Bagateli, C. J., Heiman, J. R., Rivier, J. E. et al: Effects of
endogenous testoterone and estradiol on sexual behavior in
normal young men. J Clin Endocr Metab, 78: 711, 1994
36. Gray A., Feldman A., McKinlay, J. B. et al: Age disease, and
changing sex hormone levels in middle-aged men: results of
the Massachusetts Male Aging Study. J Clin Endocr Metab,
73: 1016, 1991
37. Mulligan, T.: Why aged men become impotent. Arch Intern Med,
149: 1365, 1989
38. Becker, A. J.,
Uckert, S., Stief, C. G. et al: Cavernous and
systemic testosterone levels in different phases of human pe-
nile erection. Urology, 56: 125, 2000
RELATIONSHIP OF TESTOSTERONE TO ERECTILE FUNCION