(Table 2, Figure 2). In contrast, there was no
correlation between the sexual domain of the EPIC
questionnaire and IGF-1 levels for any single
age group, although the correlation within the
4
70-year-old age group approached significance
(
r
¼
0.87,
P
¼
0.06). In addition, no significant corre-
lation was found between testosterone level and
SHIM or EPIC score, which supports previous
findings in the literature.
32
In addition, there was
no correlation between the hormonal domain of the
EPIC questionnaire and either IGF-1 or testosterone
levels.
To control for the impact of comorbidities on
IGF-1 levels, multivariate analysis using linear
regression modeling was carried out. These data
demonstrated that only hyperlipidemia is predictive
of IGF-1 levels (
P
¼
0.02, Table 3), with other comor-
bidities not significantly impacting IGF-1 levels. The
prevalence of hyperlipidemia in our subject popula-
tion as a function of age revealed no significant
difference, suggesting that the presence of hyperlipi-
demia does not affect the findings described above.
In light of reports suggesting that elevated IGF-1
and GH levels stimulate prostate cancer cell growth,
we determined the Spearman’s rank correlation
between IGF-1 level and both total Gleason score,
as well as the high component of the Gleason score.
Our data demonstrate no significant correlation
between IGF-1 levels and Gleason score, suggesting
that while GH and IGF-1 may stimulate the growth
of prostate cancer, tumor severity is not affected in
this group of men scheduled for radical prostatect-
omy. In addition, no correlation was shown between
testosterone level and Gleason score for all ages or
in any age group (data not shown).
Discussion
The prevalence of LOH is likely to continue
increasing as the world’s population ages. Recent
work has established a role for testosterone in the
development and maintenance of erections, and
testosterone replacement has been shown to ame-
liorate the symptoms of LOH, including erectile
dysfunction.
33–35
In contrast, the role of GH in the
maintenance of sexual function is less well estab-
lished, although studies have shown a relationship
between GH and erectile function.
In this work, we demonstrate a statistically signi-
ficant correlation between self-reported, validated
measures of sexual function in men, the SHIM and
EPIC questionnaires, and IGF-1 levels, most robustly
in men
4
70, as well as in men between 50 and 59
years old. In contrast, a correlation between testo-
sterone levels and SHIM/EPIC scores is not demon-
strated. Notably, study subjects were given SHIM
0
5
10
15
20
25
30
06
0
0
IGF-1 (ng/mL)
SHIM Score
500
400
300
200
100
Figure 1
Total Sexual Health Inventory for Men (SHIM) score vs
serum insulin-like growth factor (IGF-1) level. All subjects’ IGF-1
levels plotted against their self-reported scores of sexual function
based on total SHIM score. There is a statistically significant
correlation between the two measured values (Spearman’s
correlation coefficient,
r
¼
0.31;
P
¼
0.02).
0
20
40
60
80
100
120
0
100
200
300
400
500
EPIC Sexual Domain Score
IGF-1 (ng/mL)
600
Figure 2
Total Expanded Prostate Cancer Index Composite
(EPIC) sexual domain score vs serum insulin-like growth factor 1
(IGF-1) Level. All subjects’ IGF-1 level plotted against their
self-reported scores of sexual function based on the sexual domain
portion of the EPIC questionnaire. There is a statistically
significant correlation between the two measured values (Spear-
man’s correlation coefficient,
r
¼
0.30;
P
¼
0.02).
Table 3
Multivariate analysis assessing IGF-1 levels, age and
comorbidities
P-value
Age
0.22
Testosterone
0.92
Diabetes mellitus
0.31
Hypertension
0.80
Hyperlipidemia
0.02
Coronary artery disease
0.17
Smoking
0.63
Alcohol
0.64
Depression
0.22
Abbreviation: IGF-1, insulin-like growth factor 1.
IGF-1 levels correlate with sexual dysfunction
AW Pastuszak
et al
223
International Journal of Impotence Research