Int. J. Radiation Oncology Biol. Phys., Vol. 40, No. 1, pp.

129 –133, 1998

Copyright © 1998 Elsevier Science Inc.
Printed in the USA. All rights reserved
0360-3016/98 $19.00 1 .00

PII S0360-3016(97)00554-3

● Clinical Investigation

ELUCIDATING THE ETIOLOGY OF ERECTILE DYSFUNCTION AFTER

DEFINITIVE THERAPY FOR PROSTATIC CANCER

MICHAEL J. ZELEFSKY, M.D.* AND J. FRANCOIS EID, M.D.†

* Department Radiation Oncology, Memorial Sloan-Kettering Cancer, † Department of Urology, New York Hospital
Cornell Medical Center, New York, NY

Purpose: To determine the etiology of treatment-induced erectile dysfunction among patients who underwent
surgery or radiotherapy for prostatic cancer.
Methods and Materials: Ninety-eight patients were evaluated for erectile dysfunction after definitive therapy for
prostate cancer with Duplex ultrasonography before and after intracorporal prostaglandin injection. Patients
were classified as having arteriogenic, cavernosal, mixed (arteriogenic/cavernosal), or neurogenic impotence
based upon the results of the Duplex studies.
Results: Among patients who underwent radical prostatectomy (RP), 31 (52%) had cavernosal dysfunction, 19
(32%) had arteriogenic dysfunction, 3 (5%) were classified as mixed, and 7 (12%) as neurogenic dysfunction.
Among patients treated with radiotherapy (RT), 24 (63%) had arteriogenic dysfunction, 12 (32%) had cavernosal
dysfunction, 1 (2.5%) were classified as mixed, and 1 (2.5%) as neurogenic dysfunction. A multivariate analysis
identified prior RT as the only predictor of an arteriogenic etiology (p < 0.001) and prior RP as the only predictor
of a cavernosal etiology (p < 0.04) for erectile dysfunction among these patients. In the RP and RT groups, the
median erectile responses were 70 and 65%, respectively. Arterial peak flows <25 cc/min predicted for a
suboptimal erectile response with intracavernosal prostaglandin injections. Among 47 patients with arterial peak
flows <25 cc/min, 21 (55%) had erectile responses of <70%, while for 51 patients with arterial peak flows >25
cc/min, 31 (39%) had erectile responses of <70% (p < 0.039).
Conclusions: While the etiology of erectile dysfunction after definitive therapy for prostatic cancer is likely a
multifactorial phenomenon, these data suggest that the predominant etiology among patients who undergo RT
is arteriogenic and among patients who undergo RP is veno-occlussive/cavernosal pathology. This information
may have implications for the design of more effective therapies to address erectile dysfunction in this patient
population. © 1998 Elsevier Science Inc.

Prostate cancer, Impotence, Radiation therapy, Surgery, Etiology, Erections.

INTRODUCTION ine penile injections to assess the adequacy of penile blood

The physiology of penile erection is a complex multistep
phenomenon. Maintenance of a functional erection requires
vasodilation of the penile arteries and concomitant relax-
ation of the corporal smooth muscles. This allows for sub-
stantial increases in penile blood flow and trabecular space
engorgement. As a direct consequence of the increased
venous resistance and decreased venous outflow, the cor-
pora cavernosa and corpus spongiosum become turgid. The
bulbocavernosus and bulbospongiosus muscles further con-
tract allowing for increased cavernosal rigidity. These com-
plex physiologic interactions have been described as the
veno-occlusive mechanism of erectile function (8, 13, 14).
Recent studies have demonstrated that pharmacologic inter-
ventions that enhance the corporal smooth muscle relax-
ation and increase venous outflow resistance can improve
function in patients with vasculogenic impotence (1, 3, 13).
Lue et al. have pionered the use of ultrasound and pulsed
Doppler studies in conjunction with prostaglandin/papaver-
flow and erectile function (11–14). With this approach, the
structure and function of the cavernosal arteries and dorsal
veins can be assessed. The diameter of the arteries before
and after injection therapy can be directly measured to
quantitate their distensibilty during pharamacologic ther-
apy. These diagnostic interventions have also been used to
elucidate the etiology of newly diagnosed impotence (7, 10,
16, 19). With an increased understanding of the mechanism
of erectile dysfunction among patients, more effective ther-
apeutic interventions can be designed.
Between 1994 –1996, 98 patients with erectile dysfunc-
tion after definitive therapy for localized prostate cancer
were evaluated with Duplex ultrasonography before and
after intracoporal prostaglandin injection to evalute the eti-
ology of their impotence. Patients were classified as having
arteriogenic, cavernosal, or neurogenic impotence based
upon the results of the Duplex studies. The preliminary
findings indicate that among patients who have erectile

Reprint requests to: Michael J. Zelefsky, M.D., Department of York Ave., New York, NY 10021.
Radiation Oncology, Memorial Sloan-Kettering Cancer, 1275 Accepted for publication 14 July 1997.

129
130 I. J. Radiation Oncology ● Biology
dysfunction after radiotherapy, arteriogenic impotence is
the most prevalent etiology, while among patients who
underwent radical prostatectomy, cavernosal dysfunction is
a more likely cause of impotence.
METHODS AND MATERIALS
Between 1994 –1996, 98 patients treated with surgery or
radiation therapy for stage T1c–T3 prostate cancer under-
went Duplex penile sonography with pulsed doppler anal-
ysis to evaluate the etiology of their erectile dysfunction.
Sixty (61%) patients were previously treated with a nerve-
sparing radical prostatectomy, and 38 patients (39%) were
treated radiotherapy [external beam radiotherapy (n 5 35)
or permanent 125.1 implantation (n 5 3)]. The median age of
the patients were 62 years (range: 47–78). Among the
patients treated with radiotherapy, the median dose was 70.2
Gy (range: 61.2–75.6 Gy). Twelve patients (31%) in the
radiotherapy group were treated with 3 months of neoadju-
vant androgen ablation therapy prior to therapy and one
patient in the prostatectomy group received similar therapy
prior to surgery.
All patients were initially evaluated with a detailed med-
ical history that included information regarding concomitant
usage of medications, smoking, and alcohol intake. A de-
tailed baseline sexual history was also obtained; patients
with a suspected psychogenic etiology to their erectile dys-
function were not included within this analysis and referred
for appropriate evaluation. Physical examination included
careful inspection of the corpora to rule out scarring or
deformity. Patients then underwent a combined injection
and stimulation test. A 30-guage needle was used to inject
5 mg of Prostaglandin E-1 at the base of the penile shaft. In
the standing position, the erectile response was subjectively
scored by one examiner (JFE) at 5, 15, and 30 min postin-
jection.
All patients underwent scanning of the penis in the flaccid
state with images obtained in the transverse and longitudinal
cross-section. The diameters of the right and left cavernous
arteries were measured. Following a 10.0 mg intracavernous
injection of Prostaglandin E-1, the penis was rescanned and
the diameter of the cavernous arteries were measured to
asssess distensibility in response to vasoactive medication.
For the purposes of this study, any increase ,95% of the
diameter of the cavernosal arteries was considered abnor-
mal.
After the above was performed, each of the selectively
imaged arteries underwent pulsed doppler analysis with
generation of a wave form. The peak velocities of both the
right and left cavernosal arteries were measured at 5, 10,
and 15 min, respectively. Based on the work of Lue et al.
from measurements performed on normal subjects (11, 12),
peak blood flows ,25 cm/s were considered abnormal.
Resistive indexes (which reflect cavernosal smooth mucle
cell function) were measured for both the right and left
cavernosal arteries. Compared to the flaccid state, the erec-
tile response was measured in percentage at 5, 15, and 30
● Physics Volume 40, Number 1, 1998
Table 1. Patient characteristics according to treatment received
RP RT p
Characteristic (n 5 60) (n 5 38) value
Median age 62 (range: 47–75) 69 (range: 53–78)
Smoking history 32 (54%) 25 (68%) NS*
Daily alcohol
usage 27 (46%) 16 (43%) NS*
Cardiovascular
disease 16 (26%) 16 (42%) 0.17
* NS 5 not significant .0.05
min after the prostaglandin injection was administered. Pa-
tients were discharged from the clinic after spontaneous
detumescence occurred.
Differences between the frequencies among the treatment
groups were evaluated using the Mantel log rank test for
censored data (15). The Cox proportional hazard regression
model (5) was used to determine the contribution of covari-
ates predicting for a particular etiology of erectile dysfunc-
tion (i.e., arteriogenic, cavernosal).
RESULTS
Patients characteristics:
Table 1 demonstrates the patient characteristics of the
radical prostatectomy (RP) and radiotherapy (RT) groups.
Patients in the RP group described the onset of their erectile
dysfunction as ‘‘sudden’’ in 83% (50 of 60) while 10 (17%)
characterized their erectile dysfunction as ‘‘gradual.’’ In
contrast, 78% (29 of 38) in the RT group described the onset
of their dysfunction as ‘‘gradual,’’ while 22% characterized
this as ‘‘sudden.’’ These differences were highly significant
(p , 0.0001).
Prior to the onset of erectile dysfunction, the baseline
sexual function was similar for the treatment groups. In the
RP group, the interest in sexual activity prior to surgery was
described as strong, normal or low in 27, 37, and 36%,
respectively. In the RT group, baseline interest in sexual
activity was described as strong, normal or low in 24, 46,
and 30%, respectively. In addition, the frequency of inter-
course prior to manifestation of sexual dysfunction was
similar. In the RP group, frequency of activity was reported
as #1/week, 2/week, and .2 week in 42, 24, and 34%,
respectively. In the RT group, frequency of activity was
reported as #1/week, 2/week, and .2 week in 35, 27, and
38%, respectively. In the RP group, the median time from
surgery to evaluation for erectile dysfunction was 11
months (range: 3–130 months). In the RT group, the median
time from RT to evaluation for erectile dysfunction was 14
months (range: 2–150 months).
Doppler Evaluation to assess erectile function
Overall, 43 (44%) were classified as having arteriogenic
dysfunction based upon peak penile blood flow rates of ,25
cc/min. Forty-three patients (44%) were characterized as
having cavernosal dysfunction based upon abnomal disten-
 Etiology erectile dysfunction
Table 2. Etiology of erectile dysfunction
Type Surgery RT
Dysfunction (n 5 60) (n 5 38)
Arteriogenic 19(32%) 24(63%)
Cavernosal 31(52%) 12(31%)
Arteriogenic/Cavernosal 3 (5%) 1 (3%)
Neurogenic 7(12%) 1 (3%)
sibility of the corpora cavenosa in the setting of normal
penile peak blood flow. In 32 of these 43 patients (74%), a
venous leak was documented on Doppler study. In four
patients (4%) both arteriogenic and cavernosal dysfunction
were noted. Eight patients (8%) were classified as having
neurogenic dysfunction based on normal distensibility and
peak blood flow values but poor erectile responses despite
prostaglandin intracavernosal injections.
When patients were analyzed by the therapy received for
prostate cancer, a marked difference in the etiology of
erectile dysfunction was noted. Among RP patients, 31
(52%) had cavernosal dysfunction and 19 (32%) had arte-
riogenic dysfunction; while among patients treated with RT,
24 (63%) had arteriogenic and 12 (32%) had cavernosal
dysfunction. These results are summarized in Table 2.
Patients with a prior history of myocardial infarction and
or hypertension currently on antihypertensives were classi-
fied as having a cardiovascular (CV) risk factor. When
dysfunction were classsified according to the presence of
CV risk factors, no significant differences were noted as
shown in Table 3. In addition, the use of neoadjuvant
hormonal therapy prior to therapy had no influence on the
type of erectile dysfunction observed in this group of pa-
tients. A multivariate analysis was performed to identify
variables predicting for arteriogenic and cavernosal dys-
function. Variables entered into the regression analysis in-
cluded smoking history, age .60, cadiovascular disease,
and prior RT. The only predictor of arteriogenic etiology
was prior RT (p 5 0.002) and the only predictor for a
cavernosal etiology was prior RP (p 5 0.04).
The overall erectile response after administration of int-
racavernosal pharmacotherapy was 70%. In the RP and RT
groups the median erectile response were 70 and 65%,
respectively. Arterial peak flows ,25 cc/min predicted for
a suboptimal erectile responses with intracavernosal pros-
taglandin injections. Among 47 patients with arterial peak
flows ,25 cc/min, 26 (55%) achieved erectile responses of
Table 3. Etiology of erectile dysfunction
No
Cardiovascular Cardiovascular
Type Risk Risk p report suggests that vasculogenic etiology is the most prom-
Dysfunction (n 5 66) (n 5 32) value
Arteriogenic 26(39%) 17(53%)
Cavernosal 29(44%) 14(44%)
Arteriogenic/Cavernosal 4 (6%) - osal rather than arteriogenic dysfunction. In these patients
Neurogenic 7(11%) 1 (3%)
● M. J. ZELEFSKY AND J. F. EID 131
,70%, while for 51 patients with with arterial peak flows
$25 cc/min, 20 (39%) achieved erectile responses of ,70%
p
value (p , 0.039).
0.002
0.05 DISCUSSION
-
-
Our results suggest that erectile dysfunction after prostate
radiotherapy is likely caused by disruption of the arteriolar
system supplying the corporal muscles rather than veno-
occlusive pathology. Among the 38 patients who were
treated with RT, an arteriogenic etiology was the most
common cause of posttreatement erectile dysfunction, and
this etiology was significantly more prevalent among pa-
tients who underwent RT compared to RP. On the other
hand, among patients who underwent RP, the predominant
cause of erectile dysfunction was cavernosal followed by
arteriogenic and neurogenic etiologies. These data also
highlight the fact that in some patients who are treated with
RT, cavernosal dysfunction may be the dominant etiology
of radiation-induced impotence rather than reduced penile
arterial blood flow. In this study, cavernosal dysfunction
was suggestive of disruption of the normal veno-occlusive
mechanism necessary for normal erectile function. Among
the 43 patients who were classified as having cavernosal
dysfuction, 75% had evidence of venous leakage based on
Doppler studies. It is recognized as well that inherent char-
acteristics among patients may have contributed to the
higher incidence of a specific etiology of dysfunction ob-
served for a treatment group. In general, those treated with
radiotherapy represented an older patient population with an
expected increase in atherosclerotic heart disease. The cur-
rent data suggest that such patients who develop impotence
after radiotherapy for prostatic cancer will be more likely to
have arteriogenic dysfunction.
Several prior studies have attempted to define the etiol-
ogy of erectile dysfunction after RT (2, 6, 17). Goldstein et
al. (6) performed Doppler evaluations in 23 patients who
previously underwent RT to the pelvis area and found that
all had decreased blood flow in the cavernosal arterial
system. In contrast to these findings, Mittal (17) reported on
six patients who underwent similar evaluation prior to and
6 –9 months after RT for prostate cancer. No changes in
penile blood flow were demonstrated in these patients. Al-
though this study had the advantage of having available
baseline and posttreatment data on all patients (not available
in the current study), it is nevertheless difficult to draw
conclusions given the limited number of patients studied. In
addition, 6 –9 months after treatment may be an insufficient
amount of time to fully appreciate changes in the penile
vasculature. Others have speculated the possibility of dam-
age to the nerves or veins in the pelvis after RT (3). Our
inent factor associated with RT-induced impotence.
0.20 This study also suggests that the dominant cause of
0.84 impotence among patients who underwent RP was cavern-
mechanical disruption or manipulation of the the neurovas-
132 I. J. Radiation Oncology ● Biology ● Physics
cular bundle (even during a nerve-sparing procedure) may
lead to aberrant nerve conduction, disruption of arterial or
venous blood flow. Among the 60 RP patients in the present
study, the most common etiology of erectile dysfunction
was cavernosal dysfunction with disruption of the normal
veno-occlusive mechanism, most often in the form of ve-
nous leakage. Only seven (12%) were identified as having a
neurogenic etiology for their erectile dysfunction. These
latter patients all had normal distensibility of the cavernosal
arteries and normal penile blood flow but poor responses
after PGE injections. Critical factors affecting the likelihood
of potency preservation after nerve-sparing prostatectomy
procedures include the ability to spare the neurovascular
bundle bilaterally, age of the patient, and the presence of
preexisiting erectile dysfunction. Lue has hypothesized that
among patients with borderline erectile function prior to an
operation, even minimal surgical manipulation may be re-
flected in a higher incidence of impotence (13).
In the patients in the present report, optimal erectile
responses were achieved with pharmacologic therapy. The
overall response rates after intracavernosal injections were
70 and 65% in the RP and RT groups, respectively. Several
studies have reported satisfactory responses with vasoactive
pharmacotherapy for patients with postprostatectomy erec-
tile dysfunction (1, 13, 17, 18). Pierce et al. (18) reported on
eight patients who received radiotherapy to the prostate/
pelvis and were treated with intracavernosal injection con-
sisting of a phentolamine-papavarine solution. All patients
were reported to have achieved a satisfactory erectile re-
sponse. There is a paucity of information regarding which
factors predict for a suboptimal response with intracavern-
osal pharmacotherapy. In the present report, patients with
arterial penile blood flow ,25 cc/min were more likely to
have erectile responses ,70% compared to patients with
flow velocities .25 cc/min (p 5 0.039). This is consistent
with observations that patients with known vasculogenic
REFERENCES
1. Austoni, E.; Bellorofonte, C.; Mantovani, F. Improved results
with intracavernous vasoactive drug infusion following new
surgical techniques for vasculogenic impotence. World
J. Urol. 5:182; 1987.
2. Banker, F. L. The preservation of potency after external beam
irradiation for prostate cancer. Int. J. Radiat. Oncol. Biol.
Phys. 15:219; 1988.
3. Barada, J. H.; McKimmy, R. M. Vasoactive pharmacotherapy.
In: Bennett, A. H., ed. Impotence: Diagnosis and management
of erectile dysfunction, 1st ed. Philadelphia, PA: W. B. Saun-
ders Co.; 1994:229 –250.
4. Bondil, P.; Schauvliege, T.; Nguyen Qui, J. L. Impuissance
par fuite veineuse. Reflexions a propos de 60 cas. Acta Urol.
Belg. 56:279; 1988.
5. Cox, D. R. Regression models and life tables. J. R. Stat. Soc.
B. 34:187–220; 1972.
6. Goldstein, I.; Feldman, M. I.; Deckers, P. J.; Babayan, R. K.;
Krane, R. J. Radiation associated impotence. JAMA 215:903;
1984.
7. Iacono, F.; Barra, S.; Lotti, T. Evaluation of penile deep
Volume 40, Number 1, 1998
causes of impotence or diabetes have required higher doses
of vasoactive agents to achieve satisfactory erectile re-
sponses compared to patients with psychogenic or neuro-
genic impotence with intact arterial blood flow and veno-
occlusive mechanism (13).
Elucidating the etiology of impotence after denitive ther-
apy for prostate cancer may facilitate the development of
more optimum strategies to address the patient’s erectile
dysfunction. Patients with cavernovenous insufficiency may
benefit more from therapeutic interventions that are effec-
tive in reducing venous outflow during erection rather than
simply increasing penile arterial blood flow. For patients
with significant leakage as the etiology of their erectile
dysfunction, deep dorsal vein resection/ligation, emboliza-
tion, and spongiolysis have been used with reported success
rates ranging from 10 –90% (4, 9, 20, 21). Patients with
predominant arteriogenic dysfunction will likely benefit
from therapies that further induce improved penile blood
flow through the cavernosal arteries.
Although the current report suggests that that the pre-
dominant etiology of radiation-induced impotence is arte-
riogenic, it is likely that the etiology is multifactorial with
other factors still contributing in part to the patient’s erectile
dysfunction. Many patients with preexisting atherosclerotic
disease who receive prostate irradiation may have a combi-
nation of arteriogenic, cavernosal, and neurogenic etiologies
at work. As a result, a particular therapeutic intervention
aimed at addressing one of multiple etiologies may not be
sufficient to achieve satisfactory erectile function for the
patient. As treatment-induced impotence in this patient pop-
ulation is most likely a complex, multifactorial phenome-
non, more refined diagnostic tools are needed to further
elucidate the presence and degree to which contributing
factors play a role in the erectile dysfunction so that more
effective therapeutic strageies can be employed for this
patient population.
arteries in psychogenic impotence by means of duplex ultra-
sonography. J. Urol. 149:1262–1264; 1993.
8. Lerner, S. E.; Melman, A.; Christ, G. J. A review of erectile
dysfunction: New insights and more questions. J. Urol. 149:
1246 –1255; 1993.
9. Lue, T. F. Penile venous surgery. Urol. Clin. North Am.
16:607; 1989.
10. Lue, T. F. Impotence after prostatectomy. Urol. Clin. North
Am. 17:613; 1990.
11. Lue, T. F.; Hricak, H.; Marich, K. W.; Tanagho, E. A. Evaluation
of arteriogenic impotence with intracorporeal injection of papav-
erine and the duplex ultrasound scanner. Semin. Urol. 3:43; 1985.
12. Lue, T. F.; Hricak, H.; Marich, K. W.; Tanagho, E. A. Vasculo-
genic impotence evaluated by high-resolution ultrasonography
and pulsed Doppler spectrum analysis. Radiology 155:777; 1985.
13. Lue, T. F.; Donatucci, C. F. Dysfunction of the venoocclusive
mechanism. In: Bennett, A. H., ed. Impotence: Diagnosis and
management of erectile dysfunction, 1st ed: Philadelphia, PA:
W. B. Saunders Co.; 1994:197–228.
14. Lue, T. F.; Tanagho, E. A. Physiology of erection and phar-
 Etiology erectile dysfunction
macological management of impotence. J. Urol. 137:829 –
836; 1987.
15. Mantel, N. Evaluation of survival data and two rank order
statistics arising in its consideration. Cancer Chemo. Rep.
50:163–170; 1966.
16. Meuleman, E. J. H.; Bemelmans, B. L. H.; van Asten,
W. N. J. C.; Doesburg, W. H.; Skotnicki, S. H.; Debruyne,
F. M. J. Assessment of penile blood flow by duplex ultra-
sonography in 44 men with normal erectile potency in differ-
ent phases of erection. J. Urol. 147:51–56; 1992.
17. Mittal, B. A study of penile circulation before and after
radiation in patients with prostate cancer and its effect on
impotence. Int. J. Radiat. Oncol. Biol. Phys. 11:1121–1125;
1985.
● M. J. ZELEFSKY AND J. F. EID 133
18. Pierce D. J.; Whittington, R.; Hanno, P. M. Pharmocologic
erection with intracavernosal injection for men with sexual
dysfunction following irradiation: A preliminary report. Int. J.
Radiat. Oncol. Biol. Phys. 21:1311; 1991.
19. Shabsigh, R.; Fishman, I. J.; Quesada, E. T.; Seale-Hawkins,
C. K.; Dunn, J. K. Evaluation of vasculogenic erectile impo-
tence using penile duplex ultrasonography. J. Urol. 142:1469 –
1474; 1989.
20. Treiber, U.; Gilbert, P. Venous surgery in erectile dysfunction:
A critical report on 116 patients. Urology 34:22; 1989.
21. Wespes, E.; Delcour, C.; Preserowitz, L.; Herbaut, A. G.;
Struyven, J.; Schulman, C. Impotence due to corporeal veno-
occlusive dysfunction: Long-term follow-up of venous sur-
gery. Eur. Urol. 21:115; 1992.