Erectile Dysfunction in Hypertensive Subjects
Assessment of Potential Determinants
Hypertension is often cited as a risk factor for erectile dysfunction. To clarify the relation between hypertension and erectile dysfunction, we evaluated 32 consecutive hypertensive and 78 normotensive impotent men with respect to multiple potential determinants and parameters of erectile function, including medical and sexual history, depression, hormonal profile, penile nocturnal tumescence, penile vascular supply, and pudendal nerve conduction. The hypertensive men were older, had higher body mass index, and used more medications than the normotensive men. The groups were not different with respect to the prevalence of smoking and peripheral vascular disease, but the hypertensive men had a marginally higher rate of ischemic heart disease (P=.06). The prevalence of depression, abnormal nocturnal penile tumescence, anomalous pudendal nerve conduction, and impairment in arterial supply as determined by penile brachial index were similar in the two groups. Testosterone and bioavailable testosterone levels were lower in the hypertensive men. After stratification by age and body mass index, hypertensive men younger than 50 years with body mass index less than 30 kg/m2 had significantly lower testosterone levels (12.0±1.7 versus 21.3±1.4 nmol/L, P<.02) but not bioavailable testosterone levels (3.9±0.7 versus 6.4±0.7 nmol/L, P<.17) than the corresponding normotensive group. Prolactin, follicle-stimulating hormone, and luteinizing hormone levels of the two groups were not significantly different. Contrary to common belief and with the exception of lower circulating testosterone levels, the overall analysis showed little difference between hypertensive and normotensive men with respect to a wide range of classic determinants of erectile function. Direct study of the local vascular erectile apparatus appears necessary for further elucidation of the mechanisms underlying erectile dysfunction in hypertensive men.
Hypertension is often cited as a risk factor for ED. The association is usually attributed to either antihypertensive drug therapy or vascular factors. Although it is commonly assumed that hypertension somehow predisposes men to impotence, the precise relation of hypertension and impotence remains an enigma. Available information is scant, fragmented, and often inconsistent. Contrary to common belief, Newman and Marcus1 reported a meager relationship between hypertension and ED, with a similar prevalence of impotence in hypertensive and normotensive men. Although vascular disease is commonly implicated in the pathogenesis of “hypertensive impotence,” Shabsigh et al2 and Virag et al3 found that even though hypertensive impotent men had more arterial risk factors for vascular impotence, the presence of hypertension did not affect the prevalence of ED.
Hormonal abnormalities may provide another etiologic basis for sexual dysfunction in hypertensive men. Higher prolactin levels have been reported in hypertensive men.4 5 However, whether this relative hyperprolactinemia is associated with altered sexual function has not been established. To clarify the relation between hypertension and ED, we compared normotensive and hypertensive ED men with regard to age, BMI, hormonal profile, penile arterial flow, risk factors for arterial disease, and the presence of neurological and psychological abnormalities.
One hundred fifty-eight consecutive referred men with ED (age range, 18 to 84 years) were evaluated at the Elias Sourasky–Tel Aviv Medical Center outpatient clinic for sexual dysfunction. They filled out in privacy a questionnaire addressing their general medical status; potential risk factors for sexual dysfunction, including smoking and alcohol and drug intake; and details of sexual activity and marital (or between-partner) relationships. Additionally, the possibility of coexisting depression was assessed with Beck's Depression Inventory,6 the most commonly used self-rating scale for depression. Subjects are asked to rate 21 items according to their feelings at the present time, and a cumulative score of all questions (each graded on a scale of 0 to 3) is derived. Each subject underwent a complete physical examination, with special attention paid to the vascular, neurological, and endocrine systems. Subjects with a BMI of 25 to 30 kg/m2 were considered overweight, and obesity was defined as BMI higher than 30 kg/m2.
Laboratory screening for reproductive hormones consisted of three blood samples taken 20 minutes apart in the morning. The pooled samples were analyzed for testosterone, BT, LH, FSH, and prolactin. LH, FSH, and prolactin were determined by an automated enzyme immunoassay system (ES 300, Boehringer Mannheim). Serum testosterone was measured with a double-antibody commercial radioimmunoassay (Diagnostic Products Corp). BT was measured by a modification of a method described by Tremblay and Dube.7 Tracer amounts of tritiated testosterone were added to serum aliquots (0.5 mL) that were then incubated for 30 minutes at 37°C. An equal volume of a saturated solution of ammonium sulfate was added to precipitate the globulin fraction that included the sex hormone binding globulin bound to testosterone. After separation of the sex hormone binding globulin fraction by centrifugation (1100g, 3 minutes), the supernatant was counted. The product of the percentage of labeled testosterone remaining in the supernatant and total testosterone yields the BT concentration. Urinary androsterone and etiocholanolone were determined in the morning (first void) urine as we have previously described.8
Primary hypogonadism was defined by testosterone levels less than 3 ng/mL and/or BT levels less than 1 ng/mL and LH levels more than 15 mIU/mL. Hyperprolactinemia was defined as prolactin levels higher than 20 ng/mL on at least two separate samples (collected at least 2 hours after awakening). Erectile function was assessed by recording nocturnal penile tumescence using Rigiscan II (Dacomed Corp) for two consecutive nights. Nocturnal penile tumescence was considered abnormal if at least one of the following criteria was met: change in penile diameter during erectile episodes of less than 3 cm at the base and 2 cm at the tip, duration of penile tumescence of less than 10 minutes per event in all recorded events, rigidity of less than 70%, or dissociation in the rigidity scores between the base and the tip.
Penile arterial supply was determined by measurement of the peak systolic velocity, end-diastolic velocity, and resistance index with the use of a 7.5-MHz color Doppler probe (Rimed). In each case, resistance index was derived by subtracting end-diastolic velocity from peak systolic velocity and dividing the difference by peak systolic velocity. We also calculated the penile brachial index, which is the ratio of penile systolic arterial pressure to systolic pressure in the arm. Penile brachial index, peak systolic velocity, end-diastolic velocity, and resistance index were derived with subjects at rest and again either after 5 minutes of bicycle exercise or 15 minutes after intracavernous injection of 10 μg prostaglandin E1. Both procedures result in increased arterial flow. Resting penile brachial index of less than 0.6 and/or an exercise-related or intracavernous injection–related drop in penile brachial index of at least 0.15 were considered significant markers of arterial impairment.
Penile neural function was measured by pudendal nerve conduction velocity.9 To evaluate the sensory and motor nerves of the penis, we determined latency and amplitude based on at least three evoked responses. Latency was considered abnormal if it exceeded 42 milliseconds.
Results are reported as mean±SE. Differences in the prevalence of various parameters were assessed by the χ2 test.
Among 158 ED men, 48 diabetics were excluded from the analysis. Of the remaining 110 subjects, 78 were normotensive (mean blood pressure, 137/85 mm Hg) and 32 were hypertensive (mean BP on therapy, 161/95 mm Hg) (Table 1⇓). The normotensive men had no chronic or debilitating disorders. Self-reported pretreatment BP levels showed that all hypertensive subjects had either mild or moderate essential hypertension as determined by standard criteria. The hypertensive men were older than the normotensive men (61.6±1.7 versus 52.7±1.4 years, P<.001) and also had higher BMI (29.5±0.5 versus 26.7±0.6 kg/m2, P<.01). Obesity (BMI >30 kg/m2) was more prevalent among the hypertensive men (40% versus 12.7% in the normotensive group, P<.02).
Effect of Antihypertensive Drugs
Although a broad spectrum of hypotensive agents was recorded in the hypertensive group, the most common medications were calcium channel blockers (35%), β-blockers (33%), and converting enzyme inhibitors (28%). Of note was the rare use of diuretics (3%). Of the 32 hypertensive subjects, 28 took antihypertensive drugs on a regular basis. These subjects were asked to determine the time or appearance of ED symptoms in relation to drug therapy. In only 3 subjects did the chronological relationship suggest a drug effect. Twenty subjects rejected the possibility of any association, and the remaining 6 were unable to recall the necessary details with sufficient accuracy.
Smoking and Vascular Disease
As summarized in Table 2⇓, the prevalence of smoking and clinically detectable peripheral vascular disease was quite similar in hypertensive and normotensive men. Ischemic heart disease appeared somewhat more often among the hypertensive subjects, but this difference was just short of attaining statistical significance (P=.06).
The depression scores according to Beck's questionnaire did not differ between the groups. The prevalence of depression was 6.25% and 20.5% for hypertensive and normotensive men, respectively (P=.2).
Lower mean testosterone and BT levels were found in the hypertensive group (Table 3⇓). However, there was no correlation between blood pressure and testosterone (r=.01) or blood pressure and BT (r=−.05) in our study population. Since the hypertensive men were older and more obese than the normotensive men and since both advanced age10 and obesity are associated with lower testosterone levels, data were analyzed again after stratification according to age and BMI. Hypertensive subjects younger than 50 years (mean age, 47) had significantly lower testosterone levels than the younger normotensive subjects (mean age, 40) (12.1±1.2 and 21.3±1.4 nmol/L, respectively). BT was also lower in the hypertensive men of this age group (4.3±0.6 and 6.2±0.7 nmol/L, P=.048). However, the prevalence of overt hypogonadism in the two groups was negligible, and the prevalence of hypogonadism did not differ significantly between the groups. To assess the potential contribution of obesity to the lower testosterone levels in the hypertensive group, we also analyzed separately androgen levels in men with BMI less than 30 kg/m2. In this group, hypertensive men had lower BT levels (3.7±0.3 versus 4.9±0.5 nmol/L, P<.03) but not testosterone levels (18±1.4 versus 20.4±0.7 nmol/L). In hypertensive men younger than 50 years with BMI less than 30 kg/m2, testosterone levels (12.0±1.7 nmol/L) were lower than in the normotensive nonobese subjects of the corresponding age group (21.3±1.4 nmol/L, P<.02). BT levels in younger, nonobese hypertensive and normotensive subjects were not different (3.9±0.7 and 6.4±0.7 nmol/L, respectively; P<.17). However, the small number of hypertensive subjects in this subgroup (n=4) precludes any firm conclusion in this respect.
Mean testosterone and BT levels did not differ between hypertensive men treated with β-blockers and those taking other hypertensive agents. Notably, FSH, LH, and the ratios of testosterone to LH and BT to LH of the two groups were not significantly different. The ratio of androsterone to etiocholanolone in the urine, which is a marker of 5α-reductase activity and presumably reflects the peripheral (tissue) conversion of testosterone to dihydrotestosterone,11 also did not differ between the groups (1.34±0.1 and 1.37±0.18 in the hypertensive and normotensive subjects, respectively).
Nocturnal Penile Tumescence
The prevalence of abnormal nocturnal penile tumescence was 62.5% and 53.3% in the hypertensive and normotensive men, respectively (P=.64).
Penile Vascular Studies
Basal peak systolic velocity, resistance index, and penile brachial index of the two groups did not differ (Table 4⇓). Since there was no difference between the groups when postexercise or post–intracavernous injection parameters were analyzed separately, these two maneuvers were treated as a single probe (Table 4⇓, posttest). As shown, posttest peak systolic velocity, resistance index, and penile brachial index also showed no between-group differences.
Pudendal Nerve Conduction
Pudendal nerve conduction was tested in 50 subjects (15 hypertensive and 35 normotensive). Three hypertensive men (20%) and 6 normotensive men (17%) had abnormal recordings (P=.8).
Although hypertension is a known risk factor for atherosclerotic disease, our data provide no direct evidence of a role for arterial factors in the pathogenesis of ED in mild to moderate essential hypertension. Although there was a trend for an increased frequency of coronary artery disease in the hypertensive impotent group, the prevalence of impaired penile arterial blood supply as determined by penile brachial index, peak systolic velocity, and resistance index in the hypertensive subjects was quite similar to that of the normotensive group. This finding is consistent with the report of Virag et al,3 who found a similar prevalence of abnormal penile brachial index among normotensive and hypertensive men with ED. However, the prevalence of an abnormal penile brachial index in the study of Virag et al was twofold higher than in our series, probably because of differences in the definition of the normal range.
Testosterone and BT levels were lower in the young hypertensive impotent men, yet they were not within the hypogonadal range. Furthermore, the number of men with overt hypogonadism was low in both groups. An explanation for the lower testosterone and BT levels among hypertensive subjects with ED is not readily available. Our hypertensive subjects were older than the normotensive subjects. Although this may have contributed to the observed difference, it is unlikely that it can fully explain the lower androgen concentrations in hypertensive subjects: In younger subjects (<50 years), the difference in testosterone levels between hypertensive and normotensive men was even greater than for the overall between-group comparison. Furthermore, mean testosterone levels in our young hypertensive men with ED (3.5 ng/mL) are clearly lower than in the general male population of this age. Overweight may also contribute to the lower androgen levels in hypertensive individuals, as differences in BT levels were no longer present after young obese subjects were excluded. Overweight is also associated with increased estradiol levels,12 which we did not assess. Thus, the effect of lower testosterone levels in subjects with increased circulating estrogens may be larger than that reflected by androgen levels per se because of an altered ratio of testosterone to estrogen. The possibility of a drug effect on circulating testosterone deserves consideration. Theoretically, such effects could be exerted indirectly via effects on prolactin or gonadotropins. However, we are not aware of chronic effects of any of the main hypotensive agents used by the hypertensive group on pituitary hormone secretion. Suzuki et al13 examined the effects of atenolol, nifedipine, captopril, and thiazides on sex steroids. In their study, atenolol induced a mild decrease in serum testosterone levels. In our analysis, however, the use of β-blockers had no effect on testosterone or BT concentrations. Finally, the possibility that hypertension per se, regardless of drugs and sexual function, is associated with alterations in the gonadal axis has been previously addressed. Hughes et al14 reported that both total and free testosterone levels were lower in 24 untreated individuals with uncomplicated essential hypertension compared with normotensive control subjects. Khaw and Barrett-Connor15 found an inverse relationship between blood pressure and endogenous testosterone in men. Thus, it is possible that the lower testosterone levels present in hypertensive men predispose these men to the evolution of ED.
It has been previously reported that hypertensive subjects with ED take more antihypertensive drugs at any age group.16 This is consistent with the concept that overall, the use of medications contributes to the evolution of impotence in hypertensive subjects. This contention is not supported by our findings because in 70% of our subjects, hypertensive medications were started either long before or long after the emergence of ED. Although the possibility cannot be excluded of a gradually evolving drug effect exerting a deleterious influence months to years after the initiation of treatment, we do not believe that this is likely. The changing profile of antihypertensive drugs (ie, the decrease in the use of diuretics and centrally acting agents) may have contributed to our finding of an apparent lack of association between drug therapy and impotence in the majority of subjects with mild to moderate essential hypertension.
In this study, we could not identify a simple and clear-cut pathophysiological etiology for impotence in hypertensive men. Although we did not design our study to establish whether the prevalence of impotence is increased in hypertension, it suggests that classic determinants of sexual function cannot readily explain any putative increase in the rate of impotence in hypertension. Lower androgen levels, but not overt hypogonadism, is the only difference between hypertensive and normotensive men with ED.
Our results should prompt direct study of the biology of the erectile vascular apparatus in hypertension. Recently, androgens were found to stimulate nitric oxide release in cavernous body slices, whereas lack of androgens led to decreased nitric oxide production.17 18 Subsequent studies should examine whether subtle alterations in nitric oxide production in hypertension abetted by lowering of circulating testosterone within the normal range such as we observed might form the basis for deficient cavernous sinusoidal relaxation and ED.
Selected Abbreviations and Acronyms
|BMI||=||body mass index|
Reprint requests to Naftali Stern, MD, Institute of Endocrinology, Tel Aviv–Sourasky Medical Center, Ichilov Hospital, 6 Weizman St, Tel Aviv 64239, Israel.
- Received April 8, 1996.
- Revision received May 17, 1996.
- Revision received June 17, 1996.
Sowers JR, Nyby M, Jasberg K. Dopaminergic control of prolactin and blood pressure: altered control in essential hypertension. Hypertension.. 1982;4:431-435.
Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventory for measuring depression. Arch Gen Psychiatry.. 1961;4:561-571.
Imperato MC, Ginley J, Shackletone C, Orlic S, Stoner E. C19 and C21 5β/5α metabolic ratios in subjects treated with the 5α reductase inhibitor finasteride: comparison of male pseudohermaphrodites with inherited 5α reductase deficiency. J Clin Endocrinol Metab.. 1990;70:777-782.
Lipson LG. Treatment of hypertension in diabetic men: problems with sexual dysfunction. Am J Cardiol.. 1984;27:46A-50A.
Mills TM, Wiedmeier VT, Stopper VS. Androgen maintenance of erectile function in the rat penis. Biol Reprod.. 1992;46:342-348.