Managing Erectile Dysfunction in Hypertensive Patients


Joel Handler, MD, Kaiser Permanente, 411 Lakeview Avenue, Anaheim, CA 92807

Case 1: A 49-year-old man was referred to our hypertension clinic because of uncontrolled hypertension and patient concerns regarding medication-related erectile dysfunction (ED). He had been taking antihypertensive medication for approximately 8 years, but several medications were discontinued due to side effects and the patient perceived that his sexual performance was declining because of antihypertensive medication. Throat scratchiness led to discontinuation of lisinopril; losartan was associated with dose-related headache and dizziness, permitting only 25 mg daily; amlodipine caused headache; and doxazosin was discontinued due to dizziness. On a combination of hydrochlorothiazide (HCTZ) 25 mg daily, atenolol 50 mg daily, and losartan 25 mg daily during the past 2 years, blood pressures (BPs) were 118 to 132 mm Hg/82 to 96 mm Hg. However, erections sufficient to permit successful sexual intercourse had declined from 3 to 4 times to 1 to 2 times per week. He consumed 1 to 3 glasses of wine daily, his body mass index (BMI) was 28 kg/m2, and he exercised on a treadmill for 30 to 60 minutes 3 times each week. He was a nonsmoker.

At the time of his referral, BP was 132/98 mm Hg. There were no carotid or femoral bruits, and a 12-lead electrocardiogram (ECG) was normal. Fasting glucose was 108 mg/dL (normal 70–99 mg/dL), alanine aminotranferase 42 U/L (normal <41 U/L), low-density lipoprotein 80 mg/dL (normal <100 mg/dL), high-density lipoprotein 53 mg/dL (≥40 mg/dL), potassium 4.0 mEq/L (normal 3.5–5.0 mEq/L), and creatinine was 1.2 mg/dL (normal 0.7–1.3 mg/dL). He was advised to consume no more than 1 glass of wine daily and to reduce his calories. Atenolol was changed to nebivolol 20 mg daily without recovery of his prior level of potency or improvement in BP. Prolactin level was 3.8 ng/mL (normal 2.1–17.7 ng/mL) and testosterone level was 484 ng/mL (normal 280–800 ng/dL). He scored 3 of 27 on a Public Health Questionnaire (PHQ-9) screening scale for depression (≤4 indicates minimal depression).1 HCTZ was changed to chlorthalidone 25 mg daily, losartan was discontinued, and nebivolol was changed to aliskiren 150 mg twice daily with a follow-up clinic BP of 122/78 mm Hg, but successful erections remained at a 1 to 2 per week level.

Case 2: A 61-year-old man with treated hypertension was referred to our hypertension clinic because of perceived medication-related ED. An unknown antihypertensive medication had been initiated 3 to 4 years prior to referral. BPs had been variable depending on the patient’s level of anxiety, and he had speech difficulty under stress for which he had initially been hospitalized for suspicion of transient ischemic attack, a diagnosis which was later retracted. A year prior to referral he was switched to losartan monotherapy with good BP control. However, following initiation of losartan 50 mg daily, his potency level rapidly declined over a month from 5 or 6 erections weekly to 1 or 2 weekly. The patient felt that this decline was medication-related.

He was a nonsmoker who consumed 5 glasses of red wine weekly and played golf or tennis most days of the week. BP was 134/82 mm Hg and BMI was 25 kg/m2. The patient was adamant about a change in therapy, and was taken off losartan and started on HCTZ 25 mg daily, which maintained BP control and led to satisfactorily improved erectile function.



At a National Institutes of Health consensus conference in 1993, because impotence was felt to be a vague and imprecise term, it was recommended that ED be used instead to describe male inability to achieve an erect penis. However, sexual dysfunction is used more generally to describe impairment in sexual desire, orgasmic capacity, and ejaculatory ability.2 Even ED encompasses a quantitative spectrum of abilities subject to individual patient expectation regarding frequency of successful erections. The spectrum of sexual dysfunction and the interplay of patient expectation often makes it difficult to interpret and compare studies on this topic and to decide prevalence.3,4 ED is also significantly underreported5 and prevalence estimates vary according to reporting techniques, which include direct interviews and various types of questionnaires. Direct interviews elicit lower reporting rates and patients may be less likely to report sexual difficulties to interviewers of the opposite sex.6

Vascular Origin and Risk Factors for ED

ED is often a disease of vascular origin. Normal physiology begins with central brain arousal and proceeds to transmission via preganglionic parasympathetic nerves, resulting in the local release of nitric oxide in the penis, which causes vasodilitation and filling of the cavernosal arteries.3 Female arousal also involves release of nitric oxide in the genitalia.3 Generalized atherosclerotic disease associated with hypertension can interrupt normal erectile physiology, and for that reason, coronary artery disease is commonly associated with vasculogenic ED.3 The relative risk of ED in patients with hypertension compared with normotensive patients ranges from 1.3 to 6.9.3 Patients with additional risk factors for cardiovascular disease, such as diabetes, have a greater prevalence of vasculogenic ED.7,8 As a surrogate for more severe atherosclerotic disease associated with difficult-to-control hypertension, an increasing number of antihypertensive medications is proportional to increasing rates of ED.9 Excess alcohol, prostate disease, use of tobacco, heart disease, and arthritis also reduce sexual function.9

Given the association with cardiovascular disease, evaluation of the hypertensive patient with ED needs to consider increasing age and systolic BP as the most important associated factors. In the Treatment of Mild Hypertension Study (TOMHS) of men and women aged 45 to 69 years without clinical cardiovascular disease, baseline sexual dysfunction was determined by physician and nurse interviewers asking the question “During the past 12 months, have you had a problem with sexual activity?” and follow-up questions if the initial response was “yes.” There was a positive report in 14.4% of men and 4.9% of women.10 Multiple logistic regression analysis showed that the odds ratio of baseline sexual dysfunction was 1.72 for age 60 years and older and 2.17 for systolic BP ≥140 mm Hg10 (Figure 1).

Figure 1.

 Percentage of men at baseline who reported problems with obtaining and/or maintaining an erection, by age and systolic pressure. Adapted with permission.10

How Definitive Is the Thiazide Association With ED?

Thiazide diuretics are commonly thought to disproportionally predispose to ED and sexual dysfunction compared with other antihypertensive agents, but a review of the evidence suggests that may not be the case. In the TOMHS study, chlorthalidone 15 mg was compared with placebo, acebutolol 400 mg, doxazosin 2 mg, amlodipine 5 mg, and enalapril 5 mg over 4 years.11 In this double-blinded clinical trial, a significant decline in sexual dysfunction was reported at 2 years with chlorthalidone compared with placebo that did not occur with the other comparators.10 However, there was no significant difference at 4 years. Most often, reports of sexual dysfunction on chlorthalidone did not lead to medication discontinuation. Between 12 and 48 months into the trial, there were 11 new cases of sexual dysfunction on placebo compared with only 1 new case on chlorthalidone.10 It appears from this study that chlorthalidone may accelerate age-related decline in sexual function to a mild degree in the first year of therapy, usually not enough to lead to drug discontinuation, without any difference compared with placebo in longer-term sexual function.

Long-term effects on sexual function were also reported in the Hypertension Detection and Follow-up Program (HDFP) comparing more aggressive “stepped care” therapy vs standard “referred care.” The first drug in stepped care was chlorthalidone 25 to 100 mg daily.12 The 4.4% discontinuation rate of chlorthalidone due to impotence and the 1.0% discontinuation rate due to decreased libido were not appreciably different from other drugs used in the HDFP.13 The authors noted that because multiple-drug therapies were discontinued simultaneously in some patients with side effects, the probability of a real association between a drug and particular side effect was small.13

The Medical Research Trial (MRC) was an older trial that used a higher diuretic dose than is currently recommended or available.14 A total of 22.6% of men reported impotence while taking bendrofluazide 10 mg daily compared with 10.1% of men taking placebo.15 However, MRC was a single-blind study, and physician reporting of side effects may have been biased to follow patterns of perceived medication association. Short-term ED questionnaire studies comparing exposure to thiazide vs placebo have had conflicting results.16,17 Comparing 6 antihypertensive drugs with placebo, a Veteran’s Affairs Hospital (VAH) trial did not detect drug-related impotence when HCTZ 12.5 to 50 mg was a comparator.18 Given what we know of the pathophysiology of male ED, the lack of thiazide effect on central nervous system, autonomic, and hormonal function makes it difficult to arrive at a plausible explanation for thiazide-related ED.19

An Important Role for Healthy Behaviors

Healthy behaviors such as weight management and exercise appear to mitigate drug-related ED. Although there was no benefit of weight reduction in overweight patients in TOMHS,10 weight loss, but not sodium reduction, blocked the unfavorable effect of chlorthalidone on sexual function in the Trial of Antihypertensive Interventions and Management (TAIM), a multicentered randomized placebo-controlled trial.20 Other studies have confirmed that both weight management and exercise can improve ED in middle-aged men.21–23

How Sound Is the Evidence Favoring Antihypertensive Drugs That May Improve ED?

A few classes of antihypertensive drugs have been promoted for possible beneficial effects on sexual function, particularly the peripheral adrenergic blockers, angiotensin receptor blockers (ARBs), and nebivolol. Angiotensin-converting enzyme (ACE) inhibitors and calcium channel blockers have been generally regarded as neutral.3,10 In the TOMHS study, although there was no significant difference in sexual dysfunction at either 2 or 4 years between placebo and doxazosin, there were a few cases where baseline ED was reversed on doxazosin.10 A few similar cases have been reported in other studies, perhaps due to drug effect reducing penile sympathetic nervous system input.4,24 However, peripheral adrenergic blockers should not be used as antihypertensive monotherapy,25 and the overall side effect profile affecting quality of life is not as attractive as other drugs.18

Pharmaceutical industry interest in the area of antihypertensive drug-related sexual dysfunction has been focused on ARBs and nebivolol. An increasingly accepted assessment instrument, the International Index of Erectile Dysfunction (IIEF), has been developed with pharmaceutical sponsorship to promote further research in this area.26 Although ARB action against angiotensin II is a reasonable physiologic explanation for improved tumescence, ACE inhibitors act in a similar fashion. It is difficult to explain ARB-related increased “sexual fantasies” in one study,27 and perhaps not so difficult to explain ARB-improved sexual function in an unblinded study with pharmaceutical grant sponsorship.28 The Antihypertensive Treatment and Lipid Profile in a North of Sweden Efficacy Evaluation (ALPINE) study was unable to detect a difference in sexual satisfaction between candesartan and hydrochlorothiazide.29 In all, literature favoring ARB antihypertensive treatment for sexual dysfunction is probably insufficient to recommend a proposed algorithm for switching to ARBs when dysfunction arises.3,30,31 Although ARBs may be the only antihypertensive medication class not to mention ED in the package insert,32 ED has been described in an ARB postmarketing study.33

Nebivolol is a new third-generation β-blocker with the additive effect of vasodilatation attributed to generation of nitric oxide. There have been a few pharmaceutical industry–sponsored studies showing reduced ED in small numbers of men treated with nebivolol, based on the physiologic rationale that local generation of nitric oxide is the final step in the pathway for penile erection.34–36 Larger studies will be necessary to determine whether this potential benefit is in fact significant.

How Strongly Are β-Blockers Related to ED?

First- and second-generation β-blockers have been widely associated with ED, but, again, the evidence for this association is not striking. In the previously noted MRC trial, there was no significant increase in ED compared with placebo using doses of propanolol as high as 240 mg daily.14 In a large meta-analysis of 15 randomized trials using β-blockers, which included more than 35,000 patients, the increased risk of sexual dysfunction was of borderline statistical significance, with a relative risk of 1.10 and a 95% confidence interval of 0.96 to 1.25.37 In 4 trials where withdrawal of study medication was noted, the annual absolute increase in β-blocker withdrawal due to sexual dysfunction was 2 per 1000 patients per year.37 Most of these trials included in this meta-analysis used high doses of first-generation β-blockers. In the TOMHS trial, there was no adverse effect on sexual function compared with placebo using acebutalol, a cardioselective β-blocker.10 In the VAH 6-drug trial, atenolol 25 to 100 mg daily was not associated with increased impotence compared with placebo.18

How Important Is Psychogenic ED?

Psychogenic ED has been estimated to comprise 70% of ED up to age 35 and 10% of ED past age 50, as vascular disease progresses.7“Nocebo” effects significantly influence patient reaction to medication, and the volume of possible side effects listed in package inserts increase the likelihood of adverse reactions due to such nocebo effects.38,39 As one patient reported in a focus group trying to determine the cause for statin drug nonadherence, “If you read those sheets they give you, you wouldn’t take anything. It’s scary.”40

A study by Silvestri and colleagues41 examined a nocebo effect causing ED in 3 groups of men aged 52±7 years with newly diagnosed hypertension or angina who were then administered blinded tablets of atenolol 50 mg daily. None had ED at baseline. All patients were advised that they were being prescribed a medication to reduce their chance of heart attack. Group A did not receive any additional information. Group B was told they were being given atenolol, which was a β-blocker. Group C was given the same information as group B and in addition instructed that “it may cause ED.” Results of the 90 days IIEF questionnaire showed ED incidence rates of 3% in group A, 15% in group B (P<.05), and 31% in group C (P<.01) (Figure 2). In the second phase of the trial, new cases of ED in the first phase were randomized to placebo vs sildenafil, and there was no significant difference in ED reversal.41 Therefore, the placebo effect successfully treated the nocebo effect, and ED was shown to be psychogenic in the vast majority of these middle-aged individuals when ED was associated with β-blocker initiation.

Figure 2.

 Incidence of erectile dysfunction (ED) in the different groups receiving β-blockers. Grey into white bar: patients given blinded drug to improve health. Black bar: patients informed that the drug was a β-blocker. Grid bar: patients informed that the drug was a β-blocker and that it might cause ED. SE indicates side effect. Adapted with permission.41

Review of the Two Cases: Likely Vascular and Likely Psychogenic Etiologies

A review of the two case presentations reveals common primary care ED scenarios occurring with antihypertensive medication. The first case probably had progressive vasculogenic ED associated with being a middle-aged overweight hypertensive man. An examination for occult vascular disease and risk, which included a check for bruits, an ECG, a lipid profile, and fasting glucose, were all negative. Both prolactin and testosterone levels were normal. Hyperprolactinemia resulting in ED may be drug-induced or due to hypothalamic-pituitary disease.19,42 Reduced testosterone levels have been associated with reduced sexual activity.43

This patient’s ED was not sufficient to warrant treatment with a phosphodiesterase-5 inhibitor, and continuing patient/physician collaboration to encourage weight reduction and exercise was the preferred course of action. Advancing antihypertensive therapy to achieve more effective BP control would have sufficed in the majority of patients. However, this individual had done a lot of internet research and was convinced that his ED was medication-related. That perception led to more extensive testing and a series of unsuccessful antihypertensive medication substitutions.

The second case represents an example of probable psychogenic ED. This patient’s lifestyle behavior was excellent and the sole atherosclerotic disease risk factor was age. His medical history suggested a tendency toward anxiety-related somatic illness. With longer-term follow-up, HCTZ was discontinued after a couple of years and another agent substituted, because the patient felt that HCTZ made him “tense.” The diagnostic difficulty in deciding antihypertensive drug-related ED vs psychogenic ED due to a change in drug therapy is evident in this case; the latter is far more common. Due to anxiety associated with new drug therapy, any antihypertensive agent may be associated with psychogenic ED.