Long-term Effects on Sexual Function of Five Antihypertensive Drugs and Nutritional Hygienic Treatment in Hypertensive Men and Women
Treatment of Mild Hypertension Study (TOMHS)
Problems with sexual function have been a long-standing concern in the treatment of hypertension and may influence the choice of treatment regimens and decisions to discontinue drugs. The Treatment of Mild Hypertension Study (TOMHS) provides an excellent opportunity for examination of sexual function and effects of treatment on sexual function in men and women with stage I diastolic hypertension because of the number of drug classes studied, the double-blind study design, and the long-term follow-up. TOMHS was a double-blind, randomized controlled trial of 902 hypertensive individuals (557 men, 345 women), aged 45 to 69 years, treated with placebo or one of five active drugs (acebutolol, amlodipine maleate, chlorthalidone, doxazosin maleate, or enalapril maleate). All participants received intensive lifestyle counseling regarding weight loss, dietary sodium reduction, alcohol reduction (for current drinkers), and increased physical activity. Sexual function was ascertained by physician interviews at baseline and annually during follow-up. At baseline, 14.4% of men and 4.9% of women reported problems with sexual function. In men, 12.2% had problems obtaining and/or maintaining an erection; 2.0% of women reported a problem having an orgasm. Erection problems in men at baseline were positively related to age, systolic pressure, and previous antihypertensive drug use. The incidences of erection dysfunction during follow-up in men were 9.5% and 14.7% through 24 and 48 months, respectively, and were related to type of antihypertensive therapy. Participants randomized to chlorthalidone reported a significantly higher incidence of erection problems through 24 months than participants randomized to placebo (17.1% versus 8.1%, P=.025). Incidence rates through 48 months were more similar among treatment groups than at 24 months, with nonsignificant differences between the chlorthalidone and placebo groups. Incidence was lowest in the doxazosin group but was not significantly different from the placebo group. Incidence for acebutolol, amlodipine, and enalapril groups was similar to that in the placebo group. In many cases, erection dysfunction did not require withdrawal of medication. Disappearance of erection problems among men with problems at baseline was common in all groups but greatest in the doxazosin group. Incidence of reported sexual problems in women was low in all treatment groups. In conclusion, long-term incidence of erection problems in treated hypertensive men is relatively low but is higher with chlorthalidone treatment. Effects of erection dysfunction with chlorthalidone appear relatively early and are often tolerable, and new occurrences after 2 years are unlikely. The rate of reported sexual problems in hypertensive women is low and does not appear to differ by type of drug. Similar incidence rates of erection dysfunction in placebo and most active drug groups caution against routine attribution of erection problems to antihypertensive medication.
Multiple classes of antihypertensive drugs are available for the treatment of hypertension. Achieving adherence to drug treatment with long-term management can be a challenge, and problems with sexual function can contribute to this difficulty.1 In medical practice, physicians frequently attribute sexual problems to the antihypertensive drug and modify or discontinue medication regimens to address this concern. However, scientific evidence that links antihypertensive drugs to sexual dysfunction in placebo-controlled trials is limited,2 and the question has been raised as to whether the higher rate of sexual dysfunction in hypertensive individuals is due to hypertension or to drug treatment of hypertension. In addition, direct comparison of the effects of major classes of antihypertensive drugs, especially newer agents, on sexual function has not been made. Also, most previous studies have been relatively short term (usually several weeks or up to 1 year) and have mostly examined sexual function only in men.3
In this article, we compare the effects over 4 years of five antihypertensive drugs (representing five classes of antihypertensive agents) and placebo, combined with nutritional-hygienic treatment, on sexual function in men and women aged 45 to 69 years with stage I diastolic hypertension in the Treatment of Mild Hypertension Study (TOMHS). We also explore the relationships of several baseline factors, including age, blood pressure, and previous use of antihypertensive medication, with sexual dysfunction.
TOMHS was a double-blind, placebo-controlled, randomized trial comparing six treatments for long-term care of individuals with stage I hypertension. The study was approved by an institutional review committee, and all participants gave informed consent. Procedures followed were in accordance with institutional guidelines. The design, baseline characteristics, detailed eligibility criteria, 1-year intervention, and final results for major end points have been published.4 5 6 Participants were men (n=557) and women (n=345), aged 45 to 69 years, with stage I diastolic hypertension and without clinical evidence of cardiovascular disease. Participants not taking antihypertensive medication at initial screening (stratum 1) had a diastolic pressure of 90 to 99 mm Hg averaged over three eligibility visits. Participants taking a single class of antihypertensive drug at initial screening (stratum 2) had diastolic pressure of 85 to 99 mm Hg averaged over three visits after withdrawal from medication. Exclusions for other than cardiovascular disease and blood pressure level included body weight 50% above (or 10% below) desirable weight, excess alcohol intake (>28 drinks per week), excess meals (>50%) eaten away from home, and an unwillingness or inability to comply with the lifestyle modification program. Desirable weight was calculated as 90% of the sex-specific weight of people aged 25 to 34 years.7
After randomization, participants were seen at 3-month intervals for a minimum of 4 years. Attendance at follow-up visits was excellent, averaging more than 90% for the four annual visits.
All participants received intensive nutritional-hygienic intervention aimed at weight loss, reduced dietary sodium and alcohol intake (for current alcohol drinkers), and increased leisure-time physical activity. In addition, participants were randomized to receive either placebo (n=235) or one of five drugs representing commonly used classes of antihypertensive agents: (1) a β-blocker (acebutolol, 400 mg/d, n=132), (2) a calcium antagonist (amlodipine, 5 mg/d, n=131), (3) a diuretic (chlorthalidone, 15 mg/d, n=136), (4) an α1-antagonist (doxazosin, 2 mg/d, n=134), or (5) an angiotensin-converting enzyme inhibitor (enalapril, 5 mg/d, n=135). Participants randomized to doxazosin were given a blinded dose of 1 mg/d for the first month according to the manufacturer's protocol. All drugs were administered once a day in the morning.
Medication Step-up Protocol
During follow-up, the initial dose of step 1 medication was doubled if diastolic pressure was ≥95 mm Hg on three successive visits or ≥105 mm Hg at a single visit. If elevated blood pressure persisted, chlorthalidone (15 mg/d and then 30 mg/d, if needed) was added as step 2, unless the participant was randomized to chlorthalidone, in which case enalapril (2.5 mg/d and then 5 mg/d, if needed) was added.
Assessment of Sexual Function
At baseline and at annual follow-up visits, the TOMHS attending physician asked participants questions regarding problems with sexual activity. Specifically, each participant was asked, “During the past 12 months, have you had a problem with sexual activity?” If the participant answered “yes,” additional questions were asked. Men were asked, (1) “During the past 12 months, have you experienced difficulty obtaining an erection?” and (2) “during the past 12 months, have you experienced difficulty in maintaining an erection?” Women were asked, “During the past 12 months, have you experienced difficulty having an orgasm?” Physicians also asked participants whether they had an increase, decrease, or no change in the frequency of sexual activity in the past year.
Nurse interviewers also asked a similar question at each 3-month visit regarding change in the frequency of sexual activity. Higher positive response rates to decreases in sexual activity were obtained from physician inquiries than from nurse inquiries (in both men and women), indicating likely underreporting of sexual problems in data collected by nurses. This report includes only data from responses to questions asked by the physician. Assessment from nurse interviews through 12 months have been reported.5 A brief summary comparing treatment groups for incidence of erection problems was given in the TOMHS final results.6
Assessment of Other Measures
Methods for measurement of blood pressure, collection of timed overnight urine, assessment of physical activity and alcohol drinking, and measurement of body weight have been reported.5 Relative weight was calculated as percentage of desirable weight. Baseline systolic blood pressure (SBP) was calculated as the average of pressures taken at the second and third eligibility visits. Smoking status was assessed by self-report. The current medication prescription for each participant was recorded at each 3-month follow-up visit.
The relationship of baseline factors and prevalence of sexual problems was assessed with the Mantel-Haenszel χ2 test8 with stratification by clinical center and use of antihypertensive medication (stratum) at initial screening. Logistic regression was used for assessment of the relationship of multiple factors simultaneously with sexual function. Differences among treatment groups for incidence of sexual problems during follow-up were assessed with the Mantel-Haenszel χ2 test with stratification by clinical center, stratum, and year of event (1 through 4). Probability values are reported comparing each active treatment with placebo and comparing the five active drug groups (4 df). Pairwise comparisons were made among the active treatment groups if the probability value comparing the active groups was less than .05. Analyses were done separately for incidence through 24 months and through 48 months. A separate analysis was done for men reporting any erection problems at baseline for assessment of whether treatments differed in modifying the rate of existing sexual problems. For this group, the percentage of men reporting an erection problem at any annual visit through 24 and through 48 months is reported; the complement of this percentage is the percentage of men with a reported disappearance of erection problems through the indicated visit. The very low rate of reported sexual problems in women at baseline precluded this type of analysis in women.
All analyses comparing treatment groups were intention-to-treat; ie, participants were classified according to initial randomized treatment assignment.
Baseline characteristics of TOMHS participants have been published in detail.5 Average age was 55 years, 62% were men, 20% were black, and 61% were on antihypertensive drugs at the initial screening visit. Significant changes were made in all lifestyle factors.6 Overall adherence over follow-up to initially assigned medication was high. Among participants randomized to active medication, 81% were on initial drug alone at 24 months (range among active groups, 76% to 90%) and 72% at 48 months (range, 66% to 83%). For participants randomized to placebo, these percentages at 24 and 48 months were 74% and 59%, respectively.
Table 1⇓ gives responses to sexual function questions at baseline for men and women by race, age, use of antihypertensive medication at initial screening (stratum), smoking status, relative weight, alcohol intake, and SBP level. Sexual problems were reported in 14.4% of men and 4.9% of women. Problems with achieving or maintaining an erection were reported by 12.2% of men. Older men (55 years or older) were significantly more likely to experience erection problems than younger men. More detailed age stratification showed that erection problems were substantially higher after age 60 (Figure⇓). Men taking antihypertensive drugs at initial screening (stratum 2) were approximately twice as likely to report erection problems as men not taking drugs (stratum 1). Erection problems were also reported more than twice as often in men with SBP ≥140 mm Hg than in men with SBP <140 mm Hg; the highest rate of erection problems was in men with SBP ≥140 mm Hg (Figure⇓). Stratification by age level showed that SBP ≥140 mm Hg was associated with twice the rate of erection problems for both men younger than 60 years and men 60 years old or older (data not shown). Baseline sexual function in men was not significantly related to obesity, alcohol intake, smoking, or physical activity level (data for physical activity not shown), although the small percentage of current cigarette smokers reported the highest rate of sexual problems (20%) among the subgroups considered (P=.09 versus nonsmokers). Patterns for decreased frequency of sexual activity in the past year paralleled those for erection problems for age and SBP but not for previous use of antihypertensive medication. Men previously on and not on antihypertensive medication at initial screening reported similar rates of decreased frequency of sexual activity in the previous year, despite the fact that those on medication reported more erection dysfunction.
Few women (n=7) reported a problem having an orgasm (2%), making comparisons among subgroups difficult. All seven of these women were on antihypertensive medication at initial screening. Approximately 11% of women reported a decrease in sexual activity in the previous year; no significant relationship was found between decreased frequency of sex and any baseline factor considered. Higher rates of sexual problems were found in women taking hormone replacement therapy, but this was not significantly different from rates for women not on such therapy.
Multivariate logistic regression analysis showed that age (≥60 years), previous use of antihypertensive medication, and SBP (≥140 mm Hg) were significantly and independently related to erection problems at baseline (Table 2⇓). Previous use of antihypertensive medication and SBP were somewhat stronger independent predictors of erection problems than age.
Findings During the Trial
Age and previous use of antihypertensive medication were also significantly and independently related to incidence of erection problems through 48 months, with relative risks similar to those at baseline (data not shown).
Complaints of sexual dysfunction during follow-up in women were rare. Three women in the placebo group and two women randomized to active treatment reported difficulty in having an orgasm through 48 months (data not shown).
Table 3⇓ shows the incidence of erection problems through 24 and 48 months by randomized treatment group. Included in the analyses are the approximately 88% of men who reported no erection dysfunction at baseline. Of the 69 men reporting incident erection problems sometime during follow-up, 44 (64%) reported problems at a single annual visit only. In general, the rate of erection problems was low, ranging from 6% to 17% across groups through 24 months and 11% to 18% through 48 months. At 24 months, men in the chlorthalidone group experienced the highest incidence of problems obtaining an erection (15.7%), which was significantly higher than placebo (4.9%, P<.01). Among the active treatments, only doxazosin had a lower rate (2.8%) than placebo at 24 months, but this percentage did not differ significantly from that of placebo. The acebutolol, amlodipine, and enalapril groups had incidence rates only slightly higher than the placebo group (P=NS). The test for comparison of the five active drug groups was significant (P=.04); pairwise comparisons among the active groups showed a significant difference between the doxazosin and chlorthalidone groups (P<.01). Differences among groups in incidence rates of problems maintaining an erection were similar to those observed regarding problems obtaining an erection. The rates for any erection problems in the chlorthalidone and placebo groups were 17.1% and 8.1%, respectively (P=.02). Most of the participants in the chlorthalidone group developing erection problems through 24 months experienced the problem in the first year (8 of 12 men); a majority of them continued on diuretic medication.
Differences among groups in the incidence of erection problems through 48 months were smaller than at 24 months (Table 3⇑). The difference in the incidence of problems obtaining an erection was not significant between chlorthalidone and placebo groups (16.9% versus 11.9%, P=.2), and the rate of any erection problem was nearly equal (18.3% versus 16.7%). This narrowing of the difference between the chlorthalidone and placebo groups was due to 11 participants in the placebo group, compared with only 1 in the chlorthalidone group, reporting a new erection problem at 36 or 48 months of follow-up. Of the 11 men randomized to placebo who experienced a new erection problem at 36 or 48 months, 5 had been taking antihypertensive medication in the year before reporting the sexual problem (4 on medication from their private physician). Of the 69 men reporting incident erection problems during the study, 34 were still on their initially assigned (step 1) medications at 48 months, including 8 of 13 men randomized to chlorthalidone.
Changes, withdrawals, or additions of medication appeared to have minimal effects on the intention-to-treat analysis. In the active treatment groups, 38 (79%) of the 48 men reporting incident erection dysfunction during follow-up were taking step 1 medication alone during the year before reporting the dysfunction. Of the other 10 men, 3 were on no medication, 4 were on other medications alone, and 3 were switched from step 1 medication during the previous year. In the placebo group, 16 (76%) were not on any medication during the year before reporting the erection dysfunction.
Rates of erection problems were also compared among treatment groups in the 12% of men who reported an erection problem at baseline (Table 4⇓). There was a uniform finding for five of the six groups, in that only a minority of these men reported erection problems at any visit through 48 months, ranging from 13% to 46%, with the enalapril group having a higher percentage (75%). The rate of apparent cessation of these problems was highest in the doxazosin group, with none of the eight men reporting these problems at 24 months and only one at 48 months.
The incidence of erection problems was not significantly related to change in lifestyle factors, ie, in weight, urinary sodium excretion, alcohol, and physical activity (data not shown). In the placebo group, greater decreases in SBP during follow-up were associated with a lower incidence of erection problems (P=.1); this pattern was not true in an analysis of all active groups combined.
Table 5⇓ shows the incidence of reported decrease in the frequency of sexual activity during follow-up by treatment group and sex. Higher rates of decrease in the frequency of sexual activity through 24 months were reported in men in the enalapril and chlorthalidone groups (22% to 23%); the lowest rate was in the doxazosin group (10%); none of these rates differed significantly from the placebo rate (15%). Differences between groups were smaller at 48 months, consistent with the findings for erection dysfunction.
For all groups, women reported decreased sexual activity less often than men at 24 months (Table 5⇑). At 24 months, rates ranged from 8% (acebutolol group) to 16% (amlodipine group); none of the active drug groups differed significantly from the placebo group. Rates of decrease in sexual activity through 48 months were more heterogeneous than at 24 months, ranging from 10% (acebutolol group) to 26% (amlodipine group, P=.11 versus placebo).
In TOMHS, 12.2% of men reported erection problems at baseline. The prevalence of impotence in TOMHS likely underestimates the prevalence in the general hypertensive population because of initial study exclusions for factors such as diabetes, cardiovascular disease, and high alcohol intake. There is also the possibility of underreporting of sexual dysfunction, despite efforts to minimize this by having physicians query participants. In men, erection problems at baseline were strongly related to age, SBP, and previous use of antihypertensive medication. The rate of erection problems was relatively constant until 60 years of age; for men 60 and older, it was substantially higher. For SBP, the rate was higher, with levels ≥140 mm Hg; men with SBP ≥140 mm Hg had more than twice the rate of erection problems of men with SBP <140 mm Hg. This was the case even though all participants had diastolic pressure levels in the stage I range. This relationship persisted with controls for age and previous use of antihypertensive medication. Higher rates of sexual dysfunction for men on antihypertensive medication at initial screening may be due to the medication itself or to these participants having higher “true” levels (before medication) of blood pressure.
A higher rate of erection problems at older ages has often been reported.9 10 11 12 The relationship of erection problems with blood pressure level, as found in TOMHS, is consistent with the understanding that elevated blood pressure contributes directly to impotence problems, possibly because of increased vascular disease.13 14 Part of the relationship generally found between age and sexual dysfunction may be due to blood pressure differences between older and younger men. In TOMHS, the magnitude of the effect of age on erection problems was reduced but not eliminated after adjustment for SBP.
The TOMHS design provided an opportunity for us to compare five antihypertensive drugs, representing five classes of antihypertensive agents, with placebo with regards to long-term effects on sexual function. The high rate of participants remaining on initially prescribed monotherapy over the study, especially in the first 2 years, allows for comparisons among treatments that have minimal confounding effects because of withdrawal or change in medication. Overall, incidence rates of erection problems were moderately low, approximately 10% after 24 months and 15% after 48 months. Men randomized to chlorthalidone experienced approximately twice the incidence of erection problems after 2 years compared with placebo (17% versus 8%); rates at 48 months were similar between the two groups. New erection problems in the chlorthalidone group after year 2 were rare; only one man reported a new problem in the last 2 years of the trial. New erection problems in the placebo group (and to a certain extent also in the other active treatment groups) were more common in the last 2 years. It is possible that men predisposed to impotence problems when taking chlorthalidone manifest impotency early on, whereas with other drug treatment (or no drug treatment), impotency takes a more natural course, relating to aging and perhaps blood pressure level. The small number of events in TOMHS precludes a definitive conclusion in this regard. Impotence problems did not necessarily require withdrawal of chlorthalidone; 8 of 12 men who reported erection problems through 24 months were taking chlorthalidone at the end of 48 months.
Adverse effects of thiazide and thiazide-like diuretics (eg, chlorthalidone) on sexual function have been reported in other studies15 16 17 18 19 and appear to be related to the drug class. The magnitude of the effect on erection problems reported here is similar to that reported in other studies despite the low dose of chlorthalidone used in TOMHS (15 mg/d starting dose) combined with intensive lifestyle intervention. The mechanism of increased erection problems with thiazide diuretics remains unclear, because diuretics have no known important central nervous system or autonomic nervous system effects. The overall weight loss in TOMHS participants did not ameliorate the effect of chlorthalidone on sexual function, as was observed in the Trial of Antihypertensive Interventions and Management (TAIM) study.18
Other classes of drugs used in TOMHS, the β-blocker (acebutolol), angiotensin-converting enzyme inhibitor (enalapril), and calcium channel blocker (amlodipine), did not appear to affect erection function, with the incidence rate for men randomized to any of these three drugs similar to that for men randomized to placebo. The cardioselective β-blocker acebutolol did not have the adverse effect on sexual function previously reported for nonselective β-blockers such as propranolol.1 Participants randomized to the α-blocker doxazosin experienced the lowest incidence of erection problems, although the lower rate did not differ significantly from the rate with placebo. However, the fact that for men randomized to doxazosin erection problems disappeared throughout the study for 7 of 8 men (87.5%) with erection problems at baseline, compared with 33 of 60 (55.0%) in all other groups combined, suggests that doxazosin may reduce sexual dysfunction in men. The lower rate of erection problems with doxazosin is consistent with results of the α-blocker prazosin, which was also observed in an early study to be related to a very low rate of erection problems.20
Reported decreases in sexual activity in men among treatment groups in general paralleled trends with erection problems, although differences among groups were smaller. A possible exception was the nonsignificantly greater rate of decrease in sexual activity in the enalapril group (see Table 5⇑), which appeared not be to related to erection problems (see Table 3⇑).
In TOMHS, the incidence of reported sexual problems at baseline and during follow-up among women was low. Even for the more general category of reporting decreased frequency of sexual activity, rates for women were appreciably less than for men. This may be due to the limited questions asked regarding sexual problems in women; for example, no separate questions were asked regarding problems with lubrication or with discomfort during intercourse. It may also be partly due to a certain percentage of women not being sexually active because of the lack of an available partner; in TOMHS, 35% of women (compared with 15% of men) were not married at the beginning of the study. It is also possible that the actual rate for women was higher but that women were uncomfortable with reporting these problems. Most study physicians in TOMHS were men, which may have contributed to underreporting in women. Data from other studies on the effects of antihypertensive agents on sexual function in hypertensive women are limited, especially for these drug classes, and studies that have compared sexual dysfunction among drugs usually lacked a control group.3 The rate of decreased frequency of sexual activity among women did not differ significantly by treatment group, although women randomized to amlodipine had a higher rate through 48 months compared with other groups.
A limitation of the present study is the modest sample size (approximately 70 men per treatment group free of erection problems at baseline), thus giving statistical power to detect only moderate to large differences between study groups. Another limitation was that only a few questions were asked; for example, men were not asked whether they had problems with ejaculation, a component of sexual dysfunction that has sometimes been associated with calcium channel blockers.21 Information on whether the participant was sexually active at the beginning of the study was also not gathered.
In TOMHS, the incidence of erection dysfunction in men was relatively low but was higher with diuretic treatment. The incidence of erection dysfunction with chlorthalidone appeared relatively early (first year) and was unlikely to occur after 2 years. In most cases, men reporting erection problems with chlorthalidone were able to remain on the drug. Similar rates of problems in the placebo group and most active drug groups caution against routinely attributing erection problems to antihypertensive treatment. Sexual dysfunction in hypertensive individuals may be related more to hypertension level than to drug treatment. The incidence of sexual problems reported by women was low (as measured in TOMHS), regardless of treatment, suggesting that sexual dysfunction is not a major concern in drug treatment of women with hypertension.
Funding was provided by the National Heart, Lung, and Blood Institute, National Institutes of Health (NIH R01-HL-34767) and by the following pharmaceutical firms: Boehringer-Ingelheim; Merck, Sharp & Dohme; Pfizer Inc; and Wyeth-Ayerst. The authors would like to thank Monika M. Wahi for her effort on this manuscript.
Reprint requests to Richard H. Grimm, Jr, MD, PhD, Shapiro Center for Evidence-Based Medicine, 914 S 8th St, D-5, Minneapolis, MN 55404.
- Received April 8, 1996.
- Revision received May 15, 1996.
- Revision received July 17, 1996.
Neaton JD, Grimm RH, Prineas RP, Stamler J, Grandits GA, Elmer PJ, Cutler JA, Flack JM, Schoenberg JA, McDonald R, Lewis CE, Liebson PR. The Treatment of Mild Hypertension Study: final results. JAMA. 1993:270:713-724.
US Department of Health, Education, and Welfare. Weight and Height of Adults 18-74 Years of Age: United States 1971-74. Hyattsville, Md: DHEW; 1979. DHEW publication (PHS) 79-1659.
Kinsey AC, Pomeroy WB, Martin CF. Sexual Behavior in the Human Male. Philadelphia, Pa: WB Saunders; 1948:804.
Bauer GE, Baker J, Hunyor SN, Marshall P. Side-effects of antihypertensive treatment: a placebo-controlled study. Clin Sci Mol Med. 1978;4(suppl):341S-344S.
Riley AJ, Riley EJ, Davies HJ. A method of monitoring drug effects on male sexual response: the effect of single dose labetolol. J Clin Pharmacol. 1982;14:695-700.
Bansol S. Sexual dysfunction in hypertensive men: a critical review of the literature. Hypertension. 1988;12:1-10.
Pitts NE. The clinical evaluation of prazosin: evaluation of a new antihypertensive agent. Postgrad Med. 1975;Spec No:117-127.
Suzuki H, Tominaga T, Kumagai H, Saruta T. Effects of first-line antihypertensive agents on sexual function and sex hormones. J Hypertens. 1988;6(suppl):S649-S651.