A Twin Study of Erectile Dysfunction

Background  The extent of genetic influence on erectile dysfunction (ED) is unknown. This study determines the contribution of heredity to ED in a sample of middle-aged men.

Methods  A classical twin study was conducted in the Vietnam Era Twin Registry, a national sample of male-male pairs (mean birth year, 1949) who served on active duty during the Vietnam era (1965-1975). A 1999 male health survey was completed by 890 monozygotic (MZ) and 619 dizygotic (DZ) pairs. The prevalence and heritability of 2 self-report indicators of ED, difficulty in having an erection and in maintaining an erection, are estimated.

Results  The prevalence of difficulty in having an erection is 23.3% and in maintaining an erection is 26.7%. Twin correlations for dysfunction in having an erection are 0.35 (95% confidence interval [CI], 0.28-0.41) in MZ and 0.17 (95% CI, 0.09-0.27) in DZ pairs. For dysfunction in maintaining an erection, the twin correlations in MZ and DZ pairs are 0.39 (95% CI, 0.32-0.45) and 0.18 (95% CI, 0.09-0.27), respectively. The estimated heritability of liability for dysfunction in having an erection is 35% and in maintaining an erection is 42%. The heritable influence on ED remained significant after adjustment for ED risk factors.

Conclusions  The present study demonstrates an ED-specific genetic component that is independent of genetic influences from numerous ED risk factors. The results suggest that future molecular genetic studies to identify ED-related polymorphisms are warranted.

ERECTILE DYSFUNCTION (ED), defined as the inability to achieve or maintain an erection sufficient for satisfactory sexual performance,¹ is a highly prevalent condition that is increasing with the number of older men in the population. Several factors have been shown to be related to ED prevalence, including age, hypertension, diabetes, heart disease, cigarette smoking, alcohol use, obesity, and lipid disorders.^2-16 Inverse relationships of ED with higher education and physical activity have also been observed.^2-4,7,11,17 To our knowledge, the role of genetic factors for ED has never been investigated. The purpose of this study was to estimate the prevalence and genetic influence on ED in a sample of middle-aged twin pairs.

Study participants are members of the Vietnam Era Twin (VET) Registry. The VET Registry was established in the early 1980s and consists of 7368 male-male monozygotic (MZ) and dizygotic (DZ) twin pairs, born between the years 1939 and 1957, with both members of the pair having served in the US military during the Vietnam era (1965-1975).^18,19 In 1987, a mail and telephone survey collected information to classify zygosity using an approach with an approximate 95% accuracy rate.²⁰ In total, 4774 pairs responded to the survey, of which 53.5% were MZ pairs, 43.8% were DZ pairs, and 2.7% could not be classified.

In 1999, a male health survey was sent to 10 762 VET Registry members (included 4636 pairs). With 2 mailed follow-ups, a response rate of 49.8% (5361 individuals; 1621 pairs) was achieved. The response rate was slightly higher among the MZ twins (51.2%) compared with the DZ twins (49.1%). Only pairs with ED information from both members are included in the present study, which leads to an analytic sample of 1509 twin pairs. All survey procedures were approved by the Hines VA Cooperative Studies Program Coordinating Center Human Subjects Committee.

The measurement of ED included 2 questions about the difficulty in the past month of (1) having an erection and (2) maintaining an erection. Twins rated their abilities along a 6-point scale ranging from very good to no erection. Responses to the ordinal questions were each significantly (P<.001) correlated with the self-report of a physician diagnosis of ED; for this analysis we chose the self-rated measures because it is not confounded by treatment-seeking behavior and is consistent with previous epidemiologic research.^2,3,5,12,21 Risk factor data for ED were also collected in the survey and included diabetes, hypertension, coronary heart disease, body shape, cigarette smoking, and alcohol use.

The prevalence of ED is estimated (in percentage with 95% confidence interval [CI]) for all twins and for 3 age groupings (42-49, 50-51, and 52-60 years) corresponding to approximate tertiles. For the prevalence estimates, ED is dichotomized comparing those with responses of fair, poor, very poor, and no erection with those with responses of very good and good. In the classical twin study analysis, the within-pair correlation of ED in MZ pairs (sharing 100% of their genes) is compared with the within-pair correlation of ED in DZ pairs (sharing 50% of their genes); MZ correlations greater than DZ correlations imply genetic influence. The correlations are estimated using a multilevel, mixed-effects regression model specifying a probit link function with separate random effects for MZ and DZ twin pairs.²² The significance of the difference in the MZ and DZ within-pair correlations was determined through the likelihood ratio χ² test statistic. Heritability, defined as the percentage of the phenotypic variance in the liability to ED due to genetic factors, was estimated from the within-pair correlations.²³ In this context, it is equivalent to 2(r_MZ − r_DZ). Adjustment for age, diabetes, hypertension, coronary heart disease, body shape (surrogate for obesity), cigarette smoking, and alcohol consumption was made by including these factors as covariates in the regression model. A P value less than .05 was considered significant.

Prevalence was estimated using the SAS System for Windows version 8 (SAS Institute Inc, Cary, NC). Mixed-effects regression modeling was performed using the MIXOR program.²⁴

Complete ED data are available for 890 MZ and 619 DZ pairs. The mean age of the sample at the time of the 1999 survey is 50.5 years in both the MZ and DZ pairs (Table 1). In addition, the distributions of the remaining ED risk factors used for adjustment do not differ significantly by zygosity, with the exception of alcohol use in the month prior to survey completion (P<.01). The prevalence of ED defined as difficulty in having an erection is 23.3% (Table 2). There is a significant trend in the age-specific prevalences, increasing from 21.4% in the youngest group to 24.8% in the oldest group (P = .07). The prevalence of ED defined as difficulty in maintaining an erection is 26.7% and also demonstrates a direct relationship with age (P = .02).

Table 3 presents the unadjusted twin correlations and heritability estimates for both ED measures. Overall, the correlations for difficulty in having an erection are 0.35 (95% CI, 0.28-0.41) for MZ and 0.17 (95% CI, 0.09-0.27) for DZ pairs. With respect to maintaining an erection, the correlations in the MZ and DZ pairs are 0.39 (95% CI, 0.32-0.45) and 0.18 (95% CI, 0.09-0.27), respectively. The differences in the MZ and DZ correlations for both measures of ED are highly significant. The heritability of liability for difficulty in having an erection is estimated to be 35%. For difficulty in maintaining an erection, the estimated heritability is 42%.

Adjustment for age, diabetes, hypertension, coronary heart disease, body shape, cigarette smoking, and alcohol consumption does not appreciably alter the magnitude of the correlations or the significance of the differences in the MZ and DZ correlations (Table 4). The adjusted estimated heritabilities of liability are 29% for difficulty in having an erection and 36% for difficulty in maintaining an erection.

Our results indicate that there is a genetic component involved in the etiology of ED. This component has an effect that is independent of the genetic influences of the established ED risk factors. Many of the factors responsible for ED may exert their influence through genetic as well as environmental mechanisms. These influences include the effects of known risk factors in addition to the effects of factors yet to be discovered. What is known is that age, lower education, diabetes, hypertension, heart disease, cigarette smoking, alcohol use, obesity, lack of physical activity, and lipid disorders most likely contribute to ED development.^1-16 What is not known is the precise nature of the relationships between ED and these factors, their interaction with one another, and whether there are other physical health, psychological, and lifestyle/behavioral factors that contribute to ED risk. The mechanism underlying the inheritance of ED is also not known. Possible candidates are polymorphisms for endothelial nitric oxide synthase (an enzyme involved in the production of nitrous oxide, a neurotransmitter involved in cavernosal smooth muscle relaxation) and for angiotensin-converting enzyme (involved in regulation of the penile cavernous smooth muscle tone).^25-27 Also of particular interest are expression studies of phosphodiesterase genes in human cavernous tissue and the expression of the arginase II gene in the cavernous tissue of patients with diabetes.^28-30

The present study has a number of potential limitations. The response rate was approximately 50%. While this is modest, it is similar to previous studies of this sensitive topic, including the Massachusetts Male Aging Study.^4,31-33 It is unlikely that nonresponse bias is producing the significant heritable effects—to do so would mean that the likelihood of response from ED-concordant MZ pairs relative to all MZ pairs was different than the likelihood of response from ED-concordant DZ pairs relative to all DZ pairs.

The use of self-reported ED is also of concern, yet studies have suggested that self-reported ED is the most appropriate method for assessing this condition.^34,35 In addition, it has been shown that a self-administered questionnaire such as the International Index of Erectile Function (IIEF) can provide a valid diagnosis of ED and its severity along with a valid assessment of treatment-related changes in ED severity.^21,36,37 As an indication of the validity of the ED items used in the present study, we demonstrated a highly significant association with the report of a physician diagnosis of ED.

Another possible limitation of our study is that information on the presence of a sexual partner was not collected, and the questions used in the measurement of ED are not restricted to those with partners. As a consequence, the self-report of ED could be affected by the absence of a sexual partner. However, using marital status as a surrogate for partner availability, a number of studies have suggested that marital status is not significantly related to the prevalence of ED after adjustment for age.^6,7,11,38 Lastly, our sample is relatively young and is composed solely of men who served in the military during the Vietnam era.

The independent, ED-specific genetic influence identified in the present study should be confirmed and further elucidated in longitudinal follow-up studies of our cohort into the years of highest ED prevalence. A better understanding of the etiology of ED can build on these results, concentrating on molecular genetic studies to identify polymorphisms contributing to ED development.

Corresponding author: Jack Goldberg, PhD, VA Puget Sound Health Care System, VET Registry/Seattle ERIC (MS 152E), 1660 S Columbian Way, Seattle, WA 98108 (e-mail: goldie1@u.washington.edu).

Accepted for publication February 21, 2003.

The US Department of Veterans Affairs, Cooperative Studies Program, Research and Development Service (Washington, DC) has provided financial support for the development and maintenance of the Vietnam Era Twin (VET) Registry. Numerous organizations have provided invaluable assistance in the conduct of this study, including the Department of Defense (Arlington, Va); National Personnel Records Center, National Archives and Records Administration (Washington, DC); the Internal Revenue Service (Washington, DC); National Opinion Research Center (Chicago, Ill); National Research Council, National Academy of Sciences (Washington, DC); and the Institute for Survey Research, Temple University (Philadelphia, Pa).

We thank David Penson, MD, MPH, University of Washington School of Medicine, Seattle, for his careful review of the manuscript and his insightful comments. Most importantly, we gratefully acknowledge the continued cooperation and participation of the members of the VET Registry and their families. Without their contribution this research would not have been possible.