Introduction

Erectile dysfunction (ED) is a common public health problem affecting millions of men worldwide.¹ There are substantial gaps in our knowledge of the occurrence and etiology of ED.^{2, 3} The use of tobacco and consumption of alcohol are major public health problems worldwide, and the available evidence on the role of smoking^{2, 3, 4, 5, 6, 7} and alcohol consumption^{4, 5, 8, 9, 10, 11, 12} on erectile function is controversial.

Few population-based follow-up studies have been published on the effect of smoking and alcohol consumption on erectile function. The association between these factors and ED has been assessed mainly in prevalence studies, which have serious limitations for etiological purposes. Only one population-based follow-up study has evaluated the role of alcohol intake on erectile function and failed to show an effect of alcohol consumption on ED.⁴ Only two longitudinal studies have evaluated the effect of smoking on erectile function. One showed clearly negative result³ and the other a negative overall result, but doubling of risk in a subgroup of men free of vascular disease at baseline.⁴ We estimated the effect of sociodemographic and life-style factors on the incidence of ED in a population-based sample of the 1130 men who were free of ED at baseline and followed up for a period of 5 y.

Materials and methods

Details of the Tampere Aging Male Urological Study population have been previously described elsewhere.^{13, 14, 15} In this population-based follow-up study, the target population comprised all 3152 men born in 1924, 1934 or 1944 residing in Tampere or 11 surrounding municipalities in 1994, identified from the National Population Register. Information was collected by means of a mailed self-administered questionnaire comprising items on sociodemographic status, life-style factors, medical conditions and medications, diet, physical activity, erectile capacity and concern about erection problems.

The questionnaire was mailed to all 3152 men in the target population during the first quarter of 1994. Nine men had died before receiving the questionnaire, leaving 3143 eligible subjects. An identical questionnaire was sent 3 months later to the 1433 men who did not respond to the first. A total of 2198 completed questionnaires (70%) were returned. Of them, 257 were excluded from the study, 244 due to missing data with regard to erectile function and 13 as the respondents were institutionalized or unable to respond independently. Hence, 1941 men (62%) were included in the study at baseline.

Similar questionnaires were sent 5 y later in May 1999 to 2864 men with a reminder to the 1162 who did not respond to the first within 3 months. Between 1994 and 1999, 262 men had died, six had emigrated and 38 no longer had a permanent address in the Population Register.

Overall, 1683 (53.5%) men responded to both baseline and follow-up inquiries. Of them, 241 were excluded because of missing data on erection function and 1442 (46%) were included in the follow-up sample. The 1130 men free of ED at baseline were included and the 312 with ED excluded for this analysis. The 306 men free of ED at baseline were lost to follow-up or without data on erectile function in 1999. We also estimated the effect of smoking on ED in a subgroup of 762 healthy men who were free of cerebrovascular disease, heart disease and diabetes or related medications and did not undergo prostatectomy or have spinal cord or urinary tract injury at both surveys.

ED was defined according to the National Institutes of Health Consensus Panel on ED as the inability to achieve or maintain an erection sufficient for satisfactory sexual function.¹⁶ Subjects were asked two questions on their erectile capacity: 'Have you had trouble getting an erection before intercourse begins?' and 'Have you had trouble maintaining an erection once intercourse has begun?' The two questions were combined as follows to classify the severity of ED. Normal erectile function: neither difficulty in achieving nor maintaining an erection. Minimal ED: some difficulties in achieving and/or maintaining an erection. Moderate ED: fairly frequent difficulties in achieving and/or maintaining an erection. Complete ED: intercourse does not succeed at all. ED was categorized as none or minimal vs moderate or complete for analysis.

Self-reported smoking history was obtained from the questionnaire. Men were defined as smokers if their history of smoking had lasted for at least 1 y, and as never smokers if they had never smoked or smoked for less than 1 y. Weekly consumption of alcohol was estimated in grams of absolute alcohol from self-reported consumption of beer, wine and spirits. The estimate was based on approximation of a bottle of beer, a glass of wine and other drinks as containing 11.5 g of absolute alcohol. Similarly, a bottle of wine was estimated to contain 6.3 glasses and a bottle of spirit 12.5 drinks. Coffee drinking was recorded in number of cups per day.

The incidence of ED was calculated by dividing the number of new cases occurring between baseline and follow-up surveys by the number of person-years of follow-up. Person-years were estimated by multiplying the number of men who did not develop ED by 5 y (follow-up period) and the number of those who developed ED by 2.5 y (midpoint of follow-up period).

Incidence is expressed as the number of cases per 1000 person-years. The rate ratios (RRs) were defined as the incidence of ED among men with the determinant at baseline divided by the incidence of ED in men without such characteristic. The 95 percent confidence intervals (CI) for the rates and RRs were estimated under the assumption that the number of new cases is a Poisson distributed variable using the STATA statistical software package. A multivariate Poisson regression model was used to evaluate whether selected sociodemographic and life-style factors at baseline predicted for ED at follow-up.

Tampere University Hospital and Tampere City ethical review committees approved the study protocol.

Results

Men who participated in the follow-up survey differed in some respects from those for whom no information on erectile function was available in 1999 (Table 1). The 306 men without follow-up information were older (on average 3 y older), less educated, more frequently single or divorced and more likely to live in rural or semirural areas. They also reported current smoking more frequently at baseline than the participants.

Table 1 - Distribution of sociodemographic and life-style factors among men free of ED at baseline and those without data at follow-up.

Full table

The annual incidence of ED increased 100% with each one-decade increment in age and was higher after the age of 60 y (Table 2). The incidence of ED was higher in obese men compared with those who had normal weight (RR=1.7, 95% CI: 1.1–2.5), but overweight men were not at higher risk of ED. Education, urban–rural residence, marital status, alcohol intake and coffee consumption did not have effect on ED (Table 3). There was a small, nonsignificant excess risk of ED in current (RR=1.5, CI: 0.9–2.2) and in past smokers (RR=1.2, CI: 0.9–1.7).

Table 2 - Annual incidence of ED by sociodemographic and life-style factors in the Tampere Aging Male Urological Study, 1994–1999.

Full table

Table 3 - Rate ratios for ED by sociodemographic and life-style factors, with adjustment for age and all other factors (full model).

Full table

The incidence of ED also increased weakly in current smokers, in a subgroup of healthy men who were free of comorbidity at both surveys. The RR after controlling for effects of sociodemographic and life-style factors was 1.3 (95% CI: 0.8–2.1) and 1.1 (95% CI: 0.7–1.7) for current and former smokers, respectively.

Discussion

The incidence of ED increased strongly with age, especially after the age of 60 y. It increased 100% by each one-decade increment in age for 50- to 75-y-old men. The incidence of ED increased in obese men and in those who were current smokers. Current smokers free of comorbidity were also at higher risk of ED, but not ex-smokers.

The men without follow-up information were on average older than the participants, which may have placed them at a higher risk of ED than the study sample. Therefore, our figures may underestimate the true incidence in the population; however, the RRs are likely to be unbiased. Furthermore, a longitudinal study format can provide more valid estimates on the effects of risk factors than cross-sectional prevalence studies that are prone to selection bias and unable to distinguish between determinants of disease onset and those predicting survival with disease. The prevalence of smoking in the follow-up sample was similar to that among Finnish men of these ages in general, suggesting representativeness of the study population.¹⁷

The Massachusetts Male Ageing Study (MMAS)² and a Brazilian study³ are the only population-based follow-up studies published so far. An incidence of moderate or complete ED was 26 cases per 1000 person-years in the MMAS, 39 in the current study and 66 in the Brazilian study. Those and the present study showed an increased incidence of ED with age. Marital status and alcohol consumption did not have an effect on ED in the MMAS or present study. The age-specific rates in the current study were similar to those of the MMAS. The incidence of ED in the MMAS was 30 cases per 1000 person-years for ages 50–59 y and 46 for ages 60–65 y. It was 22 cases per 1000 person-years for 50–55 y and 49 for ages 60–65 y in the present study.

Education is commonly used as a proxy of socioeconomic status in population survey. The association between socioeconomic status and ED has not been well established. Education^{2, 11, 12, 18, 19, 20} and household income^{2, 3, 12, 21} have been inversely related to ED in some studies. The effect of socioeconomic status on ED is partly mediated by life-style factors and medical conditions.²² Higher socioeconomic status has been linked to better health,²³ and low education is a marker of high prevalence of nondiagnosed diseases. In the present study, education did not have a significant effect on the incidence of ED.

Association between smoking and erectile function has been reported mainly in prevalence studies that have considerable weaknesses for elucidating the etiology of ED. Meta-analyses have shown that impotent men are significantly more likely to be current smokers than men in the general population (prevalence of smoking 40 vs 28%).^{6, 7} Yet, the results are completely equivocal.^{8, 9, 11, 12, 20, 24, 25} Of the two population-based follow-up studies, smoking did not have an effect on erectile function in the Brazilian study. In the MMAS, smoking was not associated with ED in the entire study population,² but when an effect of smoking was confined to a subgroup of men without diabetes, heart disease or related medications and prostate operation, smoking doubled the risk of ED.⁴ Only one-third of the baseline sample was included for the analysis, and the men in the subsample were younger and more educated than whole followed-up population. Furthermore, the crude incidence in the subsample was 18%, while in whole sample it was 23%.

Consistent with the MMAS and the cross-sectional study conducted in Italy,²⁶ our findings showed an elevated risk of ED for current smokers in a subsample of men without comorbidity. The relationship between smoking and ED may be due to the relationship with vascular disease,²⁷ which is thought to be a cause of ED.²⁸ The effect of smoking was evident in men without comorbidity. This suggesting that smoking is associated with the vascular damage, but when the damage is present, it does not give any additional risk. In addition, an initial, undiagnosed smoke-induced cardiovascular condition would be present with ED as the first sign.^{26, 29} Furthermore, current smokers who develop comorbidity may have stopped to smoke anymore.

In the present study, the incidence of ED for past smokers was slightly higher than nonsmokers. A recent study suggested that smoking cessation in middle-aged men did not significantly decrease the risk of ED.³⁰ In an earlier study,⁶ the excess risk in former smokers decreased substantially in the initial 2–3 y, but thereafter the risk reduction decreased. In the present study, most (80%) former smokers free of ED at baseline had stopped smoking at least 10 y before entry into the study. The effect on ED in this study had persisted after several years since the stopping of smoking. The evidence of such an association is likely due to the consistency of the relationship of smoking and endothelial disease and the strength of the association of ED with other endothelial disease.⁶ This may be related to chronic diseases caused by smoking such as cardiovascular and cerebrovascular diseases. Overall, the results of follow-up studies on smoking and ED remain conflicting.

Also, the association between alcohol consumption and erectile function is controversial.³¹ Cross-sectional studies have yielded partly conflicting results.^{5, 8, 9, 20} The effect of excessive alcohol consumption on the prevalence of minimal ED has been small,⁵ and a lower prevalence of ED has been reported among moderate drinkers (one to two drinks per day) than nondrinkers.^{10, 11, 12, 25} Only a single population-based follow-up study has evaluated the effect of alcohol consumption on the incidence of ED and it failed to show a clear effect.⁴ The present study showed that alcohol consumption does not have an effect on the incidence of ED.

Obesity was^{25, 30} or was not⁵ an independent predictor of ED. Obesity imposes a risk to vasculogenic impotence by developing chronic diseases of diabetes, hypertension, heart disease and hyperlipidemia.

In the previous studies, caffeine had a protective effect on ED^{12, 32} or was a risk factor for ED.³³ It has also been shown that coffee intake does not have an effect on ED,^{11, 34} and it is also supported by this study.

In a population-based follow-up study, we evaluated the effects of sociodemographic and life-style factors on the incidence of ED. We found that obesity and current smoking increase the incidence of ED. Current smokers without comorbidity were also at risk of ED. Sociodemographic and life-style factors, except age and obesity seem to be of limited importance in the etiology of ED.

References

1. Ayta IA, McKinlay JB, Krane RJ. The likely worldwide increase in erectile dysfunction between 1995 and 2025 and some possible policy consequences. BJU Int 1999; 84: 50–56. | Article | PubMed | ChemPort |
2. Johannes CB et al. Incidence of erectile dysfunction in men 40 to 69 years old: longitudinal results from the Massachusetts Male Ageing Study. J Urol 2000; 163: 460–463. | Article | PubMed | ISI | ChemPort |
3. Moreira Jr ED et al. Incidence of erectile dysfunction in men 40 to 69 years old: results from a population-based cohort study in Brazil. Urology 2003; 61: 431–436. | Article | PubMed | ISI |
4. Feldman HA et al. Erectile dysfunction and coronary risk factors: prospective results from the Massachusetts Male Aging Study. Prev Med 2000; 30: 328–338. | Article | PubMed | ISI | ChemPort |
5. Feldman HA et al. Impotence and its medical and psychosocial correlates: results of the Massachusetts Male Ageing Study. J Urol 1994; 151: 54–61. | PubMed | ISI | ChemPort |
6. McVary KT, Carrier S, Wessells H. Smoking and erectile dysfunction: evidence based analysis. J Urol 200; 166: 1624–1632.
7. Tengs TO, Osgood ND. The link between smoking and impotence: two decades of evidence. Prev Med 2001; 32: 447–452. | Article | PubMed | ISI | ChemPort |
8. Morales AM et al. Prevalence and independent risk factors for erectile dysfunction in Spain. Results of the epidemiologia de la dysfunction erectile masculine study. J Urol 2001; 166: 569–574. | Article | PubMed | ChemPort |
9. Green JS et al. An investigation of erectile dysfunction in Gwent, Wales. BJU Int 2001; 88: 551–553. | Article | PubMed | ChemPort |
10. Rimm EB et al. Body weight, physical activity, and alcohol consumption in relation to erectile dysfunction among US male health professionals free of major chronic diseases. In: Proceedings of the American Urological Association 95th Annual Meeting, Atlanta, April 29 May 4, 2000. J Urol 2000; 163(Suppl): 15.
11. Nicolosi A et al. Epidemiology of erectile dysfunction in four countries: cross-national study of the prevalence and correlates of erectile dysfunction. Urology 2003; 61: 201–206. | Article | PubMed | ISI |
12. Akkus E et al. Prevalence and correlates of erectile dysfunction in Turkey: a population-based study. Eur Urol 2002; 41: 298–304. | Article | PubMed | ISI |
13. Koskimaki J et al. Effect of erectile dysfunction on frequency of intercourse: a population based prevalence study in Finland. J Urol 2000; 164: 367–370. | Article | PubMed |
14. Shiri R et al. Prevalence and severity of erectile dysfunction in 50 to 75-year-old Finnish men. J Urol 2003; 170: 2342–2344. | Article | PubMed | ISI |
15. Shiri R et al. Effect of chronic diseases on incidence of erectile dysfunction. Urology 2003; 62: 1097–1102. | Article | PubMed | ISI |
16. National Institutes of Health Consensus Conference. Impotence: National Institutes of Health Development Panel on impotence. JAMA 1993; 270: 83–90. | PubMed | ISI |
17. Helakorpi S et al. Health behavior among Finnish adult population. Natl Public Health Instit, Helsinki 1994; B8: 13–95.
18. Moreira Jr ED et al. Prevalence and correlates of erectile dysfunction: results of the Brazilian study of sexual behavior. Urology 2001; 58: 583–588. | Article | PubMed |
19. Giuliano F et al. Prevalence of erectile dysfunction in France: results of an epidemiological survey of a representative sample of 1004 men. Eur Urol 2002; 42: 382–389. | Article | PubMed | ISI |
20. Parazzini F et al. Frequency and determinants of erectile dysfunction in Italy. Eur Urol 2000; 37: 43–49. | Article | PubMed | ISI | ChemPort |
21. Laumann EO, Paik A, Rosen RC. Sexual dysfunction in the United States: prevalence and predictors. JAMA 1999; 281: 537–544. | Article | PubMed | ISI | ChemPort |
22. Aytac IA et al. Socio-economic factors and incidence of erectile dysfunction: findings of the longitudinal Massachusetts Male Aging Study. Soc Sci Med 2000; 51: 771–778. | Article | PubMed | ISI | ChemPort |
23. Link BG, Phelan J. Social conditions as fundamental causes of disease. J Health Soc Behav 1995, Spec No: 80–94.
24. Mannino DM, Klevens RM, Flanders WD. Cigarette smoking: an independent risk factor for impotence? Am J Epidemiol 1994; 140: 1003–1008. | PubMed | ISI | ChemPort |
25. Moreira Jr ED, Bestane WJ, Bartolo EB, Fittipaldi JA. Prevalence and determinants of erectile dysfunction in Santos, southeastern Brazil. Sao Paulo Med J 2002; 120: 49–54. | PubMed |
26. Mirone V et al. Cigarette smoking as risk factor for erectile dysfunction: Results from an Italian epidemiological study. Eur Urol 2002; 41: 294–297. | Article | PubMed | ISI |
27. Shabsigh R, Fishman IJ, Schum C, Dunn JK. Cigarette smoking and other vascular risk factors in vasculogenic impotence. Urology 1991; 38: 227–231. | Article | PubMed | ISI | ChemPort |
28. Jeremy JY, Mikhailidis DP. Cigarette smoking and erectile dysfunction. J R Soc Health 1998; 118: 151–155. | PubMed | ChemPort |
29. Kirby M, Jackson G, Betteridge J, Friedli K. Is erectile dysfunction a marker for cardiovascular disease? Int J Clin Pract 2001; 55: 614–618. | PubMed | ISI | ChemPort |
30. Derby CA et al. Modifiable risk factors and erectile dysfunction: can lifestyle changes modify risk? JB Urol 2000; 56: 302–306. | Article |
31. Bortolotti A, Parazzini F, Colli E, Landoni M. The epidemiology of erectile dysfunction and risk factors. Int J Androl 1997; 20: 323–334. | Article | PubMed | ISI | ChemPort |
32. Diokno AC, Brown MB, Herzog AR. Sexual function in the elderly. Arch Intern Med 1990; 150: 197–200. | Article | PubMed | ISI | ChemPort |
33. Shaeer KZ et al. Prevalence of erectile dysfunction and its correlates among men attending primary care clinics in three countries: Pakistan, Egypt, and Nigeria. Int J Impot Res 2003; 15(Suppl 1): S8–S14. | Article | PubMed | ISI |
34. Berrada S, Kadri N, Mechakra-Tahiri S, Nejjari C. Prevalence of erectile dysfunction and its correlates: a population-based study in Morocco. Int J Impot Res 2003; 15(Suppl 1): S3–S7. | Article | PubMed |