If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Reprint requests: Alfredo Nicolosi, M.D., Ph.D., Department of Epidemiology, Institute of Biomedical Technologies, National Research Council, Via Fratelli Cervi 93, Segrate, Milan 20090, Italy
Affiliations
Department of Epidemiology, Institute of Biomedical Technologies, National Research Council, Milan, ItalyGertrude H. Sergievsky Center, Columbia University School of Public Health, New York, New York, USA
Specialist Reproductive Research Center, National Population and Family Development Board, Ministry of Women and Family Development, Kuala Lumpur, Malaysia
To measure the prevalence of erectile dysfunction (ED) in community-based populations in Brazil, Italy, Japan, and Malaysia and to study its association with the demographic characteristics, medical conditions, and health-related behavior.
Methods
In each country, a random sample of approximately 600 men aged 40 to 70 years was interviewed using a standardized questionnaire. All the data were self-reported. ED was assessed by the participants’ “ability to attain and maintain an erection satisfactory for sexual intercourse,” and the men were classified as not having ED if they answered “always” and as having mild, moderate, or complete ED if they answered “usually,” “sometimes,” or “never,” respectively.
Results
The age-adjusted prevalence of moderate or complete ED was 34% in Japan, 22% in Malaysia, 17% in Italy, and 15% in Brazil. The overall age-specific prevalence of moderate or complete ED was 9% for men aged 40 to 44 years, 12% for 45 to 49 years, 18% for 50 to 54 years, 29% for 55 to 59 years, 38% for 60 to 64 years, and 54% for those 65 to 70 years. The increased risk of ED was associated with diabetes, heart disease, lower urinary tract symptoms, heavy smoking, and depression and increased by 10% per year of age. It was inversely associated with education, physical activity, and alcohol drinking.
Conclusions
ED is an international problem, the prevalence and severity of which increases with age. Despite national variations in prevalence, uniform associations were found between ED and medical conditions and lifestyle habits.
Erectile dysfunction (ED), “the inability to attain and/or maintain an erection satisfactory for sexual intercourse,”
is a common problem of middle-age and older men. The prevalence of ED has been described in some communities in the United States (Olmsted County, Minnesota,
These results are difficult to compare because the investigators used different definitions of ED and sampled men within different age ranges. We undertook this cross-national survey using a single questionnaire to determine the prevalence and correlates of ED in diverse geographic, ethnic, and cultural populations.
Material and methods
The study was a population-based survey of men aged 40 to 70 years in Brazil, Italy, Japan, and Malaysia conducted between October 1997 and June 1998. On the basis of previous estimates, we determined that a total sample size of 2400 men would be adequate to yield stable age-specific estimates of the prevalence of ED. The sampling of households in each country was conducted in such a way as to compensate for the probable number of absences of eligible men and allow for nonresponses. All the sampled households were interviewed, and approximately 600 eligible men were included from each country.
A structured questionnaire was developed in consultation with the investigators, and two focus groups were held in each country to allow for adjustments reflecting local customs and culture. The method of data collection was different among the countries to elicit the most valid response given the cultural differences. The men were interviewed in person in Brazil, by telephone in Italy, by telephone and/or in person in Malaysia, and by means of self-administered, mailed questionnaires in Japan.
ED was assessed by a single question: “Using the following categories, how would you describe yourself?: Always/usually/sometimes/never able to get and keep an erection adequate for satisfactory sexual intercourse.” The responses were considered to represent “no ED,” “mild ED,” “moderate ED,” and “complete ED,” respectively. We focused on the men with moderate (ie, “sometimes able”) or complete ED (ie, “never able”), because the “usually able” category may be a mix of men with intermittent situational problems or mild persistent ED. This conservative choice would underestimate the prevalence of ED.
The questionnaire included medical, anthropometric, sociodemographic, and lifestyle data. The information was self-reported and no attempt was made to validate respondents’ answers with their medical records. The men were classified as having a specific disease if they reported having received a physician’s diagnosis or if they were taking medications for the disease in question. Prostate conditions included benign prostatic hyperplasia, a history of prostatitis, prostate surgery, and prostate cancer. Lower urinary tract symptoms (LUTS) were assessed using the International Prostate Symptoms Score and classified as absent or minor (a score of 7 or less), moderate (score of 8 to 19), or severe (score of 20 to 35).
with scores ranging from 5 to 20. A score 9 or greater (corresponding to the upper quintile) was used as indicative of the presence of depressive symptoms. Smokers included men who smoked tobacco and were classified as current or former smokers. Current smoking was analyzed on the basis of the number of tobacco units (cigarettes, cigars, pipes) consumed per day. The number of alcoholic drinks per week was calculated on the basis of the reported number of drinks of beer, wine, or spirit, weighted by the alcoholic content (beer × 0.4 + wine + spirits × 2.5). The degree of physical activity was classified as “less than average,” “average,” or “more than average,” depending on the extent and frequency at work and during leisure time. Men were considered sexually active if they reported having sexual intercourse or masturbating at least once during the preceding 6 months.
The crude and age and country-adjusted odds ratios (ORs) were calculated for each covariate. Logistic regression analysis was used to estimate the association between ED and the other variables, adjusting for the simultaneous effect of all the variables in the model.
Results
The response rate was 92% in Brazil, 72% in Italy, 51% in Japan, and 16% in Malaysia, for a total of 2513 men. Of this total, 2417 (96.2%) answered the ED question. The prevalence of ED varied across countries, but the age distribution was similar (Fig. 1). Overall, the proportion of men who could “usually” achieve and maintain an erection varied least across all age groups (between 30% and 49%). The proportion of men who could “sometimes” achieve and maintain an erection was approximately 10% until the age of 49, when it steeply increased to 41% in the oldest age group. No men with complete ED were in the youngest age group, but the prevalence of complete ED was 4% between the age of 44 and 49 years and reached 13% after 65 years. The age-adjusted prevalence of moderate or complete ED was 34.5% in Japan, 22.4% in Malaysia, 17.2% in Italy, and 15.5% in Brazil (using the age distribution of the total sample population as the standard).
FIGURE 1Age-standardized prevalence of ED by severity and country.
The age-adjusted prevalences of the demographic, social, medical, and lifestyle characteristics (with the exception of the proportion of single men) were different in the four countries, with at least one country different from the others (P <0.05; Table I). Table II shows the prevalence of the characteristics or conditions listed in Table I among the men with and without ED and the crude and adjusted (by age and country) ORs. The association between men with ED and those who reported being single or having a diagnosis of diabetes, hypertension, heart disease, depression or depressive symptoms, prostate disease, or moderate or severe LUTS was positive. No association was found between ED and current smoking but, among current smokers, the OR was higher in heavy smokers. There was an inverse association between ED and increasing educational level, physical activity, and current alcohol drinking (although not statistically significant).
TABLE ICharacteristics of study population by country (absolute number of men and age-adjusted prevalence)
Characteristic or Condition
Brazil
Italy
Japan
Malaysia
n
%
n
%
n
%
n
%
Marital status
Married
497
82.8
541
91.2
567
90.7
557
94.0
Widow, separated, divorced
70
11.6
31
4.8
26
4.2
13
1.8
Single
35
5.6
20
4.0
25
5.1
30
4.2
Sexually active
587
97.1
567
99.2
531
88.8
575
92.1
Currently employed
427
68.6
354
68.5
490
83.7
526
80.1
Education
High school not completed
398
67.2
322
51.5
210
30.5
256
48.1
High school completed
142
22.6
216
38.5
294
50.5
274
42.8
Degree
62
10.1
53
10.1
112
19.0
62
9.1
Diabetes
34
5.9
53
7.9
66
9.7
70
13.1
Hypertension
154
26.1
133
19.3
146
21.1
107
19.8
Heart disease
49
8.5
67
9.9
46
6.4
38
8.4
Ulcer
48
8.1
97
15.6
120
18.7
30
5.5
Depression
50
8.6
56
9.1
29
4.1
29
5.0
Depressive symptoms (CES-D)
199
33.2
130
22.8
68
18.0
79
11.5
Prostate disease (any)
45
8.2
82
11.8
45
6.3
11
2.0
LUTS
Absent or minor
426
70.0
522
90.4
465
90.9
556
93.9
Moderate
136
23.3
60
9.1
46
8.4
32
5.8
Severe
39
6.7
3
0.6
4
0.7
1
0.4
Smoking status
Never
173
28.2
187
32.2
92
15.1
256
43.6
Former smoker
235
40.4
200
32.2
193
29.3
65
12.1
Current smoker
193
31.4
203
35.6
321
55.6
269
44.4
Tobacco units per day (n)
None
408
68.5
387
64.4
300
45.6
324
55.2
1–30
183
29.7
189
33.4
260
44.2
246
40.1
>30
11
1.9
14
2.2
57
10.2
29
4.7
Alcoholic drinks (n)
None
191
33.0
136
23.2
190
29.8
493
82.1
1–7/wk
253
41.3
103
18.7
208
33.5
87
14.1
≥8/wk
158
25.7
353
58.1
223
36.7
19
3.8
Physical activity
Less than average
52
9.3
44
6.7
118
19.8
40
9.6
Average
89
14.6
167
27.4
178
28.4
110
18.5
More than average
461
76.1
381
65.8
326
51.7
451
71.9
Key: CES-D = Center for Epidemiologic Studies Depression Scale; LUTS = lower urinary tract symptoms.
TABLE IIAssociation between moderate or severe erectile dysfunction and sociodemographic, medical, and lifestyle characteristics: crude and adjusted (by age and country) odds ratios
Characteristic or Condition
Patients with ED
Patients Without ED
Odds Ratio
95% CI
n
%
n
%
Crude
Adjusted
LCL
UCL
Marital status
Married
493
88.8
1669
89.9
1
1
Widower, separated, divorced
37
6.7
103
5.5
1.22
1.31
0.85
2.02
Single
25
4.5
85
4.6
1.00
1.85
1.12
3.05
Absence of sexual activity
77
14.5
42
2.3
7.31
3.88
2.55
5.91
Currently employed
327
58.9
1470
79.1
0.38
0.77
0.59
1.00
Education
High school not completed
321
57.9
874
47.1
1
1
High school completed
187
33.8
739
39.8
0.69
0.71
0.56
0.90
Degree
46
8.3
243
13.1
0.52
0.47
0.32
0.69
Diabetes
88
16.0
135
7.3
2.42
1.96
1.43
2.70
Hypertension
178
32.3
362
19.5
1.97
1.45
1.15
1.84
Heart disease
81
14.7
119
6.4
2.51
1.81
1.30
2.53
Ulcer
98
17.9
197
10.6
1.82
1.31
0.97
1.76
Depression
60
10.8
104
5.6
2.05
2.21
1.53
3.21
Depressive symptoms (CES-D)
129
29.7
347
20.2
1.67
2.27
1.74
2.96
Prostate disease (any)
76
13.8
107
5.8
2.61
1.69
1.19
2.40
LUTS
Absent or minor
380
77.4
1589
88.3
1
1
Moderate
90
18.3
184
10.2
2.05
2.05
1.49
2.82
Severe
21
4.3
26
1.4
3.38
5.75
2.93
11.30
Smoking status
Never
143
26.3
565
30.7
1
1
Former smoker
179
32.9
514
27.9
1.38
0.90
0.68
1.20
Current smoker
222
40.8
764
41.5
1.15
1.03
0.79
1.34
Tobacco units per day (n)
None
331
60.1
1088
58.6
1
1
1–30
183
33.2
695
37.4
0.87
1.00
0.80
1.26
>30
37
6.7
74
4.0
1.64
1.74
1.11
2.74
Alcoholic drinks (n)
None
237
42.7
773
41.6
1
1
1–7/wk
139
25.0
512
27.5
0.89
0.70
0.59
1.03
≥8/wk
179
32.3
574
30.9
1.02
0.81
0.60
1.05
Physical activity
Less than average
97
17.5
157
8.4
1
1
Average
151
27.2
393
21.1
0.62
0.70
0.50
1.00
More than average
307
55.3
1312
70.5
0.38
0.55
0.40
0.75
Key: ED = erectile dysfunction; CI = confidence interval; LCL = lower confidence limit of adjusted odds ratio; UCL = upper confidence limit of adjusted odds ratio; CES-D = Center for Epidemiologic Studies Depression Scale; LUTS = lower urinary tract symptoms.
The multivariate logistic analysis showed that the risk of ED increased by almost 10% per year of age (Table III). The risk of ED varied across countries, even after controlling for individual characteristics. ED was associated with the duration of diabetes and heart disease (5% per year of disease duration), depression (either a history of diagnosed depression or current depressive symptoms), and heavy smoking. LUTS were associated with ED in a graded manner. The inverse association between ED, education level, alcohol use, and physical activity was confirmed.
TABLE IIIAssociation between moderate or severe erectile dysfunction and sociodemographic, medical, and lifestyle characteristics: multivariate analysis (logistic regression model)
Characteristic
Odds Ratio
95% CI
Age (per 1-yr increase)
1.10
1.08, 1.11
Education
High school not completed
1
High school completed
0.73
0.56, 0.96
Degree
0.52
0.34, 0.80
Country
Brazil
1
Italy
1.60
1.10, 2.33
Malaysia
2.16
1.45, 3.24
Japan
5.38
3.58, 8.07
Diabetes (per 1-yr duration)
1.05
1.01, 1.10
Heart disease (per 1-yr duration)
1.05
1.01, 1.09
Depression/depressive symptoms
2.09
1.60, 2.74
LUTS (IPSS score)
Absent or minor (<8)
1
Moderate (8–19)
1.80
1.28, 2.54
Severe (≥20)
4.56
2.24, 9.27
Tobacco units per day (n)
None
1
1–30
0.95
0.73, 1.24
>30
2.12
1.26, 3.56
Alcoholic drinks (n)
None
1
1–7/wk
0.74
0.53, 1.02
≥8/wk
0.73
0.53, 0.99
Physical activity
Less than average
1
Average
0.81
0.54, 1.22
More than average
0.64
0.45, 0.92
Key: CI = confidence interval; LUTS = lower urinary tract symptoms; IPSS = International Prostate Symptom Score.
The men were also asked whether they would consult a doctor or other health professional if they had ED. In Brazil, 90% said “yes”; in Italy, Malaysia, and Japan, 89%, 72%, and 31%, respectively, answered affirmatively. Of the 548 men who reported moderate or severe ED and answered this question, only 26 (5%) had been previously treated: 17 (19%) of 89 in Brazil, 6 (6%) of 97 in Malaysia, 3 (2%) of 123 in Italy, and 0 (0%) of 239 in Japan. When asked how comfortable they were in discussing their ED problem with a physician, 13% in Brazil, 25% in Italy, 57% in Japan, and 25% in Malaysia said they were not at ease.
Comment
The results of this study contribute information on the prevalence of ED and confirm that it is a widespread problem among mature men, regardless of epidemiologic, ethnic, or environmental diversities. The prevalence of ED varied among the countries, but its age pattern and associations with background diseases and behaviors were similar.
The country-specific risk may reflect the influence of factors such as genetics (which we did not study) or cultural differences in the perception of, attitudes toward, and willingness to report, ED. The difference in prevalence may also have been a result of the different ways in which the questionnaire was administered. The method of administration was chosen with the aim of obtaining valid answers in the most suitable way for each country.
There is no simple answer to explain the low response rate in Malaysia and Japan. The most plausible explanation calls for the influence of cultural factors. However, the different response rates cannot influence the study of the association between ED and possible risk factors because the survey was presented to potential subjects as a survey of the general health of adult men. As in every epidemiologic study—particularly those dealing with sensitive questions—the results are subject to a certain degree of inaccuracy, but, rather than a random variation in reporting ED prevalence, it could be expected that the prevalence would be underreported. Our findings are consistent with the available epidemiologic and biologic evidence and are strengthened by the internal validity of the study, which comes from the use of a common questionnaire, the four-step scale in the definition of ED, and the linear trends of the association between ED and health-related factors. The major limitation of the study was that the data collection was limited to self-reports. Medical conditions that may be asymptomatic are often unknown by the subject and consequently underreported, which is likely to lead to a nondifferential misclassification and the attenuation of the OR. The negative association between education and ED can be understood if we consider lower education as a marker of a high prevalence of undiagnosed diseases.
our results confirm that ED is associated with age but not an inevitable outcome of the aging process. Among the men aged 65 years or older, only 29% reported moderate or complete ED in Brazil versus 71% in Japan.
In addition to aging, ED is associated with medical and behavioral conditions. Two studies of diabetic patients reported a prevalence of ED ranging from 35% to 75%.
an etiology that is supported by our finding that it was more prevalent in men with heart disease and hypertension. Prostate disease or surgery have repeatedly been associated with ED,
and we found that a history of prostate disease and the presence and severity of LUTS were associated factors. Depression and sexual dysfunction are often reported as associated conditions, with some investigators considering depression an important factor in the etiology of ED,
Alcohol consumption inversely correlated with ED in our population. The Health Professionals Follow-up Study found a lower prevalence of ED in moderate drinkers than in nondrinkers or heavy drinkers.
Moderately active and active men had a lower prevalence of ED, and the beneficial effect of physical activity on ED has also been shown in other studies.
Most of the men complaining of ED did not seek medical attention. A major cause for that can be the reluctance to admit to this disorder. There was a difference among the studied countries that leads us to believe that local culture plays an important role. This study was carried out before oral therapy with sildenafil was available. It will be interesting to see whether the observed pattern will change after the introduction of this treatment.
Conclusions
Although the prevalence of ED varied across the four countries, it was common in all of them and increased in prevalence and severity with age. Despite the national variations in the prevalence of medical conditions and health-related behaviors, clear associations emerged between ED and other factors. To define the risk for ED better, future studies should also examine the medical correlates of ED (especially those that are frequently asymptomatic or underdiagnosed) in prospective samples.
02102-7&id=gr1.jpg){kind=link}