Incidence rates of erectile dysfunction in the Dutch general population. Effects of definition, clinical relevance and duration of follow-up in the Krimpen Study

Abstract

This study aims to describe the incidence rate of erectile dysfunction (ED) in older men in the Netherlands according to three definitions. The influence of the duration of follow-up on the incidence rate is also explored. In a large community-based follow-up study, 1661 men aged 50–75 y completed the International Continence Society sex questionnaire and a question on sexual activity, at baseline and at a mean of 2.1 and 4.2 y of follow-up. We defined ‘ED’ as a report of erections with ‘reduced rigidity’ or worse; ‘Significant_ED’ as ‘severely reduced rigidity’ or ‘no erections’; and ‘Clinically_Relevant_ED’ as either ‘ED’ reported as ‘quite a problem’ or ‘a serious problem’, or ‘Significant_ED’ reported as at least ‘a bit of a problem’. Incidence rates of ED status were calculated in those men who completed at least one period of follow-up and were not diagnosed with prostate cancer (n=1604). For ‘ED’ the incidence rate (cases per 1000 person-years) is 99 and ranges over the 10-y age groups from 77 (50–59 y) to 205 (70–78 y); for ‘Significant_ED’ these rates were 33, 21, and 97, respectively and for ‘Clinically_Relevant_ED’ 28, 25, and 39, respectively. In general, incidence rates should not vary with the duration of follow-up. However, for ‘ED’ the 4.2 y incidence rate is about 69% of the 2.1 y incidence rate. This study presents incidence rates, for the general population, as well as based on a definition of ED that takes concern/bother into account. ‘Clinically_Relevant_ED’ has a lower increase in incidence with increasing age than other definitions that do not take concern/bother into account. The phenomenon of lower incidence rates with longer duration of follow-up may account for the differences in reported incidence rates between different studies. The effects of differences related to the duration of follow-up should be taken into consideration in future incidence reports.

Introduction

Erectile dysfunction (ED) is a common and bothersome problem in older men. With the introduction of new classes of drugs and new modes of administration of existing drugs public interest in the condition has increased.^1,2,3 Nevertheless, there is still a paucity of incidence data on ED in the general population. Data are restricted to three surveys conducted in the United States,^4,5 Brazil^4,5,6 and Finland.⁷ All three surveys showed a rising incidence with advancing age. The differences in the incidence rates reported in these studies could be based on geographical or methodological difference, such as duration of follow-up or use of different definitions. Such methodological differences have accounted for the large variety in prevalence data on ED.^8,9 To make epidemiological data clinically more relevant, it has been suggested to use a definition of ED in which psychological distress, patient bother or concern is incorporated.^1,3 This approach has not yet been used in studies on the incidence of ED.^4,5,6 Therefore, this study aimed to describe the incidence rate of ED in community-dwelling older Dutch men according to three definitions, including one with concern/bother taken into account. Moreover, the influence of the duration of follow-up on the presented incidence rate is explored.

Patients and methods

The data were obtained as part of a large community-based study, that is, the Krimpen study on male urogenital tract problems and general health status; the design of this study has been described previously.¹⁰ Briefly, all men aged 50–75 y (age on reference date, n=3924) in Krimpen aan den IJssel (a Dutch municipality near Rotterdam) were studied. Men without radical prostatectomy, prostate or bladder cancer, neurogenic bladder disease or negative advice from their general practitioner (GP), and who were able to complete questionnaires and visit the health centre were invited to participate. All men entering the cohort study provided written informed consent. The Medical Ethics Committee of the Erasmus Medical Centre Rotterdam approved the study.

In the baseline part of the study 1661 participants filled out a 113-item self-administered questionnaire, which included the International Continence Society (ICS) sex questionnaire¹¹ and a question about sexual activity. The men visited the local health centre for a physical examination (including measurement of blood pressure, height and bodyweight) and to provide data on current drug use. Subsequently, they visited the urology outpatient department for additional measurements. Prostate biopsies were taken, according to a previously described protocol, to detect prostate cancer.¹²

As all participants were registered with one of the local GPs, death of a participant or change of address was registered. No attempts were made to track participants that had moved out of the area.

If no prostate cancer was detected, if the participant had not died or moved away, and if no other exclusion criteria were met the GPs were asked to send a re-invitation letter to all participants. The first and second follow-up measurements consisted of the 113-item self-administered questionnaire and a visit to the urology outpatient department. Prostate biopsies were taken according to a previously described protocol to detect prostate cancer. Figure 1 presents a flowchart accounting for each participant.

Figure 1

Flow chart of participants in Krimpen Study.

Baseline data were collected between June 1995 and December 1997. In the first follow-up period data were collected between October 1997 and February 2000 and in the second follow-up between February 2000 and May 2002. The two periods of follow-up measurements were completed with an average interval of 2.1 and 4.2 y, respectively, since baseline.

Definitions

Based on the question of the ICSMale sex questionnaire (Appendix) the following definitions were made.

‘All ED’ a report of erections with ‘reduced rigidity’ or ‘severely reduced rigidity’ or ‘no erections’; ‘Significant ED’ a report of erections with ‘severely reduced rigidity’ or ‘no erections’.⁸

‘Clinical relevant ED’ either: ‘All ED’ associated with a major concern or ‘Significant ED’ associated with a minor or major concern. In this respect, ‘minor concern’ was defined as a report of ‘a bit of a problem’; ‘major concern’ as a report of ‘quite a problem’ or ‘a serious problem’ on the associated questions.⁸

Statistical analyses

ED was studied in those men who completed the baseline measurement and were not diagnosed with prostate cancer (n=1604).¹² Depending on the definition, a proportion of these men reported ED at baseline. Those men who did not were defined as ‘ED-free men at risk’ and constitute the basis for estimation of the ED incidence rates.

Crude cumulative incidence risks (CCI) were calculated by dividing the number of new cases by the total number of participants at risk during that measurement period. CCI was calculated for all three definitions (ie All ED, Significant ED and Clinically relevant ED).

The number of person-years was defined as the number of years between respective measurements (eg from baseline to first follow-up) times the number of men at risk. In as much as incident cases occurred sometime between respective measurements (eg between baseline and first follow-up), we assumed that these cases on average became incident cases halfway their follow-up period.

We calculated the incidence rate of ED in three ways, by dividing the number of new cases by the number of person-years at risk for a certain period. First, we used the period between the baseline and the first follow-up (average 2.1 y). Second, we used the period between the baseline and second follow-up (average 4.2 y). Third, to study the possible effect of ‘unstable’ ED, we used the period between the baseline and first follow-up summed with the period between the first and second follow-up (average 2 × 2.1 y).

Age was calculated at baseline and men were categorised into 10-y age groups at baseline. The oldest age group went from 70–78 y for practical reasons. For each of the 10-y age groups the person-years and the incidence rates were calculated according to the three definitions of ED.

In total, 95% confidence intervals (95% CI) were calculated for the incidence rates assuming that the number of incident cases is Poisson distributed.

A complete case analysis was carried out. It has been discussed extensively that the complete case analysis is an incomplete analysis when missing data occur.^{13,14,15,16,17,18,19,20} Additionally, we analysed all data using IVEware for a multiple imputation strategy (MI).^{13,15,16,17,19} To impute the missing ED status IVEware used a model with all baseline data and the known ED status in first and second follow-up.

All calculations were carried out using SAS version 8.2;²¹ imputation was done with IVEware version 2.0.²² We calculated five imputed databases with 10 iterations each. Pooling of the estimates was done according to the formulas of Rubin and Schenker.¹⁸

Results

Mean follow-up time was 2.1 (range 1.8–3.3) and 4.2 (range 3.6–5.8) y for the first and second follow-up periods, respectively. In the first follow-up period the response rate was 78%, in the second follow-up period the response rate was 80%. From baseline to second follow-up the response rate was 62%.

Table 1 shows that participants with ‘Significant ED’ do not always suffer from ‘Clinically relevant ED’. Incidence rates are given in Tables 2 and 3.

Table 1 Share of clinically relevant ED and significant ED in All ED at baseline

Table 2 Incidence rates by age for three different definitions

Table 3 Effect of imputation^a as instrument for correction of bias due to loss to follow-up

All ED

Of the 1053 ‘All ED’-free men at baseline 781 and 604 men participated in the first and second follow-up period, respectively. The CCI was 18.8%/2.1 y. Incidence rates (per 1000 person-years) are presented in Tables 2 and 3. The crude incidence rate is 98.6 and varies from 76.5 in men aged 50–59 y to 205.4 in men aged 70–78 y. The ED incidence rate rises with ageing.

Significant ED

Of the 1432 ‘Significant ED’-free men at baseline in the study 1048 and 780 men participated in the first and second follow-up period, respectively. The CCI was 6.7%/2.1 y. Incidence rates (per 1000 person-years) are given in Tables 2 and 3. The crude incidence rate is 32.8 and varies from 21.2 in men aged 50–59 y to 96.6 in men aged 70–78 y.

Clinically relevant ED

Of the 1458 ‘Clinically relevant ED’-free men at baseline in the study 1062 and 791 men participated in the first and second follow-up period, respectively. The CCI was 5.7%/2.1 y. Incidence rates (per 1000 person-years) are given in Tables 2 and 3. The crude incidence rate is 28.1 and varies from 25.3 in men aged 50–59 y to 39.1 in men aged 70–78 y. The range of incidence rates between age groups is smaller than in the definitions without bother taken into account.

Discussion

This European study presents incidence rates for the general population and is, to our knowledge, the first study to present incidence rates (incident cases per 1000 person-years) based on a definition of ED that takes concern/bother into account.

The incidence increases with age, as reported in other studies.^4,6,7,23 The use of different definitions resulted in large differences of ED incidence rates (per 1000 person-years): from 99 for ‘All ED’ to 33 for ‘Significant ED’ and 28 for ‘Clinically relevant ED’.

‘Clinically relevant ED’ puts the incidence in a more relevant perspective. This definition will make a better proxy for medical consumption due to ED.

In men above 70 y, clinically relevant ED has a lower incidence than significant ED. This indicates that older men are less bothered by their ED. This has important implications in an ageing population because the prevalence is highest in the older age groups.^1,3,8,9 This is in contrast with some studies where older age was associated with more bother^24,25 but in agreement with our earlier bother/prevalence of ED analysis.⁸

There was no clearly defined time frame for the men to refer to when answering the question that we have used to assess ED status. However, the context of the complete 113-item questionnaire was such that most questions referred to the preceding month. This is a general problem of ED questionnaires that do not clearly define the timeframe, which was also the case in the two incidence studies^4,5,6 discussed in this article.

The decrease in estimates of the incidence rates of the 4.2 y follow-up in comparison with the two times 2.1 y follow-up seems to implicate that at least in some incident cases ED is a self-limiting disease. This in turn implies that incidence rates based on a follow-up period spanning several years might result in an underestimation of the problem.

Compared with the data from the Massachusetts male aging study (MMAS),^4,5 our incidence rate of ‘All ED’ is higher, whereas the incidence rates of ‘Significant- and Clinically relevant ED’ are lower, as were the previously published prevalence rates from the Krimpen study.⁸ These differences might be explained by our different methodology, but more probably our use of ‘a clinically relevant definition’ of ED may account for the differences. The present study shows that the incidence rates are heavily influenced by the definition of ED that is used.

The incidence rates in the Brazilian study were higher than those from the MMAS,⁶ but both studies used the same definitions. Our study shows that a 2.1 y longer duration of follow-up (2.1 or 4.2 y) gives a 31–50% reduction of the incidence rates depending on the definition used. The difference in follow-up time between the Brazilian-study and the MMAS is approximately 6.6 y, which could account to some extent for the ±250% difference.⁶

The Incidence rates found by the Finnish study were, depending on the definition, similar to our incidence rates where ‘Moderate ED’ lies in between the incidence rates of ‘All ED’ and ‘Significant ED’ with a follow-up of 4.2 y. They used a follow-up interval of 5 y. As the authors suggested in their comment, this might give an underestimation of the true incidence in the population.⁷

In our study, 62% loss to follow-up was higher than in the Brazilian study (average follow-up period 2.2 y response rate 87%) and than in the MMAS (average follow-up period 8.8 y response rate 68%) and than the Finnish study (average follow-up period 5 y response rate 74%). The participants in the Krimpen-study had to endure an arduous protocol, but no incentives were given; loss to follow-up analyses showed that this had minor influence on outcome measures. Especially, age-specific incidence rates of ED are minimally influenced by the loss to follow-up (unpublished data). Furthermore, after correction by imputation of the missing subjects only small changes emerged. The crude incidence rates increased by approximately 5% (range 3–8%) depending on the duration of follow-up and the definition used.

In conclusion, this population-based study presents European incidence rates of ED; these rates are lower than the MMAS incidence rates or recent European incidence rates. For the first time ‘Clinically relevant ED’ incidence rates are reported and are shown to have a lower increase with age than ED incidence based on other definitions that do not take concern/bother into account.

A longer duration of follow-up has a downsizing effect on the estimations of incidence rate, possibly because of a self-limiting character of the ED during the interval. This should be taken into consideration in future reports on ED-incidence.

Appendix 1

Table a1 Appendix: ICS male sex questionnaire—erectile questions