The prevalence of urinary incontinence in women in four European countries

Authors


S. Hunskaar, Section for General Practice, Department of Public Health and Primary Health Care, University of Bergen, Ulriksdal 8c, N-5009 Bergen, Norway.
e-mail: steinar.hunskar@isf.uib.no

Abstract

OBJECTIVE

To determine the prevalence, type and treatment behaviour of women with urinary incontinence in four European countries.

SUBJECTS AND METHODS

Data were collected using a postal survey which was sent to 29 500 community-dwelling women aged ≥ 18 years in France, Germany, Spain and the UK. Subjects were asked about the type of urinary incontinence they had experienced and their treatment behaviour.

RESULTS

Of the women who responded, 35% reported involuntary loss of urine in the preceding 30 days; stress urinary incontinence was the most prevalent type. The lowest prevalence was in Spain (23%), while the prevalence was 44%, 41% and 42% for France, Germany and the UK, respectively. About a quarter of women with urinary incontinence in Spain (24%) and the UK (25%) had consulted a doctor about it; in France (33%) and Germany (40%) the percentages were higher. Overall, < 5% of the women had ever undergone surgery for their condition. While pads were used by half of the women, there were some differences among the countries.

CONCLUSIONS

Millions of women in Europe have urinary incontinence; the consultation and treatment rates were low in the European countries included in this study.

Abbreviations
UI

urinary incontinence

ICS

International Continence Society.

INTRODUCTION

Urinary incontinence (UI) is a common condition, especially in women, and affects the physical, psychosocial, social and economic well-being of affected individuals and their families [1–3]. Different studies have shown a wide variability among prevalence rates, mainly attributed to differences in the definitions used, the design of the questionnaires, the study population and selection criteria [4]. The prevalence of UI increases with age, with a typical rate in young adults of 20–30%, a peak around middle age (prevalence 30–40%) and a steady increase in old age (prevalence 30–50%) [5,6].

The former definition of UI by the International Continence Society (ICS) as ‘a condition where involuntary loss of urine is a social or hygienic problem and is objectively demonstrable’[7] was not ideal for epidemiological studies [8,9]. However, the new ICS definition of UI as ‘the complaint of any involuntary leakage of urine’[10] makes it easier for epidemiological studies to comply with the most accepted scientific definitions in use. This new definition has not been used in previous studies across countries.

Crude prevalence studies of UI are abundant, and the 2nd International Consultation on Incontinence recommended that further studies should be conducted only with recommended and validated questionnaires, to allow the combination of data from prevalence studies with those of cofactors and predictors (analytical epidemiology) [1]. There is a need for more advanced epidemiological analyses of several factors associated with UI [3,8]. As there are convincing data suggesting that the different types of UI may reflect different pathologies and risk factors, epidemiological research should also differentiate between the types. There is no hard evidence for different incidence and prevalence of UI among Western countries and the published differences may relate to social, culture, economic or, until now, undetermined factors [11]. Because of the heterogeneity of published studies the results are difficult to compare, and most of the studies do not lend themselves easily to cross-cultural or cross-national comparisons.

We report here a large cross-national study using a validated instrument to investigate the prevalence of UI. The study includes the distribution of the main types of UI (stress, urge and mixed) and the treatment and consultation rates for female UI in four European countries, i.e. France, Germany, Spain and the UK.

SUBJECTS AND METHODS

A postal survey was mailed to 29 500 households in the four countries; the sample was drawn iteratively from households participating in national surveys in France, Germany and the UK, and similar panels in Spain. The samples were checked and were representative of the age, household size and geographical region of the general population in all countries. In the UK it was also representative of social class and in France of employment status; these variables were not recorded in the other countries.

Definitions conform to the standards recommended by the ICS [10] except where specifically noted. UI was defined as any leakage or involuntary loss of urine. Stress UI symptoms were defined as a leak or loss of urine caused by sneezing, coughing, exercising, lifting or physical activity. Urge UI symptoms were defined either as an urge to urinate but being unable to reach the toilet before leaking or having a strong sudden urge to go to the toilet to urinate with no advance warning. Mixed UI symptoms were defined as at least one stress and one urge symptom. Stress-only symptoms were those in respondents having one or more stress UI symptom, but no urge symptoms; urge-only symptoms were those in respondents having one or more urge UI symptom, but no stress symptoms.

A 13-item questionnaire was used to assess the prevalence and frequency of UI symptoms during the last 30 days (Appendix); each questionnaire was to be completed by adult female household members aged ≥ 18 years. The present analysis focuses on the prevalence of UI symptoms and some related factors, but a sample of responders was sent a further more detailed questionnaire, which will be analysed separately. The questionnaire included a question on leakage caused by laughing, but this was included with other stress symptoms in line with the latest ICS definition [10].

The questionnaire first recorded any experience of UI and related symptoms in the last 30 days, then the frequency of the symptoms in the last 7 days. Consultation with a doctor was registered, and the doctor's diagnosis if known. Use of prescription drugs, pads or other absorbent material, and previous surgery related to UI, were also recorded.

We determined three ways for a responder to be defined as having any UI. First, women who confirmed leakage and incontinence symptoms in Question 1; second, if a woman reported no symptoms on question 1 but subsequently answered ‘Yes’ to any of the specific symptoms in Question 2; third, if a response was missing on Question 1, but then subsequently ‘Yes’ was answered to any of the specific symptoms in Question 2 (Fig. 1). If a person responded ‘No’ to Questions 2a-h and her response to Question 1 was missing, then she was coded as ‘No’ on Question 1. There were 63 cases which were missing age or sex or both; these cases were not included in the analysis.

Figure 1.

The flow diagram of women included in the analysis.

The same procedure was followed to re-code answers to other questions, except Question 11, i.e. missing responses were treated as ‘No’ responses. The analysis for Question 11 about pad usage was treated in a different way; missing responses were left as missing.

The questionnaire was first constructed in English by including questions recommended for epidemiological purposes when studying UI [3]. It was translated into German, French and Spanish by professional translators. Back-translation of selected parts of the questionnaire was done to ensure homogeneity and each translated questionnaire was tested in the respective country with four respondents. Several experts in the field also checked the wording, and thus face validity was obtained.

Test-retest reliability of the German version of the questionnaire was tested in a separate study in 2002 in 269 German women who had UI. The subjects completed the same questionnaire twice with an interval of ≈ 30 days. The women were recruited from the German consumer panel used in the main study. Statistical analysis depended on the question and used a t-test for continuous variables and McNemar's test for dichotomous variables.

The questionnaire was also used in a study in the USA and a test-retest reliability study in 251 patients showed that it was reliable, with results similar to the present study of the German version (Fultz N, Burgio K, Diokno A, personal communication).

STATISTICS

The data were analysed as 5-year age groups for any UI and for type of UI by country. When appropriate, three age groups were defined (18–44, 45–59 and ≥ 60 years). The results were assessed statistically by univariate and bivariate methods. Chi-square tests were used when comparing prevalence, type of UI, treatment and consultation rates by country. Results are given as numbers and percentages with 95% CI. Statistical significance was accepted at the 1% level (P < 0.01), chosen to take account of the relatively many statistical analyses on the dataset.

RESULTS

Null (insignificant) results indicated no difference between the answers to the two questionnaires in the test-retest study in 269 German subjects, thus showing acceptable test-retest reliability. From the 25 answers, four were statistically different between the first and second completion of the questionnaire. In one of the questions 10% of subjects reported a different age, an expected result, as one in 12 women may have had a birthday during the 30-day interval. Although there were statistically different responses for another three questions the differences were small and probably reflect true differences between the periods. For example, the question ‘Did you experience any of the following symptoms in the past 30 days?’ required a positive answer the first time the questionnaire was completed for women to be eligible to continue in the study. However, 12% reported no symptoms during the second 30-day period before completing the second questionnaire. For the question asking about loss of urine caused by coughing/sneezing, 13% responded differently in the second completion. The test-retest results for the other symptom questions were not significantly different, so it could be assumed that during the 30-day interval the subjects may have coughed or sneezed less, possibly if a cough or cold resolved. The question asking if their condition was diagnosed by a physician was answered differently by 8% of respondents; this may reflect an interpretation that the question only related to diagnosis in the previous 30 days.

In all, 29 500 questionnaires were sent to women in four European countries (Table 1). After excluding invalid questionnaires (Fig. 1), 17 080 (58%) were included in the analysis. The mean response rate was 58%, but only 45% in UK; the mean age was 46.3 years and the mean, median and range of age differed little among the countries.

Table 1.  The characteristics of the responders and the prevalence of UI for the four countries
CharacteristicFranceGermanySpainUKAll*Valid responders
  • *

    Combined figures for UK, Germany, France and Spain;

  • P < 0.01 vs Germany for all ages combined;

  • P < 0.001 vs France and Germany and UK for all ages combined.

Questionnaires sent6500650010 000650029 500 
Valid responders38813824 6 444293117 080 
Response rate, %  60  59    64  45    58 
Mean (median)
[range] age, years
  44.8 (44) [18–88]  47.4 (47) [18–98]    46.3 (42) [18–99]  47.1 (46) [18–97]    46.3 (44) [18–99] 
Prevalence of UI, n (%) [95% CI]
Age group
18–24  47 (27) [21–34]  21 (25) [16–35]    33 (6) [4–8]  46 (30) [22–37]   147 (15) [13–17]   979
25–29 142 (31) [27–35]  118 (27) [23–31]    86 (15) [12–17]  69 (31) [25–37]   415 (24) [22–26] 1 709
30–34 142 (31) [27–35] 105 (31) [26–36]   148 (17) [14–19] 126 (30) [25–34]   521 (25) [23–27] 2 102
35–39 183 (37) [33–41] 138 (31) [26–35]   187 (21) [18–24] 104 (37) [32–43]   612 (29) [27–31]  2 116
40–44 184 (44) [39–49] 182 (42) [37–47]   128 (23) [19–26]  116 (42) [36–48]   610 (36) [34–38] 1 694
45–49 169 (51) [46–56] 213 (39) [35–43]    97 (22) [18–26]  114 (45) [39–51]   593 (38) [35–40] 1 566
50–54 220 (54) [49–58] 156 (42) [37–47]    89 (25) [20–29] 174 (50) [45–56]   639 (43) [40–46] 1 486
55–59 228 (57) [52–62] 109 (50) [43–56]    87 (23) [19–27] 153 (53) [47–59]   577 (45) [42–47] 1 292
60–64 232 (55) [51–60] 200 (53) [47–58]    90 (25) [20–29] 123 (50) [44–57]   645 (46) [43–48]  1 411
65–69 152 (52) [46–58] 156 (58) [52–64]   195 (34) [30–38]  92 (52) [45–59]   595 (45) [43–48]  1 315
70–74   2 (20) [0–45]  96 (53) [46–60]   138 (35) [31–40]  48 (37) [29–46]   284 (40) [36–44]   710
75–79   4 (80) [45–100]  56 (62) [52–72]    83 (38) [31–44]  36 (44) [33–55]   179 (45) [40–50]   398
80–84   6 (86) [60–100]  24 (71) [55–86]    42 (40) [31–49]  17 (50) [33–67]    89 (49) [42–57]   180
85–89   3 (75) [33–100]   1 (50) [0–100]    29 (51) [38–64]   7 (70) [42–98]    40 (55) [43–66]    73
≥ 90   0 [NA]   1 (50) [0–100]    23 (61) [45–76]   6 (67) [36–97]    30 (61) [48–75]    49
Total1714 (44) [43–46]1576 (41) [40–43] 1 455 (23) [22–24]1231 (42) [40–44] 5 976 (35) [34–36]17 080

In all, 35% of women reported that they had involuntary loss of urine (any UI) in the preceding 30 days; the prevalence by age and country is shown in Table 1, showing that Spain had a significantly lower overall prevalence than the other countries (P < 0.001). There was also a statistically significant difference in overall prevalence between France and Germany (P < 0.01). The prevalence increased with age in general, but there was a plateau between 50 and 79 years old, except for a small decrease for age 70–74 years; there were relatively few women ≥ 80 years old. The country differences in prevalence are shown by three age bands in Fig. 2. Spain had the lowest prevalence in all age groups, and had a total prevalence of only about half of that of France; the curves for France, Germany and UK were very similar.

Figure 2.

Country differences in prevalence by three age bands, for France (green), Germany (light green), Spain (red) and the UK (light red).

The type of incontinence by age is shown in Fig. 3. The prevalence of stress symptoms was the highest amongst all women who reported UI (all age groups combined) and in the middle-aged groups. Mixed symptoms increase steadily with age, as did urge symptoms, but the urge group was the smallest in most age bands. For all age bands there was a small group with other or unclassified incontinence. The relative distribution of the type of incontinence by three age bands and country is shown in Table 2. Stress UI was significantly less common in France than in the other countries and urge UI was statistically more common (both P < 0.001); the other inter-country differences are shown in Table 2.

Figure 3.

Type of incontinence by age band (other, white; urge, green; mixed, light green; stress, red).

Table 2.  The relative distribution (%) of the type of incontinence by three age bands in France, Germany, Spain and the UK. N is the total number of incontinent women and n the number of incontinent women in each age band
Age group/typeFranceGermanySpainUKAll
  1. Significance was tested only for all ages combined, using chi square tests; *P < 0.001 for France vs Germany, Spain and UK; P < 0.001 for France vs Germany, Spain and UK, P < 0.01 for Spain vs UK, and P < 0.001 for Spain vs France, Germany and UK.

18–446985645824612305
Stress 29 47 41 44  39
Urge 31 14 21 18  22
Mixed 26 27 17 25  24
Other 15 12 21 13  15
45–596174782734411809
Stress 33 41 55 42  41
Urge 24 14  11  11  16
Mixed 38 40 21 39  36
Other  5  4 13  7   7
≥ 603995346003291862
Stress 29 31 30 35  31
Urge 25 19 24 19  22
Mixed 41 46 37 40  41
Other  4  4  9  6   6
All ages
Stress* 31 40 39 41  37
Urge 27 16 21 16  20
Mixed 34 38 26 34  33
Other  9  7 15  9  10

The treatment and consultation rates for women with UI are shown in Table 3. Among the women who reported UI, about a quarter in Spain and the UK had ever consulted a physician about their condition, but the percentages were significantly higher in France and Germany (P < 0.001). Women in France and Germany were also the most likely to have had their condition diagnosed (P < 0.001), and to be receiving medication for their UI; there were significant differences between France and Spain or the UK, and between Germany and Spain (both P < 0.001) and between Germany and the UK (P < 0.01). About a half of women with UI in all the countries used pads. Taking all responders from all countries, 2.1% had ever had a surgical procedure for UI. Of the responders who had UI in the past 30 days < 5% had ever had a surgical procedure for UI, although this was significantly higher (≈ 8%) in the UK (P < 0.001 vs France and Spain, and < 0.01 vs Germany).

Table 3.  Treatment and consultation rates by country for women with UI
GroupFranceGermanySpainUKAll (95% CI)
  1. Chi square tests: *P < 0.001 for France vs Germany, Spain and UK; P < 0.001 for Germany vs Spain and UK; P < 0.001 for France vs Spain and UK, P < 0.001 for Germany vs Spain; P < 0.01 for Germany vs UK; P < 0.01 for Germany vs France and UK; P < 0.001 for Germany vs France, Spain and UK; P < 0.001 for France vs Spain; P < 0.01 for UK vs France and Spain; §P < 0.001 for UK vs France and Spain; P < 0.01 for Germany vs Spain and UK.

Women with UI, n17141576145512315976
Consultation rate*  33  40  24  25  31 (30–32)
Diagnosed by a physician*  28  36  16  19  25 (24–26)
Use of drugs for UI   7   5   2   3   5 (4–5)
Use of catheter   1   0   0   1   1 (0–1)
Use of pads  40  59  52  46  50 (48–51)
Surgery for UI§   3   5   3   8   5 (4–5)

DISCUSSION

This large study showed that the prevalence of UI and the distribution of different types of UI in France, Germany and UK were quite similar, while the prevalence was lower in Spain. We used the same questionnaire and methodological approach in the four countries, and the responders were of similar age. The overall prevalence of UI for these four European countries was ≈ 35%; this is in the typical range of results for many previous studies [3,11,12] but slightly higher than the large EPINCONT study in Norway, which reported an overall prevalence of UI of 25%[6].

A decrease in the prevalence of UI in women aged 65–79 years was also reported in other studies [4,13–17]. The reasons for this decrease are unclear but it was postulated that it might be associated with hormonal factors or a reduction in physical activity as women age. The very wide CIs in the older groups reflect that there are fewer women in those groups. The significant difference in the overall prevalence of UI between France and Germany was numerically small (44% and 41%) and the statistical significance was probably a result of the large sample sizes used.

As the same questionnaire and definitions were used the apparently lower prevalence in Spain was unexpected, but may have several possible explanations. It might be a true effect, but this seems unlikely as no major differences in risk factors between Spanish and other European women would be expected. Under-reporting of incontinence by Spanish women is a possible cause and could result from various cultural factors, e.g. embarrassment about the condition or an expectation that UI is an inevitable part of ageing [13,18]. Herzog and Fultz [19] recognized that survey questions must be asked in an appropriate way to avoid under-reporting of embarrassing information. Other Spanish studies have found very variable prevalence in different settings, from 14.5%[20] to 40%[21,22]. The effect might also result from unintended and unseen differences in the methods used in the four countries, like subtle differences in the translation of the questionnaire. However, a prevalence study of overactive bladder in six European countries [23] found a higher prevalence of this condition and urge UI in Spain than the average for the other countries, and especially higher than France.

The sample included was representative of the general population, but we did not assess whether nonresponders to the questionnaire were demographically similar to respondents; this information would have been useful in such an epidemiological study. The overall response rate was ≈ 60%, which is acceptable although not very high. In UK the response rate was 45%, which was lower than desired, and here conclusions should be drawn cautiously.

The prevalence of UI by type shows that stress UI was the commonest type overall, although the relative prevalence of mixed symptoms increased with age and was the commonest type in the population aged > 55 years. This is in agreement with the EPINCONT [6] and other large community-based studies [3]. The significantly lower rate of stress UI in French women and the corresponding higher rate of urge UI may reflect differences in diagnosis rather than true differences in prevalence, but this may be an area for future study. Because we did not ask women about the degree of bother caused by their UI we do not know how this may have affected their likelihood of responding to the questionnaire, or whether it might have introduced some bias into responses about their health-seeking behaviour.

The consultation rate of ≈ 30% is in line with other studies [19,24,25] and confirms that most people with UI have not sought help. Increasing age, increasing severity, increasing duration and urge/mixed type of UI were associated with consulting a doctor [24,26].

The major method of managing UI among community residents is the use of absorbent products [12,27,28]. There is growing evidence indicating that appropriate management can reduce the prevalence of UI [29], which is known to impose a significant financial burden on sufferers, their families and health services [30]. Only a small proportion of incontinent community-residing women have had surgery, medication or exercise regimens. The only factor that appears reliably related to whether any treatment or management is sought is the severity of the condition [3]. It is also probable that many primary healthcare providers lack confidence in managing UI, and that this contributes to under-treatment in those seeking help. There were some differences in the treatment rates among the four countries; the lowest prescription rates were in Spain (2%) and the UK (3%). These low prescription rates appear to follow the lower consultation rates in those countries, but the prescription (and surgery) rate in Spain may reflect the apparently lower prevalence of UI.

It is obvious that millions of women suffer from their UI, and that many of them can be offered good treatment [1]. It is an important point that many incontinent women have never talked to a healthcare professional about the condition. Both epidemiological and qualitative research in this field should be encouraged to understand the factors influencing patients to seek help.

ACKNOWLEDGEMENTS

We thank Dr Montserrat Espuña i Pons and Dr David Castro for their helpful comments about the prevalence of UI in Spain, and Mr Stephen Palmer for his invaluable help with the analysis. Eli Lilly and Company Limited funded the study and also employs S. Voss and D. Sykes. S. Hunskaar and G. Lose have received consultation fees, unrestricted research grants and funding for travel from Eli Lilly and Company Limited, as well as other companies engaged in incontinence research.

APPENDIX

THE QUESTIONNAIRE

Millions of people experience a leak or involuntary loss of urine. Sometimes there will be a leak or loss of urine when coughing, sneezing, laughing or during physical activity. Then again, sometimes it is more a feeling of pressure on the bladder.

1 During the LAST 30 DAYS, did you experience any of these symptoms?

2 During the LAST 30 DAYS did you experience any of the following? (Yes/No For EACH Symptom)

A leak or loss of urine due to . . .

a. Sneezing/coughing; b. Exercising; c. Lifting or physical activity; d. Laughing; e. Urge to urinate and could not get to the toilet before leaking or losing urine; f. Have a strong sudden urge to go to the toilet to urinate with no advance warning; g. Leak of urine that did not occur due to sneezing, coughing, exercising, lifting, physical activity or laughing; h. A feeling that your bladder is constantly full.

If ‘yes’ to any of these symptoms, continue. Otherwise skip to Question 13

3 During the PAST 7 DAYS, on how many days did you experience any of the symptoms listed above? (You may indicate any number from 0 to 7 days.)

4 How long have you experienced any of these symptoms?

5 Have you ever consulted with your doctor about any of symptoms listed in Question 2? (If Yes, continue. Otherwise, skip to Question 11)

6 How long did you wait from the time that you first started experiencing loss or leakage of urine until you consulted with your doctor?

7 Was your condition diagnosed by a physician?

(If Yes, continue. Otherwise, skip to Question 11)

8 With which of the following were you diagnosed?

(For EACH Condition)

Stress incontinence

Urge incontinence

Overactive bladder

Other

Don’t know

9 Are you taking any prescription medication for leakage or involuntary loss of urine?

10 Are you currently using a urinary catheter?

11 Do you use pads or other absorbent material for protection in case of loss or leakage of urine?

12 Have you been diagnosed with a urinary tract infection in the PAST 30 DAYS?

13 Have you ever undergone a surgical procedure due to leakage or involuntary loss of urine?

Ancillary