Urinary Incontinence as a Worldwide Problem

12
International Journal of Gynecology and Obstetrics 82 (2003) 327–338 0020-7292/03/$30.00 2003 International Federation of Gynecology and Obstetrics. Published by Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0020-7292Ž03.00220-0 Urinary incontinence as a worldwide problem V.A. Minassian*, H.P. Drutz, A. Al-Badr Division of Urogynecology and Reconstructive Pelvic Surgery, Department of Obstetrics and Gynecology, University of Toronto, Mount Sinai Hospital, Toronto, Ontario, Canada Received 11 December 2002; received in revised form 6 January 2003; accepted 10 January 2003 Abstract Objectives: This paper reviews the literature on the prevalence of urinary incontinence (UI) and demonstrates its impact as a worldwide problem. Methods: A MEDLINE search was performed to review population-based studies in English. Studies were grouped according to demographic variables and type of incontinence. Risk factors, help- seeking behavior, and quality of life measures were analyzed. Results: The median prevalence of female UI was 27.6% (range: 4.8–58.4%) and prevalence of significant incontinence increased with age. The commonest cause of UI was stress (50%), then mixed (32%) and finally urge (14%). Risk factors included parity, obesity, chronic cough, depression, poor health, lower urinary tract symptoms, previous hysterectomy, and stroke. Although quality of life was affected, most patients did not seek help. Conclusion: UI is a prevalent cross-cultural condition. Future studies should rely on universally accepted standardized definitions to produce meaningful evidence-based conclusions, as well as project the costs of this global healthcare problem. 2003 International Federation of Gynecology and Obstetrics. Published by Elsevier Science Ireland Ltd. All rights reserved. Keywords: Urinary incontinence; Epidemiology; Prevalence; Female 1. Introduction According to the International Continence Soci- ety (ICS) the definition of urinary incontinence (UI) changed from ‘the involuntary loss of urine that is a social or hygienic problem and is objec- tively demonstrable’ in 1979 w1x to ‘the complaint of any involuntary leakage of urine’ in 2002 w2x. The former definition is impractical for large epidemiologic studies, which are usually based on *Corresponding author. Tel.: q1-416-586-4642; fax: q1- 416-586-3208. E-mail address: [email protected] (V.A. Minassian). questionnaires or interviews. However, the present definition is too broad and potentially could include any patient with even one episode of UI in a lifetime. Despite this discrepancy, the recent ICS report states that UI should be further described by specifying frequency, severity, risk factors, social and hygienic impact, effect on quality of life, and whether or not the individual seeks help w2x. The stated reasons for this and other changes in the terminology of lower urinary tract function are to promote treatments based on symptoms, facilitate comparison of results, and help with effective communication between investigators.

Transcript of Urinary Incontinence as a Worldwide Problem

Page 1: Urinary Incontinence as a Worldwide Problem

International Journal of Gynecology and Obstetrics 82 (2003) 327–338

0020-7292/03/$30.00 ! 2003 International Federation of Gynecology and Obstetrics. Published by Elsevier Science Ireland Ltd.All rights reserved.doi:10.1016/S0020-7292!03.00220-0

Urinary incontinence as a worldwide problemV.A. Minassian*, H.P. Drutz, A. Al-Badr

Division of Urogynecology and Reconstructive Pelvic Surgery, Department of Obstetrics and Gynecology, University of Toronto,Mount Sinai Hospital, Toronto, Ontario, Canada

Received 11 December 2002; received in revised form 6 January 2003; accepted 10 January 2003

Abstract

Objectives: This paper reviews the literature on the prevalence of urinary incontinence (UI) and demonstrates itsimpact as a worldwide problem. Methods: A MEDLINE search was performed to review population-based studies inEnglish. Studies were grouped according to demographic variables and type of incontinence. Risk factors, help-seeking behavior, and quality of life measures were analyzed. Results: The median prevalence of female UI was27.6% (range: 4.8–58.4%) and prevalence of significant incontinence increased with age. The commonest cause ofUI was stress (50%), then mixed (32%) and finally urge (14%). Risk factors included parity, obesity, chronic cough,depression, poor health, lower urinary tract symptoms, previous hysterectomy, and stroke. Although quality of lifewas affected, most patients did not seek help. Conclusion: UI is a prevalent cross-cultural condition. Future studiesshould rely on universally accepted standardized definitions to produce meaningful evidence-based conclusions, aswell as project the costs of this global healthcare problem.! 2003 International Federation of Gynecology and Obstetrics. Published by Elsevier Science Ireland Ltd. All rightsreserved.

Keywords: Urinary incontinence; Epidemiology; Prevalence; Female

1. Introduction

According to the International Continence Soci-ety (ICS) the definition of urinary incontinence(UI) changed from ‘the involuntary loss of urinethat is a social or hygienic problem and is objec-tively demonstrable’ in 1979 w1x to ‘the complaintof any involuntary leakage of urine’ in 2002 w2x.The former definition is impractical for largeepidemiologic studies, which are usually based on

*Corresponding author. Tel.: q1-416-586-4642; fax: q1-416-586-3208.

E-mail address:[email protected] (V.A. Minassian).

questionnaires or interviews. However, the presentdefinition is too broad and potentially couldinclude any patient with even one episode of UIin a lifetime. Despite this discrepancy, the recentICS report states that UI should be furtherdescribed by specifying frequency, severity, riskfactors, social and hygienic impact, effect onquality of life, and whether or not the individualseeks help w2x. The stated reasons for this andother changes in the terminology of lower urinarytract function are to promote treatments based onsymptoms, facilitate comparison of results, andhelp with effective communication betweeninvestigators.

Page 2: Urinary Incontinence as a Worldwide Problem

328 V.A. Minassian et al. / International Journal of Gynecology and Obstetrics 82 (2003) 327–338

Numerous epidemiologic studies show that theincidence of UI increases with age w3–8x, with therange of prevalence estimates among communitydwelling patients varying enormously (2–58%)w4–6,9x. The lower prevalence of UI in institution-alized patients is even higher, with many authorssuggesting a prevalence of 40–60% w10–12x.Despite these high prevalence rates, UI is not astatic condition. Rather it is a dynamic conditionwhereby significant incidence rates are associatedwith equally significant remission rates, andpatients move back and forth from continence toincontinence w13x.This review describes epidemiologic studies on

UI from around the world. An attempt is made tocompare different studies, and the difficultiesencountered in such comparisons are presented.Also, the importance of UI as a national healthcare issue and a worldwide problem is stressed.Finally, recommendations for future epidemiologicstudies are presented.

2. Materials and methods

A MEDLINE literature search was performedspanning the period from January 1980 to October2002 using the key words: ‘urinary incontinence’combined with ‘epidemiology’ and ‘prevalence’.Other studies were identified by reviewing second-ary references in the original citations. Only pop-ulation-based studies in English were reviewed,and studies limited to only men or institutionalizedpatients were excluded.Only one population-based study was included

from each nation where such a publication wasavailable. In some countries, no population-basedstudies could be identified, but prevalence studiestargeting a specific group within the populationwere available. These studies were also includedfor analysis. Studies targeting specific ethnic orracial groups were also included. More than onepopulation-based study from the United States wasincluded only when originating from a differentstate and by a different author. Although someauthors have published numerous well designedprevalence studies from the same country, due tolack of space we elected to present the one study

from that country or state that surveyed the largestand most representative population sample.Studies were grouped together according to

location, year of survey, nature of populationsample, age, gender, population size, response rate,type of survey, definition, and prevalence of UI.Furthermore, studies that stratified for age andprevalence by type of UI (stress, urge, or mixed)were grouped together. Age groups were dividedinto 10-year periods ranging from the 5th year inone decade to the 4th year in the next. Studiesthat stratified their age groups from year zero toyear 9 of each decade were grouped with the nextgroup up for comparison purposes. For example,the age range 30–39 was grouped with that of35–44 and so forth. Risk factors of UI, help-seeking behavior, and quality of life measureswere analyzed.Because no unifying definition was presented in

the majority of the studies, a modification of thedefinitions reported by Hampel et al. w5x was usedto classify studies as follows:

1. Any UI in the previous 12 months (DefinitionI).

2. More than one episode of UI in a month(Definition II).

3. Two or more episodes of UI in a week (Defi-nition III).

4. Involuntary UI that is a social or hygienicproblem and is objectively demonstrable (Defi-nition IV).

5. Any UI, past or present (Definition V).

Studies not meeting any of the definitions listedwere grouped with the one closest to their defini-tion. Mean and median with range of prevalencewere calculated for the pooled data.

3. Results

Thirty-five studies w3,7,13–45x were identifiedusing the selection criteria mentioned above (Table1). Of those, 10 were from North America, eightfrom Asia, 13 from Europe, one from Africa, andthree from Australasia. Populations 5 years of ageand above from all continents and different raceswere studied. Twenty-one studies included onlywomen, whereas 14 others included both genders.

Page 3: Urinary Incontinence as a Worldwide Problem

329V.A

.M

inassianet

al./

InternationalJournal

ofG

ynecologyand

Obstetrics

82(2003)

327–338

Table 1Worldwide prevalence of urinary incontinence

Study Location Year of Population Age Sex Respondentsy Response Survey Definition Prevalence18 author survey sampling population rate type

Chiarelli w14x Australia 1996 Random sample 18q F 41 724y88 250 48% (y)a Mailed Definition I 12.8% (y)anational database 54% (m) survey 36.1% (m)

41% (o) 35% (o)Temml w15x Austria 1998–99 City voluntary 20q MyF 2498yNR NR On-site Definition II 5% (M)

free health survey survey 26.3% (F)Schulman w16x Belgium 1994–95 Stratified random 30q MyF 5269y5920 89% Home Definition I 5.2% (M)

population sample survey 16% (F)Alnaif w17x Canada 1996 ObyGyn visitors 15–19 F 332y474 70% On-site Definition I 27%

during U of T dayb surveyMoller w18x Denmark 1996 Random sample 40–60 F 2860y4000 72% Mailed Definition III 16.1%

national register surveyin two counties

Thomas w3x England Prior to 1980 Random sample 5q MyF 18 084y20 398 89% Mailed Definition II 3.3% (M)of GP practices survey 8.5% (F)

Peyrat w19x France 1998 Academic hospital 18q F 1700y2800 61% Mailed Definition I 27.5%employees survey

Lionis w20x Greece 1997 All patients in two 35–75 F 251yNR NR Personal Definition I 27.5%GP practices over interview4 months

Rekers w21x Holland Prior to 1991 Stratified random 35–79 F 1299y1920 68% Mailed Definition I 26.5%sample city register survey

Brieger w22x Hong Kong 1996 Random sample NR F 1500y3509 43% Phone Definition V 13%phone directory interview

Vinker w23x Israel Prior to 2001 Random sample of 30–75 F 418y500 84% GP office Definition V 36%GP practices survey

Bortolotti w24x Italy 1997 Random sample of 40q MyF 5488y5488 100% Phone Definition I 3% (M)GP practices interview 11% (F)

Ueda w25x Japan Prior to 2000 Random sample 40q MyF 1836y3500 53% Mailed Definition I 10.5% (M)prefecture survey 53.7% (F)population

Dolan w26x N. Ireland Prior to 1999 Random sample of 35–74 F 689y1050 66% Mailed Definition I 57%GP practices survey

Holst w27x New Zealand Prior to 1988 Random sample 18q F 851y1125 76% Phone Definition I 31%city electoral interviewregister

Lara w28x New Zealandy 1991 Random sample 18q F 556y1028 54% Mailed Definition V 29.2% (PI)cthree ethnicities electoral list from survey 31.2% (EU)

three districts 46.8% (Ma)Okonkwo w29x Nigeria Prior to 2001 Random sample 20q F 3963yNRa NR Personal Not defined 20% stress

of Gyn patients interview 22% urgeHannestad w7x Norway 1995–97 Total county 20q F 27 936y34 755 80% Mailed Definition V 25%

population surveyYarnell w30x South Wales Prior to 1981 Random sample 18q F 1000y1060 94% Personal Definition I 45%

district electoral interviewregister

Ju w31x Singapore 1989 National postal 65q MyF 919y1143 80% Personal Definition II 4.4% (M)district register interview 4.8% (F)

Page 4: Urinary Incontinence as a Worldwide Problem

330V.A

.M

inassianet

al./

InternationalJournal

ofG

ynecologyand

Obstetrics

82(2003)

327–338

Table 1 (Continued)

Study Location Year of Population Age Sex Respondentsy Response Survey Definition Prevalence18 author survey sampling population rate type

Gavira Iglesias w32x Spain 1996 Random sample 65q MyF 827y869 95% Personal Definition I 29% (M)census three interview 42% (F)municipalities

Milsom w33x Sweden 1986 Random sample 46q F 7459y10 000 75% Mailed Definition IV 16.8%city population surveyregister

Tseng w34x Taiwan 1997 Random sample 65q MyF 504y630 80% Personal Definition I 15% (M)town population interview 27.7% (F)

Swaddiwudhipong w35x Thailand 1989 Total 8 village 60q MyF 567y602 94% Personal Not defined 13.3% (M)population interview 14.5% (F)

Maral w36x Turkey Prior to 2000 Random sample 15q MyF 2053y2261 91% Personal Definition V 1% (M)district population interview 20.8% (F)

Rizk w37x United Arab Emirates 1996–97 Random sample NR F 400y448 89% Personal Definition I 20.3%from GP visits and interviewcommunity

Lagace w38x USA Michigan 1990 All patients in five 20q MyF 2830y3638 78% GP office Definition I 11% (M)GP practices survey 43% (F)

Nygaard w13x USA Iowa 1981–82 Total two county 65q F 2025y2530 80% Personal Definition V 55.1%population interview

Wetle w39x USA Mass. 1982 Total community 65q MyF 3809y4485 85% Personal Definition V 34.1% (M)population interview 44.4% (F)

Burgio w40x USA Pennsylvania Prior to 1991 Sample of city 42–50 F 541y901 60% Personal Definition V 58.4%patients with interviewdriver’s licenses

Roberts w41x USA Minnesota 1994 Random sample 50q MyF 1540y2337 66% Mailed Definition I 24% (M)county population survey 49% (F)

Brown w42x USA four states 1992–94 Random sample 69q F 7949y8366 95% Personal Definition I 41%population listing interview

Sze w43x USA N. Carolina 2000–01 All gyn patients 30q F 2370yNR NR Gyn Definition V 41% whitemedical center office 31% blackthree racial groups survey 30% hisp

Miles w44x USA five states 1993–94 Random sample 65q MyF 2660y3051 86% Personal Definition I 14.1%Hispanic population interview

Fultz w45x USA all states 1993–94 Random sample of 70q F 4221y5250 80% Phone and Definition I 23% whiteMedicare enrollees personal 16% black

interview

M, male; F, female; NR, not reported; GP, General Practitioner; gyn, gynecology; Mass, Massachusetts; hisp, Hispanic. Definition I: Any uncontrolled urine loss in theprevious 12 months. Definition II: More than one episode of UI in a month. Definition III: Two or more episodes of UI in a week. Definition IV: Involuntary loss of urine thatis a social or hygienic problem and is objectively demonstrable. Definition V: Any UI, past or present.

Three age cohorts (y, young: 18–23 years; m, middle age: 45–50 years, o, old: 70–75 years).a

U of T, University of Toronto.b

PI, Pacific Island; EU, European; Ma, Maori.c

Page 5: Urinary Incontinence as a Worldwide Problem

331V.A. Minassian et al. / International Journal of Gynecology and Obstetrics 82 (2003) 327–338

Fig. 1. Prevalence of any UI in women by age group (data from 13 studies).

Approximately 230 000 people were surveyed witha median response rate of 80% (range: 41–100%).The median prevalence of UI was 27.6% (range:4.8–58.4%) in females, and 10.5% (range: 1–34.1%) in males.The survey was conducted on total populations

in 5y35 (14%) studies, random population samplesin 15y35 (43%) studies, doctors’ patients in 8y35(23%) studies, and ‘other’ in 7y35 (20%) studies.Although the majority of the studies 19y35 (54%)defined UI as any loss of urine in the past 12months, more than five different definitions wereused. Some studies included only patients withstress urinary incontinence w36x, others includedpatients with only stress or urge incontinence w43x,and still others did not report on what definitionthey used w29,35x.Fig. 1 shows the relationship between age and

median prevalence of any UI in females pooledfrom 13 studies w3,7,14–16,21,24–27,30,33,38xwith an age span of four decades or more includingpre- and post-menopausal women. The medianprevalence of any UI had two peaks, one at the5th decade (33%) and another at the 8th decade

of life (34%). Six studies w3,7,21,27,33,38x distin-guished any or occasional UI from regular orsignificant UI. Fig. 2 shows that the medianprevalence of significant UI increased from the2nd to the 8th decades of life.Five studies presented comparisons between dif-

ferent races or ethnicities w28,31,40,43,45x. Theprevalence of UI was higher in Maori women(46.8%) compared with Pacific Island (29.2%)and European women (31.2%) (x s14.02, Ps2

0.001) w28x. Two American studies showed thatUI was higher in white (23–32%) compared withblack women (16–18%) (x )10, P-0.01)2

w40,45x. Another study showed that white womenhad higher prevalence of UI compared with blackand Hispanic women with rates of 41%, 31% and30%, respectively, P-0.001 w43x. Finally, a lowprevalence study from Singapore did not revealany significant difference in prevalence of UIbetween Chinese (5.2%), Malay (1.1%), and Indi-an (1.6%) elderly people (P)0.05) w31x.Fig. 3 includes pooled data from 14 studies

w7,15,16,19,20,22,24,25,27,28,30,34,40,41x show-ing the relationship between age and median prev-

Page 6: Urinary Incontinence as a Worldwide Problem

332 V.A. Minassian et al. / International Journal of Gynecology and Obstetrics 82 (2003) 327–338

Fig. 2. Prevalence of significant UI in women by age group (data from six studies).

alence of the three subtypes of UI in females. Theprevalence of stress UI peaked at the 4th decade,whereas urge and mixed UI peaked at the 8thdecade. Considering women of all ages, the mostcommon cause of UI was stress followed by mixedand urge with a mean prevalence of 50%, 32%,and 14%, respectively (Fig. 4).Twenty-one studies w3,13,14,16,17,19,21–

28,33,34,36–38,40,45x analyzed risk factors asso-ciated with UI. Common risk factors studied,excluding age, and their significance are listed inTable 2. Fifteen studies w3,7,16,20,21,23,25–28,30,31,38,40,41x reported on help-seekingbehavior (Table 3). There was no standardizeddefinition of severity of UI between the studies.Common reasons given for not seeking helpincluded: UI not seen as abnormal or serious, UIbeing part of the normal aging process, low expec-tation of treatment benefit, lack of knowledge asto where to seek treatment, embarrassment, hesi-tation, or fear to consult health care professionals,consultation cost too expensive, and others.Fourteen studies w7,14–16,18,20,21,27,30–

32,37,40,41x reviewed the effect of UI on quality

of life. Patients suffered social consequences, neg-ative feelings, andyor embarrassment in 8–74% ofcases w15,16,20,21,23,30,31x. UI had a moderateto severe impact on the quality of life in 10% w7xto 22% w27x of patients (Table 4). Physical andmental component summary scores of the shortform health questionnaire (SF-36) were signifi-cantly lower in incontinent compared to continentwomen w14x. In one study, UI was related to anoutwardly expressed anger w40x. UI interfered withmarital and sexual life in 7.5–33% of patientsw15,20,37x. Severity of UI was directly related toa negative quality of life w7,15,16,18,27,32x.

4. Discussion

UI remains a worldwide problem affecting wom-en of all ages and across different cultures andraces. The range of prevalence rates among thepublished studies is wide. This variation is due todifferences in definitions used, population sur-veyed, survey type, response rate, age, gender,availability and efficacy of health-care, and otherfactors w9x. Some studies included women of

Page 7: Urinary Incontinence as a Worldwide Problem

333V.A. Minassian et al. / International Journal of Gynecology and Obstetrics 82 (2003) 327–338

Fig. 3. Median prevalence of stress, urge, and mixed UI by age group.

Fig. 4. Overall mean prevalence of different types of UI (data from 14 studies).

all ages w3,7,14,15,19,27–30,36,38x, whereasothers included only elderly womenw13,31,32,34,35,39,41,42,44,45x. Also, many eld-erly women (over 65) are in nursing homes andthey are not accounted for in population-based

studies in community-dwelling patients. The prev-alence of UI would be higher if they were includedw10–12x.More than five different definitions of UI were

used in more than five different patient popula-

Page 8: Urinary Incontinence as a Worldwide Problem

334 V.A. Minassian et al. / International Journal of Gynecology and Obstetrics 82 (2003) 327–338

Table 2Urinary incontinence: risk factors in both sexes

Risk factor Numberof studies

Total no. of

Sig. NS

studies

Alcohol drinking 4 4Chronic cough 3 3Constipation 1 2 3Depression 1 1Diabetes 3 1 4Education 1 2 3Fecal incontinence 1 1 2Functional or motor impairment 5 5General health status 3 3Hormone replacement 1 1Income 2 2Lower urinary tract symptoms 3 3Menopause 1 4 5Parity (G1 birth) 11 7 18Previous hysterectomy 7 1 8Smoking 5 5Stroke 5 1 6Weight or body mass index 8 4 12

Sig., significance; NS, not significant.

Table 4Impact of urinary incontinence on the quality of life

Study Quality of life (%)

None Slight Moderate Severe

Hannestad w7x 66 24 6 4Holst w27x 12 66 17 5Iglesias w32x 77 11 9 1Temml w15x 34 48 11 7Median 50 36 10 4.5

Table 3Percent of patients seeking help by severity of UI

Study Country Help-seeking (%)18 author

Mild Severe Anyincontinence incontinence incontinence

Burgio w40x USA 26 55 19Dolan w26x N. Ireland 40 20Hannestad w7x Norway 54 26Holst w27x Australia 35Ju w31x Singapore 60Lagace w38x USA 14 41 28Lara w28x New Zealand 26Lionis w20x Greece 16Rekers w21x Holland 22 44 28Roberts w41x USA 13Schulman w16x Belgium 41Thomas w3x England 5 29 10Ueda w25x Japan 3Vinker w23x Israel 32Yarnell w30x S. Wales 50 9Median (range) 18 (5–26) 42.5 (29–55) 23 (3–60)

tions. In general, studies with a broad definitionof UI such as any loss of urine in a 12-monthperiod had a higher prevalence rate than those

defining UI over a shorter period of time such astwo or more wetting episodes in the past month(Table 1). Further evidence of this is found in thestudy by Thomas et al. whereby prevalence of UIin women was 8.5% when the definition used wastwo or more wetting episodes per month, and16.6% for less than two incontinence episodes w3x.There was also a wide variation in the popula-

tions sampled and response rates. Ideally, totalpopulations sampled in a certain geographical loca-tion with a high response rate reflect the prevalenceof UI more accurately than samples taken fromdoctors’ offices with a low response rate. Anotherimportant variable is the type of survey used andthe manner in which the questions about UI areasked. Fultz and Herzog showed that the use of

Page 9: Urinary Incontinence as a Worldwide Problem

335V.A. Minassian et al. / International Journal of Gynecology and Obstetrics 82 (2003) 327–338

an introduction and follow-up probe question aboutUI resulted in a doubling of the prevalence ratew4,46x.The relationship between age and prevalence of

any UI in women is not straight forward. Thepeaks in the 5th and 8th decades and the declinein between suggest that menopause may not havea positive influence on the overall prevalence ofUI (Fig. 1). In fact, four out of five studies thatexamined menopause as a risk factor did not findany significant correlation between UI and meno-pausal status (Table 2). However, median preva-lence of significant UI showed a gradual increaseto reach a prevalence rate of 18% by the 8thdecade (Fig. 2).Most studies on UI have been conducted on

white women. Studies that compared white andblack women showed that the prevalence of UIwas higher in the former w40,43,45x. Comparisongroups in other studies had small numbers and nosignificant trends could be established. It is note-worthy to mention, however, that although stressand urge incontinence are the common types ofUI in developed countries, vesico-vaginal fistulaeremain the most common cause of UI in develop-ing countries w47x.The prevalence of stress UI peaked at the 4th

decade and gradually declined thereafter to itslowest level by the 8th decade. However, theprevalence of urge and mixed incontinenceincreased after the 4th decade. These findings aresupported by a large population-based prevalencestudy in patients with overactive bladder (OAB)w48x. This survey of 16 776 patients from sixcountries showed that the prevalence of OAB(urgency, frequency, nocturia"urge incontinence)increased from 9% at 40–44 years to 31% at 75qyears. Such findings have to be interpreted withcaution, however. Sandvik et al. validated surveyquestions with clinical diagnosis made by a gyne-cologist after urodynamic testing w49x. The resultsshowed that stress incontinence was under-report-ed, in contrast to mixed incontinence which wasover-reported in epidemiologic studies. The per-centage of stress incontinence increased from 51to 77%; mixed incontinence reduced from 39 to11%; and urge incontinence changed only slightlyfrom 10 to 12%.

Alcohol, smoking, income, and hormonereplacement were not significantly related to UI.However, chronic cough, depression, functional ormotor impairment, general health status, lowerurinary tract symptoms, and history of stroke weresignificantly related to UI (Table 2). Two out ofthree studies failed to show that constipation orlevel of education were significant risk factors.Previous hysterectomy was a significant risk factor,but menopausal status was not. Finally, multiparityand obesity increased the risk of UI in 11y18(61%) and 8y12 (67%) studies, respectively. Moststudies that reviewed parity as a risk factor for UIdid not report on peripartum parameters includingthe mode of delivery that could have an influenceon the development of UI.The overwhelming majority of patients with UI

did not seek medical help for their condition in14y15 (93%) of the studies (Table 3). Even withsevere UI, only 42.5% of patients consulted ahealth care professional. UI remains an underre-ported and embarrassing condition across all coun-tries and cultures. Severity of UI (volume,frequency, and duration) was directly related todecreased quality of life w7,15,16,18,27,32x.Although 50% of patients reported that UI affectedtheir quality of life at least slightly (Table 4), 77%of these patients did not seek help. People are stillnot informed about available treatment modalitiesand health care professionals need to educatepatients and incorporate questions about inconti-nence in their history forms.Future prevalence studies should aim at distin-

guishing significant UI that is clinically relevantand affects the patient’s quality of life from thatUI described as rare or occasional. Studies shoulduse standardized definitions, survey representativepopulation samples, and improve response rates.Validated tools such as the severity index shouldbe used to classify UI w50x. This index is basedon the frequency of UI which is divided into fourlevels of severity (1–4) and the amount of urineloss which is divided into three levels (1–3) (Table5). By multiplying the frequency level with theamount level, a severity index is obtained: mild(1–2), moderate (3–6), severe (8–9), and verysevere (12).

Page 10: Urinary Incontinence as a Worldwide Problem

336 V.A. Minassian et al. / International Journal of Gynecology and Obstetrics 82 (2003) 327–338

Table 5Urinary incontinence severity indicatora

A. Frequency of UI Once or lessymonth Few timesymonth Few timesyweek Every day1 2 3 4

B. Amount of UI Few drops Small splashes More1 2 3

C. Severity index (A=B) Mild Moderate Severe Very severe1–2 3–6 8–9 12

Adapted from Sandvik et al. w50x.a

The impact of UI on health care costs is sub-stantial and increasing. The condition imposes asignificant financial burden on individuals, theirfamilies, and healthcare organizations. Studiesfrom the US have reported that direct health carecosts in individuals 65 years of age and olderamounted to approximately 8.2 billion in the 1980sand 16.4 billion US$ in the 1990s w51,52x. Thecost of UI on society for individuals aged 65 yearsand older was $26.3 billion w52x. In Sweden, theestimated annual cost for UI was 1.8 billionSwedish Crowns in 1990, or approximately 2% ofthe total health care costs w53x.In conclusion, UI remains a highly prevalent

cross-cultural and costly condition that affectswomen of all ages. Risk factors are numerous andthe impact on the quality of life is substantial.Only a minority of patients seek help for theircondition. Future epidemiologic studies shouldensure unifying definitions to produce meaningfulevidence-based medicine, and to project the costsinvolved in managing this global health care prob-lem with the goal of improving the quality andavailability of health care.

References

w1x Bates P, Bradley WE, Glen E, Griffiths D, Melchior H,Rowan D, et al. The standardization of terminology oflower urinary tract function. J Urol 1979;121:551–554.

w2x Abrams P, Cardozo L, Fall M, Griffiths D, Rosier P,Ulmsten U, et al. The standardization of terminology oflower urinary tract function: report from the standardi-zation sub-committee of the International ContinenceSociety. Neurourol Urodyn 2002;21:167–178.

w3x Thomas TM, Plymat KR, Blannin J, Meade TW. Prev-alence of urinary incontinence. Br Med J1980;281:1243–1245.

w4x Herzog AR, Fultz NH. Prevalence and incidence ofurinary incontinence in community-dwelling popula-tions. J Am Geriatr Soc 1990;38:273–281.

w5x Hampel C, Wienhold D, Benken N, Eggersmann C,Thuroff JW. Definition of overactive bladder and epi-demiology of urinary incontinence. Urology 1997;50:4–14.

w6x Thom D. Variation in estimates of urinary incontinenceprevalence in the community: effects of differences indefinition, population characteristics, and study type. JAm Geriatr Soc 1998;46:473–480.

w7x Hannestad YS, Rortveit G, Sandvik H, Hunskaar S. Acommunity-based epidemiological survey of female uri-nary incontinence: the Norwegian EPINCONT study. JClin Epidemiol 2000;53:1150–1157.

w8x Milsom I. The prevalence of urinary incontinence. ActaObstet Gynecol Scand 2000;79:1056–1059.

w9x Hunskaar S, Arnold EP, Burgio K, Diokno AC, HerzogAR, Mallett VT. Epidemiology and natural history ofurinary incontinence. Int Urogynecol J Pelvic FloorDysfunct 2000;11:301–319.

w10x Resnick NM, Yalla SW, Laurino E. The pathophysiologyof urinary incontinence among institutionalized elderlypersons. N Engl J Med 1989;320:1–7.

w11x Aggazzotti G, Pesce F, Grassi D, Fantuzzi G, Righi E,De Vita D, et al. Prevalence of urinary incontinenceamong institutionalized patients: a cross-sectional epi-demiologic study in a midsized city in northern Italy.Urology 2000;56:245–249.

w12x Ouslander JG, Palmer MH, Rovner BW, German PS.Urinary incontinence in nursing homes: incidence,remission and associated factors. J Am Geriatr Soc1993;41:1083–1087.

w13x Nygaard IE, Lemke JH. Urinary incontinence in ruralolder women: Prevalence, incidence and remission. JAm Geriatr Soc 1996;44:1049–1054.

w14x Chiarelli P, Brown W, McElduff P. Leaking urine:prevalence and associated factors in Australian women.Neurourol Urodyn 1999;18:567–577.

w15x Temml C, Haidinger G, Schmidbauer J, Schatzl G,Madersbacher S. Urinary incontinence in both sexes:prevalence rates and impact on quality of life and sexuallife. Neurourol Urodyn 2000;19:259–271.

Page 11: Urinary Incontinence as a Worldwide Problem

337V.A. Minassian et al. / International Journal of Gynecology and Obstetrics 82 (2003) 327–338

w16x Schulman C, Claes H, Matthijs J. Urinary incontinencein Belgium: a population-based epidemiological survey.Eur Urol 1997;32:315–320.

w17x Alnaif B, Drutz HP. The prevalence of urinary and fecalincontinence in Canadian secondary school teenagegirls: questionnaire study and review of the literature.Int Urogynecol J Pelvic Floor Dysfunct 2001;12:134–137.

w18x Moller LA, Lose G, Jorgensen T. The prevalence andbothersomeness of lower urinary tract symptoms inwomen 40–60 years of age. Acta Obstet Gynecol Scand2000;79:298–305.

w19x Peyrat L, Haillot O, Bruyere F, Boutin JM, Bertrand P,Lanson Y. Prevalence and risk factors of urinary incon-tinence in young and middle-aged women. Br J UrolInt 2002;89:61–66.

w20x Lionis C, Vlachonikolis L, Bathianaki M, Daskalopou-los G, Anifantaki S, Cranidis A. Urinary incontinence,the hidden health problem of Cretan women: reportfrom a primary care survey in Greece. Women Health2000;31:59–66.

w21x Rekers H, Drogendijk AC, Valkenburg H, Riphagen F.Urinary incontinence in women from 35 to 79 years ofage: prevalence and consequences. Eur J Obstet GynecolReprod Biol 1992;43:229–234.

w22x Brieger GM, Mongelli M, Hin LY, Chung TKH. Theepidemiology of urinary dysfunction in Chinese women.Int Urogynecol J Pelvic Floor Dysfunct 1997;8:191–195.

w23x Vinker S, Kaplan B, Nakar S, Samuels G, Shapira G,Kitai E. Urinary incontinence in women: prevalence,characteristics and effect on quality of life. A primarycare clinic study. Isr Med Assoc J 2001;3:663–666.

w24x Bortolotti A, Bernardini B, Colli E, Di Benedetto P,Giocoli Nacci G, Landoni M, et al. Prevalence and riskfactors for urinary incontinence in Italy. Eur Urol2000;37(1):30–35.

w25x Ueda T, Tamaki M, Kageyama S, Yoshimura N, YoshidaO. Urinary incontinence among community-dwellingpeople aged 40 years or older in Japan: prevalence, riskfactors, knowledge and self-perception. Int J Urol2000;7:95–103.

w26x Dolan LM, Casson K, McDonald P, Ashe RG. Urinaryincontinence in Northern Ireland: a prevalence study.Br J Urol Int 1999;83:760–766.

w27x Holst K, Wilson PD. The prevalence of female incon-tinence and reasons for not seeking treatment. N Z MedJ 1988;101:756–758.

w28x Lara C, Nacey J. Ethnic differences between Maori,Pacific Island and European New Zealand women inprevalence and attitudes to urinary incontinence. N ZMed J 1994;107:374–376.

w29x Okonkwo JE, Obionu CO, Obiechina NJ. Factors con-tributing to urinary incontinence and pelvic prolapse inNigeria. Int J Gynecol Obstet 2001;74:301–303.

w30x Yarnell JW, Voyle GJ, Richards CJ, Stephenson TP. Theprevalence and severity of urinary incontinence in wom-en. J Epidemiol Commun Health 1981;35:71–74.

w31x Ju CC, Swan LK, Merriman A, Choon TE, Viegas O.Urinary incontinence among the elderly people of Sin-gapore. Age Ageing 1991;20:262–266.

w32x Gavira Iglesias FJ, Caridad Y, Ocerin JM,Perez Del Molino Martin J, Valderrama Gama E,Lopez Perez M, et al. Prevalence and psychosocialimpact of urinary incontinence in older people of aSpanish rural population. J Gerontol Med Sci2000;55:M207–M213.

w33x Milsom I, Ekelund P, Molander U, Arvidsson L, Ares-koug B. The influence of age, parity, oral contraception,hysterectomy and menopause on the prevalence ofurinary incontinence in women. J Urol 1993;149:1459–1462.

w34x Tseng IJ, Chen YT, Chen MT, Kou HY, Tseng SF.Prevalence of urinary incontinence and intention to seektreatment in the elderly. J Formos Med Assoc2000;99:753–758.

w35x Swaddiwudhipong W, Koonchote S, Nguntra P, Chao-vakiratipong C. Assessment of socio-economic, func-tional and medical problems among the elderly in onerural community of Thailand. Southeast Asian J TropMed Public Health 1991;22:299–306.

w36x Maral I, Ozkardes H, Peskircioglu L, Bumin MA.Prevalence of stress urinary incontinence in both sexesat or after age 15 years: a cross-sectional study. J Urol2001;165:408–412.

w37x Rizk DE, Shaheen H, Thomas L, Dunn E, Hassan M.The prevalence and determinants of health care-seekingbehavior for urinary incontinence in United Arab Emir-ates women. Int Urogynecol J 1999;10:160–165.

w38x Lagace EA, Hansen W, Hickner JM. Prevalence andseverity of urinary incontinence in ambulatory adults:an UPRNet study. J Fam Pract 1993;36:610–614.

w39x Wetle T, Scherr P, Branch LG, Resnick NM, Harris T,Evans D, et al. Difficulty with holding urine amongolder persons in a geographically defined community:prevalence and correlates. J Am Geriatr Soc1995;43:349–355.

w40x Burgio KL, Matthews KA, Engel BT. Prevalence, inci-dence and correlates of urinary incontinence in healthy,middle-aged women. J Urol 1991;146:1255–1259.

w41x Roberts RO, Jacobsen SJ, Rhodes T, Reilly WT, GirmanCJ, Talley NJ, et al. Urinary incontinence in a commu-nity-based cohort: prevalence and healthcare-seeking. JAm Geriatr Soc 1998;46:467–472.

w42x Brown JS, Seeley DG, Fong J, Black DM, Ensrud KE,Grady D. Urinary incontinence in older women: who isat risk? Study of Osteoporotic Fractures ResearchGroup. Obstet Gynecol 1996;87:715–721.

w43x Sze EH, Jones WP, Ferguson JL, Barker CD, DolezalJM. Prevalence of urinary incontinence symptomsamong black, white, and Hispanic women. Obstet Gyne-col 2002;99:572–575.

Page 12: Urinary Incontinence as a Worldwide Problem

338 V.A. Minassian et al. / International Journal of Gynecology and Obstetrics 82 (2003) 327–338

w44x Miles TP, Palmer RF, Espino DV, Mouton CP, Lichten-stein MJ, Markides KS. New-onset incontinence andmarkers of frailty: data from the Hispanic EstablishedPopulations for Epidemiologic Studies of the Elderly. JGerontol Med Sci 2001;56A:M19–M24.

w45x Fultz NH, Herzog AR, Raghunathan TE, Wallace RB,Diokno AC. Prevalence and severity of urinary incon-tinence in older African American and Caucasian wom-en. J Gerontol A Biol Sci Med Sci 1999;54:M299–M303.

w46x Fultz NH, Herzog AR. Prevalence of urinary inconti-nence in middle-aged and older women: a survey-basedmethodological experiment. J Aging Health2000;12:459–469.

w47x Danso KA, Martey JO, Wall LL, Elkins TE. Theepidemiology of genitourinary fistulae in Kumasi,Ghana, 1977–1992. Int Urogynecol J Pelvic Floor Dys-funct 1996;7:117–120.

w48x Milsom I, Abrams P, Cardozo L, Roberts RG, ThuroffJ, Wein AJ. How widespread are the symptoms of anoveractive bladder and how are they managed? Apopulation-based prevalence study. Br J Urol Int2001;87:760–766.

w49x Sandvik H, Hunskaar S, Vanvik A, Bratt H, Seim A,Hermstad R. Diagnostic classification of female urinaryincontinence: an epidemiologic survey corrected forvalidity. J Clin Epidemiol 1995;48:339–345.

w50x Sandvik H, Seim A, Vanvik A, Hunskaar S. A severityindex for epidemiological surveys of female urinaryincontinence: comparison with 48-hour pad-weighingtests. Neurourol Urodyn 2000;19:137–145.

w51x Hu T. Impact of urinary incontinence on health carecosts. J Am Geriatr Soc 1990;38:392–395.

w52x Wagner TH, Hu TW. Economic costs of urinary incon-tinence in 1995. Urology 1998;51:355–361.

w53x Milsom I. The prevalence of urinary incontinence. ActaObstet Gynecol Scand 2000;79:1056–1059.