J Epidemiol Community Health: first published as 10.1136/jech.53.8.453 on 1 August 1999. Downloaded from J Epidemiol Community Health 1999;53:453–458 453 Urinary incontinence in middle aged women: childhood enuresis and other lifetime risk factors in a British prospective cohort Diana Kuh, Linda Cardozo, Rebecca Hardy Abstract and, in adulthood, on childbearing, a history of Study objective—To investigate the preva- kidney and bladder problems, body weight, hys- lence and lifetime risk factors for urinary terectomy and menopausal status, which previ- incontinence in middle aged women. ous general population studies have suggested Design—Nationally representative birth are related to incontinence in midlife.138912–16 cohort study with prospective data on childhood enuresis, measured adult height and weight, childbearing histories Methods and measures of socioeconomic status The Medical Research Council National Sur- updated at regular contacts, and measures vey of Health and Development is a socially of menopausal status, symptomatology stratified cohort of 2548 women and 2814 men and health care in midlife. followed up 19 times between their birth in March 1946 and the age of 43 years.17–19 When Setting—England, Scotland and Wales. —General population sample they were 47 years old a postal questionnaire Participants was sent to 1778 women study members with of 1333 women aged 48 years. whom the team was still in regular contact to —Fifty per cent reported Main results obtain information about the timing of the symptoms of stress incontinence and 22% menopause, health changes, and use of hor- reported symptoms of urge incontinence in mone replacement therapy.20 Of the original the previous year. Eight per cent had severe cohort 6% had died, 12% had refused to take symptoms. Women who at age 6 years had part at earlier follow ups, and 13% could not be wet in the day or several nights a week were traced. The first postal questionnaire was com- more likely to suVer severe incontinence pleted and returned by 1498 women giving a and report urge symptoms but occasional response rate of 84%. The second postal ques- bedwetting was not associated with an tionnaire, when the women were 48 years old, increased risk in adult life. Those who were asked a series of questions about their experi- older at the birth of their children and who ence of incontinence and was sent to 1486 had vaginal deliveries had an excess risk of women who had completed the first year’s stress symptoms. Heavier adult body questionnaire: the response rate was 93% weight was also a risk factor for these (1378) for this questionnaire. http://jech.bmj.com/ symptoms and for severe incontinence. Postmenopausal women were less likely to report stress symptoms. These risk fac- INCONTINENCE AND OTHER URINARY SYMPTOMS tors remained significant after taking In the second postal questionnaire, specific account of the increased reporting of questions designed to identify women with incontinence among women with high lev- symptoms of stress and urge incontinence and els of general symptomatology and gen- to estimate the severity of incontinence were Medical Research eral practioner visits, and of stress answered by 1333 women (97% of those who Council National symptoms among better educated women. returned the questionnaire). Severe inconti- on September 27, 2021 by guest. Protected copyright. Survey of Health and nence was defined as occurring twice a month Development, Conclusions—Urinary incontinence is a Department of common problem among middle aged or more over the previous year and the reported Epidemiology and women. This is the first prospective study loss of more than a few drops of urine. Women Public Health, of a general population sample to support were classified as moderately incontinent if University College the postulated link between childhood they reported one but not both of these symp- London Medical enuresis and adult incontinence. Child- toms. These groups were distinguished from School, Gower Street other women who reported milder levels of Campus, 1–19 bearing has long term adverse eVects, Torrington Place, particularly for older mothers. Over- incontinence or none at all. London WC1E 6BT weight is a common risk factor. Women who replied positively to the question DKuh (J Epidemiol Community Health 1999;53:453–458) “Do you ever lose any urine when you cough, R Hardy sneeze, laugh, run or exercise?” were classified as having symptoms of stress incontinence (hereaf- Department of Urinary incontinence is a common problem ter known as stress symptoms). Those who Urogynaecology, 1–11 King’s College, London among middle aged women. The purpose of replied positively both to the question “Do you L Cardozo this paper is to investigate lifetime risk factors for ever have an urgent and strong desire to pass urinary incontinence in women participating in urine which is diYcult to control?” and to the Correspondence to: a cohort study that has followed up a represen- follow up question “Do you ever lose any urine Dr D Kuh. tative sample of the general population since before you reach the toilet?” were classified as Accepted for publication their birth in 1946 until middle age. Prospective having symptoms of urge incontinence (hereaf- 29 January 1999 information is available on childhood enuresis ter known as urge symptoms). Women were J Epidemiol Community Health: first published as 10.1136/jech.53.8.453 on 1 August 1999. Downloaded from 454 Kuh, Cardozo, Hardy Table 1 Prevalence of urinary symptoms Severe incontinence Urge symptoms Stress symptoms No incontinence All women (n=107) % (n=299) % (n=665) % (n=603) % (n=1333) % Frequency > 2 hourly 25.2 18.1 11.9 7.0 9.9 Nocturia > 2 29.0 22.4 16.8 11.9 14.6 Dysuria 15.1 10.5 6.5 3.7 5.0 Uses pads for incontinence occasionally 32.7 21.7 17.3 n/a 8.9 frequently 29.9 11.7 8.7 4.5 Bothered by incontinence in everyday life a little 43.9 36.8 28.7 n/a 15.5 a lot 29.9 14.7 9.0 4.7 asked how often they passed urine during the health was obtained by deriving a score from a day and night, and if they used pads to protect checklist of 20 common symptoms given at 47 against the loss of urine. Urinary frequency dur- and 48 years.20 Symptoms that had bothered ing the day was defined as at least two hourly the women a little over the past 12 months were and nocturia as at least twice a night. They were scored 1 and those that had bothered them a also asked whether in the last year they had lot were scored 2. The scores for each year were experienced pain on urination (dysuria) or had standardised to give a mean of 0 and a standard been bothered by incontinence in their everyday deviation of 1, and then added together to give life, how long they had suVered from inconti- a total symptom score. The number of consul- nence and what, if anything, they thought had tations with the family doctor over the past 12 caused its occurrence. months was used as a measure of health care utilisation. POTENTIAL RISK FACTORS AND CONFOUNDERS Reports of incontinence may also be affected Childhood enuresis was defined from maternal by socioeconomic status and this was assessed reports of bedwetting “occasionally” or at least by the highest educational and training qualifi- “several nights a week” or wetting “sometimes” cations achieved by 26 years, classified by the during the day when study members were 6 Burnham scale23 and grouped into no qualifica- years old. tions (32% of women), less than O level (9%), Information on the birth of children was O level or equivalent (28%), A level or equival- updated at each adult home visit (the last one ent (25%) or degree level or equivalent (6%). occurring when survey members were 43 years old) and was available for 1239 women. In STATISTICAL ANALYSIS addition, in the third postal questionnaire (at Risk factors for stress and urge symptoms and 49 years), women were asked to recall the for severe incontinence are investigated using ÷2 number of caesarean deliveries—this was tests. Factors that were significant at the 10% obtained for 1057 women. level were included in the logistic regression Body mass index was derived from measure- analyses to identify the most powerful predictors ments of height and weight taken at 43 years by of stress and urge symptoms and severe inconti- research nurses using portable stadiometers nence. Body mass index rather than the four and scales and the same standardised condi- categories of body weight was entered into these http://jech.bmj.com/ tions as used at previous contacts.21 Using the models as a continuous variable. The results are same criteria as the Health Survey for presented for unweighted data as weighting the England,22 women were classified according to sample to allow for the original sample stratifica- their body mass index as either underweight tion had no appreciable eVects on the findings. (<20.0), of normal weight (20.1–25.0), over- weight (25.1–30.0) or obese (>30.0). Results Women with a history of kidney or bladder At 48 years over half the women (55%) infections, such as cystitis, were identified from reported some incontinence in the past 12 on September 27, 2021 by guest. Protected copyright. answers to a list of health problems completed months. Eight per cent had severe symptoms, by research nurses at the home visit when sur- 15% moderate symptoms and almost a third vey members were 43 years old. (32%) had only mild symptoms. Over one in Women reporting an absence of menstrua- five women (22%) suVered urge symptoms and tion for more than 12 months (8%) at the age half (50%) suVered stress symptoms.