Objective. Urinary incontinence is a common and highly embarrassing co
ndition among females of all age groups and has been the subject of se
veral epidemiological studies in the past. Methods. From an extensive
literature search covering the time period from 1954 to 1995, 48 epide
miological studies and several other publications dealing with prevale
nce and natural history of female incontinence were reviewed. Meta-ana
lysis of reported data was performed in respect of incontinence defini
tions, investigation methods, home country of survey and age groups. R
esults. Available data are inhomogenous and difficult to compare becau
se of differences in definitions of incontinence, target populations a
nd study design in different investigations. By grouping the studies b
y similarities in the above criteria and analysing the results for eac
h group of studies, an attempt was made to understand the great variat
ion of reported results. Differences in prevalence of incontinence wer
e identified for all examined aspects and for distinct ethnic populati
ons. The little information that exists on the incidence, spontaneous
remission rates and risk factors were used to elucidate the natural hi
story of female incontinence. Conclusion. A generally accepted definit
ion of incontinence is highly desirable and should comprise aspects of
severity and demonstrability of the condition, bother factor and impa
ct on quality of life. Furthermore, basic requirements for epidemiolog
ical surveys of incontinence such as validation of questionnaire resul
ts need to be defined and standardised to establish a sensible basis f
or useful epidemiological studies in the future.