Systematic review

Systematic review of the prevalence of faecal incontinence

A. Sharma1 , L. Yuan1 , R. J. Marshall1 , A. E. H. Merrie2 and I. P. Bissett1,2
Departments of Surgery, 1 University of Auckland and 2 Auckland City Hospital, Auckland, New Zealand
Correspondence to: Professor I. P. Bissett, Department of Surgery, University of Auckland, Private Bag 92019, Auckland, New Zealand
(e-mail: i.bissett@auckland.ac.nz)

Background: Faecal incontinence (FI) is widely recognized as a significant problem in the community.
Conjecture exists around the proportion of the population affected. This systematic review evaluated
studies reporting the community prevalence of FI in terms of methodology, design and definitions.
Methods: MEDLINE, Embase, CINAHL, the Cochrane Collaboration and National Guideline databases
were searched for studies investigating the prevalence of FI in community-based adults published from
January 1966 to February 2015. Study data, including methodology, sample size, response rate, definition
of FI and prevalence rates, were extracted on to a pro forma and appraised critically. Where possible, FI
prevalence estimates were pooled.
Results: Thirty studies were analysed from 4840 screened articles. FI prevalence estimates varied from
1⋅4 to 19⋅5 per cent. This variation was explained by differences in data collection method and two factors
within definitions of FI: type of stool and frequency of FI episodes. When these factors were accounted
for, the FI prevalence at a threshold of at least once per month for liquid or solid stool was 8⋅3–8⋅4
per cent for face-to-face or telephone interviews, and 11⋅2–12⋅4 per cent for postal surveys. The pooled
prevalence rate from studies for functional FI (defined by ROME II criteria) was 5⋅9 (95 per cent c.i. 5⋅6
to 6⋅3) per cent.
Conclusion: When comparable methodologies and definitions are used, studies produce remarkably
similar prevalence rates in different community populations. FI remains an unspoken symptom, with
lower rates reported in personal interviews compared with anonymous postal questionnaires.

Paper accepted 15 July 2016

Published online 29 September 2016 in Wiley Online Library (www.bjs.co.uk). DOI: 10.1002/bjs.10298

Introduction (such as the elderly and women), and again confirmed a

Faecal incontinence (FI) is an under-reported but common
problem, which is associated with substantial economic
burden1 – 3 . Accurate assessment of the prevalence of FI has
important implications for the development of treatments
and services4,5 . However, estimation of the severity of
the problem in the adult community population has been
difficult owing to the frequent reluctance of individuals
to volunteer their FI symptoms, despite these having a
devastating impact on their quality of life6,7 .
A previous systematic review by Macmillan and
colleagues8 in 2004 reported significant heterogeneity
in the quality of cross-sectional studies investigating FI.
In addition, it was noted that studies varied in their sen-
sitivity in detecting FI, because of inherent differences
in methodology, resulting in a wide range of prevalence
rates from 0⋅4 to 19⋅6 per cent. Pretlove and co-workers9
performed another systematic review in 2006, which
included studies involving perceived at-risk populations
significant variation in the reported FI prevalence in the
community9 .
The purpose of the present investigation was to per-
form an up-to-date systematic review of studies reporting
a prevalence estimate for FI to determine an accu-
rate estimate of the current prevalence of FI in the
community-based adult population. Particular focus was
given to review differences in response rates, study
methodologies and quality, and the definitions and
measures of FI used in the literature, to explain any
variance in the reported FI prevalence between studies.
Methods
Search strategy and study selection
The systematic literature search followed the PRISMA
guidelines10 and was designed to capture studies reporting
estimates of FI prevalence in community-based adults.

© 2016 BJS Society Ltd BJS 2016; 103: 1589–1597

Published by John Wiley & Sons Ltd
1590
Ovid MEDLINE, Embase, CINAHL, the Cochrane
Collaboration and National Guideline databases were
searched using medical subject heading (MeSH) themes
‘faecal incontinence’, ‘cross-sectional studies’ and ‘out-
come measures’ with Boolean AND/OR operators, for
studies published between January 1966 and February
2015 (inclusive) (Table S1, supporting information). Stud-
ies were limited to human subjects and adults/adolescents.
There were no language restrictions. Conference abstracts
were excluded.
Inclusion and exclusion criteria are summarized in
Table S1 (supporting information). To ensure accurate
representation of the general community-based adult pop-
ulation, studies of at-risk populations, such as the elderly or
based in clinic settings, were excluded to reduce sampling
bias. Consequently, only studies that had a lower limit for
age of inclusion of 15–50 years were selected. The purpose
of this review was to identify the overall population preva-
lence rather than to examine risk factors for FI. Studies
based on registers, for example health authority data, were
included if the register was deemed to be representative
and had adequate coverage of the general adult population.
Two reviewers independently carried out searches,
initially using titles and abstracts to exclude irrelevant
publications. Selected studies deemed to require further
evaluation were compared before the acquisition of
full manuscripts. Studies were included only with the
agreement of both reviewers following evaluation of full
manuscripts; any discrepancies were resolved by consensus
and, if required, arbitration by a third author. A manual
search of the reference lists from included articles, review
articles and commentaries was conducted to identify any
other relevant studies.
Data extraction and analysis
The same two reviewers independently extracted data from
included studies on to separate pro formas. These were
compared and a summary of the data was recorded follow-
ing consensus.
Reported prevalence rates were assessed against response
rates and study methodology (sampling, data collection
method, FI definition and study quality). An attempt was
made to estimate the pooled prevalence from studies with
sufficiently similar study methodologies and definitions.
Response rates were calculated from each study by divid-
ing the number of individuals with complete data by the
total sample size to allow uniform comparison.
Critical appraisal of study quality was carried out accord-
ing to the principles of the STROBE statement for
observational studies, and Centre for Evidence-Based
Medicine questionnaires for cross-sectional surveys11 – 14 .
© 2016 BJS Society Ltd
Published by John Wiley & Sons Ltd
A. Sharma, L. Yuan, R. J. Marshall, A. E. H. Merrie and I. P. Bissett
Quality attributes assessed included: representativeness of
community-based adult population; adequate sampling of
study population; use of a validated FI assessment tool; and
precision assessed by 95 per cent confidence intervals. One
point was given for each of: use of representative sampling;
response rate exceeding 50 per cent; use of a validated
assessment tool; and 95 per cent confidence interval for the
estimated FI prevalence of no more than 2 per cent. Studies
scoring at least 3 of 4 were considered to be of high quality.
Statistical analysis
Statistical analysis, pooling of data and construction of
graphs were done using Stata® version 13 (StataCorp,
College Station, Texas, USA). Reported prevalence rates
are presented with 95 per cent confidence intervals. Where
the confidence interval was not stated, it was calculated
using the prevalence mean and sample size, based on stan-
dard large-number sample theory for binomial data with an
‘exact’ criterion for confidence interval.
Results
A total of 4840 studies (excluding duplicates) were iden-
tified for screening, with 4786 discarded on title and
abstract alone. Of the remaining articles, 54 manuscripts
were reviewed in full, with 30 studies ultimately being
included for final analysis (Fig. 1). Studies by Whitehead
and colleagues15 and Ditah et al.16 both analysed data from
the National Health and Nutrition Examination Survey
(NHANES), but looked at different time periods so were
included as separate studies.
Table S2 (supporting information) shows the FI
prevalence estimate from each study. Prevalence rates
ranged from 1⋅4 to 19⋅5 per cent15 – 44 . All but five
studies23,24,30,32,42 were able to give 95 per cent confidence
intervals within 2 per cent of the FI prevalence estimate.
Included studies were heterogeneous with regard to mini-
mum age of recruitment, response rate, study methodology
and study definition.
Association of variables with study prevalence rates
Age
Studies included ‘adult populations’ with a varying mini-
mum age of recruitment. The majority (21 of 30 studies)
included populations with a minimum age of between 15
and 21 years. Only five studies25,29,31,37,43 included sub-
jects with minimum recruitment age of 40 years or over.
Two studies38,39 did not report a specific age range. No
relationship between the age of inclusion and FI prevalence
rate was observed.
www.bjs.co.uk BJS 2016; 103: 1589–1597
Prevalence of faecal incontinence 1591

Records identified through database searching

MEDLINE n = 2336
Embase n = 2880
CINAHL n = 648
Cochrane Collaboration n = 870

Records screened after duplicates removed

n = 4840

Records excluded
n = 4786

Full manuscripts assessed for eligibility

Excluded articles n = 27
n = 54
Duplicate study n = 4
Conferencce abstract n = 4
Not representative of adult population n = 8
Unreported FI prevalence n = 4<
Not primary study n = 5
Case–control study n = 2

Records identified from reference

lists of included studies
n=3

Studies included in qualitative

synthesis
n = 30

Studies included in pooled prevalence

analysis†
n=5

Fig. 1 PRISMA diagram showing selection of articles for review. *Attempts to contact author for original data were unsuccessful. †With
functional faecal incontinence (FI) defined by ROME II functional gastrointestinal disorder criteria

Table 1 Response rates for different survey methodologies

Postal survey Postal survey with telephone follow-up Internet survey Telephone interview Face-to-face interview

No. of studies 14 3 2 4 6
Response rate (%) 30–80⋅5 48⋅1–89 14⋅9–62⋅2 10⋅8–73 38⋅1–71⋅0
FI prevalence rate (%) 1⋅5–16⋅8 1⋅4–12⋅7 10⋅6–19⋅5 2⋅1–6⋅9 2⋅9–8⋅8

FI, faecal incontinence.

Response rates
Response rates varied from 14⋅9 to 89 per cent, with
20 of the 30 studies having response rates of 50 per
cent or greater (Table S2, supporting information). Higher
response rates were observed in studies including only
populations aged 40 years and over, and those that used
health authority registers for sampling. Response rates
were not associated with reported prevalence rates.
Study methodology
Sampling frame, method and sample size
Table S2 (supporting information) summarizes the designs
of all included studies. No association was seen between
reported prevalence rates and sampling frame, methods
and size. The commonest sampling frames used were
electoral rolls and population registers. Higher response
rates were noted for private research databases and health
authority registers.
Random sampling was used in 2715 – 33,35,36,38 – 42,44 of the
30 studies, 1119,20,22 – 26,30,41,42,44 of which used stratified
random sampling, stratifying mainly for age and sex. Three
studies15,16,36 used cluster-stratified random sampling to
include representative participants from a wide geograph-
ical area. One study43 used non-random matching to
establish a sample with characteristics representative of
the national population, and another34 used non-random
cluster sampling. The sampling frame could not be
determined for one study37 .

© 2016 BJS Society Ltd www.bjs.co.uk BJS 2016; 103: 1589–1597

Published by John Wiley & Sons Ltd
1592 A. Sharma, L. Yuan, R. J. Marshall, A. E. H. Merrie and I. P. Bissett

Table 2 Prevalence estimate of functional faecal incontinence in studies using Rome II criteria

Reference Data collection method Response rate (%) No. of respondents FI prevalence rate (%)

Boyce et al.17 Postal 62⋅2 762 7⋅6 (5⋅7, 9⋅5)

Bytzer et al.18 Postal 57⋅7 8657 4⋅2 (3⋅8, 4⋅7)
Drossman et al.22 Postal 65⋅8 5430 7⋅8 (7⋅1, 8⋅6)
López-Colombo et al.39 Face-to-face Unclear 500 4⋅6 (2⋅9, 6⋅8)
Thompson et al.36 Telephone < 50* 1149 6⋅9 (5⋅4, 8⋅4)
Overall FI prevalence range 4⋅2–7⋅6
Combined functional FI prevalence 16 498 5⋅9 (5⋅6, 6⋅3)

Values in parentheses are 95 per cent confidence intervals. *Exact response rate could not be determined owing to study methodology. FI, faecal
incontinence.

Table 3 Summary of faecal incontinence definition and effect on prevalence of faecal incontinence

Pad Lifestyle Severity FI prevalence Data collection

Reference Type of stool use Impact Staining scale Frequency rate (%) method

Thomas et al.34 Any leakage No Yes Yes No Several times a month 1⋅4 Postal + telephone
Brittain et al.29 Any leakage No No Yes Vaizey Several times a month 1⋅5 Postal
Perry et al.31 Any leakage No No Yes Vaizey Several times a month 3⋅3 Postal
Walter et al.28 Solid only* No No No No Several times a month 0⋅9 Postal
Denis et al.37 Liquid or solid No No No No No time frame 6⋅0 Telephone
De Miguel et al.41 Any leakage No No No No No time frame 8⋅8 Face-to-face
Coyne et al.43 Any stool No No No No No time frame 19⋅5 Internet
Buckley and Lapitan44 Any leakage No Yes No No No time frame 10⋅6 Internet
Lam et al.24 Any leakage† Yes No No No No time frame 15⋅0 Postal
Whitehead et al.15 Mucus, liquid, solid No No No FISI At least once per month 8⋅3 Face-to-face
Ditah et al.16 Mucus, liquid, solid No No No FISI At least once per month 8⋅4 Face-to-face
Damon et al.21 Liquid or solid Yes Yes No Wexner At least once per month 11⋅2 Postal
Sharma et al.26 Liquid or solid No No No FISI At least once per month 12⋅4 Postal + telephone
Lynch et al.30 Any leakage† Yes Yes No Wexner At least once per month 17⋅0 Postal
Nelson et al.38 Any leakage No No No No Any in past year 2⋅2 Telephone
MacLennan et al.40 Any leakage† No No No No Any in past year 2⋅9 Face-to-face
Ilnyckyj35 Liquid or solid No No No No Any in past year 3⋅7 Telephone
Aitola et al.20 Liquid or solid No No No No Any in past year 10⋅6 Postal
Kalantar et al.23 Liquid or solid No Yes No No Any in past year 11⋅2 Postal
Bartlett et al.33 Liquid or solid No No No Wexner Any in past year 12⋅7 Postal + telephone
Roberts et al.25 Any leakage No No No No Any in past year 13⋅1 Postal
Siproudhis et al.27 Any leakage† No No No No Any in past year 16⋅8 Postal
De Souza Santos and Santos32 Inadequate information No No time frame 7⋅0 Unclear
Lim et al.42 Inadequate information FISI/Wexner No time frame 4⋅7 Face-to-face

*Also provides rate of leakage of liquid stool but not a combined rate for solid and liquid stool. †Definition includes leakage of flatus (anal incontinence).
FI, faecal incontinence; FISI, Faecal Incontinence Severity Index.

Wide variations were seen in study sample sizes, ranging
from 500 to 64 749. Only nine15,16,18,26,33,36,38,39,42 of the 30
included studies provided a power calculation or justifica-
tion for the sample size used.
Data collection method
Table 1 shows the effect of data collection method
on reported prevalence rates. Postal survey with
self-administered questionnaires was the commonest
data collection method, followed by face-to-face and tele-
phone interviews. Personalized data collection methods
(telephone and face-to-face interview surveys) reported
narrower ranges of FI prevalence compared with postal
surveys (2⋅1–8⋅8 versus 1⋅5–16⋅8 per cent respectively).
Buckley and colleagues44 and Coyne et al.43 both
employed internet surveys for data collection. Reported
prevalence rates were 10⋅6 and 19⋅5 per cent. However,
both studies were limited by the use of sample frames that
inadequately represented a general adult population and
the lack of use of a validated questionnaire.
Faecal incontinence definition
A variety of definitions were used to report FI (Table S2,
supporting information). In general, definitions were based
on either principles from symptom severity scales or func-
tional symptom criteria.
Table 2 shows the prevalence rates from studies report-
ing functional FI, defined by ROME II functional

© 2016 BJS Society Ltd www.bjs.co.uk BJS 2016; 103: 1589–1597

Published by John Wiley & Sons Ltd
Prevalence of faecal incontinence 1593

25
Several times per month
At least once per month
Any in past year
No time frame
20

† †
Faecal incontinence (~)

†
15

10
˜ ˜ ˜

˜
5
˜
˜
˜
0
. 34

. 29

. 31

l. 28

. 15

. 16

. 21

. 26

. 30

. 38

. 40

j 35

. 20

. 23

l. 33

. 25

. 27

. 37

l. 41

. 24
ky
al

al

al

al

al

al

al

al

al

al

al

al

al

al

al

al
ta

ta

ta
yc
et

et

et

et

et

et

et

et

et

et

et

et

et

et

et

et
re

te

le
Iln
as

rry

ad

on

an

la

ar

ts

is

is

m
ue
et
te
ai

ita

nc

so

dh

en
to

er

La
nt
om

he

am

nn

rtl
Pe

al
itt

ig
ar
D

Ly

el

Ai

ob
la

ou

D
W

Ba
Br

M
te

Le
Sh
Th

N
D

Ka

pr
hi

e
ac

D
W

Si
M

Fig. 2 Faecal incontinence (FI) rates in studies arranged by stool frequency definition. Only postal or telephone-based FI prevalence
studies are included. *Studies with telephone or face-to-face interview; all other studies used postal data collection. †Postal studies with
flatus in definition. Studies with the definition ‘several times per month’ had lower estimates of FI prevalence. Telephone and
face-to-face interviews were associated with lower estimates of FI prevalence regardless of stool frequency in definition

gastrointestinal disorders criteria45 . Prevalence values month’ (Fig. 2) reported the lowest FI prevalence (0⋅9–3⋅3
from four studies using the ROME II criteria were com- per cent)28,29,31,34 . Of these, Walter and colleagues28
bined with the reported prevalence from Drossman and included only solid stool and had the lowest value (0⋅9
colleagues22 , who used a closely associated definition set per cent).
before the development of the ROME II working party Frequency definitions of ‘at least once
definitions. The combined functional FI prevalence was per month’15,16,21,26,30 and ‘any in the past
5⋅9 (95 per cent c.i. 5⋅6 to 6⋅3) per cent. Koloski and year’20,23,25,27,33,35,38,40 reported a similar range of preva-
co-workers19 also reported functional FI, but modified the lence rates (8⋅3–17⋅0 per cent and 2⋅2–16⋅8 per cent
time scale to present 1- and 12-month prevalence rates, so respectively). Exclusion of face-to-face or telephone inter-
their study was not included in pooled analysis. view methods (associated with lower reported prevalence
Parameters used to define FI for each of the remaining rates), and exclusion of flatus definition (associated with
24 studies are shown in Table 3. These included: minimum higher reported prevalence rates), resulted in further sim-
frequency; type of stool leakage; impact on life (pad use, ilarity of FI prevalence rates (11⋅2–12⋅4 per cent for ‘at
lifestyle impact or staining); and use of a defined severity least once per month’ and 10⋅6–13⋅1 per cent for ‘any in
scale. Only two pairs of studies used identical definitions the last year’).
(Ditah and colleagues16 and Whitehead et al.15 ; Perry and Inclusion of flatus in study definitions (anal
co-workers31 and Brittain et al.29 ). incontinence)24,27,30 was associated with the highest FI
Fig. 2 demonstrates the effects of varying definition prevalence rates (15⋅0–17⋅0 per cent) irrespective of fre-
parameters on reported FI prevalence. Frequency and type quency differences in the definition (Fig. 2). MacLennan
of stool were the main determinants of the differences and colleagues40 and Nelson et al.38 reported FI defini-
in reported FI prevalence. Studies used definitions of FI tions including flatus that had low prevalence estimates
that varied in frequency from ‘several times per day’ to (2⋅9 and 2⋅2 per cent respectively); however, both studies
‘any incontinence in the last year’. Studies with definitions also included personalized (face-to-face and telephone)
including incontinence frequency of ‘several times per data collection methods.

© 2016 BJS Society Ltd www.bjs.co.uk BJS 2016; 103: 1589–1597

Published by John Wiley & Sons Ltd
1594
Three studies21,24,30 that included pad use in their def-
inition had an estimated prevalence range of 11⋅2–17⋅0
per cent. However, two of these included any leakage in
their stool definitions. Three studies29,31,34 that included
staining in their definition reported low prevalence ranges
(1⋅3–3⋅3 per cent), but all three included ‘several times per
month’ in their definitions. Five studies21,23,30,34,44 included
lifestyle impact in their definitions, which did not influence
estimated prevalence rates.
Nine studies15,16,21,26,29 – 31,33,42 included various severity
scales in their definitions, including: Faecal Incontinence
Severity Index (FISI), St Mark’s tool by Vaisey and col-
leagues, the Wexner score by Miller et al., and one devel-
oped by Bharucha and colleagues46 – 53 . Despite use of com-
parable severity scores, estimated prevalence rates varied
widely (1⋅5–17⋅0 per cent), likely reflecting the hetero-
geneity of stool type and frequency definitions used by the
different studies.
FI prevalence at a threshold of at least once per month for
liquid or solid stool was 8⋅3–8⋅4 per cent for face-to-face
or telephone interviews, and 11⋅2–12⋅4 per cent for postal
surveys. These results excluded studies likely to under-
estimate (personalized data collection methods and incont-
inence frequency of ‘several times per month’) or over-
estimate (flatus inclusion in definition) the FI prevalence.
Study quality
Study quality was scored out of 4 (Table S2, supporting
information). In total, 19 of the 30 included studies were
deemed of high quality. Bias was minimized adequately in
only ten studies.
Validated questionnaires were used by 12
studies15,16,20,21,23 – 26,30,33,35,42 . Three15,16,21 of these did
not specifically validate their questionnaire but used the
FISI or Wexner scoring systems, both of which have
been validated previously in patient populations. Six
studies17 – 19,22,36,39 used the Rome I or Rome II functional
definition criteria for FI, which are based on expert opinion
and used widely in clinical practice and research. Twelve
studies27 – 29,31,32,34,37,38,40,41,43,44 used tools that had not
undergone validation.
Discussion
This is an up-to-date systematic review of studies inves-
tigating the prevalence of FI in community-based adult
populations. Previous studies reported wide ranges of FI
prevalence, from 0⋅4 to 19⋅6 per cent8,9 , partly owing to
the heterogeneity of study designs. The present system-
atic review showed that FI prevalence estimates among
© 2016 BJS Society Ltd
Published by John Wiley & Sons Ltd
A. Sharma, L. Yuan, R. J. Marshall, A. E. H. Merrie and I. P. Bissett
studies are comparable when research methodologies are
uniform. The large reported variations in prevalence can be
explained mainly by three factors: data collection methods,
and frequency of incontinence and type of stool leakage in
FI definition (Fig. 2).
Data collection methodology was shown to be a factor
influencing FI prevalence estimates. Lower prevalence
rates were seen when personalized data collection methods
were used (face-to-face and telephone interview) compared
with postal surveys. This may be due to under-reporting
with personalized data collection methods, an established
reason for lower reported prevalence of sensitive sub-
ject areas owing to reluctance by individuals to relay
information accurately because of embarrassment and
stigmatization6,7 . Further supporting this is the similar
prevalence reported between face-to-face and telephone
surveys. One rationale for the use of personalized data
collection methods over postal surveys is an expected
higher response rate, reducing the risk of sampling bias38 ;
however, this was not observed in the present review.
Response rates were similar across data collection methods
and were generally adequate, with two-thirds of all studies
reporting response rates of over 50 per cent.
The parameters used in the definition also affected
reported prevalence rates. Among the different parameters
used, two appeared to have the greatest influence on rates:
frequency of incontinence and type of stool. Studies using
similar data collection methods and type of stool parame-
ters in their definitions reported much lower FI prevalence
when frequency thresholds were set at ‘several times per
month’ compared with ‘at least once per month’ (1⋅4–3⋅3
per cent and 11⋅2–12⋅4 per cent respectively). This sug-
gests that defining incontinence frequency at a threshold
of ‘several times per month’ may be too restrictive to rep-
resent accurately the FI prevalence in the general adult
community, leading to underestimation. Broadening the
frequency threshold to ‘any in past year’ resulted in lit-
tle change to reported prevalence rates compared with ‘at
least once per month’. To allow enough sensitivity to reflect
the general adult population, the authors recommend the
use of ‘at least once per month’ as the frequency threshold
within the definition of FI.
The inclusion of flatus in the type of stool parameter in
FI definition was associated with higher prevalence rates,
irrespective of frequency. To ensure consistent measure-
ment, the authors recommend the use of ‘both liquid and
stool’ in the definition of FI. Measurements including fla-
tus incontinence should be reported separately as ‘anal
incontinence’.
To reflect the general adult community population, stud-
ies examining the prevalence of FI in at-risk populations,
www.bjs.co.uk BJS 2016; 103: 1589–1597
Prevalence of faecal incontinence
such as those with previous obstetric trauma or diabetes,
were not included. This review demonstrated no associa-
tion between the minimum age of inclusion and reported
prevalence rates. However, because only five of the 30 stud-
ies examined prevalence in subjects limited to 40 years and
over, and on account of the heterogeneous study designs
and definitions, no accurate age-stratified analysis could be
made to allow comment on the association between age and
reported prevalence.
Epidemiological research for population prevalence
requires considerable investment and commitment in
resources. The relative paucity of studies looking at popula-
tion prevalence of FI in low- and middle-income countries
was therefore not surprising. In more resource-constrained
settings, studies were often designed to focus on at-risk
populations, for example women following birth trauma,
so that appropriate therapy could be targeted54 – 61 . How-
ever, larger-scale prevalence research is achievable by
adapting research methods, such as the study conducted by
Zuckerman et al.62 . Further prevalence research from low-
and middle-income countries is important as there may
be differences in the geographical distribution of disease
prevalence, possibly as a result of varying rates of associated
risk factors9 . Differences in FI prevalence between ethnic
groups were not examined as a primary outcome in any of
the included studies; however, oversampling in NHANES
data sets to assess reliable FI estimates in different ethnic
groups showed no significant difference15,16 .
Although resource constraints may exist, it was noted
that the majority of studies in the present review were
able to report response rates of over 50 per cent from the
general population. Sampling frames using health author-
ity databases tended to achieve higher response rates.
Internet-based surveys provide a useful tool for simple
acquisition and transfer of data in epidemiological studies.
However, the quality of data obtained by this method was
often reduced by bias in sampling frames, limiting the gen-
eral applicability of the results to the general population
setting.
Consensus in definition is important to allow equal
comparison of data to guide management, accurately
estimate the burden of disease, and guide provision of
resources for disease prevention and management. The
use of ROME II criteria for functional FI resulted in
a narrow range of FI prevalence estimates, despite use
of different study designs. However, one weakness of
using functional FI in prevalence estimation of FI is the
possibility of underestimation, as it excludes people with
structural abnormalities (ROME II) (and neurological
abnormalities in the more recent ROME III criteria),
many of whom will be community-based adults.
© 2016 BJS Society Ltd
Published by John Wiley & Sons Ltd
1595
Frequency and type of stool parameters within defini-
tions are important variables affecting reported FI preva-
lence. Therefore, consensus definitions for these two fac-
tors are important to aid further research and policy devel-
opment. To avoid underestimation, the authors recom-
mend the use of anonymous survey collection methods and
avoiding the use of ‘several times per month’ as a frequency
parameter. Removing flatus from the definition and catego-
rizing it separately will reduce the risk of overestimation of
prevalence.
Disclosure
The authors declare no conflict of interest.
References
1 Bharucha AE, Dunivan G, Goode PS, Lukacz ES, Markland
AD, Matthews CA et al. Epidemiology, pathophysiology,
and classification of fecal incontinence: State of the Science
Summary for the National Institute of Diabetes and
Digestive and Kidney Diseases (NIDDK) workshop. Am J
Gastroenterol 2015; 110: 127–136.
2 Nelson RL. Epidemiology of fecal incontinence.
Gastroenterology 2004; 126(Suppl 1): S3–S7.
3 Deutekom M, Dobben AC, Dijkgraaf MGW, Terra MP,
Stoker J, Bossuyt PM. Costs of outpatients with fecal
incontinence. Scand J Gastroenterol 2005; 40: 552–558.
4 National Institute for Health and Clinical Excellence
(NICE). Faecal Incontinence. The Management of Faecal
Incontinence in Adults. Clinical Guideline 49. NICE: London,
2007.
5 Wagg AS, Newman DK, Leichsenring K, van Houten P.
Developing an internationally-applicable service
specification for continence care: systematic review, evidence
synthesis and expert consensus. PLoS ONE 2014; 9: e104129.
6 Johanson JF, Lafferty J. Epidemiology of fecal incontinence:
the silent affliction. Am J Gastroenterol 1996; 91: 33–36.
7 Leigh RJ, Turnberg LA. Faecal incontinence: the unvoiced
symptom. Lancet 1982; 1: 1349–1351.
8 Macmillan AK, Merrie AE, Marshall RJ, Parry BR. The
prevalence of fecal incontinence in community-dwelling
adults: a systematic review of the literature. Dis Colon Rectum
2004; 47: 1341–1349.
9 Pretlove SJ, Radley S, Toozs-Hobson PM, Thompson PJ,
Coomarasamy A, Khan KS. Prevalence of anal incontinence
according to age and gender: a systematic review and
meta-regression analysis. Int Urogynecol J Pelvic Floor
Dysfunct 2006; 17: 407–417.
10 Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred
reporting items for systematic reviews and meta-analyses:
the PRISMA statement. PLoS Med 2009; 6: e1000097.
11 von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC,
Vandenbroucke JP; STROBE Initiative. Strengthening the
Reporting of Observational Studies in Epidemiology
www.bjs.co.uk BJS 2016; 103: 1589–1597
1596
(STROBE) statement: guidelines for reporting
observational studies. BMJ 2007; 335: 806–808.
12 Sanderson S, Tatt ID, Higgins JP. Tools for assessing quality
and susceptibility to bias in observational studies in
epidemiology: a systematic review and annotated
bibliography. Int J Epidemiol 2007; 36: 666–676.
13 Munn Z, Moola S, Riitano D, Lisy K. The development of a
critical appraisal tool for use in systematic reviews
addressing questions of prevalence. Int J Health Policy Manag
2014; 3: 123–128.
14 Heneghan C. EBM resources on the new CEBM website.
Evid Based Med 2009; 14: 67.
15 Whitehead WE, Borrud L, Goode PS, Meikle S, Mueller
ER, Tuteja A et al. Fecal incontinence in US adults:
epidemiology and risk factors. Gastroenterology 2009; 137:
512–517.e2.
16 Ditah I, Devaki P, Luma HN, Njei B, Jaiyeoba C, Salami A
et al. Prevalence, trends, and risk factors for fecal
incontinence in United States adults, 2005–2010. Clin
Gastroenterol Hepatol 2014; 12: 636–643.e2.
17 Boyce PM, Talley NJ, Burke C, Koloski NA. Epidemiology
of the functional gastrointestinal disorders diagnosed
according to Rome II criteria: an Australian
population-based study. Int Med J 2006; 36: 28–36.
18 Bytzer P, Talley NJ, Leemon M, Young LJ, Jones MP,
Horowitz M. Prevalence of gastrointestinal symptoms
associated with diabetes mellitus: a population-based survey
of 15 000 adults. Arch Intern Med 2001; 161: 1989–1996.
19 Koloski NA, Talley NJ, Boyce PM. Epidemiology and
health care seeking in the functional GI disorders: a
population-based study. Am J Gastroenterol 2002; 97:
2290–2299.
20 Aitola P, Lehto K, Fonsell R, Huhtala H. Prevalence of
faecal incontinence in adults aged 30 years or more in
general population. Colorectal Dis 2010; 12: 687–691.
21 Damon H, Guye O, Seigneurin A, Long F, Sonko A,
Faucheron JL et al. Prevalence of anal incontinence in adults
and impact on quality-of-life. Gastroenterol Clin Biol 2006;
30: 37–43.
22 Drossman DA, Li Z, Andruzzi E, Temple RD, Talley NJ,
Thompson WG et al. U.S. Householder survey of functional
gastrointestinal disorders. Prevalence, sociodemography,
and health impact. Dig Dis Sci 1993; 38: 1569–1580.
23 Kalantar JS, Howell S, Talley NJ. Prevalence of faecal
incontinence and associated risk factors; an underdiagnosed
problem in the Australian community? Med J Aust 2002;
176: 54–57.
24 Lam TCF, Kennedy ML, Chen FC, Lubowski DZ, Talley
NJ. Prevalence of faecal incontinence: obstetric and
constipation-related risk factors; a population-based study.
Colorectal Dis 1999; 1: 197–203.
25 Roberts RO, Jacobsen SJ, Reilly WT, Pemberton JH,
Lieber MM, Talley NJ. Prevalence of combined fecal and
urinary incontinence: a community-based study. J Am
Geriatr Soc 1999; 47: 837–841.
© 2016 BJS Society Ltd
Published by John Wiley & Sons Ltd
A. Sharma, L. Yuan, R. J. Marshall, A. E. H. Merrie and I. P. Bissett
26 Sharma A, Marshall RJ, Macmillan AK, Merrie AE, Reid P,
Bissett IP. Determining levels of fecal incontinence in the
community: a New Zealand cross-sectional study. Dis Colon
Rectum 2011; 54: 1381–1387.
27 Siproudhis L, Pigot F, Godeberge P, Damon H, Soudan D,
Bigard MA. Defecation disorders: a French population
survey. Dis Colon Rectum 2006; 49: 219–227.
28 Walter S, Hallbook O, Gotthard R, Bergmark M, Sjodahl R.
A population-based study on bowel habits in a Swedish
community: prevalence of faecal incontinence and
constipation. Scand J Gastroenterol 2002; 37: 911–916.
29 Brittain K, Perry S, Shaw C, Matthews R, Jagger C, Potter
J. Isolated urinary, fecal, and double incontinence:
prevalence and degree of soiling in stroke survivors. J Am
Geriatr Soc 2006; 54: 1915–1919.
30 Lynch AC, Dobbs BR, Keating J, Frizelle FA. The
prevalence of faecal incontinence and constipation in a
general New Zealand population; a postal survey. N Z Med J
2001; 114: 474–477.
31 Perry S, Shaw C, McGrother C, Matthews RJ, Assassa RP,
Dallosso H et al. Prevalence of faecal incontinence in adults
aged 40 years or more living in the community. Gut 2002;
50: 480–484.
32 de Souza Santos CR, Santos VL. Prevalence of fecal
incontinence in the urban population of Pouso Alegre,
Minas Gerais, Brazil. Rev Esc Enferm USP 2011; 45:
180–186.
33 Bartlett LM, Nowak MJ, Ho Y. Faecal incontinence in rural
and regional northern Queensland community-dwelling
adults. Rural Remote Health 2014; 14: 1–17.
34 Thomas TM, Egan M, Walgrove A, Meade TW. The
prevalence of faecal and double incontinence. Community
Med 1984; 6: 216–220.
35 Ilnyckyj A. Prevalence of idiopathic fecal incontinence in a
community-based sample. Can J Gastroenterol 2010; 24:
251–254.
36 Thompson WG, Irvine EJ, Pare P, Ferrazzi S, Rance L.
Functional gastrointestinal disorders in Canada: first
population-based survey using Rome II criteria with
suggestions for improving the questionnaire. Dig Dis Sci
2002; 47: 225–235.
37 Denis P, Bercoff E, Bizien MF, Brocker P, Chassagne P,
Lamouliatte H et al. [Prevalence of anal incontinence in
adults.] Gastroenterol Clin Biol 1992; 16: 344–350.
38 Nelson R, Norton N, Cautley E, Furner S.
Community-based prevalence of anal incontinence. JAMA
1995; 274: 559–561.
39 López-Colombo A, Morgan D, Bravo-González D,
Montiel-Jarquín A, Méndez-Martínez S, Schmulson M. The
epidemiology of functional gastrointestinal disorders in
Mexico: a population-based study. Gastroenterol Res Pract
2012; 2012: 606174.
40 MacLennan AH, Taylor AW, Wilson DH, Wilson D. The
prevalence of pelvic floor disorders and their relationship to
gender, age, parity and mode of delivery. BJOG 2000; 107:
1460–1470.
www.bjs.co.uk BJS 2016; 103: 1589–1597
Prevalence of faecal incontinence
41 De Miguel M, Rodriguez JM, Ortiz H, Yarnoz MC, Marzo
J, Artieda C. Prevalence of alterations in the defecatory
function in the population of Navarra. An Sist Sanit Navar
1999; 22(Suppl 3): 17–24.
42 Lim JW, Heng C, Wong MT, Tang CL. Prevalence of
faecal incontinence in the community: a cross-sectional
study in Singapore. Singapore Med J 2014; 55: 640–643.
43 Coyne KS, Cash B, Kopp Z, Gelhorn H, Milsom I,
Berriman S et al. The prevalence of chronic constipation and
faecal incontinence among men and women with symptoms
of overactive bladder. BJU Int 2011; 107: 254–261.
44 Buckley BS, Lapitan MC. Prevalence of urinary and faecal
incontinence and nocturnal enuresis and attitudes to
treatment and help-seeking amongst a community-based
representative sample of adults in the United Kingdom. Int J
Clin Pract 2009; 63: 568–573.
45 Whitehead WE, Wald A, Diamant NE, Enck P, Pemberton
JH, Rao SS. Functional disorders of the anus and rectum.
Gut 1999; 45(Suppl 2): ii55–ii59.
46 Vaizey CJ, Carapeti E, Cahill JA, Kamm MA. Prospective
comparison of faecal incontinence grading systems. Gut
1999; 44: 77–80.
47 Talley NJ, Phillips SF, Melton J III, Wiltgen C, Zinsmeister
AR. A patient questionnaire to identify bowel disease. Ann
Intern Med 1989; 111: 671–674.
48 Talley NJ, Boyce PM, Owen BK, Newman P, Paterson KJ.
Initial validation of a bowel symptom questionnaire and
measurement of chronic gastrointestinal symptoms in
Australians. Aust N Z J Med 1995; 25: 302–308.
49 Rockwood TH, Church JM, Fleshman JW, Kane RL,
Mavrantonis C, Thorson AG et al. Patient and surgeon
ranking of the severity of symptoms associated with fecal
incontinence: the Fecal Incontinence Severity Index. Dis
Colon Rectum 1999; 42: 1525–1532.
50 Reilly WT, Talley NJ, Pemberton JH, Zinsmeister AR.
Validation of a questionnaire to assess fecal incontinence and
associated risk factors: Fecal Incontinence Questionnaire.
Dis Colon Rectum 2000; 43: 146–153.
Supporting information
Additional supporting information may be found in the online version of this article:
Table S1 Literature search strategy, and inclusion and exclusion criteria (Word document)
Table S2 Summary of included studies (Word document)
© 2016 BJS Society Ltd
Published by John Wiley & Sons Ltd
1597
51 Miller R, Bartolo DC, Locke-Edmunds JC, Mortensen NJ.
Prospective study of conservative and operative treatment
for faecal incontinence. Br J Surg 1988; 75: 101–105.
52 Jorge JM, Wexner SD. Etiology and management of fecal
incontinence. Dis Colon Rectum 1993; 36: 77–97.
53 Bharucha AE, Locke GR III, Seide BM, Zinsmeister AR. A
new questionnaire for constipation and faecal incontinence.
Aliment Pharmacol Ther 2004; 20: 355–364.
54 Adaji SE, Olajide FM. Pelvic floor distress symptoms within
9 weeks of childbirth among Nigerian women. Eur J Obst
Gynecol Reprod Biol 2014; 174: 54–58.
55 Alimohammadian M, Ahmadi B, Janani L, Mahjubi B.
Suffering in silence: a community-based study of fecal
incontinence in women. Int J Colorectal Dis 2014; 29:
401–406.
56 de Araujo MP, Takano CC, Girão MJ, Sartori MG. Pelvic
floor disorders among indigenous women living in Xingu
Indian Park, Brazil. Int Urogynecol J Pelvic Floor Dysfunct
2009; 20: 1079–1084.
57 Ge J, Lu YX, Shen WJ, Zhang Y, Li XY, Yang P et al.
[Prevalence of fecal incontinence among adult women in
Beijing district.] Chung Hua Fu Chan Ko Tsa Chih 2010; 45:
669–672.
58 Johnson K. Incontinence in Malawi: analysis of a proxy
measure of vaginal fistula in a national survey. Int J Gynecol
Obstet 2007; 99(Suppl 1): S122–S129.
59 Jokhio AH, Rizvi RM, Rizvi J, MacArthur C. Prevalence of
obstetric fistula: a population-based study in rural Pakistan.
BJOG 2014; 121: 1039–1046.
60 Okonkwo J, Obionu CN, Okonkwo CV, Obiechina NJ.
Anal incontinence among IGBO (Nigerian) women. Int J
Clin Pract 2002; 56: 178–180.
61 Islam RM, Bell RJ, Hossain MB, Davis SR. Bangladesh
Midlife Women’s Health Study (BMWHS): methods,
challenges and experiences. Maturitas 2014; 80: 89–94.
62 Zuckerman MJ, Nguyen G, Ho H, Nguyen L, Gregory
GG. A survey of irritable bowel syndrome in Vietnam using
the Rome criteria. Dig Dis Sci 2006; 51: 946–951.
www.bjs.co.uk BJS 2016; 103: 1589–1597