Int J Colorect Dis (1998) 13: 12

Springer-Verlag 1998

REVIEW

M. Bardaj F. Roset R. Camps F. Sant

M. J. Fernndez-Layos

Symptomatic colonic lipoma: differential diagnosis of large bowel tumors

Accepted: 10 September 1997

Abstract Symptomatic colonic lipomas, although unusual, continue to present difficulties in the preoperative
differential diagnosis between malignant and benign colonic neoplasm. Although new imaging techniques are
available, they are frequently diagnosed at laparotomy, and
definitive histology is required. Local excision is adequate
treatment, but segmental excision may be necessary when
there is doubt about the diagnosis, or when a complication
occurs.
Key words Colonic lipoma Differential diagnosis
Excision

Lipomas of the gastrointestinal tract are rare and were first

described by Bauer in 1757. Lipomas are the second most
common benign colonic tumor after adenomatous polyps
[1]. Most of these tumors remain entirely asymptomatic
and are usually detected either by chance during the investigation of symptoms apparently deriving from the large
bowel or are found in a large-bowel specimen removed for
some other reason. Nonetheless, they are a relatively uncommon clinical entity and sometimes present with symptoms similar to those of colonic malignancies [1, 2]. In the
Mayo Clinic series 46% of large-bowel lipomas were discovered incidentally in specimens removed for other diseases. Of these, 11% were resected because of a neoplasm
suspected of being a carcinoma, and 6% were symptomatic
[2].
Lipomas of the large bowel are uncommon fatty neoplasms, with a reported incidence ranging between 0.2%
and 4.4% [3]. The most common sites of lipomas in the
large intestine are the cecum, ascending colon, and sigM. Bardaj () F. Roset R. Camps M. J. Fernndez-Layos
Department of Surgery, Hospital General de Manresa, La Culla
Street, Manresa, E-08240 Barcelona, Spain
F. Sant
Department of Pathology, Hospital General de Manresa, La Culla
Street, Manresa, E-08240 Barcelona, Spain

moid colon, in decreasing frequency; 70% are localized to

the right hemicolon. Lipomas arise from the submucosa in
approximately 90% of cases and from the subserosa and
intermucosa in the remaining cases. Their size ranges from
2 mm to 30 cm. The majority of patients are between
40 and 70 years of age. Colonic lipomas are more common
in women than in men [1 3]. There in a predilection for
the right colon in women and the left colon in men. Multiple lipomas are noted in 10 20% of cases, particularly
when a lipoma is found in the cecum [2, 4, 5].
A lipoma smaller than 2 cm rarely causes symptoms. A
large lipoma can cause abdominal pain from partial bowel
obstruction, and bleeding can occur if the mucosa overlying the lipoma is ulcerated [2]. Colicky pain may be caused
by intermittent intussusception. Lipomas are the most common benign neoplasms that cause intussusception in adults.
Spontaneous expulsion of a sigmoid lipoma has also been
recorded.
Pathologically, lipomas are well-differentiated tumors
arising from deposits of adipose connective tissue in the
bowel. Sarcomatous changes, in colonic lipomas have not
been reported, but intermittent torsion and relative ischemia can give rise to a pseudomalignant appearance. The
preoperative diagnosis is of clinical revelance. Malignant
colonic neoplasms and colonic lipomas present in similar
age groups and may have similar symptoms. Patients with
malignant disease may undergo major surgery, including
resection of the involved segment and regional lymph
nodes while patients with colonic lipoma may undergo simple excision. Despite recent diagnostic innovations it has
been reported that the preoperative diagnostic accuracy is
only about 62%.
The radiographic contour of colonic lipomas often resembles that of carcinomas, myxomas, or neurofibromas.
Radiolucency of a lipoma, reflective of its fat content, has
been advanced as a means of identification. The squeeze
sign is considered pathognomonic for colonic lipomas,
and is characterized by the elongation of a spherical filling defect during peristalsis. Based on these radiological
signs some authors claim a preoperative diagnostic accuracy of more than 60%. Colonoscopy may allow direct vis-

ualization of the submucosal lipoma, which appears as a

mass covered by normal mucosa, and some of the endoscopic features have been described including tenting
sign (grasping the overlying mucosa), cushion sign
(flattening and restoration of the shape of the lipoma), and
the naked fat sign (extrusion of fat after biopsy of the
colonic mucosa) [2 5]. Colonoscopy is reliable for the
diagnosis of a typical lipoma, but it may prove to be of no
help when the lesion is atypical. Computed tomography
(CT) has been proposed as a noninvasive means of diagnosis [3, 6]. The CT characteristics of lipoma include a
spherical or ovoid shape; smooth, sharply demarcated
margins; and homogeneous density with CT values
between 40 and 120 HU. Two potential limitations of
CT are the size of lipomas and the partial volume averaging with surrounding soft tissue structures or stool, creating artifactually high CT attenuation values. Magnetic resonance imaging has recently been used successfully, but
further evaluation is still necessary [1]. In most cases
therefore the histological diagnosis is arrived at only after excision of the tumor. However, although most lipomas are incidental findings and require no treatment, there
is a small subgroup that do require surgical intervention,
including those with suspicion of malignancy, symptomatic lipomas, surgical emergencies such as intussusception and obstruction with ulceration and bleeding, and very rarely massive hemorrhage [8]. Because the majority
of lipomas are submucosal, endoscopic removal entails a
high risk. Colonoscopic polypectomy of a large lipoma
should be avoided since its high water content requires a
tremendous amount of heat to cut through the lipoma. Endoscopic removal is occasionally possible (especially if
pedunculated), but surgery should be the treatment of
choice [2, 3, 9].

Surgery is indicated for symptomatic lesions when a

complication (obstruction, intussusception, or bleeding)
occurs and for cases in which a carcinoma cannot be excluded with certainty. A wide range of operative techniques
have been described, including enucleation, colotomy and
excision, and segmental colonic resection. Recently a laparoscopic procedure has been reported as an alternative
to laparotomy for selected colorectal lipomas and polyps
in general [2 4, 10].

References
1. Rogy M, Mirza D, Berlakovich G, Winkelbauer F, Rauhs R
(1991) Submucous large-bowel lipomas presentation and management. Eur J Surg 157:5157
2. Taylor B, Wolff B (1987) Colonic lipomas. Report of two unusual cases and review of the Mayo Clinic experience,
19761985. Dis Colon Rectum 30:888893
3. Vecchio R, Ferrara M, Mosca F, Ignoto A, Latteri F (1996)
Lipomas of the large bowel. Eur J Surg 162:915919
4. Michowitz M, Lazbuik N, Noy S, Lazbuik R (1985) Lipoma of
the colon. A report of 22 cases. Am Surg 51:494454
5. Castro E, Stearns M (1972) Lipoma of the large intestine: a review of 45 cases. Dis Colon Rectum 15:441444
6. Megibow AM, Bosniak MA, Horowitz L (1979) Diagnosis of
gastrointestinal lipomas by CT. AJR Am J Roentgenol 133:743
745
7. Younathan C, Ros P, Burton S (1991) MR imaging of colonic
lipoma. J Comput Assist Tomogr 15:492494
8. Rodriguez D, Drehner D, Beck D, McCauley C (1990) Colonic
lipoma as a source of massive hemorrhage. Report of a case. Dis
Colon Rectum 33:977979
9. Gordon RT, Beal JM (1978) Lipoma of the colon. Arch Surg
113:897899
10. Saclarides T, Ko S, Airan M, Dillon C, Franklin J (1991) Laparoscopic removal of a large colonic lipoma. Report of a case. Dis
Colon Rectum 34:10271029

Int J Colorect Dis (1998) 13: 312

Springer-Verlag 1998

REVIEW

S. E. Green D. M. Bradburn J. S. Varma J. Burn

Hereditary non-polyposis colorectal cancer

Accepted: 23 October 1997

Abstract Hereditary non-polyposis colorectal cancer

(HNPCC) is an autosomal dominant condition in which affected individuals develop colorectal cancer or extracolonic cancers, most commonly endometrial, at an early age.
Recent advances in molecular genetics have led to the identification and sequencing of four genes thought to be responsible for the majority of cases of hereditary non-polyposis colorectal cancer. A description of the disease along
with details of the underlying genetics and pathological
features are presented. Current management and screening
policies in these pedigrees are not clearly established. This
article discusses some of the controversies in the light of
predictive testing.
Key words HNPCC Colorectal cancer RER status
Lynch syndrome
Rsum Le cancer colo-rectal hrditaire non polypode
(HNPCC) est une pathologie autosomique dominante dans
laquelle les individus atteints dveloppent des cancers
colo-rectaux et des cancers extra-coliques, le plus souvent
de lendomtre, un ge prcoce. Des progrs rcents dans
la gntique molculaire ont permis didentifier et de
prciser la squence de quatre gnes dont on pense quils
sont responsables, pour la plupart des cas, des HNPCC.
Nous rapportons une description dtaille de cette affection et des troubles gntiques sous-jacents ainsi que des
constatations pathologiques. Le traitement courant et les
rgles de dpistage ne sont pas encore clairement tablis.
Cet article discute quelques-unes des controverses la
lumire des tests prdictifs.
S. E. Green () J. S. Varma
Department of Surgery, University of Newcastle Upon Tyne,
Framlington Place, Newcastle Upon Tyne, NE2 4HH, UK
J. Burn
Department of Human Genetics,
University of Newcastle Upon Tyne,
19/20 Claremont Place, Newcastle Upon Tyne, NE2 4AA, UK
D. M. Bradburn
Department of Surgery, Wansbeck General Hospital,
Ashington, UK

Introduction

Genetic factors have long been recognised to be important

in the development of colorectal cancer. Aldred Warthin in
the late 1800s first documented the family cancer
syndrome with his description of Warthins family G
[1]. This pedigree demonstrated a marked aggregation of
bowel, endometrial and ovarian cancers. The significance
of this description was not fully realised until the description of two large pedigrees by Lynch et al. in 1966 [2]. The
syndrome was later called the hereditary non-polyposis
colorectal cancer syndrome (HNPCC), or Lynch syndrome.
HNPCC is an autosomal dominant condition which is
clinically characterised by the development of colorectal
cancer at an early age (mean age 44 years), an excess of
synchronous and metachronous tumours and a preponderance of right-sided tumours (70%) [3].

Associated tumours
Another feature seen in many of these families is the occurrence of adenocarcinomas at other sites, in particular
the endometrium, ovary, stomach, pancreas, ureter, renal
pelvis and skin [3 5].
Mecklin and Jarvinen [6] looked at the tumour spectrum in 40 HNPCC families and found colorectal (64%),
endometrial (8%), gastric (6%), biliary/pancreatic (4%)
and uroepithelial carcinomas (2%) to be the most frequent.
Vasen et al. [7], in 24 kindreds in the Netherlands, and Watson and Lynch [8], from a study of 23 HNPCC families,
likewise found carcinoma of the endometrium to be the
second commonest malignancy in HNPCC. They also reported an excess of gastric, urinary tract, small bowel, ovarian and hepatobiliary cancers. An increased incidence of
breast cancer and lung cancer have been reported in female
family members [9], but others workers have found no excess of these cancers in HNPCC [8, 10, 11]. Indeed Lynch
et al. [10, 11] found significantly fewer lung cancers oc-

curring in HNPCC families than in control groups. It is obviously more difficult to demonstrate a positive link
between these two cancers and HNPCC because of the relatively high incidence of breast and bronchogenic carcinoma in the general population. Risenger et al. [12], however, described widespread microsatellite instability in several breast carcinomas from known HNPCC kindreds. Furthermore, one family member had a germline hMLH1 mutation along with the wild-type allele in her normal tissues
but was homozygous for the mutant allele in the breast cancer tissue, suggesting that breast cancer may occur as an integral tumour in HNPCC. Beck et al. [13] also found germline HNPCC mutations in individuals from HNPCC kindreds with breast cancer but felt this may simply be the result of chance. Other studies have reported an increased incidence of lymphatic/haemopoietic cancers and sarcomas
in HNPCC [14 16] but this has not been substantiated.
Clinical characteristics
Lynchs early description subdivided HNPCC into two clinical variants: Lynch syndrome I, which is the predisposition to site specific colorectal cancer, and Lynch syndrome
II, which is characterised by the development of extracolonic malignancies. Many authorities now believe that there
is no clear distinction between these two variants, and that
they are merely manifestations of a single syndrome. This
is supported by recent data showing families with extra-colonic cancers and those with site-specific HNPCC to be
linked to the same chromosomes [17, 18].
Unlike familial adenomatous polyposis (FAP), in which
colorectal cancer is preceded by the development of hundreds to thousands of adenomatous polyps, HNPCC has no
such phenotypic features. Evidence suggests colorectal
carcinomas in HNPCC do develop from adenomatous
polyps, but that adenomas do not occur in large numbers.
Jass [19] reviewed 131 cancers from 117 HNPCC family
members. None of the cancers was of the small, superficial, de novo type, but residual adenoma (contiguous with
cancer) was present in 100% in situ cancers, 89% of cancers involving only the submucosa, 29% of cancers limited to the muscle coat and 12% of cancers extending
beyond muscle, suggesting that the majority had arisen
from adenomatous polyps. In an earlier study Jass and
Stewart [20] compared the prevalence of colorectal adenomas in 23 patients with HNPCC to that in an age-matched
forensic autopsy population. Although the incidence of
adenomas was greater in the HNPCC patients than in the
autopsy group (particularly those under 50 years of age;
30% and 5%, respectively) all the adenomas were solitary,
apart from that in one patient who had two synchronous
adenomas. Mecklin et al. [21] evaluated the histopathology of colorectal cancers and adenomas in 75 HNPCC patients with colorectal cancer compared to control patients
with sporadic colorectal cancer. They found a similar incidence of synchronous adenomas in the two groups, 19%
in HNPCC patients compared to 16% in controls. In a colonoscopic screening program Love and Morrissey [22]

found adenomas in 7/42 (17%) asymptomatic HNPCC

family members, the majority solitary.
Extra-colonic phenotypic markers such as congenital
hypertrophy of the retinal pigment epithelium (CHRPE)
and osteomas have been reported in some presumed
HNPCC pedigrees [23, 24], but these may well represent
attenuated forms of FAP [25 28].
Diagnostic criteria
In view of the absence of phenotypic markers in HNPCC
the identification of probands to date has relied on the accurate documentation of family histories. Difficulties arose
in distinguishing familial clusterings of colorectal cancer
due to shared environmental influences from those with a
true genetic predisposition. This led the International Collaborative Group on HNPCC to put forward a set of minimum criteria to define HNPCC, the Amsterdam criteria
[29]. These state that: (a) At least three relatives should
have histologically verified colorectal cancer; one of them
should be a first-degree relative of the other two. Familial
polyposis should be excluded. (b) At least two generations
should be affected. (c) In one of the relatives colorectal
cancer should be diagnosed under 50 years of age.
It should be noted that these criteria were designed to
focus research resources effectively and are often too stringent for use with small families. Percesepe et al. [30] found
that family size plays an important role in determining the
outcome of pedigree assessment; the relative risk of reaching a positive diagnosis of HNPCC according to the Amsterdam criteria increases by 24% with each additional
first-degree relative. There is also the possibility of falsepositive diagnosis in large pedigrees that may contain
chance clusters of tumours. These criteria also ignore extra-colonic malignancies as a clinical manifestation of
HNPCC, which may lead to under diagnosis of the syndrome. Beck et al. [13] looked at ten families with pedigrees suggestive of HNPCC but in which the Amsterdam
criteria were not fulfilled. DNA was screened for germline
mutations in the hMLH1 and hMSH2 genes. Mutations
were identified in six of the ten families. They concluded
that the Amsterdam criteria are inappropriate for use in a
clinical setting.
Incidence
The true incidence of HNPCC is not yet known. It has been
reported to be responsible for from less than 1% to up to
13% of all colorectal carcinomas [31 34]. Ponz de Leon
et al. [35] estimated the frequency of HNPCC from their
regional registry, using the Amsterdam criteria, to be
between 3.4 4.5%. Ghadirian et al. [36] in Montreal found
5.1% of colorectal cancers in a case control study to have
a family history compatible with HNPCC compared with
0.6% of controls. Hall et al. [37] examining all individuals under 45 years of age presenting with colorectal cancer, found that 8% of those family histories which could

be obtained fulfilled the Amsterdam criteria, and that a further 12% satisfied less strict criteria.

Genetics

Most colorectal cancers develop from benign adenomatous

polyps, via the so-called adenoma-carcinoma sequence
proposed by Morson [38]. It is thought that carcinomas develop by a stepwise progression from normal mucosa
through an adenoma to invasive cancer [39]. Fearon and
Vogelstein [39] described a sequence of genetic changes
that accompany these histological events. Sequential and
cumulative genetic damage leads first to a hyperprofilerative state, thence to adenomas and subsequently to an invasive carcinoma. This involves mutational activation of
several known oncogenes, in particular K-ras, coupled
with the mutational deactivation of several known tumour
suppressor genes: APC (chromosome 5), p53 (chromosome 17) and DCC (chromosome 18). Although Fearon
and Vogelstein [39] described a preferred sequence of mutations, this is a simplified model, and it appears to be the
total accumulation of changes that is important rather than
their order. Nevertheless, loss of APC is probably an initiating event while p53 loss appears to occur late in most
tumours.
Tumour suppressor genes behave in a recessive manner, so that both alleles must be inactivated in order for the
growth suppressive function to be eliminated (Knudson hypothesis), i. e. two hits must occur [40]. Clonal expansion
then occurs if the mutation conveys a survival advantage
to the cell. If a germline mutation in one allele of a tumour
suppressor gene is inherited, only one hit is required. Individuals with FAP have a germline mutation in the
APC gene together with mutation of the wild type allele at
the somatic level.
Chromosome localisation
In the search for the HNPCC gene(s) it seemed likely that
the culprit would be a tumour suppressor gene. However,
it became apparent from studying the molecular genetics
of tumours in HNPCC that a different mechanism of tumourigenesis exists in these individuals. Peltomaki et al.
[41] were the first to show close linkage of HNPCC to
anonymous microsatellite markers on chromosome 2, suggesting that this is the most likely physical location of the
HNPCC gene. Microsatellite markers are short nucleotide
sequences which are repeated within the DNA strand. Subsequent work demonstrated a rather unexpected finding in
that there are differences in these microsatellite markers
between tumour DNA and normal DNA, the tumour DNA
showing very variable increases in the length of these sequences, probably the result of repeat expansion or deletion. These microsatellites are normally scattered throughout the human genome, and many normally exhibit variations in the number of repeat sequences between individ-

uals (length polymorphisms), thus explaining their use in

linkage analysis. In general, however, alleles at these sites
are inherently stable, and this variability in tumour tissue
was therefore unexpected. It was consequently postulated
that replication errors, uncorrected by repair mechanisms,
had occurred in these sequences during tumour development.
Several DNA editing and repair systems exist in all cells
to ensure accurate replication of genetic material, thus preventing propagation of somatic mutations. The DNA lesions that lead to mutations are most frequently modified,
missing or mismatched nucleotides. The best defined
mismatch repair pathway is the so-called MutHLS pathway in Escherichia coli. Subsequent work has identified
human homologues of two bacterial mismatch repair proteins, MutS [42, 43] and MutL [44, 45], which have been
called hMSH2 and hMLH1. hMSH2 maps to human chromosome 2p22-21, very close to the locus described by Peltomaki et al. [41]. hMLH1, on the other hand, is located
on chromosome 3p21-23. More recently two further homologues of the prokaryotic mutL gene, hPMS1 and
hPMS2, have been identified and sequenced [46]. They
have been localised to chromosomes 2q31-33 and 7p22,
respectively.
Gene interaction
Germline mutations in all four of these genes have been
identified in affected individuals from different HNPCC
families [42, 46, 47]. In order for tumourigenesis to occur
in these individuals a second mutation is probably required
to inactivate the normal wild-type allele since it appears
that adequate amounts of the gene product are produced to
be functional in the heterozygous state. Parsons et al. [48]
have shown that lymphoblastoid cells from an HNPCC patient, whose colorectal cancer revealed multiple replication errors, were repair proficient. Casares et al. [49], using somatic cell hybrids between a tumour cell line without genomic instability at simple repeated sequences and
colon carcinomas cell lines with somatic genomic instability, demonstrated that restoration of defective mismatch repair and microsatellite stability is achieved by transfer of
a wild-type chromosome, i. e. confirming the recessive nature of the mutator phenotype. Further evidence that a second, somatic mutation is required comes from Williams
et al. [50]. They showed, using the mPAS histochemical
technique to identify a polymorphism for O-acetyltransferase activity, loss of which is due to a somatic mutation,
that the somatic mutation frequency in the normal mucosa
of patients with HNPCC is no greater than that in patients
with sporadic colorectal cancer. They therefore concluded
that germline defects in HNPCC do not result in a generalised increase in liability to mutation in normal colonic
mucosa but that a second, somatic event is required.
It seems likely that mutation of a second allele must occur at an initial or very early stage in tumour development,
inactivating repair enzymes and thus allowing mutations
to accumulate rapidly, leading to the accelerated develop-

ment of cancer in these families. This hypothesis is supported by experimental data in which Lazar et al. [51] demonstrated the presence of somatic mutations in the APC and
p53 genes in RER+ tumours from two HNPCC families.
Only three tumours were analysed (one polyp and two carcinomas). All the mutations detected were absent from patient lymphocyte-extracted DNA and could therefore be
considered as somatic mutations. The mutational pattern,
however, was interesting. In the polyp one mutation was
found in the APC gene and no mutations in the p53 gene,
but in the carcinomas six mutations were found in the
APC gene and four in the p53 gene in one and four mutations in the APC gene and none in the p53 gene in the other.
This mutational pattern is strikingly different from that observed in sporadic colorectal tumours, which are characterised by single or, at most, two different mutations in each
gene. This supports the hypothesis that in HNPCC inactivation of a DNA repair gene results in the progressive accumulation of mutations in critical genes known to be involved in colorectal tumourigenesis.
More recently, however, Parsons et al. [52] have looked
again at the non-neoplastic cells of HNPCC patients for
mismatch repair defects using more sensitive methods.
They found evidence of microsatellite instability in nonneoplastic cells in a subset of HNPCC patients. They postulated that this is due to inherited mutations of other genes
that participate in mismatch repair, with multiple germline
mutations leading to a reduction of mismatch repair activity. An alternative hypothesis is that mismatch repair gene
mutations are acting in a dominant-negative fashion; the
product of the abnormal allele interferes with the function
of the normal protein.
HNPCC is thus an end result of defects in one or more
of several mismatch repair genes. hMSH2 and hMLH1 are
thought to account for 70 90% of all cases of HNPCC.
hPMS1 and hPMS2, along with possibly other as yet unidentified DNA mismatch repair genes, account for the remainder.

Pathology

Despite the name non-polyposis colorectal cancer evidence to date suggests that the carcinomas in HNPCC, as
discussed above, arise from benign adenomatous polyps
[20, 53, 54]. Although adenomatous polyps are no more
prevalent in individuals from HNPCC pedigrees than in
the general population [20, 22, 53, 55], they do appear to
show a greater propensity for malignant transformation
[20]. In keeping with this they are also more likely to have
a villous component and to show moderate or severe dysplasia.
Comparison with sporadic colorectal carcinoma
Colorectal carcinomas in HNPCC are more often found in
the proximal colon than sporadic carcinomas, and there is

a higher incidence of synchronous and metachronous tumours. They usually have a normal diploid constitution
rather than being aneuploid, are frequently mucinous and
poorly differentiated, with areas of necrosis [3, 55 58]. It
has been suggested that, paradoxically, stage for stage the
prognosis of HNPCC tumours is relatively good [20, 59,
60], though this may in part be a selection bias by focusing on large pedigrees with surviving gene carriers. One
theory suggested to account for a good prognosis in the
presence of pathologically poor prognostic features is that
the relatively high mutational load that occurs in tumours
with defective DNA repair systems is detrimental to tumour survival [60]. Another possibility is that tumours carrying a large number of mutations stimulate a stronger immune response [31].
The majority of colorectal cancers in HNPCC [>80%)
manifest replication errors and are thus termed replication
error positive (RER+) tumours. Some 10 16% of sporadic
colorectal cancers have also been found to be RER+ [54,
62 63]. Interestingly, RER+ sporadic cancers exhibit
many of the features of colorectal tumours in HNPCC mentioned above, when compared to RER sporadic tumours
[62, 63]. In one study 80% of sporadic RER+ tumours were
proximal to the splenic flexure [54]. Some of these tumours
may represent unrecognised HNPCC or may be due to a
de novo rather than inherited germline mutation. Alternatively somatic mutations in HNPCC genes may have occurred during tumourigenesis.
Cancers at other sites
The RER phenotype is not limited to colon cancers. Peltomaki et al. [64] studied the incidence of microsatellite instability in more than 500 sporadic tumours, representing
six different types of cancer. They found that 18% of gastric cancers and 22% of endometrial cancers were RER+
whereas all of the lung, breast and testicular cancers examined were RER. Importantly, the first two cancers, as
opposed to the latter three, are part of the HNPCC tumour
spectrum. Risenger et al. [12] found microsatellite instability in both sporadic and HNPCC endometrial cancers. Microsatellite instability has also been observed in keratoacanthomas [65], a distinctive skin tumour frequently seen
in Muir Torre syndrome, which is now known to be part of
the HNPCC syndrome.
The reason for the observed site specificity of tumours
in HNPCC is still unclear. It may simply be a reflection of
the frequent exposure of these tissues to specific mutagens. The colorectal, gastric, small bowel and urothelial mucosas are often in contact with potential carcinogens, for
example heterocylic amines, unlike breast and testicular
cancers. If this were the case, however, it is somewhat surprising lung cancer is not part of the HNPCC tumour spectrum. Alternatively, Peltomaki et al. [64] suggested that the
site specificity is due to differences in the vulnerability of
various genes to replication errors. If target genes contain
mutation prone sequences at critical sites in their structure,
tumourigenesis is more likely to occur.

Replication errors have also commonly been found in

adenomas derived from patients with HNPCC. Around
60% of such adenomas have been found to be RER+ compared to only 3% of sporadic adenomas [54]. These findings provide further strong evidence that adenomatous
polyps are precursors of colorectal cancer in HNPCC.
Studying adenomas for RER+ has been suggested as a tool
to help in the diagnosis of HNPCC, especially in families
in whom a mutation has not been identified.

Management

Identification of HNPCC
Family history
Probably the most difficult hurdle in the management of
HNPCC is the identification of gene carriers. This is due
partly to the absence of a recognisable phenotype. The clinician must therefore always be alert to the possibility of
an inherited predisposition. The initial step in identification should be a routine family history of all malignancy.
One major problem with reliance on family history is the
accuracy of patient recall. Studies have shown inaccuracies in patient recollections of relatives illnesses, both
over- and underreporting pathology. Other factors which
should alert the clinican are multiple colonic tumours, extra-colonic cancers, particularly those recognised as part
of the HNPCC tumour spectrum and tumours occurring at
an early age. As discussed above, strict adherence to the
Amsterdam criteria is not appropriate as they are too stringent. All families with a pedigree suggestive of HNPCC
should be identified and referred to a geneticist.
Experience with FAP [66] suggests that the tracing of
pedigrees, counselling and recruitment of individuals for
screening is best performed by a trained genetics nurse,
along with a clinical geneticist. These pedigrees are often
complex and extensive. Contacting a healthy individual to
inform them of a potential health risk must be carried out
with tact and without the time restriction of a busy clinic.
The management of HNPCC is still not uniform, and pedigree tracing is not yet a routine concept as it is with FAP.
It has been adequately demonstrated with FAP that a regional register identifies more individuals at risk [66] and
leads to a more uniform approach to management. Most
FAP registries are now documenting cases of HNPCC, but
the increased workload (four to five times the incidence of
FAP) suggests that establishing regional registers for
HNPCC should be a goal in the near future.
Predictive testing
The recent identification and cloning of four genes,
hMSH2, hMLH1, hPMS1 and hPMS2, provides the basis
for direct mutation analysis. It is therefore now possible to
offer predictive testing to some families, but at present the

search for mutations in the mismatch repair genes is time

consuming and expensive.
Predictive testing cannot be established without adequate back-up. Pre- and post-testing counselling must be
available. The counselling sessions should educate the
family about the clinical and management aspects of
HNPCC, the risks of cancer and the consequences of receiving a gene positive test result with the broader implications, for close relatives and insurance prospects, for example [67]. Likewise the consequences of a gene negative
test result must be discussed. Family members who do not
carry the gene can be reassured and discharged from further follow-up but may find it hard to stop regular screening if they have already been enrolled into a screening program. It is also important to emphasise that they still have
the same risk as the general population of developing colorectal cancer. Those who are gene carriers should be enrolled into a screening program, as outlined below, screening for extra-colonic as well as colonic malignancies. The
decision to proceed with any gene test should be freely
made by the at-risk person after having had time to carefully consider the consequences of genetic testing. Those
deemed to have a significant risk from pedigree analysis
who decline predictive testing should also probably be
maintained on a regular screening program.
RER status
Several workers have looked at using RER status of tumours combined with family history to refine the clinical
diagnosis of HNPCC and hence to select patients for mutation analysis. The reported incidence of RER+ tumours
in HNPCC varies from 77% [68] to 95% [69] but also occurs in 13 20% of sporadic colorectal cancers. Jass et al.
[70] suggested that reliance on the clinical criteria alone
results in overdiagnosis of HNPCC. They examined
50 families, 19 in whom the Amsterdam criteria were totally fulfilled (group A) and 31 in whom the criteria were
partially fulfilled (group B). A family was designated
RER+ if at least half the tumours tested showed microsatellite instability. In group A 12 families were RER+ and 7
RER; in group B 9 families were RER+ and 22 RER. The
accepted clinical and pathological characteristics of
HNPCC were found to cluster within the 12 group A RER+
families, suggesting that the Amsterdam criteria for
HNPCC could be refined by inclusion of RER status. Muta
et al. [71], however, suggested the Amsterdam criteria
under diagnose HNPCC. They studied 56 patients with
colorectal cancer, 8 of whom fulfilled the Amsterdam criteria and 23 of whom fulfilled modified clinical criteria,
which included extra-colonic cancers, multiple and proximal cancers. They found 86% of those fulfilling the Amsterdam criteria were RER+ but 62% of those fulfilling the
modified clinical criteria were also RER+. They concluded
that the presence of microsatellite instability, in concert
with modified clinical criteria, identifies legitimate cases
of HNPCC that might otherwise be excluded by the Amsterdam criteria.

Samowitz and Slattery [72], on the other hand, looked

at RER status as a marker of HNPCC in a general population study. In this context they found that RER status was
not a useful marker of family history and concluded that it
should not be considered as evidence for an inherited syndrome. Both Dunlop et al. [73] and Liu et al. [74] in general population studies likewise found RER analysis of tumours not sufficiently discriminatory in identifying those
with a family history, unless stratifying age groups. Dunlop et al. [73], extrapolating from their data, indicated that
58% of patients under 35 years of age with colorectal cancer have RER+ tumours, 24% of which have a germline
mismatch repair gene mutation. This compares with the
non-age-stratified group where 15% have RER+ tumours
and only 1% have germline mutations. Liu et al. [74] found
18 out of 31 (58%) patients aged under 35 years with colorectal cancer to be RER+, compared to 12% of those aged
over 35 years. Twelve of those below 35 years of age were
evaluated for mutations in the mismatch repair genes, and
five (42%) were found to harbour a germline mutation.
Screening gene carriers and individuals
with a significant clinical risk of HNPCC
Screening should be performed in as streamlined a fashion as possible to minimise impact on normal life. It is
essential to minimise anxiety if compliance is to be optimised, and counselling is very important from an early stage.
This is probably best provided in the setting of a family
cancer clinic, with a specialist nurse dedicated to working
with these families. Collaboration between the surgeon or
physician and the geneticist is crucial to good management.
Data from the United States have highlighted problems due
to both patient compliance and physician delay [75]. Education of both parties is essential and a registry should have
a central role in such activities.
Screening should not be directed merely to colon cancer but also to other tumours that commonly occur in
HNPCC in order to be effective, for example, uterus and
possibly stomach and ovarian tumours in families where
there is a preponderance of these malignancies.
Colon
Flexible sigmoidoscopy is not appropriate in these patients
because of the preponderance of right-sided lesions. Double-contrast barium enema has a similar error rate in the
detection of small polyps (<10 mm) as colonoscopy
(11.7% compared to up to 10% for colonoscopy) but gives
no opportunity to biopsy or remove polyps [76 78].
Screening for faecal occult blood is mentioned merely to
be dismissed as a sole technique for such high-risk cases
because of its low sensitivity [79 81]. About 40% of cancers and 80% of adenomas are missed by a single screen
with standard guaiac slide tests.
It is difficult to specify a precise age at which to initiate
screening, but it is logical to recommend starting at

25 years of age or 5 years earlier than the earliest onset of

colon cancer in the family [10, 29]. Likewise the age at
which to discontinue colonoscopy is not certain. Gene carriers generally develop colorectal cancer at ages 15 20
years younger than the general population, but some do
present with tumours in their 7th or 8th decade [82]. Indeed a study of 41 families with HNPCC showed that 8%
of affected individuals presented with symptomatic cancers at greater than 60 years of age [83]. Thus, although
discontinuance of screening has been recommended by
some at 60 years of age [29], colonoscopic screening
should probably continue for life, unless comorbidity dictates otherwise.
The optimal screening interval is not known. Unlike
sporadic adenomas, which take around 5 years to reach
1 cm in diameter and 8 10 years to become malignant,
adenomas in HNPCC have a greater potential for growth
[20]. Indeed interval cancers have been reported in several
screened groups at 3 years [29] although these may represent missed lesions rather than de novo growths.
The range of recommendations varies from yearly colonoscopy to an interval of 2 3 years [10, 22, 29, 55, 84].
Lynch et al. [11] have recommended biannual screening
from 25 to 35 and annual screening from 35 onwards. Patient acceptance of the screening protocol is vitally important for their own further compliance and that of relatives.
If colon preparation is good and the whole colon right
round to caecum is well visualised, 2-yearly colonoscopy
is probably adequate, but in those proven by molecular genetic analysis to carry the gene defect it may be more prudent to offer annual examination until the natural history
of HNPCC is more clearly defined.
Endometrium
Endometrial carcinoma is the most commonly observed
extra-colonic cancer in HNPCC [6 8, 16]. The lifetime
risk for female members of an HNPCC family of developing endometrial cancer has been calculated to be as
high as 20 40%, compared to 3% in the general population, and the period of highest risk occurs 15 years earlier than in the general population [70, 74]. Effective
treatment is available if endometrial cancer is detected
early, and therefore it is reasonable to include this as part
of a screening program. Lynch et al. [85] have recommended yearly uterine washings for endometrial cytology and transvaginal ultrasound, but it is not possible to
say yet whether the costs of screening outweigh the benefits.
Ovary
Ovarian carcinoma is one of the less common cancers associated with HNPCC, occurring in 3% of female cases
compared to a population incidence of 1 in 5000 (0.02%)
[6]. Transvaginal ultrasound is the mainstay of ovarian
screening, performed at yearly intervals. Lynch et al. [55]

recommended yearly colour flow Doppler transvaginal

ultrasound and CA-125 tumour markers. However, women
must appreciate the limitations of ovarian cancer screening.
Other cancers
In total, gastric, pancreatico-biliary and urothelial cancers
arise in less than 10% of affected individuals [6]. Therefore it is questionable whether gastroscopy, abdominal
ultrasound and urine cytology should be included routinely
in the screening protocol. Perhaps in families in whom
other cancers have been recorded it may be appropriate to
screen for these specifically.
Surgery
Although Lynch and coworkers [85, 86] have recommended prophylactic subtotal colectomy, total abdominal
hysterectomy and bilateral salpino-oophorectomy in identified gene carries, we do not know enough at present on
the disease penetrance and expression to confidently offer
such advice. Not all gene carriers eventually develop colorectal cancer or extra-colonic malignancy. Unlike FAP,
where the penetrance is approximately 100%, the penetrance in HNPCC has been estimated to be approximately
80%, and thus 20% of individuals carrying the mutation
do not develop colorectal cancer [87]. Vasen and Wijnen
[88] reported a lifetime risk of colorectal cancer greater
than 80%, but this was in well-defined HNPCC families.
Dunlop et al. [73], on the hand, used a population-based
strategy to calculate lifetime cancer risk associated with
germline DNA mismatch repair gene mutations, irrespective of family history. Index patients were identified from
the Scottish National Cancer Registry and were diagnosed
with colorectal cancer aged 35 years or less. Those with a
family history fulfilling the Amsterdam criteria were excluded. Of 27 patients 13 were RER+ (56%) and underwent
mutation analysis for germline mutations in the hMLH1
and hMSH2 genes. A germline mutation was found in 6
(46%). From these probands 156 relatives aged over
18 years were traced. Of these, 67 (43%) carried a germline mismatch repair gene mutation. The lifetime risk of
developing cancer was calculated from these individuals.
For all cancers this was 91% in men and 69% in women.
This difference was due largely to the significantly increased risk of colorectal cancer in men, 74% compared to
30% in women, by 70 years of age. In women the risk of
endometrial cancer was greater than colorectal cancer, 42%
by age 70. Cancer incidence increased rapidly from age 40,
but many patients destined to develop cancer did not do so
until a relatively elderly age.
Considering the high incidence of colorectal cancer in
males reported by Dunlop et al. [73] (74%) and the incidence of interval cancers reported on screening programs,
prophylactic colectomy should probably be offered to male
gene carriers.

One of the cardinal features of HNPCC is the occurrence of synchronous and metachronous colorectal tumours, and therefore segmental resection is not appropriate in gene carriers. The risk of metachronous cancers is
40% at 10 years if segmental resection is performed [89].
The procedure of choice then is total colectomy with ileorectal anastomosis (IRA), which removes the maximal area
at risk, consistent with avoidance of a stoma, yet has low
morbidity and mortality [3, 10, 90]. The rectum must be
screened sigmoidoscopically following this, and for those
individuals for whom this is unacceptable a restorative
proctocolectomy is an appropriate choice.
Several important factors must be addressed when offering a prophylactic total colectomy, including whether
or not the procedure eliminates the cancer risk, its morbidity and its timing [87]. The cancer risk, although less, still
exists in that the rectum remains in situ. The risk of developing rectal cancer is in the region of 12% 12 years following total colectomy and IRA, and regular surveillance
is therefore mandatory [87]. There is also still the risk of
extra-colonic malignancies developing. Lynch et al. [85]
suggest that female gene carriers should be encouraged to
have their families early so that they can consider the option of hysterectomy and bilateral salpingo-oophorectomy
between the ages of 35 and 40 years, but they must be aware
of the possibility of developing peritoneal cystadenocarcinoma, ovarian in origin, despite having their ovaries removed. The patient must understand that total colectomy
and IRA is a major procedure with significant morbidity,
reported to be in the region of 7.8 10% in FAP [91, 92].
The optimal timing of surgery is not known, and advice
can only be based on the available data. The mean age at
presentation with colorectal cancer is the middle 40s
[3, 83], but the range is wide (14 82 years), [85]. The risks
of elective surgery increase with increasing age, but early
surgery inflicts a not insignificiant operative procedure on
individuals who may not develop cancer until their 60s or
70s, if at all.
It has taken a long time from Warthins first observations on family G to the establishment of the genetic
basis of HNPCC. Recent developments have contributed
greatly to our understanding of HNPCC, but if we are
to reduce morbidity and mortality rates, medical practitioners must become more aware of the syndrome, and
we must look towards the establishment of regional registers and management protocols for such families. Wider
availability of mutation analysis, as techniques improve,
will help clarify the natural history of the disease and
greatly assist in the management of these families in the
future.

References
1. Warthin AS (1913) Hereditary with reference to carcinoma. Arch
Intern Med 12: 546555
2. Lynch HT, Shaw MW, Magnuson CW, Larsen AL, Krush AJ
(1966) Hereditary factors in two large midwestern kindreds.
Arch Intern Med 117: 206 212

10
3. Lynch HT, Lanspa SJ, Boman BM, Smyrk T, Watson P, Lynch JF,
Lynch PM, Cristofaro G, Buto P, Tauro AV, Mingazzini P, DiGiulio E (1988) Hereditary nonpolyposis colorectal cancer Lynch
syndromes I and II. Gastroenterol Glin North Am 17: 679712
4. Lynch HT, Fusaro RM, Roberts L, Voorhees GJ, Lynch JF (1985)
Muir-Torre syndrome in several members of a family with a variant of the Cancer family syndrome. Br J Demeratol 113: 295301
5. Lynch HT, Lynch ML, Pester J, Fusaro RM (1981) The cancer
family syndrome. Arch Intern Med 141: 607611
6. Mecklin JP, Jarvinen HJ (1991) Tumour spectrum in cancer family syndrome (hereditary non-polyposis colorectal cancer). Cancer 68: 11091112
7. Vasen HFA, Offerhaus GJA, Den Hartog Jager FCA, Menko FH,
Nagengast FM, Griffioen G, van Hogezand RB, Heintz APM
(1990) The tumour spectrum in hereditary non-polyposis colorectal cancer: a study of 24 kindreds in the Netherlands. Int J
Cancer 46: 3134
8. Watson P, Lynch HT (1993) Extracolonic cancer in hereditary
non-polyposis colorectal cancer. Cancer 71: 677685
9. Itoh H, Houlston RS, Harocopos C, Slack J (1990) Risk of cancer death in first degree relatives of patients with hereditary nonpolyposis cancer syndrome (Lynch type II): a study of 130 kindreds in the United Kingdom. Br J Surg 77: 13671370
10. Lynch HT, Smyrk T, Watson P, Lanspa SJ, Boman BM, Lynch
PM, Lynch JF, Cavalieri J (1991) Hereditary colorectal cancer.
Semin Oncol 18(4): 337366
11. Lynch HT, Watson P, Smyrk TC, Lanspa SJ, Boman BM, Boland
CR, Lynch JF, Cavalieri RJ, Leppert M, White R, Sidransky D,
Vogelstein B (1992) Colon cancer genetics. Cancer 70: 13001312
12. Risinger JI, Barrett JC, Watson P, Lynch HT, Boyd J (1996) Molecular genetic evidence of the occurrence of breast cancer as an
integral tumour in patients with the hereditary nonpolyposis
colorectal carcinoma syndrome. Cancer 77: 1836 1843
13. Beck NE, Tomlinson PM, Homfray T, Hodgson SV, Harocopos
CJ, Bodmer WF (1997) Genetic testing is important in families
with a history suggestive of hereditary non-polyposis colorectal cancer even if the Amsterdam criteria are not fulfilled. Br J
Surg 84: 233 237
14. Law IP, Hollinshead AC, Whang-Peng J, Dean JH, Oldham RK,
Herberman RB (1977) Familial occurrences of colon and uterine carcinoma and of lymphoproliferative malignancies, II.
Cancer 39: 12291236
15. Love RR (1985) Small bowel cancers, B-cell leukaemia, and six
primary cancers with metastases and prolonged survival in the
cancer family syndrome of Lynch. Cancer 55: 499502
16. Lynch HT, Krush AJ (1971) Cancer family G revisited:
1895 1970. Cancer 27: 15051511
17. Hall NR, Murday VA, Chapman P, Williams MAT, Burn J, Finan PJ, Bishop DT (1994) Genetic linkage in Muir-Torre syndrome to the same chromosomal region as cancer family syndrome. Eur J Cancer 30: 180182
18. Hall NR, Williams MAT, Murday VA, Newton JA, Bishop DT
(1994) Muir-Torre syndrome: a variant of the cancer family syndrome. J Med Genet 31: 627631
19. Jass J (1995) Colorectal adenomas in surgical specimens from
subjects with hereditary non-polyposis colorectal cancer. Histopathology 27: 263 267
20. Jass JR, Stewart SM (1992) Evolution of hereditary non-polyposis colorectal cancer. Gut 33: 783 786
21. Mecklin J-P, Sipponen P, Jarvinen HJ (1986) Histopathology of
colorectal carcinomas and adenomas in cancer family syndrome.
Dis Colon Rectum 29: 849 853
22. Love RR, Morrissey JF (1984) Colonoscopy in asymptomatic
individuals with a family history of colorectal cancer. Arch
Intern Med 144: 22092211
23. Burt RW, Bishop DT, Lynch HT, Rozen P, Winawer SJ (1990)
Risk surveillance of individuals with heritable factors for colorectal cancer. Bull World Health Organ 68: 655665
24. Houlston RS, Fallon T, Harocopos C, Williams CB, Davey C,
Slack J (1992) Congenital hypertrophy of retinal pigment epithelium in patients with colonic polyps associated with cancer
family syndrome. Clin Genet 42: 1618

25. Leppert M, Burt R, Hughes JP, Samowitz W, Nakamura Y, Woodward S (1990) Genetic analysis of an inherited predisposition to
colon cancer in a family with a variable number of adenomatous
polyps. N Engl J Med 322: 904908
26. Spirio L, Olschwang S, Groden J, Robertson M, Samowitz W,
Joslyn G, Gelbert L, Thliveris A, Carlson M, Otterud B, Lynch
H, Watson P, Laurent-Puig P, Burt R, Hughes JP, Thomas G, Leppert M, White R (1993) Alleles of the APC gene: an attenuated
form of familial polyposis. Cell 75: 951957
27. Spirio L, Otterud B, Stauffer D, Lynch H, Lynch P, Watson P,
Lanspa S, Smyrk T, Cavalieri J, Howard L, Burt R, White R,
Leppert M (1992) Linkage of a variant or attenuated form of adenomatous polyposis coli to the adenomatous polyposis coli
(APC) locus. Am J Hum Genet 51: 92100
28. Evans DGR, Guy SP, Thakker N, Armstrong JG, Dodd C, Davies
DR, Babbs C, Clancy T, Warnes T, Sloan P, Taylor TV, Harris R
(1993) Non-penetrance and late appearance of polyps in families
with familial adenomatous polyposis. Gut 34: 13891393
29. Vasen HF, Mecklin JP, Watson P, Utsunomiya J, Bertario L,
Lynch P, Svendsen LB, Cristofaro G, Muller H, Khan PM, Lynch
HT (1993) Surveillance in hereditary nonpolyposis colorectal
cancer; an international cooperative study of 165 families. Dis
Colon Rectum 36: 14
30. Percesepe A, Anti M, Roncucci L, Armelao F, Marra G, Pahor
M, Coco C, Gasbarrini G, Ponz de Leon M (1995) The effect of
family size on estimates of the frequency of hereditary non-polyposis colorectal cancer. Br J Cancer 72: 1320 1323
31. Bodmer W, Bishop T, Karran P (1994) Genetic steps in colorectal cancer. Nat Genet 6: 217219
32. Mecklin JP (1987) Frequency of hereditary colorectal cancer.
Gastroenterology 93: 1021 1025
33. Marx J (1993) New colon cancer gene discovered. Science
260: 751752
34. Lynch HT (1986) Frequency of hereditary non-polyposis colorectal cancer. Gastroenterology 90: 486 492
35. Ponz de Leon M, Sassatelli R, Benatti P, Roncucci L (1993) Identification of hereditary nonpolyposis colorectal cancer in the
general population. Cancer 71: 34933501
36. Ghadirian P, Cadotte M, Lacroix A, Baillargeon J, Perret C (1993)
Colon cancer in seven siblings. Eur J Cancer 29A: 15531556
37. Hall NR, Finan PJ, Ward B, Turner G, Bishop DT (1994) Genetic susceptibility to colorectal cancer in patients under 45 years
of age. Br J Surg 81: 14851489
38. Morson BC (1974) Evolution of cancer of the colon and rectum.
Cancer 34: 845849
39. Fearon ER, Vogelstein B (1990) A genetic model for colorectal
tumorigenesis. Cell 61: 759767
40. Knudson AB (1985) Hereditary cancer, oncogenes and anti-oncogenes. Cancer Res 45: 1437 1443
41. Peltomaki P, Aaltonen LA, Sistonen P, Pylkkanen L, Mecklin
JP, Jarvinen H, Green JS, Jass JR, Weber JL, Leach FS, Petersen GM, Hamilton SR, de la Chapelle A, Vogelstein B (1993)
Genetic mapping of a locus predisposing to human colorectal
Cancer. Science 260: 810812
42. Leach FS, Nicolaides NC, Papadopoulos N, Liu B, Jen J, Parsons
R, Peltomaki P, Sistonen P, Aaltonen LA, Nystrom-Lahti M, Guan
X-Y, Zhang J, Meltzer PS, Yu J-W, Kao F-T, Chen DJ, Cerosaletti KM, Fournier REK, Todd S, Lewis T, Leach RJ, Naylor SJ, Weissenbach J, Mecklin J-P, Jarvinen H, Petersen GM, Hamilton SR,
Green J, Jass J, Watson P, Lynch HT, Trent JM, de la Chapelle A,
Kinzler KW, Vogelstein B (1993) Mutations of a mhutS Homolog
in hereditary nonpolyposis colorectal cancer. Cell 75: 12151225
43. Fishel R, Lescoe MK, Rao MRS, Copeland NG, Jenkins NA,
Garber J, Kane M, Kolodner R (1993) The human mutator gene
homology MSH2 and its association with hereditary nonpolyposis colon cancer. Cell 75: 10271038
44. Bronner CE, Baker SM, Morrison PT, Warren G, Smith LG,
Lescoe MK, Kane M, Earabino C, Lipford J, Lindblom A, Tannergard P, Bollag RJ, Godwin AR, Ward DC, Nordenskjold M,
Fishel R, Kolodner R, Liskay M (1994) Mutation in the DNA
mismatch repair gene homologue hMLH2 is associated with
hereditary non-polyposis colorectal cancer. Nature 368: 258261

11
45. Lindblom A, Tannergard P, Werelius B, Nordenskjold M (1993)
Genetic mapping of a second locus predisposing to hereditary
non-polyposis colorectal cancer. Nat Genet 5: 279282
46. Nicolaides NC, Papadopoulos N, Liu B, Wei Y-F, Carter KC,
Ruben SM, Rosen CA, Haseltine WA, Fleishmann RD, Fraser
CM, Adams MD, Venter JC, Dunlop MG, Hamilton SR, Peterson GM, de la Chapelle A, Vogelstein B, Kinzler KW (1994)
Mutations in two PMS homoloques in hereditary nonpolyposis
colon cancer. Nature 371: 7580
47. Papadopoulos N, Nicolaides N, Wei Y-F, Ruben SM, Carter KC,
Rosen CA, Haseltine WA, Fleischmann RD, Fraser CM, Adams
MD, Venter JC, Hamilton SR, Peterson GM, Watson P, Lynch
HT, Peltomaki P, Mecklin J-P, de la Chapelle A, Kinzler KW,
Vogelstein B (1994) Mutation of mutL homolog in hereditary
colon cancer. Science 263: 16251628
48. Parsons R, Li G-M, Longley MJ, Fang W, Papadopoulos N, Jen
J, de la Chapelle A, Kinzler KW, Vogelstein B, Modrich P (1993)
Hypermutability and mismatch repair deficiency in RER+ tumour cells. Cell 75: 12271236
49. Casares S, Ionov Y, Ge H-Y, Stanbridge E, Perucho M (1995)
The microsatellite mutator phenotype of colon cancer cells is often recessive. Oncogene 11: 2303 2310
50. Williams GT, Geraghty JM, Campbell F, Appleton MAC, Williams ED (1995) Normal colonic mucosa in hereditary non-polyposis colorectal cancer shows no generalised increase in somatic mutation. Br J Cancer 71: 1077 1080
51. Lazar V, Grandjouan B, Bognel C, Couturier D, Rougier P,
Bellet D, Bressac-de Paillerets B (1994) Accumulation of
multiple mutations in tumour suppressor genes during colorectal tumourigenesis in HNPCC patients. Hum Mol Genet
3: 2257 2260
52. Parsons R, Li G-M, Longley M, Modrich P, Liu B, Berk T, Hamilton SR, Kinzler KW, Vogelstein B (1995) Mismatch repair deficiency in phenotypically normal human cells. Science
268: 738 740
53. Lanspa SJ, Smyrk TC, Lynch HT (1990) The colonoscopist and
the Lynch syndromes. Gastrointest Endosc 36(2): 156158
54. Aaltonen LA, Peltomaki P, Mecklin JP, Jarvinen H, Jass JR, Gren
JS, Lynch HT, Watson P, Tallqvist G, Juhola M, Sistonen P, Hamilton SR, Kinzler KW, Vogelstein B, de la Chapelle A (1994)
Replication errors in benign and malignant tumours from hereditary nonpolyposis colorectal cancer patients. Cancer Res
54: 16451648
55. Lynch HT, Smyrk TC, Watson P, Lanspa SJ, Lynch JF, Lynch
PM, Cavalieri RJ, Boland CR (1993) Genetics, natural history,
tumour spectrum, and pathology of hereditary nonpolyposis
colorectal cancer: an updated review. Gastroenterology
104: 15351549
56. Kouri M, Laasonen A, Mecklin J-P, Jarvinen H, Franssila K, Pyrhonen S (1990) Diploid predominance in hereditary nonpolyposis colorectal carcinoma evaluated by flow cytometry. Cancer 65: 1825 1829
57. Frei JV (1992) Hereditary nonpolyposis colorectal cancer
(Lynch syndrome 11): diploid malignancies with prolonged survival. Cancer 69: 1108 1111
58. Mecklin J-P, Sipponen P, Jarvinen HJ (1986) Histopathology of
colorectal carcinomas and adenomas in cancer family syndrome.
Dis Colon Rect 29: 849 853
59. Lynch HT, Lynch PM, Albano WA, Lynch JF (1981) The cancer syndrome: a status report. Dis Colon Rectum 24: 311322
60. Albano WA, Recabaren JA, Lynch HT (1982) Natural history of
hereditary cancer of the breast and colon. Cancer 50: 360363
61. Shibata D, Peinado MA, Ionov Y, Malkhosyan M, Perucho M
(1994) Genomic instability in repeated sequences is an early somatic event in colorectal tumorigenesis that persists after transformation. Nat Genet 6: 273281
62. Peltomaki P, Lothe RA, Aaltonen LA, Pylkkanen L, NystromLahti M, Seruca R, David L, Holm R, Ryberg D, Haugen A,
Brogger A, Borresen A-L, de la Chapelle A (1993) Microsatellite instability is associated with tumours that characterize the
hereditary non-polyposis colorectal carcinoma syndrome. Cancer Res 53: 58535855

63. Kim H, Jen J, Vogelstein B, Hamilton SR (1994) Clinical and

pathological characteristics of sporadic colorectal carcinomas
with DNA replication errors in microsatellite sequences. Am J
Pathol 145: 148156
64. Ionov Y, Peinado MA, Malkhosyan S, Shibata D, Perucho M
(1993) Ubiquitous somatic mutations in simple repeated sequences reveal a new mechanism for colonic carcinogenesis. Nature 363: 558561
65. Halling KC, Honchel R, Pittelkow MR, Thibodeau SN (1995)
Microsatellite instability in keratoacanthoma. Cancer 76:
1765 1771
66. Rhodes M, Chapman PD, Burn J, Gunn A (1991) Role of a regional register for familial adenomatous polyposis: experience
of the northern region. Br J Surg 78: 451542
67. Petersen GM (1996) Genetic counselling and predictive testing
for colorectal cancer risk. Int J Cancer 69: 53 54
68. Aaltonen LA, Peltomaki P, Leach FS, Sistonen P, Pylkkanen L,
Mecklin J-P, Jarvinen H, Powell SM, Jen J, Hamilton SR, Petersen GM, Kinzler KW, Vogelstein B, de la Chapelle A (1993)
Clues to the pathogenesis of familial colorectal cancer. Science
260: 812 816
69. Konishi M, Kikuchi-Yanoshita (1996) Molecular nature of colon tumours in hereditary non-polyposis colorectal cancer, familial polyposis and sporadic colon cancer. Gastroenterology
111: 307 317
70. Jass JR, Cottier DS, Jeevaratnam P, Pokos V, Holdaway KM,
Bowden ML, Van de Water NS, Browett PJ (1996) Diagnostic
use of microsatellite instability in hereditary non-polyposis
colorectal cancer. Lancet 346: 1200 1201
71. Muta H, Noguchi M, Perucho M, Ushio K, Sugihara K, Ochiai
A, Nawata H, Hirohashi S (1996) Clinical implications of microsatellite instability in colorectal cancers. Cancer 77: 265 270
72. Samowitz WS, Slattery ML (1995) Microsatellite instability in
human colonic cancer is not a useful indicator of familial colorectal cancer. Gastroenterology 109: 1765 1771
73. Dunlop MG, Farrington SM, Carothers AD, Whyllie AH, Sharp
L, Burn J, Liu B, Kinzler KW, Vogelstein B (1997) Cancer risk
associated with germline DNA mismatch repair gene mutations.
Hum Mol Genet 6: 105 110
74. Liu B, Farrington SM, Peterson GM, Hamilton SR, Parsons R,
Papadopoulos N, Fujiwara T, Jen J, Kinzler KW, Wyllie AH, Vogelstein B, Dunlop MG (1995) Genetic instability occurs in the
majority of young patients with colorectal cancer. Nat Med
1: 348 352
75. Lynch HT, Smyrk TC, Lanspa SJ, Jenkins JX, Cavalieri J, Lynch
JF (1993) Cancer control problems in the Lynch syndromes. Dis
Colon Rectum 36: 254260
76. Thoeni RF, Menuck L (1977) Comparison of barium enema and
colonoscopy in the detection of small colonic polyps. Diagn Radiol 124: 631635
77. Fork FT (1981) Double contrast enema and colonoscopy in polyp
detection. Gut 22: 971977
78. Williams CB, Hunt RH, Loose H, Riddell RH, Sakai Y, Swarbrick ET (1974) Colonoscopy in the management of colonic
polyps. Br J Surg 61: 673682
79. Dunlop MG (1992) Screening for large bowel neoplasms in individuals with a family history of colorectal cancer. Br J Surg
79: 488494
80. Rozen P, Fireman Z, Ron E, Lynch HT (1986) Family history of
colorectal cancer as a marker of potential malignancy within a
screening program (abstract). Gastroenterology 90: 1608
81. Houlston RS, Murday V, Harocopos C, Williams CB, Slack J
(1990) Screening and genetic counselling for relatives of patients
with colorectal cancer in a family cancer clinic. BMJ 301:
366368
82. Vasen HFA, den Hartog Jager FCA, Menko FH (1989) Screening for hereditary nonpolyposis colorectal cancer: A study of
22 kindreds in the Netherlands. Am J Med 86: 278 281
83. Vasen HFA, Taal BG, Griffioen G, Nagengast FM, Cats A, Menko FH, Oskam W, Kleibeuker JH, Offerhaus GJA, Khan PM
(1994) Clinical heterogeneity of familial colorectal cancer and
its influence on screening protocols. Gut 35: 12621266

12
84. Fleischer DE, Goldberg SB, Browning TH (1989) Detection and
surveillance of colorectal cancer. JAMA 261: 580 585
85. Lynch HT, Smyrk T, Lynch JF (1996) Overview of natural history, pathology, molecular genetics and management of HNPCC
(Lynch syndrome). Int J Cancer 69: 38 43
86. Lynch HT (1996) Is there a role of prophylactic subtotal colectomy among hereditary non-polyposis colorectal cancer germline mutation carriers? Dis Colon Rectum 39: 109 110
87. Rodriguez-Bigas MA (1996) Prophylactic colectomy for gene
carriers in hereditary non-polyposis colorectal cancer. Has the
time come? Cancer 78: 199 201
88. Vasen HFA, Wijnen JT (1996) Cancer risk in families with hereditary non-polyposis colorectal cancer diagnosed by mutation
analysis. Gastroenterology 110: 1020 1027

89. Fante R, Roncucci L, Di Gregorio C, Tamassia MG, Losi L, Benatti P, Pedroni M, Percesepe A, De Pietri S, Ponz de Leon M
(1996) Frequency and clinical features of multiple tumours of the
large bowel in the general population and in patients with hereditary nonpolyposis colorectal carcinoma. Cancer 77: 2013 2021
90. Mecklin J-P, Jarvinen H (1993) Treatment and follow-up strategies in hereditary non-polyposis colorectal carcinoma. Dis Colon Rectum 36: 927929
91. Herrera L (1991) Extracolonic manifestations of FAP. Oncology 5: 31 33
92. Nugent KP, Northover J (1994) Total colectomy and ileorectal
anastomosis. In: Phillips RKS, Spigelman AD, Thompson JPS
(eds) Familial adenomatous polyposis, 1st edn. Arnold, London,
79 91

Int J Colorect Dis (1998) 13: 1316

Springer-Verlag 1998

O R I G I N A L A RT I C L E

A. Mellgren A. Lpez I. Schultz B. Anzn

Rectocele is associated with paradoxical anal sphincter reaction*

Accepted: 15 August 1997

Abstract Rectocele is a frequent finding in constipated

patients. However, constipation is not always relieved by
rectocele repair, which may be due to other overlooked reasons for constipation. The study was designed to investigate patients with rectocele, in order to elucidate concomitant colorectal disorders and their association with rectocele. One hundred and twelve female patients suffering
from severe constipation and rectal emptying difficulties
were investigated using defecography, electrophysiology,
anorectal manometry and colon transit time. Fifty-six patients with rectocele demonstrated by defecography were
compared with 56 patients without rectocele, but with other
abnormal findings at defecography. The frequency of paradoxical anal sphincter reaction (PSR) was higher in patients with rectocele (60%) than in patients without rectocele (24%). The present study supports an association
between rectocele and PSR. We suggest that constipated
patients with a rectocele should be investigated thoroughly
before rectocele repair is considered. Further studies on the
effect of biofeedback training in patients with rectocele and
PSR are indicated.
Key words Rectocele Constipation Paradoxical
and sphincter reaction
Rsum La mise en vidence dune rectocle est frquente chez des patientes constipes. Le traitement chirurgical de la rectocle ne gurit toutefois pas dans tous les
* Read in part at the XVIth Biennial Congress of the International
Society of Colon and Rectal Surgeons, Lisboa, April 14 to 18, 1996
A. Mellgren ()1 I. Schultz
Department of Surgery, Danderyd Hospital,
S-182 88 Stockholm, Sweden
A. Lpez B. Anzn
Department of Gynecology and Obstetrics,
Karolinska Institutet at Danderyd Hospital,
Stockholm, Sweden
Present address:
University of Minnesota, Division of Colon and Rectal Surgery,
2550 University Avenue West #313N, St. Paul, Minnesota
55114-1084, USA

cas la constipation qui peut tre due dautres raisons mconnues. Ltude prsente a t entreprise pour tudier chez
les patientes porteuses dune rectocle sil existe des lsions colo-rectales concomittantes et dans quelle mesure elles associes la rectocle. Cent-douze femmes souffrant
de constipation svre et de troubles de lvacuation rectale ont t investigues laide de dfcographies, de mesures lectrophysiologiques, de manomtries ano-rectales
et de dterminations du temps de transit colique. Cinquante-six patientes chez lesquelles la dfcographie a mis
en vidence une rectocle ont t compares avec 56 patientes sans rectocle mais porteuses dautres anomalies
la dfcographie. La frquence dune raction paradoxale
du sphincter anal est plus souvent observe chez des patientes porteuses dune rectocle (60%) que chez les patientes sans rectocle (24%). Cette tude supporte lide
dune association entre la rectocle et la raction paradoxale du sphincter anal. Nous suggrons que des patientes
constipes porteuses dune rectocle doivent faire lobjet
dinvestigations compltes avant que lon envisage la correction chirurgicale de la rectocle. Dautres tudes sur
lefficacit du biofeedback en cas de rectocle et de ractions paradoxales du sphincter anal devraient tre entreprises.

Patients with rectocele may present with a variety of symptoms, thereby necessitating consultation with both gynecologists and surgeons. A rectocele may be isolated or part
of a complete genital prolapse.
Symptoms of rectocele include a bearing-down sensation, incomplete rectal emptying, a sensation of rectal pressure and sometimes vaginal symptoms from the herniation
itself [1, 2]. The role of a rectocele in constipation is controversial. At defecography about 25% of patients with defecation disorders have a rectocele, but many women with
rectocele are not constipated [3].
A transvaginal approach for rectocele repair is advocated by several gynecologists [2, 4, 5]. Using this approach gynecologic symptoms are often improved. How-

14

ever, if the patient is constipated preoperatively, the outcome

is sometimes less favorable. In 1967 Marks [6] reported that
correction of the vaginal deformity alone did not provide
sufficient relief of constipation. Concomitant surgery of anorectal pathology and repair of rectocele by endorectal approaches have therefore been proposed [1, 6 11].
Arnold et al. [12] however, found no difference in symptomatic outcome regarding constipation after rectocele repair by transvaginal and endorectal approaches. They suggested that patients should be investigated preoperatively
by transit studies, anal manometry, and defecography to
identify constipation due to slow colonic transit or outlet
obstruction caused by reasons other than a rectocele.
The aim of the present study was to determine the degree to which patients with rectocele, have concomitant
colorectal disorders and to assess any such association with
rectocele.

Patients and methods

Patients
From 1987 to 1991 inclusive 178 female patients prospectively
underwent defecography owing to severe constipation and rectal
emptying difficulty. One hundred and twelve patients (63%) had radiologic abnormalities rectal intussusception, rectal prolapse, rectocele and or enterocele including. These patients were divided into
two groups based on the presence or absence of a rectocele.
Group 1 (Rc): This included 56 patients with rectocele (Table 1). The
mean age was 50 years (range 19 81 years). Five patients had previously undergone hysterectomy and 32 were postmenopausal.
Group 2 (No Rc): In this group there were 56 patients without rectocele who had other abnormalities on defecography (Table 1). The
mean age was 53 years (range 29 75 years). Six patients had previously undergone hysterectomy and 30 were postmenopausal.
Defecography
A modification of the technique described by Brodn and Snellman
[13] was used. The patient had a barium meal 1.5 hours prior to the
examination. When this reached the small bowel, thick constrast medium with a consistency similar to feces was injected into the rectum.
Table 1 Abnormal radiologic findings on defecography in 112
female-patients with rectal emptying difficulties
Defecographic findings
Rectal intussusception alone (RI)
Rectal prolapse alone (RP)
Rectocele alone (Rc)
Enterocele alone (Ec)
RI + Rc
RI + Ec
RP + Rc
RP + Ec
Rc + Ec
RI + Rc + Ec
RP + Rc + Ec
Total

No of patients
20
3
36
0
14
22
0
11
3
2
1
112

(18%)
(3%)
(32%)
(0%)
(12%)
(20%)
(0%)
(10%)
(3%)
(2%)
(1%)

A viscous contrast medium was simultaneously instilled into the vagina. The patient was seated on a commode placed and exposed to a
fluoroscopic unit. Iron plates were used for contrast leveling. Left lateral views of the pelvis were recorded during fluoroscopy by video.
A rectocele was diagnosed when the anterior rectal and posterior
vaginal wall herniated into the lumen of the vagina an by how much?
Rectal intussusception was defined as a circumferential descent
of the entire thickness of the rectal wall, which might extend into the
anal canal but not through the anal verge. Rectal prolapse was defined
as a circumferential descent of the entire thickness of the rectal wall
seen coming out through the anus. An enterocele was diagnosed when
small bowel was present between the vagina and rectum.
Electrophysiology
Electrophysiologic analysis provides information on the function of
the pelvic floor musculature and its innervation. Electrophysiologic
assessment was carried out according to the method described by
Swash et al. [14, 15]. Conventional needle electromyography (EMG)
was recorded in the external anal sphincter (EAS) bilaterally and in
the puborectalis muscle. EMG was considered pathologic indicating
a peripheral nerve lesion if: 1) the activity during maximal voluntary contraction (squeezing) was reduced to such an extent that only single discharges of motor unit potentials were recorded instead
of a normal interference pattern, or 2) a moderately reduced interference pattern contained a considerable number of polyphasic motor unit potentials of high amplitude (> 2 mV).
A paradoxical anal sphincter reaction (PSR) was present if: 1)
maximal straining increased the on-going EMG activity, or 2) maximal straining did not decrease the on-going EMG-activity and no
closing reflex was seen after completed straining. Fibre density (FD)
was measured in the EAS by single fibre EMG recordings at 20 different locations on each side. Normal values for FD were derived
from the literature [14, 16, 17] and the normal limits (mean 2 SD)
for different age groups were 1.52 (< 30 years), 1.82 (30 65 years),
1.88 (66 70 years) and 2.20 (71 85 years). Pudendal nerve terminal motor latency (PNTML) was determined on both sides using a
special electrode (Dantec St. Marks Pudendal Electrode 13L40). The
upper limit of normal was 2.5 ms.
Anorectal manometry
The procedure has been previously described by Holmstrm et al.
[18]. With this method maximal anal resting pressure (MRP) less
than 50 mm Hg, maximal anal squeeze pressure (MSP) less than
65 mm Hg and maximal tolerable volume (MTV) less than 150 ml
or more than 400 ml were considered pathologic. Rectal sensibility was considered abnormal if the patient had no sensation of rectal filling following insufflation of the rectal balloon with 150 ml
of air.
Colon transit time
Colon transit time was estimated according to a modification of the
methods of Hinton et al. [19] and Keighley and Shouler [20]. The
patients ingested a capsule containing markers with a meal and five
days later a plain x-ray of the abdomen was taken. The number of
markers in the colon was estimated. A residue of more than 40% in
the colon was considered pathologic.
Statistical methods
The analyses were performed by the Department of Medical Information Processing, Karolinska Institutet, Stockholm.
Chi-squared analysis was used when comparing defecographic
and electrophysiologic findings, frequency of constipation, previous

15
hysterectomy and delayed colon transit time in patients with and
without rectocele.
Normal distribution of the data was checked and Students unpaired t-test was used when comparing MRP and MSP.

Results

The frequency of previous hysterectomy did not significantly differ between patients with (9%) or without rectocele (11%).

Table 2 Pelvic floor electromyographic parameters in constipated

female-patients with and without rectocele

PSRa
Pathologic PNTMLb
Both EMG and FDc
indicate peripheral
nerve damage
a
b
c

Electrophysiology
The frequency of PSR was higher in patients with rectocele (60%) than in patients without (24%) (Table 2). There
were no statistical differences in the frequencies of pathologic PNTML and pathologic findings indicating peripheral neuropathy on EMG or FD between the two groups of
patients (Table 2).
Defecography
Thirty-six percent (20/56) of patients with rectocele had
additional radiological abnormalities on defecography
(Table 1). The frequencies of enterocele, rectal intussusception, and rectal prolapse were lower in patients with
rectocele than in patients without (Table 3).

Patients
with
rectocele

Patients
without
rectocele

60% (26/43)
16% (5/31)

24% (10/41)
32% (8/25)

P<0.001
N.S.

17% (6/35)

29% (9/31)

N.S.

Paradoxical and sphincter reaction

Pudendal nerve terminal motor latency
Fibre density

Table 3 The incidence of rectal intussusception, rectal prolapse,

and enterocele in constipated female-patients with and without rectocele

Rectal intussusception
Rectal prolapse
Enterocele

Patients
with
rectocele
(n = 56)

Patients
without
rectocele
(n = 56)

29%
2%
11%

75%
25%
59%

Table 4 Anal sphincter pressures in patients with and without rectocele

Anorectal manometry
MRP was higher in patients with rectocele than in patients
without (P < 0.01) (Table 4). There was no statistical difference of MSP between patients with and without rectocele (Table 4).
Colon transit time
The proportion of patients with delayed colon transit time
did not differ significantly between the two groups. Eleven
(30%) of the 31 patients with rectocele had delayed colon
transit time compared with 13 (32%) of 41 patients without.

Discussion

Patients with rectocele often complain of constipation [3, 9]

as the present study has also demonstrated. The vector force
created by the valsalva manoeuvre is partially dissipated
through the rectovaginal septum [9], and as a result the patients must strain harder to defecate. It is, however, often
difficult to ascertain whether the rectocele is responsible
for symptoms or whether there might be another cause.
Physiologic evaluation of patients with rectocele has there-

P<0.001
P<0.001
P<0.001

MRP (mm Hg) (mean SD)

MSP (mm Hg) (mean SD)

Patients
with
rectocele
(n = 56)

Patients
without
rectocele
(n = 56)

72 19
72 26

61 23
64 26

P < 0.01
N.S.

fore been suggested [12, 21, 22], before considering surgical repair.
We have previously reported [23] that paradoxical
sphincter reaction (PSR) is frequently found in patients with
rectocele. The present study demonstrates an association
between rectocele and PSR and this association may indicate that PSR is a causative factor in the formation of the
rectocele. Straining and emptying efforts against a contracted pelvic floor may facilitate development of rectocele.
This association might also be one of the reasons for suboptimal results of surgical repair.
Different approaches to the treatment of PSR have been
proposed [24 27]. The best results are reported using EMG
feedback [28 30], aiming to teach patients to relax the pelvic floor muscles during straining. The present study indicates that some patients with rectocele might benefit from
biofeedback.
The present study does not support an association
between rectocele and pathologic reduced colon transit although this was abnormal in a high frequency of the patients. It may nevertheless be useful to include colonic transit studies in the preoperative assessment since these pa-

16

tients might have a less favorable outcome after rectocele

repair [22].
Defecography is necessary in the evaluation of patients
to confirm the clinical diagnosis and visualize concurrent
abnormalities.
Surprisingly rectocele was negatively associated with
enterocele in the present study. Nichols [31] reported rectocele and enterocele often to be found together on gynecologic examination, but this could not be verified in the
present study. This might be explained by the reported difficulty of distinguishing enterocele from rectocele on gynecologic examination [32, 33] or that rectocele and enterocele both being due to weakness of the rectovaginal septum may compete for the same anatomic space.
Surgical repair of rectocele is not always satisfactory.
Surgical technique is important [22] and preoperative and
radiologic assessment will help to exclude patients whose
constipation is due to reasons other than rectocele.

Conclusions

The present study supports an association between rectocele and paradoxical anal sphincter reaction. Constipated
patients with rectocele should be investigated thoroughly
before surgical repair is considered. Defecography will confirm the clinical diagnosis and might show concurrent abnormalities. Electrophysiologic assessment might demonstrate PSR and transit studies will help to identity patients
with a less favorable outcome after surgical repair.
Further prospective studies on the effect of biofeedback
in patients with rectocele and PSR are indicated.
Acknowledgements We are indebted to Elisabet Berg, Department
of Medical Information Processing, Karolinska Institutet, Stockholm
for help with the statistical analyses.

References
1. Sehapayak S (1985) Transrectal repair of rectocele: an extended armamentarium of colorectal surgeons. A report of 355 cases. Dis Colon Rectum 28: 422433
2. Nichols DH, Randall CL (1989) Vaginal surgery. 3rd edn. Williams & Wilkins, Baltimore
3. Nichols DH (1991) Surgery for pelciv floor disorders. Surg Clin
N Am 71: 927946
4. Marek CB (1969) Transverse repair for rectocele. South Med J
62: 749752
5. ster S, Astrup A (1981) A new vaginal operation for recurrent
and large rectocele using dermis transplant. Acta Obstet Gynecol Scand 60: 493495
6. Marks MM (1967) The rectal side of the rectocele. Dis Colon
Rectum 10: 387388
7. Pitchford CA (1967) Rectocele: a cause of anorectal pathologic changes in women. Dis Colon Rectum 10: 464466
8. Sullivan ES, Leaverton GH, Hardwick CE (1968) Transrectal
perineal repair: an adjunct to improved function after anorectal
surgery. Dis Colon Rectum 11: 106114
9. Capps WF (1975) Rectoplasty and perineoplasty for the symptomatic rectocele: a report of fifty cases. Dis Colon Rectum
18: 237244

10. Khubchandani IT, Sheets JA, Stasik JJ, Hakki AR (1983) Endorectal repair of rectocele. Dis Colon Rectum 26: 792796
11. Block IR (1986) Transrectal repair of rectocele using obliterative suture. Dis Colon Rectum 29: 707711
12. Arnold MW, Stewart WR, Aguilar PS (1990) Rectocele repair.
Four years experience. Dis Colon Rectum 33: 684687
13. Brodn B, Snellman B (1968) Procidentia of the rectum studied
with cineradiography: a contribution to the discussion of causative mechanism. Dis Colon Rectum 11: 330347
14. Neill ME, Swash M (1980) Increased motor unit fibre density
in the external and sphincter muscle in ano-rectal incontinence:
a single fibre EMG study. J Neurol Neurosurg Psychiatry 43:
343347
15. Kiff ES, Swash M (1984) Slowed conduction in the pudendal
nerves in idiopathic (neurogenic) faecal incontinence. Br J Surg
71: 614616
16. Percy JP, Neill ME, Kandiah TK, Swash M (1982) A neurogenic factor in faecal incontinence in the elderly. Age Ageing
11: 175179
17. Henry MM, Snooks SJ, Barnes PRH, Swash M (1985) Investigation of disorders of the anorectum and colon. Ann R Coll Surg
Engl 67: 355360
18. Holmstrm B, Brodn G, Dolk A, Frenckner B (1986) Increased
anal resting pressure following the Ripstein operation: a contribution to continence? Dis Colon Rectum 29: 485487
19. Hinton JM, Lennard-Jones JE, Young AC (1969) A new method for studying gut transit time using radioopaque markers. Gut
10: 842847
20. Keighley MRB, Shouler PJ (1984) Abnormalities of colonic
function in patients with rectal prolapse and faecal incontinence.
Br J Surg 71: 892895
21. Siproudhis L, Dautrme S, Robert A, Bretagne J, Heresbach D,
Raoul J, Gosselin M (1993) Dyschezia and rectocele a marriage of convenience? Physiologic evaluation of the rectocele in
a group of 52 women complaining of difficulty in evacuation.
Dis Colon Rectum 36: 10301036
22. Mellgren A, Anzn B, Nilsson B-Y, Johansson C, Dolk A, Gillgren P, Bremmer S, Holmstrm B (1995) Results of rectocele
repair, a prospective study. Dis Colon Rectum 38: 713
23. Johansson C, Nilsson B-Y, Holmstrm B, Dolk A, Mellgren A
(1992) Association between rectocele and paradoxical sphincter response. Dis Colon Rectum 35: 503509
24. Wasserman IF (1964) Puborectalis syndrome (rectal stenosis due
to anorectal spasm) Dis Colon Rectum 7: 8798
25. Keighley MRB, Shouler P (1984) Outlet syndrome: is there a
surgical option? J R Soc Med 77: 559563
26. Kamm MA, Hawley PR, Lennard-Jones JE (1988) Lateral division of the puborectalis muscle in the management of severe
constipation. Br J Surg 75: 661663
27. Hallan RI, Melling J, Womack NR, Williams NS, Waldron DJ,
Morrison JFB (1988) Treatment of anismus in intractable constipation with botulinum A toxin. Lancet II: 714717
28. Lestr B, Penninckx F, Kerremans R (1991) Biofeedback defaecation training for anismus. Int J Colorectal Dis 6: 202207
29. Kawimbe BM, Papachrysostomou M, Binnie NR, Clare N,
Smith AN (1991) Outlet obstruction constipation (anismus)
managed by biofeedback. Gut 32: 11751179
30. Wexner SD, Cheape JD, Jorge JMN, Heymen S, Jagelman DG
(1992) Prospective assessment of biofeedback for the treatment
of paradoxical puborectalis contraction. Dis Colon Rectum
35: 145150
31. Nichols D (1972) Types of enterocele and principles underlying choice of operation for repair. Am J Obstet Gynecol
40: 257263
32. Kelvin FM, Maglinte DDT, Hornback JA, Benson JT (1992) Pelvic prolapse: assessment with evacuation proctography (defecography). Radiology 184: 547551
33. Mellgren A (1994) Enterocele, rectocele and constipation,
Stockholm

Int J Colorect Dis (1998) 13: 1720

Springer-Verlag 1998

O R I G I N A L A RT I C L E

R. Mibu Y. Oohata M. Ishikawa

M. Sakai M. Tanaka

Terminal ileal transposition procedure in ileoanal anastomosis

following proctocolectomy

Accepted: 18 July 1997

Abstract We introduced a terminal ileal transposition

procedure (TITP) in ileal pouch-anal anastomosis, in
which a 50 to 70 cm isolated ileal segment 20 to 40 cm
from the ileocecal valve was interposed between the terminal ileum and the anus. Twelve patients underwent this
procedure in two or three-staged operations. Mean stool
frequency per 24 hours was 4.41.7, and stool consistency
was formed and soft in all patients at the mean of 13 months
after TITP. We observed neither surgical technique-related
complications nor metabolic disorders, except for iron deficiency anemia, during and after the operations. The serum
level of vitamin B12 significantly increased after the operation in eight patients (P<0.05). TITP has advantages
such as preventing the terminal ileum from metabolic dysfunction due to pouchitis, avoiding sacrifice of the terminal ileum in the two-staged operation, and obviating the
need for reconstruction of ileostomy in the three-staged operation. It may also promote intestinal absorption and reduce late metabolic complications.
Key words Terminal ileal transposition procedure
Anastomosis Metabolic complications
Rsum Nous avons introduit une technique de transposition de lilon terminal (TITP) lors de la confection dune
anastomose ilo-anale dans laquelle un segment ilal de 50
70 cm est isol une distance de 20 40 cm de la valve
ilo-caecale et est interpos entre lilon terminal et lanus.
Douze patients ont subi une telle intervention au cours de
deux ou trois temps opratoires. La frquence moyenne des
exonrations par 24 heures est de 4,4+1,7 et la consistance
des selles est moule et molle chez tous les patients en moyenne 13 semaines aprs la TITP. Nous navons pas observ
de complications dues la technique chirurgicale ni de dsordre mtabolique lexception dune anmie ferriprive
au cours et aprs des interventions chirurgicales. Le niveau
de vitamine B12 tait significativement lev aprs
R. Mibu () Y. Oohata M. Ishikawa M. Sakai M. Tanaka
Department of Surgery I, Kyushu University Faculty of Medicine,
Maidashi 3-1-1, Fukuoka 812-82, Japan

lopration chez 8 patients. La technique de la TITP a des

avantages tels que la prvention du dveloppement dune
pouchite par troubles mtaboliques de lilon terminal,
dpargner le sacrifice de lilon terminal dans la technique en deux temps opratoires et dviter la ncessit dune
reconstruction dune ilostomie dans la technique en trois
temps opratoires. La TITP permet aussi damliorer labsorption intestinale et de rduire les complications mtaboliques tardives.

Introduction

Proctocolectomy and ileoanal anastomosis have become

an established option for patients with ulcerative colitis or
familial adenomatous polyposis (FAP). In this procedure,
the introduction of various ileal pouches increased the neorectal reservoir capacity, resulting in a decrease of bowel
movements. However, since this procedure consists of a
pouch using the terminal ileum, it risks impairing terminal
ileal functions such as active absorption of bile acids and
vitamin B12 whenever acute or chronic severe inflammation occurs in the pouch. In our series, gallstones developed in two of three patients with the ileoanal anastomosis [1]. Vitamin B12 absorption is also affected in not less
than one-third of patients who have undergone this procedure [2]. To avoid using the terminal ileum for a pouch, we
introduced a terminal-ileal transposition procedure (TITP)
in which an isolated ileal segment was interposed between
the terminal ileum and the anus. The first use of this procedure was reported by Stringel [3] in 1985, who anastomosed an isolated jejunal loop to the anus without creating a pouch. We also reported the physiological and histological changes in TITP using an animal model [4 6].
Recently, Chu et al. [7] showed that the ileojejunal
transposition procedure, in which the distal ileum was
transported into the proximal, middle, or distal portions of
the remaining small intestine, produced an increase of mucosal weight and DNA content of the transposed ileum,
and proliferation of the intestinal mucosa. Hotokezaka

18

et al. [8] also demonstrated in an animal model that the

small intestine, except for a pouch, had higher absorptive
capacity for sodium, chloride and bile acids in TITP than
that in the conventional ileoanal anastomosis. All these reports suggest that TITP has an advantage over the conventional procedure in intestinal adaptation following proctocolectomy.
In the present study, we report our surgical methods and
preliminary clinical results of TITP.

Patients and methods

Patients
Between October 1988 and July 1995, twelve patients (eight women,
four men) of median age 38 (range 14 66) years underwent TITP
following proctocolectomy by the two or three-staged operation. Nine
patients had ulcerative colitis and three had FAP. Surgery was indicated for patients with acute fulminant colitis, steroid related side effects, neoplasia, and for patients with diffuse rectal adenomas in FAP.
Methods
Eight patients who underwent an elective two-staged procedure fasted, and blood samples for serum vitamin B12 level were collected
preoperatively and at a mean of 10.2 (range 7 14) and 28.7 (range
19 52) months after TITP. The serum level of vitamin B12
was measured by a commercial laboratory (SRL, Tokyo, Japan)
using competitive protein binding assay (normal value g 247 to
938 pmol/l). The hemoglobin level was also measured before, 6 and
12 months after surgery.
Statistical analysis
All data are expressed as meanstandard deviation. Wilcoxon signedrank test was used for statistical analysis. A P value of <0.05 was
taken as significant.

Technique
Two-staged procedure: This procedure was indicated for patients
having chronic colitis without high dose long term administration of
steroids. In the first stage, nearly all of the colon was excised leaving the caecum and about 5 cm of ascending colon. A rectal mucosectomy was performed to just above the dentate line, leaving about
a 5 cm rectal muscle cuff (Fig. 1). The right colic and ileocolic vessels were preserved in order to assure an adequate blood supply to
the terminal ileum. A 50 to 70 cm ileal segment was isolated,
20 to 40 cm from the ileocecal valve, with a vascular pedicle and a
J-shaped pouch 15 cm long at its distal portion was constructed. Mesenteric lengthening was performed by division of the vascular arcades and transverse incision of the peritoneum in a J pouch. The
proximal end of the isolated ileal segment was closed and attached
to the tenia of the cecum. An appendectomy was usually performed.
The apex of the J pouch was anastomosed to the anus. Finally, an
end ascending colostomy was constructed in the right lower quadrant. In the second stage, more than three months later, the opening
made by the ascending colostomy was closed, the retained colon was
excised to complete the proctocolectomy, and the terminal ileum was
anastomosed to the proximal end of the interposed ileal segment to
maintain intestinal continuity (Fig. 1 D).
Three-staged procedure: This procedure was indicated for patients either receiving high doses of steroids for severe colitis or in
association with a anal fistula. As first, a subtotal colectomy was performed, followed by a Brooke type end ileostomy and mucous fistula of the sigmoid colon (Figure 1 B). Each patient with an anal fis-

Fig. 1 A D Counterclockwise direction shows the two-staged operation and clockwise direction the three-staged operation. A The
first stage of the two-staged operation. Ascending colostomy and isolated ileal J pouch-anal anastomosis are performed following subtotal colectomy. The proximal end of the isolated ileum is closed.
B The first stage of the three-staged operation. A Brooke type ileostomy and mucous fistula of the sigmoid colon are constructed. C The
second stage of the three-staged operation. An isolated ileum with a
J pouch is anastomosed to the dentate line after excision of the retained colon and rectum. D Completion of ileoanal anastomosis. Finally, the terminal ileum is transposed above the isolated ileal segment
tula underwent a subtotal colectomy, followed by a fistulectomy
combined with an ascending colostomy and mucous fistula of the
sigmoid colon. For those patients receiving high dose steroids, one
month after cessation of steroids, the retained colon and rectal mucosa was excised, and an isolated ileal segment anastomosed to the
dentate line with creation of J pouch (Fig. 1 C). The proximal end of
the isolated ileal segment was closed and fixed to the abdominal wall
near the ileostomy. More than three months after the second operation, the opening made by the ileostomy or ascending colostomy was
closed and a proctocolectomy completed followed by an end to end
anastomosis between the terminal ileum and the proximal end of the
interposed ileal J pouch (Fig. 1 D).

Results

Of twelve patients, eight underwent the two-staged procedure and four the three-staged procedure. One of patients
who underwent the three-staged procedure had an anal fistula with chronic ulcerative colitis. The mean length of the
transposed ileum was 29 (range 20 40) cm, whilst the
mean length of the isolated ileal segment was 60 (range
50 70) cm. These lengths were determined by the branching of the mesenteric vessels.
Early postoperative complications consisted of compartment syndrome of the lower extremity in one patient

19

and wound infection in another. We observed no other early

complications such as intestinal obstruction, intestinal
ischaemia, acute pouchitis or stoma-related complications.
The mean duration of follow-up was 44 (range 13 82)
months. The mean stool frequency per 24 hours was
8.93.0 (range 4.0 14.0) one month after TITP and decreased over time (6 months, 4.82.0; 13 months, 4.41.7).
We occasionally found seepage at night in five patients one
month after TITP. Three of five patients needed pads at
night but had no perineal dermatitis at 13 months after surgery. As for stool consistency, all patients had formed and
soft faeces at approximately 13 months after TITP.
The haemoglobin level was not affected by TITP (before, 10.71.9 g/dl; 6 months, 11.91.5 g/dl; 13 months,
12.22.3 g/dl). In three of the twelve patients however, a
low hemoglobin level (less than 10 g/dl) due to iron deficiency was found before and after TITP and alleviated by
oral iron supplementation. No other late complications
such as intestinal obstruction, pouchitis, cholelithiasis, or
urolithiasis were observed during the period of follow-up.
In eight patients undergoing an elective procedure operated, the serum level of vitamin B12 significantly increased
after TITP (before, 434.4146.6 pmol/l; 10 months,
713.8365.9 pmol/l, P = 0.02 vs before; 29 months;
945.7501.9 pmol/l, P = 0.04 vs before). The patients evaluated the result of the operation by the assessment of stool
frequency, soiling frequency, daily life and social life. Ten
patients were completely satisfied, and two patients were
fairly satisfied.

Discussion

Our terminal-ileal transposition procedure resulted in no

serious early or late complications relating to the surgical
technique or metabolic function because the design of this
procedure is practical and protects the terminal ileum from
physiological dysfunction due to pouchitis, and from being sacrificed.
With respect to an isolated intestinal segment, Stringel
[3] described a 60 to 70 cm isolated jejunal loop without
documenting its location in relation to the Treitzs ligament
or ileocecal valve, while we used a 50 to 70 cm ileal segment 20 to 40 cm from the ileocecal valve. Chu et al. [7]
demonstrated that the more proximally the terminal
ileum was transposed, the greater the proliferation effect
of the intestinal mucosa. However, in these cases the distal site of the isolated ileal portion needed to be anastomosed to the anus, thus the transposed site or length of the
terminal ileum was limited. We could not determine the
ideal site or length of the transposed loop. Our technique
would be practical for the transposition of the terminal ileum because it does not require sacrifice of the small intestine and main arterial branches.
We used an ascending colostomy in the two-staged operation to minimise the segment of the small intestine to
be sacrificed when the stoma was closed. Although the ascending colon is not available for colostomy when it was

severely affected, the disease in the cecum and ascending

colon is usually mild in chronic ulcerative colitis and sparse
in FAP. Therefore, in most cases, the ascending colon can
be used for colostomy for several months until completion
of restorative proctocolectomy in the two-staged procedure. On the other hand, ielostomy is a safe and logical
procedure for patients with the fulminant type of ulcerative colitis. Ileostomy is also indicated for patients with associated carcinoma of the right-side colon. Another advantage of our procedure is that it obviates the need to reestablish the ileostomy in the second operation of the threestaged procedure. Moreover, the end stoma allows for easy
management, since it is associated with few stoma-related
complications.
We adopted TITP in order to preserve the functions of
the terminal ileum by avoiding inflammation of the pouch.
In addition, we suspect that this procedure may promote
intestinal adaptation, as recent studies suggest [7, 8].
MKoma [2] reported in conventional ileoanal anastomosis that the deficiency of vitamin B12 began within the first
six months of pouch function and substitutional therapy
with vitamin B12 was necessary. However, our preliminary series showed that TITP enhanced the serum vitamin
B12 level ten months after TITP. The effect of TITP was
also beneficial in terms of stool consistency and general
condition after surgery. The postoperative frequency of
bowel movements and seepage seemed to be comparable
with that of the conventional procedure. Although we did
not observe any instances of pouchitis in the current series, conclusions about pouchitis in TITP must await a future study. Even if pouchitis occurs and the pouch is excised, the terminal ileum can be saved in TITP. The ileal
pouch-anal anastomosis has a tendency toward intestinal
malabsorption of bile acids or vitamin B12, which may result in cholelithiasis or macrocytic anemia [9]. Cholelithiasis or macrocytic anemia did not occur during the followup period in this series. It seems that these diseases can be
prevented by preserving the terminal ileum. We need further experience performing TITP and long-term follow-up
to verify the advantages of this procedure in terms of intestinal adaptation and prevention of later complications
of ileoanal anastomosis, compared with conventional
ileoanal anastomosis.
In conclusion, TITP has important advantages over the
conventional ileoanal anastomosis using the terminal ileal
pouch. It may prevent the terminal ileum from metabolic
dysfunction due to pouchitis, it minimizes the loss of the
terminal ileum in the two-staged procedure, and it obviates the need for reconstruction of ileostomy in the threestaged procedure. TITP may also enhance intestinal absorption following proctocolectomy and reduce metabolic
complications.

References
1. Mibu R, Makino I, Chijiiwa K (1995) Gallstones and their composition in patients with ileoanal anastomosis. J Gastroenterol
30: 413 415

20
2. MKoma AE (1994) Follow-up results of hematology data before and after restorative proctocolectomy (clinical outcome). Dis
Colon Rectum 37: 932 937
3. Stringel G (1985) Pull-through with isolated jejunal loop for ulcerative colitis. J Pediatr Surg 20: 661 663
4. Mibu R, Itoh H, Nakayama F (1987) Effect of total colectomy
and mucosal proctectomy on intestinal absorptive capacity in
dogs. Dis Colon Rectum 30: 47 51
5. Nakahara S, Itoh H, Mibu R, Ikeda S, Nakayama F (1988) Regional difference in intestinal adaptation after total colectomy as
judged by the changes of mucosal Na-K ATPase, cyclic AMP,
and transmural potential difference. Dis Colon Rectum
31: 523 528

6. Hotokezaka M, Nakahara S, Nakamura K, Mibu R (1994) Morphology following proctocolectomy in dogs: effect of introduction of a neocolon using an interposed jejunal segment. Eur Surg
Res 26: 179 186
7. Chu KU, Tsuchiya T, Ishizuka J, Uchida T, Townsend CM,
Thompson JC (1995) Trophic response of gut and pancreas after
ileojejunal transposition. Ann Surg 221: 249 256
8. Hotokezaka M, Nakahara S, Iwamoto T, Chijiiwa K, Mibu R
(1996) Effect of terminal ileal transposition on intestinal absorption following proctocolectomy. Br J Surg 83: 486 492
9. Santos MC, Thompson JS (1993) Late complications of the ileal
pouchanal anastomosis. Am J Gastroenterol 88: 3 10

Int J Colorect Dis (1998) 13: 2126

Springer-Verlag 1998

O R I G I N A L A RT I C L E

L. Marzio F. A. Ciccaglione M. Falcucci

M. G. Malatesta L. Grossi N. Travaglini S. Guerri

Relationship between anal canal diameter

and pressure evaluated simultaneously by endosonography
and manometry in normal human subjects
Accepted: 27 June 1997

Abstract The study investigated the relationship between anal canal size and anal canal pressure measured simultaneously by anal endosonography and an electronic
pressure probe. Twelve normal subjects were studied. Anal
endosonography was performed using a 7.5-Mhz rotating
transducer of 2 5 cm focal length covered with a sylastic
balloon filled with degassed water (25 ml). During anal endosonography an electronic manometric probe was passed
along the side of the probe and positioned in the anal canal. The ultrasonic image was frozen when maximal anal
pressure was seen at basal, squeeze, and minimal pressure
during straining. An image was also obtained at maximal
anal relaxation after rectal distension with a balloon filled
with 150 ml air. The results showed that anal canal pressure was significantly and linearly correlated with anal canal diameter (P<0.001).
Key words Anal canal Pressure Endosonography
Sphincter Squeezing Straining
Rsum Cette tude analyse les relations entre la dimension et la pression du canal anal mesure simultanment
par chographie endo-anale et enregistrement lectronique
des pressions. Douze sujets normaux ont t tudis.
Lendo-sonographie anale a t ralise au moyen dune
sonde de 7,5 Mhz rotative avec une longuer focale da 2
5 cm recouverte dun ballon de silastic rempli de 25 ml
deau dgazee. Durant lchographie endo-anale, un capteur lectronique de pression a t pass le long de la sonde
et positionn dans le canal anal. Limage chographique a
t fige lorsque la pression anale maximale tait obenue
au repos et leffort de contraction et lorsque la pression
a t minimale durant lexpulsion. Une image a t obtenue
L. Marzio () F. A. Ciccaglione M. Falcucci
M. G. Malatesta L. Grossi N. Travaglini
Unit di Medicina Interna and Gastroenterologia,
G. dAnnunzio University, Casa di Cura Pierangeli,
Largo L. Pierangeli 1, I-65100 Pescara, Italy
S. Guerri
Urology Unit, Casa di Cura Pierangeli,
Largo L. Pierangeli 1, I-65100 Pescara, Italy

galement en phase maximale de relaxation aprs distension rectale par un ballonnet rempli de 150 ml dair. Ces
rsultats montrent que la pression du canal anal est corrle
de manire significative et linaire avec le diamtre du canal anal (P<0.001).

Introduction

Anal endosonography demonstrates the anatomy of the

internal and external anal sphincter [1, 2] and permits the
diagnosis of sphincter muscle defects [3, 4]. Other studies
have shown a correlation between internal and sphincter
thickness and resting pressure of the sphincter [5]. There
is, however, no information on resting anal pressure at different diameters of the anal canal.
The present study aimed at determining whether a relationship exists between anal canal diameter and anal canal
resting pressure. The former was measured by ultrasound
and the latter by an electronic pressure probe. Recordings
were made at rest, on squeezing and straining, and during
rectal distension. The study was carried out in normal subjects.

Materials and methods

Anal endosonography was performed using a 7.5-Mhz rotating transducer of 2 5 cm focal length (Bruel & Kjaer, Oslo, Norway).
During anal endosonography an electronic manometric probe
(3 mm external diameter) was passed alongside the endosonographic probe and positioned in the anal canal. The probe was made by
three electronic pressure sensors 3 mm apart arranged at 120 around
the circumference of the probe. The pressure sensor was attached to
a portable unit (Gastroscan II, Solothurn, Switzerland) with a 4-MB
RAM memory. Pressure recordings were made at the precise location of the frozen ultrasonographic image.
The ultrasonographic probe was covered by a latex condom and
filled with 25 ml degassed air. This created an acoustic window
between the probe and the anal canal wall, enabling the anal canal
diameter to be measured (Fig. 1). Due to the elastic property of the

22
Fig. 1 Anal manometry and
endosonography simultaneously taken in basal state in a normal subject. The endosonographic image was taken at
open air (A, arrow) and when a
steady anal canal pressure was
observed at manometry (B, arrow). Anal canal diameters
(A, B) were calculated as indicated in the endosonographic
picture. The endosonographic
probe was covered with a latex
balloon filled with 25 ml of water

latex and the water within it, any change in anal canal size at rest and
during straining, squeezing, and relaxation in response to rectal balloon distension could be easily observed and measured on the ultrasonographic screen.
The ultrasonographic and mounted pressure probes were introduced into the rectum, and were slowly withdrawn until the maximal
anal canal pressure was recorded. After stabilization of the pressure
at that level the subject was asked to squeeze and strain alternatively on three consecutive occasions at an interval of 3 min. The endosonographic image was frozen at the maximal basal resting pressure
(Fig. 1), at maximum squeeze pressure (Fig. 2), at minimal pressure
on straining (Fig. 3), and at minimal and canal pressure induced by
rectal distention (Fig. 4). Finally, an image was taken after withdrawing the probe from the anal canal to calculate the diameter of the water-filled cover at atmospheric pressure (Fig. 1). On ultrasonography
the anal canal assumed an ellipsoid shape and the anteroposterior and
transverse diameters were recorded (mm; Fig. 1). Data were computed by one of the authors (F. C.) who was unaware of the pressure

measured. The mean anterioposterior and transverse diameters was

calculated. Anal canal pressure during basal, squeezing, straining, and
150 ml rectal distension was measured in a separate session in each
subject without the ultrasonographic probe. Four sets of recording
were made in the 12 subjects including one each for pressure at rest,
during squeeze and straining and following rectal distension.
Data were analyzed using Students t test. The correlation coefficient of linear regression was used in determining the relationship
between anal diameter and pressures.

Results

Mean anal canal pressure and diameter during squeezing,

straining, and at 150 ml rectal distension were significantly
different from that measured at rest (Table 1). The data

23
Fig. 2 Anal manometry and
endosonography in a normal
subject. The endosonographic
pictures are taken during stable
anal canal pressure (A, arrow)
and at maximal squeezing pressure (B, arrow)

showed that anal canal pressure was inversely correlated

with anal canal diameter (P<0.001, Fig. 5). Age was not
correlated with either anal canal diameter or pressure at
rest (r = 0.014914, NS). Anal canal pressure was measured
with or without the endosonographic probe inserted were
similar (Table 2).

Discussion

This study showed that anal canal diameter as measured

by anal endosonography is inversely correlated with anal

Table 1 Pressure and diameter of the anal canal measured by means

of endosonography and manometry (n = 12)
Normal subjects

Basal
Squeezing
Straining
Rectal distention
* P<0.05 vs. basal

Pressure
(mBar)

Diameter
(mm)

8718
14512 *
3810 *
239 *

195
133 *
297 *
268 *

24
Fig. 3 Anal manometry and
endosonography in a normal
subject. The endosonographic
pictures are taken during stable
anal canal pressure (A, arrow)
and at minimal straining pressure (B, arrow)

Table 2 Pressure of the anal canal (mBar) measured with and without the endosonographic probe (n = 12)
Normal subjects

Basal
Squeezing
Straining
Rectal distension
* NS vs. with probe

With probe

Without probe

8718
14512
3810
239

8911 *
15218 *
346 *
228 *

canal pressure. Studies by ultrasound anal canal behavior

during rectal distension, squeezing, and straining may give
useful information on pressure.
Thickness of the internal and external and sphincter can
be reliably measured by ultrasound with values not differing from those measured by magnetic resonance [6]. It
seems also that the thickness of the internal and sphincter
is inversely correlated with resting anal pressure, suggesting that the elderly and those with incontinence have a thick
sphincter with low resting pressure [7, 8]. The substitution
of muscle fibers by collagen may account for the increased

25
Fig. 4 Anal manometry and
endosonography in a normal
subject. The endosonographic
pictures are taken during stable
anal canal pressure (A, arrow)
and at minimal pressure recorded during rectal distension with
150 ml of air (B, arrow)

Fig. 5 Correlation coefficient

and regression line between
anal canal size (mm) and anal
canal pressure (mBar) taken
during basal state, squeezing,
straining, and 150 ml of rectal
distension during 48 simultaneous observations with manometry and endosonography

26

muscle thickness and reduced capability to generate pressure in the anal canal. This is supported by evidence that
the lower the electrical activity of the external anal sphincter during squeezing, the thicker is the internal anal sphincter muscle [9]. While the determination of internal and
sphincter thickness is easy, imaging by ultrasound the external anal sphincter is less clear [5]. Our study shows that
the anal canal diameter is correlated well with the pressures generated by the combined internal and external
sphincters. It is appreciated that the diameter measured by
ultrasound is essentially a distortion due to the latex balloon. The presence of the endosonographic probe does not,
however, appear to influence anal canal pressures since no
differences were found when the pressures were measured
with or without the endosonographic probe.
Our study provides evidence that the anal canal of normal human subjects changes in size at variations in pressure in an inverse linear manner.
Acknowledgement Part of this paper was presented at the American Gastroenterological Association Meeting, San Francisco,
California, USA, 19 22 May 1996.

References
1. Law PJ, Bartam CI (1989) Anal endosonography: technique and
normal anatomy. Gastrointest Radiol 14: 349 353
2. Pittmann JS, Benson TB, Sumners JE (1990) Physiologic evaluation of the anorectum a new ultrasound technique. Dis Colon
Rectum 33: 476 478
3. Sultan AH, Kamm MA, Talbot IC, Nicholls RJ, Bartam CI (1994)
Anal endosonography for identifying sphincter defects confirmed
histologically. Br J Surg 81: 463 465
4. Felt-Bersma RJF, Cuesta MA, Koorevaar M, Srijers RLM, Meuwissen SGM, Derksen EJ, Wesdorp RIC (1992) Anal endosonography: relationship with anal manometry and neurophysiologic
tests. Dis Colon Rectum 35: 944 949
5. Emblem R, Dhaenens G, Stien R, Morkrid L, Aasen AO, Bergan
A (1994) The importance of anal endosonography in the evaluation of idiopathic fecal incontinence. Dis Colon Rectum
37: 42 48
6. Schafer A, Enck P, Furst G, Kahn T, Frieling T, Lubke HJ (1994)
Anatomy of the anal sphincters. Comparison of anal endosonography to magnetic resonance imaging. Dis Colon Rectum
37: 777 781
7. Burnett SJD, Bartram CI (1991) Endosonographic variations in
the normal internal sphincter. Int J Colorect Dis 6: 2 4
8. Swash M, Gray A, Lubowski DZ, Nicholls RJ (1988) Ultrastructural changes in internal anal sphincter in neurogenic fecal incontinence. Gut 29: 1692 1698
9. Law PJ, Kamm MA, Bartrams CI (1991) Anal endosonography
in the investigation of fecal incontinence. Br J Surg 78: 312 314

Int J Colorect Dis (1998) 13: 2731

Springer-Verlag 1998

O R I G I N A L A RT I C L E

S. Carter M. Winslet

Delay in the presentation of colorectal carcinoma: a review of causation

Accepted: 10 October 1997

Abstract This review paper studies the reasons behind

the delay in presentation of colorectal carcinoma. Such delay can occur at three different levels: delay on the part of
the patient, delay by the general practitioner prior to patient referral and delay incurred at the hospital after patient
referral. There is some evidence that patient delay has decreased in recent years; general practitioner delay is disputed by some authors, but there is considerable evidence
that it exists. Hospital delay can be influenced by the general practitioner; there is evidence that it has increased in
recent years and is likely to increase further with the current pressure on in-patient beds.
Key words Delay Presentation Colorectal carcinoma
Rsum Cette revue tudie les raisons expliquant le retard dans la mise en vidence des cancers colorectaux. Ces
retards peuvent survenir trois moments diffrents au
cours de lvolution: retard de la part du patient, retard de
la part du praticien avant dadresser le patient et retard
dhospitalisation imputable lhpital une fois que le patient a t annonc. Il y un certain nombre dvidences
prouvant que le retard imputable au patient a diminu au
cours des dernires annes; le retard imputable au praticien est controvers par certains auteurs mais il y a une vidence considrable qui existe. Le retard dhospitalisation
peut tre influenc par le praticien; il y a des vidences que
ce retard a augment au cours des dernires annes et quil
va encore augmenter en raison des pressions exerces en
vue de fermer des lits dhospitalisation.

S. Carter () M. Winslet
University Department of Surgery,
Royal Free Hospital, School of Medicine,
Pond Street, London NW3 2QG, UK

Introduction

Colorectal carcinoma is responsible for approximately

19,000 deaths per annum in the United Kingdom [1]. Overall relative 5-year survival is only 37% despite the introduction of improved surgical techniques and the use of adjuvant chemotherapy [2]. This is due predominantly to the
fact that the disease is often at an advanced stage at diagnosis, and there is scant evidence of a shift in the Duke
stage at presentation over the past 20 years [3, 4]. A major factor that may contribute to the low incidence of earlystage tumours at diagnosis is the delay that can occur at
three different levels: the patient, the general practitioner
and the hospital.

Patient delay

The figures reported for patient delay in the literature are

variable. MacArthur and Smith [5] reported that 45% of
patients had consulted a physician for their bowel symptoms by 1 month from onset, and that a further 28% had
seen one by 3 months, whilst Macadam [6] reported a range
in patient delay from 1 week to 2 years. There is, however,
some consensus of opinion that less delay is associated
with carcinoma of the colon (approx. 3 months) than with
carcinoma of the rectum (4 months) [7]. The reasons for
this are not clear, particularly considering the high incidence of over bleeding in the latter group. Unfortunately,
patients are often oblivious to their symptoms such as rectal bleeding as few individuals regularly look at their stools
or at the toilet paper on defecation [3]. Even when patients
are aware of their symptoms, they may not consult a physician [8] or do so only after a considerable delay [9].
There are several reasons why patients do not seek a
medical opinion: the symptom is not deemed serious
enough, it is considered too unpleasant or embarrassing,
the cause is already known, a wait-and-see policy is adopted, investigation is feared, or a serious underlying condi-

28

tion is suspected; there may also be an association with

dietary indiscretion, a lack of faith in physicians, or a lack
of time to consult one [7, 9, 10]. There is evidence that
some patients are habitual procrastinators, and in patients
who have a positive family history of cancer the past experience of relatives may well lead them to have a negative attitude to the medical profession [11].
Berkanovic [12] noted that women and younger individuals are more likely to experience symptoms that can
be related to cancer, but that overall less than half of these
individuals seek a medical opinion. Those who are more
likely to consult a physician have certain characteristics:
they are more likely to be single, older, on low income, of
lower social class, female and to have been advised by
family or friends to see a physician [12]. Other studies
[10, 13], however, do not concur with all of these characteristics, in particular wih reference to marital state,
social class and income. Dent et al. [10], found no association between patient delay and ethnic group, English
comprehension, occupation, income, marital state, or personality traits.
A study by Hackett et al. [11] addressed the question of
why patients with a subsequent diagnosis of cancer initially consulted their physician and found that pain, incapacity, fear, advice from a friend or relative and attending for
a routine medical check-up are the main reasons. In fact in
this study the highest rate of detection of cancer was in patients attending for a routine check-up. Age per se was not
found to be related to delay, but delay was inversely related to social class, with less delay being associated with
higher social class. In contrast, other studies [13, 14] report increased anxiety about health in the lower social
classes; this may be linked to a lack of information, with
evidence that physicians communicate differently with patients in different socio-economic groups [13].
The patients dilemma is compounded by the fact that
bowel symptoms in the general population are common
and often indistinguishable from those of colorectal carcinoma. In their study of a healthy population Dent et al.
[15] reported a variable bowel habit in almost one-third
of the study sample; rectal bleeding had been experienced
by 16% of individuals in the preceding 6 months, almost
one in five reported tenesmus, and one in ten reported rectal pain on defecation. Other studies [13, 14] concur with
these findings. There do, however, appear to be certain
symptoms that prompt early consultation, including abdominal pain, nausea and vomiting, whereas weight loss
and rectal pain do not, as the latter is often interpreted by
the patient as being associated with haemorrhoids [5].
With regard to rectal bleeding, no association has been reported between patient delay and whether the blood is confined to the toilet paper, separate or admixed with stool,
nor is there any reported statistically significant association between patient delay and whether self-medication
has been tried or not [10].
A study by Holliday et al. [7] on 200 patients with colorectal carcinoma reported that 50% professed ignorance of
their symptoms, whilst only 7% thought that their symptoms may be indicative of cancer. In this series 90% had

discussed their symptoms with relatives or friends but only 22% had consulted a physician. On the positive side,
there is some evidence that patient delay has decreased
over the past decade, with one series reporting a median
delay of 16 weeks in 1984 compared to 6 weeks in 1993
[16], although the reasons for this are unclear.

General practitioner delay

Current practice by general practitioners may contribute to

further diagnostic delay in colorectal carcinoma. In response to a questionnaire sent to general practitioners concerning symptoms, management and referral practice, only
60% indicated that they would immediately refer a patient
with dark red rectal bleeding to hospital whilst 7% reported
that they would not immediately refer a patient with any
type of rectal bleeding [17]. Furthermore, a review of outcome when patients consulted a general practitioner with
bowel symptoms, revealed that one-third had not been examined at all, approximately one-third had had an abdominal examination, and fewer than half had had a rectal examination. In the subsequent referral letter to a hospital
practitioner almost 90% of general practitioners made no
mention of their findings on examination [18].
The reported reasons why general practitioners are reluctant to perform rectal examinations are multifactorial,
including, reluctance on the part of the patient, expectation of a repeat rectal examination at the hospital, unsufficient time and lack of a chaperone [19, 20]. These reasons
were particularly prevalent in the case of general practitioners who were women, and who worked in small
practices, especially in inner London where the waiting
time for an urgent out-patient appointment was less than
2 weeks [19].
The incidence of patient examination may not increase
despite repeated presentation with the same symptoms
[20]. In a study of 100 consecutive patients with established carcinoma of the colon including those who presented as emergencies to hospital [7] three-quarters had already consulted the general practitioner but had not been
investigated or referred to hospital. In the same study of
patients with carcinoma of the rectum who had presented
with bowel symptoms to their general practitioner only approximately one in five had undergone rectal examination,
and half of these were referred to hospital only after consulting their physician on three or more occasions for the
same symptoms.
No statistically significant association has been found
between the decision whether to examine the patient and
the duration of symptoms, the type of practice, the age, sex
or social class of the patients, or how long they have been
registered with the practice [5, 6]. However, delay on the
part of the general practitioner in referring a patient is increased if the patient had only one symptom, had diarrhoea,
or was of low social class. In contrast, if the patient had
constipation or was middle class, general practitioner delay was less likely [5].

29
Table 1 Dukes stage at presentation and delay at general practitioner (CP) and hospital levels (percentages) in carcinoma of the
colon (n = 100) and carcinoma of the rectum (n = 100; from [22] with
kind permission of the Lancet)
Delay

Dukes stage
A

Hospital delay

Colon
None
GP or hospital

78
22 a

17
2

41
9

20
11

Rectum
None
GP or hospital

78
22 b

16
2

39
6

23
14

a
b

income to be earnt by carrying out surgical procedures

[21]. In the current climate of fund-holding contracts in
general practice, these considerations may become more
important.

32% at GP level, 68% at hospital level

82% at GP level, 18% at hospital level

There is considerable variation in the length of general

practitioner delay. In the series reported by MacArthur and
Smith [5] one-third of patients were referred immediately,
one-half had been referred after 1 month, and the remainder were referred more than 3 months after the initial consultation. There appears to be a greater delay in patients
with carcinoma of the rectum, although the reason for this
is unclear. A relationship between the incidence of physical examination and referral time has been suggested, with
the same authors reporting that one-third of patients who
had not had an abdominal or rectal examination (or both)
had a median delay of 90 days prior to referral, whereas in
those examined there was a delay of only 1 3 days [5]. It
is not clear, however, whether the examination per se revealed a lesion that warranted further investigation and referral, or whether delay is less likely in general practitioners who tend to examine their patients. In contrast to the
above, Kingston et al. [16] found little evidence to support
general practitioner delay, with a median referral time of
1 2 weeks in 1984 and less than 1 week in 1993.
Whilst digital rectal examination may increase the diagnostic yield, the value of direct visualization of the anorectum in a primary health care setting is more controversial. Rubin [21] reviewed the use of proctoscopy in general practice in the north of England and reported that just
under one-third of practices possessed a proctoscope, and
of those that did only three-quarters of the general practitioners used it. The reasons cited for not performing
proctoscopy are similar to those outlined for digital rectal examination, with the additional reasons of lack of
training in the use of the instrument and lack of confidence
in interpretation of the findings. Others consider that both
proctoscopy and rigid sigmoidoscopy are inappropriate
procedures in the primary care setting or are considered a
specialist procedure [21]. In view of the inability to exclude carcinoma of the colon and rectum without direct
visualization of the whole of the left colon, routine proctoscopy or sigmoidoscopy alone is difficult to justify in a
primary health care setting. It is noteworthy, however, that
larger general practices gave a more positive response in
the survey, where money saving was paramount, as was

Delay in diagnosis is not solely the responsibility of the

patient and general practitioner (Table 1). Hospital delay,
however, can be influenced by the general practitioner in
two main ways: firstly, by inappropriate referral to departments within the hospital other than surgery. Holliday and Hardcastle [7] reported a series of 200 consecutive patients with colorectal carcinoma including emergency admissions, of whom almost one in five had been
referred to other specialities such as medicine or gynaecology. Furthermore, 65% of these patients had not had
a rectal examination. Secondly, delay may occur if a hospital referral letter is suggestive of haemorrhoids, for example, a young patient referred with rectal bleeding with
a presumptive diagnosis by the general practitioner.
Dixon et al. [23] reported that patients in this category
are commonly given a routine out-patient appointment,
incurring a further delay of 9 weeks. In this series almost
one in three of the patients who were eventually treated
by abdominoperineal resection had not undergone a rectal examination by their general practitioner and were referred with a diagnosis of haemorrhoids. By contrast, if
they had been referred with a letter suggesting a diagnosis of carcinoma they were seen as out-patients within 2
weeks. The type of symptoms experienced by patients
does not appear to be associated with any particular
delay except for the symptom of rectal bleeding [5].
MacArthur and Smith [5] report that 45% of patients with
rectal bleeding are seen within 1 month of referral, compared to almost three-quarters of patients with no such
history.
Other reported reasons for delay at hospital level are
the presence of concomitant disease, slow cross-referral
between hospital departments, problems with the diagnostic accuracy of special investigations [24] and administrative error when no hospital appointment had been sent [16,
25]. Unfortunately, there is evidence to suggest that delay
at hospital level has increased in recent times, from a median delay in 1984 of 1 week to a delay of more than
6 weeks in 1993 [16]. This trend in delay in admission to
hospital is likely to increase in line with the current increased pressure on in-patient beds.

The relationship between delay, stage at presentation

and prognosis

The study by Rowe-Jones [22] clearly illustrates how delay can affect the stage at presentation. One hundred patients with carcinoma of the colon and 100 with carcinoma

30

of the rectum were studied with regard to the duration of

symptoms (patient delay) when there was no medical (general practitioner and or hospital) delay, and the duration of
symptoms when medical delay was present. Patient delay
was similar in the two groups, and thus the significance of
medical delay in relation to prognosis could be evaluated
in this study by comparing the Dukes stage at diagnosis.
Half of the patients with carcinoma of the colon with medical delay had stage C disease at diagnosis and almost twothirds of those with carcinoma of the rectum with medical
delay had stage C disease. There was little evidence of a
shift in the Dukes stage at presentation. Other authors have
observed that when patients with colorectal carcinoma
present as an emergency the duration of symptoms is
often short, less than 3 months, and that in these patients
there is an increased percentage of stage D tumours, which
suggests that these tumours are intrinsically more malignant [26].

tient, may well go some way to improving the outcome for

this most common of cancers.

Summary

There is considerable evidence that there is delay at patient, general practitioner and hospital levels prior to the
diagnosis of colorectal carcinoma. This has been present
for several years and does not seem likely to decrease in
the near future but rather to increase particularly in the
case of delay at hospital level. There are potential means
to try and reduce this delay at all three levels, but there
does appear to be a group of colorectal tumours that are
biologically aggressive and associated with a shorter duration of symptoms, and it seems unlikely that the prognosis in this group of patients will be improved by earlier
diagnosis.

Possible means of reducing delay in presentation

References

Increased patient education with regard to observation and

awareness of their symptoms, with emphasis on those
which should prompt them to seek an early medical opinion, may help to reduce the overall delay in diagnosis and
treatment. Unfortunately, the traditional type of education
campaign may not be helpful to a significant group of patients who deliberately and consciously delay consulting
their general practitioners about their symptoms. Such patients may well be resistant to any education programmes
[11]. In education programmes that have been designed to
worry individuals into being aware of symptoms that may
be attributable to cancer, any fear generated may be counteracted by denial, which in turn may prolong delay. Thus
in such campaigns awareness can be used as an effective
means only if a positive message that cancer can be curable accompanies it [11].
General practitioners have a daunting task in identifying at risk patients, and locally generated protocols from
their allocated hospital coloproctology unit may well reduce this burden. The risk of repeated attendance with the
same symptoms needs to be highlighted. More frequent
routine examination of patients by general practitioners
should be encouraged as this may also reduce the delay in
investigation and treatment, as evidenced by studies citing
the long-term experience with the routine use of the Papanicolaou test and carcinoma of the cervix. In addition,
they should be encouraged to communicate their findings
on examination and to give an informed diagnosis in the
hospital referral letter.
Hospital delay may be reduced by setting up more openaccess flexible sigmoidoscopy clinics for direct referral of
patients with left-sided colonic symptoms [25]. At the very
least the guidelines for managing colorectal cancer set out
by the Royal College of Surgeons of England should be
adhered to [27]. This pragmatic and relatively inexpensive
approach, with emphasis on both the physician and the pa-

1. Austoker J (1994) Screening for colorectal cancer. BMJ 309:

382 386
2. Atkin W, Cuzick J, Northover J, Whynes D (1993) Prevention
of colorectal cancer by once-only sigmoidoscopy. Lancet 341:
736 740
3. Shapiro S (1992) Goals of screening. Cancer 70: 1252 1258
4. Kwok S, Lau W, Carey P, Kelly S, Leung K, Li A (1996) Prospective evaluation of laparoscopic-assisted large-bowel excision
for cancer. Ann Surg 223: 170 176
5. MacArthur C, Smith A (1984) Factors associated with speed of
diagnosis, referral and treatment in colorectal cancer. J Epidemiol Community Health 38: 122 126
6. Macadam D (1979) A study in general practice of the symptoms
and delay patterns in the diagnosis of gastrointestinal cancer.
J R Coll Gen Pract 29: 723 729
7. Holliday H, Hardcastle J (1979) Delay in diagnosis and treatment of symptomatic colorectal cancer. Lancet I: 309 311
8. Crosland A, Jones R (1995) Rectal bleeding prevalence and consultation behaviour. BMJ 311: 486 488
9. Byles J, Redman S, Hennrikus D, Sanson-Fisher R, Dickinson
J (1992) Delay in consulting a medical practitioner about rectal
bleeding. J Epidemiol Community Health 46: 241 244
10. Dent O, Goulston K, Tennant C, Langeluddecke P, Mant A, Chapuis P, Ward M, Bokey L (1990) Rectal bleeding patient delay
in presentation. Dis Colon Rectum 33: 851 857
11. Hackett T, Cassem N, Raker T (1973) Patient delay in cancer.
N Eng J Med 289: 14 20
12. Berkanovic E (1982) Seeking care for cancer relevant symptoms. J Chron Dis 35: 727 734
13. Waitzkin H, Stoeckle J (1972) The communication of information about illness. Adv Psychosom Med 8: 180 215
14. Goldsen R, Gerhardt P, Handy V (1957) Some factors related to
patient delay in seeking diagnosis for cancer symptoms. Cancer
10: 1 7
15. Dent O, Goulston K, Zubrzycki J, Chapuis P (1986) Bowel
symptoms in an apparently well population. Dis Colon Rectum
29: 243 247
16. Kingston R, Jeacock J, Walsh S (1995) Treatment delay times
for patients with colorectal cancer. Abstract from the summer
meeting British Association of Surgical Oncology
17. Vellacott K, Roe A, Mortensen N (1987) An evaluation of a direct access flexible fiberoptic sigmoidoscopy service. Ann R
Coll Surg Engl 69: 149 152

31
18. Springall R, Todd I (1988) General practitioner referral of patients
with lower gastrointestinal symptoms. J R Soc Med 81: 87 88
19. Hennigan T, Franks P, Hocken D, Allen-Mersh T (1990) Rectal
examination in general practice. BMJ 301: 478 480
20. MacArthur C, Smith A (1983) Delay in the diagnosis of colorectal cancer. J R Coll Gen Pract 33: 159 161
21. Rubin G (1992) Endoscopy facilities in general practice. BMJ
304: 1542 1543
22. Rowe-Jones D, Aylett S (1965) Delay in treatment in carcinoma of colon and rectum. Lancet II: 973 976
23. Dixon A, Thornton-Holmes J, Cheetham N (1990) General
practitioners awareness of colorectal cancer a 10 year review.
BMJ 301: 152 153

24. Heald R (1977) Causes of delay in the diagnosis of cancer of

colon and rectum. Practitioner 219: 339 342
25. Jones R (1992) Investigating lower bowel symptoms in general
practice. BMJ 304: 1521 1522
26. Waldron R, Donovan I, Drumm J, Mottram S, Tedman S (1986)
Emergency presentation and mortality from colorectal cancer in
the elderly. Br J Surg 73: 214 216
27. Royal College of Surgeons of England (1996) Guidelines for the
management of colorectal cancer. Association of Coloproctology of Great Britain and Ireland. R Coll Surg Engl

Int J Colorect Dis (1998) 13: 3238

Springer-Verlag 1998

O R I G I N A L A RT I C L E

N. Saito K. Koda N. Takiguchi K. Oda H. Soda

M. Nunomura H. Sarashina N. Nakajima

Surgery for local pelvic recurrence after resection of rectal cancer

Accepted: 10 October 1997

Abstract This retrospective study evaluated outcome

with regard to procedure, local control, and survival after
curative surgical resection with and without preoperative
radiotherapy for local pelvic recurrence. A total of 58 consecutive patients with local pelvic recurrence of rectal cancer after previous curative resection for primary tumors
were reviewed. Of these, 36 underwent both initial resection and follow-up in our department; the remaining 22 had
initial surgery and follow-up elsewhere. Of the 58 patients
27 underwent curative re-resection, 9 had palliative resection, and 22 were treated by conservative therapy. Among
the 27 patients with curative resection 17 received preoperative radiotherapy (40 Gy) plus surgery and 10 surgery
only. No patients were lost to follow-up; median followup time was 36.3 months. The overall rate of curative resection was 46.6%: 55.6% in our own follow-up group and
31.8% in the others. With regard to surgical procedure, abdominoperineal resection (APR) with or without sacral
resection was standard following previous low anterior resection, and total pelvic exenteration (TPE) with or without sacral resection was common following APR. There
was a high incidence of morbidity (71.4%) after TPE. Rerecurrence was observed in 12 (44.4%) after curative reresection. There was local re-recurrence in 6 (22.2%). The
local re-recurrence rate was 11.8% (n = 2) with radiotherapy plus surgery, and 40.0% (n = 4) with surgery alone. The
estimated 5-year survival following curative re-resection
was 45.6% (61.2% with radiotherapy plus surgery, 29.6%
with surgery alone). Both survival and local control with
radiotherapy plus surgery tended to be better than with surgery alone. Thus, in selected patients pelvic local recurrence of rectal cancer can be re-resected curably by APR
or TPE (with or without sacral resection) combined with
preoperative radiotherapy.

N. Saito () K. Koda N. Takiguchi K. Oda H. Soda

M. Nunomura H. Sarashima N. Nakajima
First Department of Surgery,
Chiba University School of Medicine, 1-8-1, Inohana, Chuo-Ku,
Chiba 260, Japan

Key words Rectal cancer Local pelvic recurrence

Preoperative radiotherapy Total pelvic exenteration
Abdominoperineal resection Sacral resection
Rsum Le but de cette tude rtrospective est dvaluer
le devenir des interventions chirurgicales, le contrle local et la survie aprs rsection chirurgicale curative avec
ou sans radiothrapie pr-opratoire de rcidives pelviennes locales. Mthodes: cinquante-huit patients conscutifs
porteurs dune rcidive pelvienne locale dun cancer du
rectum, aprs rsection curative pour tumeur primaire, ont
t revus. Ces patients ont t classs dans un groups soumis un follow-up (36 patients) et les autres (22 patients).
Les patients du groupe follow-up avaient subi la rsection
initiale et le suivi post-opratoire dans notre dpartement.
Vingtsept des cinquante-huit patients ont subi une r-rsection curative; neuf ont subi une rsection palliative et
vingt-deux ont t traits par un traitement purement conservateur. Chez les vingt-sept patients avec rsection curative, dix-sept ont reu de la radiothrapie pr-opratoire
(40 Gy) XRT+ chirurgie et 10 patients qui nont subi
quun geste chirurgical isol. Aucun patient na t perdu
de vue. Le temps moyen de survie est de 36,3 mois.
Rsultats: la moyenne de rsection curative est de
46,6% : 55,6% dans le groupe avec follow-up et 31,8% dans
le reste des cas. En ce qui concerne le procd chirurgical,
lamputation abdomino-prinale (APR) avec ou sans rsection sacre est loption standard faisant suite une rsection antrieure, et lexentration pelvienne totale (TPE)
avec ou sans rsection sacre a t ralise frquemment
aprs une amputation abdomino-prinale. Aprs une
exentration pelvienne totale, la morbidit est leve
(71,4%), une deuxime rcidive survient chez
12 malades (44,4% aprs une re-rsection curative). Une
deuxime rcidive survient chez 6 patients (22,2%). Le taux
de re-rcidive locale est de 11,8% (2 patients) chez ceux
qui ont subi une radiothrapie per-opratoire + chirurgie
et de 40% (4 patients) dans le groupe ayant subi la chirurgie seule. La survie 5 ans estime aprs re-rsection curative est de 45,6% (61,2% dans le groupe avec radiothrapie et chirurgie, 29,6% dans le groupe avec chirurgie

33

seule). La survie et le contrle local chez les patients ayant

sub une radiothrapie pr-opratoire + chirurgie est meilleure que le groupe ayant subi une chirurgie seule. Conclusion: dans un groupe de patients slectionns, une rcidive pelvienne locale du cancer peut tre re-rsque de
manire curative par amputation prinale ou exentration
pelvienne totale (avec ou sans rsection sacre) en combinaison avec une radiothrapie pr-opratoire.

Introduction

Local regional disease in the pelvis is the most common

pattern of recurrence after rectal cancer surgery. Locoregional recurrence after radical surgery for rectal cancer varies from 20 30% in a number of series to less than 10%,
and even below 5%, in others [1 8]. The vanable incidence reported in the literature is due to different site, stage,
and character of the primary tumor, and to the surgical technique, i. e., total mesorectal excision and degree of lymphadenectomy including lateral dissection [9 11]. It also
depends on the adjuvant therapy, such as radiotherapy
(XRT), hyperthermia, and chemotherapy, or combinations
[12 17]. Whether these therapeutic options, such as extended radical surgery combined with XRT, can improve
survival, local control, and pelvic recurrence is unclear. Although XRT can reduce the development of local recurrence and provide palliation for localized disease, it is not
curative. In the study by Cass et al. [18] the survival rate
after local regional recurrence was less than 4% after 4
years in patients with rectal cancer treated without surgery,
and most patients died from isolated disease.
Several series report the resection of local recurrence
following rectal cancer [9, 10, 19 21]. For recurrence at
the anastomotic site resection, usually by adominoperineal
resection (APR), can be curative [22]. There are now several published series incorporating abdominal sacral resection, including extended pelvic exenteration for recurrent
rectal cancer [23 30]. This concept was advanced years
ago by Brunschwig [29], who combined exenteration with
resection of the bony segments of the pelvis and reported
good results with regard to survival and local control.
The goal of this retrospective study was to evaluate results in treating local pelvic recurrence after radical
surgery for rectal cancer, with regard to local control and
survival after surgery with curative intent and the effect of
curative resection combined with preoperative radiotherapy.

Patients and methods

Fifty-eight patients were treated for local pelvic recurrence of rectal
cancer in our Department between April 1981 and September 1995.
There were 45 males and 13 females, with ages ranging from 44 to
75 months (median 64). In these 58 patients the time between initial
operation and diagnosis of local pelvic recurrence ranged from 10 to
56 months (median 22). No patient had received preoperative or post-

operative XRT in initial treatment. Patients were classified into two

groups. Of the 58 patients 36 were operated in our Department for
cure by low anterior resection or APR, and follow-up was also carried out in our Department (group 1). The other 22 patients were operated on and followed-up in other hospitals and came to our Department for treatment of local recurrence (group 2). Follow-up was more
rigorous in group 1 than in group 2. The regimen of our follow-up
is presented in Table 1. Follow-up was performed at intervals of 3
months for the first 2 years after the operation.
Local pelvic recurrence was defined as relapsed tumor growth
within the pelvic cavity, including the anastomotic site, after a curative operation for primary rectal cancer. Histopathological findings
of recurrence were obtained in all cases by endoscopy, echo, and
computed tomography (CT) guided biopsy. To determine whether
surgical resection could be performed the patients were evaluated by
tumor markers such as carcinoembryonic antigen, liver and pelvic
ultrasonography (US), CT of the upper abdomen and pelvic, magnetic resonance imaging (MRI) of the pelvis, barium enema or colonofiberscopy (CF), chest X-ray, and bone scan. Local recurrence
was evaluated by US, CT, MRI, and if possible by positron emission
tomography (PET). US and CT focusing on the liver, and chest Xray and bone scan were undertaken to exclude detectable distant metastases. In cases of suspicious local recurrence imaging-guided biopsy and PET were performed for definite diagnosis.
Patients with distant metastases, demonstrable extensive invasion to the lateral pelvic wall, presenting with obstruction of bilateral femoral vein, and with bilateral sciatic pain caused by local pelvic recurrence were not considered for an operation with curative intent. In accordance with this, 27 patients had resection with curative
intent, and 9 underwent palliative resection. The surgical procedure,
such as tumor resection, APR with or without sacral resection, and
total pelvic exenteration (TPE) with or without sacral resection, was
determined by the degree of tumor growth in the patient. Patients
undergoing curative resection were classified into two groups. Of the
27 patients with curative resection 17 received preoperatively XRT
as adjuvant therapy: the total dose of preoperative XRT was 40 Gy
in each patient; the usual regime was 2 Gy, five fractions a week.
The other 10 patients underwent surgery alone. Patients who had mucinous carcinoma or severe complications, or who rejected XRT,
were excluded from the protocol for XRT plus surgery; these underwent surgery without XRT. Median follow-up time was 36.3 months
(range 3 118) in patients receiving curative resection.
There was no difference in surgical procedure between the group
receiving XRT + surgery and the group receiving surgery alone. The
pelvis was filled with a Dexon mesh after resection in the patients
undergoing TPE. In 9 patients undergoing palliative resection APR
was performed in 4 patients, TPE in 2, and tumor resection in 3 to
reduce clinical symptoms. In 22 patients with conservative therapy
chemo-radiotherapy with or without hyperthermia was performed.
In the 27 patients with curative resection for local pelvic recurrence
the mean follow-up time after surgery was 36.3 months (range
3 118) in the XRT + surgery group, and 36.4 months (range 6 77)
in the surgery alone group.
The number of patients receiving curative resection, palliative
resection, and conservative therapy are shown in Table 2. A curative
resection was performed in 27 cases (46.6%) and a palliative resection in 9 (15.5%) of the 58 patients with local pelvic recurrence. In
group 1, 20 of the 36 patients (55.6%) underwent curative resection
but only 7 of the 22 patients (31.8%) in group 2. Table 2 also presents
the relationship of stage and degree of lymph node metastases in primary cancer to treatment of local recurrence. In the
58 patients with local recurrence 41 had been in Dukes stage C at initial treatment; among these curative resection was performed in only 16 (39.0%). Only 3 of 12 patients with lateral lymph node metastases at initial treatment could undergo curative resection. In
17 recurrent patients with Dukes stage A or B at initial treatment
11 (64.7%) were able to undergo curative resection.
The operative procedures for local pelvic recurrence by each initial treatment are shown in Table 3. In 27 patients with curative resection for local recurrence TPE with or without sacral resection was
performed in 14 (51.9%), APR with or without sacral resection in 10
(37.0%), and tumor resection in 3. In 15 patients with a sphincter

34
Table 1 Follow-up regimen: frequency of examinations per year
Dukes
stage
A
B
C1
C2

Tumor
marker

Pelvis

4
4
4
4

Barium enema
or CF

US

CT

MRI

1
1
2
2

1
1
2
2

1
1, 2
1, 2

Chest
X-ray

1
1
1, 2
1, 2

Liver

1
1
2
2

US

CT

1
3
2
3

1
2
1
2

Table 2 Treatment of pelvic recurrence. (n1 corresponds to Dukes C1 , n2 ~ corresponds to Dukes C2 , including lateral lymph nodes metastases)
Total (n)

Curative resection

Palliative resection

Conservative therapy

Total
Group 1
Group 2

58
36
22

27
20
7

46.6
55.6
31.8

9
4
5

15.5
11.1
22.7

22
12
10

37.9
33.3
45.5

Initial Dukes stage

A
B
C
n1
n2 ~
Lateral (+)

7
10
41
24
17
12

4
7
16
10
6
3

57.1
70.0
39.0
41.7
35.3
25.0

1
1
7
7
0
0

2
2
18
7
11
9

Table 3 Operating procedure in pelvic recurrence. (Conservative therapy: SPO n = 4, APR n = 18; resection rate: SPO 20/25, 89%; APR
16/33, 48.5%; P<0.05)
Initial procedure

Total (n)

Secondary procedure
Tumor resection
(n)

APR

TPE

Curative resection
Total
SPO
APR

27 a
15
12

3
2
1

10 b
10

37.0
71.4

14 b
3
11

51.9

84.5

Palliative resection
Total
SPO
APR

9
5
4

4
4

2
1
1

a
b

Preoperative XRT: 17 cases

Combined with sacral resection: APR, 2 cases; TPE, 8 cases

preserving operation (SPO) as initial treatment and 10 (71.4%) had

a curative resection by APR. In 12 with APR as initial treatment 11
(84.6%) were cured by TPE. In 10 of these 27 patients (37.0%)
8 underwent TPE combined with sacral resection and 2 patients received APR with sacral resection because local recurrence was at or
very close to the os sacrum (below S2 level) and the urinary bladder
or the prostate.
The surgical procedure with or without preoperative radiotherapy for local pelvic recurrence, the sites of re-recurrence after curative resection, and survival were reviewed. Staging in primary rectal cancer was performed in accordance with Dukes classification
[31], and the degree of lymph node metastases was in accordance
with Japanese Society for Cancer of the Colon and Rectum [32]. For
statistical analysis the 2 and Fischer exact tests were applied to all
relationships, and the Kaplan-Meyer method was used to compute

survival and corresponding curves. Survival functions for subgroups

were compared using the log-rank method.

Results

No patients were lost to follow-up. Re-recurrence after the

curative resection was documented in 27 patients (Table 4).
Twelve (44.4%) developed local re-recurrent tumors and
distant metastases; six (22.2%) of these had local and six
(22.2%) had distant metastases. In patients with local re-recurrence two received XRT + surgery (re-recurrence rate

35
Table 4 Sites of failure of curative resection for recurrent rectal cancer
Treatment

Total

Local
regional recurrence

Liver
metastases

Lung
metastases

Bone
metastases

Total

12 (44.2%)

6 a (22.2%)

Secondary procedure
XRT+surgery b (n = 15)
Surgery alone c (n = 10)

5
7

2
4

2
4

3
2

3
2

Initial procedure
SPO (n = 15)
APR (n = 12)

7
5

4 (26.7%)
2 (16.7%)

4
2

1
4

3
2

a
b
c

Three were cases of tumor resection at secondary treatment

Mean follow-up: 36.3 months
Mean follow-up: 36.4 months

Fig. 1 Survival curves for cases of pelvic recurrence

Fig. 2 Comparison of survival following curative resection in

XRT + surgery and surgery alone

11.8%) and four received surgery alone (re-recurrence rate

40.0%). Five of six patients had synchronous distant metastases, and four (14.8%) had bone metastases. There was no
significant difference in pattern of re-recurrence between
two groups. There were four local re-recurrences (26.7%)

in 15 patients with initial SPO. There were two local re-recurrences (16.7%) in 12 with initial APR: There was no significant difference between initial procedures. However, all
three patients who underwent local tumor resection in the
secondary procedure developed locally re-recurrent tumors.
Overall survival curves in patients undergoing this treatment are shown in Fig. 1. In the 27 patients with curative
resection there were six (22.2%) 5-year survivors, four of
whom had received XRT + surgery (TPE, one; TPE with
sacral resection, one; APR with sacral resection, two), and
two received surgery alone (TPE, one; TPE with sacral resection, one). One of the two patients in the latter group
died of local re-recurrence with distant metastases
65 months after the operation. Five are alive and well without evidence of disease. The estimated 5-year survival in
this group of 27 patients was 45.6%. There were no 5-year
survivors among those receiving palliative resection or
conservative therapy. There was a significant difference in
the survival curves between the patients with curative resection and the patients with palliative resection and conservative therapy. Actuarial survival curves of the patients
with curative resection for treatment group (XRT + surgery, surgery alone) are shown in Fig. 2. The estimated
5-year survival in XRT + surgery group was 61.2% and
that in the surgery alone group was 29.6%; there was no
significant difference in the curves between the two groups.
Although no perioperative deaths occurred in this series, complications were common in those undergoing TPE
or TPE with sacral resection (n = 14), with ten patients
(71.4%) having perioperative complications (Table 5).
Seven of eight patients with XRT + TPE (87.5%) and all
patients with TPE and sacral resection had serious complications. The most common complications were infection
and poor wound healing. Serious complications such as fistula and cardiovascular abnormalities were observed in the
groups receiving XRT + TPE and TPE with sacral resection. The morbidity rate (71.4%) was significant. Intraoperative blood loss was extensive (average 6900 ml, range
4000 12 600 ml), and the duration of operation was long
(average 13 h, range 8.5 26 h) in these procedures, especially in TPE with sacral resection.

36
Table 5 Morbidity in 14 patients with TPE with or without sacral
resection for pelvic local recurrence. (Intraoperative blood loss: average 6900 ml, range 4000 12,600 ml; duration of operation: average 13 h, range 8.5 26 h)
Complications

n
a

Perioperative morbidity
Infection
Abscess
Sepsis
Wound complications
Infection
Separation
Fistula
Small bowel/large bowel
Ureteral
Cardiovascular complications
Deep venous thrombosis
Postoperative bleeding
Pulmonary insufficiency
Hepatic failure
a

10 (71.4%)
7
3
6
4
3
2
3
2
3
1

XRT + TPE, 7/8 (87.5%); TPE alone, 3/6 (50.0%)

Discussion

Local pelvic recurrence of rectal cancer is a major problem. The prognosis for patients who develop pelvic recurrence is generally poor, with 5-year survival ranging from
2% to 30% [18, 33, 34]. Surgical resection remains the only
potentially curative treatment in the majority of these patients. As long ago as 1959 Bacon and Berkley reported a
group of 93 patients with local recurrence [35]. In this report 38 patients underwent a curative resection, and
5-year survival was 37%. Recent studies suggest that the
addition of external or intraoperative XRT surgical resection improves survival and local control. However, the risk
of further local recurrence remains approximately 50%
[36 40]. The study by Wanebo et al. [41] on pelvic resection of recurrent rectal cancer reported that pelvic recurrence was resected safely with the expectation of long-term
survival in 33% of cases.
Resection of local recurrence probably also offers the
best chance for improved local control and palliation. For
recurrence at the anastomotic site resection can be curative or palliative by APR with or without sacral resection.
However, there is less chance of resection in local recurrence following radical APR. In our study the resection
rate (curative and palliative) was 80.0% in the SPO group
and 48.5% in the APR group. There was a significant difference in the resectability between two initial treatment
groups. This may in part be due to the fact that it is sometimes difficult to detect local recurrence early and to distinguish local recurrence from scar tissue formation. Most
studies report the role of early detection of local recurrence in resectability and prognosis after resection. In our
study the rate of curative resection in local recurrence was
46.6%: 55.6% in group 1 but only 31.8% group 2. There
was a significant difference in resectability between two
groups. The mean size of recurrent tumors in the resected
specimen was 3.6 cm (range 2.2 5.1 cm) in group 1 and

4.9 cm (range 3.8 6.9 cm) in group 2. These results suggest that our follow-up (group 2) was performed more intensive and was useful for the early detection of local recurrence. However, a simple and more effective followup must be considered. There were few chances for curative resection in the patients with local recurrence who
had undergone an extended radical surgery for the lateral
spread lymph node metastases in initial treatment since
most had local recurrence with the demonstrable extensive involvement of the lateral pelvic wall with distant
metastases. There were few chances for curative resection in recurrent cases that had had an extended radical
resection for a considerable advanced tumor in the initial
treatment.
In the curative procedure for local pelvic recurrence
APR with or without sacral resection was the standard procedure for patients with previous SPO, and TPE with or
without sacral resection was the standard operation for patients with previous APR, because several foci of local recurrence were frequently found in the pelvis, and it was
impossible to obtain sufficient surgical margins by other
procedures. In our experience these pathological findings
are observed in 4 of 16 specimens resected by TPE. Recurrence invading the urogenital organs can be resected
successfully by TPE. When recurrence invades or is adjacent to the sacrum, a combined sacral resection below S2
or S3 is required to obtain radical (curative, negative) margins. All of the patients receiving local excision for pelvic
recurrence developed local re-recurrent tumors in our
study. Local excision (tumor resection) was not a curative
procedure for local recurrence after rectal cancer surgery.
TPE with or without sacral resection was performed in
11 of 12 patients with previous APR in this study. However, this radical procedure for local recurrence had extensive perioperative morbidity (pelvic abscess, sepsis,
wound infection, and separation, bowel/urinary fistula,
pulmonary insufficiency, postoperative bleeding, deep venous thrombosis, hepatic failure), extensive blood loss (average 6900 ml, range 4000 12 600 ml), and long duration
of operation (average 13 h, range 8.5 26 h).
Complications were common in patients receiving TPE
for local recurrence (9 of 14 patients, 64.3%), and there
were four operative deaths (perioperative mortality 9.5%)
in the series using by this procedure. In the study of pelvic resection by Wanebo et al. [41] perioperative mortality was 8.5%, and the majority of their cases had operative
complications. Therefore careful selection of patients is
necessary for these procedures. All three of our patients
with local tumor resection for recurrence developed local
re-recurrence in the pelvis within 2 years of re-operation.
Local tumor resection therefore does not seen an appropriate procedure for local recurrence as curative margins (whether microscopically clear) were not obtained by this
procedure. Since there was the potential for microscopic
or minimal gross residual disease in the pelvis, TPE with
or without sacral resection was a viable procedure for patients who developed pelvic recurrence following previous APR. Complete removal of recurrence resulted in improved survival.

37

In the study by Wanebo et al. [23, 41] 24 of 47 patients

with recurrent rectal cancer had tumors potentially amenable to complete resection by pelvic resection. Median survival was 36 months; 47 of 53 patients with pelvic resection for recurrent rectal cancer were cured, and long-term
survival for 4 years was achieved in 14 of 43 patients
(33%). Recent studies suggest that it is possible to obtain
better survival and local control if external or intraoperative radiotherapy is added to surgical resection of local recurrence. However, the risk of further local recurrence is
still approximately 50% [3643]. In our study a curative
resection combined with preoperative XRT was performed
in 17 patients, and a curative resection without XRT in 10.
Five-year survival was 61.2% in those receiving XRT +
surgery and 29.6% in those receiving surgery alone. Although there was no significant difference between the two
groups, survival in the former group was better than that
in the latter. The local re-recurrent rate was relatively low
(2 of 17 patients, 11.8%) in the XRT + surgery group but
was high (4 of 10 patients, 40.0%) in the surgery alone
group. Although there was no statistically significant difference between two groups, those in the group receiving
XRT + surgery tended to fare better than those receiving
surgery alone for local control. Cytoreduction and downstaging seemed to be obtained by the preoperative XRT
(total dose 40 Gy) in widespread recurrent disease of the
pelvic cavity, and sufficient negative margins were
achieved for the majority of those in the XRT + surgery
group by resection. These results suggest tumor downstaging by preoperative XRT and radical resections enhances
local control and perhaps improves long-term survival.
Many patients, however, developed recurrence including distant metastases after these extended radical procedures. Recurrence developed in 12 of 27 patients (44.4%),
who in general had a very poor course. Five of six patients
with local re-recurrence had distant metastases, and these
distant metastases were multiple. Also, the propensity for
lung and bone metastases (5 of 12 patients) was higher than
usually expected in patients with advanced primary rectal
cancer. Although XRT + surgery was able to achieve preoperative tumor cytoreduction and to provide radical margins, it was not successful in decreasing the incidence of
distant metastases. This high incidence of distant metastases was presumably caused in part by intraoperative manipulation. Therefore, as noted by Lowy et al. [42], it seems
that further combined therapy, such as chemoradiation or
other new treatments, is necessary to prevent distant micrometastatic disease and to improve long-term survival.
Local recurrence following APR is frequently more extensive and invades the pelvic viscera and lateral or posterior
pelvic wall. Sometimes it may be resected successfully by
TPE with or without sacral resection combined with preoperative XRT. However, these procedure are extensive
and entail morbidity and mortality. Candidates for these
operations should therefore be selected very carefully. Determining relative contraindication is influenced by the initial Dukes stage of the primary cancer and the type of initial surgery. In this regard there is little benefit from this
procedure in patients who had had a low-grade, poorly dif-

ferentiated adenocarcinoma, with extensive lymph node

metastases, and an extensive radical APR.

Conclusions

Pelvic local recurrence of rectal cancer can be resected curably by APR and TPE with or without sacral resection in
selected patients. TPE with or without sacral resection
should be standard procedure for local recurrence following previous APR. The estimated 5-year survival in our curative patients was 45.6%. Survival following surgery
combined with preoperative XRT also appears superior to
that after surgery alone. Although relatively few patients
received curative resection for local pelvic recurrence in
this study, it seemed that early detection of local recurrence, careful patient selection, and wide excision with preoperative radiotherapy were important for obtaining good
results.

References
1. Nicholls RJ, Ritchie JK, Wadsworth J, Parks AG (1979) Total
excision or retrative resection for carcinoma of the middle third
of the rectum. Br J Surg 66: 625 627
2. Luke M, Kirkegaard P, Lendorf A, Christiansen J (1983) Pelvic
recurrence rate after abdominoperineal resection and low anterior resection for rectal cancer before and after introduction of
the stapling technique. World J Surg 7: 616 619
3. Williams NS, Johnson D (1984) Survival and recurrence after
sphincter saving resection and abdominoperineal resection for
carcinoma of the middle third of the rectum. Br J Surg
71: 278 282
4. Gillen P, Peel AL (1986) Comparison of the mortality, morbidity and incidence of local recurrence in patients with cancer treated by either stapled anterior resection or abdominoperineal resection. Br J Surg 73: 339 341
5. Vlasak JW, Wagner D, Passaro E, Wilson SE (1989) Local recurrence after curative resection of rectal cancer: a comparison
of low anterior resection and abdominoperineal resection. J Surg
Oncol 41: 236 239
6. Fick TE, Baeten CG, von Meyenfeldt MF, Obertop H (1990) Recurrence and survival after abdominoperineal and low anterior
resection for rectal cancer, without adjunctive therapy. Eur J Surg
Oncol 16: 105 108
7. Wolff BG (1992) Lateral margins of resection in adenocarcinoma of the rectum. World J Surg 16: 467 469
8. MacFarlane JK, Ryall RD, Heald RJ (1993) Mesorectal excision
for rectal cancer. Lancet 341: 457 460
9. Sanella NA (1976) Abdominoperineal resection following anterior resection. Cancer 38: 378 381
10. Segall MM, Goldberg SM, Nivatvongs S, Balcos EG, Nemer
FD, Schottler JL, Christenson CE, Rothenberger DA (1981) Abdominoperineal resection for recurrent cancer following anterior resection. Dis Colon Rectum 24: 80 84
11. Vassilopoulos PP, Yoon JM, Ledesma EJ, Mittelman A (1981)
Treatment of recurrence of adenocarcinoma of the colon and rectum at the anastomotic site. Surg Gynecol Obstet 152: 777 780
12. Pahlman L, Glimelius B (1990) Pre- or postoperative radiotherapy in rectal and rectosigmoid carcinoma: report from a randomized multicenter trial. Ann Surg 211: 187 195
13. Stockholm Rectal Cancer Study Group (1990) Preoperative
short-term radiation therapy in operable rectal cancer. Cancer
66: 49 55

38
14. Marsh PJ, James RD, Schofield PF (1994) Adjuvant preoperative radiotherapy for locally advanced rectal carcinoma. Dis Colon Rectum 37: 1205 1214
15. Fujimoto S, Takahashi M, Endo F, Shrestha RD, Kokubun M,
Takai M, Okui K (1991) A clinical pilot study combining surgery with intraoperative pelvic hyperthermochemotherapy to
prevent the local recurrence of rectal cancer. Ann Surg
213: 43 47
16. Korenaga D, Matsushita T, Adachi Y, Mori M, Matsuda H, Kuwano H, Sugimachi K (1992) Preoperative hyperthermia combined with chemotherapy and radiotherapy for patients with rectal carcinoma may prevent early local pelvic recurrence. Int J
Colorect Dis 7: 206 209
17. Willett CG, Shellito PC, Tepper JE, Eliseo R, Convery K, Wood
WC (1991) Intraoperative electron beam radiation therapy for
primary locally advanced rectal and rectosigmoid carcinoma.
J Clin Oncol 9: 843 849
18. Cass AW, Million RR, Pfaff WW (1976) Patterns of recurrence
following surgery alone for adenocarcinoma of the colon and
rectum. Cancer 37: 2861 2865
19. Polk HC, Spratt JJ (1979) Results of treatment of perineal recurrence of cancer of the rectum. Cancer 43: 952 956
20. Hojo K (1986) Anastomotic recurrence after sphincter saving
resection for rectal cancer; length of distal clearance of the bowel. Dis Colon Rectum 29: 11 14
21. Schiessel R, Wunderlick M, Herbst F (1986) Local recurrence
of colorectal cancer; effect of early detection and aggressive surgery. Br J Surg 72: 342 344
22. Nymann T, Jess P, Christiansen J (1995) Rate and treatment of
pelvic recurrence after abdominoperineal resection and low anterior resection for rectal cancer. Dis Colon Rectum 38: 799 802
23. Wanebo HJ, Gaker DL, Whitehill R, Morgan RF, Constable WC
(1987) Pelvic recurrence of rectal cancer; options for curative
resection. Ann Surg 205: 482 495
24. Takagi H, Morimoto T, Hara S (1986) Seven cases of pelvic exenteration combined with sacral resection for locally recurrent
rectal cancer. J Surg Oncol 32: 184 188
25. Pearlman NW, Donahue RE, Steigmann GV, Ahnen DJ, Sedlacek SM, Braun TJ (1987) Pelvic and sacropelvic exenteration
for locally advanced or recurrent anorectal cancer. Arch Surg
122: 537 541
26. Touran T, Frost DB, OConnell TR (1990) Sacral resection: operative technique and outcome. Arch Surg 179: 555 560
27. Temple WJ, Ketchman AS (1991) Sacral resection for control
of pelvic tumors. Am J Surg 163: 370 374
28. Wanebo HJ, Marcove RC (1981) Abdominosacral resection of
locally recurrent rectal cancer. Am Surg 194: 458 471

29. Brunschwig A, Barber K (1969) Pelvic exenteration combined

with resection of segments of bony pelvis. Surgery 65: 417 420
30. Sugarbaker PH (1982) Partial sacrectomy for en bloc excision
of rectal cancer with posterior fixation. Dis Colon Rectum
25: 708 711
31. Dukes CE (1932) The classification of cancer of the rectum.
J Pathol Bacteriol 35: 323 339
32. Japanese Society for Cancer of the Colon and Rectum (1994)
General rules for clinical and pathological studies on cancer of
the colon, rectum and anus. The 5th edition, Kanehara & Co.,
Ltd., Tokyo, pp 12 25
33. Hoffman JP, Riley L, Carp NZ, Litwin S (1993) Isolated locally recurrent rectal cancer; a review of the incidence, presentation, and management. Semin Oncol 20: 506 519
34. Turk PS, Wanebo HJ (1993) Results of surgical treatment of nonhepatic recurrence of colorectal carcinoma. Cancer 71:
4267 4277
35. Bacon HE, Berkley JL (1959) The rationale of re-resection for
recurrent cancer of the colon and the rectum. Dis Colon Rectum
2: 549 554
36. OConnell MJ, Childs DS, Moertel CG, Holbrock MA, Schutt
AJ, Rubin RE Jr (1982) A prospective controlled evaluation of
combined pelvic radiotherapy and methanol extraction residue
of BCG (MER) for locally unresectable or recurrent rectal carcinoma. Int J Radiat Oncol Biol Phys 8: 1115 1119
37. Gastrointestinal Tumor Study Group (1985) Prolongation of the
disease-free interval in surgically treated rectal carcinoma.
N Engl J Med 312: 465 472
38. Ketcham AS (1985) The management of recurrent rectal carcinoma. Can J Surg 28: 422 424
39. Beart RW Jr, Martin JK Jr, Gunderson LL (1986) Management
of recurrent rectal cancer. Mayo Clin Proc 61: 448 450
40. Jones DJ, Zaloudik J, James RD, Haboubi N, Moore M, Schofield PF (1986) Predicting local recurrence of carcinoma of the
rectum after preoperative radiotherapy and surgery. Br J Surg
76: 1171 1175
41. Wanebo HJ, Kones JR, Vezeridis MP, Cohen S, Wrobleski DE
(1994) Pelvic resection of recurrent rectal cancer. Ann Surg
220: 586 597
42. Lowy AM, Rich TA, Skibber JM, Dubrow RA, Curley SA (1996)
Preoperative infusional chemoradiation, selective intraoperative
radiation, and resection for locally advanced pelvic recurrence
of colorectal adenocarcinoma. Ann Surg 223: 177 185
43. Wiggers T, Vries MR, Veeze-Kuypers B (1996) Surgery for local recurrence of rectal carcinoma. Dis Colon Rectum 39:
323 328

Int J Colorect Dis (1998) 13: 3942

Springer-Verlag 1998

O R I G I N A L A RT I C L E

S. Causeret Y. Franois J. B. Griot B. Flourie

F. N. Gilly J. Vignal

Prophylactic pancreaticoduodenectomy for premalignant duodenal polyposis

in familial adenomatous polyposis

Accepted: 29 October 1997

Abstract The frequency of duodenal adenomas in patients with, familial adenomatous polyposis is high.
Duodenal adenoma has malignant potential, and duodenal
adenocarcinoma is one of the main causes of death in patients who have had previous proctocolectomy. A conservative approach to the treatment of duodenal adenomas
(nonsteroidal anti-inflammatory drugs, endoscopy, polypectomy through duodenotomy) is inefficient and unsafe.
When invasive cancer occurs in duodenal adenomas, the
result of surgery is poor. We have performed prophylactic
pancreaticoduodenal resection (PDR) for nonmalignant
severe duodenal polyposis in five patients since 1991. No
operative mortality was observed. One patient developed
a pancreatic fistula which was successfully managed by
medical treatment. The mean follow-up was 35 months.
All five patients are still alive and have a good functional
outcome. Prophylactic PDR may be indicated in familial
adenomatous polyposis when duodenal polyposis is severe. Stages III and IV of Spigelmans classification, periampullary adenoma, age above 40, and family history of
duodenal cancer are factors that may lead to the decision
to perform prophylactic PDR.
Key words Familial adenomatous polyposis Duodenal
adenomas Prophylactic pancreaticoduodenal resection
Rsum Lincidence dadnomes duodnaux chez des
patients atteints de polypose familiale (F. A. P.) est levee. Les adnomes du duodnum ont un potentiel de malignit et les adnocarcinomes du duodnum sont une des
causes principales de dcs chez les patients qui avaient
subi au pralable une procto-colectomie. Une approche
conservatrice du traitement des adnomes duodnaux (mdication anti-inflammatoire non strodienne, endoscopie,
S. Causeret Y. Franois J. B. Griot F. N. Gilly J. Vignal ()
Department of Surgery, Hpital Lyon Sud,
F-69495 Pierre Benite Cedex, France
B. Flourie
Department of Gastroenterology, Hpital Lyon Sud,
F-69495 Pierre Benite Cedex, France

polypectomie par duodnotomie) est insuffisante et peu

sre. Si un cancer invasif se dveloppe dans un adnome
du duodnum, le rsultat de la chirurgie est mauvais. Nous
avons ralis des rsections pancratico-duodnales prophylactiques pour des polyposes duodnales svres non
malignes chez 5 patients depuis 1991. Aucune mortalit
opratoire nest dplorer. Un patient a dvelopp une fistule pancratique qui a t traite avec succs mdicalement. Le follow-up moyen est de 35 mois. Les cinq patients sont toujours en vie et ont un bon rsultat fonctionnel.
En cas de F. A. P., la rsection pancraticoduodnale prophylactique peut tre indique lorsquune polypose duodnale est prsente. Les stades III et IV de la classification
de Spigelman, un adnome pri-ampullaire, plus de 40 ans
dge et une histoire familiale de cancer du duodnum sont
des facteurs qui peuvent conduire la dcision de raliser
une rsection pancratico-duodnale prophylactique.

Introduction

Familial adenomatous polyposis (FAP) is a premalignant

condition characterized by numerous colorectal adenomas.
In addition, patients with FAP frequently develop upper
gastrointestinal tract adenomas, particularly in the duodenum [1 13]. As with colorectal polyps, duodenal adenomas are premalignant lesions [14]. Upper gastrointestinal
endoscopy is required to diagnose and treat duodenal lesions at early stage [4, 6, 8, 15 23], as once invasive cancer has developed, prognosis is poor [23].
We report five cases of prophylactic pancreaticoduodenal resection (PDR) performed in patients with FAP for
nonmalignant duodenal polyposis.

Patients and methods

Five women with a past history of FAP have undergone PDR for duodenal polyposis since 1991. Patient characteristics are reported in
Table 1.

40
Table 1 Patient characteristics: age (years) at diagnosis of different stages of the disease and intestinal status when PDR was performed. (IRA Ileo-rectal anastomosis, IAA Ileo-anal anastomosis)
Patient
no.

FAP

Duodenal
disease

PDR

Intestinal
status

1
2
3
4
5

16
40
19
21
28

38
52
27
38
33

38
53
27
38
36

Ileostomy
Ileostomy
IRA
IRA
IAA

Table 2 Duodenal polyposis: Spigelmans classification. (Stage I

score 1 4, Stage II score 5, 6, Stage III score 7, 8, Stage IV score
9 12)

Adenomas
Size (mm)
Histology
Dysplasia

14
14
Tubulous
Mild

5 20
5 10
Tubulovillous
Moderate

>20
>10
Villous
Severe

Duodenal polyposis was diagnosed by routine gastroduodenoscopy in three patients; another patient presented with a history of
gastroesophageal reflux and another with an episode of cholangitis.
Gastroduodenoscopy was performed with forward and side-viewing
endoscopes. The endoscopic appearance of the papilla was noted,
and biopsies were taken systematically from the polyps. The histopathological examination was performed by the same pathologist.
To assess the severity of the duodenal polyposis the classification of
Spigelman was used (Table 2), and endoscopic aspect of the papilla was noted (Table 3).
Before PDR the duodenum was inspected directly two patients
by duodenotomy. In one of them enteroscopy was used to identify
possible adenomas in the small intestine. Resection of the first jejunal loop was performed in four cases because of distal duodenal
involvement. The reconstruction, according to Child, started with
the pancreatic anastomosis. In all cases pancreatic tissue was friable, and the diameter of the Wirsung duct was small (2 3 mm) in
all cases but one. Wirsungojejunostomy was performed in four cases and was intubated by a unexteriorised silastic catheter. Two anterior and posterior layers secured the pancreatic section to the jejunal serosa. The fifth patient, in whom no Wirsung duct was seen,

Table 3 Duodenal polyposis

(endoscopic and histological
feature) and stage (Spigelmans classification)

Table 4 Histopathological
examination of the specimens
and stage (Spigelmans
classification)

underwent a terminolateral pancreato-jejunostomy. The common

bile duct was normal in diameter except in one patient in whom it
was enlarged. A terminolateral cholangio-jejunostomy was performed in all cases. The pylorus was retained in all patients, and digestive tract continuity was restored by a terminolateral duodenojejunostomy.

Results

Histopathological examination of the specimens confirmed the absence of cancer and revealed a more severe
polyposis (as regards number of polyps, size, and dysplasia) than expected preoperatively. Four patients were classified stage IV postoperatively compared to only two preoperatively (Tables 3, 4).
No mortality was observed. The postoperative course
was uneventful in three patients. In one it was complicated
by a delay in bowel transit that required nasogastric suction for 16 days. A pancreatic fistula, with a daily output
up to 500 ml, occurred in one patient; this was successfully
managed by suction, inhibition of pancreatic secretion
(Sandostatine for 23 days), and total parenteral nutrition.
The five patients were discharged respectively on the 13th,
14th, 14th, 28th, and 66th postoperative days.
Late outcome
The mean follow-up of the five patients was 35 months (respectively 2, 19, 42, 42, and 69 months). All five patients
are still alive and have undergone regular endoscopy of the
upper digestive tract; this has also been performed of the
retained rectum in two patients. After PDR no gastric adenomas were observed. On the other hand, the rectum displayed recurrent polyposis which was dealt with either by
endoscopic electrocoagulation (in one patient) or by proctectomy and a J ileal pouch (in another).
Table 5 shows the functional digestive assessment, need
for special died and weight before and after PDR.

Patient
no.

Duodenal
adenomas (n)

Size
(mm)

Papilla
aspect

Histology

Dysplasia

Stage

1
2
3
4
5

3
6
3
3
>20

>10
>10
>10
>10
>10

Abnormal
?
Abnormal
Abnormal
Abnormal

Tubulovillous
Tubulovillous
Tubulovillous
Tubulovillous
Tubulovillous

Moderate
Moderate
Moderate
Moderate
Severe

III
IV
III
III
IV

Patient

Adenomas
(n)

Max. size
(mm)

Papilla
aspect

Ampulla
localization

Histology

Dysplasia

Stage

1
2
3
4
5

6
3
5
15
>20

40
18
30
30
70

Tumor
Normal
Tumor
Tumor
Tumor

Yes
No
Yes
Yes
Yes

Tubulovillous
Tubulovillous
Tubulovillous
Tubulovillous
Tubulovillous

Moderate
Moderate
Severe
Moderate
Severe

IV
III
IV
IV
IV

41
Table 5 Functional digestive
assessment, requirements for
special diet and weight, before
and after PDR

Patient

1
2
3
4
5

Bowel movement
(n/24 h)

Ileostomy output
(l/24 h)

Special diet

Before

Before

After

Before

After

Before

After

<1
<1

<1
<1

No
No
No
No
No

No
No
Yes
No
No

51
65
66
44
50

51
58
69
45
44

1, 2
13
4, 5

After

1, 2
13
5, 6

Comments

Many rationales have been put forth for PDR in cases

of duodenal polyposis in patients with FAP.

Weight
(kg)

under evaluation [27]. Endoscopic methods appear generally to be restricted to small, few, and easily accessible
lesions.
Polypectomy through duodenotomy has the same limits as endoscopic treatment. Penna et al. [28] studied 12
patients and reported recurrence of duodenal polyposis in
all patients after a mean follow-up of 13 months.

Duodenal polyposis is a precancerous condition

FAP is known to involve the upper digestive tract [1 14].
Recent studies [1, 4, 6, 7] have shown that nine of ten patients with FAP experience duodenal involvement during
the course of the disorder. Gastric lesions consist of glandulocystic polyposis and are not at risk of cancer [3, 4, 6,
7, 10]. In Western countries the risk of gastric cancer is not
higher in FAP patients than in the general population [6,
7, 20]. On the other hand, duodenal polyps are adenomas
and consequently may advance to dysplasia and cancer
[23]. The frequency of duodenal cancer in FAP is 2% [2,
20, 22]. The risk of cancer increases with age and is higher
in cases of familial history of duodenal cancer [24]. The
two main causes of death in patients who have undergone
prophylactic proctocolectomy are currently desmoid tumors and duodenal cancer [15, 19].
When cancer arises in duodenal adenomas, the result of
resection is poor [2, 22]. Beckwith et al. [22] report a mean
a survival of 13 months in 4 patients (three radical pancreatico-duodenectomies and one palliative bypass).
Patients with duodenal involvement require regular endoscopic examination. The periampullary area is at risk of
adenomas [6, 7]. Duodenoscopy with side-viewing endoscope is mandatory as well as systematic biopsies even in
cases with normal macroscopic appearance [3, 16].
Conservative approach is inefficient or unsafe
Nonsteroidal anti-inflammatory drugs such as Sulindac
have no role in treating severe duodenal polyposis [25, 26].
They may slow the evolution to malignancy [26], and any
effect seems to be restricted to polyps less than 2 mm in
diameter [25].
Endoscopy allows electrocoagulation of adenomas,
but the size, site, and number of polyps limit its therapeutic application. Ampullary localization may risk biliary
or pancreatic duct injury, and even if polyps are destroyed, electrocoagulation does not prevent recurrence.
Photodynamic laser, a new endoscopic technique, is still

PDR is an available option

Pancreaticoduodenectomy is an aggressive surgical approach [29] that allows radical resection of the ampullary
area. Although morbidity and mortality of PDR may be acceptable, this procedure is not to be undertaken lightly and
should be restricted to patients with a high risk of cancer
and with lesions preventing full resection by using conservative procedures. The risk of malignancy increases with
severity of polyposis, as assessed by the Spigelmans classification in cases of periampullary lesions, with age, and
with a familial history of duodenal cancer [16, 24]. Preoperative staging of duodenal involvement seems to be
underestimated. In our series three patients in whom polyposis had been initially classified as stage III were reclassified as stage IV postoperatively (Tables 3, 4). Spigelmans classification should not be the only method for assessing severity. The site of polyps is also important since
periampullary adenomas are at higher risk of cancer, and
endoscopic resection is difficult.
In FAP biliary secretion has a promoting effect on adenomatosis, which involves mainly the periampullary area
[30, 31] . Retaining the pylorus may be of relevance; pylorus preservation prevents biliary reflux and may avoid
its harmful impact on gastric mucosa. Furthermore function of the stomach may be less impaired. In the present
series the functional status of the digestive tract was not
impaired after PDR. All patients regained their preoperative weight (Table 4), and no was malnutrition was observed. No patient developed diabetes or signs of pancreatic secretion impairment.

Conclusion

PDR in duodenal polyposis must not be a routine procedure but rather should be reserved for patients at high risk
of cancer. This risk is related to the severity of polyposis

42

(stages III and IV of Spigelmans classification) and to the

periampullary localization. Age above 40 years and family history of duodenal cancer are also to be taken into account and may indicate the need for PDR.

References
1. Blow S, Alm T, Fausa O, Hulterantz R, Jrvinen H, Vasen H
(1995) Duodenal adenomatosis in familial adenomatous polyposis. Int J Colorect Dis 10: 43 46
2. Adedeji OA, Trescoli-Serrano C, Garcia-Zarco M (1995) Primary duodenal carcinoma. Postgrad Med J 71: 354 358
3. Goedde TA, Rodrigues-Bigas MA, Herrera L, Petrelli NJ (1992)
Gastroduodenal polyps in familial adenomatous polyposis. Surg
Oncol 1: 357 361
4. Church JM, McGannon E, Hull-Boiner S, Sivak MV, Van Stolk
R, Jagelman DG, Fazio VW, Oakley JR, Lavery IC, Milsom LW
(1992) Gastroduodenal polyps in patients with familial adenomatous polyposis. Dis Colon Rectum 35: 1170 1173
5. Franois Y, Durous E, Bonvoisin S, Descos L, Vignal J (1992)
Atteinte gastroduodnale au cours de la polypose adnoma-teuse
familiale. Rev Fran gastro-entrol 282: 263 267
6. Domizio P, Tablot IC, Spigelman AD, Williams CB, Phillips
RKS (1990) Upper gastrointestinal pathology in familial adenomatous polyposis: result from a prospective study of 102 patients. J Clin Pathol 43: 738 743
7. Spigelman AD, Williams CB, Talbot IC, Domizio P, Phillips
RKS (1989) Uper gastrointestinal cancer in familial adenomatous polyposis. Lancet II: 783 785
8. Alexander JR, Andrews JM, Buchi KN, Lee RG, Becker JM,
Burt RW (1989) High prevalence of adenomatous polyps of the
duodenal papilla in familial adenomatous polyposis. Dig Dis Sci
34: 167 170
9. jerskog B, Myrvold HE, Nilsson LO, Philipson BM, Ahrn C
(1987) Gastroduodenal and ileal polyps in patients treated surgically for familial polyposis coli with proctocolectomy and continent ileostomy. Acta Chir Scand 153: 681 685
10. Sarre RG, Frost AG, Jagelman DG, Petras RE, Sivak MV, Mc
Gannon E (1987) Gastric and duodenal polyps in familial adenomatous polyposis: a prospective study of the nature and prevalence of upper gastrointestinal polyps. Gut 28: 306 314
11. Tonelli F, Nardi F, Bechi P, Taddei G, Gozzo P, Romagnoli P
(1985) Extra colonic polyps in familial polyposis coli and
Gardners syndrome. Dis Colon Rectum 28: 664 668
12. Jrvinen H, Nyberg M, Peltokallio P (1983) Upper gastrointestinal tract polyps in familial polyposis coli. Gut 24: 333 339
13. Ranzi T, Castagnone D, Velio P, Bianchi P, Polli EE (1981)
Gastric and duodenal polyps in familial polyposis coli. Gut
22: 363 367
14. Spigelman AD, Talbot IC, Penna G, Nugent KP, Phillips RKS,
Costello C, DeCosse JJ (1994) Evidence for adenoma-carcinoma sequence in the duodenum of patients with familial adenomatous polyposis. J Clin Pathol 47: 709 710
15. Yao T, Iida M, Ohsato K, Watanabe H, Omae T (1977) Duodenal lesion in familial polyposis of the colon. Gastroenterology
73: 1086 1092

16. Nugent KP, Spigelman AD, Phillips RKS (1996) Risk of extracolonic cancer in familial adenomatous polyposis. Br J Surg
83: 1121 1122
17. Debinski HS, Spigelman AD, Hatfield A, Williams CB, Phillips
RKS (1995) Upper intestinal surveillance in familial adenomatous polyposis. Eur J Cancer 31A: 1149 1153
18. Sawada T, Muto T (1995) Familial adenomatous polyposis:
should patients undergo surveillance of the upper gastrointestinal tract. Endoscopy 27: 6 11
19. Nugent KP, Spigelman AD, Williams CB, Talbot IC, Phillips
RKS (1994) Surveillance of duodenal polyps in familial adenomatous polyposis: progress report. J R Soc Med 87: 704 706
20. Nugent KP, Spigelman AD, Phillips RKS (1993) Life expectancy after colectomy and ileorectal anastomosis for familial adenomatous polyposis. Dis Colon Rectum 36: 1059 1062
21. Offerhaus GJ, Giardiello FM, Krush AJ, Booker SV, Tersmette
AC, Kelley NC, Hamilton SR (1992) The risk of upper gastrointestinal cancer in familial adenomatous polyposis. Gastroenterology 102: 1980 1982
22. Noda Y, Watanabe H, Iida M, Narisawa R, Kurosaki I, Iwafuchi M, Satoh M, Ajioka Y (1992) Histologic follow-up of ampullary adenomas in patients with familial adenomatosis coli.
Cancer 70: 1847 1856
23. Beckwith PS, Van-Heerden JA, Dozois RR (1991) Prognosis of
symptomatic duodenal adenomas in familial adenomatous polyposis. Arch Surg 126: 825 827
24. Sanabria JR, Croxford R, Berck TC, Cohen Z, Bapat BV, Gallinger S (1996) Familial segregation in the occurrence and severity of periampullary neoplasms in familial adenomatous
polyposis. Am J Surg 171: 136 140
24. Debinski HS, Trojan J, Nugent KP, Spigelman AD, Phillips RKS
(1995) Effect of sulindac on small polyps in familial adenomatous polyposis. Lancet 345: 855 856
26. Nugent KP, Farmer KCR, Spigelman AD, Williams CB, Phillips RKS (1993) Randomized controlled trial of the effect of
sulindac on duodenal and rectal polyposis and cell proliferation
in patients with familial adenomatous polyposis. Br J Surg
80: 1618 1619
27. Mlkvy P, Messmann H, Debinski H, Regula J, Conio M,
MacRobert A, Spigelman AD, Phillips RKS, Bown SG (1995)
Photodynamic therapy in familial adenomatous polyposis: a
pilot study. Eur J Cancer 31A: 1160 1165
28. Penna C, Phillips RKS, Tiret E, Spigelman AD (1993) Surgical
polypectomy of duodenal adenomas in familial adenomatous
polyposis: experience of two European centres. Br J Surg 80:
1027 1029
29. Balladur P, Penna C, Tiret E, Vaillant JC, Gailleton R, Parc R
(1993) Pancreatico-duodenectomy for cancer and precancer in
familial adenomatous polyposis. Int J Colorectal Dis 8: 151 153
30. Scates DK, Venitt S, Phillips RKS, Spigelman AD (1995) High
pH reduces DNA damage caused by bile from patients with familial adenomatous polyposis. Gut 71: 354 358
31. Spigelman AD, Owen RW, Hill MJ, Phillips RKS (1991)
Biliary bile acid profiles in familial adenomatous polyposis.
Br J Surg 78: 321 325
32. Chung RS, Church JM, van Stolk R (1995) Pancreas-sparing
duodenectomy: indications, surgical technic and results. Surgery
117: 254 259

Int J Colorect Dis (1998) 13: 43 47

Springer-Verlag 1998

O R I G I N A L A RT I C L E

L. Khler D. Rixen H. Troidl

Laparoscopic colorectal resection for diverticulitis

Accepted: 10 October 1997

Abstract This study evaluated outcome in patients

undergoing laparoscopically assisted sigmoid resection
for diverticular disease. A total of 29 consecutive patients
were treated surgically for colonic diverticulitis; in 27 of
these laparoscopy was performed. The review of medical
records from a control group of 34 patients undergoing
open resection were used for comparison. The conversion
rate was 7.5%. Using the laparoscopic technique the duration of surgery was longer (165 vs. 121 min, P<0.05),
blood loss less (182 vs. 352 ml, P<0.05), and subsequent
blood transfusion less (0 vs. 61%). The incidence of complications following laparoscopic resection was lower
(two anastomotic leakages, two wound infections) than in
the conventional group. Convalescence in the laparoscopic group was more rapid and hospital stay shorter
(7.9 vs. 14.3 days, P<0.05). In the laparoscopic group patients expressed less pain at rest and in motion. The cost
of the laparoscopically assisted procedure was less than
that of conventional resection (7185 vs. 8975 DM). In this
series laparoscopically assisted sigmoid resection for diverticulitis proved safe. Recovery was faster, hospital stay
was shorter, and patients expressed less pain than in conventional open surgery.

de comparaison. Le taux de conversion est de 7,5%. Par

laparoscopie, la dure de lintervention chirurgicale a t
plus longue (165 vs 121 minutes, P<0,05), les pertes sanguines (182 vs 352 ml, P<0,05) et les transfusions sanguines moindres (0 vs 16%). Lincidence des complications
suivant la laparoscopie a t infrieure (2 fuites anastomotiques, 2 infections de plaie) en comparaison la chirurgie conventionnelle. La convalescence du groupe laparoscopique a t plus rapide et la dure dhospitalisation
rduite (7,9 vs 14,3 jours, P<0,05). Dans le groupe laparascopique, les douleurs au repos et aux mouvements ont
t infrieures. Let cot du geste chirurgical laparoscopique a t infrieur en comparaison la rsection conventionelle (7185 vs 8975 DM). Dans cette srie, la rsection
sigmodienne par laparoscopie pour maladie diverticulaire
est sre et plus rapide, la dure de gurison est plus rapide,
la dure dhospitalisation plus brve et les patients ont prsent moins de douleurs dans le groupe laparoscopique en
comparaison avec la chirurgie conventionelle.

Key words Diverticular disease Laparoscopically

assisted sigmoid resection Open sigmoid resection
Recovery Morbidity

Colonic diverticulitis is a common disease in Western

countries. Medical therapy is indicated for the first episode
of colonic diverticulitis, except in the young. Sigmoid resection may be considered following the first episode in
young patients, when complications arise, and after second episodes in elderly patients [1 4]. The role of laparoscopic surgery in diverticulitis is still undetermined. Despite its potential advantages multiple factors have impeded the universal adoption of the laparoscopic approach
for treating patients with colorectal disease [5].
Here we report our early experience with the laparoscopic approach for diverticular disease and place this in
context of experience with the conventional approach.

Rsum Le but de cette tude est dvaluer le devenir des

patients ayant subi une sigmodectomie laparoscopique
pour maladie diverticulaire. Vingt-neuf patients conscutifs ont t traits chirurgicalement pour une diverticulite
colique. Vingt-sept des cas ont pu tre oprs par laparoscopie, Ltude des dossiers cliniques dun groupe-contrle de 34 patients subissant une rsection ouverte a servi

L. Khler () D. Rixen H. Troidl

II Department of Surgery, University of Cologne,
Ostmerheimerstrasse 200, D-51109 Cologne, Germany

Introduction

44

Methods and materials

Patients
All patients with colonic diverticulitis who were electively treated
by surgery were followed prospectively for 18 months (November
1995 April 1997). We attempted to operate on all patients by the laparoscopic technique. The indications for operation were published
in 1990 [1] and have not changed following the introduction of laparoscopy. A sigmoid resection was performed if the patient had two
or more attacks of left-sided pain with fever, leukocytosis, and radiological signs of diverticulities or: (a) one attack with obstruction,
(b) one attack with urological signs, (c) one attack under the age of
55 years of age, (d) nonexclusion of carcinoma, and (e) a persistent
diverticular mass.
During the 18-month period 29 consecutive patients (15 males,
14 females) were operated on for diverticular disease. With the exception of two cases all operations were performed by the same surgeon.
Two patients underwent a conventional procedure due to anticipated
severe adhesions after previous abdominal surgery. The remaining
27 patients were operated on laparoscopically; two conversions were
necessary due to adhesions. Thus 25 of the 29 (86%) patients were operated on laparoscopically. The conversion rate was 7.5%.
The mean age was 56.6 years (range 38 84). Mean body mass
index (BMI) was 27.9 kg/m2 (range: 21.7 54.2). The ASA classification was as follows: 50% II, 45% III, 5% IV. Twelve patients had
previous abdominal surgery (e. g. cholecystectomy, appendectomy,
gastrectomy).
Medical records of 34 patients operated on for colonic diverticulitis before the laparoscopic technique was introduced in our clinic were used for comparison.
The two groups were comparable for age, gender, BMI, and history of prior abdominal surgery. The indication for operation did not
differ during the two time periods.
Operative technique
After establishing a pneumoperitoneum the optic trocar was inserted 1 cm above the umbilicus in the midline. Another 10-mm trocar
was positioned in the left lower quadrant and two further trocars
(10- and 12-mm) were placed in the lower right quadrant. Complete
mobilization of the sigmoid was performed intra-abdominally, with
a combination of electrocautery and sharp dissection. The left ureter was always identified. The mesentery was divided intracorporeally with a linear cutting stapler. The distal margin of bowel resection
was at the rectosigmoid junction, which was also divided intra-abdominally using a Multifire Endo-Gia (USSC, Norwalk, CT). The
bowel was then extracted through a transverse muscle-splitting incision (4 6 cm) in the lower left quadrant, and proximal resection
was completed. The anvil from a 31-mm Premium Plus CEEA stapler (USSC) was placed extracorporeally after performing a pursestring suture. After reestablishing the pneumoperitoneum the stapler
was introduced peranally, and the anastomosis was completed.
The integrity of the anastomosis was checked with methylene
blue colored water, and the completeness of the resection rings was
examined.
The conventional technique did not differ from the laparoscopic
technique with the exception that 60% of the anastomose were handsewn. All patients received subcutaneous heparin and perioperative
antibiotics.
The following pre-, intra-, and postoperative data were documented.
Preoperative: Age, gender, BMI, ASA classification, previous operations, colonoscopy and X-ray results, blood chemistry
Intraoperative: Name of surgeon, number, and position of trocars,
intra-abdominal pressure, length of minilaparotomy, mobilization
of splenic flexure, estimated blood loss, duration of operation,
technique of vascular control, form of anastomosis, stapler diameter, control of anastomotic integrity, intraoperative complications, drainage, reason for conversion

Postoperative: Length of postoperative nasogastric intubation,

reoperation, transfer to ICU, blood transfusion, duration of abdominal drainage, first oral intake, first bowel movement, wound
infection, bleeding, anastomotic insufficiency, obstruction, intraabdominal abscess, peritonitis, thrombosis, pulmonary embolism, urinary tract infection, pain at rest and in motion by a visual analog scale (VAS) ranging from 0 (no pain) to 10 (extreme
pain), pain medication, hospital stay, and length of resected
bowel.
Statistics
Means were calculated and analyzed for both groups using the Wilcoxon rank-sum test and Fisher exact test. A P value <0.05 was considered significant.

Results

There was no intraoperative death in either group. Table 1

presents the total duration of the operative procedure, estimated intraoperative blood loss, length of resected bowel,
percentage of splenic flexures mobilized and need of conversion. Total operating time was significantly longer in
patients undergoing laparoscopic resection than in those
with open laparotomy [165 min (80 210) vs. 121 min
(75 205), P<0.05]. Blood loss was significantly less in
the laparoscopic group [182 ml (50 600) versus 352 ml
(160 410), P<0.05]. The splenic flexure was mobilized in
Table 1 Comparison of intraoperative data in laparoscopic and
conventional sigmoid resection for diverticular disease
Laparoscopic
resection
(n = 25)
Duration of operation (min)
165 (80 210) *
Estimated blood loss (ml)
182 (50 600) *
Patients transfused (%)

Patients converted (%)

7.5
Splenic flexure mobilized (%) 71
Length of resected bowel (cm) 22.4 (19 28)

Open
resection
(n = 34)
121 (75 205)
352 (160 410)
16

74
24.4 (21 27)

* P < 0.05

Table 2 Postoperative complications following laparoscopic and

conventional sigmoid resection for diverticular disease

Reoperation
Bleeding
Ureter damage
Anastomotic insufficiency
Pneumonia
Urinary tract infection
Thrombosis
Wound infection
a

Laparoscopic
resection
(n = 25)

Open
resection
(n = 34)

2/25 a

2/25

2/34 a
4/34
6/34
2/34
7/34

Clinically symptomatic leak successfully treated by i. v. antibiotics and parenteral nutrition, no reoperation required

45
Table 3 Postoperative course following laparoscopic and conventional sigmoid resection for diverticular disease
Laparoscopic
resection
(n = 25)

Open
resection
(n = 34)

First bowel movement

(days after surgery)

3.7 *
(3 5)

5.3
(3 7)

Days until regular diet

4.1 *
(3 10)

5.8
(4 10)

Patients postop on ICU

1/25 *

16/34

Days on ICU (mean)

0.05 *

1.5
(1 14)

Length of postoperative
stay (days)

7.9 *
(6 14)

14.3
(9 23)

* P<0.05

paroscopic group. Two patients developed symptomatic

leakage at the anastomosis proven by colonoscopy; both
were successfully treated by intravenous antibiotics and
parenteral nutrition, and no reoperation was required. Two
patients developed a wound infection of the minilaparotomy site. Following open resection two symptomatic anastomotic leakages occurred. Both were treated conservatively, without the need for reoperation. Four patients developed postoperative pneumonia, six had urinary tract infection, two developed a thrombosis, and seven experienced wound infection at the laparotomy site.
The postoperative course is shown in Table 3. Patients
undergoing laparoscopic bowel resection resumed a regular diet sooner than patients undergoing surgery [4.1
(3 10) vs. 5.8 (4 10), P<0.05] and first bowel movement
occurred earlier [3.7 (3 5) vs. 5.3 (3 7), P<0.05]. One
patient out of 25 was transferred to ICU after laparoscopic
resection while this was necessary in 16 out of 34 patients
having conventional surgery.
The postoperative stay was significantly shorter after
laparoscopic sigmoid resection [7.9 days (6 14) vs. 14.3
days (9 23) after conventional surgery P<0.05].
Patients after laparoscopic sigmoid resection had less
pain at rest and in motion than those of the open resection
group (Fig. 1). The need for and the length of analgesia
was reduced.
The postoperative course of the two converted patients
was without complication. The first bowel movement appeared within 6 days after surgery, and the length of postoperative stay was 10 and 13 days, respectively.

Discussion

Fig. 1 a Post-operative pain at rest following laparoscopic

(A, n = 25) and conventional ( n = 34) sigmoid resection for diverticular disease. b Post-operative pain in motion following laparoscopic (A, n = 25) and conventional ( n = 34) sigmoid resection
for diverticular disease

71% and 74% of laparoscopic and nonlaparoscopic patients. The length of the resected bowel did not differ
between the two techniques.
Postoperative complications in each of the treatment
groups are shown in Table 2. Four were recorded in the la-

Laparoscopic techniques should be assessed critically. Frazier and Mosteller [6] emphasize such outcomes of treatment as morbidity, stabilization or restoration of function,
pain, the financial impact of treatment, applicability of the
technique, patient satisfaction, overall quality of life, and
unwanted side effects related to the treatment itself. The
outcome of patients undergoing laparoscopic sigmoid resection for diverticular disease is not well documented in
the literature, however. Only a few publications [7, 8] have
examined at laparoscopic treatment of diverticular disease
in depth. Few series address morbidity, mortality, outcome,
or cost specifically.
A substantial selection bias is found in all laparoscopic
series. Ramos and coworkers [9] considered 200 patients
with various colorectal diseases and concluded that laparoscopic surgery was inappropriate in 105, who were operated on by standard surgery. Laparoscopic surgery, was attempted in 95 cases and was successful in 62, with convertion in 33. Thus only 62 of 200 patients (31%) were suitable for laparoscopic surgery. Until now no precise data on
laparoscopy for diverticular disease have been published.
In our own series we were able to perform a laparoscopic sigmoid resection in 25 of 29 cases (86%). More experience, better instruments, and more training may in-

46

crease the number of procedures performed laparoscopically.

Due to different selection criteria and variable technical skill of the surgeons converted conversion rates vary
worldwide between 5% and 34%.
Only few reports have detailed the course of patients
undergoing converted laparoscopic procedures. While we
observed no increase in complications in the small converted group, Slim et al. [10] report a higher postoperative
morbidity rate (50% vs. 21%) and more anastomotic leakages (25% vs. 8%) in the converted group. Careful preoperative patient selection and rapid decision to convert in
cases of difficulty are thus important. However, we must
remember that the converted group is also a highly selected
subgroup.
The overall postoperative complication rates including
mortality following laparoscopic colorectal resection are
lower than those with conventional techniques. Ortega
et al. [11] reported a complication rate of 13%, compared
to one of 29% following conventional resection. In our series of 25 patients the overall complication rate was less
following laparoscopic resection. There was a reduction
not only in the complication rate but also in the severity of
observed complications, for example, the consequences of
an infected median laparotomy with the need to reopen the
wound were not comparable to those of an infected reopened trocar site. Furthermore, long-term complication
rates, for example, for incisional hernias are likely to be
lower after laparoscopic surgery.
Laparoscopic sigmoid resection took about 45 min
longer in our hands. More experience and better instruments may reduce the duration of surgery. Longer operative time does not, however, automatically mean a detrimental effect for the patient. Blood loss, the amount of narcotics required, and number of patients needed to be transfered to the ICU was reduced in our setting.
The major advantage of laparoscopic surgery is the
more rapid convalescence of patients. The passing flatus
and stool occurs earlier, oral intake can be reinstated earlier, and patients are more mobile following laparoscopic
resection [12 14]. Postoperative pain at rest and in motion is significantly less after laparoscopic resection than
with conventional resection. However, placebo effects may
result in high rates of good outcome that are beyond any
specific treatment effect. The true reason for diminished
pain following the laparoscopic technique remains unknown in the absence of randomized controlled trials [15].
This study showed unequivocally that postoperative hospital stay is reduced following laparoscopic colorectal resection. Others report similar results [9, 16, 17].
The cost of health care delivery has become an important issue. One of the potential advantages of laparoscopic
surgery is shorter hospitalization. This is offset by the increased cost of disposable instruments and longer operating time. Wissing et al. [18] have demonstrated that median laparotomy is followed by a 10% infection rate, 15%
incidence of incisional hernia, and 3% wound dehiscence
rate. A prolonged hospital stay due to wound infection
would cost about 500 DM per day at our hospital. Some of

Table 4 Cost of laparoscopic and conventional sigmoid resection

(from [20])

Operative time
Instruments
Clinic costs
(surgeon, nurse)
Hotel costs
Wound morbidity
Time off work
Total costs

Conventional
resection

Laparoscopic
resection

120 min = 600 DM

100 DM
14 days = 3325 DM

165 min = 825 DM

2000 DM
8 days = 1800 DM

14 days = 1750 DM
1000 DM
20 days = 2200 DM
8975 DM

8 days = 960 DM

15 days = 1600 DM
7185 DM

the incisional hernias subsequently require surgery and relaparotomy for wound dehiscence, which is also expensive. This increased wound morbidity would cost about
100,000 DM per 100 patients operated on. Thus 1000 DM
would be saved per patient if patients underwent laparoscopic procedure [19]. In calculating the cost of a laparoscopic sigmoid resection, this wound morbidity must be
considered (Table 4). Liberman [7] and Bruce [8] published comparable data. However, lost-benefit and cost-effectiveness analyses for laparoscopic and conventional
techniques are still missing

References
1. Mennigen R, Vestweber KH, Ure B, Troidl H (1990) Kolondivertikulitis: Wann operieren? Wie operieren? Medwelt 41:
721 725
2. Corman ML (1983) Colon and rectal surgery, 3rd edn. Lippincott, Philadelphia
3. Rothenberger D, Waltz O (1993) Surgery for complicated diverticulities. Surg Clin North Am 73: 975 992
4. Schauer P, Ramos R, Ghiatas A (1992) Virulent diverticular disease in young obese men. Am J Surg 164: 443 451
5. Forde KA, Hultn L (1996) Laparoscopy in colorectal surgery.
Surg Endosc 10: 1039 1040
6. Frazier HS, Mosteller F (1995) Medicine worth paying for. Assessing Medical Innovations. Havard University Press, Cambridge
7. Liberman MA, Phillips EH, Carroll BJ, Fallas M, Rosenthal R
(1996) Laparoscopic colectomy vs traditional colectomy for diverticulitis. Surg Endosc 10: 15 18
8. Bruce CJ, Coller JA, Murray JJ, Schoetz DJ, Roberts PL, Rusin
LC (1996) Laparoscopic resection for diverticular disease. Dis
Colon Rectum 39: S1 S6
9. Ramos JM, Beart RW, Goes R, Ortega AE, Schlinkert RT (1995)
Role of laparoscopy in colorectal surgery. Dis Colon Rectum
38: 494 501
10. Slim K, Pezet D, Riff Y, Clark E, Chipponi J (1995) High morbidity rate after converted laparoscopic colorectal surgery. Br J
Surg 82: 1406 1408
11. Ortega AE, Beart RW, Steele GD (1995) Laparoscopic bowel
surgery registry. Dis Colon Rectum 38: 681 686
12. Begos DG, Arsenault J, Ballantayne GH (1996) Laparoscopic
colon and rectal surgery at a VA hospital. Surg Endosc
10: 1050 1056
13. Lacy AM, Garcia-Valdecasas JC, Piqu JM (1995) Short-term
outcome analysis of a randomized study comparing laparoscopic vs. open colectomy for colon cancer. Surg Endosc 9:
1101 1105

47
14. Tate JJT, Kwok S, Dawson JW, Lau WY, Li AKC (1993) Prospective comparison of laparoscopic and conventional anterior
resection. Br J Surg 80: 1396 1398
15. Turner JA, Deyo RA, Loeser JD, Von Korff M, Fordyce WE
(1994) The importance of placebo effects in pain treatment and
research. JAMA 271: 1609 1614
16. Gellmann L, Salky B, Edye M (1996) Laparoscopic assisted colectomy. Surg Endosc 10: 1041 1044
17. Wexner SD, Reissmann P, Pfeifer J, Bernstein M, Geron N
(1996) Laparosocopic colorectal surgery. Surg Endosc 10:
133 136

18. Wissing J, van Vroonhoven TJMV, Eeftinckl-Schattenkerk M,

Veen HF, Ponsen RJG, Jeekel J (1987) Fascia closure after midline laparotomy: results of a randomized trial. Br J Surg
74: 738 741
19. Eypasch E, Khler L, Troidl H (1995) Viszerosynthesen bei
laparoskopischen Eingriffen Konzepte und Kosten-NutzenRelation, Kongressband, Langenbeck
20. Lefering R, Troidl H, Ure BM (1994) Entscheiden die Kosten?
Chirurg 65: 317 325

Int J Colorect Dis (1998) 13: 4851

Springer-Verlag 1998

O R I G I N A L A RT I C L E

I. N. Nomikos

Necrotizing perineal infections (Fourniers disease):

old remedies for an old disease

Accepted: 29 October 1997

Abstract Perineoscrotal gangrene (Fourniers disease, a

specific type of necrotizing fasciitis) is a rare and potentially fatal clinical entity characterized by progressive
spread of necrosis in the skin and subcutaneous tissues
combined with severe systemic sepsis. We analyzed retrospectively seven patients with perineoscrotal gangrene, illustrating the various clinical presentations, problems in
management, and specific approaches to therapy. On admission all patients were evaluated as being in critical condition, having severe systemic disease and rapidly advancing gangrene. After hemodynamic stability was achieved
broad spectrum antibiotics were started, and aggressive
surgical treatment initiated. In all the patients the infection
originated from the anorectal area and proved to be polymicrobial in nature. Six patients survived and were able to
be discharged from hospital there was one death on the fifth
postoperative day due to pulmonary embolism. The crucial points in the management of this infection remains
early diagnosis, wide and repeated surgical dbridements,
and appropriate antibiotic therapy.
Key words Fourniers disease Perineoscrotal soft
tissues infection
Rsum La gangrne prino-scrotale (maladie de Fournier fascite ncrosante spcifique) est une affection rare
et potentiellement fatale caractrise par une extension
progressive de la ncrose la peau et au tissu sous-cutan
combin avec des signes septiques systmiques. Nous
avons analys rtrospectivement sept patients porteurs
dune gangrne prinoscrotale qui illustrent les diffrentes prsentations cliniques, les problmes dans la gestion
et lapproche spcifique du traitement. A ladmission, tous
les patients ont t jugs comme tant dans un tat critique,
prsentant une maladie systmique svre et une gangrne
rapidement volutive. Aprs stabilisation hmo-dynamique, une antibiothrapie large spectre a t dbute et un
traitement chirurgical agressif instaur. Rsultats: dans
I. N. Nomikos ()
55 Psarron str. Koridallos, Piraeus, GR-18120 Greece

tous les cas linfection dbutant dans la rgion ano-rectale

a prouv une origine polymicrobienne. Six patients ont survcu et ont pu quitter lhpital; un patient est dcd au 5e
jour post-opratoire dune embolie pulmonaire. Conclusion: les points cruciaux dans la prise en charge de linfection restent un diagnostic prcoce, un dbridement chirurgical extensif et rpt et une antibiothrapie adquate.

Introduction

Perineoscrotal gangrene is a rare, rapidly progressive and

potentially fatal infection of the anorectal, perineal, and
scrotal regions, secondary to anorectal and/or periourethral pathology. This syndrome was first described by
Fournier in 1883 as an unexplained gangrene of the perineum; since then several cases have been reported in the
world literature [11]. Although the local findings are pathognomonic of the disease, its presentation as systemic
illness or as an acute abdomen may be misleading. If not
recognized and treated early in its course, this infection
may be fatal.
We report our experience in treating seven patients with
Fourniers disease, illustrating the various clinical presentations and specific approaches to treatment.

Material and methods

Seven patients with perineal soft tissue necrosis have been treated
by the author over a 13-year period (1981 1994). In all cases the diagnosis was made on the basis of fulminating progression of the infection from the perineum to the scrotum and lower abdominal wall
and to the labia majora (in the two female patients of the present series). The local extension of the disease and the perioperative course
of the patients are presented in Table 1. Five patients had signs of
systemic sepsis upon admission. Soon after initial resuscitation and
antibiotic coverage all patients were submitted to surgical treatment.
The surgical regimen consisted of unroofing of all infected areas, drainage of pus, dbridement of necrotic tissues, and establishing a wide communication between the ischiorectal fossae and the

49
Table 1 Data on patients with Fourniers disease. (IRF Ischiorectal fossa, FIR femoro-inguinal region, SCR scrotum, LAW lower abdominal wall, LM labia majora, PRS pararectal space)
Patients

Sex

Age
(years)

Predisposing
factors

Septic
shock
on admission

Extent of soft
tissue necrosis

Postoperative
complications

Colostomy

Perineal
reconstruction

Outcome

29

Obesity,
delayed hosp.
transfer

Yes

IRF, FIR
SCR

Septicemia

No

Surgical

Survived

67

Delayed hosp.
transfer

Yes

IRF, SCR
LAW

Liver dysfunction,
paralytic ileus,
temporary anal
incontinence

No

Surgical

Survived

77

Latent rectal
perforation

No

IRF, LM

Paralytic ileus,
pulmonary
embolism

Yes

49

Obesity,
delayed hosp.
transfer

Yes

IRF, PRS
SCR

Cecum perforation,
gangrenous
cholecystitis

No

Surgical

Survived

57

Diabetes,
delayed hosp.
transfer

Yes

IRF, PRS
LM

Septicemia,
temporary anal
incontinence

Yes

Spontaneous

Survived

42

Unknown

Yes

IRF, FIR, SCR

LAW

No

Spontaneous

Survived

57

Unknown

No

IRF, SCR

No

Surgical

Survived

Died

scrotum or the space underlying the labia majora and/or the lower
abdominal wall. The wounds were packed open, and the dressings
were changed at least twice daily. During each dressing replacement
course the wound was irrigated with normal saline (or tap water, if
the patient was able to walk to the bathroom). Depending on the extent of the perineal or perineoscrotal wound either spontaneous healing by secondary intention or reconstructive surgical procedure was
attempted after the appearance of healthy granulation tissue at the
base of the wound.

Results

There were five men and two women, ranging in age from
29 to 77 years (mean 54). A factor predisposing to fulminant progression of perineal gangrene was found in all
but two patients. A delay in initial admission to hospital and
consequent surgical intervention were considered in four
patients as a determining factor. In contrast to the mean
interval between initial symptoms and diagnosis of 6.3 days
(range 3 11), for these four patients (patients 1, 2, 4, and
5 of Table 1) this was 9.2 days. Two of them were obese,
and one suffered from insulin-dependent diabetes. One patient developed perineal gangrene due to accidental rectal
perforation made during a cathartic enema (Fig. 1).
In all patients the infection originated from the anorectal area and proved to be polymicrobial in nature (Table 2).
Broad-spectrum antibiotics active against gram-positive
cocci, facultative anaerobic gram-negative rods, and anaerobic organisms were administered immediately after
samples of pus or tissue were taken for culture. The patients received intravenous antibiotic therapy for a mean
period of 11.5 days. The most common combination used

Fig. 1 Edema, redness, and necrosis of overlying skin were the local findings in a 77-year-old woman with Fourniers disease after accidental rectal perforation during a cathartic enema

was high-dose penicillin or suitable analogue, an aminoglycoside, and metronidazole.

Five patients were mentally confused on admission,
and as their septic process subsided, their neurological
status progressively improved. Fecal diversion was

50
Table 2 Bacteriology of patients with Fourniers gangrene (n = 7)
n
Aerobes
E. coli
S. faecalis
Proteus mirabilis
Klebsiella pneumoniae
S. epidermidis

5
3
3
2
1

Anaerobes
Bacteroides fragilis
Bacteroides melaninogenicous
Gram positive cocci
Clostridium spp.

6
3
3
2

deemed necessary in two patients. Temporary anal incontinence appeared in two patients. Two patients developed
long-lasting adynamic ileus. One patient underwent a right
hemicolectomy and 10 days later had a cholecystectomy
for gangrenous alithiasic cholecystitis. One patient developed hepatic dysfunction during his septic syndrome.
A spontaneous closure of the wound was achieved only
in two patients (patients 5 and 6 of Table 1). No reconstructive procedure was undertaken in the others, but a delayed
primary closure of the scrotal and perineal skin was possible without tension after the appearance of granulation
tissue at the base of the wound. There was one death due
to pulmonary embolism. The remaining six patients survived the septic process and were discharged from hospial
(mortality 14.2%). During the follow-up period (from
6 months to 12 years) only one of the survivors developed
fisulae in ano 4 months after his initial management.
Table 1 presents the basic data of our patients with
Fourniers disease.

Discussion

Fourniers gangrene is a potentially fatal soft tissue infection of the perineum that rapidly invades the surrounding
tissues. Alcoholism, diabetes mellitus, and obesity have
been cited as main predisposing factors. However, delay
in seeking medical treatment seemed to be a major factor
in our series. In aged, immunocompromized, or debilitated
patients the disease may be silent, whereas the clinical picture of septic syndrome predominates in younger adults.
The potential causes that are directly related to the disease
can be classified into two groups: (a) infections of cutaneous, urogenital, or rectal origin and (b) complications of
surgical procedures such as hemorrhoidectomy, orchiectomy, herniorrhapy, or vasectomy [1, 2]. No apparent cause
is detected in 30% of cases. Although the disease predominates in males, its appearance in two women in this
series may be due to an accidental event (rectal perforation by a cathartic enema) in one patient and in the other
to delay in hospital admission (9 days after the onset of
symptoms) and the coexistence of insulin-dependent diabetes mellitus which caused an insidious course.

Scarpas fascia necrosis allows for the spread of inflammation to the abdominal wall (as occurred in two of our
patients) [3]. Usually the scrotal skin is affected by the septic process (due to obliterating endarteritis) secondary to
microbial invasion of subcutaneous tissue. In contrast, the
testes are invariably spared of the infection because of their
separate blood supply [1].
Persistent paralytic ileus, as seen in three of our patients,
is an interesting finding not mentioned in the literature.
The etiology of ileus remains obscure; reflex sympathetic
activity mediated by extensive scrotal inflammation
probably plays a role. Elevated levels of circulating catecholamines may also inhibit peristalsis and cause
adynamic ileus, as has been suggested previously [3, 4].
Gram stain is most helpful in demonstrating the characteristic synergism of aerobes (gram-positive cocci and
gram-negative bacilli) and anaerobes. Proper and extensive dbridement of all necrotic tissue is an absolute requirement for a successful outcome. Likewise, sinus tracts
must be fully unroofed. The incision may require extension into the buttocks, highs, perineal region, or inguinal
canal depending upon the areas involved. The wound
should remain open and covered with dressings, which
should be changed two or more times daily. The wound
must be irrigated each time with normal saline or tap water with simultaneous dbridement of all necrotic tissue.
Simultaneous adjuvant supportive treatment is a prerequiste for a successful outcome.
With appropriate treatment, granulation tissue appears
in the base of the wound, and primary wound closure is
possible in most of the cases. Since the scrotal skin is elastic and has the capability of rapid regeneration, large defects in this area can heal spontaneously. However, may
need to be reconstructed surgically by one of the following techniques: delayed primary surgical closure of the
scrotum, skin grafting, implantation of the testes in a subcutaneous abdominal or medial thigh pocket, and rotating
flaps from the inguinal region.
None of our patients received hyperbaric oxygen as an
adjuvant treatment although this has been suggested as an
initial step for the control of local disease. The value of hyperbaric oxygen therapy has been debated in the literature,
with some studies finding a survival benefit from its use
and others finding no benefit. Hyperbaric oxygen could
play a favorable role as an adjuvant modality of treatment
to lower the mortality and morbidity of the disease. Its use
is based mainly on: (a) the presence of anaerobic and
facultative anaerobic bacteria in most of the cases,
(b) the presence of endarteritis and septic vasculitis as a
pathological feature leading to ischemia, and (c) the
marked edema which is due to the inflammatory process
and to venous thrombosis [5]. However, although evidence
for decreased mortality seems convincing when hyperbaric
oxygen is used for clostridial myonecrosis, no survival
benefit has been well demonstrated by a prospective, randomized, controlled trial for hyperbaric oxygen used for
nonclostridial necrotizing soft tissue infections [6].
Despite optimal medical and surgical treatment, mortality rates in Fourniers gangrene range between 8% and

51

40% in many series [1, 4, 7]. The manifestations of the

disease and the involvement of a variety of systems sometimes delay recognition. However, the survival of the patient is possible after a combined local and systemic approach [8]. The successful outcome in six out of seven patients in the present series justifies the use of old fashioned
but aggressive remedies for this old and potentially fatal
disease.

References
1. Lauks SS II (1994) Fourniers gangrene. Surg Clin North Am
74: 1339 1352

2. Viddelear AC, Lycklama Nijeholt GA (1992) Lethal Fourniers

gangrene following vasectomy. J Urol 147: 1613 1614
3. Rudolph R, Saloway M, De Palma RG, Persky L (1975)
Fourniers syndrome: synergistic gangrene of the scrotum. Am J
Surg 129: 591 596
4. Wolach MD, Mac Dermott JP, Stone AR, De Vere White RW
(1989) Treatment and complications of Fourniers disease. Br J
Urol 64: 310 314
5. Eltorai IM, Hart GB, Strauss MB, Montroy R, Juler GL (1986)
The role of hyperbaric oxygen in the management of Fourniers
gangrene. Int Surg 71: 53 58
6. Elliott DC, Kufera JA, Myers Roy AM (1996) Necrotizing soft
tissue infections. Risk factors for mortality and strategies for
management. Ann Surg 224: 672 683
7. Paty R, Smith A (1992) Gangrene and Fourniers gangrene. Urol
Clin North Am 19: 149 162
8. Nomikos IN (1997) Anorectal abscesses: need for accurate anatomical localization of the disease. Clin Anat 10: 239 244

Int J Colorect Dis (1998) 13: 52

Springer-Verlag 1998

L E T T E R TO T H E E D I TO R

Improved determination of terminal pudendal latency by modification

of the standard electrode

Dear Sir,
Electrical stimulation by superficial
electrodes has been established in the
diagnostic workup of pudendal neuropathy [1 3]. The latency of the terminal segment of the pudendal nerve
can be determined by this method
with St Marks pudendal electrode
(Dantec Company, Skovlunde, Denmark). With this standard electrode,
the stimulating electrodes are placed
at the tip of the examining finger and
the recording electrodes at the base.
This arrangement can lead to a methodical error, owing to the fact that
the different and indifferent electrodes are attached to the base of the
finger such that stimulation of the
right pudendal nerve gives a negative
recording of the potential. In contrast,
stimulation of the left pudendal nerve
yields a positive response since the
finger changes sides. In this case, considerable portions of the right external anal sphincter muscle are adding
to this potential. This may explain
why many researchers are measuring
the latency of the right pudendal
nerve only.

We have replaced the stimulating

electrode by a recording electrode by
changing the two outer metal strips in
the adapter. This modification allows
the standard St Marks electrode to be
used for adequate stimulation of both
right and left pudendal nerve shown
by an immediate and negative response from the baseline in each case
(Fig. 1). This is clearer with no inter-

ference from the right external and

sphincter muscle.
For a proper evaluation of bilateral
pudendal neuropathy, stimulation of
both sides can be carried out with the
stimulating electrode positioned over
the muscle. The modification described helps to fulfil this requirement even when using the standard
St Marks pudendal electrode.
W. Jost ()
Aukammallee 33
D-65191 Wiesbaden

B
5

References

B
6

Fig. 1 Track B5: Positive response of the

potential from the baseline after stimulation
of the left pudendal nerve (using the electrode in the usual way). Track B6: Negative
response of the potential from the baseline
by using the adapter

1. Brindley GS (1981) Electroejaculation:

its technique, neurological implications
and uses. J Neurol Neurosurg Psychiatry 44: 918
2. Kiff ES, Swash M (1984) Normal proximal and delayed distal conduction in the
pudendal nerves of patients with idiopathic (neurogenic) faecal incontinence. J Neurol Neurosurg Psychiatry
47: 820823
3. Kiff ES, Swash M (1984) Slowed conduction in the pudendal nerves in idiopathic (neurogenic) faecal incontinence. Br J Surg 71: 614616

Int J Colorect Dis (1998) 13: 53

Springer-Verlag 1998

L E T T E R TO T H E E D I TO R

Internal anal sphincter repair

Sir,
We enjoyed reading the article by
Leroi et al. (Int J Colorect Dis 12:
243245), which clearly demonstrates that the management of anorectal incontinence in patients with
isolated internal anal sphincter injury is difficult. In keeping with the
findings of the above study, we have
disappointing failures with direct
internal anal sphincter repair. By
contrast, our results with anoplasty
[1] which comprise filling the defect
in the anal canal with skin and sub-

cutaneous tissue by raising a flap of

perianal and buttock skin and subcutaneous tissue have been encouraging to date. We believe that this
procedure prevents leakage by allowing the residual internal anal
sphincter to effect complete closure
of the anal canal around the inlaid
soft tissue buttress. Based on our experience, we recommend that the option of anoplasty deserves to be included in the armamentarium for the
management of isolated anal sphincter injury producing anorectal incontinence.

Reference
1. Morgan AR, Patel B, Beynon J, Carr ND
(1997) Surgical management of anorectal incontinence due to internal anal
sphincter deficiency. Br J Surg 84:
226230

Yours sincerely,
F. Abbasakoor () A. R. Morgan
J. Beynon N. D. Carr
Department of Surgery
Singleton Hospital
Swansea SA2 8QA
UK

Int J Colorect Dis (1998) 13: 5455

Springer-Verlag 1998

L E T T E R TO T H E E D I TO R

Pathological features of rectal cancer after preoperative radiochemotherapy

Dear Sir,
We appreciated very much the paper
by Dworak et al. [1] and feel flattered that their classification of rectal cancer regression after preoperative radiotherapy (RT) and chemotherapy (CT) is quite similar to that
which we reported [2] 2 years ago in
this journal for rectal cancer patients
receiving preoperative RT. This
system was adopted by our pathologist (S.A.) strictly following Mandards [3] proposal for staging the response of the esophageal carcinoma
to the preoperative therapy.
To date we have treated 53 patients, with the following results: (a)
a complete pathological disappearance of the tumor in 3 patients (6%)
and (b) a regression greater than 50%
in an other 36 patients. These results
are in agreement with data in the literature [4 8] which report a complete response rate in 9% of patients
(range 5 14%) receiving preoperative RT.
However, the main questions that
we are now faced with differ from
those of Dr. Dworak. We think that it
is important to demonstrate whether:
(a) preoperative RT and CT are able
to decrease local recurrence and distant metastases compared to a control
group, (b) tumor regression after preoperative therapy is associated with
a more favorable outcome, and (c) it
is possible to predict in some way
(e.g., by biological markers) the response to therapy prior to initiating
it. We cannot overemphasize the re-

gression of the tumor mass per se if

the tumor was already considered
suitable for surgical resection at the
clinical presentation (T2 or T3, not
fixed to adjacent structures).
We think that the potential benefit is in achieving sufficient tumor regression to be able to perform a
sphincter-saving operation. This appears to be true even if it is difficult
to demonstrate it. In fact, whereas in
our historical control, patients with
cancer of the low-middle rectum, had
a sphincter-saving procedure in 40%
of cases [9], in our recent experience
with preoperative RT CT a conservative procedure was possible in
about 80%, with a local recurrence
rate (as single failure) of 4% after 3
years.
Our study protocol is now addressed particularly to analyzing the
impact of RT CT on lymph node
status in the awareness that metastatic nodes are one of the most important indicators of a an adverse
prognosis, and that removal of metastatic perirectal nodes is perhaps
better than leaving them [10].
In this regard, we have noted that
in our institute the mean number of
lymph nodes detected in surgical
specimens was 16.8 after preoperative RT + CT, 22 after RT only, and
42.3 in surgical patients without
prior therapy (unpublished data).
Acknowledgement Partially supported
by a grant of the Associazione Italiana Ricerca Contro il Cancro.

F. Bozzetti ()
Department of Digestive Tract Surgery
Istituto Nazionale per lo Studio e la Cura
dei Tumori
Via Venezian, 1
I-20133 Milano, Italy
S. Andreola
Department of Pathology
Istituto Nazionale per lo Studio e la Cura
dei Tumori
Via Venezian, 1
I-20133 Milano, Italy
L. Bertario
Department of Digestive Tract Surgery
Istituto Nazionale per lo Studio e la Cura
dei Tumori
Via Venezian, 1
I-20133 Milano, Italy

References
1. Dworak O, Keilholz L, Hoffmann A
(1997) Pathological features of rectal
cancer after preoperative radiochemotherapy. Int J Colorect Dis 12:1923
2. Bozzetti F, Andreola S, Rossetti C,
Zucali R, Meroni E, Baratti D, Bertario
L, Doci R, Gennari L (1996) Preoperative radiotherapy for resectable cancer
of the middle-distal rectum: its effect on
the primary lesion as determined by
endorectal ultrasound using flexible
echo colonoscope. Int J Colorect Dis
11: 283286
3. Mandard AM, Dalibard F, Mandard JC,
Marnay J, Henry-Amar M, Petiot JF,
Roussel A, Jacob JH, Segol P, Samama
G, Ollivier JM, Bonvalot S, Gignoux M
(1994) Pathologic assessment of tumor
regression after preoperative chemoradiotherapy of esophageal carcinoma.
Cancer 73: 26802686
4. Minsky BD, Cohen AM, Ender WE,
Paty P (1995) Sphincter preservation
with preoperative radiation therapy and
coloanal anastomosis. Int J Radiat
Oncol Biol Phys 31: 553559

55
5. Mohiuddin M, Marks G (1991) High
dose preoperative irradiation for cancer
of the rectum, 19761988. Int J Radiat Oncol Biol Phys 20: 3743
6. Marks G, Motriuddin M, Masoni L
(1993) The reality of radical sphincter
preservation surgery for cancer of the
distal 3 cm of rectum following highdose radiation. Int J Radiat Oncol Biol
Phys 27: 779783

7. Berger C, De Muret A, Garaud P, Chapet S, Bourlier P, Reynaud-Bougnoux

A, Dorval E, De Calan L, Huten N, Le
Floch O, Calais G (1997) Preoperative
radiotherapy (RT) for rectal cancer:
predictive factors of tumor downstaging and residual tumor cell density
(RTCD): prognostic implications. Int J
Radiat Oncol Biol Phys 37: 619627
8. Otmezguine Y, Grimard L, Calitchi E,
Despretz J, Mazeron JJ, LeBourgeois
JP, Pierquin B, Julien M (1989) A new
combined approach in the conservative
management of rectal cancer. Int J Radiat Oncol Biol Phys 17: 639545

9. Bozzetti F, Mariani L, Miceli R, Doci

R, Montalto F, Andreola S, Gennari L
(1996) Cancer of the low and middle
rectum: local and distant recurrences
and survival in 350 radically resected
patients J Surg Oncol 62: 207213
10. Bozzetti F, Gennari L (1995) Local recurrences from rectal cancer: impact of
previous surgery. Tumori 81 [Suppl]:
135140