Key words
ileostomy - ileostomy closure - contrast enema
Introduction
The display of the aboral segment of a double-barrelled ileostomy using a contrast
enema is a conventional radiological examination carried out within the scope of the
preoperative planned surgical ileostomy loop closure [1]. Here, on the one hand, the passage of the contrast medium applied via the aboral
segment is tracked for the assessment of potential pathological changes such as intestinal
stenosis or suture insufficiencies in the region of a loop ileostomy; on the other
hand, a check is made of the continence of liquid contrast media with sphincter resting
pressure and in the Valsalva maneuver. Although several studies have questioned the
usefulness of a routine fluoroscopy examination of the aboral stoma, and postulate
a limitation of the indication to specific issues [2]
[3]
[4]
[5], this examination is performed regularly in the clinical routine. Although guidelines
on pathologies that often require an ileostomy, such as Crohn's disease or the treatment
of colorectal carcinoma suggest ileostomy; however, there is no detailed recommendation
regarding preoperative diagnostics prior to ileostomy closure in the guidelines. For
example the S3 guideline “Diagnosis and therapy of Crohn’s disease” only indicates
that “in general, the desire to avoid a permanent stoma is more important than the
risk of relapse” [6]; the guideline provides no exact indications or contraindications for reversal.
Since, in the context of preoperative preparation for planned reversal, modern imaging
methods, sonography, endoscopy and comprehensive clinical examination, including a
digital rectal examination are performed, this retrospective study will evaluate the
clinical value of a radiological examination of the loop ileostomy using monocontrast
based on a large cohort of a large-scale university hospital. In particular the study
will review the extent to which this method as the sole diagnostic measure actually
influences the surgical procedure in the context of other examinations.
Materials and Methods
Patient Selection
A retrospective analysis was performed of all monocontrast studies carried out at
a clinic of offering greatest extent of care (University Hospital Regensburg) between
January 2005 and February 2011; the study focused on monocontrast to display the diverting
loop of patients with an artificial anus. Compilation of the cohort was based on the
RIS (Radiological Information System; MEDOS 7.42; NEXUS/DIS GmbH Frankfurt/ Main,
Germany) and included all patients who underwent a monocontrast examination during
the above-mentioned time period. Monocontrast studies of non-double-barreled stomata
and premature termination of the examination were excluded.
Examinations
From January 2005 to July 2007 all examinations were carried out on a Sireskop 5/45
type Polydoros 80S (Siemens AG, Erlangen, Germany). Until July 2007a Polystar unit
(Siemens AG, Erlangen, Germany) was available on an interim basis. An AXIOM Artis
dMP device (Siemens AG, Erlangen, Germany) with a flat detector system has been used
since August 2007. Under-table devices were used for all examinations which were carried
out by a physician with a radiological fellowship in conjunction with a specialist
in radiology. A written radiological finding was prepared after the examination and
documented in the RIS.
First, control images were acquired prior to administration of contrast. After probing
the aboral segment of the ileostomy with a 14F bladder catheter, the application of
0.5 to 2 liters of a radiopaque iodine-based hyperosmolar contrast agent (Gastrolux®RE, 180 mg / ml iodine-based solution, Sanochemia Diagnostics, Neuss, Germany). The
contrast medium flow in the colon was shown under fluoroscopy and documented using
conventional radiographs. Colon flexures and the sigma were freely projected in the
usual technique ([Fig. 1]). If a constriction was detected, after ruling out contraindications (cardiac arrhythmia,
glaucoma, prostate hyperplasia for butylscopolamine or pheochromocytoma, hypersensitivity
to glucagon or lactose), 40 mg of butylscopolamine or 1 – 2 mg glucagon was administered.
[Fig. 2] shows an example of a real stenosis also verified by computed tomography. Finally,
continence behavior was examined by fluoroscopy using a Valsalva maneuver.
Fig. 1 After the contrast application using the aboral segment of the ileostomy the complete
colon is distended. The white arrows mark a true stenosis.
Fig. 2 a A true stenosis revealed in the monocontrast colonic examination. b Verification of this true stenosis by CT: the white arrow marks the stenosis in the
same patient within the coronary reconstruction of the pelvis.
Data Collection
Approval of the ethics committee was obtained for the retrospective data analysis.
The patients provided their written consent. Patient-related parameters were extracted
from radiological findings in the RIS as well as the documents in the hospital information
system. These included demographic data, primary diseases and related operations.
In addition, parameters related to the radiological findings of the monocontrast study
were recorded as technical data (dose area product, fluoroscopy time). The radiological
findings documented in writing were examined with regard to the presence of constrictions
in the examined intestinal section. Any intravenous spasmolysis as well as the evaluation
of the constriction after spasmolysis were likewise documented. A constriction was
considered functional if the drug spasmolysis led to distension. If the stenosis did
not open under spasmolysis, the narrowing was evaluated as a real stenosis. In order
to assess the therapeutic consequences of the radiological findings, an analysis was
performed based on medical patient documentation in the hospital information system
to determine whether reversal surgery could be foregone due to the stenosis, or whether
further studies were needed to verify or rule out the stenosis. Rectal continence
or incontinence was classified as follows: 0 – retained continence, 1 – limited to
moderate incontinence, 2 – complete incontinence. Here as well, the existence of a
therapeutic effect based on the documentation in the hospital information system was
analyzed to determine whether reversal surgery was ruled our due to radiologically
documented incontinence of fluid contrast agent. In addition, the presence of anastomotic
leakage was reviewed and the consequences for further therapy analyzed. [Fig. 3] shows an example of anastomotic leakage verified by computed tomography. Adjunctive
pathological changes, e. g. diverticula or fistulas were additionally documented in
the findings. The date of surgery and the time span between the fluoroscopic examination
and reversal operation were documented. Finally, the individual indication or contraindication
for the ileostomy reversal was recorded for each patient.
Fig. 3 Contrast enema in a lateral view a and a CT of the same patients (sagittal reconstruction) after a low anterior resection
after rectum carcinoma. The white arrow marks a contrast depot representing a cavity
after insufficiency of the anastomosis.
The acquired data were grouped qualitatively using a spreadsheet (Excel 2013, Microsoft,
Redmond, WA, USA) and evaluated with descriptive statistics.
Results
Extraction of data from the RIS yielded a total of 364 patients and 413 examinations.
Evaluation of physician reports in the hospital information system (SAP-R/ 3 IS-H/
i.s.h.med) resulted in the cohort exclusion of 84 colon examinations using a rectal
contrast agent application, evidence of 40 Hartmann’s stumps and ten end-stage colostomies
as well as one evidence of fistula. The examination was terminated in four cases and
consequently could not be included in the assessment. The reasons for this were a
premature termination of the examination by the patient, incorrect intubation of the
oral segment of the stoma, and the technical failure of probing as well as pronounced
incontinence resulting in the end of the examination. This resulted in a final cohort
with 252 patients and 274 examinations. The study group was divided into two subgroups
for further evaluation based on physician's letters in the hospital information system:
Group I with 225 (89 %) patients and 244 examinations in which an artificial anus
reversal was performed after images were acquired of the loop ileostomy, and Group
II with 27 patients (11 %) and 30 examinations without subsequent artificial anus
reversal ([Table 1]). [Fig. 4] presents an overview of the patient selection.
Table 1
Patient cohort overview.
|
total patients
|
total cases
|
patient group
|
description of group
|
number of patients
|
number of cases
|
|
252
|
274
|
I
|
patients with artificial anus closure
|
225
|
244
|
|
II
|
patients without artificial anus closure
|
27
|
30
|
Fig. 4 Flow chart of patients’ selection and subgroup definition.
The patient population included 95 women (38 %) and 157 men (62 %) with a median age
of 59.5 years (18 – 83 years). Analysis of the age distribution of the study cohort
showed a gender-independent age peak in the seventh decade of life (n = 73, 29 %,
[Fig. 5]).
Fig. 5 Age distribution of evaluated patients.
Analysis of the underlying diseases, leading to the initial operation with a double-barrelled
ileostomy, demonstrated a broad spectrum of benign and malignant entities. As a number
of primary diseases were present in some patients, the total number of underlying
diseases exceeded that of the patients. Two patients had, in addition to ulcerative
colitis, colon cancer; one patient also had cervical cancer. In 184 patients (72 %)
there was malignant primary disease, most frequently rectal cancer (n = 132), followed
by colon cancer (n = 30). [Table 2] contains a list of the individual malignant primary diseases. At the time of the
examination, 71 patients had a benign primary disease (28 %). The most common type
of benign disease was chronic inflammatory disease (n = 35), namely ulcerative colitis
in 26 patients and Crohn's disease in nine patients. The second most common benign
disease was diverticulitis or diverticulosis of the sigmoid in nine patients, with
seven patients having already experienced the complication of a perforation. The entire
range of benign primary diseases is presented in [Table 3].
Table 2
Overview of malignant primary diseases.
|
malignant primary disease
|
number of patients
|
|
rectal cancer
|
132
|
|
colon cancer
|
30
|
|
pseudomyxoma, mesothelioma
|
6
|
|
ovarian cancer
|
5
|
|
cancer of appendix
|
5
|
|
stomach cancer
|
2
|
|
pancreatic cancer
|
1
|
|
other
|
3
|
Table 3
Overview of benign primary diseases.
|
benign primary disease
|
number of patients
|
|
ulcerative colitis
|
26
|
|
Crohn’s disease
|
9
|
|
diverticulosis(-itis)
|
9
|
|
adenoma
|
7
|
|
FAP
|
4
|
|
ileus
|
4
|
|
fistula, abscess
|
4
|
|
polytrauma
|
3
|
|
colon ischemia
|
2
|
|
other
|
3
|
Likewise, the type of surgery performed to treat the underlying disease requiring
the ileostomy was analyzed ([Fig. 6]). Rectal resection was the most frequent operation (n = 140), colorectal resections
were performed in 35 patients, and 26 patients had undergone proctocolectomy. In 20
patients only a double-barrelled ileostomy was performed without organ resection.
In this case the ileostomy was protective, for example in the case of multiple abscesses
and fistulas, in the occurrence of an ileus, in the case of inadequate wound healing
in the pelvis, prior to planned radiochemotherapy, because of unexplained GI hemorrhage
with mucus lining ulceration.
Fig. 6 Overview of applied surgical techniques for ileostomy.
An analysis of the type of anastomosis indicated a rectal anastomosis in 63 % of cases
(n = 158); 22 % demonstrated colon anastomosis (n = 56), and there were multiple intestinal
anastomoses in 7 % of cases (n = 18). In 8 % of patients (n = 20) no anastomosis was
performed, the ileostomy was protective.
Evaluation of the technical data showed a median dose area product (DAP) of 2067 μGy
× m2 (range 12.5 – 75 637 μGy × m2) and a median fluoroscopy time of 234 seconds (range 1 – 1386 s). The mean was 2994.7
µGy × m2. Using a conversion factor of 0.26 mSv/ cGy × cm² for gastrointestinal fluoroscopic
examinations results in a mean value of 7.8 mSv (SD +/– 13 mSv) for monocontrast images
of the loop ileostomy with a median of 5.4 mSv [7].
Stenosis
In 74 of 274 examinations (27 %), narrowing of the aboral stoma was documented before
administration of a spasmolytic. 200 examinations (73 %) were unremarkable with respect
to the presence of stenosis. Intravenous spasmolysis was performed in 59 of the 74
examinations (80 %) with documented stenosis. After spasmolysis, stenosis persisted
in 20 of the 59 patients (34 %) and was evaluated as a real stenosis. In 39 examinations
(66 %) no stenosis was visible after spasmolysis, so that the constriction was considered
functional. In 15 examinations (20 %) spasmolysis was ruled out due to contraindications.
In these examinations, the constrictions were evaluated by the examiners as real in
13 cases (87 %) and twice (13 %) as functional stenosis. This resulted in a total
of 241 findings (88 %) without radiologically detectable constriction and 33 examinations
(12 %) with imaged stenosis.
A therapeutic effect was found in 15 of the 33 examinations with real stenosis. In
13 patients the ileostomy reversal was not initially performed. In two patients, resection
of the stenosis in the region of the anastomosis in the course of ileostomy reversal.
In the remaining 18 studies, the radiologically diagnosed stenosis had no therapeutic
relevance to the surgical procedure.
Anastomotic Leakage
Suspicion of anastomotic leakage was described in 15 of 274 examinations. Computed
tomography of the abdomen was additionally performed on all these patients. Leakage
was confirmed in eight of the 15 patients. In the remaining seven patients, the findings
of CT and proctoscopy were summarized postoperatively. Thus there were eight cases
of anastomotic leakage and 266 examinations without evidence of leakage.
In seven of the eight positive findings (88 %), the diagnosed anastomotic leakage
had immediate therapeutic consequences: in six patients, closure of the existing stoma
was dispensed with, and in another patient the double-barrelled ileostomy was converted
into a terminal colostomy.
In one patient, the radiological suspicion of anastomotic leakage could not be confirmed
in the proctoscopy so that closure of the ileostomy was performed.
Incontinence
The analysis of 274 examinations with respect to rectal incontinence for liquid contrast
agent revealed that continence was completely maintained in 194 examinations (71 %).
Incontinence was documented in 80 examinations (29 %). A therapeutic consequence could
be derived in seven examinations (9 %); in these patients, closure was initially ruled
out. In 73 studies (91 %) the radiologically documented incontinence did not have
any therapeutic consequences, and closure surgery was performed.
Secondary Findings
Diverticula were the most frequently documented secondary findings (n = 54). An intraluminal
contrast media recess was described five times: the causes were air pockets (n = 2)
or scybala (n = 1); in two patients the etiology could not be definitively explained
even after a second review of the images. Fistulas were described in four examinations.
Post-inflammatory changes were described for one patient. The documented secondary
findings were not relevant to therapy in any case.
Further Therapy and Patient Progress
Closure surgery was performed in 225 patients of the cohort (89 %) after monocontrast
imaging of the aboral stoma or later; these patients are summarized on Group I. In
27 patients (11 %), the stoma was not reversed; these patients were assigned to Group
II.
The median time interval between the imaging of the loop ileostomy and the reversal
operation was 4 days (range 0 – 332 days, mean value 18 days).
In the group of 225 patients with ileostomy closure (Group I), the indication for
the reversing operation was based on radiographic fluoroscopic examination in 34 of
the 225 patients (15 %). In 63 patients (28 %) the radiological findings and a further
examination (e. g. manometry, proctological examination) led to a decision. In 16
patients (7 %) the indication for surgery was determined by colonic monocontrast and
two to three additional examinations. In 20 patients (13 %) it was not possible to
determine retrospectively which examinations were instrumental in establishing the
indication for surgery. A reversal operation was performed in 92 patients (36 %),
although the examination showed stenosis, leakage or incontinence; the radiological
findings were not considered in the decision-making process. The results are shown
in [Fig. 7].
Fig. 7 Decision making for further procedures within group 1.
In the group without artificial anus closure (Group II), reversal surgery was not
performed in five of 27 patients (19 %) due to radiological findings. In four patients
(15 %) the radiological findings and a further examination contributed to a decision.
For three patients (11 %), the decision to forgo reversal surgery was based on the
findings of three examinations (rectoscopy, manometry and monocontrast). Findings
of fluoroscopic examinations had no influence of the decisions made for seven patients
(26 %). Based on written documentation, it was not possible to determine retrospectively
which examinations influenced clinical decisions in the cases of eight patients (30 %).
Discussion
In our study, we analyzed 252 patients with protective ileostomy who between January
2005 and February 2011 were examined using contrast-enhanced fluoroscopic examinations
of the aboral segment of the ileostomy. The aim of the study was to describe a typical
patient collective in which this examination was performed at a full-service hospital,
while evaluating the clinical relevance and value of this fluoroscopic examination
prior to surgical reversal. In particular, the extent to which surgical decision-making
is influenced by radiological findings should be evaluated.
There are already several studies which discuss the importance of conventional radiographic
examination of the diverting segment or other intestinal sections using monocontrast
as part of routine diagnosis compared to other examinations methods.
As early as 1990, Thoeni et al. [8] demonstrated that in the special case of the ileo-anal pouch, monocontrast-based
fluoroscopic examination did not offer any clear advantage compared to computed tomography
and scintigraphy in the detection of postoperative complications. In a group of 57
patients with an ileo-anal pouch, the investigators showed that fluoroscopic examination
did not lead to diagnostic gains with respect to complications such as pouchitis,
abscess, or fistula formation. Hong et al. in 2012 [3] reported similar results in an investigation of 145 patients with protective ileostomy
performed after deep anterior resection whose anastomosis was examined clinically,
endoscopically and using contrast enema. It was found that the fluoroscopic examination
in patients with colorectal or colo-anal anastomosis did not provide any additional
information in preoperative diagnosis compared to clinical examination and proctoscopy.
The investigators did not consider monocontrast imaging of the aboral stoma as a routine
necessity, especially since the radiological findings of the pouchogram of the three
patients of the group who developed postoperative complications were unremarkable
in all cases. Da Silva et al. [2] evaluated the fluoroscopy of 84 patients with a colon pouch and protective ileostomy
prior to planned closure surgery. They found a low sensitivity in the prediction of
postoperative complications. This sensitivity was 0 % for anastomotic leakage, 33 %
for stenosis and 50 % for predicting fistula formation. The course of therapy was
influenced in only one case. An 2007 additional study by Khair et al. [5]. showed a similar result. They retrospectively evaluated the data of 81 patients
with deep anterior rectal resection and placement of an ileostomy prior to planned
closure surgery. A preoperative monocontrast examination was performed on 69 of the
81 patients. Their results also show that routine imaging of the aboral segment or
anastomosis using monocontrast plays a subordinate role in the absence of clinical
suspicion of leakage. In their study Cowan et al. [9] evaluated the data of patients with anterior rectal resection and ileostomy without
postoperative complications and colo-anal/rectal anastomosis accessible to a clinical
examination; fluoroscopy was performed on only 35 of 59 patients. An operation to
close the ileostomy was planned for all patients. The investigators concluded that
contrast imaging prior to a reversal operation is not necessary in cases of complication-free
progress and a deep anastomosis allowing accessible clinical examination. In a group
of 211 patients, Kalady et al. [3] demonstrated that contrast-enhanced imaging of deep anastomoses before ileostomy
closure surgery was often used for confirmation in clinical suspicion of leakage,
but the study did not detect any anastomotic leakage that had not already been clinically
suspected. Thus the examination did not provide any diagnostic benefit to planning
therapy for the patient. Lawal et al. [10] investigated the utility of routinely performed ileostomy imaging prior to closure
in a cohort of 26 children and adolescents. They also found no diagnostic gain in
comparison to clinical examination or after consideration of the individual medical
history and postulated that fluoroscopic examination of ileostomies should be restricted
to patients with suspected leakage, especially since children should be subjected
to as limited radiation exposure as possible. Only Hrung et al. in their study published
in 1997 [11] demonstrated a good detection rate of leakage not previously detected clinically.
They retrospectively evaluated 59 contrast studies of 40 patients after proctocolectomy
and placement of an ileoanal pouch and an ileostomy. Routine postoperative contrast
imaging of the stoma was performed in 37 patients. Three clinically unsuspected anastomotic
leakages were detected.
To our knowledge, the present study is the first attempt to evaluate the direct impact
of study results on clinical decision-making regarding planned reversal surgery. This
issue was examined using a large retrospective study cohort. Thus, not only patients
with colon or rectal resections were included, but also patients in which a protective
ileostomy was performed without intestinal resection such as after trauma or to protect
an irradiation region.
The decision for or against ileostomy closure was reviewed. Of 225 patients who had
surgery to close the ileostomy, fluoroscopic examination was decisive in the decision
for 34 of them (15 %). In 80 patients (36 %) the radiological findings and at least
one further examination (e. g. manometry, proctological examination) led to a decision.
The examination was irrelevant for 92 patients (41 %); the reversal operation was
performed even though the examination yielded findings such as anastomotic leakage,
stenosis or leakage of liquid contrast. Although the monocontrast examination was
the sole basis for decision-making in only 15 % of the examinations, it was a contributing
factor in a total of 50 % of cases and not taken into account in 41 %. It is not possible
to retrospectively evaluate the significance of the fluoroscopic examination when
several other examination modalities contributed to decision-making.
In patient Group II (in whom reversal surgery was not performed), the radiological
findings, together with at least one additional examination were decisive with respect
to seven patients (26 %). The findings had no influence of the decisions made for
seven patients (26 %). In summary, it can be noted that the radiological findings
of the fluoroscopic examination was significant for the decision not to perform stoma
closure in 35 % of patients, and had no impact in 26 % of cases.
The result of the fluoroscopic examination played a subordinate role with respect
to radiologically detected incontinence, since this leakage had therapeutic consequences
in only 9 % of examinations. Here the patient’s desire appears to be crucial for decision-making,
especially if the patient is compliant and can improve the continence function through
Kegel exercises, for example. Stadelmaier et al. [12] also showed that there was no significant correlation between postoperative continence
and continence for liquid contrast media in a monocontrast study of a cohort that
had undergone ileostomy reversal after rectal carcinoma. Likewise regarding these
results, the authors of this study, when describing a radiologically-detected anastomotic
leak, ascribe a rather subordinate role to the examination in the evaluation of the
clinical decision-making.
Fluoroscopic examination of the loop of the ileostomy appears be significant primarily
in the case of anastomotic leakage. In 15 examinations there was suspicion of anastomotic
leakage which in all cases was confirmed by computed tomography. A comparison of the
two radiological examinations showed suspicion of anastomotic leakage in eight cases
(53 %); the remaining findings showed postoperative changes. However in all eight
of the cases, the findings were consequential. In 7 patients closure surgery was not
performed. A proctoscopy performed on one patient could not confirm the finding, and
the surgery was performed.
Stenosis detected in the fluoroscopic examination exhibited a therapeutic consequence
in 45 % of cases, and none in 55 %. These results neither proved nor disproved the
important of radiological examination. In summary, the results of the monocontrast
examination of the diverting segment of the ileostomy is important in the context
of routine clinical diagnostics, possibly because of rapid and universal availability
as an initial exploratory method to determine any contraindications would result in
further diagnosis in the case of indistinct findings and additional examinations to
clarify special issues.
In addition, radiation exposure for both patient and examiner in the course of monocontrast
examination is a factor to be considered. An effective dose of an average of 7.8 mSv
in our cohort is only slightly lower than, for example, computer tomography of the
abdomen and pelvis, which taking into account the DLP values specified by the Federal
Office for Radiation Protection [13] and a conversion factor of approx. 0.015 according to Deak et al. [14] is at about 10 mSv. Hirofuji et al. [15] evaluated the dosage of fluoroscopic examinations of the lower GI tract. They found
an effective dose of 10.7 mSv for an analog fluoroscopic examination, and about 9.4 mSv
using a digital system. These values were not exceeded in our study.
Its retrospective design was a limitation of our study. There were numerous documentation
discrepancies between the different electronic patient documentation systems (RIS
and PACS) as well as documentation gaps, the cause of which could not be determined
retrospectively. Nevertheless, we consider the evaluation of 252 consecutive patients
over a more than 5-year observation period as sufficiently representative in order
to be able to adequately classify the significance of this radiological fluoroscopic
examination clinically.
Conclusions
The evaluation of the clinical therapeutic consequences of radiological imaging of
the loop ileostomy using monocontrast at a full-service hospital demonstrated that
a single conventional radiological examination of this segment does not generally
appear to have a direct influence on decision-making regarding ileostomy reversal,
and in many instances supplementary examinations are called for. However in the course
of routine diagnosis it is useful as an initial exploratory method to ascertain any
contraindications which would result in further diagnosis in the case of indistinct
findings and additional examinations to clarify special issues.
Clinical Relevance of the Study
-
Frequently used as an exploratory procedure in the course routine diagnosis.
-
Generally few therapeutic consequences of fluoroscopic examination
-
Application with respect to special clinical issues, particularly anastomotic leakage.
