Journal of Pediatric Surgery 49 (2014) 8790

Contents lists available at ScienceDirect

Journal of Pediatric Surgery

journal homepage: www.elsevier.com/locate/jpedsurg

Loop versus divided colostomy for the management of anorectal malformations

Omar Oda, Dafydd Davies, Kimberly Colapinto, J. Ted Gerstle
Division of General and Thoracic Surgery, The Hospital for Sick Children, Toronto, ON, Canada

a r t i c l e

i n f o

Article history:
Received 17 September 2013
Accepted 30 September 2013
Key words:
Anorectal malformation
Loop colostomy
Divided colostomy

a b s t r a c t
Purpose: The purpose of this study was to compare the clinical outcomes of loop and divided colostomies in
patients with anorectal malformations (ARM).
Methods: We performed a retrospective cohort study reviewing the medical records of all patients with ARM
managed with diverting colostomies between 2000 and 2010 at our institution. Independent variables and
outcomes of stoma complications were analyzed by parametric measures and logistic regression.
Results: One hundred forty-four patients managed with a colostomy for ARM were evaluated (37.5% females,
50.7% loop, 49.3% divided). The incidence of patients with loop and divided colostomies who developed
stoma-related complications was 31.5 and 15.5%, respectively (p=0.031). The incidence of prolapse was 17.8
and 2.8%, respectively (p=0.005). Multivariable-logistic regression controlling for other signicant
independent variables found loop colostomies to be positively associated with the development of a stoma
complication (OR 3.13, 95%CI (1.09, 8.96), p=0.033). When individual complications were evaluated, it
was only stoma prolapse that was more likely in patients with loop colostomies (OR 8.75, 95%CI (1.74, 44.16),
p=0.009).
Conclusion: Because of the higher incidence of prolapse, loop colostomies were found to be associated with a
higher total incidence of complications than divided stomas. The development of other complications,
including urinary tract infections (UTIs) and megarectum, were independent of the type of colostomy
performed.
2014 Elsevier Inc. All rights reserved.

Since the rst diverting stoma performed for the treatment of

imperforate anus in 1783 [1], the site and the type of the least
troublesome stoma in the surgical management of anorectal malformations (ARM) have been major subjects for discussion amongst
pediatric surgeons. Clinical studies have established the high
morbidity associated with neonatal colostomy and that transverse
colostomy has a higher rate of complications than sigmoid colostomy
[2,3]. However debate continues regarding the type of the diverting
colostomy. This study compares clinical outcomes of loop and divided
colostomies performed as part of the surgical management of ARM.
1. Materials and methods
Research Ethics Board approval (Application No: 1000031264,
Protocol No: 1000020656) was obtained to perform a retrospective
cohort study to review the medical records of all patients who
presented with ARM and were managed with a diverting colostomy
between November 2000 and November 2010. Patients with ARM
who underwent primary corrective surgery without a diverting
colostomy were not included in this study. Data were collected from
Corresponding author. Division of General and Thoracic Surgery, The Hospital for
Sick Children, 555 University Avenue, Room 1526, Toronto, Ontario M5G 1X8, Canada.
Tel.: +1 416 813 7500; fax: +1 416 813 7477.
E-mail address: ted.gerstle@sickkids.ca (J.T. Gerstle).
0022-3468/$ see front matter 2014 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jpedsurg.2013.09.032

admission, progress, operative, discharge, and follow up clinic notes.

Radiology reports were also reviewed. Demographic data collected
included sex, type of ARM, and age at stoma creation. From the
operative reports we determined whether a loop or divided
colostomy was created. A loop colostomy was dened when the
bowel wall continuity was partially preserved. A divided colostomy
was dened when bowel continuity was completely disrupted
regardless of the distance between the proximal and the distal ends
of the stoma. For statistical purpose, stoma creation was considered
early when age at stoma creation was less than 6 days, and late when
age at stoma creation was equal or more than 6 days. The level of the
colostomy (sigmoid, descending or transverse colon), the start of
stoma function (days), and the period of time (months) the patient
had the stoma were also determined. For those patients who had
repair of the ARM at our institution, we noted whether a distal
contrast study was performed prior to anorectoplasty and whether it
showed the presence of a megarectum, as determined from the
radiologists report. We determined if the patients developed
complications from their stomas (for the time from stoma creation
till stoma closure); these complications included retraction, prolapse,
parastomal hernia, obstruction, and need for revision. The development of megarectum and UTIs were particular variables of interest;
one of our goals was to establish if there was a relationship between
them and the type of the colostomy, hence they were studied as
potential stoma complications. For those who had repair of the ARM at

88

O. Oda et al. / Journal of Pediatric Surgery 49 (2014) 8790

our institution we documented if these patients developed any

complications following anorectoplasty other than stoma complications. These included wound infection, anal stricture that required
dilatation under general anesthesia or repeated anoplasty, and
recurrence of recto-urinary or recto-genital stula. A more detailed
review of patients who developed UTIs was also performed; data
included the presence of associated urinary tract malformations or
dysfunction and wether these UTI episodes occurred before or after
anorectoplasty and division of recto-urinary stula. For those who had
their stomas closed at our institution, we documented if they
developed wound infections after closure of the colostomy.
Statistical analysis was performed using STATA SE Version 10.0.
The demographics and rates of complications were compared
between patients initially managed with loop stomas to those
managed with divided stomas. Categorical and continuous variables
were analyzed with Fishers Exact and Wilcoxon Rank Sum Tests,
respectively looking for differences between the two groups. Logistic
regression analysis was performed. Along with the type of colostomy
(loop versus divided), independent variables of sex, type of ARM, age
at stoma creation, level of stoma, and duration of stoma were
evaluated for associations with the outcomes of stoma complications
(retraction, prolapse, parastomal hernia, obstruction, need for stoma
revision, and the development of megarectum and UTIs). We also
examined their impact on complications following ARM repair and
stoma closure including wound infection, development of anal
stricture or recurrent stula. In order to prevent the inclusion of
variables with no outcome association in our multiple variable logistic
regression analysis, only independent variables with associations
meeting signicance levels of p less that 0.1 were included to control
for their effects. This was repeated for each outcome measure if the
number of events was sufcient [4]. The results of the multiple
variable logistic regression analysis were only considered signicant if
they reached a p-value less than 0.05.
2. Results
Over the 10 year study period 461 children with all types of ARM
(low, intermediate, and high) were treated at our institution. 317
patients underwent a primary repair of the ARM without a colostomy
and were excluded from the study. 144 patients were managed with a
colostomy. 73 (50.7%) patients had a loop colostomy, and 71 (49.3%)
patients had a divided colostomy. 112 (77.7%) had sigmoid colostomy,
14 (9.7%) had descending colostomy, and 18 (12.5%) had transverse
colostomy. 138 (95.8%) patients underwent anorectoplasty at our
institution: 112 underwent a posterior saggital anorectoplasty
(PSARP), 18 underwent a laparascopic assisted anorectoplasty, and
in 7 patients anorectoplasty was performed with an open abdominoperineal approach. The operative report was not found for one patient
and we were unable to determine what type of anorectoplasty he had,
despite it being performed at our institution. These, and the
demographic data are shown in Table 1. 118 (85.5%) of repaired
patients underwent a distal contrast study prior to anorectoplasty,
and 130 (94.2%) had their colostomies closed at our institution. With
respect to demographics and other data shown in table 1 including
age at stoma creation (early vs. delayed), there were no statistically
signicant differences between patients who had a loop and those
who had a divided colostomy.
Stoma related complications occurred in 34 (23.6%) patients
(some with more than one) giving a total of 61 complications in 144
patients. The types and rates of complications are summarized in
Table 2. Patients with loop stomas were signicantly more likely to
develop complications (p=0.031) and the number of complications
was higher in this group (0.002). When comparing the rates of
developing each individual complication, only the rate of stoma
prolapse was found to be statistically higher in patients with loop
stomas (p=0.005).

Table 1
Demographics.
Variable

Loop

n (%)
Sex Male, n (%)
Type of Anorectal Malformation
Perineal Fistula, n (%)
Vestibular Fistula, n (%)
Cloaca, n (%)
Recto-urethral Fistula, n (%)
Recto-Bladder Neck Fistula, n (%)
H-Type Fistula, n (%)
Atresia without Fistula, n (%)
Level of Colostomy
Sigmoid, n (%)
Descending, n (%)
Transverse, n (%)
Age at Colostomy, median
(IQR) (days)
Delayed Stoma Creation
N 6 days old, n (%)
Duration of Stoma, months
(n=130), median (IQR)
Type of Anorectoplasty (n=138)
PSARP, n (%)
Laparoscopic Assisted
anorectoplasty, n (%)
Abdominoperineal Approach, n (%)
Undened

73 (50.7) 71 (49.3) 144 (100%)

42 (46.7) 48 (53.3) 90 (62.5)
0.232
0.600
4 (5.5)
8 (11.3)
12 (8.3)
16 (22.0) 11 (15.5) 27 (18.8)
7 (9.6)
8 (11.3)
15 (10.4)
26 (35.6) 22 (31.0) 48 (33.3)
6 (8.2)
7 (9.9)
13 (9)
2 (2.7)
0 (0)
2 (1.4)
12 (16.4) 15 (21.1) 27 (18.8)
0.615
56 (76.7) 56 (78.9) 112 (77.8)
6 (8.2)
8 (11.3)
14 (9.7)
11 (15.1) 7 (9.9)
18 (12.5)
2 (1, 3)
2 (1, 3)
2 (1, 2)
0.914

Divided

Total

p-value
C
C

17 (22.3) 13 (18.3) 30 (20.8)

0.540 C

8 (6, 10)

0.085

7 (5, 10)

8 (6, 10)

0.091 C
61 (87.1) 51 (75.0) 112
5 (7.1)
13 (19.1) 18
3 (4.3)
1 (1.4)

4 (5.9)
0 (0)

7
1

C = Fishers Exact test.

= Wilcoxon Rank Sum Test.

Anorectoplasty was performed at our institution on 138 patients.

There were a total of 26 complications following anorectoplasty
which occurred in 22 (15.3%) patients. These are summarized in Table
3 and were no more frequent in either group.
The results of the univariable logistic regression found that sex,
type of ARM, age at stoma creation, and duration of the stoma were
neither positively associated with the development of complications
nor the development of any of the specic complications evaluated;
having not met our predetermined level of signicance (p=0.1) they
were not included in the multivariable analysis.
The level of colostomy was found to be positively associated with
the development of stoma complications. As shown in Table 4,
sigmoid colostomies were most favorable, whereas transverse
colostomies had the highest association with the development of
complications (p=0.006, OR 4.33, 95% CI (1.53, 12.24)). Again,
prolapse accounted for the majority of total complications (p b
0.005, OR 13.61, 95% CI (3.69, 50.20)). This met our predetermined
level of signicance (pb0.1), hence it was included in our model to
control for its effects when determining the signicance of the type of
stoma on specic complications.
Multiple-variable, logistic regression analysis was performed to
determine the association between the type of stoma and the

Table 2
Stoma-related complications.
Type of Complication

Loop

Divided

Total

p-value C

Retraction, n (%)
Prolapse, n (%)
Obstruction, n (%)
Parastomal Hernia, n (%)
Need for Stoma Revision, n (%)
Megarectum, n (%)
Urinary Tract Infection, n (%)
Number of complications
Number of patients who developed
one or more complications

1 (1.4)
13 (17.8)
0 (0)
2 (3.0)
7 (9.6)
5 (8.2)
12 (16.4)
40
23 (31.5)

3 (4.2)
2 (2.8)
2 (2.8)
0 (0)
6 (8.4)
3 (5.3)
5 (7.0)
21
11 (15.5)

4 (2.8)
15 (10.4)
2 (1.4)
2 (1.5)
13 (9.0)
8 (6.8)
17 (11.8)
61
34 (23.6)

0.363
0.005
0.241
0.497
1.000
0.718
0.120
0.002
0.031

C = Fishers Exact Test.

= Statistically signicant value.

O. Oda et al. / Journal of Pediatric Surgery 49 (2014) 8790

Table 3
Complications following anorectoplasty.
Type of Complication

Loop

Divided

Total

P-valueC

Wound infection after

anorectoplasty, n (%)
Anal stricture, n (%)
Fistula recurrence, n (%)
Number of complications
after anorectoplasty
Number of patients who developed
one or more complication after
anorectoplasty, n (%)

9 (12.9)

4 (5.9)

13 (9.4)

0.244

4 (5.7)
2 (2.9)
15

6 (8.8)
1 (1.5)
11

10 (7.2)
3 (2.2)
26

0.529
1.000
0.739

13 (17.8)

9 (12.7)

22 (15.3)

0.487

89

Table 5
Multiple-variable logistic regression examining the association of of loop colostomies
with the development of stoma related complications, controlling for the level of
colostomy.
Complication

Loop colostomies
P-value

Retraction
Prolapse
obstruction
Parastomal hernia
Need for revision
Megarectum
UTI
Patients with complication

C = Fishers Exact Test.

development of complications, controlling for effects of the level of

the colostomy. The results are summarized in Table 5. Having a loop
colostomy was found to be positively associated with the development of complications (p=0.033, OR 3.13, 95% CI (1.09, 8.96)) and
stoma prolapse (p=0.009, OR 8.75, 95% CI (1.73, 44.16)), but not for
any of the other complications. Neither the type nor the level of
colostomy was associated with the development of complications
following anorectoplasty and stoma closure.

3. Discussion
Low types of ARMs can be repaired without a protective diverting
colostomy. Although many reports have shown the safety of one stage
corrective procedure for intermediate and high types of ARMs, a
diverting colostomy is still considered the rst step in the surgical
management of these malformations [5,6]. By itself, creating a
colostomy is a minor surgical procedure, but with potentially
signicant morbidity [79]. Complications include, but are not limited
to, retraction, prolapse, parastomal hernia, bowel obstruction, skin
excoriation, need for revision, and anastomotic leak and wound
infections following stoma closure. The type of a diverting colostomy
chosen depends on healthcare resources, surgeons training, personal
experience and preference [10]. Some surgeons at our institution, and
many other pediatric surgeons in the world, recommend a divided
sigmoid colostomy in the left lower abdominal quadrant (LLQ) with a
sufcient skin bridge between proximal stoma and distal mucous
stula that permits the appliance to be tted on the proximal stoma
allowing complete diversion of stool. They believe that complete stool
diversion will prevent the development of megarectum, UTI, and
wound infection after anorectoplasty [3,8]. Proponents of loop
colostomy claim that a well-fashioned loop colostomy may not lead
to such complications, and that a loop colostomy is easier to create
and close; it also has the advantage of having better cosmetic results,
particularly if it is fashioned at the site of umbilicus [11]. Some
surgeons try to combine the alleged advantages of a divided stoma
with the cosmetic advantage of umbilical incision by dividing the
sigmoid colon and fashioning an end colostomy at the site of
umbilicus and a mucus stula in the LLQ. They claim that umbilical
incision also allows for a better visualization of pelvic structures and

0.363
0.009
0.107
0.756
0.580
0.125
0.033

OR

95%CI

8.75

1.74, 44.16

3.13

1.09, 8.96

OR: Odds Ratio.

95% CI: 95% Condence Interval.
- = Insufcient data to perform regression.
= Statistically signicant value.

thus is a better option, particularly in complex ARMs and cloacas

where additional procedures may be required. Although it is not
mentioned in literature, it is the practice of many other surgeons to
perform a double barrel sigmoid colostomy in the LLQ. These surgeons
believe that disrupting bowel continuity is all that is required for
complete stool diversion, and that a double barrel stoma can serve
all purposes.
The various types of stoma, the ongoing debate with regard to the
best and least troublesome stoma, and the insufcient evidence in the
literature created the basis for this study.
Due to study limitations, including retrospective design and
limited patient population (n=144), it was impractical to study
each of the aforementioned stoma preferences separately. Thus, we
divided all the patients into only two main groups: the rst group
included those who had a loop colostomy (where bowel wall
continuity was partially preserved), and the second group included
those who had a divided colostomy (where bowel continuity was
completely disrupted regardless of distance between the proximal
stoma and the distal mucous stula).
In summary, our results conrm the high incidence of stoma
related complications in general, and the fact that sigmoid colostomies are more favorable than transverse colostomies. The results of
our study also conrm that loop colostomies have a higher rate of
complications than divided colostomies; this is principally related to
prolapse. When other complications (retraction, obstruction, parastomal hernia, need for revision, post-anorectoplasty and post-stomaclosure wound infections) were studied separately, we found no
difference between the two groups. Lastly, our results showed that
development of a megarectum had no relation to the stoma type (loop
vs divided). It is conceivable that a retracted loop colostomy could
lead to stool accumulation in the distal pouch and the development of
a megarectum. However, it is equally conceivable that a wellfashioned loop colostomy may not lead to such complications [8],
which questions the need for bowel division and separation of
bowel ends.

Table 4
Univariable logistic regression examining the association of the level of the colostomy with the development of stoma complications.
Level of
Colostomy

Development of a Complication

Sigmoid
Descending
Transverse

0
0.053
0.006

OR

95%CI

OR: Odds Ratio.

95% CI: 95% Condence Interval.
= statistically signicant value.

Number of Complication

OR

95%CI

p
4.33

1.53, 12.24

0
0.607
0.047

Prolapse

OR

95%CI

5.84
13.61

1.22, 27.78
3.69, 50.20

p
3.19

1.01, 10.03

0.027
b0.005

90

O. Oda et al. / Journal of Pediatric Surgery 49 (2014) 8790

Another controversy that was examined by the study was the

association between the type of colostomy and the development of UTIs
in patients with ARM. Although reports have shown no association
between stoma type and UTIs in these patients [7], a common belief
amongst pediatric surgeons is that loop colostomy can lead to a higher
incidence of UTIs through the existing recto-urinary stula [8]. Although
our results showed a higher incidence of UTIs amongst those who had a
loop colostomy (16.4% versus 7.0%), this was not statistically signicant
(p=0.120). Of interest was the additional nding that the majority of
patients who developed UTIs had associated urinary tract anomalies or
dysfunction, including neurogenic bladder and vesicourethral reux.
The presence of urinary tract anomalies or dysfunction was not initially
a variable of interest in our study, thus it was not included in our
analysis. However, in post-hoc review only one of the 17 patients who
developed UTIs had no associated urinary tract anomalies or dysfunction; and 14 had UTI episodes after division of the recto-urinary stula
and repair of the ARM. A published collective review of 41 cases of
epididymo-orchitis complicating ARM supports our ndings[12].
According to the author, division of the recto-urinary stula was
curative only in one third of cases, which suggested that the presence of
other risk factors, including persistent mesonephric duct syndrome,
urethroejaculatory duct reux, vasovesical ectopia, neurovesical dysfunction and urethral stricture-stenosis were the major risk factors for
the development of epididymo-orchitis in those patients. We think that
UTIs in patients with ARM are mainly the result of existing urinary tract
anomalies or dysfunction, rather than the type of the diverting
colostomy. The fact that there was no difference in the development
of wound infections after anorectoplasty in our groups may indirectly
support this point of view.
4. Conclusion
Because of a higher incidence of prolapse, loop colostomies are
associated with a higher total rate of complications than divided
stomas. Other complications, including megarectum and UTIs are
independent of stoma type. Generally speaking, the best type of the
diverting sigmoid colostomy has not been dened yet. It remains the
responsibility of the surgeon to determine if the benets of cosmesis
and easy closure warrant the increased risk of prolapse associated
with loop colostomies.
References
[1] Yesildag E, Muniz RM. Buyukunal: How did the surgeons treat neonates with
imperforate anus in the eighteenth century? Pediatr Surg Int 2010;26(12):114958.
[2] Chandramouli B, Srinivasan K, Jagdish S, et al. Morbidity and Mortality of
colostomy and its closure in children. J Pediatr Surg 2004;49(4):5969.
[3] Levitt MA, Kant A, Pena A. The morbidity of constipation in patients with anorectal
malformations. J Pediatr Surg 2010;45(6):122833.
[4] Peduzzi P, Concato J, Kemper E, et al. A simulation study of the number of events
per variable in logistic regression analysis. J Clin Epidemiol 1996;49:13739.
[5] Vick LR, Gosche JR, Boulanger SC, et al. Primary laparascopic repair of high
imperforate anus in neonatal males. J Pediatr Surg 2007;42:187781.
[6] Iwai N, Fumino S: Surgical treatment of anorectal malformations. Surg Today 2012
Nov 30 [Epub ahead of print].
[7] Patwardhan N, Kiely EM, Drake DP, et al. Colostomy for anorectal malformation:
high incidence of complications. J Pediatr Surg 2001;36(5):7958.
[8] Pena A, Migatto-Krieger M, Levitt MA. Colostomy in anorectal malformations: a
procedure with serious but preventable complications. J Pediatr Surg 2006;41(4):
74855.
[9] Chirdan LB, Uba FA, Ameh EA, et al. Colostomy for high anorectal malformation: an
evaluation of morbidity in a developing country. Pediatr Surg Int 2008;24:40710.
[10] Chowdhary SK, Chalapathi G, Narasimhan KL, et al: An audit of neonatal colostomy
for high anorectal malformation: the developing word perspective.
[11] Hamada Y, Takada K, Nakamura Y, et al. Temporary umbilical loop colostomy for
anorectal malformations. Pediatr Surg Int 2012;28:11336.

[12] Raveenthiran V, Sam CJ. Epididymo-orchitis complicating anorectal malformations: collective review of 41 cases. J Urol 2011;186(4):146772.

Discussion
Discussant: Mr. Edward Kiely (London, UK): Can I ask you at what age
did you close the stomas?
How long were the stomas present? Sometimes if a loop stoma is
present for many months prolapse is very common. If it is not
present for many months, prolapse usually does not occur.
Response: Dr Omar Oda: If you will look at this table, we have here the
duration of the stoma in months. In the loop group it was eight
months and in the divided group it was seven months.
Mr. Edward kiely: I think if a loop stoma is present for less than three
months you have no prolapse. Its a very simple stoma to make if I
can make a comment. Its an easy stoma to close, and it is relatively
trouble free if the reconstruction is to be done early. I dont believe
the stool comes out of the proximal stoma and looks around for the
distal stoma to go burrowing down straight away. I dont believe
the distal stoma sucks it up either, so if you want to make a simple
stoma and are going to close it soon, I have a strong preference for
loop stoma. I see that Dr. Pena may be on his feet. (laughter).
Moderater: Alberto has long been a champion of the divided
colostomy. Dr. Pena, would you have a comment?
Discussant: Dr Alberto Pena (Cincinnati, Oh): Yes. We studied in
retrospect over 800 colostomies done in different parts of the
world, so we have been exposed to all kinds of colostomies.
Prolapse may occur in separated stomas or in loop colostomies. It
doesnt depend on the type of colostomy. We found that it depends
on the location. If you open the colostomy in a mobile portion of
the colon it most likely will prolapse regardless of the type that you
perform. If you do it in a xed portion of the colon, then it will not
prolapse. And if you do it in a mobile portion of the colon, then you
are obligated to x the colon inside the abdomen in an adequate
way.
When you analyze urinary tract infections, of course it makes
sense to believe that a loop colostomy allows the potential
contamination of the urinary tract, but you have to separate the
patients because there are patients that will have urinary tract
infections for many other reasons. Patients that have tethered
cord, absent sacrum, vesicoureteral reux, cloacas we have to
completely separate.
Dr. Omar Oda: With respect, our data did not reveal any difference
between urinary tract infections between the loop stoma and the
divided stoma. When we analyzed our patients, those with
recurrent UTIs, we found that 16 of the 17 had another urinary
tract malformation, mainly vesicoureteral reux which makes it
more reasonable to infer that these urinary tract infections are
related mainly to urinary tract malformation rather than the other
issue. Of the 17 patients, 14 developed urinary tract infection even
after repair of the malformation and division of the rectourinary
stula which suggests the rectourinary stula may not have much
to do with these infections. This has been reported also by other
studies.