Journal of Pediatric Surgery (2010) 45, 12281233

www.elsevier.com/locate/jpedsurg

The morbidity of constipation in patients with

anorectal malformations
Marc A. Levitt , Adrien Kant, Alberto Pea
Colorectal Center for Children, Cincinnati Children's Hospital, Division of Pediatric Surgery, University of Cincinnati,
Cincinnati, Ohio 45229, USA
Received 16 February 2010; accepted 23 February 2010

Key words:
Anorectal malformation;
Imperforate anus;
Fecal incontinence;
Morbidity;
Constipation;
Megasigmoid

Abstract
Background: Constipation in anorectal malformations (ARM) is extremely common, particularly in the
lower types. Failure to adequately treat it can lead to significant morbidity.
Methods: From our series of over 2000 patients with ARM, we reviewed 398 with good prognosis for
bowel control and a tendency toward constipation; rectoperineal fistula (63), rectovestibular fistula
(114), rectobulbar urethral fistula (104), imperforate anus with no fistula (46), rectal atresia or stenosis
(9), and cloaca with a common channel below 3 cm (62). Those lost to follow-up, not yet toilet-trained
(b3 years old), or with poor prognostic features were excluded. We compared morbidities in patients we
operated on and managed primarily (group A, n = 268) to those managed at other institutions who
suffered from constipation or incontinence and were referred to us for treatment (group B, n = 130).
Those we managed primarily were subjected to an aggressive senna-based laxative program, started
after their primary repair or after colostomy closure.
Results: Morbidities associated with constipation were higher in the referral group and included fecal
impaction (7.8% vs 38.5%), overflow pseudoincontinence (4.9% vs 33.8%), and megacolon (14.6% vs
54.6%). A loop or transverse colostomy (4.9% vs 9.2%), stoma or anorectal stricture, or a stenotic fistula
(2.2% vs 28.5%) were contributing factors. Adequate laxative treatment with, in certain cases, resection
of a megarectosigmoid (2.6% vs 23.1%) enabled many pseudoincontinent children to achieve bowel
control (reported previously). Unneeded colorectal biopsies (1.9% vs 16.2%), Hirschsprung's-type
pullthroughs (0% vs 3.1%), and, in retrospect, unneeded antegrade continent enema procedures (0% vs
3.1%) were higher in Group B. Overall, 19.8% of Group A and 66.2% of Group B experienced
constipation-related morbidities.
Conclusion: The morbidity of constipation in ARM includes fecal impaction, megacolon, incontinence,
and performance of unneeded surgeries. Inadequate treatment, the type of the original colostomy, and
postoperative anal or stomal stricture as well as stenotic fistulae were key contributing factors. Children
with ARM and good prognosis for bowel control are at the greatest risk for severe constipation and its
consequences. With recognition and aggressive, proactive treatment, we have found that these
morbidities can be reduced.
2010 Elsevier Inc. All rights reserved.

Corresponding author. Tel.: +1 513 636 3240; fax: +1 513 636 3248.
E-mail address: marc.levitt@cchmc.org (M.A. Levitt).
0022-3468/$ see front matter 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jpedsurg.2010.02.096

The morbidity of constipation in patients with anorectal malformations

Constipation in anorectal malformations (ARM) is
extremely common, particularly in the lower types with
good prognosis for bowel control [1-5], which include
imperforate anus with no fistula, rectal atresia or stenosis,
rectoperineal fistula, males with rectourethral fistula at the
bulbar level, and females with rectovestibular fistula or
cloaca with a common channel of less than 3 cm.
Constipation in its most serious forms can produce
overflow pseudoincontinence, frequently related to a dilated
rectosigmoid [6-10]. Failure to recognize or adequately
treat this associated hypomotility in ARM patients can lead
to significant morbidity [4,6-10], which we believe is
largely preventable.
Minor childhood constipation is often dismissed as dietrelated or behavioral. Diet does have some impact on colonic
motility, and treating constipation with stool softeners may be
sufficient in mild cases, but their therapeutic value is
negligible in the most serious forms of constipation. The
rectosigmoid normally stores the stool, and active peristaltic
waves occur every 24 to 48 hours indicating that it is time to
empty. A normal individual, with an intact continence
mechanism, feels this sensation, contracts their sphincter,
and then decides when to relax the voluntary sphincter
mechanism to allow evacuation. The ability to have a
voluntary bowel movement depends on the combined
integrity of the anal canal, sphincters, and motility [11].
Patients with anorectal malformations lack a normal anal
canal, have variably deficient sphincters, and have an
accompanying motility disorder, usually hypomotility
[8,11]. Management for this depends on treatment of
constipation with laxatives to provoke peristalsis and to
overcome the dysmotility. A child with a correctly repaired
good-prognosis anorectal defect, normal lumbosacral spine,
good sphincters, and an intact rectosigmoid should be fecally
continent [1], so that if such a patient is soiling, we can expect
that they are not being adequately treated for constipation.
Hirschsprung's disease is sometimes suspected when a
child with ARM has severe constipation, and some clinicians
perform biopsies [12,13]. In fact, Hirschsprung's disease is
no more common in patients with ARM than in the general
population [8]. Some surgeons have even done Hirschsprung's type pullthroughs in these patients, but loss of the
rectum (a common part of certain pullthroughs for ARM
done mostly before 1985 [14]) leads to loose stool and
requires treatments that slow down the colon. Unfortunately,
most of these patients become fecally incontinent, because,
in addition to the lack of a true anal canal and deficient
sphincters, they now no longer have a rectal reservoir and,
thus, cannot retain loose stool or rely on proprioception for
help with a voluntary bowel movement [15]. Therefore, such
a resection when constipation is misinterpreted for Hirschsprung's can lead to fecal incontinence.
For patients with true fecal incontinence, a bowel
management program is an artificial way to keep them
clean with daily enemas [11,15]. For most ARM patients
(75%), with the potential for continence, medical manage-

1229

Table 1 Indicators of prognosis for bowel control in anorectal

malformation patients
Good prognostic features

Poor prognostic features

Normal sacrum / spine

Abnormal sacrum
No presacral mass
Myelomeningocele
Good buttock crease
Some types of ARM
Good anal dimple
Rectoprostatic urethral fistula
Some types of ARM
Rectobladder neck fistula
Rectoperineal fistula
Cloacal exstrophy cloaca
Rectobulbar urethral fistula N3 cm common channel
Rectovestibular fistula
Complex defects
Cloaca b 3 cm common
channel
Rectal atresia or stenosis
Imperforate anus without
fistula

ment consists of avoidance and treatment of constipation, as

well as effective potty training strategies [11]. Therefore,
only a small minority require washouts and, thus, would need
a Malone or cecostomy.
In this review, we sought to objectively assess the
consequences of constipation in ARM who have good
potential for bowel control (Table 1). We hypothesized that
there is a lower incidence of morbidity when constipation is
managed early and effectively.

1. Methods
From our series of over 2000 patients with ARM, we
reviewed the charts of 398 patients with good potential for
fecal continence (Table 2) and a tendency toward constipation, including those with rectoperineal fistula (63), rectovestibular urethral fistula (114), rectobulbar urethral fistula
(104), cloaca with common channel below 3 cm (62),
imperforate anus with no fistula (46), and rectal atresia or
stenosis (9). These patients historically have a good
prognosis for bowel control. [1] Those lost to follow up,
not yet toilet-trained (or b3 years old), or with poor
prognostic features (Table 1) were excluded. We compared
morbidities in patients we operated on and managed
primarily with an aggressive constipation avoidance protocol
(group A, n = 268) to those operated on, treated primarily or
managed postoperatively at other institutions who suffered
from constipation or incontinence and were referred to us for
treatment, operation or reoperation (group B, n = 130) [25].
In group A, constipation was treated with the following
protocol: a senna-based laxative was begun after primary
repair or after colostomy closure. Stool softeners were
avoided because they make the stool too soft, and they do not
help to provoke a bowel movement. We found stimulant
laxatives in this patient group more effective as they are
designed to increase the motility of the colon. Although often

1230
Table 2

M.A. Levitt et al.

Constipation-related morbidities in groups A and B

Constipation-related morbidities in Groups A and B

Group A n = 268 Group B n = 130
Fecal impaction
21
Pseudoincontinence
13
Megacolon
39
Sigmoid resection
7
Hirschsprung's biopsy 5
Hirschsprung's
0
pullthrough
Cecostomy/
0
appendicostomy

7.8%
4.9%
14.6%
2.6%
1.9%
0.0%

50
44
71
30
21
4

38.5%
33.8%
54.6%
23.1%
16.2%
3.1%

0.0%

3.1%

therapy and resection of a massive megarectosigmoid (2.6%

vs 23.1%) enabled many pseudoincontinent children to
achieve bowel control (a finding that we have reported
previously) [8,11]. Unneeded colorectal biopsies were
performed in 21 patients (1.9% vs 16.2%). Hirschsprung'stype pullthroughs were performed in four patients. Four had
an antegrade continent enema procedure which, in retrospect,
after successful laxative treatment, they did not need.
Overall, 19.8% of group A and 66.2% of group B patients
experienced morbidities related to their constipation.

3. Discussion
discussed as a concern by parents and physicians, we are
unable to find any long term problems related to senna usage.
We sometimes add pectin to the regimen which gives the
stool a little bulk and makes the laxative more efficient. The
goal was 1 to 2 well-formed stools per day.
To ensure regular emptying, the parents are instructed to
give the laxative each day and then observe the child for the
next 24 hours. If the patient does not have a bowel movement
in the 24 hours, or does not empty completely, it means the
laxative dose was not enough, and it must be increased. An
enema is given in order to remove the stool produced during
the previous 24 hours, and the laxative dose is increased. The
routine of giving an enema, if needed, and increasing the
laxative are then continued. The child should stool daily. We
monitor them with an occasional abdominal radiograph to
confirm colonic emptying.
We defined overflow pseudoincontinence as soiling in a
constipated patient who, once treated adequately with
laxatives, has voluntary bowel movements and stops soiling.
Such patients have benign malformations, a normal sacrum
and spine, and good sphincters, yet they soil. Fecal impaction
was defined as episodes requiring hospitalization for
disimpaction. Those patients with poor prognosis defects,
an abnormal sacrum and spine, a flat bottom, and poor
sphincters were considered to be in the true incontinence
group and were not part of this review. Group B patients
were referred for a variety of reasons. Those with good
prognosis for bowel control but with constipation-related
morbidities were analyzed. The protocol for their treatment
and evaluation has been reported on previously [8,11,15].

We recognize that this review is inherently flawed because

there is no way to know the denominator of patients with
ARM who have constipation, and we do not know precisely
what treatments were employed before referral to our center.
Clearly, the problem of constipation is very common
particularly in the benign types, and we found that, in many
patients, it was not recognized or treated aggressively [1]. It
was our intention therefore to show that when constipation is
proactively managed from early in life, and a good bowel
movement pattern (1-2 well-formed stools per day and
objective, radiologic confirmation of emptying) is established, children with ARM with good prognosis should potty
train successfully [1]. We hypothesized that when constipation is not adequately treated, and patients present after many
years, they can have overflow pseudoincontinence and other
constipation related morbidities. We do not understand why
many of these patients suffer from different degrees of
dilation of the distal colon, usually a megarectosigmoid (Figs.
1 and 2A), but surmise that a hypomotility disorder exists that
interferes with complete emptying of the rectosigmoid, which
leads to progressive dilation. This process sometimes may
start in utero as the newborn colon can be found already
dilated (Fig. 1) [7,8,16,17], but usually, this is a process that
develops over time.

2. Results
Morbidities associated with constipation included episodes of fecal impaction (group A = 7.8% vs group B =
38.5%), usually with overflow pseudoincontinence (4.9% vs
33.8%), and varying degrees of megacolon (14.6% vs
54.6%) (Table 2). A loop or transverse colostomy (4.9% vs
9.2%), and/or stomal or anorectal stricture or stenotic fistula
(2.2% vs 28.5%) were key contributing factors. Laxative

Fig. 1

Megarectosigmoid in a newborn.

The morbidity of constipation in patients with anorectal malformations

Fig. 2

1231

A, Megasigmoid. B, Colorectal anastomosis after sigmoid resection. J Pediatr Surg. 1993; 28: 1-5.

Constipation in anorectal malformations is a selfperpetuating disease. We know that dilation of a hollow

viscus leads to poor peristalsis [18], which can explain how
constipation leads to fecal retention and results in increasingly greater degrees of megasigmoid, which then exacerbates the constipation in a vicious cycle. We surmised that
with an early and aggressive laxative regimen, we could
potentially interrupt this cycle.
Colonic dilation can also occur when the distal limb of a
colostomy is very long or not cleaned out, or from stool

spillage, when it is not completely diverted from the

proximal stoma [19]. This dilation leads to more constipation. We also observed varying degrees of megacolon in
patients with postoperative stomal or anal strictures, in
patients who did not perform or continue anal dilations, and
in children with stenotic perineal or vestibular fistulas in
whom the anoplasty was delayed.
Soiling of the underwear in an older child is an ominous
sign of bad constipation and may represent overflow
pseudoincontinence. This phenomenon was more common

Fig. 3 A, Contrast enema showing megarectosigmoid. B, Absent (previously resected) rectosigmoid with straight colonic segment from the
splenic flexure to the perineum (note the haustrations in the pelvis). Printed with kind permission from Curr Probl Surg. 2002; 39: 661-732.

1232
in the patients whose degree of constipation was unrecognized and is a common source of referral to our Center [15].
These patients behave as if they are fecally incontinent, but
when the constipation is treated adequately, the great
majority regain bowel control [11]. Of course, this clinical
presentation of soiling may also occur in a patient with a poor
prognosis ARM, and an associated spinal or sacral anomaly.
For such a patient who has true incontinence, bowel
management with a daily enema is needed [15].
Most of the time, the dose of laxative that the patient
needs in order to empty the rectosigmoid completely, as
demonstrated radiologically, can be achieved. At that dose,
the pseudoincontinent patient should stop soiling, because by
successfully emptying the distal colon each day, the child
will remain clean until the laxative provokes the next
voluntary bowel movement(s). Occasionally, in the process
of increasing the amount of laxative, a patient may be unable
to tolerate (as evidenced by cramping, nausea, or emesis) the
amount required to empty the colon daily. Such a patient may
be a candidate for surgical intervention, a sigmoid resection.
The megarectosigmoid is resected, and the descending colon
is anastomosed to the residual rectum (Fig. 2). We attribute
the smaller incidence of megarectosigmoid (Fig. 3A) found
in our patients to our early and vigilant treatment of
constipation. Our incidence was not zero, however, which
is potentially explained by a congenital dilation, or our own
inadequate treatment and/or follow-up care. After a sigmoid
resection, the amount of laxative required to treat these
children can be significantly reduced or even eliminated
[6,7]. The more localized the dilation of the rectosigmoid, the
better the results of its resection [7]. These patients are not
cured, however, and must be followed up closely, because
the remaining distal colon can redilate.
Before performing this operation, it is mandatory to
confirm that the patient is definitely suffering from overflow
pseudoincontinence rather than true fecal incontinence with
constipationthat is, whether the child remains clean and
has effective voluntary bowel movements provoked by
adequate laxative treatment. Failure to make this distinction
may lead to an operation in which a fecally incontinent
constipated child is changed to one without an adequate
reservoir and a tendency to have loose stool, which is much
more difficult to manage.
Unfortunately, this is what happened to many ARM
patients who were initially treated by abdominoperineal and/
or endorectal pullthrough operations [14] used commonly
before the introduction of the PSARP. Loss of the rectal
reservoir (Fig. 3B) led many of these children to lose their
potential for continence, making them dependent on enemas
[15]. Greater understanding and appreciation of this fact,
along with widespread practice of the posterior sagittal
approach, has considerably reduced the frequency of this
outcome [20]. It still occurs, however, when similar
operations are undertaken in attempts to treat severe
constipation. Faced with severe or recurrent constipation in
a child with an ARM, some physicians suspect Hirsch-

M.A. Levitt et al.

sprung's disease and, as noted, obtain rectal biopsies to rule
out this possibility. Some even proceeded with rectosigmoid
resections. Unfortunately, this leads to loose stools and fecal
incontinence [15]. We believe that, in the absence of
characteristic contrast studies or episodes of enterocolitis,
the likelihood of Hirschsprung's disease in conjunction with
a known anorectal malformation is very low and does not
warrant biopsy or rectosigmoid resection [8].
Other clinicians attempted to treat unmanageable
constipation with antegrade enemas through a catheterizable
stoma or button device [21,22]. We have found, though, that
once their constipation is effectively treated medically with
laxatives, patients can become continent and turn out not to
need washouts at all. We reserve antegrade enemas
procedures for incontinent patients who require a daily
washout [11].
Despite the limitations of this review, mainly owing to the
fact that the referred patients are a non-representative group,
as they are referred because they are having problems, the
trends we have documented are clear. We regard these
morbidities as mostly preventable, despite the fact that some
were present in our own patients. In our group, morbidities
were mainly related to the patient having rectosigmoid
dilation starting at birth, or from inadequate follow up. With
an aggressive constipation avoidance and treatment protocol,
we were able to see a lower rate of morbidity.
The morbidity of constipation in ARM is an impressive
list, including impaction, megacolon, pseudoincontinence,
and performance of unindicated or unnecessary surgeries. In
addition to inadequate treatment, other preventable causes
include type and location of the original colostomy
(transverse and loop colostomies) postoperative stricture
and delayed treatment of a narrow perineal or vestibular
fistula. Noncompliance or poor follow-up also contributes to
increased morbidity from constipation.
It is imperative to understand that children with anorectal
malformations with the best prognosis for bowel control are at
the greatest risk for severe constipation and its consequences.
With appropriate recognition and aggressive, proactive
treatment, these morbidities can be significantly reduced.

References
[1] Levitt MA, Pea A. Imperforate anus and cloacal malformations. In
Pediaric Surgery, 5th ed, Holcomb GW & Murphy JP (editors)
Elsevier Saunders, Philadelphia. pp. 337-59 [In Press].
[2] Heinen F. The surgical treatment of low anal defects and vestibular
fistulas. Semin Pediatr Surg 1997;64:204-16.
[3] Rintala R. Fecal incontinence in anorectal malformations, neuropathy, and miscellaneous conditions. Semin Pediatr Surg 2002;11:
75-82.
[4] Rintala R, Lindahl H, Rasanen M. Do children with repaired low
anorectal malformations have normal bowel function? J Pediatr Surg
1997;32:823-6.
[5] Pakarinen M, Koivusalo A, Lindahl H, et al. Prospective controlled
long-term follow-up for functional outcome after anoplasty in boys
with perineal fistula. J Pediatr Gastroenterol Nutr 2007;44:436-9.

The morbidity of constipation in patients with anorectal malformations

[6] Keshtgar A, Ward H, Richards C, et al. Outcome of excision of
megarectum in children with anorectal malformation. J Pediatr Surg
2007;42:227-33.
[7] Pea A, El-Behery M. Megasigmoid: a Source of pseudo-incontinence
in children with repaired anorectal malformations. J Pediatr Surg 1993;
28:1-5.
[8] Pea A, Levitt MA. Colonic inertia disorders in pediatrics. Curr Probl
Surg 2002;39:661-732.
[9] Smith E. The identification and management of anorectal anomalies
the factors ensuring continence. Prog Pediatr Surg 1976;9:7-40.
[10] Kim H, Gow K, Penner J, et al. Presentation of low anorectal
malformations beyond the neonatal period. Pediatrics 2000;105:e68.
[11] Levitt MA, Pea A. Update on pediatric fecal incontinence. Eur J
Pediatr Surg 2009;19:1-9.
[12] Kiesewetter W, Sukarochana K, Sieber W. The frequency of
aganglionosis associated with imperforate anus. Surg 1965;58:
655-71.
[13] Stevenson J, Herron P. Hirschsprung's disease: a rare cause of
constipation in patients following surgical correction of imperforate
anus: report of 2 cases. Am Surg 1967;33:555-8.
[14] Kiesewetter W. Imperforate anus: the role and results of the sacroabdomino-perineal operation. Ann Surg 1966;164:655-61.
[15] Bischoff A, Levitt M, Bauer C, et al. Treatment of fecal incontinence
with a comprehensive bowel management program. J Pediatr Surg
2009;6:1278-84.
[16] Holschneider A, Koebke J, Meier-Ruge W, et al. Pathophysiology of
chronic constipation on anorectal malformations. Long-term results
and preliminary anatomical investigations. Eur J Pediatr Surg 2001;11:
305-10.
[17] Stephens F. Rectal ectasia: primary and secondary associated with
anorectal anomalies. Birth Defects Org Artic Ser 1988;24:99-104.
[18] Nixon H. An experimental study of propulsion in isolated small
intestine and applications to surgery in the newborn. Ann R Coll Surg
Engl 1960;27:105-24.
[19] Pea A, Migotta-Krieger M, Levitt MA. Colostomy in anorectal
malformations: a procedure with serious but preventable complications. J Pediatr Surg 2006;41:748-56.
[20] Holschneider A, Pfrommer W, Gerresheim B. Results in the treatment
of anorectal malformations with special regard to the histology of the
rectal pouch. Eur J Pediatr Surg 1994;4:303-9.
[21] Malone P, Ransley P, Kiely E. Preliminary report: the antegrade
continence enema. Lancet 1990;336:1217-8.
[22] Rakic S, Baeten C. Laparoscopic appendicostomy with cecoplication
for antegrade colonic enema procedure. Surg Laparosc Endosc
Percutan Tech 2004;14:165-6.

Discussion
Gail Besner, MD (Columbus, OH): Could you comment on
the three patients that were referred to you from elsewhere
who had pull-through procedures that were not indicated?
Marc Levitt, MD (Cincinnati, OH) (response): These
patients had severe constipation as their only symptom.

1233

These are patients that underwent biopsies. Those

biopsies were done too low so there may have been
absence of ganglion cells, but there was no comment on
whether the nerves are hypertrophic or not. Those patients
ultimately underwent rectosigmoid resections and pullthroughs, and sadly, those patients became incontinent
because of the hypermotility induced by that resection.
Obviously, they didn't have the greatest anal canal or
sphincter by nature of their anorectal malformation and
could not handle the hypermotility. This is something that
needs to be avoided. These patients did not have
Hirschprung's disease.
Daniel Von Allmen Md (Cincinnati, Ohio): Do you think
there was a selection bias since the patients who were
referred may have been more complex?
Marc Levitt, MD (response): I believe that this is certainly a
factor. We do not intend in any way to compare the two
groups. These are obviously complex patients. We have
no idea what the denominator is. My only point of
utilizing the two groups was to show that with an
aggressive management scheme from colostomy closure
on, you can dramatically reduce the morbidity. Even in
the group managed in our institution, we did not come
close to zero, there was still a significant amount of
morbidity, but you can dramatically reduce it if you're
very aware of it and treat it aggressively.
Don Nakayama, MD (Macon, Ga): What kind of laxatives
do you use on your patients?
Marc Levitt, MD (response): We use senna-based laxatives.
Most of the things on the market are actually stool
softeners, and we regularly stop the stool softeners and
change them over to something with a stimulant
component, something with senna, I think in these
patients, the problem is not hard stool but rather motility
of the colon. So giving them a stool softener just makes the
stool soft, it doesn't really help it come out, and they have
a colon full of stool that doesn't empty, and in fact, may
empty in more of a leakage kind of way and they're worse.
Many patients will tell you that they're much worse on a
stool softener such as MiraLAX; once you put them on
something stimulating, it actually empties the distal part of
the colon and they do a lot better. It's important to
remember that distinction. We routinely are stopping stool
softeners and switching patients to laxatives.