THE AMERICAN JOURNAL OF GASTROENTEROLOGY
12, 1999
© 1999 by Am. Coll. of Gastroenterology
Published by Elsevier Science Inc.
A Randomized, Prospective Study to
Evaluate the Efficacy and Acceptance of Three
Bowel Preparations for Colonoscopy in Children
Ahmed Dahshan, M.D., Chuan-Hao Lin, M.D., John Peters, D.O., Ronald Thomas, Ph.D., and
Vasundhara Tolia, M.D.
The Division of Pediatric Gastroenterology, Department of Pediatrics, Children’s Hospital of Michigan,
Wayne State University School of Medicine, Detroit, Michigan
OBJECTIVE: We performed a prospective, randomized, sin-
gle-blind study in children undergoing colonoscopy to eval-
uate the acceptance and efficacy of three different bowel
preparations.
METHODS: Seventy patients (ages 3–20 yr, 38 males) were
randomly assigned to one of the three study preparations:
Magnesium citrate with X-prep and clear liquid diet for 2
days (group A); Dulcolax for 2 days and Fleet enema
without dietary restriction (group B); and Golytely 20 ml/kg
(up to 1 L) per hour for 4 h with clear liquid diet for 1 day
(group C). Endoscopists blinded to bowel preparation
graded the adequacy of colon cleansing. The preparations
were rated by patients for tolerance, willingness to retake
them, adverse effects, and compliance.
RESULTS: Data analysis using Fisher exact test and trend
test showed that colon cleansing in groups A and C was
superior to that in group B (p ⬍ 0.0001) and better in group
C than A (p ⬍ 0.075). Overall tolerance and compliance
were significantly better for groups A and B than group C
(p ⬍ 0.003), but not different between A and B. More of
group B patients were willing to retake the preparation than
in group C (p ⬍ 0.002) and group A (p ⬍ 0.05), but this was
not different between groups A and C. Adverse effects were
reported more frequently by patients in group C than in
groups A and B (p ⬍ 0.01).
CONCLUSIONS: Although the least well tolerated, Golytely
provided the best cleansing. Dulcolax without dietary re-
striction provided unsatisfactory colon cleansing. Magne-
sium citrate with X-prep was acceptable and provided good
cleansing. (Am J Gastroenterol 1999;94:3497–3501.
© 1999 by Am. Coll. of Gastroenterology)
This work was presented in part at the North American Society for Pediatric
Gastroenterology and Nutrition’s annual meeting, October 1997, Toronto, Canada,
and published as an abstract (J Pediatr Gastroenterol Nutr 1997;25:462).
Vol. 94, No.
ISSN 0002-9270/99/$20.00
PII S0002-9270(99)00672-3
INTRODUCTION
Colonoscopy is an established procedure for investigating
large bowel and terminal ileal diseases in infants and chil-
dren (1–3). An adequately clean colon is essential for mean-
ingful visualization during colonoscopy. Inability to visual-
ize the mucosa may lead to missed diagnosis and could
potentially increase the risk of peritoneal contamination if
the complication of perforation were to occur (1). Cleansing
is especially necessary if polypectomy is contemplated be-
cause mixed gases can be present with residual fecal matter,
which can ignite during electrocautery (2).
Bowel preparation in children undergoing colonoscopy
has been traditionally achieved by using clear liquid diet for
2–3 days, followed by laxatives and multiple enemas on the
evening before and in the morning of the procedure (3).
Although bowel preparation with balanced lavage solutions
(BLS) had been shown to be safe and effective in a pediatric
population, its use has not been as popular as in adults (4, 5).
Young children and adolescents frequently object to the
taste and volume of liquid that must be consumed for an
adequate preparation. Nausea, vomiting, and bloating asso-
ciated with intake of BLS, with subsequent poor compli-
ance, frequently hamper a successful preparation for a clean
colon and nasogastric administration has to be used in some
of the younger patients (5). Furthermore, 1–2 days of clear
liquid diet is usually needed with BLS; this is not easily
accepted.
There is a need for easier, alternative methods for cleans-
ing the colon in children. A recent, uncontrolled pediatric
study evaluated another method of bowel preparation that
did not require any dietary restrictions utilizing an oral
laxative, bisacodyl (Dulcolax), for 2 days and a phosphate
enema administered on the day of the procedure, with re-
portedly good acceptability and adequate bowel preparation
(6). Because there are only a few well-controlled studies in
a pediatric population evaluating the efficacy and accep-
tance of bowel preparations, we performed a prospective,
randomized, single-blind controlled study to evaluate three
methods of bowel preparation in our patients requiring
colonoscopy.
3498 Dahshan et al.
MATERIALS AND METHODS
Patients
Eligible patients undergoing elective colonoscopy were re-
cruited to the study and were randomly assigned by a
computer-generated sequence to receive one of the three
bowel preparation regimens detailed here. Exclusion criteria
included patients ⬍2 yr in age; acutely ill patients with
colitis or emergency colonoscopy; known allergy to bisaco-
dyl, senna derivatives, or polyethylene glycol; or the pres-
ence of metabolic, renal, and cardiac conditions; or failure to
consent to the study. Those who refused to participate were
assigned to the BLS group, as that is our standard prepara-
tion.
This study protocol was approved by the Pediatric Insti-
tutional Review Board of the Wayne State University. In-
dications for colonoscopy included the evaluation of chronic
diarrhea, persistent rectal bleeding, suspected or known
inflammatory bowel disease, and suspected polyps. All pa-
tients were randomized in a single-blind fashion with the
endoscopy nurse determining the assignment of bowel prep-
aration after informed consent. The study was conducted at
Children’s Hospital of Michigan between January 1, 1996
and June 30, 1997. Written instructions were provided to the
families regarding the method of bowel preparation after
verbal explanation. All patients were fasted from midnight
before the day of colonoscopy. Patients underwent colonos-
copy under conscious sedation with meperidine and diaze-
pam or under general anesthesia based on their age, coop-
eration, and family’s preference. All the procedures were
performed by three staff endoscopists (C.H.L., J.P., V.T.)
using Olympus CF 100L, PCF 140L, and PCF 20 scopes
(Olympus America Inc., Melville, NY). All the patients
underwent continuous monitoring of heart rate, respiration,
blood pressure, and oxygen saturation during and for 1/2 h
after the procedure.
Procedures
Group A patients received clear liquid diet for 2 days with
magnesium citrate and X-prep (130 mg sennosides Senna
fruit, sugar, 7% alcohol) (Gray Pharmaceutical, Norwalk,
CT). In the morning of the day before colonoscopy, this
group received magnesium citrate in the dose of 6 oz for
children ⬎5 yr of age and 4 oz for those ⬍5 yr of age, and
X-prep 2.5 oz. Both of these medications were given orally.
Group B patients received Dulcolax (Ciba Consumer
Pharmaceuticals, Woodbridge, NJ) on two consecutive days
with a Fleet enema (C. B. Fleet, Inc., Lynchburg, VA) on the
morning of the procedure, without any dietary restriction.
Dulcolax was given orally once daily in the morning for 2
days before the colonoscopy in the dose of 15 mg for
children ⬎12 yr, 10 mg for those between 5 and 12 yr, and
5 mg for those ⬍5 yr of age. Fleet enema in the amount of
118 ml for children ⱖ12 yr and 59 ml (Pediatric Fleet) for
those ⬍12 yr was given at home on the morning of the
procedure, with instructions to keep the child lying on the
AJG – Vol. 94, No. 12, 1999
Table 1. Cleansing Adequacy Grading Score by the Endoscopist
1 ⫽ Excellent: no fecal matter, or nearly none, seen in the colon
2 ⫽ Good: small amounts of thin liquid fecal matter seen and
suctioned easily, mainly distal to the splenic flexure
3 ⫽ Fair: moderate amounts of thick liquid to semisolid fecal
matter seen and suctioned from the colon, including
proximal to the splenic flexure
4 ⫽ Poor: large amounts of solid fecal matter were found,
precluding a satisfactory study; unacceptable preparation
left side for 15 min, as per the manufacturer’s recommen-
dations.
Group C patients were assigned the BLS, Golytely
(Braintree Laboratories, Braintree, MA). Golytely was ad-
ministered orally on the day before the procedure at the rate
of 20 ml/kg per hour up to a maximum of 1 L/h for 4 h, or
until the stool was clear, along with clear liquid diet for 1
day.
Assessment
Patients and parents were given a questionnaire and inter-
viewed on the day of the procedure to evaluate their accep-
tance and compliance with their assigned bowel preparation
method, diet, and willingness to retake it in the future if
needed. Compliance was assessed by self-report and parents
verified amount of the preparation medication actually con-
sumed and adherence to the dietary restrictions. Possible
adverse effects of the preparations including bloating, ab-
dominal pain, nausea or vomiting, and anal or rectal dis-
comfort with enemas, as well as resulting diarrhea, fecal
incontinence, sleep disturbance, or having to miss school or
work; any other adverse effect as a result of the preparation
were also assessed.
The adequacy of the preparation was graded on a scale of
1 to 4 according to previously established criteria as listed in
Table 1 (4). The endoscopists remained blinded to method
of preparation and the questionnaire results. Note was also
made regarding the level of insertion of the colonoscope and
if the terminal ileum was canulated.
Statistical Analysis
Items evaluated on parents’ questionnaires and endoscopist
evaluation forms for the three groups were discrete and
nominal in nature. Therefore, the Fisher exact test for trend
and ␹2 test (two-tailed) were used for analysis. Findings
were considered statistically significantly if a p ⬍ 0.05 was
obtained. The power analysis for the primary endpoint of the
study was conducted for a 3-⫻-2 cross-tabulation table.
Power was calculated using SPSS Sample Power version 1.0
(Borenstein, Cohen, and SPSS Inc.).
We understand that inclusion of patients who refused
participation in the BLS group may introduce the possibility
of bias into the study in that these patients who refused to
participate may have been somewhat different than patients
who did. In this regard, all ␹2 analyses were run twice and
compared for outcome. One analysis was run using the
AJG – December, 1999
Table 2. Summary of the Three Groups’ Characteristics
Group A Group B Group C
Number of patients 20 19 31
Gender, M:F 11:9 10:9 17:14
Race, C:AA:O 9:9:2 11:7:1 15:16:0
Age in years, 13.9 (3–20) 12.9 (3.2–20) 13.6 (4.2–19.6)
mean (range)
M ⫽ male; F ⫽ female; C ⫽ caucasian; AA ⫽ African-American; O ⫽ other races.
sample size of 70 patients, which included 10 patients as-
signed without randomization to the BLS group, and a
second was run with these 10 patients removed, in essence
using only those patients randomly assigned to one of the
three preparation types.
RESULTS
Seventy patients were recruited for the study; however, only
60 consented, so the remaining 10 were automatically as-
signed to BLS preparation and were considered part of
group C. They were willing to participate in the question-
naire part of the study. There were 38 males and 32 females.
Thirty-five patients were white, 32 were black, and three
were of other races. The mean age of the group was 13.5 yr
(range, 3–20.2 yr). There were 20 patients in group A, 19 in
group B, and 31 in group C. Patients in the three groups
were well matched for age, race, and gender. Table 2 lists
the demographics of patients in the three groups. The cecum
was reached in 62 of the 70 procedures; the terminal ileum
was canulated in 25; seven were aborted because of poor
preparation; and in one the cecum was not reached due to
technical difficulty unrelated to the cleansing (because of
severe distal colitis). Thirty-nine patients had the procedure
done under general anesthesia and 31 under intravenous
sedation. Bowel preparation was rated as excellent in 23
patients, good in 23, fair in 17, and poor in seven patients.
Preparation quality is depicted as bar graphs in Figure 1.
Group C had significantly superior preparation than group B
(p ⬍ 0.0001) and better preparation than group A (p ⬍
0.075). Patients in group A had significantly cleaner bowel,
compared with group B (p ⬍ 0.002). The rate of failure of
Figure 1. Comparison of adequacy of cleaning of the three bowel
preparations.
Pediatric Bowel Preparations 3499
Figure 2. Overall tolerance of patients with the three bowel prep-
arations.
the preparation because of inadequate visualization of mu-
cosa, leading to cancellation of the colonoscopy and re-
scheduling, was unacceptably high in group B patients (7/
19). Thus, overall, both groups A and C had superior
cleansing than group B, with group C having the cleanest
colon.
Patient tolerance of different preparations is shown in
Figure 2. Patients in group B had significantly better toler-
ance than those in group C (p ⬍ 0.0005), and it was also
better for group A than C (p ⬍ 0.004). Group B preparation
was better tolerated than A; however, the difference was not
statistically significant (p ⫽ 0.2). Willingness to retake the
preparation as a measure of acceptance and convenience
was also significantly higher in group B patients than either
groups A (p ⬍ 0.05) or C (p ⬍ 0.002). However, there was
no significant difference between groups A and C (p ⫽
0.09).
Compliance with group B preparation was significantly
better than preparation C (p ⬍ 0.003), and that of A was
better than C (p ⬍ 0.04), as shown in Figure 3. Although
group B patients reported better compliance than those in
group A, the difference failed to reach statistical signifi-
cance (p ⫽ 0.07).
Adverse effects were reported significantly more fre-
quently in group C than in group B, and included vomiting
(p ⬍ 0.01), sleep disturbance (p ⬍ 0.05), and absence from
school or work (p ⬍ 0.008). Group A patients also reported
Figure 3. Compliance of patients with the three bowel prepara-
tions.
3500 Dahshan et al.
Figure 4. Adverse effects with the three bowel preparations.
fewer adverse effects than those in group C, including
vomiting (p ⬍ 0.02), sleep disturbance (p ⫽ 0.25), and
missing school (p ⫽ 0.06), but this was not significantly
different from group B (p ⫽ 0.29, p ⫽ 0.1, and p ⫽ 0.18,
respectively, for the same parameters). These data are
shown in Figure 4.
DISCUSSION
Colonoscopy is now a routine procedure for the evaluation
of the large bowel for diagnostic and therapeutic indications.
The outcome of the procedure is clearly dependent on the
adequacy of the cleansing rendered by the bowel prepara-
tion used. The achievement of good cleansing is difficult
even in adults, who understand the need for adequate visu-
alization during the colonoscopy, but it is a marathon task in
children, whose ability to understand or cooperate varies
with their age and cognitive abilities. Several bowel prep-
arations have been used in children undergoing colonos-
copy, with variable limitations due to tolerance, acceptance,
and reliable cleansing (1, 4 –7). Dietary restrictions, large
volumes of lavage solutions, and adverse effects pose a
greater problem in children, thereby affecting compliance
and consequently the adequacy of the preparation. There-
fore, the continued search for an ideal bowel preparation is
needed particularly for the pediatric population.
Abubaker et al. suggested that bisacodyl with Fleet en-
ema appeared to be an acceptable preparation for colonos-
copy in children, with reportedly good compliance and
cleansing while eliminating the need for dietary restriction
(6). We wanted to confirm these observations and hence
designed this prospective study. Our findings regarding the
efficacy of bisacodyl were extremely disappointing, as the
quality of colonic cleansing was poor. Our study showed
very poor cleansing with this preparation although it was the
most acceptable of the three regimens. One of the main
objectives of achieving a good preparation is to ensure that
the patient does not have to return for a second colonoscopy,
because the preparation itself can be arduous. Furthermore,
a second procedure is inconvenient, costly, and subjects the
patient to the rigors of an alternate bowel-cleansing pro-
gram. Although the regimen needs to be comfortable for the
AJG – Vol. 94, No. 12, 1999
patient, the quality of bowel cleansing can not be compro-
mised. This was clearly a glaring problem with the patients
in the Dulcolax/Fleet enema arm of our study, as the quality
of cleansing was poor in almost half of these patients,
necessitating abortion of the procedure. When the endos-
copy nurses noted that the patients assigned to this regimen
were being rescheduled because of poor preparation, they
discussed these findings with a referee who reviewed these
concerns and recommended that it would be unethical to
continue to recruit patients in this arm of the study because
of the significantly increased risk of repeating the procedure.
The high failure rate of this bowel preparation led to early
termination of the study.
Adverse effects were reported in all groups; however,
none were serious. Of these, vomiting, sleep disturbances,
and absence from school were the most frequent, especially
in group C. The overall tolerance to the preparation and
willingness to retake it in the future reflected the anxiety that
the patients experienced with the dietary restrictions im-
posed, the acceptance of the taste and volume of the laxative
administered, and the perceived adverse effects of the reg-
imen. Our results differ significantly from those reported in
the previous study evaluating Dulcolax without dietary re-
strictions in children, and raises serious doubts regarding its
reproducibility and the adequacy of the cleansing (6). Pain
and discomfort with administration of the enemas was re-
ported in that study as well.
Another regimen for bowel preparation used in adults
includes oral sodium phosphate (7–10). A relatively small
volume of this buffered oral saline laxative, available as a
nonprescription medication, is administered with water on
the evening before and in the morning of the procedure, with
need for dietary restrictions for at least 1 day. It produces
excellent cleansing but has been associated with transient
fluid and electrolyte disturbances (9 –13). It may also induce
mucosal changes mimicking inflammatory bowel disease at
endoscopy (14). Ischemic colitis has also been reported with
its use (15). A small pediatric study by da Silva and co-
workers recently reported on the comparable efficacy of
bowel preparation using orally administered sodium phos-
phate or balanced lavage solution administered via a naso-
gastric tube in doses similar to ours (7). Although the former
laxative was better tolerated, it should be noted that theirs
was a small study involving only 29 children. One of their
interpretations was that if the nasogastric tube could be
avoided, the volume of laxative was a relatively minor
consideration in determining tolerance. Although we did not
study this saline laxative in our study, it clearly requires
further validation. Moreover, none of the patients in our
study needed the lavage solution administered via a naso-
gastric tube; they were able to drink it, albeit with a lot of
perseverance and patience. Although the use of a nasogas-
tric tube obviously compounds the discomfort, as observed
by da Silva et al., we believe that large volumes of unpal-
atable solution and attendant nausea were major factors
AJG – December, 1999
contributing to decreased acceptance of lavage preparation
in our population.
Another recently advocated method of colon preparation
in adults is pretreatment with magnesium citrate in combi-
nation with PEG electrolyte lavage solution to achieve com-
parable cleansing with significantly less volume of BLS
(16). It also requires patients to receive clear liquid diet for
1 day. It was reported to be better accepted, with fewer
adverse effects than with the traditional doses of electrolyte
lavage solution in adults. Such regimens have not yet been
studied in children.
In conclusion, none of the three preparations examined
for bowel cleansing in children undergoing colonoscopy
was ideal. An ideal bowel preparation for colonoscopy may
continue to elude us. The combination of X-prep and mag-
nesium citrate with clear liquid diet had reasonable accept-
ability, tolerance, and bowel cleansing, and clearly emerged
as the overall winner in our study. Our data also suggest that
dietary restriction, with intake of only clear liquids for at
least 1 day, is a very important part of any good preparation,
as the presence of solid feces precludes satisfactory cleans-
ing. Further studies with some of the newer combination
preparations need to be conducted in children in the quest of
a bowel preparation that is easy to comply with, is accept-
able and efficacious, and does not pose any additional risk
in the presence of multisystem disease.
Reprint requests and correspondence: Ahmed Dahshan, M.D.,
F.A.A.P., Division of Pediatric Gastroenterology and Nutrition,
Department of Pediatrics, The University of Oklahoma Health
Science Center, 2815 South Sheridan Road, Tulsa, OK 74129.
Received Sep. 29, 1998; accepted Aug. 9, 1999.
REFERENCES
1. Beneroch LM, Rudolph CD. Pediatric endoscopy. Semin Gas-
trointest Dis 1994;5:32– 46.
2. Rossi T. Endoscopic examination of colon in infancy and
childhood. Pediatr Clinics of North Am 1988;35:331–56.
Pediatric Bowel Preparations 3501
3. Wyllie R, Kay MH. Colonoscopy and therapeutic intervention
in infants and children. GI Endosc Clin North Am 1994;4:
143– 60.
4. Tolia V, Fleming S, Dubois R. Use of golytely in children and
adolescents. J Pediatr Gastroenterol Nutr 1984;3:468 –70.
5. Sondheimer JM, Sokol RJ, Taylor SF, et al. Safety, efficacy
and tolerance of intestinal lavage in pediatric patients under-
going diagnostic colonoscopy. J Pediatr 1991;119:148 –52.
6. Abubaker K, Goggin N, Gomally S, et al. Preparing the bowel
for colonoscopy. Arch Dis Child 1995;73:459 – 61.
7. da Silva MM, Briars GL, Patrick MK, et al. Colonoscopy
preparation in children: Safety, efficacy, and tolerance of high-
versus low-volume cleansing methods. J Pediatr Gastroenterol
Nutr 1997;24:33–7.
8. Wendell K, Tsen TN, Dies DF, et al. Oral sodium phosphate
versus sulfate free PEG electrolyte lavage solution in outpa-
tient preparation for colonoscopy: A prospective comparison.
Gastrointest Endosc 1996;43:42– 8.
9. Vanner SJ, MacDonald PH, Paterson WG, et al. A randomized
prospective trial comparing oral sodium phosphate with stan-
dard PEG-based lavage solution (Golytely) in the preparation
of patients for colonoscopy. Am J Gastroenterol 1990;85:
422–7.
10. Kolts BE, Lyles WE, Achem SR, et al. A comparison of the
effectiveness and patient tolerance of oral sodium phosphate,
castor oil, and standard electrolyte lavage for colonoscopy
preparation. Am J Gastroenterol 1993;88:1218 –23.
11. Cohen SM, Wexner SD, Binderow SR, et al. Prospective,
randomized endoscopic-blinded trial comparing precolono-
scopic bowel cleansing methods. Dis Colon Rectum 1994;37:
689 –96.
12. Henderson JM, Barnett JL, Kim DT, et al. Single day, divided-
dose oral sodium phosphate laxative versus intestinal lavage as
preparation for colonoscopy: Efficacy and patient tolerance.
Gastrointest Endosc 1995;42:238 – 43.
13. Lieberman DA, Ghormley J, Flora K. Effect of oral sodium
phosphate colon preparation on serum electrolytes in patients
with normal serum creatinine. Gastrointest Endosc 1996;43:
467–9.
14. Zwas RF, Cirillo NW, El- Serag HB, et al. Colonic mucosal
abnormalities associated with oral sodium phosphate solution.
Gastrointest Endosc 1996;43:463– 6.
15. Oh JK, Meiselman M, Lataif Jr LE. Ischemic colitis caused by
oral hyperosmotic saline laxatives. Gastrointest Endosc 1997;
45:319 –21.
16. Sharma VK, Steinberg EN, Vasudeva R, et al. Randomized,
controlled study of the pretreatment with magnesium citrate on
the quality of colonoscopy preparation with PEG electrolyte
lavage solution. Gastrointest Endosc 1997;46:541–3.