Diagnosis of Hirschsprung’s

Disease

A. N. O’Donovan1 OBJECTIVE. Our aims were to determine the validity of using low-osmolality water-
G. Habra2 soluble contrast enemas (WSCE) in neonates and infants with suspected Hirschsprung’s
S. Somers1 disease (HD) and to devise a scoring system that uses a checklist of radiologic signs to
0. E. Malone3 determine the probability of HD.
A. Rees4 MATERIALS AND METHODS. The records of all patients referred by pediatric sur-
A. L. Winthrop5 geons from 1988 through 1992 for the radiologic investigation of possible HD were retro-
spectively reviewed. Thirty-eight patients who were from 2 days to 9 months old were
studied; 20 of them were neonates (less than 1 month old). Of all the patients, 24 under-
went WSCE and the other 14 underwent barium enemas. For all patients, HD had been
diagnosed by rectal biopsy or excluded by biopsy, clinical follow-up, or both. Radiographs
were read by a gastrointestinal radiologist who used a checklist of diagnostic criteria
reported in the literature. The sensitivity and specificity of the findings were compared
with those in the literature.
RESULTS. Of the I 8 patients with HD, 12 were neonates. All reported radiologic diag-
nostic criteria were seen; the frequency, sensitivity, and specificity of the findings were
reported. Twenty percent (n = 2) of HD patients in the WSCE group (n = 10) had negative
findings. Two of the 1 2 neonates developed colonic perforation, one during the enema and
the other within 24 hr of the procedure.
CONCLUSION. WSCE has a sensitivity and specificity equivalent to those of the
barium enema for the detection of HD. For the two patients with perforation, the use of
WSCE was of considerable benefit, avoiding the problems associated with barium spillage
into the peritoneal cavity. A scoring system for diagnostic enemas is feasible.

H irschsprung’s
a congenital,
disease
functional
(HD) is
intes-
infants
ogist’s
with
opinion
abnormal
then
bowel
forms part
habits. A radiol-
of the data
tinal obstruction attributable to used by the clinician (not always a pediatric
the absence of ganglia in the submucosal surgeon) to decide whether manometry, rectal
and intramuscular plexuses of the gas- biopsy, or both are indicated to confirm or
Received September 28, 1995; accepted after revision trointestinal tract [1]. The radiologic fea- exclude HD. Not all of these patients have
March 11, 1996. tures of HD on barium enemas (BE) and HD; a spectrum of conditions may be present.
1 Department of Radiology, Health Science Centre, Mc- delayed radiographs have been described The first aim of this study was to deter-
Master University, 1200 Main St W., Hamilton, Ontario [2, 3]. In the l980s, the sensitivities, spec- mine the validity of using low-osmolality
L8N 3Z5, Canada. Address correspondence to A. N.
O’Donovan.
ificities, and predictive values of these water-soluble contrast enemas (WSCE)
radiologic findings were critically evalu- instead of BE in neonates and infants for the
2Henry Ford Hospital, Detroit, Ml 48202-2689.
ated [4, 5]. It has been shown that a radio- investigation of HD. The second aim was to
3St. Vincent’s Hospital, Dublin 4, Ireland.
logic study alone is not a sensitive enough develop a scoring system to improve the
4College of Medicine, University of Wales, Cardiff CF4 tool to exclude HD. Manometry, rectal diagnostic accuracy of WSCE for the detec-
4XN, Wales. mucosal biopsy, or both are required for tion of HD.
5Division of Pediatric Surgery, St Louis Children’s Hospi-
an accurate diagnosis [4-71.
tal, St. Louis, MO 63110.
In many centers, a contrast enema is Materials and Methods
AJR 1996;167:517-520
readily available at any time and is the first Our retrospective study included all neonates
0361-803X/96/1672-517 examination requested for the evaluation of and infants who were referred for radiologic eval-
© American Roentgen Ray Society neonates with distal bowel obstruction or uation because of symptoms and signs suspicious

AJR:167, August 1996 517

 O’Donovan et al.

for HD from the pediatric surgical division at the colonic inucosa are shown in Figures 1-3. Signif- believe that such agreementjustifies the use
McMaster University Health Science Centre from icant retention of contrast agent was considered of WSCE in neonatal and infant popula-
1988 through 1992. In the radiology department present when the study radiologist saw contrast tions. In our WSCE group of patients, the
of that institution, diagnostic neonatal enemas are agent proximal to the sigmoid colon at 24 hr [4].
only radiologic sign that was not identified
performed with low-osmolality contrast medium An example of delayed evacuation and fine
by the study radiologist was a cobblestone
(ioxaglate sodium meglumine IHexabrix 160; mucosal serrations is shown in Figure 4.
appearance of the colonic mucosa. In our
Mallinckrodt Canada, Pointe Claire, Canadal or The occurrences of the various criteria from
series, this sign was seen on a single BE in a
iohexol [Omnipaque 300; Winthrop Pharmaceuti- the literature were recorded. The sensitivity and
cals, New York, NY]). The results of the enemas specificity of each criterion were calculated and patient with HD. It is obviously rare: the
and delayed radiographs and the medical records compared with those in the literature [4, 5]. other reported series [4, 5] do not provide
of these neonates and infants were reviewed. either a sensitivity or a specificity for the
Patients were included in the study when all cobblestone appearance of the colonic
radiographs were available and HD either had Results mucosa. Its absence from our WSCE group
been confirmed or excluded by rectal biopsy or
Table I shows the final diagnoses in our in no way detracts from our results or from
had been excluded by a combination of the enema
patient population. The frequency of occur- the validity of the use of WSCE in neonates
results, manometry, and close clinical follow-up.
rence of abnormal enema findings is shown and infants.
Of the 52 patients referred for evaluation during
in Figure 5. Two cases of colonic perforation were
the study period, 38 satisfied the study entrance
By combining the results of our two found in our series: one occurred during the
criteria (i.e., all radiographs and records were
available). There were 23 male and 15 female groups (BE and WSCE patients), we found enema, and the other occurred within 24 hr
patients. Twenty patients (53%) were neonates. that five (28%) of the 18 patients with HD of the procedure. In both cases, water-solu-
Fourteen enemas were perfonned with barium, had normal enema results. ble contrast medium was present in the
and 24 were performed with a water-soluble con- Table 2 shows the sensitivities and speci- colon at the time of perforation. Thus, the
trast agent. ficities of the radiologic criteria for the com- problems associated with barium spillage
Radiographs were read by an experienced gas- bined groups and for patients who underwent into the peritoneal cavity were avoided.
trointestinal radiologist unaware of the results of
WSCE. Our results are in good agreement
the pathology reports. The study radiologist knew
with those previously published [4, 5]. Scoring System
that not all patients in the study had HD and had
recently reviewed the relevant literature I I . 4, 5]. The second aim of our study was to
The radiologist used a checklist of radiologic cri- WSCE Patients develop a scoring system to aid radiologists
teria described in the literature, including transi-
Ofthe 24 patients who underwent WSCE, in making an informed decision with regard
tion zone, irregular contractions, rectosigmoid
10 had HD. Two (20%) of these 10 patients to the presence or absence of HD. Any such
index (maximum width of the rectum divided by
had normal enema results. The sensitivities reporting scheme will have limitations
maximum width of the sigmoid; abnormal if <I),
and specificities of the radiologic signs for because of high false-negative rates (20-
spasm, mucosal cobblestone pattern on WSCE,
mucosal irregularity, serrations, and retention of these patients are shown in Table 2 and can 28%). Nevertheless, we believe that a scor-

contrast agent on delayed radiographs [I , 4, 5]. be compared with those for all patients. ing system is worthwhile, particularly for
Examples of a typical transition zone, short-seg- We found good agreement in the sensi- conveying a level of diagnostic probability
ment HD, and cobblestone appearance of the tivities and specificities for both groups. We to our clinical colleagues.

Fig. 1.-Radiograph from water-soluble contrast enema Fig. 2.-Radiograph from water-soluble con- Fig. 3.-Radiograph from barium enema in neonate with
in 1-month-old female neonate with Hirschsprung’s dis- trast enema in 9-month-old male infant with Hirschsprung’s disease shows mucosal cobblestone pat-
ease shows rectosigmoid transition zone (filled arrow). biopsy-proven short-segment Hirschsprung’s tern in transverse colon. Transition zone was near junction
Note associated irregular contractions and mucosal ir- disease shows short tight stricture in distal of sigmoid and descending colon. Cobblestone pattern was
regularity (open arrows). sigmoid colon. originally considered likely related to colitis. It persisted for
several months after resection of aganglionic segment.
There was no clinical evidence of colitis atthattime. Etiology
of cobblestone pattern is unknown.

518 AJR:167, August 1996

 Diagnosis of Hirschsprung’s Disease

%
8Oi
70
60

50

40

30

20

io
0
Serrat Spasm Contract M Irreg
Diagnostic Criteria

- HD (n-18) Non-HD (n=20)

Fig. 5.-Bar graph shows frequency of occurrence of some abnormal enema findings in this study. 17 = transition
zone, RSI = rectosigmoid index, Serrat = serrations, Contract = irregular contractions, M lrreg = mucosal irregu-
larity, Cbls = cobblestone pattern of mucosa, HD = Hirschsprung’s disease.

fi11SensitMty and Specificity of Radlologlc Criteria

All Enemas (BE and WSCE) WSCE

Fig. 4.-Radiograph from barium enema in infant with Finding
Hirschsprung’s disease shows delayed evacuation of % Sensitivity % Specificity % Sensitivity % Specificity
barium. Barium
column extends proximal to sigmoid Transition zone 65 60 80 64
colon. Note multiple fine serrations along colonic mar-
lrregularcontractions 22 90 20 100
gins (arrows).
Rectosigmoid index 17 65 80 71
Spasm 33 85 40 93
Final Diagnoses for 38
Cobblestone appearance 5 100
!leonates and Infants with of mucosa
iowel Dysfunction Mucosal irregularity 22 95 20 100
Serrations 50 90 50 93
Diagnosis No (%)ofPatients
Delayedevacuation 66 20 71 33
Hirschsprung’s#{231}Iisease -
y.4i;1(41) -., -.
‘
Note-BE = barium enema, WSCE = water-soluble contrast enema.
Neonates,:,.. ‘ 12 (32) !
Infants . ‘ ‘#{149}
.-. -f #{149}_tt,

Meconium plug syndrome ? .

should benefit their clinical colleagues.
Functional megacolon :; .11(29) Table 2 shows that four radiologic signs

Duplication cyst. : f(j have high specificities for HD: irregular

lntussusception :.(; r contractions, spasm, mucosal irregularity,
and serrations. However, their sensitivities
are low. As such, their identification should
We looked at all eight radiologic criteria: prompt a careful search for other signs of
those seen in our series of patients and those HD; in cases of equivocal scores, their pres-
that have been reported in the literature ence may suggest a higher probability.
(Table 2). Our scoring system was simple.
When a sign was present, the study radiolo-
gist recorded a score of 1; when a sign was Discussion
absent, the score was 0. Therefore, the maxi- In this study, we have shown that the use should be supplemented with other clinical
mum score for a patient was 8. With regard of WSCE in the neonatal and infant popula- tests, including rectal biopsy. Nevertheless,
to the rectosigmoid index, when the rec- tion in no way alters the diagnostic accuracy the radiologic enema has an important role
tosigmoid index was less than 1, the study of the enema (when WSCE is compared in the investigation of neonates and infants
radiologist recorded a score of 1 ; when the with BE). In fact, the false-negative rate was with colonic problems.
rectosigmoid index was greater than 1 , the reduced. The false-negative rate in our series Colonic perforation is not uncommon in
score was 0. Table 3 summarizes the results (20-28%) resembled that reported by Smith HD, being reported in 3-4% of affected
of our scoring system for all patients. and Cass (24%) [7]. Our results also agreed patients [8, 9]. In our series, two neonates
Although this system is not perfect, we with those of other published series [4, 5]. developed colonic perforation while their
believe that it will help radiologists calculate Because of the high false-negative rate colons still contained contrast agent. For
a level of probability for their reports, which (20-28%) of the radiologic enema, this test these two neonates, the use of WSCE pre-

AJR:167, August 1996 519

 O’Donovan et al.

vented barium spillage into the peritoneal who performs neonatal and infant enemas 4. RosenlIeld NS, Ablow RC, Markowitz RI, et al.
cavity. We believe that the risk of perfora- only occasionally. Hirschsprung’s disease: accuracy of the barium
enema examination. Radiology 1984;150:393-400
tion in the neonatal population justifies the In this study, we have shown that, like
5. Taxman U, Yulish BS, Rothstein FC. How use-
use of low-osmolality contrast agents. As we barium, low-osmolality contrast agents can
ful is the barium enema in the diagnosis of infan-
have shown in this study, radiologic diag- be used to reveal the features of HD. We tile Hirschsprung’s disease? Am J Dis Child
nostic signs can be as easily revealed with recommend such agents. We also encourage 1986; 140:881-884
these agents as with barium. Furthermore, radiologists to use our scoring system, 6. Lanfranchi GA, Bazzocchi G, Federici 5, et al.
we believe that the margin of safety is which can provide the referring physician Anorectal manometry in
diagnosis the of
Hirschsprung’s disease-comparison with clini-
greater when low-osmolality contrast with an estimate of the probability of HD.
cal and radiological criteria. Am J Gastroenterol
agents are used.
1984;79:270-275
The scoring system outlined in Table 3 References 7. Smith GHH, Cass D. Infantile Hirschsprung’s
provides a means of attaching a level of 1. Momoh iT. Short-segment Hirschsprung’s dis- disease-is a barium enema useful? Pediatr Surg
probability to a radiology report that should ease. Trop Dna 1988:18:16-19 lot 1991;6:3l8-321
2. Berman CZ. Roentgenographic manifestations of 8. Newman B, Nussbaum A, Kirkpatrick JA Jr.
benefit both the radiologist and the referring
congenital megacolon (Hirschsprung’s disease) Bowel perforation in Hirschsprung’s disease.
physician. The system is easy to use and
in early infancy. Pediatrics 1956:18:227-238 AiR 1987;148:l 195-1197
does not require any alteration in the way 3. McDonald RG, Evans WA. Hirschsprung’s dis- 9. Swenson 0, Sherman JO, Fisher JH. Diagnosis
the enema is performed. The system should ease: roentgen diagnosis in infants. Am J Dis of congenital megacolon: an analysis of 501
prove particularly useful to a radiologist Child 1954;87:575-585 patients. J Pediatr Surg 1973;8:587-594

520 AJR:167, August 1996