19th Congress of AAPS

176 ASIAN JOURNAL OF SURGERY VOL 29 NO 3 JULY 2006

2006 Elsevier. All rights reserved.

Diagnostic Scoring System of Hirschsprungs Diseasein the Neonatal Period

Weihong Guo, Qinming Zhang, Yongwei Chen and Dawei Hou, Department of Surgery, Beijing ChildrensHospital Affiliated to Capital University of Medical Sciences, Beijing, China.

OBJECTIVE: Eighty to ninety percent of Hirschsprungs disease (HD) patients present in newborns. However,the diagnosis of HD in the neonatal period remains difficult. Our present study aims to propose a diagnostic

scoring system and hope this will increase early diagnosis of HD and avoid unnecessary rectal biopsy.

METHODS: In the first study period, 57 suspected HD patients (03 months) completed our predeter-mined study protocol in which barium enema (BE), rectal manometry (RM) and full-thickness rectal

biopsy were performed. Symptoms, signs and investigations were analysed for their correlation with HD

diagnosis. A HD diagnostic scoring system was developed according to the statistical results and was

assessed in 74 patients in the second study period.

RESULTS: Forty-five patients were diagnosed with HD in the first study period. A HD scoring systemwas developed in which delayed meconium, tight anus, BE and RM were diagnostic factors. A cut-off

point of 3 provided 84% of HD patients score > 3, whereas 75% non-HD patients score 3 (p < 0.05). Inthe second study group, patients with score 3 were selected for rectal biopsy.CONCLUSION: In the neonatal period, using a HD scoring system may help to select patients for further invasive investigation so that unnecessary biopsy can be avoided. [Asian J Surg 2006;29(3):1769]

Key Words: Hirschsprungs disease, neonatal period

Introduction

Symptoms in 8090% of Hirschsprungs disease (HD)

patients present in the neonatal period.1 With the

advances in neonatal anaesthesiology and surgical care,

the use of the primary endorectal pull-through procedure

in the management of neonates with HD represents

a significant change from the classical approach to

its treatment.2,3 Nevertheless, the diagnosis of HD in

the neonatal and early infant period remains difficult.

Barium enema (BE) and rectal manometry (RM) are

the two common noninvasive tests used in the diagnosis

of HD since they can be performed as screening tests

in outpatient clinics and at bedside.46 Rectal biopsy

is believed to be the gold standard for the diagnosis

of HD. However, there are about 13% of cases of suction

biopsy with false-negative results because of inadequate

biopsies, and full-thickness biopsy may result in sev-

ere complications such as bleeding, perforation and

sepsis.7,8 Alizai et al have reported that only 1217% of

constipation patients who underwent rectal biopsy

were HD patients.9 Therefore, nearly 80% of patients

received unnecessary surgical procedure with the risks

of complications. The purpose of this prospective study

is to propose a scoring system based on the evaluation

of diagnostic factors, in which key clinical features,

physical examination and diagnostic investigations are

included. We hope this will increase early diagnostic

Address correspondence and reprint requests to Dr Weihong Guo, Department of Surgery, Beijing Childrens Hospital,No.56, Nan Li Shi Road, Xi Cheng District, Beijing 100045, China.E-mail: guo-weihong@tom.com Date of acceptance: 28 February 2005

DIAGNOSIS OF HIRSCHSPRUNGS DISEASE

ASIAN JOURNAL OF SURGERY VOL 29 NO 3 JULY 2006 177

acumen and avoid unnecessary rectal biopsy in the neo-

natal period.

Patients and methods

Between February 2002 and March 2003, 57 suspected

HD patients (03 months) were admitted to the Depart-

ment of Surgery, Beijing Childrens Hospital, and com-

pleted our predetermined study protocol. Clinical features

such as delayed passage of meconium/constipation,

abdominal distension, vomiting, diarrhoea and intestinal

obstruction were carefully recorded. Delayed passage of

meconium was defined as failure to pass meconium within

48 hours or the duration was more than 3 days. Rectal

examination, BE and RM were performed in all patients.

Tight anus was considered as an impacted or resistant

feeling when the little finger was inserted into the rectum.

The radiological diagnosis was made independently by

radiologists and the patients were classified into HD

(including short, classic, long segment and total colonic

aganglionosis) and non-HD (colitis) groups. Using a six-

channel probe (MMS UPS2020, The Netherlands), a typical

rectoanal inhibitory reflex (RAIR) was defined as relax-

ation wave with a decrease of 70% of the pressure whenincreasing stimulation volume (1030 mL) was given.10

Failure to induce RAIR with stimulation of maximum

volume repeated three times was considered abnormal.

All 57 patients underwent full-thickness rectal biopsy and

definitive diagnosis of HD was made according to the

pathological results in which ganglion cells and/or nerve

fibres were directly visualized by H&E staining. Additional

immunohistochemical staining for Cathepsin D provided

more information of immature ganglion cells.11 Data

were evaluated for specificity, sensitivity and predictive

value in HD diagnosis. Based on the results of one-way

ANOVA (x2 with p < 0.05 considered statistically signifi-cant), the important factors correlating with HD diag-

nosis were selected and analysed by logistic regression

study. Parameters with good correlation coefficients were

considered to be independent diagnostic factors. A scor-

ing system was developed in which important factors

were assigned a score of 1, and independent factors were

scored as 2.

In the second study period (April 2003June 2004), 74

suspected HD patients (03 months) were studied and the

scoring system was used in all patients to aid in selecting

patients for rectal biopsy.

Results

In the first study period, 45 patients were diagnosed with

HD according to their pathological results. Their clinical

presentations included delayed passage of meconium/

constipation, vomiting, abdominal distension, diarrhoea

and intestinal obstruction. Most of the patients had more

than one of the above symptoms.

As shown in the Table, delayed passage of meconium

occurred in 40 HD patients (88%), the predictive diagnostic

value (including both HD and non-HD diagnoses) of

delayed meconium was 0.77. Of 33 patients who had tight

anus, 30 were finally diagnosed with HD (91%). About 90%

of BE-diagnosed HD patients matched their pathological

diagnosis. The specificity of BE was 89% in the neonatal HD

patients. Short-segment HD was associated with a higher

rate (67%) of false diagnosis and was easily confused with

colitis in BE study. RM study showed a high sensitivity

(88%) and specificity (84%) in HD diagnosis. Again, short-

segment HD was associated with a higher false-positive

result in RAIR test. Compared with the other clinical fea-

tures, delayed meconium, tight anus, BE and RM showed

statistically significant values in HD diagnosis (x2 test,

p < 0.05), making them important predictors in the neona-tal period. When these four important factors were analysed

by logistic regression study, BE and RM had good correla-

tion coefficient (r = 0.9) and were considered as independ-ent factors. A HD diagnostic scoring system was developed

in which delayed meconium and tight anus were assigned

a score of 1, whereas BE and RM were scored as 2.

When the diagnostic score was given to 45 HD

patients, we found that a cut-off point of 3 as shown in

Table. Sensitivity, specificity and predictive value of clinical features in HD diagnosis

Sensitivity Specificity Predictive

(%) (%) value

Delayed meconium* 83 88 0.77

Vomiting 77 38 0.30

Abdominal distension 83 33 0.26

Diarrhoea 86 13 0.11

Intestinal obstruction 80 18 0.14

Tight anus* 91 67 0.68

Barium enema* 90 89 0.84

Rectal manometry* 88 84 0.73

*One-way ANOVA, p< 0.05; logistic regression analysis (r= 0.9).

the Figure provided 84% of HD patients score > 3, whereas75% non-HD patients score 3 (p < 0.05). Out of sevenHD patients whose scores were 3, five patients were clas-sified as having short-segment HD, one had classic and

one had total colonic aganglionosis.

The scoring system was then used in 74 suspected HD

neonates in the following study. There were 57 patients

with score > 3 and all of them had typical signs of HD inBE study. Primary or staged procedures were performed

according to their X-ray classification. Full-thickness

biopsies were performed in 17 patients with score 3.Histopathological results showed that 10 patients had

HD and they eventually underwent surgical treatment.

Another four patients were diagnosed with intestinal

neuronal dysplasia, and three patients had non-HD-

related disease.

Discussion

Diagnosing HD in the neonatal period is a challenge for the

paediatric surgeon. Unlike in older HD patients who have

constipation as a leading feature, the main clinical symp-

toms in neonatal HD patients are variable and include

abdominal distension, vomiting, diarrhoea, sepsis or

intestinal obstruction. These symptoms alone have little

predictive value in HD diagnosis since they may occur in

many other diseases. Rectal suction biopsy has become a

conventional method for the definitive diagnosis of HD.

Some surgeons prefer doing aggressive suction biopsy

in all suspected HD patients. However, it is an invasive

procedure that may cause severe complications. It may

also have false-negative results since it depends on the

surgeons and pathologists technique and experience

with HD diagnosis. BE and RM studies are valuable non-

invasive investigations that still play very important roles

in HD diagnosis. We believe that combined features of

symptoms, signs and diagnostic investigations would

help objectively identify those patients who really require

rectal biopsy for HD diagnosis.

Our statistical analysis suggested that a history of

delayed meconium combined with any of the above clini-

cal presentations must alert a paediatric surgeon to the

possibility of HD. A tightness or feeling of resistance on

rectal examination was another representative sign of HD.

However, it depends on the surgeons experience and is a

subjective sign. Our results showed that BE and RM were

reliable diagnostic investigations and were independent

factors in neonatal HD diagnosis. BE study should be per-

formed in all suspected patients since it can locate the

transition zone and help to classify HD. If the BE result

plus the clinical features make the patients score > 3, it isprobably a typical HD patient and rectal biopsy is not nec-

essary. If the score is 3, further study such as RM or rectalbiopsy needs to be performed to exclude HD. We prefer

doing the noninvasive RM study first. With this addi-

tional study, a total score > 3 will support the HD diagno-sis, otherwise, a total score 3 suggests that rectal biopsyneeds to be performed. This scoring system helped us to

approach suspected HD neonates rationally using symp-

toms, signs and noninvasive investigations to arrive at a

decision on whether rectal biopsy should be taken. This

proved to be very useful in our second study period in

which the rate of rectal biopsy was reduced to as low as

20% of the suspected cases and the overall HD diagnostic

accuracy was nearly 100%. Short-segment HD is easy to

be confused with non-HD diseases on BE or RM study.

Rectal biopsy should be performed in patients with lower

score for the definitive diagnosis.

Our results have demonstrated that most neonatal

HD patients could be identified by the combined features

on history, physical examination and investigations. HD

diagnostic scoring system may help the paediatric surgeon

to select patients for further invasive investigations such

as suction biopsy or full-thickness biopsy so that unnec-

essary biopsy can be avoided. Biopsies and further inves-

tigative tests are required for those patients with scores

3, short-segment HD has to be excluded.

GUO et al

178 ASIAN JOURNAL OF SURGERY VOL 29 NO 3 JULY 2006

0

10

20

30

40

50

Patie

nts,

n

HD Non-HDDefinitive diagnosis

> 3

3

Figure. Comparison of HD and non-HD neonatal patientsaccording to the diagnostic scoring system. A cut-off point of 3 provided 84% of HD patients with score > 3, whereas 75% of non-HD patients had score 3 (p < 0.05) in the first studyperiod.

DIAGNOSIS OF HIRSCHSPRUNGS DISEASE

ASIAN JOURNAL OF SURGERY VOL 29 NO 3 JULY 2006 179

References

1. Babu R, Camadoo L, Drake DP, et al. Rectal biopsy in the investi-

gation of constipation. Arch Dis Child 1999;81:189.

2. Teitelbaum DH, Cilley RE, Sherman NJ, et al. A decade of experi-

ence with the primary pull-through for Hirschsprung disease in

the newborn period: a multicenter analysis of outcomes. Ann

Surg 2000;232:37280.

3. Warner BW. Single-stage operations for Hirschsprungs disease:

pushing the envelope. Gastroenterology 2001;120:1299301.

4. Osatakul S, Patrapinyokul S, Osatakul N. The diagnostic value

of anorectal manometry as a screening test for Hirschsprungs

disease. J Med Assoc Thai 1999;82:11005.

5. Emir H, Akman M, Sarimurat N, et al. Anorectal mano-

metry during the neonatal period: its specificity in the diag-

nosis of Hirschsprungs disease. Eur J Pediatr Surg 1999;9:

1013.

6. Reid JR, Buonomo C, Moreira C, et al. The barium enema in con-

stipation: comparison with rectal manometry and biopsy to

exclude Hirschsprungs disease after the neonatal period. Pediatr

Radiol 2000;30:6814.

7. Cusick EL, Buick RG. Injury to the common iliac artery during

suction rectal biopsy. J Pediatr Surg 1995;30:1112.

8. Rees BI, Azmy A, Nigam M, et al. Complication of rectal suction

biopsy. J Pediatr Surg 1983;18:2735.

9. Alizai NK, Batcup G, Dixon MF, et al. Rectal biopsy for

Hirschsprungs disease: what is the optimum method? Pediatr

Surg Int 1998;13:1214.

10. Lopez-Alonso M, Ribas J, Hernandez A, et al. Efficiency of the

anorectal manometry for the diagnosis of Hirschsprungs dis-

ease in the newborn period. Eur J Pediatr Surg 1995;5:1603.

11. Abu-Alfa AK, Kuan SF, West AB, et al. Cathepsin D in intes-

tinal ganglion cells. A potential aid to diagnosis in suspected

Hirschsprungs disease. Am J Surg Pathol 1997;21:2015.