- Email: [email protected]
Comparison of functional outcomes of duhamel and transanal endorectal coloanal anastomosis for Hirschsprung’s disease
Comparison of Functional Outcomes of Duhamel and Transanal Endorectal Coloanal Anastomosis for Hirschsprung’s Disease By J.L. Minford, A. Ram, R.R. Turnock, G.L. Lamont, S.E. Kenny, R.J. Rintala, D.A. Lloyd, and C.T. Baillie Liverpool, England
Purpose: The aim of this study was to determine the morbidity and medium-term functional outcome of the Duhamel operation and laparotomy and transanal endorectal coloanal anastomosis (TECA) for Hirschsprung’s disease (HSCR). Methods: The study populations were 34 consecutive children who underwent the Duhamel operation (or Lester Martin modification) and 37 who had the TECA. Demographic details were obtained by case note review, and functional outcome was determined by a combination of outpatient interview, questionnaire, and telephone enquiry. Results: There was no difference between the groups with respect to age, gender, and length of aganglionic segment. Seventy percent presented as neonates (Duhamel, 24 of 34; TECA, 26 of 37). A single-stage primary pull-through was performed in 17 of 37 children in the TECA group, and in 1 of 34 from the Duhamel group. There was a single perioperative death in the Duhamel group and an unrelated, late death in the TECA group. Postoperative enterocolitis was seen in 13 of 37 TECA children and in a single child from the Duhamel group. A stricture of the pull-through segment was seen in 7 of 37 children after TECA and required temporary
S
UBSEQUENT TO SWENSON’S description of his definitive operation for Hirschsprung’s disease in 19481 controversy has existed regarding the best operative treatment. Until 1995, the Duhamel procedure was the favoured operation at Alder Hey. However, longterm results were disappointing with a failure rate of 10% and a satisfactory functional outcome in only 42% of children.2 This prompted a review of their management and a move toward an alternative approach. In 1993, Rintala and Lindahl3 described the transanal, endorectal, coloanal anastomosis (TECA). This had previously been used in adults for the treatment of malignancy, ulcerative colitis, and polyposis coli and was adapted as a pull-through procedure for Hirschsprung’s disease. This method is similar to the Soave-Boley pullthrough, combining a laparotomy with transanal mucosectomy and results in a shorter rectal muscular cuff. The TECA was adopted as the preferred operation at Alder Hey starting in 1995. In 2001, Shankar et al4 reported early outcome data from 2 institutions. Initial results were favourable with normal bowel function in 76% of children, but the number at our institution in whom continence could be assessed was small. This study reports medium-term outcome of the TECA proJournal of Pediatric Surgery, Vol 39, No 2 (February), 2004: pp 161-165
diversion in 2 of 9. Late division of a rectal spur was required in 6 of 33 Duhamel children. Requirement for late myectomy was the same in both groups (Duhamel 3 of 33, TECA 4 of 37). Complications requiring stoma formation occurred in 5 of 37 after TECA and 2 of 33 after the Duhamel operation. Two children from the TECA group and 1 from the Duhamel group remain diverted. One child from each group required a re–pull-through procedure. Two patients were lost to follow-up in the TECA group, leaving 34 children in this group and 33 in the Duhamel group in whom functional outcome could be assessed. Functional outcome was similar in the 2 groups. Conclusions: TECA and Duhamel procedures have similar medium-term functional outcomes. TECA has a high incidence of postoperative enterocolitis and transient stricture formation but is suitable for single-stage neonatal treatment of HSCR. J Pediatr Surg 39:161-165. © 2004 Elsevier Inc. All rights reserved. INDEX WORDS: Hirschsprung’s diseases, transanal endorectal pull-through, Duhamel pull-through.
cedure comparing results with an age-matched group of children who underwent the Duhamel procedure. MATERIALS AND METHODS
TECA Group Thirty-seven consecutive children identified for the period from January 1995 to December 1998 formed the TECA group. Laparotomy with mobilisation of the colon and rectum was performed before transanal mucosectomy and coloanal anastomosis.
Duhamel Group Thirty-four consecutive children who underwent the Duhamel procedure (with or without the Lester Martin modification) between January 1987 and December 1990 formed the second group. Demographic data including age at presentation and final assess-
From Alder Hey Children’s Hospital and University of Liverpool, Liverpool, England. Presented at the 50th Annual Congress of the British Association of Paediatric Surgeons, Estoril, Portugal, July 15-18, 2003. Address reprint requests to Mr C.T. Baillie, Department of Paediatric Surgery, Alder Hey Children’s Hospital, Eaton Rd, Liverpool L12 2AP, England. © 2004 Elsevier Inc. All rights reserved. 0022-3468/04/3902-0007$30.00/0 doi:10.1016/j.jpedsurg.2003.10.004 161
162
BAKER ET AL
Table 1. Summary of Scoring System 1 2 3 4 5
Enterocolitis
Normal bowel habit Soiling ⬍1/wk Soiling ⬎1/wk Daily soiling or need for enemas ACE, permanent stoma or major revision surgery
Modified from Shankar et al 2001.3
ment, gender, and length of aganglionosis were collected by case note review. The incidence of postoperative complications (enterocolitis, stricture, and rectal spur) and subsequent surgical procedures were noted. Enterocolitis was defined as abdominal distension associated with loose offensive stool and general malaise, treated by rectal washout, with or without dilatation and oral or intravenous antibiotics.5-7 Stricture was defined as a narrowing of the anastomosis or neorectum, requiring treatment in the form of repeated dilatations, stricturoplasty, or diversion. An analogue scoring system (Table 1)4 was used to provide a functional outcome score for each child, using a combination of questionnaire follow-up, patient interview, and case note review. Satisfactory outcome was defined as a score of 1 or 2. Poor outcome was defined as 3, 4, or 5.
Statistical Analysis Data were analysed using SPSS for Windows (V.6.1.2 SPSS Inc, Chicago, IL). Mann Whitney U test and Fishers Exact test were used to compare nonparametric variables. Multivariate analysis was used to assess potential risk factors. P values of less than .05 were considered significant. Data are presented as median (range) unless otherwise stated.
RESULTS
Patient Demographics There was no significant difference in age at completion of functional outcome score between the 2 groups (TECA, 6 years [range 4 to 11]; Duhamel, 7 years [range 4 to 9]; P ⫽ .22; Mann Whitney U test). Similarly, there were no significant differences in gender, length of aganglionic segment (rectosigmoid, long segment, total colonic) or percentage neonatal presentation (Table 2). A single-stage pull-through was performed in 17 of 37 (45%) of the TECA group and 1 of 34 (3%) of the Duhamel group (P ⫽ .0001; Fishers Exact test). Morbidity and Mortality There was 1 perioperative death in the Duhamel group and 1 late, unrelated death from cardiac disease in the TECA group (Table 3).
The frequency of postoperative enterocolitis was significantly higher in the TECA group, occurring in 13 of 37 (35%) children compared with 1 of 33 (3%) of the Duhamel group (P ⫽ .001; multivariate). Across both groups, single-stage procedure was a positive predictor of enterocolitis (7 of 18 [38%] children after single-stage procedure versus 7 of 52 [14%]; P ⫽ .02, multivariate). Gender and length of aganglionosis showed no significant association with enterocolitis (P ⫽ .9; P ⫽ .57, respectively; multivariate). Stricture A stricture occurred in 7 of 37 (19%) patients in the TECA group and no patients in the Duhamel group (P ⫽ 0.01; multivariate). One child required 2 attempts at stricturoplasty (anoplasty and transperitoneal stricturoplasty), and his planned stoma closure was delayed (eventual functional outcome score 3); a second child responded to formation of a stoma and repeated dilatations (final functional outcome score 4). All other children responded to dilatations. Rectal Spur This well-recognized complication of the Duhamel procedure was seen in 6 of 33 (18%) children in that group. Transanal division of the spur using a stapling device was performed in all; 1 child required 2 such procedures. Five of these children had poor functional outcome (score ⱖ3). Myectomy Late myectomy was performed in 4 of 37 children in the TECA group and 3 of 33 children in the Duhamel group for symptoms suggesting outlet obstruction (presumed sphincter achalasia). Final functional outcome scores were satisfactory in 2 of 4 of the TECA group and 1 of 3 of the Duhamel group. Late Stoma Formation and Operative Failure Late stoma formation was required in 5 children in the TECA group and 2 in the Duhamel group (P ⫽ .44 Fishers exact). In the TECA group this was temporary in 3 children (for stricture, enterocolitis, and colovaginal
Table 2. Demographic Details of Duhamel and TECA Groups
Age at scoring (yr) Male to female ratio Rectosigmoid:long segment:total colonic Neonatal presentation
Duhamel Group (n ⫽ 34)
TECA Group (n ⫽ 37)
P Value
7 (4-9) 25:9 (2.7:1) 24:7:3 24/34 (71%)
6 (4-11) 27:10 (2.7:1) 27:8:2 26/37 (70%)
.22 .96 .95 .76
OUTCOMES AFTER DUHAMEL AND ENDORECTAL PULL-THROUGH
Table 3. Incidence of Complications in the Duhamel and TECA Groups Duhamel Group (n ⫽ 34)
Enterocolitis Stricture (requiring diversion) Rectal spur Myectomy Stoma formation
1/33 (3%) 0/33 6/33 (18%) 3/33 (9%) 2/33 (6%)
TECA Group (n ⫽ 37)
13/37 (35%)
P Value
.001
7/37 (24%) (2/7)
.01
4/37 (11%) 5/37 (14%)
.56 .44
fistula). A successful repeat pull-through procedure was performed in the child with colovaginal fistula (eventual functional outcome 3). Two children remain diverted, both with a stoma for enterocolitis; in 1 of these a repeat TECA is planned. In the Duhamel group, a late stoma was formed for technical failure at initial operation (1) and severe constipation (1). The former has been closed after ileo-anal J pouch anastomosis (final functional outcome 1). Functional Outcome Functional outcome data were obtainable for 33 (100%) children in the Duhamel group and 34 (94%) children in the TECA group. Two children in the latter group were lost to follow-up (Table 4). Functional outcomes were similar for the 2 groups (Fig 1). A satisfactory outcome was seen in 16 of 33 (48%) Duhamel children and 14 of 34 (41%) TECA children (P ⫽ .08; Fishers Exact test). There was no significant difference between scores for the TECA sub-
Fig 1. group.
Functional continence scores for Duhamel and TECA
163
Table 4. Functional Outcome in the Duhamel and TECA Groups Functional Continence Score
Duhamel Group (n ⫽ 33)
TECA Group (n ⫽ 34)
1 2 3 4 5
12 4 9 6 2
7 7 7 10 3
groups of 1-stage and staged operation (P ⫽ .48; Fishers Exact test). Gender, length of aganglionic segment, and neonatal diagnosis were not associated with a worse outcome (P ⫽ .44; P ⫽ .72; P ⫽ .37, respectively; multivariate). DISCUSSION
A number of operative strategies have been described for Hirschsprung’s disease. All are perceived to have relative merits and weaknesses, supported in some cases by medium and long-term outcome data.2,8-10 Evaluation of different techniques is complex, and results of large single-surgeon series may reflect the merits of the surgeon rather than the operation. Randomised controlled trials are not feasible, and, where an institution has adopted a new technique, only historical comparison is possible. In contrast to establishing the safety of a new technique, assessment of functional outcome is complicated. Lack of consensus regarding socially acceptable norms, relative insensitivity of questionnaire tools, and lack of a generally accepted functional continence score all contribute to this. The inevitable delay between treatment and assessment of continence at age 4 prolongs the process of comparison. Nevertheless, along with changes in practice comes the responsibility to audit clinical outcome, and this demands robust data regarding complications and functional outcome from well-matched groups. We have chosen to use a modified analogue score to provide objective data for comparison. Using this scoring system there is no significant difference in functional continence between our 2 study groups. The incidence of post–pull-through enterocolitis reported in the literature varies widely, with some studies reporting rates as high as 32% to 42%.7,10 We have found enterocolitis to be significantly more frequent in the TECA group (35%) compared with the Duhamel group (3%). There are a number of possible explanations for this apparent increase in incidence. The diagnosis of enterocolitis remains subjective. Changing criteria or thresholds for making the diagnosis could account for differences in rates. Alterations in intestinal flora over time could play an aetiological role. The rate of enterocolitis in this series is disappointing given the theoretical advantage of the shorter muscular cuff in the TECA.
164
BAKER ET AL
Evolving treatment strategies for Hirschsprung’s-associated enterocolitis, such as selective digestive decontamination, probiotic supplementation, and sodium cromoglycate,11 combined with the more usual treatment of antibiotics, rectal dilatation, and washout, may decrease the severity and frequency of episodes. It remains to be seen if enterocolitis affects functional outcome in the long term. Single-stage primary neonatal pull-through without formation of colostomy is gaining popularity,5,12 with possible benefits in the development of early anorectal function. Avoidance of stoma-associated complications and reduced hospital stay are other potential advantages. Although the TECA approach is easily applicable to single-stage surgery, the development of smaller stapling devices has enabled neonatal one-stage Duhamel procedures to be performed with good initial results.13,14 The development of laparoscopic-assisted,15 and an entirely transanal TECA16 offer potential advantages of reduced
intraabdominal complications and improved cosmesis. Effects on functional outcome have yet to be shown. The finding in this series that single-stage surgery is an independent risk factor for the development of enterocolitis is of some concern. However, there was no difference in eventual functional outcome between singlestage and staged procedures. We conclude that the TECA is suitable for the singlestage neonatal treatment of Hirschsprung’s disease, with similar medium-term functional outcome to the Duhamel procedure. TECA has a higher incidence of postoperative enterocolitis and transient stricture formation. Further investigation and follow-up are required to determine whether these complications affect long-term functional outcome. This study clearly illustrates the need for thorough, structured follow-up of large cohorts of children with Hirschsprung’s disease to determine the best treatment modality.
REFERENCES 1. Swenson O, Bill AH: Resection of the rectum and rectosigmoid with preservation of the sphincter for benign spastic lesions producing megacolon: An experimental study. Surgery 24:212-220, 1948 2. Baillie CT, Kenny SE, Rintala RJ, et al: Long-term outcome and colonic motility after Duhamel procedure for Hirschsprung’s disease. J Pediatr Surg 34:325-329, 1999 3. Rintala RJ, Lindahl H: Transanal endorectal coloanal anastomosis for Hirschsprung’s disease. Pediatr Surg Int 8:128-131, 1993 4. Shankar KR, Losty PD, Lamont GL, et al: Transanal endorectal coloanal surgery for Hirschsprung’s disease: Experience in two centers. J Pediatr Surg 35:1209-1213, 2000 5. Teitelbaum DH, Coran AG: Enterocolitis. Semin Pediatr Surg 7:162-169, 1998 6. Coran AG, Teitelbaum DH: Recent advances in the management of Hirschsprung’s disease. Am J Surg 180:382-387, 2000 7. Hackham DJ, Filler RM, Pearl RH: Enterocolitis after the surgical treatment of Hirschsprung’s disease: Risk factors and financial impact. J Pediatr Surg 33:830-833, 1998 8. Bai Y, Chen H, Hao J, et al: Long-term outcome and quality of life after the Swenson procedure for Hirschsprung’s disease. J Pediatr Surg 37:639-642, 2002
9. Bourdelat D, Vransky P, Page´ s R, et al: Duhamel operation 40 years after: A multicentric study. Eur J Pediatr Surg 7:70-76, 1997 10. Teitelbaum DH, Cilley RE, Sherman NJ, et al: A decade of experience with the primary pull-through for Hirschsprung’s disease in the newborn period. Ann Surg 232:372-380, 2000 11. Rintala RJ, Lindahl H: Sodium cromoglycate in the management of chronic or recurrent enterocolitis in patients with Hirschsprung’s disease. J Pediatr Surg 36:1032-1035, 2001 12. Bianchi A: One stage neonatal reconstruction without stoma for Hirschsprung’s disease. Semin Pediatr Surg 7:170-173, 1998 13. van der Zee DC, Bax KN: One-stage Duhamel-Martin procedure for Hirschsprung’s disease: A 5-year follow-up study. J Pediatr Surg: 1434-1436, 2000 14. Mir E, Karaca I, Gu¨ nsar C, et al: Primary Duhamel-Martin operations in neonates and infants. Pediatr Int 43:405-408, 2001 15. Georgeson KE, Cohen RD, Hebra A, et al: Primary laparoscopic-assisted endorectal colon pull-through for Hirschsprung’s disease: A new gold standard. Ann Surg 229:678-682, 1999 16. Langer JC, Seifert M, Minkes RK, et al: One-stage Soave pull-through for Hirschsprung’s disease: A comparison of the transanal and open approaches. J Pediatr Surg 35:820-822, 2000
Discussion A. Coran (Ann Arbor, MI): I believe that the use of historical groups creates problems. The major problem is that when you talk about enterocolitis it is as if it is in the eyes of the beholder, like beauty. People don’t recognise enterocolitis and haven’t recognised enterocolitis until very recently. Even though Dr Swanson described it in 1956 with Dr Chapman, over the next 20 years the incidence of enterocolitis that was reported in the various series was very low. It is only recently as people began to look for it that they began to realise that its incidence is very much higher after any form of pull-through. Since you have an historical group in the Duhamel that was
done in an earlier period, I really don’t believe the number that you have for enterocolitis. I am sure that it is much higher, it just wasn’t recognised by the observers. I think, therefore, that it is a long deductive leap to take to say that the new operation (transanal) has a higher incidence. In fact, we did a study that was reported in the Annals of Surgery a couple of years ago of 80 consecutive newborns undergoing a primary pull-through in the newborn period, some were transanal at that time and some were done transabdominally, and we compared these with 103 patients, done at the same time, with a staged procedure, all with the endorectal. The incidence
OUTCOMES AFTER DUHAMEL AND ENDORECTAL PULL-THROUGH
165
of entercolitis was really statistically not different, and we took into account the differences between observers’ definition of enterocolitis. What do you define as enterocolitis, and how did you grade it? J.L. Baker (response): I am very reassured by your comments. We defined enterocolitis as abdominal distension, general unwellness with our without a pyrexia, and any other symptoms that could be attributed to enterocolitis. I must admit our definition was very soft. We defined this group as patients who were treated for enterocolitis. So I agree with you that our threshold for diagnosis in this group was very low. We didn’t have a specific grading system, but I think looking at our results, we included many patients with very mild enterocolitis who would have been missed out of the previous Duhamel group. A. Coran (Ann Arbor, MI): If you look at a couple of papers from 1995 in the Journal of Pediatric Surgery you will see that there are histologic definitions of enterocolitis that one could use, and, in fact, you could actually go back with your series, both the Duhamel and a transanal, and look for the various histologic grades, which go from 1 to 5, and a paper that we wrote in 1995, and get a handle on that issue, more accurate data.
J.L. Baker (response): I have read that paper and it is referred to in the paper we have written. However, a very small number of these children had actually got operative specimens and would have required biopsy to make that diagnosis. A. Hadidi (Cairo, Egypt): I am afraid that many the people in the audience may get confused when you say TECA. Is this technique similar to the totally transanal approach described by De la Torre in Mexico, or a different technique? My impression is that it is a different technique as I understood from the Liverpool team. I think that people need to know that. J.L. Baker (response): The procedure that most of these children underwent was a combined abdominal procedure similar to a SOAVE with a transanal dissection of the mucosal cuff. A. Hadidi (Cairo, Egypt): In our Unit in Egypt, we have operated on more than 110 patients with the total transanal approach and probably our criteria for enterocolitis is different, but those who needed admission are less than 3%. Do you split the mucosal cuff or not? If you leave it intact, this may explain the high incidence of enterocolitis.
Purpose: The aim of this study was to determine the morbidity and medium-term functional outcome of the Duhamel operation and laparotomy and transanal endorectal coloanal anastomosis (TECA) for Hirschsprung’s disease (HSCR). Methods: The study populations were 34 consecutive children who underwent the Duhamel operation (or Lester Martin modification) and 37 who had the TECA. Demographic details were obtained by case note review, and functional outcome was determined by a combination of outpatient interview, questionnaire, and telephone enquiry. Results: There was no difference between the groups with respect to age, gender, and length of aganglionic segment. Seventy percent presented as neonates (Duhamel, 24 of 34; TECA, 26 of 37). A single-stage primary pull-through was performed in 17 of 37 children in the TECA group, and in 1 of 34 from the Duhamel group. There was a single perioperative death in the Duhamel group and an unrelated, late death in the TECA group. Postoperative enterocolitis was seen in 13 of 37 TECA children and in a single child from the Duhamel group. A stricture of the pull-through segment was seen in 7 of 37 children after TECA and required temporary
S
UBSEQUENT TO SWENSON’S description of his definitive operation for Hirschsprung’s disease in 19481 controversy has existed regarding the best operative treatment. Until 1995, the Duhamel procedure was the favoured operation at Alder Hey. However, longterm results were disappointing with a failure rate of 10% and a satisfactory functional outcome in only 42% of children.2 This prompted a review of their management and a move toward an alternative approach. In 1993, Rintala and Lindahl3 described the transanal, endorectal, coloanal anastomosis (TECA). This had previously been used in adults for the treatment of malignancy, ulcerative colitis, and polyposis coli and was adapted as a pull-through procedure for Hirschsprung’s disease. This method is similar to the Soave-Boley pullthrough, combining a laparotomy with transanal mucosectomy and results in a shorter rectal muscular cuff. The TECA was adopted as the preferred operation at Alder Hey starting in 1995. In 2001, Shankar et al4 reported early outcome data from 2 institutions. Initial results were favourable with normal bowel function in 76% of children, but the number at our institution in whom continence could be assessed was small. This study reports medium-term outcome of the TECA proJournal of Pediatric Surgery, Vol 39, No 2 (February), 2004: pp 161-165
diversion in 2 of 9. Late division of a rectal spur was required in 6 of 33 Duhamel children. Requirement for late myectomy was the same in both groups (Duhamel 3 of 33, TECA 4 of 37). Complications requiring stoma formation occurred in 5 of 37 after TECA and 2 of 33 after the Duhamel operation. Two children from the TECA group and 1 from the Duhamel group remain diverted. One child from each group required a re–pull-through procedure. Two patients were lost to follow-up in the TECA group, leaving 34 children in this group and 33 in the Duhamel group in whom functional outcome could be assessed. Functional outcome was similar in the 2 groups. Conclusions: TECA and Duhamel procedures have similar medium-term functional outcomes. TECA has a high incidence of postoperative enterocolitis and transient stricture formation but is suitable for single-stage neonatal treatment of HSCR. J Pediatr Surg 39:161-165. © 2004 Elsevier Inc. All rights reserved. INDEX WORDS: Hirschsprung’s diseases, transanal endorectal pull-through, Duhamel pull-through.
cedure comparing results with an age-matched group of children who underwent the Duhamel procedure. MATERIALS AND METHODS
TECA Group Thirty-seven consecutive children identified for the period from January 1995 to December 1998 formed the TECA group. Laparotomy with mobilisation of the colon and rectum was performed before transanal mucosectomy and coloanal anastomosis.
Duhamel Group Thirty-four consecutive children who underwent the Duhamel procedure (with or without the Lester Martin modification) between January 1987 and December 1990 formed the second group. Demographic data including age at presentation and final assess-
From Alder Hey Children’s Hospital and University of Liverpool, Liverpool, England. Presented at the 50th Annual Congress of the British Association of Paediatric Surgeons, Estoril, Portugal, July 15-18, 2003. Address reprint requests to Mr C.T. Baillie, Department of Paediatric Surgery, Alder Hey Children’s Hospital, Eaton Rd, Liverpool L12 2AP, England. © 2004 Elsevier Inc. All rights reserved. 0022-3468/04/3902-0007$30.00/0 doi:10.1016/j.jpedsurg.2003.10.004 161
162
BAKER ET AL
Table 1. Summary of Scoring System 1 2 3 4 5
Enterocolitis
Normal bowel habit Soiling ⬍1/wk Soiling ⬎1/wk Daily soiling or need for enemas ACE, permanent stoma or major revision surgery
Modified from Shankar et al 2001.3
ment, gender, and length of aganglionosis were collected by case note review. The incidence of postoperative complications (enterocolitis, stricture, and rectal spur) and subsequent surgical procedures were noted. Enterocolitis was defined as abdominal distension associated with loose offensive stool and general malaise, treated by rectal washout, with or without dilatation and oral or intravenous antibiotics.5-7 Stricture was defined as a narrowing of the anastomosis or neorectum, requiring treatment in the form of repeated dilatations, stricturoplasty, or diversion. An analogue scoring system (Table 1)4 was used to provide a functional outcome score for each child, using a combination of questionnaire follow-up, patient interview, and case note review. Satisfactory outcome was defined as a score of 1 or 2. Poor outcome was defined as 3, 4, or 5.
Statistical Analysis Data were analysed using SPSS for Windows (V.6.1.2 SPSS Inc, Chicago, IL). Mann Whitney U test and Fishers Exact test were used to compare nonparametric variables. Multivariate analysis was used to assess potential risk factors. P values of less than .05 were considered significant. Data are presented as median (range) unless otherwise stated.
RESULTS
Patient Demographics There was no significant difference in age at completion of functional outcome score between the 2 groups (TECA, 6 years [range 4 to 11]; Duhamel, 7 years [range 4 to 9]; P ⫽ .22; Mann Whitney U test). Similarly, there were no significant differences in gender, length of aganglionic segment (rectosigmoid, long segment, total colonic) or percentage neonatal presentation (Table 2). A single-stage pull-through was performed in 17 of 37 (45%) of the TECA group and 1 of 34 (3%) of the Duhamel group (P ⫽ .0001; Fishers Exact test). Morbidity and Mortality There was 1 perioperative death in the Duhamel group and 1 late, unrelated death from cardiac disease in the TECA group (Table 3).
The frequency of postoperative enterocolitis was significantly higher in the TECA group, occurring in 13 of 37 (35%) children compared with 1 of 33 (3%) of the Duhamel group (P ⫽ .001; multivariate). Across both groups, single-stage procedure was a positive predictor of enterocolitis (7 of 18 [38%] children after single-stage procedure versus 7 of 52 [14%]; P ⫽ .02, multivariate). Gender and length of aganglionosis showed no significant association with enterocolitis (P ⫽ .9; P ⫽ .57, respectively; multivariate). Stricture A stricture occurred in 7 of 37 (19%) patients in the TECA group and no patients in the Duhamel group (P ⫽ 0.01; multivariate). One child required 2 attempts at stricturoplasty (anoplasty and transperitoneal stricturoplasty), and his planned stoma closure was delayed (eventual functional outcome score 3); a second child responded to formation of a stoma and repeated dilatations (final functional outcome score 4). All other children responded to dilatations. Rectal Spur This well-recognized complication of the Duhamel procedure was seen in 6 of 33 (18%) children in that group. Transanal division of the spur using a stapling device was performed in all; 1 child required 2 such procedures. Five of these children had poor functional outcome (score ⱖ3). Myectomy Late myectomy was performed in 4 of 37 children in the TECA group and 3 of 33 children in the Duhamel group for symptoms suggesting outlet obstruction (presumed sphincter achalasia). Final functional outcome scores were satisfactory in 2 of 4 of the TECA group and 1 of 3 of the Duhamel group. Late Stoma Formation and Operative Failure Late stoma formation was required in 5 children in the TECA group and 2 in the Duhamel group (P ⫽ .44 Fishers exact). In the TECA group this was temporary in 3 children (for stricture, enterocolitis, and colovaginal
Table 2. Demographic Details of Duhamel and TECA Groups
Age at scoring (yr) Male to female ratio Rectosigmoid:long segment:total colonic Neonatal presentation
Duhamel Group (n ⫽ 34)
TECA Group (n ⫽ 37)
P Value
7 (4-9) 25:9 (2.7:1) 24:7:3 24/34 (71%)
6 (4-11) 27:10 (2.7:1) 27:8:2 26/37 (70%)
.22 .96 .95 .76
OUTCOMES AFTER DUHAMEL AND ENDORECTAL PULL-THROUGH
Table 3. Incidence of Complications in the Duhamel and TECA Groups Duhamel Group (n ⫽ 34)
Enterocolitis Stricture (requiring diversion) Rectal spur Myectomy Stoma formation
1/33 (3%) 0/33 6/33 (18%) 3/33 (9%) 2/33 (6%)
TECA Group (n ⫽ 37)
13/37 (35%)
P Value
.001
7/37 (24%) (2/7)
.01
4/37 (11%) 5/37 (14%)
.56 .44
fistula). A successful repeat pull-through procedure was performed in the child with colovaginal fistula (eventual functional outcome 3). Two children remain diverted, both with a stoma for enterocolitis; in 1 of these a repeat TECA is planned. In the Duhamel group, a late stoma was formed for technical failure at initial operation (1) and severe constipation (1). The former has been closed after ileo-anal J pouch anastomosis (final functional outcome 1). Functional Outcome Functional outcome data were obtainable for 33 (100%) children in the Duhamel group and 34 (94%) children in the TECA group. Two children in the latter group were lost to follow-up (Table 4). Functional outcomes were similar for the 2 groups (Fig 1). A satisfactory outcome was seen in 16 of 33 (48%) Duhamel children and 14 of 34 (41%) TECA children (P ⫽ .08; Fishers Exact test). There was no significant difference between scores for the TECA sub-
Fig 1. group.
Functional continence scores for Duhamel and TECA
163
Table 4. Functional Outcome in the Duhamel and TECA Groups Functional Continence Score
Duhamel Group (n ⫽ 33)
TECA Group (n ⫽ 34)
1 2 3 4 5
12 4 9 6 2
7 7 7 10 3
groups of 1-stage and staged operation (P ⫽ .48; Fishers Exact test). Gender, length of aganglionic segment, and neonatal diagnosis were not associated with a worse outcome (P ⫽ .44; P ⫽ .72; P ⫽ .37, respectively; multivariate). DISCUSSION
A number of operative strategies have been described for Hirschsprung’s disease. All are perceived to have relative merits and weaknesses, supported in some cases by medium and long-term outcome data.2,8-10 Evaluation of different techniques is complex, and results of large single-surgeon series may reflect the merits of the surgeon rather than the operation. Randomised controlled trials are not feasible, and, where an institution has adopted a new technique, only historical comparison is possible. In contrast to establishing the safety of a new technique, assessment of functional outcome is complicated. Lack of consensus regarding socially acceptable norms, relative insensitivity of questionnaire tools, and lack of a generally accepted functional continence score all contribute to this. The inevitable delay between treatment and assessment of continence at age 4 prolongs the process of comparison. Nevertheless, along with changes in practice comes the responsibility to audit clinical outcome, and this demands robust data regarding complications and functional outcome from well-matched groups. We have chosen to use a modified analogue score to provide objective data for comparison. Using this scoring system there is no significant difference in functional continence between our 2 study groups. The incidence of post–pull-through enterocolitis reported in the literature varies widely, with some studies reporting rates as high as 32% to 42%.7,10 We have found enterocolitis to be significantly more frequent in the TECA group (35%) compared with the Duhamel group (3%). There are a number of possible explanations for this apparent increase in incidence. The diagnosis of enterocolitis remains subjective. Changing criteria or thresholds for making the diagnosis could account for differences in rates. Alterations in intestinal flora over time could play an aetiological role. The rate of enterocolitis in this series is disappointing given the theoretical advantage of the shorter muscular cuff in the TECA.
164
BAKER ET AL
Evolving treatment strategies for Hirschsprung’s-associated enterocolitis, such as selective digestive decontamination, probiotic supplementation, and sodium cromoglycate,11 combined with the more usual treatment of antibiotics, rectal dilatation, and washout, may decrease the severity and frequency of episodes. It remains to be seen if enterocolitis affects functional outcome in the long term. Single-stage primary neonatal pull-through without formation of colostomy is gaining popularity,5,12 with possible benefits in the development of early anorectal function. Avoidance of stoma-associated complications and reduced hospital stay are other potential advantages. Although the TECA approach is easily applicable to single-stage surgery, the development of smaller stapling devices has enabled neonatal one-stage Duhamel procedures to be performed with good initial results.13,14 The development of laparoscopic-assisted,15 and an entirely transanal TECA16 offer potential advantages of reduced
intraabdominal complications and improved cosmesis. Effects on functional outcome have yet to be shown. The finding in this series that single-stage surgery is an independent risk factor for the development of enterocolitis is of some concern. However, there was no difference in eventual functional outcome between singlestage and staged procedures. We conclude that the TECA is suitable for the singlestage neonatal treatment of Hirschsprung’s disease, with similar medium-term functional outcome to the Duhamel procedure. TECA has a higher incidence of postoperative enterocolitis and transient stricture formation. Further investigation and follow-up are required to determine whether these complications affect long-term functional outcome. This study clearly illustrates the need for thorough, structured follow-up of large cohorts of children with Hirschsprung’s disease to determine the best treatment modality.
REFERENCES 1. Swenson O, Bill AH: Resection of the rectum and rectosigmoid with preservation of the sphincter for benign spastic lesions producing megacolon: An experimental study. Surgery 24:212-220, 1948 2. Baillie CT, Kenny SE, Rintala RJ, et al: Long-term outcome and colonic motility after Duhamel procedure for Hirschsprung’s disease. J Pediatr Surg 34:325-329, 1999 3. Rintala RJ, Lindahl H: Transanal endorectal coloanal anastomosis for Hirschsprung’s disease. Pediatr Surg Int 8:128-131, 1993 4. Shankar KR, Losty PD, Lamont GL, et al: Transanal endorectal coloanal surgery for Hirschsprung’s disease: Experience in two centers. J Pediatr Surg 35:1209-1213, 2000 5. Teitelbaum DH, Coran AG: Enterocolitis. Semin Pediatr Surg 7:162-169, 1998 6. Coran AG, Teitelbaum DH: Recent advances in the management of Hirschsprung’s disease. Am J Surg 180:382-387, 2000 7. Hackham DJ, Filler RM, Pearl RH: Enterocolitis after the surgical treatment of Hirschsprung’s disease: Risk factors and financial impact. J Pediatr Surg 33:830-833, 1998 8. Bai Y, Chen H, Hao J, et al: Long-term outcome and quality of life after the Swenson procedure for Hirschsprung’s disease. J Pediatr Surg 37:639-642, 2002
9. Bourdelat D, Vransky P, Page´ s R, et al: Duhamel operation 40 years after: A multicentric study. Eur J Pediatr Surg 7:70-76, 1997 10. Teitelbaum DH, Cilley RE, Sherman NJ, et al: A decade of experience with the primary pull-through for Hirschsprung’s disease in the newborn period. Ann Surg 232:372-380, 2000 11. Rintala RJ, Lindahl H: Sodium cromoglycate in the management of chronic or recurrent enterocolitis in patients with Hirschsprung’s disease. J Pediatr Surg 36:1032-1035, 2001 12. Bianchi A: One stage neonatal reconstruction without stoma for Hirschsprung’s disease. Semin Pediatr Surg 7:170-173, 1998 13. van der Zee DC, Bax KN: One-stage Duhamel-Martin procedure for Hirschsprung’s disease: A 5-year follow-up study. J Pediatr Surg: 1434-1436, 2000 14. Mir E, Karaca I, Gu¨ nsar C, et al: Primary Duhamel-Martin operations in neonates and infants. Pediatr Int 43:405-408, 2001 15. Georgeson KE, Cohen RD, Hebra A, et al: Primary laparoscopic-assisted endorectal colon pull-through for Hirschsprung’s disease: A new gold standard. Ann Surg 229:678-682, 1999 16. Langer JC, Seifert M, Minkes RK, et al: One-stage Soave pull-through for Hirschsprung’s disease: A comparison of the transanal and open approaches. J Pediatr Surg 35:820-822, 2000
Discussion A. Coran (Ann Arbor, MI): I believe that the use of historical groups creates problems. The major problem is that when you talk about enterocolitis it is as if it is in the eyes of the beholder, like beauty. People don’t recognise enterocolitis and haven’t recognised enterocolitis until very recently. Even though Dr Swanson described it in 1956 with Dr Chapman, over the next 20 years the incidence of enterocolitis that was reported in the various series was very low. It is only recently as people began to look for it that they began to realise that its incidence is very much higher after any form of pull-through. Since you have an historical group in the Duhamel that was
done in an earlier period, I really don’t believe the number that you have for enterocolitis. I am sure that it is much higher, it just wasn’t recognised by the observers. I think, therefore, that it is a long deductive leap to take to say that the new operation (transanal) has a higher incidence. In fact, we did a study that was reported in the Annals of Surgery a couple of years ago of 80 consecutive newborns undergoing a primary pull-through in the newborn period, some were transanal at that time and some were done transabdominally, and we compared these with 103 patients, done at the same time, with a staged procedure, all with the endorectal. The incidence
OUTCOMES AFTER DUHAMEL AND ENDORECTAL PULL-THROUGH
165
of entercolitis was really statistically not different, and we took into account the differences between observers’ definition of enterocolitis. What do you define as enterocolitis, and how did you grade it? J.L. Baker (response): I am very reassured by your comments. We defined enterocolitis as abdominal distension, general unwellness with our without a pyrexia, and any other symptoms that could be attributed to enterocolitis. I must admit our definition was very soft. We defined this group as patients who were treated for enterocolitis. So I agree with you that our threshold for diagnosis in this group was very low. We didn’t have a specific grading system, but I think looking at our results, we included many patients with very mild enterocolitis who would have been missed out of the previous Duhamel group. A. Coran (Ann Arbor, MI): If you look at a couple of papers from 1995 in the Journal of Pediatric Surgery you will see that there are histologic definitions of enterocolitis that one could use, and, in fact, you could actually go back with your series, both the Duhamel and a transanal, and look for the various histologic grades, which go from 1 to 5, and a paper that we wrote in 1995, and get a handle on that issue, more accurate data.
J.L. Baker (response): I have read that paper and it is referred to in the paper we have written. However, a very small number of these children had actually got operative specimens and would have required biopsy to make that diagnosis. A. Hadidi (Cairo, Egypt): I am afraid that many the people in the audience may get confused when you say TECA. Is this technique similar to the totally transanal approach described by De la Torre in Mexico, or a different technique? My impression is that it is a different technique as I understood from the Liverpool team. I think that people need to know that. J.L. Baker (response): The procedure that most of these children underwent was a combined abdominal procedure similar to a SOAVE with a transanal dissection of the mucosal cuff. A. Hadidi (Cairo, Egypt): In our Unit in Egypt, we have operated on more than 110 patients with the total transanal approach and probably our criteria for enterocolitis is different, but those who needed admission are less than 3%. Do you split the mucosal cuff or not? If you leave it intact, this may explain the high incidence of enterocolitis.