- Email: [email protected]
One-stage neonatal pull-through to treat Hirschsprung's disease
One-Stage Neonatal By Duncan
Pull-Through T. Wilcox,
James Bruce, John Bowen, Manchester, England
0 Introduction: Traditionally Hirschsprung’s disease has been treated by a three-stage procedure. This approach has been associated with stoma problems, three episodes of hospitalization, and a theoretical disadvantage, because of the delayed passage of feces via the rectum, in controlling defecation. The aim of this study was to assess the results of one-stage neonatal pull-through for the treatment of Hirschsprung’s disease. Methods: This was a IO-year retrospective review of all patients treated with a neonatal pullthrough for Hirschsprung’s disease. Information was obtained from the patients’ records. All descriptive data are expressed as mean f standard deviation. Results: Fifty-one patients were treated in the IO-year period studied. The mean gestational age was 39.6 f 1.7 weeks, weight was 3.3 + 0.54 kg, with 76% being boys. The surgery was performed at 10.3 f 5.8 days of age. A Swenson procedure was carried out in 38 patients and a Soave in 13. The extent of the disease was rectosigmoid in 39, and the remaining were sigmoid, descending colon, and total colonic. The mean length of the surgery was 181 f 40 minutes. Blood transfusions were required in 13 patients with a mean volume of 55 mL given. The complications observed were: an anastomotic leak in two cases, requiring laparotomy and colostomy (both in patients with Down’s syndrome), and wound infection in three cases. There were two delayed deaths both associated with other congenital anomalies. In total, three patients required further surgery; two had colostomies, one requiring a redo pull-through and one had an ileostomy for enterocolitis. The total length of hospital stay was 20.4 + 11.2 days. Mean follow-up was 3.73 f 2.9 years. In 21 patients the follow-up was greater than 4 years and in this group none were totally incontinent. Four patients complained of soiling at least once a week, and 17 had normal bowel control. Conclusion: Hirschsprung’s disease can be successfully treated in the neonatal period with a one-stage pull-through. The shortand long-term results are as good as those with the three-stage procedure, with the child usually benefitting by not having a stoma and a shorter hospital stay. Copyright Q 1997 by W. B. Saunders Company INDEX Soave.
WORDS:
Hirschsprung’s
disease,
neonate,
Swenson,
S
INCE THE RECONSTRUCTIVE operation to treat Hirschsprung’s disease was first described by Swenson and Bill in 1948, neonates have been treated soon after diagnosis with a defunctioning colostomy.’ When patients were older a definitive pull-through procedure was performed followed by colostomy closure. This three-stage procedure has been advocated because early reports suggestedit was unsafe to perform a definitive operation during the neonatal period.* With advancesin neonatal, anaesthetic,and surgical care, a safe one-stage definitive procedure has become possible.3.JHere we report a IO-year experience of Hirschsprung’s disease
Journal
ofPediatric
Surgery,
Vol 32, No 2 (February),
to Treat Hirschsprung’s
1997: pp 243-247
and Adrian
Disease
Bianchi
treated, during the neonatal period. with a one-stage pull-through. MATERIALS
AND METHODS
This is a retrospective review of patients’ notes. All neonates diagnosed with Hrrschsprung’s disease between 1984 and 1994 treated by two consultants were included. Informatton was obtained from the patients’ notes, and all children had been reviewed wtthin the last 12 months. Diagnosis was achieved in all patients by histological examination of a rectal biopsy. Patrents m whom the diagnosis was suspected or confirmed received preoperative antibiotics (usually Augmentin) and 4 to 6 hourly rectal decompresston (by rectal stimulation). At a convenient time, and when the child’s general health had returned to normal, a primary pull-through procedure was performed. Two pull-through procedures were used in this series, the Swenson’ and the Soave (endorectal pull-through).j Early in the series or when there was concern over the anastomosts, a covering colostomy or tube cecostomy was performed. For the pull-through, patients were placed m a supine lithotomy position, with the perineum fully exposed at the end of the operating table. The abdomen was entered through a transverse lower abdommal mciston and a neonatal Denis Browne retractor applied. Muscle biopsy spectmens were taken from the anttmesentertc surface of the colon, wtthout entering the lumen, startmg in the rectum and workmg proximally. The bropsy results were assessed mtraoperatively by frozen section until normal ganghonated bowel was Identified. In the Swenson procedure the rectosigmoid was devascularized with bipolar diathermy keepmg close to the bowel wall. Dtssection was down to the level of the external sphincter posteriorly and laterally; anteriorly dissectton stopped approximately I .5 cm more proxtmal. Sufftcrent length was obtained by dtvrding the infenor mesenteric pedicle, being careful to preserve the marginal vessels The level to be resected was marked wrth silk sutures and the bowel was mtussuscepted through the anus wtth a Babcock forceps. In the endorectal pull-through procedure. the rectostgmoid was devascularized wtth bipolar diathemry down to the peritoneal reflection. At a convement place a circumferential mctston was made through the serosa and muscular layers. The submucosal dissection was performed wrth a combinatton of bipolar drathermy and cotton swab dissectton. The submucosal drssection was helped by proximal tractton on the colon. Submucosal dissectron was contmued up to the external sphmcter, then the bowel was intussuscepted through the anus. The aganghomc bowel was excused. A tension-free oblique cola-anal anastomosts was performed using absorbable monofilament sutures.
From the Royal Manchester Chddrenk Hogitul, Hospital Rd. Manchestet: UK. Presented at the 27th Atmual Meetmg of the American Pedutrtc Surgwxd Association. Sun Dtego. California. May 20-23. 1996. Address reprrnt requests to Duncan T Wilcox. FRCS. Great Ormond Street Hospital for Children NHS Trust, Great Ormond St. London WC1 3JH. UK. Copyght o 1997 bx W.B. Saunders Company 0022.34621/97/3202-0019$03.00/O
243
244
WILCOX
Postoperatively antibiotics were continued for 5 days and feeding began once bowel function had returned. At 2 weeks the child was reassessed and, If necessary, anal dilatation was started using Hegar’s dilators. All patients were reviewed regularly as outpatients.
RESULTS
During the lo-year period studied, 51 patients were diagnosed and treated for Hirschsprung’sdiseasewith a primary one-stage pull-through procedure. The mean gestational age was 39.6 weeks (range, 31 to 42 weeks), birth weight was 3.30 kg (range, 1.6 to 4.4 kg), and 76% of the neonateswere boys. Eight children had associated anomalies:six had Down’s syndrome, one had associated limb anomalies,and one had microcephaly, absentcorpus callosum, patent ductus arteriosus,and hydronephrosis. The presentingclinical featureswere abdominaldistension in 88%, vomiting in 50%, and a delay to pass meconium (defined as greater than 24 hours) in 68%. Contrast enemaswere not used to make the diagnosis, and all patients had the diagnosis made by histological examination of a rectal biopsy. The site of Hirschsprung’s diseasewas rectosigmoid in 39, distal colon in 11, and total colonic in one patient. An endorectal pull-through procedure was performed in 13patients and a Swensonprocedure wasperformed in 38. The age at which surgery was performed was 10.3 days (range, 2 to 34), the total time of the procedure (including the time to examine the frozen sections, usually 30 to 45 minutes) was 181 minutes (range, 90 to 270 minutes). Transfusion was necessaryfor 13 neonates (meanvolume was 55 mL, range 10 to 100 mL). A stoma was placed in 22% (11 of 5 l), and in the first 10 patients it was policy to perform a tube cecostomy. Postoperative complications included the following. (1) Two patients had anastomoticleaks, both occurring in children with Down’s syndrome. Thesewere treated with colostomy and dilatation of the resulting anastomotic stricture in one, and a redo pull-through in one. (2) Three patients developed wound infections and one followed with superficial dehiscenceof the wound. (3) Postoperative enterocolitis occurred in four patients, which was successfullytreated with rectal decompressionand antibiotics in three, and ileostomy in one. (4) Two children presented with adhesive obstruction and one required a laparotomy. There were no postoperative deaths,but two patients have died, one secondary to a cardiac anomaly and the secondfrom suddeninfant death syndrome. Total length of hospital stay was 20.4 days (range, 6 to 72). These 5 1 patientshave beenfollowed-up for a meanof 3.7 years (range, 6 monthsto 10 years). In 21 patients the follow-up is for greater than 4 years. Among these patients, four are soiling at least once a week, and the remaining 17 are described by their parents as having normal bowel control.
ET AL
DISCUSSION
The conventional approachto the surgical treatment of Hirschsprung’sdiseaseconsistsof a diverting colostomy after diagnosis in the neonate, and at 1 year of age a definitive pull-through procedure with subsequentclosureof colostomy.’ This approach waspopularized when neonatal pre-, intra- and postoperative care was in its infancy and consequentlyhas provided safe and satisfactory results.2However, with advances in preoperative, anaestheticand postoperative care of the neonate, complex surgery can now safely be performed during the neonatal period. This view has been supported in our seriesbecausecomplications occurred in only 22% of cases,with two anastomoticleaks, four casesof enterocolitis and two casesof adhesiveobstruction. Theseresults compare favorably with other recently reported seriesin which a more conventional treatment path was taken.5 Before embarking on one-stage pull-through procedures for the treatment of Hirschsprung’s diseaseit is essentialto have experienced pathology support. In this procedurethe level of bowel excision is totally dependent on the histopathological study of frozen sections. The time taken for histological examination was between 30 and 45 minutes for each child, dependingon the number of biopsy specimenssent. This in general was between three and five specimens.During this time further preparation and dissection could be started, but frequently there was a short delay in the surgery. Therefore this procedure can only be considered with the back up of rapidly processedand expertly interpreted frozen section biopsies. Surgically good exposure can be obtained in the neonate becauseof the shallow pelvis3 allowing dissection both in the extraserosal dissection performed in a Swensonprocedure and in the submucosalplane usedin endorectal pull-through procedure.In our institution both techniqueswere usedwith similar success.Some authors have reported that it can be necessary to perform the terminal dissectionwith the bowel intussusceptedthrough the anus.4This was not our experience, and in all cases the entire dissection was performed through the pelvis, this is in contrast to definitive pull-through procedures performed in infancy. Early in this seriestube cecostomies were used to protect the coloanal anastomosis,but after the first ten patients this was stopped.Diverting the fecal streamis only considerednow if the viability of the anastomosisis questioned. Anastomotic leakshave occurred in two of 5 1 patients. Anastomotic leaks were treated with a diverting colostomy followed by a further pull-through in one patient and anal dilatation of the subsequentstricture in the second patient. Interestingly, both patients who developed an anastomoticleak had Down’s syndrome, representing a leak rate of two out of six patients. Previous
ONE-STAGE
PULL-THROUGH
245
FOR HD
studies have also shown a poor outcome in Hirschsprung’s disease associated with Down’s syndrome.6 Despite this our policy has still been to offer a primary pull-through procedure to those neonates with both Hirschsprung’s disease and Down’ syndrome. Postoperative enterocolitis occurred in four patients, which is similar to other reported series.4.7 Enterocolitis was successfully treated with antibiotics. no enteral intake, and rectal decompression in three neonates. For one this was insufficient and an ileostomy was necessary. Some centers now recommend routine postoperative rectal washouts to reduce the incidence of postoperative enterocolitis.’ This was not performed in this series. but the overall postoperative enterocolitis incidence remains similar. Long-term bowel continence can be considered in 21 children who were over four years of age. In these children none were totally incontinent and four complained of soiling in excess of once a week. These results are similar if not better than other reported series.7J These results suggest that it is possible to operate within the neonatal pelvis, using both the Swenson and endorectal pull-through technique, without damaging the structures responsible for continence. In addition it is possible to speculate that the passage of feces through the anal canal soon after birth allows these children to develop a more
normal control of continence because of formation and continued use of the required neural circuits. These circuits may not develop sufficiently in those children who are left with a diverting stoma. This mechanism would be similar to that seen in experimental animals in whom an eye patch is applied immediately after birth and who remain blind once the patch is removed because they have not had the required stimulation to produce the necessary neural networks.9 This series. like others, demonstrates that it is possible to perform a primary one-stage pull-through procedure to treat Hirschsprung’s disease during the neonatal and early infant period.-‘,‘“,” Postoperative complications occur but at a similar rate regardless of age at surgery. These results of this series also suggest that long-term bowel continence can be achieved after a neonatal pull-through, with the same or greater frequency to those children in whom the definitive operation was delayed.8 In addition, a one-stage pull-through offers the advantages of one hospitalization in the majority of cases and the subsequent reduction in hospital costs. But the major advance is performing the reconstruction without the need for a stoma. In our experience parents and staff consider this to be a significant improvement in terms of parental bonding, cosmesis, elimination of stoma1 complications, and possibly the improvement of long-term continence.
REFERENCES I. Swenson 0. Bdl AH Jr: Resectton of rectum and rectostgmoid with preservatton of the sphmcter for bemgn spastic lessons producmg megacolons: An experimental study Surgery 24.2 12-220. 1948 2. Swenson 0. Sherman JO, Fisher SH. et al: The treatment and postoperative complicatrons of congemtal megacolon. A 25 year follow up. Ann Surg 182.266-270. 1975 3. Shanbhogue LKR, Btancht A. Experience wtth primary Swenson resection and pullthrough for neonatal Htrschsprung’s disease. Pediatr Surg Int 5:446-448, 1990 4. Cdley RE. Statter MB. Htrschl RB. et al: Detinmve treatment of Hirschsprung’s disease in the newborn wtth a one stage procedure Surgery 115:551-6. 1994 5. Jung PM. Hirschsprung’s disease. One surgeon’s experience in one institution. J Pediatr Surg 30:646-65 I, 1995 6. Qumn FM, Surana R, Puri P: The mfluence of trisomy 21 on
outcome in children with Hirschsprung’s disease. J Pediatr Surg 79 781-783,199-l 7. Marty TL, Seo T. Sulh\an JJ, et al: Rectal irrtgatton for the preventton of postoperative enterocolitts in Hirschsprung’s disease. J Pediatr Surg 30652-653. 1995 8. Help HA, de Vrtea X. Bremer 1. et al: Long-term anorectal functton after Duhamel operation for Hirschsprung’s disease J Pediatr Sumc 30:330-432, 1995 Y Cyander M, Mitchell DE. Prolonged sensmvny to monocular deprivation m dark reared cats. J Neurophyatol-43: 1026-1040. 1980 IO. Carcassone M. Guys JM. Morrison-Lancombe G. et al Management of Hnschsprung’s disease: Curative surgery before 3 months of age. J Pediatr Surg 21.1032-103-l. 1989 I I. So HB. Schwartr DL. Becker JM. et al: Endorectal “pullthrough” without prelimmary colostomy m neonates with Htrschsprung’s disease. J Pedtatr Surg I5 -170-47 I ~ 1980
Discussion WW Neblett: (Nashville, TN): Dr Wilcox. have you comparedyour resultswith the more traditional approach of an initial colostomy and then delayed pull-through procedure?As 1 calculated from the data in the abstract you had a 5% incidence of anastomotic leakage and a 19% incidence of some impairment of continence. I wonder how that would comparewith your earlier results. D.ir: Wilcox (response):The answerto your questionis
no, and the reasonfor that is that this representeda total experience of the attending surgeonswho were performing theseprocedures.It was felt inappropriate to compare this with other surgeonsworking in the sameinstitution who were performing a more standard-stagedapproach. However, I am aware of data from London in which they have compared 20 patients, 20 who had a one-stage procedure and 20 who had a more traditional approach.
246
There was no significant difference in the outcome between those two groups. The incidence of an anastomotic leak of 5% is similar to other reported series using a more traditional approach and an incontinence rate of 19% is certainly within the bounds of other recorded series, some of which have gone up to 50% after Hirschsprung’s disease. TC. Moore (Torrrance, CA): I would like to comment on the last comment here in terms of long-term continence. I think that’s an extremely important part of this observation. In Bianchi’s classic 1990 paper (Pediatric Surgery International 5:446-448) with one consultant doing all the operationsin 25 patients, the median age at presentationwas 3 days, the median ageat operation was 14 days, there were no deaths and no incontinence, and also very importantly there was no mention of intestinal neuronal dysplasia.It’s my feeling that the overwhelming most important part of the care of Hirschsprung’saswith imperforate anus in the newborn is to do the surgical repair in the early hours of life and without a prior colostomy asthe authors have suggestedhere becauseof the very narrow window of opportunity, which is only a few days or hours after birth where the development of normal activity-driven anal-brain neocortical neurocircuitry. This important window was first demonstratedin newborn kittens in whom suture closure of one eyelid at birth rendered them permanently blind in this eye, a finding not encountered in adult cats. This important observation in related studies won the investigators Hubel and Wiesel the Nobel prize in 1981. More recent studieshave shown this phenomenonto apply as well to the somatic sensory neocortex of the brain in postbirth development of neoneurocircuitry aswell asto the visual striate neocortex (Nature 364:623, 1993 and Nature 375:325, 1995). I would like to ask the essayists if they, in this experience, have encounteredneuronal intestinal dysplasia, and what is the level of continence? It would take a good deal of time to determine this. This discussant harbors a suspicion that the elusive intestinal neuronal dysplasiamay well be an artifact generatedby colostomy, and delayed operation induced faulty postbirth anocortical neurocircuitry development. Do you have, in this experience, any intestinal neuronal dysplasia? Thank you. D. T Wilcox (response): Thank you for your comments. The answer to that is no, we have no experience with neuronal dysplasia, although we weren’t specifically looking for it, and we don’t routinely perform biopsieson these patients after their procedure unless there is a problem. J.2. Jona (Evanston, IL): I have somedifficulties with your premisesand the results. First of all, you stated in the beginning that most of thesechildren will need three
WILCOX
ET AL
hospitalizations. Most conventional pull-through operations are done with two, a colostomy and then a primary pull-through unprotected. Secondly, you stated that the purpose is to try and avoid colostomy, yet 20% of your patients endedup with a colostomy anyway. Thirdly, you were trying to minimize hospitalization, yet we seethat your patients on the average spent about 10 days in the hospital after surgery. I stand here only to state that this paper hasonly historical significance. With the introduction of laparoscopic pull-through in the neonates, all of the problems that I pointed to in your presentationare practically eliminated. 0.7: Wilcox (response): The first 10 patients had plannedstomas;only one of the last 43 patientsrequired a stoma. Concerning the laparoscopicapproach, I look forward to hearing your short- and long-term results with much interest. A.G. Coran (Ann Arbol; MI): We’ve been doing the endorectal pull-through in newbornssince 1989 and now have accumulated over 30 casesand have not seenthe sameextended length of stay, the need for blood transfusions, nor the incidence of incontinence. My question is were you able to stratify the resultsbetween the endorectal pull-through and the Swenson?I know the numbers are small and they are skewed toward the Swenson,but did you notice any difference in terms of length of stay, enterocolitis, or leaks? 0.7: Wilcox (response): With regard to the leaks, both leaks occurred in patients who had a Swenson pullthrough rather than the endorectal pullthrough. The length of hospital stay was similar between the two groups. The endorectal pullthrough hasbeenperformed only in the last 4 years and therefore no patient was over 4 years of age. It was difficult to assesscontinence in those patients. J.-M. Laberge (Montreal, Quebec): In Montreal we also like the one-stagepull-through, but not necessarilyin the neonatalperiod. Often we will sendthe patients home on irrigations and perform them when the operating room is available at a few weeks of age. We haven’t had any problems with preoperative enterocolitis and we don’t useantibiotics preoperatively either. One thing that we have seenoccasionally in the very young patients is they seemto have an internal sphincter spasm postoperatively. They are very distended, and becauseof that, we will leave a Malecot or mushroom catheter in place to let somegas passfor the first couple of days. Our operative stay for the pull-through without any stoma is around 4 days in the hospital, so it is also quite short. My third comment is just to publicize a great idea that
ONE-STAGE
PULL-THROUGH
FOR
HD
one of our former chief residents had. If you have a leak from the anastomosis, rather than do a stoma, we converted one patient with a leak into a classical Soave: we just mobilized the IMA and let the bowel hang out from the anus. The child went home, and a few weeks later we performed the resection and the final anastomosis, a nice way to avoid a stricture and re-pull-through later on. J.J. W&man (Encino, CA): One major disadvantage of not performing a colostomy in a newborn is giving up the ability to know that the bowel being pulled through is normal. The transition zone can be of variable length or oblique. By doing a primary pull-through in the newborn you are totally relying on the frozen section to be sure normal bowel is being pulled through. On occasion a colostomy is placed in what was thought to be normal bowel, but because of malfunctioning we subsequently
247
find that the colostomy is not in normal bowel. I believe there is an advantage in knowing that the infant with a colostomy is thriving, thus indicating that the colostomy site of the future pull-through is in perfectly normal bowel. This paper assumes that everybody else does threestage operations. Those of us who perform a colostomy on the newborn usually do two-stage operations. Thank you. D.T Wilcox (response): Well I agree it is very important that the pathologists have good experience performing frozen section. In the literature are a number of reported cases of missed segments or neuronal intestinal dysplasia leading to constipation even in those performed by a more classic procedure. Both ways are not without trouble with regard to leaving small sections of aganglionic bowel behind or pulling the transition zone through.
Pull-Through T. Wilcox,
James Bruce, John Bowen, Manchester, England
0 Introduction: Traditionally Hirschsprung’s disease has been treated by a three-stage procedure. This approach has been associated with stoma problems, three episodes of hospitalization, and a theoretical disadvantage, because of the delayed passage of feces via the rectum, in controlling defecation. The aim of this study was to assess the results of one-stage neonatal pull-through for the treatment of Hirschsprung’s disease. Methods: This was a IO-year retrospective review of all patients treated with a neonatal pullthrough for Hirschsprung’s disease. Information was obtained from the patients’ records. All descriptive data are expressed as mean f standard deviation. Results: Fifty-one patients were treated in the IO-year period studied. The mean gestational age was 39.6 f 1.7 weeks, weight was 3.3 + 0.54 kg, with 76% being boys. The surgery was performed at 10.3 f 5.8 days of age. A Swenson procedure was carried out in 38 patients and a Soave in 13. The extent of the disease was rectosigmoid in 39, and the remaining were sigmoid, descending colon, and total colonic. The mean length of the surgery was 181 f 40 minutes. Blood transfusions were required in 13 patients with a mean volume of 55 mL given. The complications observed were: an anastomotic leak in two cases, requiring laparotomy and colostomy (both in patients with Down’s syndrome), and wound infection in three cases. There were two delayed deaths both associated with other congenital anomalies. In total, three patients required further surgery; two had colostomies, one requiring a redo pull-through and one had an ileostomy for enterocolitis. The total length of hospital stay was 20.4 + 11.2 days. Mean follow-up was 3.73 f 2.9 years. In 21 patients the follow-up was greater than 4 years and in this group none were totally incontinent. Four patients complained of soiling at least once a week, and 17 had normal bowel control. Conclusion: Hirschsprung’s disease can be successfully treated in the neonatal period with a one-stage pull-through. The shortand long-term results are as good as those with the three-stage procedure, with the child usually benefitting by not having a stoma and a shorter hospital stay. Copyright Q 1997 by W. B. Saunders Company INDEX Soave.
WORDS:
Hirschsprung’s
disease,
neonate,
Swenson,
S
INCE THE RECONSTRUCTIVE operation to treat Hirschsprung’s disease was first described by Swenson and Bill in 1948, neonates have been treated soon after diagnosis with a defunctioning colostomy.’ When patients were older a definitive pull-through procedure was performed followed by colostomy closure. This three-stage procedure has been advocated because early reports suggestedit was unsafe to perform a definitive operation during the neonatal period.* With advancesin neonatal, anaesthetic,and surgical care, a safe one-stage definitive procedure has become possible.3.JHere we report a IO-year experience of Hirschsprung’s disease
Journal
ofPediatric
Surgery,
Vol 32, No 2 (February),
to Treat Hirschsprung’s
1997: pp 243-247
and Adrian
Disease
Bianchi
treated, during the neonatal period. with a one-stage pull-through. MATERIALS
AND METHODS
This is a retrospective review of patients’ notes. All neonates diagnosed with Hrrschsprung’s disease between 1984 and 1994 treated by two consultants were included. Informatton was obtained from the patients’ notes, and all children had been reviewed wtthin the last 12 months. Diagnosis was achieved in all patients by histological examination of a rectal biopsy. Patrents m whom the diagnosis was suspected or confirmed received preoperative antibiotics (usually Augmentin) and 4 to 6 hourly rectal decompresston (by rectal stimulation). At a convenient time, and when the child’s general health had returned to normal, a primary pull-through procedure was performed. Two pull-through procedures were used in this series, the Swenson’ and the Soave (endorectal pull-through).j Early in the series or when there was concern over the anastomosts, a covering colostomy or tube cecostomy was performed. For the pull-through, patients were placed m a supine lithotomy position, with the perineum fully exposed at the end of the operating table. The abdomen was entered through a transverse lower abdommal mciston and a neonatal Denis Browne retractor applied. Muscle biopsy spectmens were taken from the anttmesentertc surface of the colon, wtthout entering the lumen, startmg in the rectum and workmg proximally. The bropsy results were assessed mtraoperatively by frozen section until normal ganghonated bowel was Identified. In the Swenson procedure the rectosigmoid was devascularized with bipolar diathermy keepmg close to the bowel wall. Dtssection was down to the level of the external sphincter posteriorly and laterally; anteriorly dissectton stopped approximately I .5 cm more proxtmal. Sufftcrent length was obtained by dtvrding the infenor mesenteric pedicle, being careful to preserve the marginal vessels The level to be resected was marked wrth silk sutures and the bowel was mtussuscepted through the anus wtth a Babcock forceps. In the endorectal pull-through procedure. the rectostgmoid was devascularized wtth bipolar diathemry down to the peritoneal reflection. At a convement place a circumferential mctston was made through the serosa and muscular layers. The submucosal dissection was performed wrth a combinatton of bipolar drathermy and cotton swab dissectton. The submucosal drssection was helped by proximal tractton on the colon. Submucosal dissectron was contmued up to the external sphmcter, then the bowel was intussuscepted through the anus. The aganghomc bowel was excused. A tension-free oblique cola-anal anastomosts was performed using absorbable monofilament sutures.
From the Royal Manchester Chddrenk Hogitul, Hospital Rd. Manchestet: UK. Presented at the 27th Atmual Meetmg of the American Pedutrtc Surgwxd Association. Sun Dtego. California. May 20-23. 1996. Address reprrnt requests to Duncan T Wilcox. FRCS. Great Ormond Street Hospital for Children NHS Trust, Great Ormond St. London WC1 3JH. UK. Copyght o 1997 bx W.B. Saunders Company 0022.34621/97/3202-0019$03.00/O
243
244
WILCOX
Postoperatively antibiotics were continued for 5 days and feeding began once bowel function had returned. At 2 weeks the child was reassessed and, If necessary, anal dilatation was started using Hegar’s dilators. All patients were reviewed regularly as outpatients.
RESULTS
During the lo-year period studied, 51 patients were diagnosed and treated for Hirschsprung’sdiseasewith a primary one-stage pull-through procedure. The mean gestational age was 39.6 weeks (range, 31 to 42 weeks), birth weight was 3.30 kg (range, 1.6 to 4.4 kg), and 76% of the neonateswere boys. Eight children had associated anomalies:six had Down’s syndrome, one had associated limb anomalies,and one had microcephaly, absentcorpus callosum, patent ductus arteriosus,and hydronephrosis. The presentingclinical featureswere abdominaldistension in 88%, vomiting in 50%, and a delay to pass meconium (defined as greater than 24 hours) in 68%. Contrast enemaswere not used to make the diagnosis, and all patients had the diagnosis made by histological examination of a rectal biopsy. The site of Hirschsprung’s diseasewas rectosigmoid in 39, distal colon in 11, and total colonic in one patient. An endorectal pull-through procedure was performed in 13patients and a Swensonprocedure wasperformed in 38. The age at which surgery was performed was 10.3 days (range, 2 to 34), the total time of the procedure (including the time to examine the frozen sections, usually 30 to 45 minutes) was 181 minutes (range, 90 to 270 minutes). Transfusion was necessaryfor 13 neonates (meanvolume was 55 mL, range 10 to 100 mL). A stoma was placed in 22% (11 of 5 l), and in the first 10 patients it was policy to perform a tube cecostomy. Postoperative complications included the following. (1) Two patients had anastomoticleaks, both occurring in children with Down’s syndrome. Thesewere treated with colostomy and dilatation of the resulting anastomotic stricture in one, and a redo pull-through in one. (2) Three patients developed wound infections and one followed with superficial dehiscenceof the wound. (3) Postoperative enterocolitis occurred in four patients, which was successfullytreated with rectal decompressionand antibiotics in three, and ileostomy in one. (4) Two children presented with adhesive obstruction and one required a laparotomy. There were no postoperative deaths,but two patients have died, one secondary to a cardiac anomaly and the secondfrom suddeninfant death syndrome. Total length of hospital stay was 20.4 days (range, 6 to 72). These 5 1 patientshave beenfollowed-up for a meanof 3.7 years (range, 6 monthsto 10 years). In 21 patients the follow-up is for greater than 4 years. Among these patients, four are soiling at least once a week, and the remaining 17 are described by their parents as having normal bowel control.
ET AL
DISCUSSION
The conventional approachto the surgical treatment of Hirschsprung’sdiseaseconsistsof a diverting colostomy after diagnosis in the neonate, and at 1 year of age a definitive pull-through procedure with subsequentclosureof colostomy.’ This approach waspopularized when neonatal pre-, intra- and postoperative care was in its infancy and consequentlyhas provided safe and satisfactory results.2However, with advances in preoperative, anaestheticand postoperative care of the neonate, complex surgery can now safely be performed during the neonatal period. This view has been supported in our seriesbecausecomplications occurred in only 22% of cases,with two anastomoticleaks, four casesof enterocolitis and two casesof adhesiveobstruction. Theseresults compare favorably with other recently reported seriesin which a more conventional treatment path was taken.5 Before embarking on one-stage pull-through procedures for the treatment of Hirschsprung’s diseaseit is essentialto have experienced pathology support. In this procedurethe level of bowel excision is totally dependent on the histopathological study of frozen sections. The time taken for histological examination was between 30 and 45 minutes for each child, dependingon the number of biopsy specimenssent. This in general was between three and five specimens.During this time further preparation and dissection could be started, but frequently there was a short delay in the surgery. Therefore this procedure can only be considered with the back up of rapidly processedand expertly interpreted frozen section biopsies. Surgically good exposure can be obtained in the neonate becauseof the shallow pelvis3 allowing dissection both in the extraserosal dissection performed in a Swensonprocedure and in the submucosalplane usedin endorectal pull-through procedure.In our institution both techniqueswere usedwith similar success.Some authors have reported that it can be necessary to perform the terminal dissectionwith the bowel intussusceptedthrough the anus.4This was not our experience, and in all cases the entire dissection was performed through the pelvis, this is in contrast to definitive pull-through procedures performed in infancy. Early in this seriestube cecostomies were used to protect the coloanal anastomosis,but after the first ten patients this was stopped.Diverting the fecal streamis only considerednow if the viability of the anastomosisis questioned. Anastomotic leakshave occurred in two of 5 1 patients. Anastomotic leaks were treated with a diverting colostomy followed by a further pull-through in one patient and anal dilatation of the subsequentstricture in the second patient. Interestingly, both patients who developed an anastomoticleak had Down’s syndrome, representing a leak rate of two out of six patients. Previous
ONE-STAGE
PULL-THROUGH
245
FOR HD
studies have also shown a poor outcome in Hirschsprung’s disease associated with Down’s syndrome.6 Despite this our policy has still been to offer a primary pull-through procedure to those neonates with both Hirschsprung’s disease and Down’ syndrome. Postoperative enterocolitis occurred in four patients, which is similar to other reported series.4.7 Enterocolitis was successfully treated with antibiotics. no enteral intake, and rectal decompression in three neonates. For one this was insufficient and an ileostomy was necessary. Some centers now recommend routine postoperative rectal washouts to reduce the incidence of postoperative enterocolitis.’ This was not performed in this series. but the overall postoperative enterocolitis incidence remains similar. Long-term bowel continence can be considered in 21 children who were over four years of age. In these children none were totally incontinent and four complained of soiling in excess of once a week. These results are similar if not better than other reported series.7J These results suggest that it is possible to operate within the neonatal pelvis, using both the Swenson and endorectal pull-through technique, without damaging the structures responsible for continence. In addition it is possible to speculate that the passage of feces through the anal canal soon after birth allows these children to develop a more
normal control of continence because of formation and continued use of the required neural circuits. These circuits may not develop sufficiently in those children who are left with a diverting stoma. This mechanism would be similar to that seen in experimental animals in whom an eye patch is applied immediately after birth and who remain blind once the patch is removed because they have not had the required stimulation to produce the necessary neural networks.9 This series. like others, demonstrates that it is possible to perform a primary one-stage pull-through procedure to treat Hirschsprung’s disease during the neonatal and early infant period.-‘,‘“,” Postoperative complications occur but at a similar rate regardless of age at surgery. These results of this series also suggest that long-term bowel continence can be achieved after a neonatal pull-through, with the same or greater frequency to those children in whom the definitive operation was delayed.8 In addition, a one-stage pull-through offers the advantages of one hospitalization in the majority of cases and the subsequent reduction in hospital costs. But the major advance is performing the reconstruction without the need for a stoma. In our experience parents and staff consider this to be a significant improvement in terms of parental bonding, cosmesis, elimination of stoma1 complications, and possibly the improvement of long-term continence.
REFERENCES I. Swenson 0. Bdl AH Jr: Resectton of rectum and rectostgmoid with preservatton of the sphmcter for bemgn spastic lessons producmg megacolons: An experimental study Surgery 24.2 12-220. 1948 2. Swenson 0. Sherman JO, Fisher SH. et al: The treatment and postoperative complicatrons of congemtal megacolon. A 25 year follow up. Ann Surg 182.266-270. 1975 3. Shanbhogue LKR, Btancht A. Experience wtth primary Swenson resection and pullthrough for neonatal Htrschsprung’s disease. Pediatr Surg Int 5:446-448, 1990 4. Cdley RE. Statter MB. Htrschl RB. et al: Detinmve treatment of Hirschsprung’s disease in the newborn wtth a one stage procedure Surgery 115:551-6. 1994 5. Jung PM. Hirschsprung’s disease. One surgeon’s experience in one institution. J Pediatr Surg 30:646-65 I, 1995 6. Qumn FM, Surana R, Puri P: The mfluence of trisomy 21 on
outcome in children with Hirschsprung’s disease. J Pediatr Surg 79 781-783,199-l 7. Marty TL, Seo T. Sulh\an JJ, et al: Rectal irrtgatton for the preventton of postoperative enterocolitts in Hirschsprung’s disease. J Pediatr Surg 30652-653. 1995 8. Help HA, de Vrtea X. Bremer 1. et al: Long-term anorectal functton after Duhamel operation for Hirschsprung’s disease J Pediatr Sumc 30:330-432, 1995 Y Cyander M, Mitchell DE. Prolonged sensmvny to monocular deprivation m dark reared cats. J Neurophyatol-43: 1026-1040. 1980 IO. Carcassone M. Guys JM. Morrison-Lancombe G. et al Management of Hnschsprung’s disease: Curative surgery before 3 months of age. J Pediatr Surg 21.1032-103-l. 1989 I I. So HB. Schwartr DL. Becker JM. et al: Endorectal “pullthrough” without prelimmary colostomy m neonates with Htrschsprung’s disease. J Pedtatr Surg I5 -170-47 I ~ 1980
Discussion WW Neblett: (Nashville, TN): Dr Wilcox. have you comparedyour resultswith the more traditional approach of an initial colostomy and then delayed pull-through procedure?As 1 calculated from the data in the abstract you had a 5% incidence of anastomotic leakage and a 19% incidence of some impairment of continence. I wonder how that would comparewith your earlier results. D.ir: Wilcox (response):The answerto your questionis
no, and the reasonfor that is that this representeda total experience of the attending surgeonswho were performing theseprocedures.It was felt inappropriate to compare this with other surgeonsworking in the sameinstitution who were performing a more standard-stagedapproach. However, I am aware of data from London in which they have compared 20 patients, 20 who had a one-stage procedure and 20 who had a more traditional approach.
246
There was no significant difference in the outcome between those two groups. The incidence of an anastomotic leak of 5% is similar to other reported series using a more traditional approach and an incontinence rate of 19% is certainly within the bounds of other recorded series, some of which have gone up to 50% after Hirschsprung’s disease. TC. Moore (Torrrance, CA): I would like to comment on the last comment here in terms of long-term continence. I think that’s an extremely important part of this observation. In Bianchi’s classic 1990 paper (Pediatric Surgery International 5:446-448) with one consultant doing all the operationsin 25 patients, the median age at presentationwas 3 days, the median ageat operation was 14 days, there were no deaths and no incontinence, and also very importantly there was no mention of intestinal neuronal dysplasia.It’s my feeling that the overwhelming most important part of the care of Hirschsprung’saswith imperforate anus in the newborn is to do the surgical repair in the early hours of life and without a prior colostomy asthe authors have suggestedhere becauseof the very narrow window of opportunity, which is only a few days or hours after birth where the development of normal activity-driven anal-brain neocortical neurocircuitry. This important window was first demonstratedin newborn kittens in whom suture closure of one eyelid at birth rendered them permanently blind in this eye, a finding not encountered in adult cats. This important observation in related studies won the investigators Hubel and Wiesel the Nobel prize in 1981. More recent studieshave shown this phenomenonto apply as well to the somatic sensory neocortex of the brain in postbirth development of neoneurocircuitry aswell asto the visual striate neocortex (Nature 364:623, 1993 and Nature 375:325, 1995). I would like to ask the essayists if they, in this experience, have encounteredneuronal intestinal dysplasia, and what is the level of continence? It would take a good deal of time to determine this. This discussant harbors a suspicion that the elusive intestinal neuronal dysplasiamay well be an artifact generatedby colostomy, and delayed operation induced faulty postbirth anocortical neurocircuitry development. Do you have, in this experience, any intestinal neuronal dysplasia? Thank you. D. T Wilcox (response): Thank you for your comments. The answer to that is no, we have no experience with neuronal dysplasia, although we weren’t specifically looking for it, and we don’t routinely perform biopsieson these patients after their procedure unless there is a problem. J.2. Jona (Evanston, IL): I have somedifficulties with your premisesand the results. First of all, you stated in the beginning that most of thesechildren will need three
WILCOX
ET AL
hospitalizations. Most conventional pull-through operations are done with two, a colostomy and then a primary pull-through unprotected. Secondly, you stated that the purpose is to try and avoid colostomy, yet 20% of your patients endedup with a colostomy anyway. Thirdly, you were trying to minimize hospitalization, yet we seethat your patients on the average spent about 10 days in the hospital after surgery. I stand here only to state that this paper hasonly historical significance. With the introduction of laparoscopic pull-through in the neonates, all of the problems that I pointed to in your presentationare practically eliminated. 0.7: Wilcox (response): The first 10 patients had plannedstomas;only one of the last 43 patientsrequired a stoma. Concerning the laparoscopicapproach, I look forward to hearing your short- and long-term results with much interest. A.G. Coran (Ann Arbol; MI): We’ve been doing the endorectal pull-through in newbornssince 1989 and now have accumulated over 30 casesand have not seenthe sameextended length of stay, the need for blood transfusions, nor the incidence of incontinence. My question is were you able to stratify the resultsbetween the endorectal pull-through and the Swenson?I know the numbers are small and they are skewed toward the Swenson,but did you notice any difference in terms of length of stay, enterocolitis, or leaks? 0.7: Wilcox (response): With regard to the leaks, both leaks occurred in patients who had a Swenson pullthrough rather than the endorectal pullthrough. The length of hospital stay was similar between the two groups. The endorectal pullthrough hasbeenperformed only in the last 4 years and therefore no patient was over 4 years of age. It was difficult to assesscontinence in those patients. J.-M. Laberge (Montreal, Quebec): In Montreal we also like the one-stagepull-through, but not necessarilyin the neonatalperiod. Often we will sendthe patients home on irrigations and perform them when the operating room is available at a few weeks of age. We haven’t had any problems with preoperative enterocolitis and we don’t useantibiotics preoperatively either. One thing that we have seenoccasionally in the very young patients is they seemto have an internal sphincter spasm postoperatively. They are very distended, and becauseof that, we will leave a Malecot or mushroom catheter in place to let somegas passfor the first couple of days. Our operative stay for the pull-through without any stoma is around 4 days in the hospital, so it is also quite short. My third comment is just to publicize a great idea that
ONE-STAGE
PULL-THROUGH
FOR
HD
one of our former chief residents had. If you have a leak from the anastomosis, rather than do a stoma, we converted one patient with a leak into a classical Soave: we just mobilized the IMA and let the bowel hang out from the anus. The child went home, and a few weeks later we performed the resection and the final anastomosis, a nice way to avoid a stricture and re-pull-through later on. J.J. W&man (Encino, CA): One major disadvantage of not performing a colostomy in a newborn is giving up the ability to know that the bowel being pulled through is normal. The transition zone can be of variable length or oblique. By doing a primary pull-through in the newborn you are totally relying on the frozen section to be sure normal bowel is being pulled through. On occasion a colostomy is placed in what was thought to be normal bowel, but because of malfunctioning we subsequently
247
find that the colostomy is not in normal bowel. I believe there is an advantage in knowing that the infant with a colostomy is thriving, thus indicating that the colostomy site of the future pull-through is in perfectly normal bowel. This paper assumes that everybody else does threestage operations. Those of us who perform a colostomy on the newborn usually do two-stage operations. Thank you. D.T Wilcox (response): Well I agree it is very important that the pathologists have good experience performing frozen section. In the literature are a number of reported cases of missed segments or neuronal intestinal dysplasia leading to constipation even in those performed by a more classic procedure. Both ways are not without trouble with regard to leaving small sections of aganglionic bowel behind or pulling the transition zone through.