Hirschsprung disease (HD), a neural crest disorder (1) characterized by aganglionosis beginning in the rectum and extending to various lengths proximally, occurs in approximately 1 in every 5000 to 10,000 live births (2,3). Surgical management has been evolving, particularly with the introduction of laparoscopic and transanal techniques (4). Irrespective of the surgical technique used, up to 60% of children with HD experience a protracted defecation disorder (DD) and/or enterocolitis after initial HD corrective surgery (CS) (5–7). These DDs may include fecal incontinence (5), constipation (6), at times associated with episodes of abdominal distention and emesis (8), and/or enterocolitis (9). Postoperative enterocolitis may lead to increased morbidity and hospitalizations as well as increased mortality (9). Full bowel continence and resolution of symptoms are often achieved by late adolescence, but not in all cases (10), and there is a significant negative effect on the child and the child's family when dealing with these postoperative DDs (11,12).
Given the effect and prevalence of DDs and/or enterocolitis after HD CS in children, various diagnostic and management strategies have been suggested to evaluate specific (eg, fecal incontinence alone) postoperative DDs in this population (13–15). These strategies include various radiologic, manometric, dietary, pharmaceutical, and surgical interventions. Evaluation of a comprehensive systematic algorithm combining strategies and determining the ability to affect long-term clinical outcome in all children with all postoperative DDs has not been completed, however.
The purpose of the present study was to determine the diagnoses, management, and long-term clinical outcomes following a well-defined systemic algorithm incorporating a variety of diagnostic and management strategies to address DDs in children with HD after CS presenting to a tertiary care pediatric gastroenterology clinic.
Following institutional review board approval, medical records of all children status post-CS for HD presenting to a tertiary care pediatric gastroenterology clinic at a tertiary medical center between 1998 and 2006 were reviewed. Children with a physician's diagnosis of constipation, fecal incontinence, and/or enterocolitis and documented follow-up after initial evaluation were included.
A diagnostic algorithm (Fig. 1) that first excludes potential anatomic etiologies via physical examination and/or barium enema was systematically followed in all children as part of the routine clinical practice. Although a barium enema was preferred, at times given the child's anxiety, an examination under anesthesia was undertaken. All of the children were tested for the presence of Clostridium difficile toxin in the stool and treated appropriately. The senior author (S.N.) was involved in the evaluation and management of all of the patients. Children with anal stenosis, strictures, or a transition zone were referred for further surgical evaluation.
Following initial evaluation, children were classified into 1 of 2 groups: obstructive symptoms (OS) (eg, constipation, straining with defecation, abdominal distention) and/or enterocolitis, or nonretentive fecal incontinence based on radiologic results and clinical impression. Children classified as having OS and/or enterocolitis were started on an aggressive bowel regimen. Children with nonretentive fecal incontinence received full medical evaluations and were started on oral agents (Imodium, fiber supplementation) to slow colonic transit. Children with abdominal distention and/or recurrent enterocolitis were started on a 5-day course of antibiotics (metronidazole or amoxicillin with clavulanic acid) monthly for the duration of the symptom.
If initial management was unsuccessful per physician assessment, then patients underwent further anorectal function evaluation (Fig. 2), which included anorectal manometry, flexible sigmoidoscopy, and rectal suction biopsy. Clostridium botulinum toxin was injected after completion of a flexible sigmoidoscopy in children with anal sphincter pressure above 50 mmHg, who continued with OS and/or enterocolitis (16). Anorectal manometry was completed before all C botulinum toxin injections. Surgical referral was completed if residual aganglionosis or an anatomic etiology (eg, stricture) was identified. Children with intractable or recurring symptoms despite aggressive medical therapy were referred for surgical consultation after discussion between the primary pediatric gastroenterologist and the family.
Anorectal manometry was performed as previously described using a continuously perfused catheter using a low-compliance pneumo-hydraulic system (Model ARM2; Arndorfer Medical Specialties, Greendale, WI) (16). Colonic manometry was performed in children with continued symptoms despite completing an anorectal evaluation with directed therapy (Fig. 2) (16). Injection of Botox (C botulinum toxin A, Allergan, Irvine, CA) with a 1-mL syringe and a 25- to 30-gauge needle, 6 U/kg (up to 100 U total), was divided and injected equally into 4 quadrants at the level of the dentate line, as previously described (8).
Management was based on the diagnostic findings and clinical course (Figs. 1 and 2). Laxative bowel regimens included oral combinations of senna, polyethylene glycol 3350, magnesium hydroxide, and bisacodyl. In patients with fewer than 2 bowel movements per week and/or abdominal distention causing discomfort despite the use of an aggressive oral laxative regimen, rectal therapies were recommended. These included either normal saline irrigations or bisacodyl suppositories or sodium phosphate enemas.
Final Diagnoses Responsible for the Symptoms
Diagnoses responsible for the symptoms were determined based on the result of diagnostic testing and management response. Children with OS that improved significantly (eg, no longer having constipation or fecal incontinence episodes related to stool retention) following an initial bowel regimen were classified as having nonintractable constipation. Children who improved following anal sphincter C botulinum injection were classified as having nonrelaxing anal sphincter as the reason for the symptoms, if another cause (eg, colonic dysmotility, persistent aganglionosis) was not subsequently identified. Children with abnormalities on colonic manometry were classified as having colonic dysmotility. Children undergoing a surgical procedure received a final diagnosis based on the indication for the procedure (eg, residual aganglionosis). Children with nonretentive fecal incontinence received a final diagnosis reflecting the etiology of the incontinence (eg, food intolerance).
Clinical outcome was determined based on symptom severity and management needed at the time of last follow-up as compared to baseline. The 4 categories were defined as follows:
1. Poor: No improvement or worsening in the number of bowel movements and/or fecal accidents, or an enterocolitis episode within 3 months of the last documented follow-up.
2. Fair: Improvement in the number of bowel movements with need for rectal therapies (eg, enema, suppository, irrigation), or fecal incontinence episodes more than once per week, or an enterocolitis episode between 3 to 6 months of the last documented follow-up.
3. Good: Improvement in the number of bowel movements with continued usage of oral laxatives, or fecal incontinence episodes less than once per week, or an enterocolitis episode between 6 to 12 months of the last documented follow-up.
4. Excellent: Improvement in the number of bowel movements or other OS without need for medications, or elimination of all fecal incontinence episodes, or last enterocolitis episode more than 1 year from the last documented follow-up.
Results are presented as mean ± standard error. Percentages are rounded to the nearest decimal point. χ2 analysis was used to analyze differences in categorical variables between groups. Excellent and good long-term outcomes were classified as being favorable. Binomial logistic forward stepwise regression analysis with favorable outcome as the dependent factor incorporating sex, age at presentation, symptom type (eg, enterocolitis), Down syndrome, extent of original aganglionosis, type of original pull-through (eg, Duhamel), and final diagnosis was performed. SPSS was the statistical software used (SPSS Inc, Chicago, IL).
Fifty-seven children were identified. The mean age was 5.1 ± 0.6 years. Forty-six (81%) were boys and 11 (19%) girls. Seven (12%) had Down syndrome, 5 (9%) had developmental delay of unknown etiology, and 1 (2%) had Bardet-Biedl syndrome. Of the 57 children, 30 (53%) had OS alone (eg, constipation, straining with defecation, abdominal distention), 14 (25%) had both OS and postoperative enterocolitis episodes, 7 (12%) had postoperative enterocolitis alone, and 6 (11%) had nonretentive (nonoverflow) fecal incontinence.
The original extent of aganglionosis was short segment (rectosigmoid) in 40 (70%), long segment (through to ascending colon) in 8 (14%), and total colonic with or without small bowel involvement in 9 (16%). Surgical pull-through procedures included Soave in 36 (63%), Duhamel in 11 (19.4%), Swenson in 7 (12.3%), and unknown (documentation from outside hospital unavailable) in 3 (5.3%).
Diagnostic studies completed within this population are found in Table 1. The findings for children undergoing a barium enema were notable for 5 identifying a stricture, 6 identifying a transition zone, and 10 identifying colonic dilatation. All of the children undergoing an anorectal manometry (n = 45) did not have a recto-anal inhibitory reflex. The mean sphincter pressure in those undergoing an anorectal manometry was 90.2 ± 3.3 mmHg (range 40–140). Six (10.5%) children underwent colonic manometry evaluation. Of these, 2 were normal; 3 had high-amplitude propagating contractions only in the proximal colon, and 1 demonstrated complete colonic inertia.
Ten (17.5%) children with OS responded well to an aggressive laxative regimen and were classified as having nonintractable constipation. Eight (14%) children were found to have a mechanical obstruction (eg, stricture), and 7 (12%) were found to have residual aganglionosis. Twenty-two (38.6%) children were classified as having a nonrelaxing internal anal sphincter (IAS) as the primary reason for their OS. Four (6.8%) were identified as having colonic dysmotility.
Of the 6 children with nonretentive fecal incontinence, 2 had food intolerances that were identified; 2 had bacterial overgrowth as a contributing factor as identified via lactulose breath testing and response to antibiotic therapy, and the remaining 2 had abnormally rapid colonic transit.
Thirty-seven children (69.1%) received anal sphincter C botulinum (Botox) injections. The mean number of injections in those undergoing at least 1 Botox injection was 2.8 ± 0.3. Of the 37 children undergoing Botox injections, 33 (89.2%) demonstrated an initial short-term improvement.
Twenty-two (38.6%) children with HD after CS underwent at least 1 further surgical procedure, with 10 (17.5%) undergoing multiple procedures. A listing of surgical procedures performed can be found in Table 2. All of the children with residual aganglionosis underwent repeat pull-through. A repeat pull-through was also performed in other select cases (eg, failed previous stricturoplasty). The original type of surgical procedure did not predict the need for a repeat surgical procedure (P = 0.47). Four (7.0%) failed all therapy and required diverting ileostomies or colostomies.
Long-term Clinical Outcomes
The mean follow-up period in this cohort was 41.4 ± 4.5 months. Thirty-eight (66.7%) children had a favorable (excellent or good) long-term outcome. Children who underwent a subsequent surgical procedure (14/22) were as likely to have a favorable outcome as those following medical therapy (24/35) alone (P = 0.78). Children with any form of developmental delay including Down syndrome or Bardet-Biedl syndrome (6/13) were as likely as those without developmental delay (32/44) to have a favorable outcome (P = 0.09) (Fig. 3).
Multivariate logistic regression analysis determined that children with enterocolitis as a presenting symptom were more likely to have a favorable outcome (18/21) than those without enterocolitis (20/36) (P < 0.05). Other factors, including sex, age at presentation, Down syndrome, any form of developmental delay, extent of original aganglionosis, type of original pull-through (eg, Duhamel), and final diagnosis, were not found to predict long-term clinical outcome.
Two of the 6 (33%) children with nonretentive fecal incontinence had a favorable outcome at the time of the last follow-up. The 1 child with nonretentive fecal incontinence with identified low anorectal sphincter pressure had a poor long-term outcome.
To our knowledge, this is the first study to evaluate long-term clinical outcomes as well as diagnoses and management strategies when following a comprehensive systematic algorithm addressing all protracted DDs in children with HD after CS. A variety of diagnoses were made, including identification of residual aganglionosis and other anatomic problems that required surgical correction. Management using the algorithm was dependent on the underlying etiology of the symptoms, and approximately two-thirds of the cohort had had a good or excellent long-term clinical outcome.
Previous authors have reported that the need for reoperation following initial pull-through CS for HD may be relatively high, with estimates ranging between 26% and 34% (17,18). The rate of need for reoperation in our population is slightly higher (38%) than we had initially expected, given that all of the children evaluated in our gastroenterology clinic were past the initial perioperative period. More than 25% of children required further surgical management for either a mechanical obstruction (eg, stricture) or residual aganglionosis. This may have been partly the result of the nature of the referral population to a tertiary care center because these patients were likely to have failed previous more conservative measures and/or their symptoms may have been more pronounced. Nonetheless, communication with a pediatric surgeon is clearly needed in the routine care of these children.
The cause for a relatively high number of children presenting with residual aganglionosis is unknown; it may be the result of surgical technique (eg, transition zone pull-through), chronic unidentified enterocolitis, or loss of ganglion cells after pull-through (19,20). Efforts directed toward minimizing postoperative aganglionosis are already being investigated through modification of surgical techniques (10). In the future, progenitor cell transplantation to repopulate areas of aganglionosis after CS with functional ganglion cells may be possible, and any such advances would change the algorithm accordingly (21).
Patients with postoperative HD may be at greater risk for DDs given a postsurgical neorectum with likely decreased sensation and accommodative capacity, continued nonrelaxation of the IAS (16), and higher propensity for colonic dysmotility (22,23). Despite these apparent disadvantages, certain children with OS and/or enterocolitis responded well to an aggressive oral laxative regimen alone. This supports the view that children with postoperative HD may develop a pattern of defecation avoidance because of trauma or previous discomfort similar to children with functional constipation (6). We support an initial approach of using oral laxative regimens with common behavioral approaches (eg, toilet sitting) in this population of children in an attempt to avoid further unnecessary interventions.
In theory, all children with HD have a nonrelaxing IAS, which may cause a functional obstruction. Attribution of symptoms to this functional obstruction following improvement with local Botox injection therapy has been supported by us and others (16,20,24). Given Botox anal sphincter injection is a local therapy, and children who responded had not previously done so with more conservative measures, we believe that attribution of symptoms to a nonrelaxing IAS is appropriate. We recently reported our long-term experience of using Botox in children with nonrelaxing IAS (16) and believe that it may be a useful therapy within an overall therapeutic algorithm. Other centers may have more experience with anal sphincter myectomy (25), although we prefer to avoid myectomy when possible given higher long-term fecal incontinence rates, and the natural tendency of the anal sphincter to become hypotensive in adults over time with HD after CS (26). Future incorporation of topical therapies (eg, glycerin trinitrate) to relax the anal sphincter may prove to be beneficial (27).
Nonretentive fecal incontinence in our population was less common. In the 6 patients within this cohort, the differential diagnosis was broadened and encompassed small bowel bacterial overgrowth, food intolerance, and rapid colonic transit. These diagnoses suggest that children with HD with nonretentive fecal incontinence may benefit from full medical evaluations beyond the use of agents to slow intestinal transit alone.
The only baseline predictive factor for a favorable outcome using this algorithm was enterocolitis as a presenting symptom, suggesting children with this presentation are most likely to benefit from the management within the algorithm. The vast majority of children with enterocolitis had a favorable outcome in the present study. This may in part be due to the ability of many of the interventions to improve colonic transit and prevent obstruction. Moreover, we postulate that enterocolitis is less likely to have a functional/behavioral component and as such may be more directly amenable to the rendered therapies.
The primary weakness of the present study is that it is retrospective. As such, factors such as follow-up intervals and compliance with medication regimens prescribed were not accounted for. Symptom improvement was subjective and based on history as recorded by the primary gastroenterologist. However, some of the nonstandardized factors may have been ameliorated by the fact that 1 of the investigators (S.N.) was involved in the care of all of the patients, that all of the patients with postoperative HD with DDs at our institution are referred to the pediatric gastroenterology program, and that a systematic algorithm was followed in all of the patients. Another limitation of the present study is that it incorporates numerous diagnostic techniques and interventions. As such, in theory, an effective intervention may be buried within other less effective or even harmful interventions. It is for this reason, however, that we chose to evaluate the systematic algorithm as a whole. Given that the algorithm was followed in the entire studied cohort in a manner consistent with daily clinical care, we hope that this model will serve as an early step toward further work in comparative effectiveness in the care of these children. In the future, changes in the algorithm with new clinical and/or technical advances may be made and outcomes compared accordingly.
We propose that the stepwise algorithm is generalizable and can be followed in the majority of medical centers that treat children with HD. Evaluation of the child's anatomy followed by appropriate categorization into obstructive or nonretentive symptoms will lead to appropriate management. Referral to another institution will vary depending on the availability of therapies such as anal sphincter Botox, or the need for a more advanced colonic manometry evaluation. In our cohort, only a minority of patients underwent colonic manometry, and these patients did not have a superior long-term outcome. As such, only a small number of patients would need to be referred for this evaluation using the algorithm.
In conclusion, the results of the present study suggest that following a comprehensive systematic algorithm for protracted DDs in children with postoperative HD may identify diagnoses that require specific therapies and leads to favorable long-term outcomes for the majority of children evaluated.
1. Iwashita T, Kruger GM, Pardal R, et al. Hirschsprung disease is linked to defects in neural crest stem cell function. Science 2003; 301:972–976.
2. Adzick NS, Nance ML. Pediatric surgery. First of two parts. N Engl J Med 2000; 342:1651–1657.
3. Russell MB, Russell CA, Niebuhr E. An epidemiological study of Hirschsprung's disease and additional anomalies. Acta Paediatr 1994; 83:68–71.
4. Georgeson KE, Robertson DJ. Laparoscopic-assisted approaches for the definitive surgery for Hirschsprung's disease. Semin Pediatr Surg 2004; 13:256–262.
5. Catto-Smith AG, Coffey CM, Nolan TM, et al. Fecal incontinence after the surgical treatment of Hirschsprung disease. J Pediatr 1995; 127:954–957.
6. Hyman PE. Defecation disorders after surgery for Hirschsprung's disease. J Pediatr Gastroenterol Nutr 2005; 41(Suppl 1):S62–S63.
7. Menezes M, Corbally M, Puri P. Long-term results of bowel function after treatment for Hirschsprung's disease: a 29-year review. Pediatr Surg Int 2006; 22:987–990.
8. Ciamarra P, Nurko S, Barksdale E, et al. Internal anal sphincter achalasia in children: clinical characteristics and treatment with Clostridium botulinum toxin. J Pediatr Gastroenterol Nutr 2003; 37:315–319.
9. Menezes M, Puri P. Long-term outcome of patients with enterocolitis complicating Hirschsprung's disease. Pediatr Surg Int 2006; 22:316–318.
10. Bax KN. Duhamel Lecture: the incurability of Hirschsprung's disease. Eur J Pediatr Surg 2006; 16:380–384.
11. Yanchar NL, Soucy P. Long-term outcome after Hirschsprung's disease: patients' perspectives. J Pediatr Surg 1999; 34:1152–1160.
12. Mills JL, Konkin DE, Milner R, et al. Long-term bowel function and quality of life in children with Hirschsprung's disease. J Pediatr Surg 2008; 43:899–905.
13. Dasgupta R, Langer JC. Evaluation and management of persistent problems after surgery for Hirschsprung disease in a child. J Pediatr Gastroenterol Nutr 2008; 46:13–19.
14. Keshtgar AS, Ward HC, Clayden GS, et al. Investigations for incontinence and constipation after surgery for Hirschsprung's disease in children. Pediatr Surg Int 2003; 19:4–8.
15. Levitt MA, Martin CA, Olesevich M, et al. Hirschsprung disease and fecal incontinence: diagnostic and management strategies. J Pediatr Surg 2009; 44:271–277.
16. Chumpitazi BP, Fishman SJ, Nurko S. Long-term clinical outcome after botulinum toxin injection in children with nonrelaxing internal anal sphincter. Am J Gastroenterol 2009; 104:976–983.
17. Fortuna RS, Weber TR, Tracy TF Jr, et al. Critical analysis of the operative treatment of Hirschsprung's disease. Arch Surg 1996; 131:520–524, discussion 524–5.
18. Weber TR, Fortuna RS, Silen ML, et al. Reoperation for Hirschsprung's disease. J Pediatr Surg 1999; 34:153–157.
19. Schweizer P, Berger S, Schweizer M, et al. Repeated pull-through surgery for complicated Hirschsprung's disease—principles derived from clinical experience. J Pediatr Surg 2007; 42:536–543.
20. Langer JC. Persistent obstructive symptoms after surgery for Hirschsprung's disease: development of a diagnostic and therapeutic algorithm. J Pediatr Surg 2004; 39:1458–1462.
21. Lindley RM, Hawcutt DB, Connell MG, et al. Human and mouse enteric nervous system neurosphere transplants regulate the function of aganglionic embryonic distal colon. Gastroenterology 2008; 135:205–216.
22. Di Lorenzo C, Solzi GF, Flores AF, et al. Colonic motility after surgery for Hirschsprung's disease. Am J Gastroenterol 2000; 95:1759–1764.
23. Pensabene L, Youssef NN, Griffiths JM, et al. Colonic manometry in children with defecatory disorders. Role in diagnosis and management. Am J Gastroenterol 2003; 98:1052–1057.
24. Minkes RK, Langer JC. A prospective study of botulinum toxin for internal anal sphincter hypertonicity in children with Hirschsprung's disease. J Pediatr Surg 2000; 35:1733–1736.
25. Blair GK, Murphy JJ, Fraser GC. Internal sphincterotomy in post-pull-through Hirschsprung's disease. J Pediatr Surg 1996; 31:843–845.
26. Heikkinen M, Rintala R, Luukkonen P. Long-term anal sphincter performance after surgery for Hirschsprung's disease. J Pediatr Surg 1997; 32:1443–1446.
27. Tiryaki T, Demirbag S, Atayurt H, et al. Topical nitric oxide treatment after pull through operations for Hirschsprung disease. J Pediatr Gastroenterol Nutr 2005; 40:390–392.