Current Understanding of Hirschsprung-Associated Enterocolitis: Pathogenesis, Diagnosis and Treatment

Abstract

Hirschsprung-Associated Enterocolitis (HAEC) was described in 1886 by Harald Hirschsprung and is a potentially deadly complication of Hirschsprung Disease. HAEC is classically characterized by abdominal distension, fever, and diarrhea, although there can be a variety of other associated symptoms, including colicky abdominal pain, lethargy, and the passage of blood-stained stools. HAEC occurs both pre-operatively and post-operatively, is the presenting symptom of HSCR in up to 25% of infants and varies in overall incidence from 20% to 60%. This article reviews our current understanding of HAEC pathogenesis, diagnosis, and treatment with discussion of areas of ongoing research, controversy, and future investigation.

INTRODUCTION:

Hirschsprung disease (HSCR, Online Mendelian Inheritance in Man #142623) is a common cause of neonatal bowel obstruction and was first described in 1886 by Harald Hirschsprung(1). HSCR can lead to the feared complication of Hirschsprung-associated enterocolitis (HAEC), which is the actual condition that Hirschsprung originally described in his initial report. HAEC is classically characterized by abdominal distension, fever, and diarrhea, although there can be a variety of other associated symptoms, including colicky abdominal pain, lethargy and the passage of blood-stained stools. At the time of his original description, Hirschsprung presented two children with constipation from birth who died after developing marked abdominal distension and loose stools; these would become the first reported cases HAEC. Although the concept of HAEC was alluded to in the literature in the 1950s by Burnard, Fisher and Swenson, and Dorman, it was not until 1962 that Bill and Chapman presented the first definitive description of the condition(2–5).

HAEC occurs both pre-operatively and post-operatively and is the presenting symptom of HSCR in up to 25% of infants(6–9). The incidence of enterocolitis ranges from 20% to 60%, with most prospective series identifying higher incidences than previously recognized(6, 10–13). HAEC can occur at any time from the neonatal period onwards into adulthood, although it appears to be less frequent in school-aged children. Additionally, the incidence appears to be independent of which surgical procedure for definitive management of HSCR has been performed. Recurrent HAEC can occur even in the presence of a diverting colostomy and is termed “diversion enterocolitis”(14, 15).

The diagnosis of HAEC has been predominantly based on clinical judgment. This has led to highly variable reports in the incidence rates of HAEC, ranging from 25–30% in recent series to as high as 60% in the past. Pre-operative incidence is estimated at 6–60%, and post-operative incidence ranges from 25–42%(13, 16, 17). The overall mortality rate associated with HAEC ranges from 1–10%, with the majority of deaths occurring in newborns prior to definitive operation(9). These newborn infants are at the highest risk of mortality among patients with HAEC. It is unclear whether these infants have higher mortality rates due to delays in diagnosis, immature immune responses, or some other unidentified mechanism. Fortunately, the mortality rate appears to be declining over time, likely due to increasing recognition of this devastating complication and prompt initiation of therapies such as rectal decompression, vigorous fluid resuscitation and antibiotic therapy.

However, despite the improvement in mortality rates in HAEC, the morbidity has a profound impact with prolonged hospitalization with a mean of 13 days (ranging from 6 to 29 days). Teitelbaum et al. found that neonates with HAEC have a mortality rate of 5% and a morbidity rate of 30%, and their hospitalization is twice as long as neonates without HAEC(18). The medical management of HSCR children with HAEC is 2.5 times costlier than of those without HAEC. Moreover, HSCR patients who develop HAEC have worse long-term bowel function than those that never develop HAEC, possibly secondary to inflammation-driven alterations to the enteric nervous system (ENS)(9).

This article reviews our current understanding of HAEC pathogenesis, diagnosis, and treatment with discussion of areas of ongoing research, controversy, and future investigation.

RISK FACTORS FOR HAEC

Many risk factors for HAEC have been identified including delay in the initial diagnosis of HSCR, male gender, a family history of HSCR, and the presence of trisomy 21. Delay in the diagnosis of HSCR leads to a higher incidence of HAEC as the presenting symptomatology(19). In the neonatal period, the incidence of HAEC increases from 11% in the first week of life to 24% after.

The best-established risk factor for HAEC is trisomy 21 with double the incidence of HAEC compared to non-trisomy 21 HSCR patients(19–21). The combination of HSCR and trisomy 21 is associated with a higher incidence of postoperative morbidity, prolonged hospitalization and poor long-term bowel function. Infants with trisomy 21 have an intrinsic immune deficiency due to both decreased cytotoxic T-lymphocytes and derangement in humoral function(22) which may explain their increased risk of HAEC. Of patients with trisomy 21, ~50% develop HAEC as opposed to 29% in the normal population(21, 23).

Other risk factors include family history of HSCR, male sex, delay in diagnosis of HSCR, and other genetic syndromes. Some have also postulated that the occurrence of a single episode of HAEC can alter intrinsic intestinal immunity leading to an increased risk of further episodes(10). There has been conflicting evidence on whether or not the length of disease is related to recurrent HAEC, with longer disease involvement postulated to have a higher rate of recurrence(5, 6, 10, 24). Studies have shown that HAEC is significantly more common in patients with aganglionic segments longer than the sigmoid(10, 24). Neonates with total colonic aganglionosis may present with perforation of the ganglionic bowel. However, some studies on this condition have found no difference as regards length of the aganglionic bowel(5, 6, 20).

There is no evidence that type of pull-through or presence of stoma after pull-through are related to the incidence of post-operative HAEC(25). Swenson reported an HAEC incidence of 21% after pull-through in a 40-year follow-up(26). However, Wildhaber et al. demonstrated no correlation between the incidence of HAEC and the type of pull-through performed(27). Others have noted similar findings(25, 28, 29). Additionally, no increase in HAEC has been found in the postoperative period after a primary pull-through without stoma formation(30).

After pull-through surgery, known risk factors for HAEC include anastomotic leak or stricture, and postoperative intestinal obstruction secondary to adhesive disease. These increase the relative risk of HAEC by nearly three-fold(17, 31). Finally, although HAEC does occur with a diverting colostomy/enterostomy, its incidence appears to be substantially lower.

PATHOPHYSIOLOGY

The last decade has seen marked progress in our understanding of the pathogenesis of HAEC. Building in a long history of clinical series postulating HAEC mechanisms, recent investigators have identified a sequential pathogenesis: 1) dysmotility, 2) dysbiosis of the microbiota, 3) altered intestinal barrier function, and 4) impaired mucosal immune responses, resulting in HAEC.

Dysmotility

The role of intestinal motility in the pathogenesis of inflammatory bowel disease is well described, with acute episodes of intestinal inflammation in Crohn’s disease associated with decreased motility(32, 33).

One of the earliest proposed causes for HAEC was impaired intestinal motility leading to functional obstruction with either subsequent bacterial stasis, overgrowth and translocation, or distension and ischemia(5, 34, 35). Bill and Chapman argued that partial mechanical obstruction was involved in the pathogenesis of HAEC causing mechanical dilatation of the proximal bowel leading to fecal stasis, resulting in further dilatation and subsequent mucosal ischemia and bacterial invasion (5). They suggested that enterocolitis only occurs in dilated ganglionic proximal bowel. However, this theory does not explain the enterocolitis that can occur in distal colon even with a defunctionalizing proximal stoma, the occurrence of HAEC in postoperative patients or histological evidence of enterocolitis in aganglionic bowel(26, 36).

More recently, structural analysis of the ENS has advanced our understanding of how dysmotility may contribute to HAEC pathogenesis. The ENS participates in host defense by modulation of secretory function and propulsion of luminal contents, thereby diluting and purging pathogens(37, 38). Using the EdnrB^NCC−/− mouse model of Hirschsprung disease, investigators have noted decreased neuronal density in the ganglionated bowel(39). Further, there is a shift in neurotransmitter phenotypes, with over-representation of nitrergic (relaxation) neurons and under-representation of cholinergic (contractility) neurons, altogether suggesting reduced numbers of neurons and a shift away from those that can function to propel luminal contents efficiently. Another group corroborated these findings in the EdnrB^−/− mouse model and extended the analysis to include human tissues, with a positive correlation between increased nitrergic neurons and post-operative enterocolitis(40). Finally, recently work has identified similar findings of small ganglion size and over-representation of nitrergic neurons in the ganglionated bowel of patients, with the extent of these imbalances correlating with medium-term patient-reported quality of life surveys(41). Together, these findings suggest that the ENS in the remaining ganglionated bowel after surgery for Hirschsprung disease may be insufficient to support normal function. Additionally, these studies open the possibility of more detailed analysis of pathologic specimens to include neurotransmitter subtypes, which may aid in counseling patients or families about future expectations for bowel function(41).

Dysbiosis of the Microbiota

Infectious etiologies have been linked to enterocolitis by several historical studies. With the advent of newer techniques for next generation sequencing and identifying components of the microbiome, our understanding has shifted from focusing on single organism etiologies to considering the concept of “dysbiosis.” Dysbiosis can be defined as 1) decreased diversity of species, 2) loss of beneficial species, and 3) increased pathobionts.

Current therapy for HAEC, consisting of systemic antibiotics, rectal irrigations and bowel rest (nothing to eat or drink), conceptually supports dysbiosis as a component of HAEC pathogenesis (42). Systemic antibiotics are used empirically in HAEC, with metronidazole recommended because of activity against anaerobes and Clostridium difficile, an organism that was once thought to be associated with HAEC (43). Further, fecal stasis is treated by daily (or multiple times daily) rectal irrigations aimed at reducing the fecal burden. In the last decade, mouse model studies have shed light on the relationship between host microbiome and development of HAEC.

The most common genetic defects associated with HSCR are mutations of rearranged during transfection (Ret) and endothelin receptor B (EdnrB), which are both required for NCC migration and ENS formation(44). Mice with mutations in EdnrB demonstrate congenital megacolon with absent distal ganglion cells similar to the majority of patients with HSCR and are an excellent model of HSCR(45, 46). Using the EdnrB^−/− mouse model, Ward et al. demonstrated increasing microbiome diversity over 24 days, with a greater increase in HSCR mice versus wild type(47). They identified clusters of microbiota in each group, showing that wild type (non-HSCR) and HSCR mice had distinct microbiomes. Further, HSCR mice were found to have higher levels of Bacteroidetes and Firmicutes than controls. Similarly, Pierre et al. found evidence of comparable microbiomes between HSCR mice and controls early in the neonatal period, with divergence of the microbiota between HSCR and controls as the onset of HAEC approached(48). HSCR mice expressed increased Bacteroidetes and Clostridium species, and E.coli was found only in HSCR mice. Both studies also showed decrease in Lactobacillus over time in HSCR mice. Another group demonstrated that EdnrB^−/− median survival can be extended to 36 days by changing to a liquid diet and the addition of oral antibiotics (49) and we have observed survival beyond 50 days in EdnrB^NCC−/− mice that are raised pseudo-germ free (on broad spectrum antibiotics) [data not shown], further supporting a role for dysbiosis of the microbiome in the development of HAEC.

Next-generation sequencing approaches have recently been applied to patients, although most series are limited to single patients or small groups. DeFilippo et al. used amplified ribosomal DNA restriction analysis to demonstrated distinct changes in the microbiota of a single child as he progressed from pre-enterocolitis through the acute episode, and onto resolution(50). Yan et al. used this technique in two patients with HAEC and two without and found different bacterial clustering in the patients with HAEC as compared to those without(51). Recently, Neuvonen et al. reported 16S findings on 34 patients with HSCR, ranging from 3–34 years of age (52). Key findings of their study correlate with those reported in the mouse models, including the hallmarks of dysbiosis: altered diversity, failure of maturation of the community, increased pathobionts and decreased commensals. Additionally, Shen et al. reported 16S findings on 30 HSCR patients compared to patients undergoing non-bowel surgery, and noted a decrease in Bifidobacteria and Lactobacilli (53), similar to the mouse models. Ongoing studies are beginning to investigate a potential role for alterations in the fungal communities of the gut and their contribution to HAEC pathogenesis.

Intestinal barrier dysfunction

The intestinal barrier plays a critical role in maintaining host health. The amount of mucus produced during HAEC is obvious and dramatic. Goblet cells produce mucus, which helps to maintain epithelial integrity by serving as a scaffolding for bactericidal and bacteriostatic proteins. Abnormalities in mucin production, therefore, may contribute to the pathogenesis of HAEC.

In 1981, Akkary et al studied rectal biopsies of HSCR patients and found a marked increase in the volume of mucin compared to control tissues(54). Teitelbaum et al hypothesized that HSCR involves an alteration in the composition of mucin in the colon, such that there is mucin retention and crypt dilation, and proposed a histologic grading system which is unique to HSCR and cystic fibrosis (18). A separate study of mucin turnover showed that HSCR patients who developed HAEC had turnover rates 7-fold lower than HSCR patients that did not develop HAEC(55). Additionally, Hirschsprung patients also have decreased MUC-2, the predominant mucin expressed in the human colon, and MUC-2 is undetectable in patients with HAEC(56, 57). This may suggest an intrinsic defect that allows for bacterial adherence and translocation. Finally, Thiagarajah et al compared tissue from HSCR patients and the distal colon of EdnrB^−/− mice and found increased goblet cell numbers when compared to controls(58). They then used trans-epithelial resistance measurements to assess for functional differences and noted that trans-epithelial resistance and fecal dehydration were increased in the distal colon of EdnrB^−/− mice. Additionally, they found increased mucus viscosity and therefore impaired particle diffusion in null mice. Taken together, these findings suggest that alterations in mucus production and function may play a role in the development of HAEC.

Impaired mucosal immunity

Abnormal immune cell function has been implicated in the development of HAEC. Secretory IgA immunoglobulin provides a major immunological barrier in the gastrointestinal tract. IgA is the predominant immunoglobulin at all levels in the intestinal tract both in the lumen and within the wall. Albanese et al. have shown that secreted IgA binds to bacteria and prevents bacterial translocation across an intact segment of viable intestinal tissue(59). In the late 1980’s, Wilson-Storey et al. conducted a series of studies which demonstrated that patients affected by HSCR have impaired transfer of secretory IgA across the GI mucosa(60, 61). Specifically, they noted that although HSCR patients had increased IgA in their buccal mucosal tissue, there was a significant decrease of secretory IgA in the saliva. Similarly, the plasma cells in the lamina propria of the bowel were found to have significantly increased levels of IgA, IgM and IgG in HAEC bowel compared to non-HAEC bowel(62). Those same HAEC patients were found to have decreased luminal IgA, suggesting decreased production or impaired transport into the lumen.

Interestingly, the finding of decreased luminal IgA has also been observed in EdnrB^NCC−/− mice, in which EdnrB is deleted only in neural crest cells which form the ENS(46). The EdnrB^NCC−/− mouse develops colorectal aganglionosis and HAEC similar to human HSCR. One studied showed not only that luminal IgA is reduced, but that this finding is specific to the gut, with normal levels of bronchial and nasal IgA observed in these animals(46). The gut-specific decrease in luminal sIgA is accompanied by accumulation of IgA⁺ plasma cells in the lamina propria (46, 63). Together, these results suggest decreased IgA production or transport in HSCR/HAEC as a potential therapeutic target.

Mucosal neuroendocrine cells (NE) mediate intestinal function through synthesis and storage of neuroendocrine neuropeptides and biogenic amines which act as chemical messengers(64). Soeda et al. demonstrated that NE cells are increased in the aganglionic segment of bowel in HSCR as opposed to the ganglionated bowel and normal controls(65). They noted a marked reduction in NE cells in ganglionated bowel in HAEC compared to those without. These diminished NE cells may represent an impaired immune response or a deficiency which may facilitate the initialization of inflammation(66). This impaired immune response theory is echoed in trisomy 21. The combination of HSCR and trisomy 21 is associated with a higher incidence of enterocolitis with 50% of patients with trisomy 21 and HSCR developing HAEC in contrast to 29% among the normal population(44). Infants with trisomy 21 have an intrinsic immune deficiency due to both decreased cytotoxic T-lymphocytes and derangement in humoral function which may explain their increased risk of HAEC(22).

Histological evidence of enterocolitis consists of a number of features including crypt abscesses, leukocyte aggregates, ulceration and Paneth cell metaplasia(67). Paneth cells are normally present in the small bowel and secrete lysozymes which digest the bacterial wall membranes. Their presence in HAEC colon suggests an attempt at reinforcement of the mucosal immunity. ICAM-1 is a cell surface intercellular adhesion glycoprotein which is involved in leukocyte recruitment when inflammation occurs. Kobayashi et al. have demonstrated that ICAM-1 shows increased expression in the endothelium of both the ganglionated and aganglionic bowel in patients with HAEC(68). This emphasizes the importance of endothelial cell activation in HAEC pathogenesis. Elhalaby et al. postulated that the occurrence of a single episode of HAEC can alter intrinsic intestinal immunity by causing a chronic change to the mucosa to an increased the risk of further episodes(10). This would help to explain the lower but real recurrence rate of HAEC following a “diversion” colostomy or a successful pull-through.

Splenic lymphopenia is also thought to contribute to an etiology of impaired immunity. This was first described in the EdnrB^−/− mouse model by Cheng et al(69). These animals have abnormal splenic architecture and reduced total lymphocytes in the spleen. Specifically, they have a relative reduction in B as compared to T lymphocytes, as well as a negative correlation between splenic lymphocyte counts and intestinal inflammation on histologic analysis. This finding was confirmed in the EdnrB^NCC−/− model, with the additional discovery of a decrease in marginal zone B-lymphocytes, suggesting impaired B lymphocyte development or trafficking from the spleen to the Peyer’s Patches of the small intestine(46). Another group attempted to understand the contribution of the EdnrB genotype to the clinical expression of HAEC by performing bone marrow transplants from EdnrB animals to Rag2^−/− recipients and inducing bowel obstruction in wild type animals(70). They concluded that stress from obstruction resulted in similar lymphocyte alterations to those seen in HAEC models. However, they found that after surgical relief of obstruction, EdnrB^−/− mice still carried a 40% risk of developing HAEC(71).

Summary

Dysmotility, dysbiosis of the microbiota, altered intestinal barrier function, and impaired mucosal immune responses all appear to contribute to the pathogenesis of HAEC. ENS dysfunction can result in microbiome dysbiosis through impaired motility and stasis. When followed by impaired intestinal barrier function and an abnormal mucosal immune response, HAEC develops [Figure 1]. Each point in this sequence is of potential interest for targeted prevention or therapy.

Figure 1. Current understanding of Hirschsprung-Associated Enterocolitis pathogenesis.

Dysmotility, dysbiosis of the microbiota, altered intestinal barrier function, and impaired mucosal immune responses all appear to contribute to the pathogenesis of HAEC. ENS dysfunction can result in microbiome dysbiosis through impaired motility and stasis. When followed by impaired intestinal barrier function and an abnormal mucosal immune response, HAEC develops. Each point in this sequence is of potential interest for targeted prevention or therapy.

DIAGNOSIS

The clinical presentation of HAEC can be highly variable in both symptoms and severity(42). Classic manifestations include abdominal distension, fever, and diarrhea. The spectrum of presentations is large and includes non-specific signs and symptoms, likely contributing to the variable incidence of HAEC observed in the literature. Due to the difficulty in making a definitive diagnosis and potential for morbidity or mortality with late diagnosis and under-treatment, most practitioners make a presumptive diagnosis and initiate therapy, which may result in over-treatment(72).

Mild cases may present with fever, mild abdominal distension, and diarrhea, mimicking viral gastroenteritis. More severe cases may include lethargy, rectal bleeding, and obstipation. In the neonate the classical presentation consists of a history of constipation from birth associated with occasional loose foul-smelling stools and progressive abdominal distension(10, 12). Among neonates with HSCR, 16–33% present with diarrhea. The presence of diarrhea is considered pathognomonic of enterocolitis but varies in incidence from 17–93% of patients with HAEC. A markedly distended hyper-resonant abdomen occurs in 32–83%, vomiting in 9–76%, pyrexia in 12–54%, and less commonly rectal bleeding in 5–9% of patients with HAEC(72, 73). Due to the heterogeneity of clinical presentation, providers must be alert to the possibility of HAEC in all patients with HSCR, as well as in neonates presenting with signs of distal bowel obstruction.

Rectal examination, either by digit or soft catheter, is both diagnostic and therapeutic, resulting in a characteristically explosive foul smelly stool and gaseous decompression. Patients after a pull-through operation or those with a diverting stoma will present in the same fashion. The significant morbidity associated with HAEC occurs with the toxic megacolon which is characterized by bilious vomiting, fever, dehydration, marked abdominal distension, and signs of shock. Fortunately, bowel perforation is a rare complication occurring in only 2–3% of patients(10).

Current diagnostic practice involves excluding other causes of colitis such as necrotizing enterocolitis in the infant and infectious colitis in older children. Stool studies and Clostridium difficile testing can be helpful to rule out the latter. Although in most patients the diagnosis can be made easily on clinical evaluation, certain radiographic findings have been associated with HAEC in the context of a suspicious clinical history. Simple anterior-posterior and lateral decubitus abdominal radiographs can show thickening of the bowel wall, mucosal irregularity (“sawtooth” appearance), dilated bowel loops, air-fluid levels, “cutoff” sign in the rectosigmoid colon, pneumatosis, pneumoperitoneum and evidence of toxic megacolon (grossly dilated colonic loop). Contrast enema should be avoided during episodes of HAEC due to the risk of perforation.

A major barrier in the care of Hirschsprung patients has been the lack of a standardized method of diagnosing HAEC. In 2009, a large group of gastroenterologists and surgeons participated in a Delphi process to generate a diagnostic scoring system for HAEC(74, 75). Using history, physical exam findings, laboratory findings and imaging, they arrived at a 16-item list. Each item was assigned 1–2 points, with a summed score of 10 or greater being diagnostic of HAEC. Despite multidisciplinary expert input into the development of this definition, it was not designed for clinical use and has not been widely adopted.

The American Pediatric Surgery Association’s (APSA) Hirschsprung Disease Interest Group also proposed a staging system for HAEC(42). This system classifies HAEC into three stages based on many of the same history, physical exam, and radiologic features, and is aimed towards aiding in both the diagnosis and management of HAEC. However, it is hampered by the same primary weakness of the Delphi criteria – it relied on expert opinion in its development.

Recently, another collaborative reviewed the medical records of 116 children across five centers using the 16 Delphi criteria with the goal of creating a clinically useful scoring system(76). The most common positive criteria included distended abdomen (31%), diarrhea with explosive stool (24%), diarrhea with foul-smelling stool (23%) and lethargy (19.8%). On multivariate analysis, diarrhea with explosive stool, decreased peripheral perfusion, lethargy, and dilated loops of bowel were independently associated with suspected HAEC episodes. Based on the calculated sensitivities and specificities for each score point, they demonstrated that a cutoff score of 4 points maximized the sensitivity (83.7%) and specificity (98.6%) in diagnosing HAEC. Using the same patient cohort, they then devised a new risk score based on these four criteria(73).

More recently, another group utilized four international centers to obtain data on 200 HSCR patients with a total of 1450 encounters, including 369 HAEC episodes(72). In their retrospective series, 57% of HSCR patients experienced one or more episodes of HAEC. Long segment colonic disease was associated with a higher risk of HAEC on univariate analysis. They identified six variables (fever, bloody diarrhea, obstipation, distention, dilated loops of bowel on abdominal x-ray, and leukocytosis) that were significantly associated with HAEC on multivariate analysis. Using published diagnostic cutoffs from prior scoring systems, they demonstrated that the sensitivity/specificity for existing systems were found to be 38.2%/96% for the original Delphi criteria and 56.4%/86.9% for Frykman’s score using four criteria. The new proposed scoring system demonstrated an improved sensitivity of 67.8% and specificity of 87.9%. They suggested that the new score outperformed the existing scores by decreasing under-diagnosis in this patient cohort.

Ongoing studies from multiple groups are attempting to improve the performance of these scoring systems by gathering prospective data and comparing the results head-to-head.

TREATMENT

There is currently no evidence-based, standard of care guideline or algorithm for treatment of HAEC. Therapy is non-specific and aimed at treating symptoms rather than a known etiology. Fluid resuscitation and correction of electrolyte abnormalities are critical in initial management. Additional management strategies may include dietary changes, antibiotics, rectal irrigations, and intensive care unit admission.

Treatment regimens should be tailored to the providers’ clinical judgement of the severity of disease. The APSA HSCR interest group published guideline for the treatment of HAEC based on their diagnostic guidelines (42). Three grades of disease severity are defined as suspected HAEC (grade 1), definite HAEC (grade 2), and severe HAEC (grade 3). The APSA guidelines recommend that a patient with grade 1 HAEC may be safely treated as an outpatient with oral metronidazole and oral hydration. The optimal dosing, frequency and duration of antibiotics for HAEC has not been determined. Rectal irrigations (washouts) can be considered in patients with abdominal distension or incomplete evacuation. Rectal washouts are performed using a large-bore soft catheter with multiple side holes. The tube is well lubricated and advanced into the colon. In preoperative HAEC the tube should be passed into the transition zone if technically possible. Repeated tube decompression and gentle rectal washouts with 10 mL/kg aliquots of warm or room temperature normal saline can make a significant clinical impact on these patients.

For grade 2 cases, inpatient or outpatient management is left to the provider’s clinical judgement. Dietary restriction options include clear liquids or nothing by mouth. Broad spectrum antibiotics and rectal irrigations/washouts are recommended. Patients with severe cases of HAEC (grade 3) should be strongly considered for intensive care unit admission, bowel rest, broad spectrum antibiotics and rectal irrigations. These patients may require proximal diversion if there is failure to improve with non-operative management. Clinical deterioration in the neonate, particularly those with long-segment disease in which washouts have a high failure rate, may require an emergency colonic decompression and diversion.

Recurrent HAEC

Treatment of recurrent HAEC begins with identifying the underlying cause. The workup for underlying etiology begins with assessing for causes of obstructive symptoms(77). Anatomic etiologies can be identified with contrast enema, physical examination under anesthesia, and rectal biopsies to confirm the presence of ganglionated bowel. Anatomic abnormalities, such as anastomotic stricture, transition zone pull-through, or Duhamel spur should be treated surgically. Redo pull-through operations when appropriate appear to be as effective as primary procedures in terms of continence and stooling frequency and can decrease episodes of HAEC(78).

After excluding anatomic etiologies, non-relaxation of the internal anal sphincter should be considered. The use of botulinum injections for treatment of post-operative HAEC has shown some promising results. In one study, 14 of 18 patients with persistent constipation, obstructive symptoms, or recurrent HAEC showed improvement in bowel function and 5 of these had improvement that lasted longer than 6 months(79). Multiple studies have shown a reduction in hospitalizations for HAEC following botulinum injections(80). However, it is difficult to predict which patients will respond, and long-term outcomes have not been well studied.

Prophylactic measures against HAEC

After pull-through surgery, some surgeons recommend routine anal dilations. Gao et al reported an enterocolitis rate of 2/34 (6%) after using routine dilations for three months after surgery(81). However, recent data questions these findings. A review by Temple et al compared rates of stricture development and enterocolitis among children with HSCR and anorectal malformation undergoing either weekly calibration of the anastomosis by a surgeon versus daily dilation by parents and observed no differences(82). A separate review of HSCR patients had similar findings(83).

Concerns over the mortality rate due to fulminant enterocolitis in the postoperative period led Marty et al. to suggest routine postoperative rectal washout to decrease both the incidence and the severity of episodes of enterocolitis following definitive surgery(84). They recommend a policy of rectal irrigation performed twice daily by the parents commencing 2 weeks following surgery continuing for 3 months, followed by once daily for 3 months. This policy reduced their incidence of HAEC from 36% (34 of 95 patients) to 10% (4 of 40 patients). A Spanish study of 37 children with HSCR treated between 1978 and 2005 found similar results(85).

Further research is needed to investigate what role probiotics might play in the prevention of HAEC. In one study, children undergoing surgery for HSCR were randomized to probiotic versus placebo post-operatively; this study did not show differences in HAEC rates between the two groups(16). In contrast, another group similarly randomized patients and treated them for 4 weeks(86). The probiotic group had reduced incidence and severity of HAEC over the following 3 months. A recent meta-analysis concluded that probiotics do not reduce the risk of HAEC(87).

CONCLUSION AND FUTURE DIRECTIONS

HAEC remains a diagnostic and therapeutic challenge, despite recent advances in our understanding of the pathophysiology. Current research is focusing on a number of avenues including patient-specific microbiome analysis and targeted probiotic therapy and use of stem cell therapy to restore bowel function in HSCR(88). Further studies into the underlying mechanisms of disease, accurate methods of diagnosis, and optimal treatment strategies will be needed to improve our ability to care for HSCR patients with HAEC.