Ileostomy Prolapse in Children with Intestinal Dysmotility

Background A relationship between intestinal motility and ileostomy prolapse has been suggested but not demonstrated objectively. Aims This study evaluated the association between ileostomy prolapse and intestinal dysmotility in children. Methods IRB-approved retrospective review of 163 patients with ileostomies (1998–2014) at a single institution. Patients were categorized as having clinical dysmotility as a primary diagnosis (n = 33), clinically suspected dysmotility based on underlying diagnosis (n = 60), or intestinal dysmotility unlikely (n = 70) at the time of ileostomy present. Intestinal manometry was categorized as normal (n = 13) or abnormal (n = 10). Primary outcome was pathologic stoma prolapse. Multivariate analysis using a logistic regression model and log-rank test to compare stoma prolapse rates over time between motility groups were used. Results Clinical diagnosis of dysmotility (p ≤ 0.001) and manometric findings of dysmotility (p = 0.024) were independently associated with stoma prolapse. Clinical dysmotility correlated with manometric findings (κ = 0.53). Prolapse occurred in 42% of patients with dysmotility, 34% of patients with suspected dysmotility, and 24% of patients with normal motility. One-year prolapse-free stoma “survival” was 45% for dysmotility, 72% for suspected dysmotility, and 85% for intestinal dysmotility unlikely groups (p = 0.006). Conclusions Children with intestinal dysmotility are at great risk for stoma prolapse. Intestinal manometry could help identify these patients preoperatively.


Introduction
Chronic intestinal pseudo-obstruction (CIPO), first described in 1978, is a syndrome characterized by severe congenital gastrointestinal dysmotility and frequent episodes mimicking mechanical obstruction. Despite medical and surgical therapy, the morbidity and mortality associated with intestinal motility remain high [1,2]. When surgery is required, both proximal (gastrostomy or jejunostomy) and distal (ileostomy or colostomy) decompression are frequently attempted [1]; however, these are not curative procedures and have high rates of complication [2,3].
Ileostomy creation has been associated with frequent morbidities in both adults and children, including skin excoriation, bleeding, obstruction, stenosis, fistula formation, prolapse, malnutrition, and reoperations [4][5][6]. Prolapse is reported in 13% of adult ileostomies, and in the setting of CIPO, intestinal reoperations have been associated with increased mortality [3,6]. A relationship between intestinal motility and ileostomy prolapse has been suggested anecdotally, yet this link has not been demonstrated objectively [7]. The purpose of this study is to evaluate the incidence of ileostomy prolapse in children with and without intestinal dysmotility.
The hypothesis of this study was that patients with clinical and/or manometric small intestinal dysmotility would have a higher rate of ileostomy prolapse than patients with normal motility. To allow for a robust analysis, dysmotility was assessed with the use of both diagnostic categories based on clinical evaluation and intestinal manometry studies (n = 23). Prior to data analysis, each patient was reviewed by a gastroenterologist specializing in motility disorders to determine the best diagnostic category (L.R.). Patients with a clinical diagnosis of dysmotility or CIPO were placed in the "primary dysmotility" category (n = 33). Patients with a separate primary diagnosis but documentation of suspected dysmotility and patients with a diagnosis which frequently predisposes them to dysmotility (NEC, gastroschisis, atresia, aganglionosis, and cystic fibrosis) were categorized as "suspected dysmotility" (n = 60). All remaining patients were categorized as "intestinal dysmotility unlikely" (inflammatory bowel disease, constipation, anorectal malformations, and abdominal neoplasm; n = 70). Of note, patients with intractable constipation were included in the "intestinal dysmotility unlikely" category because they had no upper intestinal symptoms. All of these patients had their constipation symptoms resolved completely with placement of an ileostomy. This grouping scheme was validated by using Cohen's kappa coefficient to evaluate agreement between dysmotility categorization and the "gold-standard" diagnostic modality, intestinal manometry. Manometry studies were performed at the discretion of a gastroenterology staff both prior to and after initial ileostomy creation following a standardized protocol that includes 3 phases: fasting motility recording for 3 hours followed by feeding challenge when tolerated (either oral or via enteral tube) and a medication challenge (IV erythromycin to stimulate antral contractions and subcutaneous octreotide to stimulate phase III of the migratory motor complex).
The primary outcome was any pathologic stoma prolapse, defined as a prolapse episode requiring a physicianperformed intervention (manual reduction with or without anesthesia, stoma revision, or stoma takedown) to achieve resolution. Patients were tracked until pathologic prolapse was documented, the stoma was reversed for other indications, or the patient was lost to follow-up. Demographic and clinical characteristics were assessed using Mann-Whitney U tests for continuous variables and Fisher's rank test for categorical variables. Multivariate analysis was performed by a logistic regression model for each possible predictor of prolapse. Variables with a p value ≤ 0.20 on univariate analysis were included in the final model. Kaplan-Meier survival tables and curves were generated and factors compared using the log-rank (Mantel-Cox) test.

Results
Clinical characteristics of the 163 patients with ileostomies included in this study are described in Table 1. At the time of ileostomy creation, these patients had a median age of 42 months (IQR 1.3-162). Median follow-up duration was 8 months (IQR 3-24). Most patients (76%) had end ileostomies while 9% had double-barrel ileostomies, and 15% had loop ileostomies.
The primary diagnosis requiring ileostomy creation for each patient is listed in Table 2. Of the total population, 33 patients had CIPO as a primary diagnosis and indication for ileostomy creation. 60 patients had dysmotility suspected based on either underlying diagnosis or clinical assessment. The remaining 70 patients had an anatomic or other indication for ileostomy creation unrelated to small bowel motility. 10 (43%) of 23 patients who underwent manometry testing were diagnosed as dysmotile. The dysmotility categories correlated well with manometric findings (κ = 0 53).

Discussion
In this investigation, 163 children with end ileostomies were followed for a median of 8.4 (IQR 3-24) months to evaluate incidence of stoma prolapse, which was observed in 38 (24%) of all patients. Each patient was categorized into three groups based on clinical intestinal dysmotility, and this classification was validated by comparison to intestinal manometry results.
Multivariate regression demonstrated that intestinal manometry results diagnostic of dysmotility were associated with higher rates of prolapse. Significant agreement (κ = 0 53) was observed between manometry results and motility categorization, validating this approach for comparison. Prolapse appears to be primarily an early complication of ileostomy creation, occurring at a median of two months (IQR 0.3-3.5) and with only three occurrences of pathologic prolapse occurring more than one year after ileostomy creation. Consequently, the median follow-up duration of patients without prolapse was significantly longer than the median time to prolapse (p < 0 001). To further assess the possible interaction of dysmotility and prolapse, Kaplan-Meier analysis was performed. One-year prolapse-free stoma "survival" was poorest in patients with a primary diagnosis of intestinal dysmotility (p = 0 006).
Chronic intestinal pseudo-obstruction is described as an amalgam of congenital syndromes characterized by severe intestinal dysmotility with frequent symptoms of obstruction [8,9]. Diagnosis is difficult and must follow careful elimination of other underlying pathologies including mechanical obstruction and aganglionosis [9,10]. CIPO is frequently accompanied by abnormal biopsy or intestinal manometry results [10,11]. Distinct variations of neurogenic and myogenic CIPO are reported, though many patients have idiopathic disease [11,12]. Some series of familial disease have also been reported [13].
Ileostomy complications have frustrated patients and surgeons for as long as ileostomies have been performed [4,6]. The morbidity of ostomies is especially high in infants and children [5,[18][19][20]. In particular, ileostomy prolapse has proven a frequent complication [7,14,21]. Many authors have suggested a variety of approaches to prevent or correct prolapse, including suture fixation of the bowel to the abdominal wall (stoma pexy), divided loop ileostomy, subcutaneous tunneling of bowel, Deflux® (hyaluronic acid/dextranomer, Salix Pharmaceuticals Inc., Raleigh, NC) injection, and stoma revision or takedown [21][22][23][24][25][26][27][28][29]. As none of these interventions have consistently proven effective, the potential risks and benefits of these options need to be very carefully assessed for each patient. This study was unable to identify any surgical technique which successfully reduced the incidence of stoma prolapse. Certainly, the lack of standardization in operative technique may contribute to the lack of any findings. In addition, it is possible, due to the retrospective nature of this analysis, that true differences could not be detected.
While intestinal motility correlated well with the clinical groups created, not all patients underwent upper intestinal manometry. The utility of preoperative manometry in identifying children at high risk for prolapse may be beneficial. Prospective investigations are necessary to determine if employment of this test may result in avoidance of an otherwise planned ileostomy or in more informed preoperative counseling of patients and families.
There are limitations to this study. Our study was retrospective in nature, and we cannot control for specific variations between surgeons for technical approach to ileostomy creation. Certainly, a prospective study controlling for surgical technique may determine if specific surgical techniques help prevent prolapse. Only 23 patients had manometry studies, limiting the number of patients diagnosed via this method. Prolapse may have influenced clinicians' decision making to diagnosis dysmotility for patients that were diagnosed postoperatively; in the future, preoperative manometry studies may help confirm dysmotility prior to stoma formation. Additionally, while this study did have a 25% loss to follow-up, this is not unexpected in a long-term retrospective analysis. Of note, the median observation period for patients lost to follow-up was 30.6 (IQR 14.0-63.4) months, which is longer than the median time to prolapse by an order of magnitude, suggesting that lack of followup in these patients is unlikely to have affected the outcome of the analysis. These data corroborate findings by Irtan et al. who have suggested an association between CIPO and stoma prolapse in a small group of patients with ileostomies and colostomies [7]. Ileostomy prolapse appears to be a relatively frequent and early complication in patients with clinical and/or manometric dysmotility. Chronic intestinal pseudo-obstruction and other dysmotility syndromes result in the need for complex management, and patients undergoing ileostomy for decompression may see benefits which outweigh the potential morbidity of ileostomy prolapse. The knowledge that stoma prolapse is indeed seen with higher frequency in this patient population helps to inform surgical decision making and counseling and will hopefully result in more focused investigations such as evaluation of the increased use of manometry to predict ileostomy morbidity. Management by a multidisciplinary team (including pediatric Overall, a logrank test determined that dysmotility category significantly differed with respect to the rate of prolapse (χ 2 = 16 828; p < 0 001). Furthermore, a log-rank test established that confirmed dysmotility significantly differed from suspected dysmotility (χ 2 = 5 386; p = 0 020) and unlikely dysmotility (χ 2 = 16 926; p < 0 001); however, suspected dysmotility was not found to significantly differ from unlikely dysmotility (χ 2 = 2 511; p = 0 11). surgery, gastroenterology, nutrition, social work, and transplant surgery) with experience in management of CIPO may be of benefit to limit high-risk procedures and aid future prospective investigation.

Disclosure
An earlier version of this work was presented as abstract in the APSA 46th Annual Meeting in 2015.

Conflicts of Interest
There are no competing interests to declare.