Clinical Characteristics of Pediatric Intussusception And Predictors of Surgery And Bowel Resection In Affected Patients


 Surgery is required for the treatment of intussusception when enema reduction is unsuccessful, or when the patient develops peritonitis, bowel perforation, or intestinal damage. We aimed to evaluate the clinical and laboratory parameters that may be used to predict the need for bowel resection in children with intussusception. This observational retrospective study included children who were admitted to the pediatric emergency room with intussusception. Multivariate logistic regression models were used to evaluate factors associated with bowel resection. In total, 584 children with intussusception were admitted to the pediatric emergency room; 129 of these children underwent surgery. Multivariate analysis revealed the following independent predictors of surgery for intussusception: abdominal pain (odds ratio [OR] = 0.372; p = 0.013), bloody stool (OR = 3.553; p = 0.044), and hyponatremia (OR = 4.12; 95% p = 0.003). Symptoms for at least 2 days before surgery (OR = 6.863; p = 0.009), long intussusception (OR = 5.088; p = 0.014), pathological lead point (OR = 6.926; p = 0.003), and intensive care unit admission (OR = 11.777; p = 0.001) were factors independently associated with bowel resection. These findings can be used to identify patients at high risk of needing surgery and bowel resection.


Introduction
Intussusception is the major cause of intestinal obstruction in infants and children younger than 3 years, with an annual incidence of 1-2 cases per 1,000 children younger than 3 years 1,2 . It occurs when one segment of the bowel (i.e., lead point or intussusceptum) invaginates into the distal bowel lumen (intussuscipiens), resulting in venous congestion, bowel wall edema, and bowel obstruction. Surgery for intussusception is required if enema reduction is unsuccessful; it is also required in cases of peritonitis, bowel perforation, or intestinal damage. Surgery for intussusception is required in 1.4-56% of affected patients 3,4 ; bowel resection is required in 9-59% of affected patients [5][6][7] . Postoperative complications occur in 8-12% of patients with intussusception 8 . Bowel resection is often associated with longer duration of symptoms, postoperative complications, and prolonged hospital admission 4,[8][9][10] . Understanding the characteristics of intussusception and risk factors for emergency bowel resection can alert clinicians to the correct diagnosis and need for thorough diagnostic tests.
The clinical manifestations of intussusception in children are varied and non-speci c [11][12][13] . The classic triad of pediatric intussusception includes vomiting, abdominal pain, and bloody currant jelly stool/palpable abdominal mass; however, this occurs in fewer than half of affected patients 14 .
Therefore, it is challenging to diagnose intussusception in the pediatric emergency room (PER). Identifying the need for surgery or bowel resection in children with intussusception will help clinicians to determine the prognosis at hospital discharge and provide appropriate care in the PER. Clinical symptoms and signs, laboratory ndings, and imaging ndings in the PER may identify patients who require surgery or bowel resection. Although children who require surgery have worse outcomes than children who undergo successful pneumatic reduction, there have been limited reports regarding the clinical and laboratory parameters at triage that predict the need for surgery or bowel resection in patients with intussusception.
In this study, we compared the clinical and laboratory parameters in the PER of children with intussusception, with the aim of identifying predictors of the need for surgery or bowel resection. We aimed to provide insights into the different clinical presentations of children with intussusception who require surgery or bowel resection; we also aimed to evaluate the predictors of need for surgery or bowel resection, which would allow prompt determination of patient outcome.

Methods
Study setting and patient selection.
This retrospective, observational study was conducted in a PER where 30,000 children (aged 0-18 years) are treated annually. We screened the medical records of patients who presented between January 2010 and December 2015 to identify patients who were assigned the International Classi cation of Diseases, Ninth Edition (ICD-9) diagnostic code of intussusception. This study was approved by the Institutional Review Board of our hospital (201700560B0). All methods were performed in accordance with the relevant guidelines and regulations.
The data were collected, reviewed, de-identi ed, and anonymously analyzed by the authors, and the ethics committee waived the requirement of informed consent because of the anonymized nature of the data and scienti c purpose of the study.

Study design.
Information was collected from the electronic database regarding the clinical presentation and vital signs at triage, imaging ndings, and surgical records of children who underwent surgery. The initial clinical presentation and vital signs were recorded by trained triage nurses who categorized patients according to their care needs upon arrival in the PER. We excluded patients who were not hospitalized or did not have intussusception (e.g., patients with ovarian cyst, renal tumor, appendicitis, or hematological cancer). The parameters evaluated for potential association with bowel resection included sex; age; clinical presentation at triage; ndings of laboratory tests, X-ray imaging, and computed tomography imaging performed in PER; pneumatic reduction records; surgical records; and pathological ndings. The following clinical de nitions were used: fever, body temperature ≥ 38°C; tachypnea, respiratory rate ≥ 24 times/min; tachycardia, heart rate ≥ 121 beats/min; and cyanosis, saturation ≤ 89%. Abdominal pain was recorded as mild, moderate, or severe when the pain score was < 4, 4-7, or ≥ 8 in a child, respectively; abdominal pain was also recorded in the event of inconsolable irritable crying in an infant. Additional clinical de nitions were as follows: hyponatremia, serum sodium ≤ 134 mEq/L; hypochloremia, serum chloride ≤ 100 mEq/L; hypokalemia, serum potassium < 3.5 mmol/L; and hyperglycemia, blood glucose ≥ 100 mg/dL. Pneumatic reduction was considered the rst-line treatment for patients with intussusception; surgery was considered if pneumatic reduction failed. Prolonged time to surgery was de ned as the presence of symptoms for at least 2 days before surgery. Prolonged length of intussusception was de ned as intussusception lesion length > 15 cm. Pathological lead point was de ned as the pathological source of the intussusception identi ed during surgery.

Statistical methods.
Descriptive statistics were used to evaluate differences in clinical characteristics and outcomes between patients with intussusception who did and did not require surgery or bowel section. Categorical variables were compared using Pearson's chi-squared test or Fisher's exact test if the expected cell size was < 5. Univariate analysis was performed to identify predictors of poor outcomes among children who underwent surgery or bowel resection. Predictors identi ed as signi cant on univariate analysis (p < 0.05) were included in subsequent multivariate logistic regression analysis. A p-value < 0.05 was considered statistically signi cant.

Comparisons of surgery and non-surgery with intussusception.
Compared with children in the non-surgery group, greater numbers of children in the surgery group were aged < 6 months (p = 0.004); had less abdominal pain at triage (p = 0.006); and had greater frequencies of vomiting (p = 0.005), bloody stool (p = 0.002), and tachypnea (p = 0.017) at triage. Compared with children in the non-surgery group, children in the surgery group had greater frequencies of hyponatremia (p < 0.001), hypochloremia (p < 0.001), and hyperglycemia (p = 0.002). Surgery was associated with ICU admission (p < 0.001) and prolonged hospitalization (p < 0.001). No signi cant differences were observed between groups in terms of sex or the heart rate at triage ( Table 2).
Comparisons of bowel resection and non-bowel resection with intussusception. Table 3 summarizes the demographic and clinical characteristics of patients in the bowel resection and non-bowel resection groups. Compared with children in the non-bowel resection group, greater numbers of children in the bowel resection group were aged < 6 months (p = 0.044) and had cyanosis at triage (p=0.002). The operative ndings revealed that compared with patients who required limited bowel resection, patients who required extensive bowel resection were more likely to have gangrene (p < 0.001), pathological lead point (p < 0.001), and ileoileal disease (p =0.007); they were less likely to have ileocolic disease (p = 0.001). Bowel resection was also associated with ICU admission (p < 0.001) and prolonged hospitalization (p < 0.001).
Factors of outcomes. Table 4   were associated with bowel resection. Bowel resection was not associated with age, sex, vital signs at triage, abdominal pain, vomiting, bloody stool, laboratory ndings, imaging ndings, or intussusception type.

Discussion
In this study, presentation to the PER with intussusception involving bloody stool, hyponatremia, and hypochloremia was associated with a greater risk of need for surgery, suggesting that electrolyte imbalance should be recognized early and treated empirically by means of isotonic uid replacement.
Furthermore, we found that prolonged time to surgery, pathological lead point, and long intussusception were independent predictors of bowel resection in children with intussusception.
Similar to the ndings in previous studies 3,7,14 , our study showed that boys were more commonly affected by intussusception. However, sex did not in uence the risk of surgery or bowel resection. Importantly, we found that presentation with bloody stool was associated with greater risk of surgery for intussusception. A previous study also reported that children with intussusception who present with bloody stool have greater risk of pneumatic reduction failure 15 .
A small proportion of children with intussusception complain of abdominal pain, probably because patients who have more severe disease may not have the energy to cry. Lack of energy in children may be caused by many diseases, including intussusception. Therefore, complete evaluation of lethargic children is essential. Notably, a lack of abdominal pain may lead to delayed recognition of intussusception by caregivers and medical practitioners, leading to delayed treatment. Infants with abdominal pain are often unable to verbally express their pain; therefore, the clinician and triage nurse must make judgments based on observations reported by caregivers. Although it is controversial whether abdominal pain alone should prompt consideration of acute abdominal pathology, our study identi ed abdominal pain as a predictor of surgery. Presence of abdominal pain was associated with successful enema reduction, possibly because it alerted the caregiver and clinicians to the possibility of intussusception, leading to timely diagnosis and treatment 15 . Additional predictors of surgery (i.e., other than abdominal pain) should be considered while deciding the appropriate plan for intussusception management.
Hyponatremia has been proposed as a predictor of intestinal gangrene in pediatric small bowel volvulus 16 , ischemic bowel in patients with small bowel obstruction 17 , and gangrene in acute cholecystitis 18 . The cause of hyponatremia in patients with intussusception is not well understood. During the early period of intestinal obstruction, uid loss into the lumen is evident, but electrolyte imbalance does not occur 19 .
Hypovolemia is a weak stimulus for antidiuretic hormone release; however, the body prioritizes volume over osmolality when there is a substantial decrease in intravascular uid volume, which may enhance antidiuretic hormone secretion 20 . Additionally, uid loss related to vomiting, third spacing, or systemic response related to bowel in ammation may cause hypovolemia. The increased risk of surgery for intussusception involving hyponatremia and hypochloremia requires early correction of the volume de cit caused by gastrointestinal uid loss via vomiting, bloody stool, or third spacing. Hypovolemic hyponatremia caused by gastrointestinal disease is treated by isotonic saline administration and the correction of underlying disease 21 .
Our study results were consistent with the ndings of previous studies in which prolonged time to surgery, long intussusception, and pathological lead point were more common in patients who required bowel resection than in patients who did not 5,8,22,23 . Delayed diagnosis was associated with signi cantly increased morbidity rate, probably because prolonged intussusception causes bowel ischemia, gangrene, perforation, or peritonitis; these are indications for bowel resection 4,10 . Pathological lead point and long intussusception are the major causes of pneumatic or hydrostatic reduction failure and delayed surgical reduction. Our study revealed that ICU admission was an independent predictor of bowel resection. Critically ill patients with bowel gangrene, respiratory failure requiring mechanical ventilation, and multiple organ dysfunction were more likely to undergo ICU admission 24 .
The strength of this study was that it involved a comprehensive analysis of potential predictors of bowel resection in pediatric patients with intussusception. However, this study had some limitations. First, our results were obtained through analysis of a hospital-based registry, which limits the generalizability of the conclusions. Further prospective studies with large sample sizes would be more representative of the general population. Second, this study did not include an exhaustive examination of all potential risk factors for bowel resection; we only assessed the parameters that were routinely documented and thoroughly recorded. Finally, we did not assess the accuracy of triage records because of limitations regarding the retrospective study design. The purpose of this study was to identify predictors of bowel resection among parameters that are commonly recorded in the PER. Prospective data collection in future studies may help to understand the signi cance of these predictors with respect to intussusception outcomes. Further long-term, prospective cohort studies that examine additional potential risk factors are warranted.

Conclusions
In this study of children with intussusception, independent risk factors for surgery comprised hyponatremia, hypochloremia, and bloody stool; lack of abdominal pain likely re ected more severe disease. Children with symptoms for 2 days or longer, long intussusception, pathological lead point, or ICU admission had a signi cantly greater risk of bowel resection. These results may help to design targeted interventions that raise awareness regarding the risk of bowel resection among patients who present to the PER.

Declarations
Ethics approval and consent to participate. The present study was approved by the Chang-Gung Memorial hospital ethics committee/institutional review board (201700560B0) and was exempted from informed consent requirements owing to its retrospective design