Intussusception (IS) is one of the most common paediatric intestinal emergencies in infants and toddlers (1). It is defined as a process in which a segment of intestine telescopes into the adjoining intestinal lumen eventually leading to perforation of the intestine if left untreated. The pathogenesis of IS is thought to be secondary to an imbalance of longitudinal forces along the intestinal wall with a mass, for example, lymph nodes, acting as lead in combination with a disorganised pattern of peristalsis. An important aetiologic factor in the development of IS seems to be viral as well as bacterial gastroenteritis even though only retrospective analyses currently support the present view (2).
Management is considered to be primarily conservative by nonsurgical reduction, whereas the mode of reduction remains highly controversial (3). During the last 3 decades there have been numerous studies published on the optimal management of IS—favouring either pneumatic (4–7) or hydrostatic reduction (8–10)—as well as incidence estimation with a wide range from 50/100,000 child-years reported for children younger than 1 year to as high as 224/100,000 (3,11). Most of these are retrospective or—if not—single-centre experiences and all of these studies have in common that they experience a lack of strict adherence to a consistent case definition, making it difficult, if not impossible, to compare individual results. It was not until 2004 that standardised case definition criteria were defined by the Brighton Collaboration (BC) group (12). In addition, except for 1 recent small study performed in Switzerland (13), all of the published studies aiming to determine IS incidence used a retrospective, mainly International Classification of Diseases (ICD) code–based design, which may significantly overestimate the incidence of IS as recently demonstrated by Kohl et al (14,15).
The interest in an exact incidence estimation based on standardised case definition criteria has been renewed due to the recent introduction of 2 new-generation rotavirus (RV) vaccines, RotaTeq (Sanofi Pasteur MSD, Lyon, France) and Rotarix (GlaxoSmithKline, London, UK), because the first RV vaccine RotaShield was withdrawn in 1998 due to a possible association with the development of IS (16,17). Even though 2 large-scale safety vaccine trials did not show any association with IS (18,19), for the new vaccines exact and reliable epidemiological data are necessary as a prerequisite for further postlicensure monitoring.
With the present study we aimed to provide clear epidemiological data for paediatric IS and to identify factors possibly associated with a higher risk for surgery. Furthermore, we aimed to compare the various techniques used for nonsurgical reduction and their potential impact on outcome.
The study was designed as a cross-sectional, prospective, hospital-based, multicentre surveillance study and conducted from January 2006 to December 2007. The data collection was performed by the German Paediatric Surveillance Unit (ESPED) founded in 1992, which includes all of the German paediatric hospitals. These are contacted on a monthly basis to report on certain diseases under investigation. Positive returns received by the unit's scientific coordinator at the ESPED headquarters are then forwarded to the appropriate investigator, who then contacts the reporting clinician directly to request completion of a data collection form (Fig. 1) consisting of specific questions concerning clearly defined epidemiological and clinical data.
For the present study, 37 German tertiary paediatric units were excluded because of their high level of specialisation—for example, paediatric cardiac surgery units, chronic care facilities, and rehabilitation centres; no reports of IS cases could be expected. Of the remaining 336 German paediatric hospitals and independent paediatric surgery units (n = 26), all of them agreed to participate in the present study. Therefore, in theory, assuming an underreporting rate of 0%, all children admitted in Germany for IS during the study period would have been included in our study.
Inclusion and Exclusion Criteria
Patients were included when the following eligibility criteria were met:
1. Younger than 15 years old at the time of enrolment into the study
2. Fulfillment of the BC criteria for definite IS (12)
3. IS during the study period
Exclusion criteria were as follows:
1. Diagnosis based on clinical assessment only
2. Undetermined means for diagnosis
Case definition followed the BC criteria (12). The assessment of IS cases was initially performed by the attending physician on presentation, who then received a detailed questionnaire from the study coordinator. Based on the present questionnaire, all of the cases were reviewed and confirmed by the principal study coordinators (S.W. and A.J.). If necessary, additional information was obtained in a face-to-face meeting. Datasets of reported cases were categorised according to the BC criteria into levels of diagnostic certainty by 2 independent observers (S.W. and A.J.), followed by calculation of interobserver agreement (kappa value) as described previously (20). Only definite IS cases, level 1 according to BC criteria, were subjected to further analysis.
Estimation of True Incidence
Because data collection by surveillance bears the risk of considerable underreporting, a random sample of participating clinics were contacted a second time to determine the fraction of unreported cases by a thorough review.
Briefly, in a random sample of 31 clinics with the same distribution of hospital characteristics regarding the number of beds, and associated paediatric radiology and surgery units as in the primary study group, all of the IS cases were collected separately by searching the ICD code in the clinic register and the medical record. ICD-10 codes K56.1 and K38.8 as well as operation and procedure codes for nonsurgical reduction 8.122.x were used to identify cases. All of the medical records of patients with IS were then reviewed by staff members of the respective hospitals in cooperation with the principal investigators (S.W. and A.J.). Based on these data we calculated the true IS incidence, taking into account the empirical distribution of the reporting quote and using the maximum likelihood estimation as described previously by Trampisch et al (21).
To ensure data quality, double entry of the data was performed. For estimation of the true IS incidence a maximum likelihood approach was used, and associating confidence intervals were constructed by bootstrap methods as described by Trampisch et al (21). Briefly, we determined the population under risk, that is, the number of children of the corresponding age group within the catchment area of a given hospital using 5 different calculations to model the probability that a specific hospital is visited in cases of IS. We then performed a maximum likelihood estimation of the global incidence, taking into account the population within the catchment area of the hospital for a fixed population model, the number of all of the cases within the catchment area of the hospital, the number of reported cases of the hospital, and the reporting quota of the hospital. Confidence bounds were obtained with a bootstrap approach.
The univariate distribution of categorical variables is specified by absolute frequencies and percentages. The quantitative variables are highly skewed and therefore non-normally distributed, and thus described by median and interquartile range (IQR). To rule out non-Gaussian distribution we used the Kolmogorov-Smirnov test. To assess the effect of possible influencing factors on the success of conservative management a multiple logistic regression analysis was performed. P < 0.05 was considered significant. Cases with missing values were excluded from analysis. As a consequence there are differences in case numbers depending on the analysed variable, which is specifically mentioned in the corresponding tables and figures. Statistical analyses were performed using SAS version 9.2 (SAS Institute, Cary, NC).
Description of the Study Setting and Population
Of the 336 German hospitals initially included in the present study, 319 (95%) reported to the ESPED headquarters on a regular monthly basis. Two hundred thirty-four of all departments reported cases of IS (69.6%). Two hundred twelve of the 234 departments returned at least 1 completed follow-up questionnaire for at least 1 definite IS case (90.6%). From January 1, 2006 to December 31, 2007, 1593 cases of IS were reported. For 1353 reported IS cases (84.9%) the follow-up questionnaires were returned (ie, cases referred to as complete in the present report). A total of 143 cases were identified as either duplicates or misreports. Additionally, 5 patients older than 15 years as well as 5 cases classified as possible IS according to the BC criteria were excluded from final analyses. Overall, 1200 children with at least 1 definite episode of IS remained (Fig. 2). All of the analyses were performed in these 1200 definite IS cases.
Incidence and Epidemiological Characteristics
Of the 1200 patients with definite IS, 787 were boys and 413 were girls; the male-to-female ratio was 1.9:1. The median age was 21.1 months (IQR 10–38 months). The age of the patients ranges from the first month of life to 14 years 9 months. Six hundred forty-seven (53.9%) patients were younger than 2 years of age, 367 (30.6%) between 2 and 4 years, and 186 (15.5%) 4 years and older. The median duration of hospitalisation for IS was 3 days (IQR 2–5 days).
Table 1 shows the results and detailed data of the incidence calculations in different age groups using the population data for 2006 provided by the German Federal Office of Statistics (22). The estimated crude incidence (based on the 1200 definite completed cases) for children younger than 2 years was 23.8/100,000 childyears, whereas the total incidence for children younger than 15 years was 5.2/100,000 child-years. Only minor seasonal fluctuations were observed without statistical significance. Using the separately collected IS cases in the random sample of 31 clinics as a second source, we calculated the proportion of underreporting to be about 50% with only slight variations in each age group and estimated the true incidence to be 10.4 [95% CI 8.6–12.1] for children under 15 years of age and 51.5 [41.7–61.1] for children under 2 years of age.
Clinical Presentation and Outcome
In 1193 cases (99.4%) clinical symptoms on admission were documented, including abdominal resistance/distension, vomiting, raspberry defecation, pallor, and the like. The most common presentation was abdominal pain with 1 (24.1%) or more additional symptoms (61.0%)—vomiting, abdominal pain, and pallor being among the most frequently occurring (Fig. 3). The proportion of patients presenting with abdominal pain exclusively was 13.9%, whereas presentation with the classical “Ombrédanne” triad—consisting of colicky abdominal pain, abdominal mass, and bloody tinged defecation—was only found in 5.1% of cases, as shown in Figure 3.
On admission 209 patients (17.4%) presented with concurrent acute gastroenteritis, 89 (7.4%) had signs of acute respiratory tract infections, and 41 (3.4%) received antibiotics 7 days before admission. In addition, 16 patients (1.3%) were diagnosed as having Henoch-Schonlein purpura, and in 10 cases (0.8%), IS was due to acute lymphoma. Exact data on treatment were available in 1175 cases with an overall need for surgical reduction of 24.6% (n = 289), whereas the ratio was higher in patients presenting with acute gastroenteritis (31.5%, n = 66) and Henoch-Schonlein purpura (62.5%, n = 10). In 50 cases, surgery was performed due to suspected intestinal malformation. Twenty-three (46%) of these patients were intraoperatively found to have Meckel diverticulum, 10 (20%) showed other pathologies (intestinal duplication, lymphoma, polyps), and in 17 (34%) no intestinal pathology could be found.
The outcome was known in 1189 of the cases. 96.6% of the patients (n = 1149) were discharged without complications; for the remaining 39 cases (3.3%), minor complications were noted, for example, prolonged hospital stay after surgery or wound infection. One patient, a previously healthy 32-month-old boy who presented after a 72-hour history of vomiting and abdominal pain with cardiorespiratory failure due to severe shock died due to IS (0.08%).
In our study, 86.2% of the children received at least 1 attempt of nonsurgical reduction with exact information about the mode of reduction obtained in 98.2% of these children (n = 995). The most common mode of reduction was hydrostatic reduction (n = 546, 54.9%), followed by radiographically guided barium enema (n = 268, 26.9%), and pneumatic reduction under x-ray control (n = 109, 11.0%). Success rates were 89% for pneumatic reduction, 72.8% for reduction with barium enema, and 80.4% for hydrostatic reduction using 0.9% sodium chloride, whereas successful conservative management was defined as no surgery within the hospital stay.
We then performed a multivariate analysis of several parameters including specialisation of the clinic, number of cases per year, age of the patient, symptoms at presentation as well as time interval between onset of symptoms and first attempt of reduction, and mode of reduction. Cases with missing variables (n = 151) and cases in which multiple or other agents for reduction were used (n = 72) were excluded from the present analysis.
Patients older than 4 years and children presenting with blood on rectal examination had an increased risk for secondary surgery (OR 2.6, CI 1.5–4.7). The success of nonsurgical management also significantly depended on the mode of reduction. Compared with pneumatic reduction, all of the other methods showed an increased risk for secondary surgery—3.7-fold (1.6–8.8) with barium enema and 2.8-fold (1.2–6.4) with hydrostatic reduction using 0.9% sodium chloride solution. The results are summarised in Table 2.
With regard to the level of specialisation of the hospital we did not find any influence on the success rate of nonsurgical reduction, whereas a time interval between onset of symptoms and first attempt of reduction of more than 5 hours, which was the case in 71.1% of the cases, was associated with a 2-fold (1.2–3.1) increase in the risk of secondary surgery (Table 2). According to this, children with IS who initially presented within 5 hours after onset of symptoms but were then transferred for nonsurgical reduction from a smaller paediatric hospital to a tertiary care facility had a significantly higher risk for secondary surgery.
To test whether the risk for failure of nonsurgical reduction may be predicted by the severity of clinical symptoms at presentation, all of the patients were divided into 5 groups according to symptom severity using a semiobjective approach based on the assessment by 3 independent paediatric gastroenterologists who did not have any information about the outcome. Retrospectively, stratification parameters into the 5 groups were as follows: slight—abdominal pain only; mild—abdominal pain in association with at least 1 more unspecific clinical sign or finding; moderate—abdominal pain in association with more serious clinical signs such as paleness or crying; severe—any clinical signs and findings in combination with bloody stool; critical—if presenting with clinical signs of shock. Cases with different categorisations were excluded from the present analysis. As can be seen in Figure 4, nonsurgical reduction was only unsuccessful in about 10% in patients classified as presenting with minor symptoms. This number rose significantly to 26.5% in children considered severely compromised, but changed dramatically to 50% if patients were classified as critically ill (Fig. 4). In this subset of patients a partial resection of bowel was necessary in 20% of cases.
RV Vaccination in the Study Population
In Germany, RV vaccines were licensed only in October 2006 and are currently not officially recommended. Therefore, the number of RV-vaccinated patients was expected to be low. Between January and December 2007, 5 cases of IS following RV vaccination were reported through the ESPED surveillance. Remarkably, 3 of these 5 patients (60%) were vaccinated much later than recommended by the regulatory authorities, that is, beyond 6 months of age. In addition, 4 of 5 patients (80%) received surgical treatment for their IS.
IS cases following RV vaccination are also recorded by the German Adverse Event Reporting System (AERS), provided by the Paul Ehrlich Institute on a voluntary basis using an open-source software tool comparable to the systems commonly used by the US National Institutes of Health. Between January and December 2007, 1 case of IS was reported to the Paul Ehrlich Institute, which was also reported to the ESPED surveillance. During the same period 5 cases were reported to the ESPED surveillance. Based on these data the reporting sensitivity of the national German AERS for IS after RV vaccination can be estimated to be 20% at best compared with the ESPED surveillance. Considering the underreporting in the ESPED surveillance, this number even declines to about 10%.
We performed a nationwide prospective surveillance study using the BC case definition criteria for IS collecting data on epidemiological and clinical characteristics as well as management and outcome of IS. Because underreporting is an inherent limitation of any surveillance study a separate second smaller cohort was obtained for its assessment. The combination of this capture–recapture analysis, the consideration of the empirical distribution of the reporting quote, and the use of a maximum likelihood estimation (12) for the calculation of the underreporting makes our estimates most likely to be valid. The derived incidence for IS was calculated to be 60.4/100,000 in children younger than 1 year, which is in accordance with the recently published incidence of 56/100,000 in children younger than 1 year seen in the small Swiss surveillance study by Buettcher et al (13). It represents the first reliable incidence estimation obtained in a large prospective nationwide surveillance suitable for further postmarketing monitoring of IS after general introduction of RV vaccines.
We also investigated other factors possibly associated with IS. Similar to the Swiss surveillance study, we found only a few patients presenting with associated gastroenteritis (17.4%) compared to previous reports (23). The significance of this finding, however, is clearly limited because most of the German hospitals do not analyse stool samples for pathogens routinely and we also did not specifically ask to document mesenterial lymphadenitis. In 1.3% of the cases IS occurred as complication of Henoch-Schonlein purpura. Interestingly 62.5% of those required surgical reduction compared with an overall rate of 24.6%.
Another factor just recently suggested to be associated with an increased rate of IS is antibiotic therapy (24,25). In our study population only a minority of patients (3.4%) received antibiotics in a 7-day interval before admission for IS as compared with the 25% in the study of Spiro et al (25). An explanation for this divergence may be that both retrospective studies used an ICD-based code approach to identify cases of IS and did not adhere to a strict case definition, which clearly has the potential to confound results. In addition, the study presented by Bines et al (23) was performed in the United States in a population with a much higher antibiotic use and more common bacterial resistance compared with European countries; in the Danish study by Hviid et al (24) the IS incidence of 36.8/100,000 child-years in children younger than 2 years was surprisingly low. Therefore, even though our study was not designed to identify a possible association of antibiotic therapy with IS and did not include a control group, we believe that our results demonstrate the need for further prospective studies concerning the present topic.
We further analysed factors associated with the need for surgical reduction providing the first data from a prospective multicentre study design comparing the various techniques used for nonsurgical reduction. Most important, we found that pneumatic reduction was significantly more effective than all of the other commonly used methods. The mean observed success rate was almost 90%, which is slightly higher than the most recently reported single-centre experiences by Lui (26) or Rubi et al (27). One possible explanation for this finding may be that air is more slowly reabsorbed by the intestinal mucosa and does not stimulate intestinal motility as compared with normal saline or water, thus acting like a splint immobilising the colon and terminal ileum. Based on the present observation, we calculated that a general introduction of pneumatic reduction as a first-line treatment has the potential of preventing 104 (CI 46–161) surgeries each year in Germany, namely in a population of very young patients often highly susceptible for short- and long-term complications. As a second factor the time interval between onset of symptoms and first attempt of reduction proved to be critical for the success of conservative management, whereas the grade of specialisation had no impact provided it is a paediatric hospital, as shown by Jen et al (28). As a consequence, patients suspected to have IS should be admitted to the next available paediatric department which is accessible in all German regions within 1 hour and transfer of patients should be avoided. In addition, consistent with previous reports we observed an increased risk for secondary surgery in older children and if they presented with blood on rectal examination (29–31).
We only observed 5 cases of IS in timely association with RV vaccination, which does not allow any conclusion. However, we noticed that during the same time interval only 1 case was reported to the German national AERS. Even though this is a comparison of 2 different approaches, which may limit the value of its result, the reporting sensitivity of the AERS seems to be only 20%, not taking into account the underreporting rate in our study. This is not surprising because several previous reports showed that reporting sensitivities depend on a number of factors, such as clinical seriousness, temporal proximity to vaccination, and the like. Rosenthal et al demonstrated in 1995 that reporting sensitivities ranged from 72% for poliomyelitis after the oral poliovirus vaccine to <1% for rash and thrombocytopenia after the MMR vaccine (32). In 2001 Verstraeten et al performed a capture–recapture analysis of IS cases which occurred in temporal association with RRV-TV vaccination between December 1998 and June 1999 (33). They calculated sensitivity of the AERS of the FDA at that time to be 47% for IS as an adverse event after RV vaccination. This must be taken into consideration for postmarketing monitoring of the RV vaccines.
In summary, the present study is the largest prospective study on IS in children using standardised case definitions. It provides comparative data on optimal management obtained in a multicentre study design analysing for the first time factors essential for optimal management. It defines a baseline IS incidence that can be considered to be rather representative for European countries. This knowledge is a prerequisite for postlicensure surveillance studies after general introduction of the new RV vaccines.
The authors thank Carin Cohen and Claire Newbern from GSK Bio, as well as Géraldine Dominiak–Felden and Tarik Derrough from Sanofi Pasteur MSD.
1. Blanch AJ, Perel SB, Acworth JP. Paediatric intussusception: epidemiology and outcome. Emerg Med Australas 2007; 19:45–50.
2. Nylund CM, Denson LA, Noel JM. Bacterial enteritis as a risk factor for childhood intussusception: a retrospective cohort study. J Pediatr 2010; 156:761–765.
3. Huppertz HI, Soriano-Gabarró M, Grimprel E, et al
. Intussusception among young children in Europe. Pediatr Infect Dis J 2006; 25:S22–S29.
4. Wang G, Liu XG, Zitsman JL. Nonfluoroscopic reduction of intussusception by air enema. World J Surg 1995; 19:435–438.
5. Sanz N, Sanchez M, Garcia AJ, et al
. Intussusception: barium vs. pneumatic reduction. Cir Pediatr 1996; 9:21–24.
6. Shiels WE, Maves CK, Hedlund GL, et al
. Air enema for diagnosis and reduction of intussusception: clinical experience and pressure correlates. Radiology 1991; 181:169–172.
7. McDermott VG, Taylor T, Mackenzie S, et al
. Pneumatic reduction of intussusception: clinical experience and factors affecting outcome. Clin Radiol 1994; 49:30–34.
8. Wang GD, Liu SJ. Enema reduction of intussusception by hydrostatic pressure under ultrasound guidance: a report of 377 cases. J Pediatr Surg 1988; 23:814–818.
9. Crystal P, Hertzanu Y, Farber B, et al
. Sonographically guided hydrostatic reduction of intussusception in children. J Clin Ultrasound 2002; 30:343–348.
10. Soun DS, Choy PC, Tok TS, et al
. Reduction of intussusception by ultrasound-guided saline enema. Zhonghua Min Guo Xiao Er Ke Yi Xue Hui Za Zhi 1996; 37:257–262.
11. Bines J, Ivanoff B. Acute Intussusception in Infants and Children: A Global Perspective. Geneva: World Health Organization; 2002.
12. Bines JE, Ivanoff B, Justice F, et al
. Clinical case definition for the diagnosis of acute intussusception. J Pediatr Gastroenterol Nutr 2004; 39:511–518.
13. Buettcher M, Baer G, Bonhoeffer J, et al
. Three-year surveillance of intussusception in children in Switzerland. Pediatrics 2007; 120:473–480.
14. Kohl LJ, Streng A, Grote V, et al. Intussusception-associated hospitalisations in Southern Germany. Eur J Pediatr
15. Cortese MM, Staat MA, Weinberg GA, et al
. Underestimates of intussusception rates among US infants based on inpatient discharge data: implications for monitoring the safety of rotavirus vaccines. J Infect Dis 2009; 200:S264–S270.
16. Murphy TV, Gargiullo PM, Massoudi MS, et al
. Intussusception among infants given an oral rotavirus vaccine. N Engl J Med 2001; 344:564–572.
17. Bines JE, Kohl KS, Forster J, et al
. Acute intussusception in infants and children as an adverse event following immunization: case definition and guidelines of data collection, analysis, and presentation. Vaccine 2004; 22:569–574.
18. Ruiz-Palacios GM, Perez-Schael I, Velazquez FR, et al
. Safety and efficacy of an attenuated vaccine against severe rotavirus gastroenteritis. N Engl J Med 2006; 354:11–22.
19. Vesikari T, Matson DO, Dennehy P, et al
. Safety and efficacy of a pentavalent human-bovine (WC3) reassortant rotavirus vaccine. N Engl J Med 2006; 354:23–33.
20. Tapiainen T, Bar G, Bonhoeffer J, et al
. Evaluation of the Brighton Collaboration case definition of acute intussusception during active surveillance. Vaccine 2006; 24:1483–1487.
21. Trampisch H, Trampisch M, Klaassen-Mielke R, et al. Hospital-based, prospective, multicentre surveillance to determine the incidence of intussusception in children aged <15 years in Germany. Paper presented at: 30th Annual ISCB 2009. Abstract 134.
22. German Federal Office for Statistics. Genesis-Tabelle Bevölkerungsstand nach Geschlecht und Altersjahren in Kreisen und kreisfreien Städten Stichtag 31.12.2006.
23. Bines JE, Liem NT, Justice FA, et al
. Risk factors for intussusception in infants in Vietnam and Australia: adenovirus implicated, but not rotavirus. J Pediatr 2006; 149:452–460.
24. Hviid A, Svanstrom H. Antibiotic use and intussusception in early childhood. J Antimicrob Chemother 2009; 64:642–648.
25. Spiro DM, Arnold DH, Barbone F. Association between antibiotic use and primary idiopathic intussusception. Arch Pediatr Adolesc Med 2003; 157:54–59.
26. Lui KW, Wong HF, Cheung YC, et al
. Air enema for diagnosis and reduction of intussusception in children: clinical experience and fluoroscopy time correlation. J Pediatr Surg 2001; 36:479–481.
27. Rubi I, Vera R, Rubi SC, et al
. Air reduction of intussusception. Eur J Pediatr Surg 2002; 12:387–390.
28. Jen HC, Shew SB. The impact of hospital type and experience on the operative utilization in pediatric intussusception: a nationwide study. J Pediatr Surg 2009; 44:241–246.
29. Barr LL, Stansberry SD, Swischuk LE. Significance of age, duration, obstruction and the dissection sign in intussusception. Pediatr Radiol 1990; 20:454–456.
30. Stephenson CA, Seibert JJ, Strain JD, et al
. Intussusception: clinical and radiographic factors influencing reducibility. Pediatr Radiol 1989; 20:57–60.
31. Katz M, Phelan E, Carlin JB, et al
. Gas enema for the reduction of intussusception: relationship between clinical signs and symptoms and outcome. AJR Am J Roentgenol 1993; 160:363–366.
32. Rosenthal S, Chen R. The reporting sensitivities of two passive surveillance systems for vaccine adverse events. Am J Public Health 1995; 85:1706–1709.
33. Verstraeten T, Baughman AL, Cadwell B, et al
. Enhancing vaccine safety surveillance: a capture–recapture analysis of intussusception after rotavirus vaccination. Am J Epidemiol 2001; 154:1006–1012.