Abstract
OBJECTIVE. We attempted to obtain baseline data on the incidence of intussusception and its association with gastroenteritis in a cross-sectional observational study in children.
METHODS. Admissions to all 38 pediatric units in Switzerland because of intussusception were reported to the Swiss Pediatric Surveillance Unit from April 2003 to March 2006. Patient and disease characteristics were assessed prospectively with the use of a standardized questionnaire based on the case definition for intussusception developed by the Brighton Collaboration. Completeness of reporting was verified through capture-recapture analysis.
RESULTS. There were 294 patients with reported intussusception; 35 cases were excluded for various reasons, and 29 additional patients were identified through International Classification of Diseases, 10th Revision, codes. After capture-recapture analysis, we estimated underreporting to the Swiss Pediatric Surveillance Unit to be 32% and we calculated a true number of 381 intussusception episodes. The highest level of diagnostic certainty was reached by 248 patients, and 20 fulfilled level 2 criteria; for the remaining 20 patients, available information was insufficient. The mean age of the patients was 2.7 years. The yearly mean incidence of intussusception was 38, 31, and 26 cases per 100000 live births in the first, second, and third year of life, respectively, with no apparent seasonality. Seventy patients had a history of coinciding gastroenteritis, and 5 of 61 tested positive for rotavirus. Spontaneous devagination was observed for 38 patients; enemas reduced intussusception successfully in 183 cases, whereas surgical treatment was required in 67. All patients recovered without sequelae.
CONCLUSIONS. This is the first prospective nationwide surveillance of intussusception in childhood using a standardized case definition. Most cases occurred beyond infancy, and association with rotavirus gastroenteritis was rare.
Intussusception is defined as the invagination of a proximal segment of intestine into a distal segment of intestine, usually ileocolic.1 This results in obstruction of bowel passage, constriction of the mesentery, and obstruction of the venous blood flow, which is characterized by sudden onset of colicky abdominal pain. The mean age at diagnosis ranges from 7 to 14 months in studies including different age groups.2–4 Estimates for yearly incidence range from 0.5 to 2.24 cases per 1000 infants or young children.3–9 Intussusception peaks have been observed to coincide with seasonal viral gastroenteritis. One study in France noted a significant peak of intussusception cases in spring,6 and a retrospective hospital chart study in Tanzania found that most of the cases presented from January to March and from July to September.10 However, the great majority of studies did not find any seasonality in the occurrence of intussusception.11–21 Diagnosis is made through ultrasonography and/or radiographically guided enema; the latter is also the treatment of choice unless symptoms have been present for a long time and the child's condition is critical. In that case, and if repositioning via enema fails, surgery is necessary. The cause of intussusception is unknown in most cases.22
In 1999, an orally administered, live attenuated rotavirus vaccine was licensed in the United States and elsewhere (RotaShield; Wyeth-Ayerst, Marietta, PA). Shortly after implementation of mass immunization in the United States, concerns regarding an association of rotavirus vaccine with intussusception were raised (in October 1999); after additional investigation, the vaccine was withdrawn from the market.23 New orally administered, live attenuated rotavirus vaccines have been licensed recently in the United States, Europe, and elsewhere. During prelicensure studies, no evidence for an increased risk for intussusception was found for these new vaccines.16,24 However, local epidemiologic baseline data on intussusception are important for safety monitoring after widespread use of these vaccines.
With this background, the primary goal of our study was to determine the incidence and clinical and epidemiologic characteristics of intussusception in children in Switzerland. A secondary goal was to assess a possible association between rotavirus gastroenteritis and intussusception.
METHODS
Case Definition
Cases with a discharge diagnosis of intussusception were categorized by the central study investigators (Drs Buettcher and Baer) in levels of diagnostic certainty, according to the case definition developed by the Brighton Collaboration Working Group for Intussusception.25 Briefly, level 1 requires the demonstration of invagination of the intestine at surgery and/or through either air- or liquid-contrast enema, the demonstration of an intraabdominal mass on abdominal ultrasound scans, with specific characteristic features and proven to be reduced by hydrostatic enema on postreduction ultrasound scans, and/or the demonstration of invagination of the intestine at autopsy. If level 1 criteria are not fulfilled, then various combinations of major and minor clinical criteria may define cases at levels 2 and 3 (see ref 25 for details).
Concurrent gastroenteritis and concurrent upper respiratory tract infection were defined as the presence of vomiting and ≥1 loose bowel movement and rhinitis, conjunctivitis, pharyngitis, and/or acute otitis media, respectively, at the time of intussusception. In accordance with the Brighton Collaboration case definition,25 typical findings on abdominal radiographs were defined as fluid levels and dilated bowel loops. Intermittent intussusception was defined as intussusception that was present at level 1, 2, or 3 of diagnostic certainty at admission and resolved spontaneously during the hospital stay without any intervention (such as hydrostatic enema) or was no longer present during surgery.
Swiss Pediatric Surveillance Unit
The Swiss Pediatric Surveillance Unit (SPSU) was established in 1995 to assess the epidemiologic features of selected childhood diseases leading to hospitalization. It is operated under the auspices of the Swiss Pediatric Society and the Swiss Federal Office of Public Health. The study period for intussusception surveillance was from April 1, 2003, to March 31, 2006.
This was a cross-sectional observational study. Before the start of the surveillance, all 38 pediatric and associated pediatric surgery units in Switzerland received information on the project, including the study protocol and a sample of the standardized questionnaire (SQ) (available from the authors on request), which included questions on demographic characteristics, clinical signs and symptoms, results of ultrasonography, radiology, and/or surgery, and final outcome, including complications.
At the end of each month, reporting forms were mailed to all units. For each reported case of intussusception, the respective attending pediatrician or surgeon had to complete the SQ and send it to our central study office. On receipt, the SQs were reviewed for completeness by the central study investigators (Drs Buettcher and Baer). In cases with missing or inconclusive data, the respective attending pediatrician or surgeon was contacted immediately via telephone or fax machine, to clarify the issue. After clarification, data were entered into the database.
In Switzerland, children with suspected intussusception are admitted to the hospital, where appropriate management (including radiographic diagnosis and all treatment modalities) can be provided as necessary. The likelihood that cases were managed on an outpatient basis is very low.
Patients Identified According to International Classification of Diseases, 10th Revision (ICD-10) Codes
To evaluate completeness of SPSU reporting and to refine incidence estimates, we performed a capture-recapture analysis after completion of the study. The second independent data source was medical charts of children admitted to 8 of the 38 units during the study period. The 8 units were selected according to their numbers of reported cases, that is, 2 with the highest numbers of reported cases, 2 with median numbers of reports, 2 with the lowest numbers of reports (but >0), and 2 that reported no cases. Cases were identified from the hospitals' databases with ICD-10 codes K56.1 (acute intussusception of the intestine) and K38.8 (acute intussusception of the appendix). All pertinent records of previously unreported cases of intussusception were reviewed by staff members of the respective hospitals and were entered into the SQ by the staff members or by the central investigators (Drs Buettcher and Baer).
Statistical Analyses
Statistical analyses were performed with SPSS 13.0.0 (SPSS Inc, Chicago, IL). Data on the general population during the study period were obtained from the Swiss Federal Office of Statistics.26 Independent proportions were compared by using the Pearson χ2 test. P values of <.05 were considered significant.
Reported cases were categorized into levels of diagnostic certainty, with the Brighton Collaboration case definition for acute intussusception, independently by 2 of us (Drs Buettcher and Baer), followed by calculation of interobserver agreement (κ value) as described previously.27 The capture-recapture analysis was performed according to the Chapman-Wittes adjustment of the Lincoln-Petersen maximal likelihood estimate.28
RESULTS
Surveillance and Capture-Recapture Analysis
A total of 294 cases were reported via SPSU during the 3-year study period. Thirty-five reports needed to be excluded (mainly because of duplicate reporting of patients who were transferred from a local pediatric unit to a nearby larger unit for additional treatment; in those instances, the reports from the unit where the patient was finally treated were used for analysis). With the remaining 259 cases and an additional 29 cases identified through ICD-10 codes in the capture-recapture analysis, a total of 288 cases were available for final analyses (Fig 1). Twenty-seven (9%) of 288 questionnaires were incomplete and needed to be clarified through contact with the responsible physician or surgeon.
Profile of study population. IS indicates intussusception. a Category as defined by the Brighton Collaboration.25
The results of the capture-recapture analysis are presented in Table 1. The completeness of SPSU reporting and ICD-10 code identification was 0.68 and 0.76, respectively (P = .978). Therefore, a 32% rate of underreporting to the SPSU can be assumed, and the estimated true number of cases of intussusception leading to hospitalization during the 3-year study period is 381 rather than 259.
Capture-Recapture Analysis Using SPSU Reporting and ICD-10 Codes as Independent Data Sources
Epidemiologic and Clinical Characteristics
Of the 288 patients, 194 (67%) were male; 8 (3%) were former preterm infants; 39 (13%) had comorbidities, with cystic fibrosis (n = 6) and constipation (n = 5) being most frequent; and 18 (6%) had a history of previous intussusception (13 with 1 previous episode and 5 with >1 previous episode). The mean age of the patients was 2.7 years (median: 1.9 years; interquartile range: 0.9–3.4 years). Infants (ie, <12 months of age) represented the largest single age group, and more cases occurred in the second half, compared with the first half, of the first year of life (Fig 2).
Age distribution of all intussusception cases (n = 288). Inset, Age distribution of intussusception cases in infants (0–12 months; n = 83).
The mean duration of hospitalization was 4.3 days (median: 2.5 days; interquartile range: 1–5 days). At admission, 70 cases (24%) presented with concurrent gastroenteritis; in 5 (8%) of 61 cases in which specific tests were performed, rotavirus was detected in stool samples. The distributions of all cases and cases of intussusception associated with gastroenteritis were similar across seasons (Fig 3), with no apparent peak between December and April, when most cases of rotavirus gastroenteritis occur in Switzerland.29 An additional 45 children (16%) had coinciding illnesses other than gastroenteritis, with upper respiratory tract infections (n = 19) being most frequent.
Seasonal distribution of all intussusception cases (black bars; n = 288) and cases associated with gastroenteritis (gray bars; n = 70), according to month of occurrence (cumulative for 3 study years).
During the study period, there were an average of 72550 live births per year in Switzerland.26 The mean yearly incidence rates of intussusception in the first, second, and third year of life were 38, 31, and 26 cases per 100000, respectively, with some variation from year to year (Fig 4). With the capture-recapture analysis taken into account, however, the true incidence rates can be estimated to be 56, 46, and 38 cases per 100000, respectively.
Incidence (cases per 100000 live births per year) of intussusception according to study year and age group in children up to 3 years of age.
The most frequently observed symptoms of patients were abdominal pain (92%), vomiting (53%), and pallor (45%) (Table 2). “Red currant jelly” stools were seen in 36 cases (13%). Infants (n = 83), being the largest age group, presented mainly with abdominal pain (n = 67; 81%), vomiting (n = 56; 66%), and bloody stools (n = 47; 57%).
Symptoms of Intussusception in Different Age Groups
Of the cases (n = 281) for which the onset of symptoms before admission to the hospital was known (in hours), 146 (52%) had symptoms for <24 hours, 54 (19%) for >24 hours but <48 hours, and 81 (29%) for >48 hours. When we compared patients hospitalized within 24 hours after the onset of illness with those with illness for >24 hours before admission, red currant jelly stools were present in similar proportions (27 of 146 patients, 18%, versus 20 of 135 patients, 15%; P = .41), whereas a palpable abdominal mass was observed more frequently for patients with early presentation (77 of 146 patients, 53%, versus 49 of 135 patients, 36%; P = .006).
Ultrasonography was the most frequently used method for diagnosis, and it confirmed intussusception for 236 (85%) of 278 patients. Intermittent intussusception was found for 31 patients (11%), ultrasound ruled out intussusception for 2 patients (1%), and findings were inconclusive for 9 patients (3%). Abdominal radiographs were performed for 90 patients; 67 (74%) revealed typical findings of intussusception and another 19 (21%) were consistent with intussusception, as were 3 of 4 abdominal computed tomograms. Of the 188 cases with specific information available, 156 (83%) were ileocolic, 31 (16%) were ileoileal, and 1 (0.5%) was colocolic.
Outcomes
Of the 288 patients, 183 (63%) were treated successfully through hydrostatic reduction of intussusception, whereas 67 (23%) required surgical treatment (after unsuccessful conservative treatment for 36 patients, after relapse for 14, and as the primary intervention for 17). Meckel diverticulum (n = 6) and intestinal lymphoma (n = 3) were discovered as obvious causes of intussusception during surgery. Intussusception occurred as a complication of gastrointestinal vasculitis with Henoch-Schönlein purpura in 7 patients; in the great majority of the remaining cases (272 of 288 cases, 94%), the underlying cause of intussusception remained unknown. Thirty-eight patients (13%) recovered with spontaneous devagination early during their hospital stays.
The categorization of intussusception cases according to the level of diagnostic certainty (Brighton Collaboration case definition) is presented in Table 3. The great majority of cases fulfilled the criteria for the highest level of diagnostic certainty, and agreement between the 2 independent observers (Drs Buettcher and Baer) who reviewed and classified the reported cases was 0.96 (κ value). All patients recovered and were discharged without apparent sequelae.
Categorization of Intussusception Cases According to Brighton Collaboration Case Definition
DISCUSSION
We surveyed prospectively the spectrum of clinical characteristics, management, and outcomes of intussusception in children and adolescents in Switzerland over a 3-year period, with the use of standardized criteria defined by the Brighton Collaboration.25 Epidemiologic and clinical characteristics of intussusception were studied previously,2–4,6,30 but none of those studies was performed prospectively and standardized criteria were not used to define cases.
In contrast, our study involved prospective nationwide surveillance of intussusception in children, and we used standardized criteria for data collection. Two recently published studies used a similar approach.11,19 However, those studies included only children <2 years of age, whereas our surveillance included children of any age. An additional unique strength of our study was the attempt to ascertain the completeness of reporting through capture-recapture analysis using 2 data sources. Surprisingly, the ICD-10 code search captured only 76% of cases reported to the SPSU, which is remarkably similar to previous SPSU surveillance of varicella hospitalizations, for which the corresponding figure was 78%.31 In our experience, this is most likely explained by coding mistakes in the hospital databases.
Underreporting is an inherent limitation of any surveillance study. Capture-recapture methods allow estimation of the degree of underreporting (ie, the number of unknown cases) and thus estimation of the true number of cases. Our estimate is likely to be valid, because the probability of cases being recorded in one sample was independent of the probability of cases being recorded in the other, and cases were equally likely to be captured in either sample.
As members of our group (with the use of data from the first year of this study) and others showed previously, the case definition for intussusception developed recently by the Brighton Collaboration is useful and reliable,27 and 86% of reported cases reached the highest level of diagnostic certainty. The triad of abdominal pain, palpable abdominal mass, and red currant jelly stools has been described as the typical presentation of intussusception.5 In our study, abdominal pain, vomiting, and pallor were the most prominent symptoms at presentation, whereas a palpable abdominal mass (38%) and red currant jelly stools (13%) were less common. It could be argued that the latter would be signs of progressed intussusception. However, the presence of red currant jelly stools was not dependent on the duration of illness, and an abdominal mass was more often palpable in patients with early (rather than late) presentation to the hospital.
Spontaneous devagination of intussusception has been reported to occur in 4% to 10% of cases.32,33 We observed this phenomenon in 13% of our cases. Possibly, comparatively early presentation of children with abdominal pain to hospitals and rapid availability of imaging techniques allowed sensitive detection of cases. Early relapse of intussusception after conservative intervention occurred for 13% of patients, and 23% required surgery during the course of their illness, which is in the range of previous observations.3,6,32,34
Only a minority of cases (24%) in this prospective study had associated gastroenteritis and, despite the high rate of rotavirus testing (61 of 70 cases), only 5 cases of rotavirus infection were identified. Rotavirus infection has a distinct seasonal pattern in Switzerland, with yearly peaks between December and April,29 but no such seasonality for intussusception was observed during the 3-year study period. In accordance with previous reports, this indicates a lack of apparent association between these 2 entities.11,16,24,35–37
We found the highest incidence of intussusception in infants (ie, 56 cases per 100000). This is in accordance with previous studies,3,4,8,19,30 but the rate is in the lower part of the range from earlier reports that reported specifically the incidence in infants, from Venezuela (incidence: 24 cases per 100000),5 Chile (incidence: 33 cases per 100000),38 Vietnam (incidence: 302 cases per 100000),11 Australia (incidence: 71–131 cases per 100000),11,18 United Kingdom (incidence: 66–100 cases per 100000),4,9 Israel (incidence: 224 cases per 100000),3 and Singapore (incidence: 60 cases per 100000).19 It is also low in the range (incidence: 50–230 cases per 100000) from studies that reported overall incidence but not rates for specific age groups.5–8 Although most previous studies found the majority of all cases in infants,3–5,8,19,30 only 29% of our cases belonged to that age group, and incidences in the second and third years of life were similar to those in infants. The fact that we did not restrict our surveillance to young children probably explains this difference. Consequently, the mean age (2.7 years) and median age (1.9 years) of our patients were considerably higher than those in other reports (mean age: 7–14 months).2–4 Interestingly, a gradual and unexplained increase in the mean age of patients with intussusception has been observed in Singapore,19 Taiwan,5 Denmark8 and the United States13 in the recent past; this epidemiologic change might have had occurred in Switzerland earlier, but there are no data available for comparison.
CONCLUSIONS
This study provides the first estimate of the incidence of intussusception in children in Switzerland obtained by using a standardized case definition and data generated through prospective surveillance. The data were obtained throughout the nation and should be useful for postmarketing safety surveillance if orally administered rotavirus vaccines are used widely in the near future.
Given the comparatively low rate of intussusception associated with RotaShield in the past (15 cases of intussusception after administration of ∼1.5 million doses, of which 13 cases occurred after the first dose39) and the small number of live births in Switzerland per year (∼72000), the likelihood of determining an association of intussusception with any new rotavirus vaccine will be limited. However, standardized assessment of intussusception episodes is feasible, and internationally standardized postlicensure studies would allow data comparability beyond any single country's pertinent population.
Acknowledgments
An unrestricted research grant was provided by Glaxo SmithKline.
We thank Dr Terhi Tapiainen for assisting in project coordination during the first study year; Daniela Beeli, SPSU secretary, for coordinating the report forms; Esther Schilling, study secretary at the University Children's Hospital Basel, for keeping track of the reports; the SPSU representatives of the respective pediatric and pediatric surgical units in Switzerland (C. Aebi, E. Antonelli, W. Baer, B. Berclaz, M. Bianchetti, M. Bittel, A. Blumberg, H.-U. Bucher, L. Buetti, E. Bussmann, O. Carrel, P. O. Cattin, A. Corboz, G. Délèze, P. Diebold, P. Dolivo, F. Farron, M. Gehri, T. Gehrke, C.A. Haenggeli, P. S. Hüppi, P. Imahorn, C. Kind, W. Kistler, B. Knöpfli, O. Lapaire, B. Laubscher, U. Lips, N. Lutz, A. Malzacher, J. McDougall, J.-L. Micheli, J. Mayr, M. Mönkhoff, V. Pezzoli, B. M. Regazzoni, L. Reinhard, F. Renevey, P. Rimensberger, H. Roten, C. Rudin, J. W. Salomon, V. Schlumbom, M. Schwöbel, G. Staubli, C. Stüssi, R. Tabin, B. Wildhaber, J. Wisser, M. Wopman, Z. Zachariou, U. G. Zeilinger, A. Zemmouri, and U. Zimmermann) for providing reports; and all of the dedicated physicians for taking care of the patients and helping to complete the questionnaires.
Footnotes
- Accepted April 3, 2007.
- Address correspondence to Ulrich Heininger, MD, Division of Pediatric Infectious Diseases, University Children's Hospital Basel, PO Box 4005, Basel, Switzerland. E-mail: ulrich.heininger{at}ukbb.ch
Drs Buettcher and Baer contributed equally to this study.
The authors have indicated they have no financial relationships relevant to this article to disclose.
REFERENCES
- Copyright © 2007 by the American Academy of Pediatrics
In this issue
Jump to section
Related Articles
Cited By...
- Characteristics of Hospitalized Children With SARS-CoV-2 in the New York City Metropolitan Area
- Movement Disorder Associated With Foreign Body Ingestion
- A young woman with a jejuno-jejunal intussusception
- Nosocomial Rotavirus Infections: A Meta-analysis
- Risk Stratification of Children Being Evaluated for Intussusception
- The Crying Infant: Diagnostic Testing and Frequency of Serious Underlying Disease
- Importance of On-time RotaTeq Vaccination and Long-term Active Surveillance