Intussusception hospitalizations incidence in the pediatric population in Italy: a nationwide cross-sectional study

The study was conducted on the Italian pediatric population aged 0 to 15 years; cases were identified from hospitalization records. Only children with an intussusception discharge diagnosis in the study period were included. Two different sub-cohorts were identified for the main analyses, i.e. infants aged 0 to < 1 years and children aged ≥ 1 year of age.

In Italy the NHS is provided universally free. A hospital database collects all hospital discharge records (including day-hospital/day-surgery admission). Only routinely collected information was used in the study and data were analyzed through a unique, anonymised, personal identifier. The following data were extracted for the present analyses from each record: age, gender, region of residence, admission and discharge dates, diagnosis (primary and secondary) and procedures [coded according to the International Classification of Diseases (ICD), 9^thRevision],status at discharge (deceased/non-deceased).

Records with the following intussusceptionICD-9 codes, either as primary or secondary diagnosis, were selected from the national database hospital discharge records: 560.0 (intussusception) and 543.9 (other and unspecified appendix diseases). The analysis was restricted to incident cases identified from 1 January 2002 to 31 December 2012; thus, only the first intussusception-associated hospitalizations (considered as index dates) were included.

To identify potential risk factors for intussusception, all hospitalization records in the 6-month period preceding the index date (the first intussusception diagnosis) for each identified case were retrieved. Hospitalization records in 2001 were used to collect the medical history of intussusception case incidents in 2002. A predefined list for the leading points pathology was used with the related ICD-9 codes (see Additional file 1: Annex 1).

As a proxy for the case severity, ICD-9 procedure codes were used to identify those intussusception hospitalizations requiring surgical or radiological intervention. Surgical intervention was defined by procedure codes 46.80 to 46.82, and radiological intervention by codes 96.29 and 96.39. All the procedure codes were coupled to an intussusception code, namely in the same hospital discharge form of the intussusception incident case. Three categories were defined: i) intussusception (identified by diagnosis); ii) intussusception requiring surgery (identified by diagnosis and surgical procedures); iii) intussusception requiring non-surgical intervention (identified by diagnosis and non-surgical procedures). The in-hospital mortality was also evaluated (number of children who died in-hospital following the first intussusception episode).

Consecutive intussusception hospitalizations for the same child were identified for up to 1 year following the first episode of intussusception (incident case). The same ICD-9 codes applied to identify incident cases were used to retrieve recurrence of intussusception episodes. Three categories of recurrence were defined: i) early recurrence: cases with hospital readmission ≤7 days from the first episode; ii) medium term recurrence: cases with hospital readmission between 8 and 30 days from the discharge; iii) late recurrence: cases with hospital readmission after 30 days from the discharge and ≤1 year.

The annual intussusception incidence per 100,000 infants/children was calculated using as a reference the Italian population resident figure provided by the National Institute for Statistics, for the period 2002 to 2012 [23]. Births were assumed to be evenly distributed throughout the year, when hospitalization rates by age (in months) were calculated. Intussusception period specific incidence rates associated with hospitalization were estimated. Annual incidence rates were calculated after being adjusted by age and gender, through direct standardization methods, using the 2012 population as a reference to take into account the ongoing demographic variation.

The incidence rate ratios (IRRs), and the 95 % Confidence Intervals (CI) were calculated through the Poisson regression. Statistical analyses were performed using STATA software (version 11.2; Stata Corp LP, College Station, TX, USA) and IBM SPSS Statistics for Windows (Version 22.0; Armonk, NY: IBM Corp.).

The description of the characteristics of the overall pediatric population enrolled is shown in Table 1. The intussusception incidence rate was 21 per 100,000 children and was higher among boys than girls (23 per 100,000 vs 19 per 100,000, respectively); thus boys had a statistically significant increased probability to experience intussusception hospitalization (IRR: 1.16; 95 % CI 1.13–1.20). Rates of intussusceptions also varied substantially by age, with the highest incidence rate occurring in infants <1 year of age (39 per 100,000 infants). However, rates were low for infants <14 weeks (19 per 100,000 infants in the group aged 6 to 14 weeks), then increased rapidly, peaking at 60 per 100,000 for infants aged 25 to 32 weeks, then decreasing for children ≤6 years of age; a rise in the hospitalization incidence rates occurred also within age groups 10 to 12 years (26 per 100,000 children). The probability of intussusception hospitalization became statistically significant higher from 15 weeks of age to ≤52 weeks (the risk was quadrupled for infants aged 25 to 32 weeks, IRR: 3.94, 95 % CI 3.15–4.93.). The incidence rates measured in the overall pediatric population varied by geographical areas, showing a marked increase in rates in the south and central Italy, with respect to the north of Italy (IRR south: 2.65; 95 % CI 2.56–2.74; IRR central: 1.61; 95 % CI 1.54–1.68).

In the vast majority of cases (87.7 %) non-surgical, or -radiological procedures were reported, shown in Table 1.

Overall, 8 children hospitalized for intussusception died in-hospital (3 in the first year of life), with an in-hospital mortality rate of 0.39 per 1000, in the study population. The incidence rate of in-hospital mortality following intussusception in infants (<1 year of age) was 0.5 per million infants (Additional file 1: Figure S1).

About 4 % of the children hospitalized with an intussusception diagnosis had a recurrence within 1 year after the first episode (Table 1). The 41.6 % of children experiencing a recurrence were hospitalized ≤30 days from the intussusception incident episode (Additional file 1: Figure S2).

Only 291 (1.4 %) children showed at least one known intussusception risk factor in the 6-month period prior to the incident episode (Table 2). The most frequent risk factors were presented in Table 2 and included gastroenteritis (44.3 % of the cases) followed by inflamed appendix, and Henoch-Schonlein purpura (35.1 % and 9.6 %, respectively).

The annual incidence rate in children <1 year of age was highest in 2002 and decreased steadily to 2004 falling from 50.2 to 39.2 per 100,000 infants, rates then remained stable from 2005 to 2012, ranging from 33.0 to 40.1 per 100,000 infants (Fig. 2a). Stable temporal trends were observed when considering incidence rates by gender, with constantly higher incidence rates in boys in each year.

Similar temporal trends were observed when the analysis was conducted among children aged 1–15 years (Fig. 2b), although the incidence was lowest in this population (ranging from 17.6 to 22.6 per 100,000 children in the considered period).

The annual trends of incidence rates by geographical areas remained overall stable (Figs. 3a,b). Specifically, when considering the infants cohort, the highest incidence rates were observed in each year in the central Italy: decreasing from 66.6 per 100,000 infants in 2002 to 58.5 in 2012. In the central Italy, the highest incidence rate was observed in Tuscany and Umbria; region specific incident rates in infants were showed in Additional file 1: Figure S3a.

With regard to the pediatric cohort over 1 year of age, a constant higher incidence trend in the period 2002 to 2012 was observed in south of Italy which was 3 fold higher compared to the northern (Fig. 3b). The highest rates were observed in the Sicily, Puglia and Basilicata (Additional file 1: Figure S3b).

The attempt in comparing intussusception hospitalization rates in infants across different time intervals, before and after 2007 (before and after RV use in Italy), stratified by age in months did not show any increases in the first year of age during post-vaccine introduction years, when compared with pre-vaccine introduction rates (Additional file 1: Table S1). Overall, the IRR of intussusception was 0.92 in 2007–2008 (95 % CI: 0.83–1.03) and 0.89 in 2009–2010 (95 % CI: 0.79–1.00) and 0.87 in 2011–2012 (95 % CI: 0.77–0.97).