A descriptive review of selected nonviral enteric illnesses reported in children in Quebec between 1999 and 2006

1Institut national de santé publique du Québec; 2Unité de recherche en santé publique, Centre Hospitalier Universitaire de Québec, Québec; 3Laboratory for Foodborne Zoonoses, Public Health Agency of Canada, Faculté de Médecine Vétérinaire, Université de Montréal, Montréal; 4Département de pédiatrie, Centre hospitalier universitaire de Québec, Québec; 5INRS-Institut-Armand-Frappier, Laval, Québec Correspondence and reprints: Dr Patrick Levallois, Institut national de santé publique du Québec, 945, Avenue Wolfe, B4-56, Québec, Québec G1V 5B3. Telephone 418-650-5115 ext 5216, fax 418-654-3144, e-mail patrick.levallois@msp.ulaval.ca Diarrheal disease remains an important cause of morbidity and mortality in children worldwide (1). The estimated incidence rates in developing countries range from 3.5 to 7.0 cases/child/year during the first two years of life, and from two to five cases/child/year for the first five years (2). In the United States (US), more than 20 million cases of diarrhea are estimated annually in children younger than five years of age, leading to approximately 2.5 million visits to a doctor, 220,000 hospitalizations and approximately 40 deaths (3). Bacterial enteric pathogens are second to rotaviruses among the known causes of diarrheal disease in infants and young children (4). Campylobacter and Salmonella are commonly identified causes of diarrheal disease in children (4,5). Because of their virulence, infections by Escherichia coli O157 and Yersinia enterocolitica are other important pathogens commonly reported, but occur with relatively less frequency in this age group (6-8). The most important intestinal parasites infecting North American children include Giardia lamblia (intestinalis) and Review

In Quebec, infections attributed to Giardia, Salmonella, Campylobacter, E coli, Yersinia and Cryptosporidium represent up to 90% of all notifiable nonviral gastroenteritis reported in regional public health registries (16).
More insight into the epidemiological characteristics of gastroenteritis caused by these leading pathogens, and specifically those pertinent to young children, may help to enhance current control programs within communities (3).In considering the importance of targeting our interventions for this age class, there are, however, very few published population-based studies to date that give a specific description and analysis of the basic epidemiology pertinent to infectious gastroenteritis in children (17,18), with most of them being case-control studies with small population subsets or data obtained through sentinel surveillance systems.
The objective of the present study was to review the patient characteristics, and the seasonal and geographical patterns of nonviral enteric illnesses in children most commonly reported in Quebec between 1999 and 2006.

Data sources
Records of reported cases were extracted from the diseases registry of the Regional Public Health Departments (RPHD) of the Quebec Ministry of Health and Social Services.In Quebec, all notifiable diseases are reported to the RPHD through an electronic database system designed to gather and manage a specific list of diseases that must be reported by physicians or chiefs of microbiological laboratories.Quebec has 18 health regions, each of which is mandated by the provincial Ministry of Health to forward reportable diseases data to a central registry managed by the provincial public health laboratory.
Cases were defined as children who were between zero and four years of age with a laboratory diagnosis of infection by Giardia, Salmonella, Campylobacter, Yersinia, invasive E coli and/or Cryptosporidium (Appendix 1), and a notification by the physician or laboratory director to the RPHD between 1999 and 2006.The total population of children, between zero and four years of age from 1999 to 2000, was estimated and projected by the Quebec Statistics Institute, based on municipal population using the 1996 general census data compiled by Statistics Canada.The estimated population was updated each year using the general census data from 2001 and 2006.

validation, exclusion and ethics
Cases from northern health regions (northern Quebec, Nunavik and James Bay Cree territories) and those that occurred outside Quebec were excluded from the study.A validation was also performed according to standard disease definitions, geographical location and possible association with known outbreaks (19).Intra-and inter-regional duplication of each case was avoided by verifying the date of onset, postal code, birth date and sex.Physician diagnosis and laboratory results were also cross-checked for all records.A minimum time interval between two episodes for the same child was estimated to avoid any link between reports (one to 10 days for Campylobacter and E coli, three to 10 days for Yersinia, one to four weeks for Giardia, one to 12 days for Cryptosporidium, and three to 90 days for Salmonella).The study has been independently reviewed and approved by the Ethics Review Board of the Centre Hospitalier Universitaire de Québec, Quebec.

Descriptive analyses
Incidence rates of Giardia, Salmonella, Campylobacter, other pathogens (Y enterocolitica, C parvum and invasive E coli) and all selected pathogens (Giardia, Salmonella, Campylobacter, E coli, Yersinia or Cryptosporidium) were calculated according to two age groups (zero to one year, and one to four years), by sex and year, and season and year.The age categorization was mainly designed with two major considerations: statistical efficiency and children behaviour, and vulnerability in relation to the epidemiology of these diseases.Therefore, children between zero and one year of age, and children between one and four years of age were grouped together for descriptive analyses.Rates were expressed as cases per children-years.Seasonal variation and trend analyses were performed using a moving average algorithm available in Microsoft Excel 2003 (Microsoft Corporation, USA).The incidence rates of Giardia, Salmonella, Campylobacter and all selected pathogens were further calculated according to four categories of municipalities based on their population sizes: the census metropolitan area of Montreal; other census metropolitan areas with municipalities of more than 100,000 inhabitants; census agglomerations with municipalities ranging from 10,000 to 100,000 inhabitants; and census areas with municipalities of fewer than 10,000 inhabitants.A descriptive analysis of the geographical distribution for the cumulative incidence rate of cases associated with the three main pathogens was performed using the empirical Bayes smoothing method to account for various population sizes and small denominators.Maps were constructed using the Jenks natural break algorithm to categorize incidence rates per health area (ArcGis version 9.3, ESRI, USA).The Moran's I statistic was estimated to assess evidence of global spatial autocorrelation (clustering) of incidence rates over the study region.Computation of the Moran's I statistic was performed using GeoDa software version 0.9.5-I (GeoDa Center for Geospatial Analysis and Computation, USA) (20).

Case occurrence
There were a total of 3,039,085 children-years estimated from 1999 to 2006 in Quebec.Over the study period, a total of 5104 validated cases were extracted from the RPHD registries.From these, 36 were excluded due to their location in the northern regions (northern Quebec, Nunavik and James Bay Cree territories).The remaining 5068 cases were included in the analyses and were divided into two age groups.A group representing cases in children younger than one year of age, which included 277 cases (5.5%); and one representing children between one and four years of age, which included 4791 cases (94.5%).The incidence rates for children younger than one year of age ranged from 0.4 to 5.8 cases/100,000 children-years.Cases attributed to Giardia, Salmonella and Campylobacter represented a total of 94.3% of all studied cases, with similar incidence rates for each of these three pathogens averaging 50 cases/100,000 children-years.In comparison, the total number of cases attributed to Yersinia, Cryptosporidium and E coli represented 5.7% of all cases, with incidence rates of less than 10 cases/100,000 children-years.Higher rates were also observed in the one-to four-year-old group compared with the younger than one-year-old group.
Detailed incidence rates per pathogen and according to age group are presented in Table 1.Incidence rates of all selected pathogens and Campylobacter cases revealed a significant (P<0.0001)difference between sexes, showing higher rates in males than females (Table 2).The incidence rates of Giardia and Salmonella cases were similar for both sexes.

Population size of municipalities and spatial distribution
The incidence rates of all the pathogens, and especially those of Salmonella and Campylobacter, seemed to increase with the population size of the municipalities, except for the incidence rates of areas that included municipalities with fewer than 10,000 inhabitants (Table 3).This 'v-shape' pattern was clearer for Campylobacter cases, whose rates in areas including municipalities of fewer than 10,000 inhabitants were significantly higher than those of the census metropolitan area of Montreal.However, the incidence rates of Giardia cases were significantly lower in areas with municipalities of fewer than 10,000 inhabitants, but seemed to be high in other census metropolitan areas with municipalities of more than 10,000 inhabitants.
The spatial distribution of incidence rates for all selected pathogens is presented in Figure 1.Visual assessment revealed a heterogeneous distribution of incidence rates in the province, with areas of higher incidence located in the southern health regions.However, the global Moran's I index of 0.09 indicated a significant (P=0.001)spatial autocorrelation of incidence rates.

trends and seasonal variations
Temporal variations in the incidence rates of Giardia, Salmonella and Campylobacter cases exhibited regular seasonal patterns with consistent peaks during summer for Salmonella and Campylobacter and autumn for Giardia.Seasonal peaks in the incidence rates of Giardia cases were low from 1999 to 2002, and higher from 2003 to 2006.The incidence rates of Campylobacter cases peaked slightly during two summers of the study period -2005 and 2006.All of these incidence rates were stable throughout 1999 to 2006 (Figure 2).

DIsCussIOn
Excluding viral enteric gastroenteritis, Giardia, Salmonella and Campylobacter were the most common enteric pathogens notified in children younger than five years of age in Quebec,  accounting for approximately 50 cases/100,000 children-years for each of these pathogens.The importance of these enteric pathogens in children was also highlighted previously in Quebec (21), and by Koehler et al (22) in the US using FoodNet data collected between 1996 and 1998.The incidence rate of reported cases of Giardia (54.2 cases/100,000 children-years) estimated in the study was approximately five times higher than the estimated rate for the general population (12 cases/100,000 personsyears) in Quebec (23).Similarly, the incidence rate of reported Salmonella cases in the study was approximately 2.5 times greater than the rate estimated for the general population in Quebec (24).Cases were defined (19) and validated based on the standard definition and time criteria, such as laboratory ascertainment and time frames for each pathogen, to avoid double reporting of the same case (25).Several studies in Canada that reviewed reported cases of Giardia found that children younger than five years of age had the highest incidence rates (26), and that males showed the highest rates among all age groups (26,27).The highest rate of Campylobacter cases was observed  in the zero-to four-year age group and the 20 to 39 year age group, with slightly higher rates occurring in males than in females (28).The investigation of age, sex and residencespecific incidence rates of Salmonella enteritidis cases in Michigan (USA) found a higher incidence rate of S enteritidis cases associated with children younger than five years of age compared with the national average incidence rate.Children younger than one year of age and those between one and four years of age were at a higher risk of S enteritidis infection than adults between 15 and 39 years of age (29)(30)(31).
Global assessment of the occurrence of reported enteric pathogen cases revealed similar incidence rates in several other studies, with the highest incidence rates of gastroenteritis being in children zero to four years of age (32)(33)(34)(35).The reasons behind the higher incidence rates in children for illnesses linked to these pathogens are unclear.However, three broad sets of factors may cause much higher incidence rates of reports in young children for these enteric diseases: greater contact, exposure and infectious doses directly linked to specific behaviours of young children and their interaction with other people and their environment (30,31); greater susceptibility of young children to become sick attributable to gut immaturity and immune system naivety; and greater reporting fraction (less under-reporting) in this age group due to parental concerns and greater severity of symptoms.Indeed, Glass et al (36) estimated that 16.5 million children younger than five years of age experience between 21 and 37 million episodes of diarrhea annually.Of these, 2.1 to 3.7 million lead to a physician visit, a total of 220,000 patients are hospitalized and 325 to 425 children die.The major cost of diarrhea is attributable to the high number and cost of hospitalizations, because approximately 10.6% of hospitalizations in this age group are due to diarrhea (36).Although our study did not investigate these factors and their relative contributions to the final incidence rates of reported cases, these data suggest that preventive and interventional measures aimed specifically at children could bring the greatest benefit with respect to reducing the disease burden caused by these pathogens in the Quebec population.
An interesting relationship seems to emerge in the present study between the incidence rates of Salmonella and Campylobacter infections and the population size of the municipalities where the reported cases resided.For municipalities of more than 10,000 people, our data suggests a proportional increase of reporting incidence rates with the size of the underlying population, which contrasts with municipalities of fewer than 10,000 inhabitants for which incidence rates of these reported cases seemed the highest.This 'v-shape' pattern in the relationship between incidence rates and population sizes leads us to believe that there may be two broad influences associated with this factor: one relating to the environmental exposure of populations living in small municipalities where water source, proximity to livestock or wildlife, and lifestyle may contribute to increase infections.This pattern may also be related to various parameters of social and material deprivation and ethnocultural factors (genetic or behavioural) that seem to be linked to larger municipalities in Quebec.If we simplify the terminology by designating areas as 'rural', where municipalities have fewer than 10,000 inhabitants, and remaining areas as 'urban', we note large differences in the age structure of incidence rates of Campylobacter cases between the two areas.In rural areas, the incidence rates of Campylobacter cases are higher in almost every age and sex category, with the greatest difference observed in the zero-to four-year age group, in which rates in rural males were 7.3 times higher than in urban males, and 6.95 times higher in rural females compared with urban females (28).Any significant differences in the incidence rates of Salmonella cases were found to occur between urban and rural dwellers or between sexes (29)(30)(31).Socioeconomic conditions can explain differences between rural and urban settings, although the association of enteric pathogens (bacteria and protozoa) with low socioeconomic conditions is not clearly established.
Our results were consistent with these findings, but we found that the incidence rates of Salmonella cases were higher in urban areas of Montreal.This may be explained by a larger number of socioeconomically deprived children in urban areas (37).High incidence rates of Giardia cases have also been reported in urban areas, but spatial analysis revealed that the highest incidence rates around Montreal were in areas where agricultural activities were prominent.This is consistent with several studies performed in Ontario (26,38).It may also be explained by the increased awareness, reporting and access to health services in urban areas, which could account for the larger number of cases reported in those areas (26).
We also consistently demonstrated, similar to several other studies, a seasonal pattern peaking in early autumn for Giardia (26,27,39,40).The autumnal pattern could be explained by the long incubation period of Giardia cysts.Most of the Giardia infections also occurred during outdoor activities involving more contact with contaminated water and/or more personal contact between friends and families during the summer holidays.
Different clinical microbiology laboratories may use different methodologies and standards, and reporting differences may also exist.If a given area is served by several laboratories and some of these use more sensitive methodologies than others, this will be reflected in an uneven distribution of cases (41 ).
Outbreaks caused by the enteric pathogens mentioned in the present study may introduce a separate level of complication in the analyses.Our descriptive geographical analysis was performed with the sole objective of visualizing geographical areas with higher risk of reports.In such an instance, outbreaks need to be taken into account -their presence may bias case distribution because it relates to the local geographical factors (41).In the present study, only cases tagged as 'sporadic' were included for analysis and cases known to be related to outbreaks were registered in different databases that were not included in the study.
One common limit inherent to studies such as the present one, which uses public health surveillance data, relates to the under-reporting of cases compiled in administrative systems.Estimating the real magnitude of infections can present challenges when studying acute gastrointestinal infection risk factors, planning health programs and prioritizing health promotion.This bias has been studied in Ontario (42) and British Columbia (43).For each case of acute gastrointestinal infection reported in these provinces, an estimated 313 to 347 cases occur in the community, respectively.In another study, it was estimated that for every reported case of verotoxigenic E coli, Salmonella and Campylobacter, there were approximately 10 to 47, 13 to 37, and 23 to 49 cases, respectively, in the Canadian population (44).In our study, the incidence rates were certainly underestimated and should be considered as conservative measures.Although correcting these values with ratios published for other provinces could be performed, it is likely that there are systematic differences in underestimation fractions across the provinces when reporting cases to public health authorities; this should be further investigated before attempting such corrections.To date, no previous investigation of nonviral enteric gastroenteritis in children between zero and four years of age has been performed in Quebec, and results presented here should be interpreted with caution.Further population-based studies with individual data on risk factors are needed for a better understanding of the epidemiology of these conditions in young children in Canada.

Figure 2 )Figure 1 )
Figure 2) Giardia, Salmonella and Campylobacter empirical Bayes smoothing incidence rates (per 100,000 children-years) showing seasonal variation and trends in children between zero and four years of age (1999 to 2006)

ACKnOWLeDGeMents:
The authors thank Professors Yves Brousseau and Marie-Hélène Vandersmissen, and Mr Serge Duchesneau, GIS technician from the Geography Department at Laval University, Laval, Quebec.They also thank Mr Stéfano Biondo, cartographic documents and geospatial data documentation adviser at the Laval University library, and Ms Stéphanie Brazeau and Mr Philippe Martin of the Université de Montreal (Montreal, Quebec).Furthermore, they thank the Directors of the Regional Public Health Departments of the Quebec Ministry of Health and Social Services for providing the data.suPPORt:The present research was jointly supported by the Fonds québécois de la recherche sur la nature et les technologies (FQRNT) and the Public Health Agency of Canada.The research work was performed at the Institut national de santé publique du Québec.APPenDIx 1: DIseAses DeFInItIOn Giardiasis Confirmed case.Presence of one of the two conditions: 1. Presence of cysts or trophozoits of Giardia lamblia in the stool, in the duodenal fluid or in the biopsy of small intestine; or 2. Detection of G lamblia antigens in the stool using the enzyme immunoassay or the direct fluorescent antibody technique.Infection due to Campylobacter Confirmed case.Isolation of one of the Campylobacter species in an appropriate clinical sample.salmonellosis Confirmed case.Isolation of Salmonella (excluding Salmonella typhi or paratyphi) from an appropriate clinical sample.

Table 3 Urban/rural distribution of selected bacterial and parasitic infection incidence rates (IR) of children between zero and four years of age in Quebec (1999 to 2006) areas Giardia Salmonella Campylobacter all selected pathogens*
*Including Giardia, Salmonella, Campylobacter, Escherichia coli, Cryptosporidium and Yersinia