Diarrhoeal diseases constitute a major public health problem, particularly in the developing world, where the rate of mortality and morbidity is very high 1. The World Health Organization (WHO) has estimated that 1.5 billion episodes of diarrhoea occur every year in developing countries, resulting in 3 million deaths 2. In Mexico, a governmental study conducted in the year 2003 reported 4556 cases caused by intestinal infectious 3. The etiological agents of diarrhoea described in epidemiological studies are transmitted as waterborne and foodborne.

Some foodborne pathogens have been recently considered as emerging diseases 4, despite the fact they have been known since a long time ago. For example, outbreaks of salmonellosis have been described for many decades, and yet their incidence have increased over the last 25 years. Diarrhoeal infections can be caused by many etiological agents, but mainly by enterobacteria such as Escherichia coli, Salmonella spp., Shigella spp., Campylobacter jejuni and Vibrio cholerae; as well as parasites such as Entamoeba histolytica and Giardia intestinalis, and some rotaviruses are also important agents 5.

Escherichia coli is considered as the etiological agent for many diseases including some affecting the urinary tract and intestine. The classification of diarrhoegenic E. coli strains is based on their virulence properties, and comprises six groups: Enterotoxigenic E. coli (ETEC), Enteropathogenic E. coli (EPEC), Enteroinvasive E. coli (EIEC), Enterohaemorragic E. coli (EHEC), Enteroaggregative E. coli (EAggEC) and Diffuse Adhering E. coli (DAEC) 6. Salmonella spp. is a facultative, gram negative, flagellated member of the Enterobacteriaceae family. The most extensive accepted classification of Salmonella strains is based on the diversity of two differentially expressed H flagellar antigens: flagellin phase I and phase II antigens (codified by fliC and fljB genes), and the O antigens of the bacterial lipopolysaccharide, both determined by serotyping 7. Until now, 2501 serotypes have been described 8; which turns Salmonella classification into a complex and laborious process in the clinical laboratory; therefore, several PCR based methods have recently been developed, and were reported to be a simple, highly sensitive, fast and reliable alternative when compared to traditional clinical laboratory methods 910.

Shigella is a Gram negative, non-motile, rod-shaped bacteria, closely related to E. coli and Salmonella, and it is the etiological agent of human shigellosis and dysentery, which is characterized by severe diarrhoea with the presence of blood in the faeces. Classification of Shigella is based on serotyping, and comprise the following groups and serotypes: Serogroup A (S. dysenteriae) 12 serotypes, serogroup B (S. flexneri) 6 serotypes, serogroup C (S. boydii) 23 serotypes and serogroup D (S. sonnei) with only one serotype 11.

Diarrhoea can also be caused by protozoa such as Entamoeba histolytica and Giardia intestinalis, these zoonotic parasites are frequently transmitted by consumption of water contaminated with infective cysts 12.

The purpose of this study was to determine the prevalence of five enteropathogen diarrheogenic agents namely E. coli, Salmonella spp., Shigella spp., E. histolytica/E. dispar and Giardia intestinalis in Mexico City. Stool specimens obtained from children patients from several communities in Mexico City were examined; the association patterns between different pathogens and its correlation with occurrence of diarrhoea were also described.

The current study used a variety of diagnostic methods which helped estimate 100% prevalence of the enteropathogens in stool samples from children presenting diarrhoea symptoms. In addition, a high rate for multiple infections 156/300 (52%) was observed. However, not only the patient group had enteropathogens, also 50/80 (62.5%) members of the control group were positive for either one or two parasites (E. histolytica/E. dispar and Giardia intestinalis). Although only 4 (5%) of the control group were positive for diarrhoegenic E. coli (Table 3), they did not have any symptoms by the time this study was conducted. One of the strains detected in the controls was E. coli O157:H7, which has been recently described as an emerging pathogen worldwide 615. The high rate of enteropathogens detected in this study among both groups, reflects the importance of monitoring on a daily basis the most vulnerable population, such as low economic level children groups.

E. histolytica/E. dispar were the most frequently found enteropathogen in both, the patient (70.3%) and the control group (43.7%) (Table 1). Globally, amoebiasis is wide spread in approximately 20% of the world's population; 10% of those individuals get sick, and 0.1- 0.25% of them die. E. histolytica constitutes the third cause of death for parasitic diseases 16. On the other hand, G. intestinalis a protozoa which causes symptomatic infections mainly in children under 12 years old was observed in 33% of the patient group, and 20% for the control group (Table 1). These results are in agreement with a previous study that reported G. intestinalis in 29.9% of a population group in Mexico City 17. An interesting observation is that although 63.7% of the members of the control group were positive for E. histolytica/E. dispar and/or G. intestinalis (Table 1), none of them presented symptoms of disease.

We used three multiplex PCR methods that accurately detected specific genetic virulence markers of Salmonella spp., Shigella spp., and E. coli groups ETEC (eltB and estA gene fragments), EPEC (eaeA and bfpA gene fragments), VTEC (eaeA and vt1or vt2 genes), EIEC/Shigella (ial region), and E. coli O157:H7 (fliC H7 flagellar antigen and a characteristic O157:H7 DNA fragment previously described) 18. EIEC and Shigella are closely related regarding some of their virulence factors and phenotypic properties 1920, the multiplex PCR method employed in this study can not be used alone for a definitive discrimination between these two enteropathogens since it recognizes the ial locus, which can be found in both, EIEC and Shigella spp. Therefore, ial positive amplification isolates were considered as positive for EIEC if they showed no agglutination when challenged against specific antiserum for Shigella spp. Also, because eltB and ial amplicons are very close regarding their molecular size (322 and 320 bp respectively) (Fig. 1), two separate PCR reactions were done, each containing specific primers against eltB or ial in order to avoid the overlapping of amplification products, that might lead to false positives.

Diarrhoegenic E. coli was present in 32% of the patient group and 5% of the control group (Table 3); these data are consistent with a previous study conducted in Hanoi, Vietnam, where diarrhoegenic E. coli was detected by multiplex PCR, in 22.5 % of the patient group and 12% for the control group 21. The presence of diarrhoegenic E. coli groups has been reported world wide and recognized as one of the major causes of deceases involving enteropathogens in children 2223242526. In Mexico and other developing countries in the world, ETEC is the most prevalent diarrhoegenic E. coli group 2728. It is considered as an important pathogen in children, especially during the first six months, where the isolation rate ranges from 10 to 30% 29. In this study, the presence of ETEC in the patient group (13.3%) was the highest compared to the other E. coli groups detected (Table 1). However, this rate is lower than what has been reported in some parts of the world, which reaches up to 20.7% 3031. It seems that, regardless of the location, the presence of eltB gene (encoding thermo labile toxin) is a definitive advantage for ETEC, since in our study, the virulence gene distribution rate for eltB was high (6.6%). The same observation was reported by another studies conducted in Sweden 13 and Vietnam 21, where eltB was the most commonly found ETEC virulence gene in the studied groups. The second most commonly found diarrhoegenic E. coli strain was EPEC (9.3%), which has been reported responsible for a high rate of mortality and morbidity among children, especially in developing countries where poor sanitary conditions prevail 6. For example, in countries such as Mexico 3233 and Brazil 34 up to 40% of diarrhoeal episodes in children are due to EPEC. Interestingly, regarding virulence gene homogeneity distribution, it seems that the situation for EPEC is not the same as for ETEC. A higher percentage of eaeA+/bfpA- isolates (atypical strains) was found compared to eaeA +bfpA+ (typical strains) (Table 3), which differs from others studies conducted in Vietnam, where most of the isolates were eaeA +bfpA+ 21.

VTEC E. coli, an enteropathogen distributed worldwide, that has been more extensively studied in USA 35 and Europe 36 was also detected in this study. One strain of VTEC was detected in the controls whereas in the patient group it had a prevalence rate of 8.6% (26/300) (Table 1), of which 17/26 (65.4 %) possessed vt1, vt2 and eaeA; 19.2% (5/26) possessed vt2 and eaeA and 15.3% (4/26) vt1 and eaeA; being vt1 the most abundant genotype within this group (Table 3). These results differ from those reported by Svenungsson et al., 13 where vt1 was in fact the less commonly found genotype in their samples.

E. coli O157:H7 is an emergent pathogen causing the haemolytic – uremic syndrome, and is considered as an important foodborne source of intestinal infection causing more than 73,000 diarrhoeal episodes in the United States every year 1537. We found one E. coli O157:H7 in the control group, however, no symptoms of diarrhoea were reported by the time this study was conducted. There is no explanation for this, but it might be possible that the infection was in an early stage when the stool sample was collected, or that the O157:H7 was a non-producing toxin strain. EIEC was the least detected in the patient group (1%) (Table 1), which is in agreement with a previous study conducted in a surrounding area of Mexico City where prevalence was 0.85% 28. It has been estimated that EIEC in developing countries including Latin American is rare 538. The prevalence rate for Shigella sonnei (1.6%) and Shigella flexneri (1%) in the current study was very low (Table 1) when compared other studies 1321. However, in other parts of the world, several Shigella outbreaks have been described, and the main feature is their ability to spread in the population, due to its very low infection doses (~10 bacterial cells per host) 1920; therefore, a low prevalence of cases of Shigella should not be underestimated.

To detect Salmonella spp., two separate reactions containing primers complementary to all internal variable regions of the fliC and fljB genes, which codifies phase I and phase II flagellar antigens (Fig. 2), were used. Both multiplex PCR systems have been previously tested in hundreds of samples from different parts of the world, and were reported to be accurate, fast and reliable alternative to other traditional diagnostic methods used in the clinical laboratory 910. In our study, both stool sample groups showed S. ohio and S. typhimurium as the predominant serovars; nevertheless, many other serovars were present in both groups of samples in a considerable proportion (Table 1), which suggest that Salmonella diversity is more in comparison to E. coli, for which ETEC was considerably predominant (Table 1). Salmonellosis is one of the biggest challenges to public health all over the world. In the year 2000, a study conducted in Mexico showed S. typhimurium and S. ohio as the most and less commonly found serotypes of Salmonella respectively 39. This may be related to the way the outbreak was studied, since we focused on most common serotypes in children (age < 12 years) and not in the whole population. This study is intended to investigate some interesting observations that deserve further studying. The results revealed that all stool samples from the patients and the majority of the control group were positive for at least one enteropathogen and that 52% of stool samples had multiple enteropathogens. All of the 300 stool samples tested except 2 had either E. histolytica/E. dispar or G. intestinalis and when an enterobacteria was present, it was most likely to be a Salmonella serotype. In addition, only 12 out of 300 stool samples tested E. coli presence was in association with a protozoa in the absence of any type of Salmonella spp. (Table 2).

It has been reported that S. typhimurium virulence is enhanced when exposed to the rumen protozoa 40; this hypervirulent phenotype is linked to the presence of the Salmonella genomic island SGI1 integron 41. Also, it has been shown that SGI1 is widely spread in other S. enterica serovars; and although rumen possesses different protozoa diversity than the parasites studied here, some Salmonella isolates from human infections have shown to posses such island 42. It is tempting to speculate that a similar virulence enhancement process might occur in human digestive trait; possibly through E. histolytica/E. dispar or G. intestinalis. Whether those Salmonella strains studied in this work possesses or expresses such genomic island needs further study.

The study of enteropathogen associations may lead to a better understanding of the etiology of diarrhoea, and therefore in prescribing more suitable treatments in cases of outbreaks. This study revealed that 52% of the patients were infected by more than one enteropathogen, notably E. histolitica/E. dispar and Salmonella ohio. These results are useful for clinicians in improving the empiric treatment used in such cases.

The author(s) declare that they have no competing interests.

GLP and EM carried out the molecular studies; OG, JA and EN characterized bacterial isolates, and identified the parasites; SV conceived of the study, and participated in its design and coordination, and drafted the manuscript. All authors read and approved the final manuscript.