Shiga Toxin-producing Escherichia coli in Peruvian Children With Bloody Diarrhea

Abstract

Shiga toxin-producing Escherichia coli (STEC) is not routinely sought in clinical laboratories in developing counties. Among 131 bloody diarrhea samples in Peruvian children <5 years of age, STEC was found in 9.2% and was associated with absence of fever, an observation that may increase suspicion of these pathogens. Because of the significant prevalence of STEC locally, proper diagnostics methods should be implemented in the region.

Bloody diarrhea is frequently treated with antibiotics without a specific etiologic diagnosis. Important infectious causes of bloody diarrhea include Shigella, Salmonella, Campylobacter, Shiga toxin-producing Escherichia coli (STEC), other diarrheagenic E. coli, and Entamoeba histolytica. Management is often difficult in resource-limited settings because diagnostic methods may be unavailable or prohibitively expensive, and because treatment with antimicrobial agents can be dangerous when STEC are the cause of illness. STEC, particularly serotype O157:H7, is the main cause of hemolytic uremic syndrome.¹ Use of antimicrobial agents to treat these infections might actually increase the risk of hemolytic uremic syndrome.² It is clear that STEC commonly cause illness in Argentina³ and Chile,⁴ but there are fewer data regarding their frequency in other parts of South America.⁵ We sought to determine the prevalence and clinical characteristics of STEC infection in Peruvian children with bloody diarrhea.

METHODS

We conducted a prospective study in pediatric patients younger than 5 years of age with bloody diarrhea at the emergency room of Hospital Nacional Cayetano Heredia in Lima, Peru. Consecutive patients were enrolled during a period of 16 months. Epidemiologic and clinical data were collected, and the patient's stool sample was taken on the first day of evaluation. Routine stool cultures were performed to determine the presence of Shigella and Salmonella. We used a multiplex real-time polymerase chain reaction (RT-PCR) for detection of each category of diarrheagenic E. coli⁶ using specific primers for the following genes: AggR for enteroaggregative E. coli, ipaH for enteroinvasive E. coli, eae for enteropathogenic E. coli, lt and st for enterotoxigenic E. coli, daaD for diffusely adherent E. coli, and stx1, sxt2, and eae for STEC. Additionally, an enzyme-linked immunosorbent assay (ELISA) for Shiga toxin (Stx) was performed on stools to detect Stx1 and Stx2 (Premier EHEC).

RESULTS

We enrolled 131 children with bloody diarrhea between April 2007 and August 2008. The median age of the children was 15 months (range, 1–69); 59% (77/131) were 1 year or older (Table 1).Stool cultures, RT-PCR, and ELISA for STEC were done with all patients samples. Pathogens were found in 53% (70/131) of samples; infections in which more than one organism was detected (coinfections) occurred in 15% of patients (20/131). The pathogen isolation rate was higher in children older than 1 year (65% vs. 37%, P = 0.002). The most commonly isolated pathogens were Shigella (18%) and enteropathogenic E. coli (11%). Shigella isolation was significantly more common in older children (31% vs. 0%, P < 0.001). Fever and vomiting were significantly more frequent in children older than 1 year old, often associated with Shigella infections. Among 24 patients with Shigella, 23 (96%) had fever and/or vomiting; however, half of these patients had a coinfection with a diarrheagenic E. coli pathotype. The most frequent species of Shigella were S. flexneri (38%, 9/24), S. boydii (33%, 8/24), and S. sonnei (25%, 6/24).

Table 1: Clinical Characteristics and Pathogens Isolated in Bloody Diarrhea Patients Using Routine Culture, RT-PCR, and ELISA for STEC

STEC infection, measured by the presence of Stx, was detected in 12 patients (9%); the diagnosis was made by ELISA in all cases. Only 2 of these 12 patients had a positive RT-PCR in the analysis of 5 colonies; one had an stx1+/eae+ STEC strain and the other had an stx2+/eae− strain. Fever was significantly less common among STEC-positive patients than patients without this pathogen (34% vs. 67%, P < 0.05). Five of 12 STEC-positive children (42%) had a family member with diarrhea in relation to 29 of 119 (24%) of STEC-negative children, although not statistically different.

DISCUSSION

STEC isolates were identified in 9% of pediatric patients with bloody diarrhea. This rate is higher than in North American studies.⁷ The overall pathogen documentation was high in this study (53%) compared with other studies of bloody diarrhea. Use of several detection methods (routine culture, PCR, and ELISA) resulted in recognition of STEC at a higher rate than culture on MacConkey Sorbitol media alone.⁸

What distinguishes our study from others is the detection of all groups of diarrheagenic E. coli by routine use of PCR; this allowed definition of etiology for half of all bloody diarrhea cases. The pathogen isolation rate was significantly higher in children older than 1 year. In a recent cohort study conducted in more than 1000 children (age range, 2–12 months) in Lima, Peru,⁹ the diarrheagenic E. colias a group were the most common pathogens and their isolation was more frequent in samples from older infants, as in the current study. This may be due to the decreased exposure to milk-related protective factors and increased exposure to contaminated food and water, as dietary options becomes broader.

The diagnosis of STEC is challenging; ELISA detects the presence of the toxins with good specificity and sensitivity, although lower than cell culture.^7,8 We used the Premier EHEC ELISA which has a sensitivity of 96% and a specificity of 99%.⁷ In our study, only 2 of the 12 ELISA-positive samples were also positive by PCR. This discrepancy could be explained by the high frequency of antibiotic use before to the emergency room visit (28%). We do not recommend routine use of this method because the prevalence of STEC in watery diarrhea cases in Peruvian children is very low.⁹

In our study, STEC-infected children had fever significantly less often than patients without this infection (P < 0.05), as has been reported elsewhere.^7,9,10 Although the study samples were small, 5 of 12 patients infected with STEC had a family member with diarrhea and 1 of them had bloody diarrhea. This demonstrates the possibility of STEC transmission in households,¹⁰ and highlights the importance of considering family contacts for possible effective intervention for prevention of STEC infection.

Our study has several limitations; first, we may not have done optimal studies for all pathogens associated with bloody diarrhea (ie, Campylobacter, Yersinia, Balantidium coli, etc). For example, our studies for Campylobacter may not have been optimal. However, other pathogens such as Yersinia and parasites associated with bloody diarrhea have historically been rare in our setting. Second, due to the cross-sectional nature of the study, we were unable to follow all of the patients to determine their outcome. This is a small study and not necessarily reflects the overall incidence of pathogens in other areas in Peru; however, it is representative of the periurban and urban communities of Lima.