Etiology of diarrhea in pediatric outpatient settings

Background: The frequency with which bacteria cause diarrhea evaluated in ambulatory settings is often unknown. We attempted to determine the microbiologic etiology of diarrhea in a private pediatric practice (site A) and a clinic serving largely immigrant children (site B) and to establish guidelines for bacterial culture. Methods: Children with diarrhea were prospectively enrolled, and their stools were examined for diarrheagenic bacteria, viruses and parasites. Results: A total of 123 and 103 children were enrolled at sites A and B, respectively. Stools from all (100%), 126 (55.8%), 104 (46.0%) and 75 (33.2%) were tested for bacterial enteric pathogens, parasites, Clostridium difficile toxin and viruses, respectively. Of the 75 patients whose stool underwent complete testing, 36 (48%) contained at least 1 definitive or plausible pathogen. Twelve stools (5.3%) tested positive for bacteria [Campylobacter jejuni (n = 7), Yersinia enterocolitica, Shigella flexneri, Shigella sonnei, Salmonella serogroup D and Salmonella Braenderup (n = 1 each)]. One contained Blastocystis hominis, 8 contained C. difficile toxin and 16 contained viruses (9 rotavirus, 5 adenovirus and 2 astrovirus). Visible fecal blood (P = 0.029), increased stool frequency (P = 0.035), abdominal tenderness (P = 0.011) and fecal white (P < 0.001) or red blood cells (P = 0.002) were associated with bacterial infection. All children with stool yielding diarrheagenic bacteria or C. difficile toxin had at least 1 of these factors, but so did 75% of children without these agents (positive predictive value, 11%; negative predictive value, 100%; sensitivity, 100%; specificity, 25%). Conclusions: The bacterial diarrhea prevalence is similar to that in other ambulatory studies, although the spectrum differs. Exclusion criteria for stool testing in diarrhea remain elusive. Studies to determine the etiology of unexplained diarrhea and cost-effective algorithms for diarrhea diagnosis, are needed.