ABSTRACT
OBJECTIVE: To determine the viral etiology of community-acquired diarrhea in children admitted to hospitals and presenting in emergency departments, pediatric practices and child care centers from November 1, 1997, to June 30, 1998. STUDY DESIGN: Children with diarrhea were identified in a prospective multisite cohort study and analyzed according to age, gender and duration of hospitalization. Stools were tested for rotavirus by enzyme immunoassay and for all other enteric viruses by electron microscopy. RESULTS: Of the 2524 children identified with diarrhea, stools of 1386 (55%) were tested by enzyme immunoassay for rotavirus, and of these 1365 (54%) were screened by electron microscopy for all identifiable enteric viruses. Rotavirus was found in 32% (n = 437), adenovirus in 4% (n = 55), torovirus in 3% (n = 44), Norwalk-like viruses in 2% (n = 25) and astrovirus (n = 14) and calicivirus (n = 7) in fewer than 1% of the specimens tested. The proportion of rotavirus was significantly higher in children 12 to 23 months of age (43% of tested stools, n = 159) and 24 to 35 months of age (38% of tested stools, n = 64) (P < 0.001) than in any other age group. Toroviruses were found to approximately the same extent in children > or =36 months of age (6% of tested stools, n = 19) as those <36 months of age. Rotavirus (36% of tested stools, n = 375, P < 0.0005) and torovirus (4% of tested stools, n = 43, P < 0.004) were most often found in hospitalized patients. In contrast Norwalk-like viruses (P < 0.001) and astroviruses (P < 0.01) were more commonly detected in specimens from patients who presented to physicians' offices and who were symptomatic for gastroenteritis in child care centers. CONCLUSION: This study demonstrates that although all known gastroenteritis viruses were diagnosed in symptomatic children, rotavirus was the etiologic agent in most cases of diarrhea managed in the community and in the hospital.
