Background: Knowledge about causes of acute diarrhea among children in developing countries is insufficient. Molecular methods might improve diagnostics of infectious gastroenteritis, but due to the high sensitivity, findings may be difficult to interpret.
Methods: Feces samples from Rwandan children 0.5-5.0 years of age, with diarrhea for <96 hours (patients, n = 544) or without diarrhea for 14 days (controls, n = 162), were analyzed by real-time polymerase chain reaction targeting 17 pathogens.
Results: At least 1 agent was detected in 94% of patients and in 79% of controls, with higher rates in sick children for rotavirus (42% vs. 2%, P < 0.0001) and enterotoxigenic Escherichia coli (ETEC)-estA (21% vs. 9%, P = 0.0006). Detection rates did not differ significantly for adenovirus (39% vs. 36%), ETEC-eltB (29% vs. 30%), Campylobacter (14% vs. 17%) or Shigella (13% vs. 10%), but for Shigella the threshold cycle (Ct) values were lower (pathogen loads were higher) in sick children than in controls. By multivariate analysis, including gender and age, detection of rotavirus (P < 0.0001), ETEC-estA (P = 0.001), Shigella (P = 0.004) and norovirus genogroup II (P = 0.009) was associated with symptomatic infection, and a Ct value below a cutoff (in the range 28-29) improved identification of ETEC-estA, Shigella and norovirus genogroup II.
Conclusion: Real-time polymerase chain reaction can detect essentially all diarrheagenic agents, and provides Ct values that improve identification of clinically relevant infections.