Abstract

BACKGROUND AND AIMS: The optimum diagnostic investigation for patients with acquired immunodeficiency syndrome (AIDS) and diarrhoea is not known. Often no pathogen is detected and it is unclear whether this is because pathogens are absent in some patients or the investigations used fail to detect them. The hypothesis that AIDS related diarrhoea is usually due to an infection, which can be identified by a simple diagnostic strategy based on the results of intensive investigation of a cohort of such patients, was investigated. METHODS: 155 patients with AIDS and chronic diarrhoea underwent contemporaneous examination of stools, duodenal, jejunal, and rectal biopsy specimens and duodenal aspirate for bacterial, protozoal, and viral pathogens. A decision tree analysis was used to determine the best sequential diagnostic strategy for clinicians. RESULTS: 128 of 155 patients investigated (83%) had at least one potential pathogen. The presenting clinical features could not predict the presence or site of the pathogens. Stool analysis identified the most pathogens (93 of 199, 47%). Rectal biopsy was essential for the diagnosis of cytomegalovirus and adenovirus. Duodenal biopsy was as helpful as jejunal biopsy and detected some treatable pathogens missed by other methods. Electron microscopy, impression smears, and duodenal aspirate yielded little extra information. If gut biopsy was reserved for patients without a stool pathogen, some treatable pathogens would have been missed. CONCLUSION: Most patients with AIDS and chronic diarrhoea have at least one gut pathogen, which can be identified by stool analysis and light microscopic examination of duodenal and rectal biopsies. Some pathogens will be missed unless all these investigations are done on all such patients.