Objectives: To identify pathogens associated with diarrhoea in HIV-infected persons and their HIV-uninfected family members.
Design: Prospective cohort study.
Setting: Rural community in eastern Uganda.
Subjects: Eight hundred and seventy nine HIV-infected adults (74% females and median age 35 years (IQR, 29-41) and 2771 HIV-uninfected family members (51% females and median age 11 years (IQR 6-16) were included.
Main outcome measures: Using microscopy and culture, stools were tested for parasites, bacteria and bacterial-antimicrobial-susceptibility. Logistic regression models, adjusting for age, CD4 cells, season, household clustering and use of safe-water system were used for relationships between pathogens, diarrhoea and HIV.
Results: Persons with HIV had similar pathogens in diarrhoeal (69%) and nondiarrhoeal stools (57%). Most diarrhoea was not associated with identifiable aetiology; the population attributable risk of diarrhoea for known diarrhoea pathogens was 32%. Enteric bacteria (19%), enteropathogenic or enterotoxigenic E. coli (8%), Aeromonas species (7%), Strongyloides stercoralis (8%) and Cryptosporidium parvum (5%). HIV-infected, stools had more Cryptosporidium parvum than HIV-uninfected (OR 2.64, 95% CI 1.43-4.87). Most bacteria were resistant to commonly used antimicrobials irrespective of HIV status.
Conclusions: Irrespective of HIV-status, aetiologies of majority of their diarrhoea in Uganda cannot be identified by microscopy and culture. Bacterial pathogens isolated have high resistance to common antimicrobials. Empiric treatment should be tailored to local bacterial-resistance patterns.