Le d’eficit immunitaire du SIDA pr’edispose à de nombreuses infections à germes opportunistes et non opportunistes. Nombre de ces germes d’eterminent des affections gastro-intestinales conduisant à une diarrh’ee et à une malabsorption. Compte tenu du potentiel l’ethal de certains de ces pathogènes, il est souhaitable d’identifier les agents infectieux curables. La d’etection d’un certain nombre de pathogènes n’ecessite, pour des raisons th’eoriques, des biopsies orient’ees par l’endoscopie (Mycobacterium avium, CM). Le diagnostic d’autres bact’eries (salmonelles, Campylobacter, Shigelles, Clostridium difficile) repose sur l’examen des selles sans examen endoscopique. Une troisième cat’egorie (Cryptosporidium, Isospora belli, Giardia, Entamoeba histolytica) peut être identifi’ee soit par l’examen des selles, soit par l’examen du liquide d’aspiration j’ejunal, soit enfin par des biopsies orient’ees endoscopiquement.

We studied prospectively 132 patients with acquired immunodeficiency syndrome to define the spectrum of enteric pathogens during this disease, with special reference to the correlation between the lesions, the infections, and the symptoms. Forty-four percent of the patients harbored at least one enteric pathogen: the most frequently recovered were Cryptosporidium (28), cytomegalovirus (16), Entamoeba histolytica (13), Giardia lamblia (9), and Mycobacterium avium intracellulare (7). Patients harboring pathogens were more likely to be diarrheics (69%) than patients without a pathogen (38%; P=0.01) and more likely to have endoscopic lesions (29%) than patients without a pathogen (4%; P<0.001). The most common pathogen associated with diarrhea was Cryptosporidium. Cytomegalovirus, Entamoeba histolytica, and Salmonella typhimurium were each significantly associated with endoscopic lesions. Patients with cytomegalovirus infection tended to have a greater incidence of ulcer than patients without cytomegalovirus infection. Stool analysis diagnosed 61% of the infections, while endoscopy diagnosed 44%. Seven percent were recognized by stool analysis and endoscopy. When considering the 24 patients in whom accurate diagnosis warranted endoscopic biopsies, stool examination alone would have given an incomplete diagnosis in 14 patients (due to the presence of polyinfection). The frequency of inaccurate diagnosis of infection by stool determination alone, plus the development of new antiviral agents that suppress cytomegalovirus, may favor the earlier application of endoscopic evaluation in these patients.