Objective: evaluate the association between the clinical conditions: dysentery and fever > 39oc, with clinical improvement in children with acute diarrhea of a possible bacterial origin (DBO) Methods: case and controls study formed in a cohort in children 1 of the variables: dysentery, fever >39°c, convulsions, systemic inflammatory response syndrome (SIRS), immunosupression, contagious notion). The cases were children with DBO in whom diarrhea was > 14 days (persistent diarrhea) and/or required antibiotics (no clinical improvement). The controls were children with DBO in whom the diarrhea was auto limited to 39°c (60.13%), convulsions (3.92%), SIRS (6.54%), immunosupression (1.96%), contagious notion (17.65%). Enteropathogens were isolated in 33 (21.6%), shiguella (18,3%), non typhi salmonella (2%), salmonella typhi (1.3%), adenovirus (0,65%), rotavirus (3,3%), E. hystolitica trophozoits (13,7%). 29.5% received antimicrobials for management. The odds ratio (CI95%) for no clinical improvement in children with dysentery was 2.9(1.3-6.6) after adjustment for confounding factors. Conclusions: disentery among children with DBO is associated with non clinical improvement. Variables such as fever > 39C, contagious notion, age <24 months were not associated with clinical improvement so we suggest not to use these criteria for the use of antibiotics.