Fecal leukocytes are widely used to identify invasive diarrhea and to make then the decision of prescribing or not antibiotics. This test has been hardly assessed in small hospitals of developing countries with efficient laboratory processes. We aimed to assess the diagnostic performance of different thresholds of fecal leukocytes in children under-five with acute diarrhea. Retrospective study of clinical and laboratory records in the Pediatric Emergency Hospital, Lima, Peru. All cases with a stool culture and fecal leukocytes independently and systematically performed were studied. Sensitivity, specificity, predictive values, likelihood ratios (LR), and receiver operating characteristics (ROC) curves were calculated. Out of 1,804 stool samples assessed, 901 (49,9%) were positive for one or more bacterial entheropathogens. Sensitivity (Sn), specificity (Sp), and positive LR varied for different thresholds: more than 5 (S: 93.2%, Sp: 21.9%, LR+:), more than 20 (Sn: %, Sp: %,+LR: ), more than 50 (Sn: 74.9%, Sp: 56.7%, +LR: 1.73), and more than 100 fecal leukocytes per high power field (Sn: 60.7%, Sp: 71.9%, LR+: 2.17). The general area under the ROC curve was 0.69 (CI 95%: 0.67-0.72). Diagnostic performance of fecal leukocytes is suboptimal and may not warrant its continued use in developing settings, as it promotes antibiotic abuse, and on the other hand increases the risk of overlooking patients with invasive diarrhea who may Benedit from antibiotic treatment. Combination of epidemiological and clinical data with either fecal leukocytes or fecal lactoferrin may provide a more efficient approach.