Endotracheal aspirates (ETAs) from mechanically ventilated pediatric patients frequently are cultured as part of an evaluation for suspected sepsis. There are now well-defined criteria for rejecting low-yield ETAs from adults, but it is uncertain whether the same criteria can be applied to ETAs from children. Therefore, we compared the Gram stain and culture results for 361 consecutive ETA specimens collected from pediatric patients over a 1-year period. Results for patients for whom a blood culture was performed within 48 h of the time that a culture of ETA was performed were also reviewed. Gram stains were examined under x100 magnification to quantitate the number of polymorphonuclear neutrophils and squamous epithelial cells (SECs) per low-power field and under x1,000 magnification for the presence of organisms. No organisms were seen by Gram staining in 225 (62%) of the ETAs. Culture of these specimens rarely yielded useful information: 52% were sterile, 32% grew rare to 1+ quantities of expected respiratory flora only, 12% grew rare to 1+ quantities of gram-negative rods mixed with expected respiratory flora, and only 10 (4%) yielded a pure or predominant growth of a potential respiratory pathogen. Unlike adult patients, we did not find the number of SECs to be a useful screening criterion. Only 17 (5%) of the ETAs had greater than 10 SECs per low-power field, and 5 (29%) of these yielded pure growth of a gram-negative rod. When blood culture results were positive, they correlated with ETA culture results in only 6 of 10 cases. On the basis of our findings, the absence of organisms on Gram staining is a useful criterion for rejecting ETAs from pediatric patients for culture and would have excluded 62% of the specimens from further processing.