Résumé
This study was aimed to analyze the results of false positive reaction in bacterial detection of blood samples with BacT/ALERT 3D system, to evaluate the specificity of this system, and to decrease the false positive reaction. Each reaction flasks in past five years were processed for bacteria isolation and identification. When the initial cultures were positive, the remaining samples and the corresponding units were recultured if still available. 11395 blood samples were detected. It is worthy of note that the incubator temperature should be stabilized, avoiding fluctuation; when the cultures were alarmed, the reaction flasks showed be kept some hours for further incubation so as to trace a sharply increasing signal to support the judgement of true bacterial growth. The results indicated that 122 samples (1.07%) wee positive at initial culture, out of them 107 samples (88.7%) were found bacterial, and 15 samples (12.3%) were found nothing. The detection curves of positive samples resulted from bacterial growth showed ascent. In conclusion, maintenance of temperature stability and avoidance of temperature fluctuation in incubator could decrease the occurrence of false-positive reaction in detection process. The reaction flasks with positive results at initial culture should be recultured, and whether existence of a sharply ascending logarilhimic growth phase in bacterial growth curve should be further detected, which are helpful to distinguish false-positive reactions from true positive, and thus increase the specificity of the BacT/ALERT system.