Blood cultures detected as positive by the automated system but negative by microscopy and subculture are considered as "false-positives." Several causes have been identified, including hyperleukocytosis or the presence of fastidious bacteria, but as many cases remain unexplained we aimed to investigate the false positives occurring in our laboratory. We retrospectively collected data on blood cultures received over a period of 12months to determine factors associated with the false-positive vials. We then prospectively validated our findings on the false-positive results occurring over a 3.5-month period. We finally applied scanning electron microscopy (SEM) on 63 false positives and molecular approaches on a subset of them. In the retrospective study, 154 (85%) of the 181 false-positive identified were positive following less than 4h of incubation and were considered as "early false-positives." By performing ROC curves on these early false positives, we demonstrate that the absolute number of leukocytes is in fact the most discriminating factor of early false positivity (p < 0.001). This phenomenon can be the consequence of either a high blood culture volume (p < 0.001) or hyperleukocytosis (p < 0.001). In the prospective study, the use of a threshold of 219 million of leukocytes per vial enabled the identification of 97% of the early false positives. Finally, SEM and specific qPCR enabled three additional identifications while 16S rRNA/nanopore sequencing enabled the detection of Helicobacter cinaedi bacteremia and a polymicrobial infection. A high absolute number of leukocytes in blood cultures explains most false positives, thereby making it possible to target additional microbiological investigations.