Objective:To investigate and distinguish the useful index of recognizing blood culture contamination and bloodstream infection when the blood cultures were positive in clinical practice, and to help clinicians to recognize blood culture contamination.Methods:A retrospective study was conducted to analyze the clinical data and distribution of bacteria of blood culture from 303 patients with positive blood culture from January to December 2021 in Huashan Hospital, Fudan University. Age, distribution of departments, positive time of blood culture, neutrophil ratio, procalcitonin and C-reactive protein (CRP) were compared between bloodstream infection group and blood culture contamination group. Mann-Whitney U test and chi-square test were used for statistical comparisons. The factors for recognizing blood culture contamination were analyzed by multivariate logistic regression. Results:A total of 303 patients with positive blood culture were enrolled and divided into two groups, including 237 cases in bloodstream infection group and 66 cases in blood culture contamination group, with the contamination rate of 21.78%. There was statistically significant difference in the distribution of departments between the two groups ( χ2=11.28, P=0.024). The proportion of age≥60 years old (52.7%(125/237) vs 36.4%(24/66)), the neutrophil ratio(0.86(0.79, 0.91) vs 0.82(0.75, 0.88)), the levels of procalcitonin (1.05(0.23, 6.64) μg/L vs 0.17(0.11, 0.72) μg/L) and CRP (60.55(24.83, 132.83) mg/L vs 40.64(8.98, 95.83) mg/L) in the bloodstream infection group were higher than those in blood culture contamination group, and the differences were statistically significant ( χ2=5.54, U=9 523.00, 10 906.00 and 9 278.50, respectively, all P<0.05). Among the 303 patients with positive blood culture, 253 patients had a single positive blood culture, 41 patients had twice positive blood cultures and nine patients had three times positive blood cultures. The highest contamination rate of detected pathogen was coagulase-negative staphylococcus (54.72%(29/53)) in the patients with single positive blood culture. The positive time of blood culture in the bloodstream infection group was 3(3, 4) d, which was shorter than 4(3, 4) d in the blood culture contamination group, and the difference was statistically significant ( U=6 521.00, P=0.026). Multivariate logistic regression analysis showed that when the time to positivity ≤3 days and procalcitonin level ≥2.00 μg/L, the positive blood culture results were more likely to be bloodstream infection (odds ratio ( OR)=2.16, 1.96, respectively, both P<0.05). Conclusions:Bacteria associated with blood culture, time to positivity and clinical index all play important roles in recognizing blood culture contamination. When the time to positivity ≤3 days and procalcitonin level ≥2.00 μg/L, the positive blood culture results are more likely to be bloodstream infection.