目的：研究肝素结合蛋白（heparin-binding protein，HBP）检测在辅助诊断血流感染中的价值。方法：选取上海交通大学医学院附属瑞金医院2021年1月至2022年1月期间收治的207例血培养阳性住院患者及94例血培养阴性住院患者。检测血培养标本采集当日患者的HBP、降钙素原（procalcitonin, PCT）、C反应蛋白（C-reactive protein, CRP）水平及白细胞计数（white blood cell count, WBC）、中性粒细胞计数（absolute neutrophil count, ANC）、中性粒细胞/淋巴细胞计数比值（neutrophil-lymphocyte count ratio, NLCR）。比较血培养阳性组与阴性组间各指标水平的差异，并比较不同病原菌血流感染患者的HBP水平差异，用受试者操作特征曲线（receiver operator characteristic curve， ROC曲线）评估上述指标单独及联合检测对血流感染的辅助诊断价值。结果：血培养阳性患者的所有指标水平均显著高于血培养阴性患者。革兰阳性菌组血流感染患者的HBP水平与革兰阴性菌组患者相比，差异无统计学意义（中位数，70.26 ng/mL比89.15 ng/mL）,但二者均显著高于真菌感染患者（中位数22.90 ng/mL）（P<0.05）。诊断血流感染的ROC曲线显示，HBP的曲线下面积最大（0.742)，灵敏度为66.80%，特异度为74.40%，诊断效能显著优于WBC（0.622，灵敏度64.70%，特异度59.80%）、ANC（0.651，灵敏度51.60%，特异度75.60%）和NLCR（0.618，灵敏度72.10%，特异度48.80%）（P<0.05），且在排除恶性肿瘤或造血系统疾病患者后，HBP仍有相近的诊断效能（0.708，灵敏度60.50%，特异度78.70%）。HBP联合PCT检测是辅助诊断血流感染的最优组合，曲线下面积达0.751。结论：HBP检测对血流感染诊断有较好的参考价值。PCT是HBP联合检测诊断血流感染的优选项。

Objective: To investigate the diagnostic value of heparin-binding protein (HBP) in bloodstream infections. Methods: Two hundred and seven inpatients with positive blood culture results and 94 inpatients with negative results admitted to Ruijin Hospital, Shanghai Jiao Tong University School of Medicine from January 2021 to January 2022 were enrolled. The levels of HBP, procalcitonin (PCT), C-reactive protein (CRP), white blood cell count (WBC), absolute neutrophil count (ANC), and neutrophil-lymphocyte count ratio (NLCR) were detected on the day of blood culture specimens were collected. The levels of each biomarker between the blood culture positive group and the negative group were compared. The comparison of HBP levels in patients with bloodstream infection between different pathogens was also conducted. The receiver operator characteristic curve (ROC curve) were used to evaluate the diagnostic effectiveness of the above biomarkers and combined detection for bloodstream infection. Results: The levels of all the biomarkers were significantly elevated in patients with positive blood culture compared with those with negative results. Moreover, HBP levels in patients infected with Gram-positive bacterial were not significantly different from those infected with Gram-negative bacterial (median 70.26 ng/mL vs 89.15 ng/mL), but both of them were significantly higher than those in patients infected with fungi (median 22.90 ng/mL). The ROC curves for the diagnosis of bloodstream infections showed that HBP demonstrated a higher area under the curve (0.742, sensitivity 66.80%, specificity 74.40%), compared with WBC (0.622, sensitivity 64.70%, specificity 59.80%), ANC (0.651, sensitivity 51.60%, specificity 75.60%) and NLCR (0.618, sensitivity 72.10%, specificity 48.80%) significantly (P<0.05). After excluding patients with malignancies or hematopoietic disorders, the diagnostic value of HBP showed a similar result (0.708, sensitivity 60.50%, specificity 78.70%). The detection of combined HBP and PCT was an optimal choose for bloodstream infection, with an area under the curve of 0.751. Conclusions: HBP possesses a significant diagnostic value for bloodstream infections, and combination of HBP and PCT may be the preferred approach for accurate diagnosis.