我们旨在描述肿瘤患者中胆源性血流感染（BSIs）的特征和结果。其次，我们分析了复发BSI发作的危险因素。我们对肿瘤患者中的所有胆源性BSIs发作进行了前瞻性收集（2008-2019年）和回顾性分析，并进行了 logistic 回归分析。我们制定了一个规则将患者分成复发胆源性BSI风险群。共记录了291名肿瘤患者中400个胆源性BSIs。最常见的致病菌是大肠埃希氏菌（42%）和克雷伯菌（27%），其中86个（21.5%）是由多重耐药的革兰氏阴性杆菌（MDR-GNB）引起的。多重耐药的革兰氏阴性杆菌的比例随时间增加。总体上，73名患者发展了118次复发性BSI发作。复发性BSI发作的独立危险因素包括先前的抗生素治疗（OR 3.781，95% CI 1.906-7.503），胆道支架（OR 2.232，95% CI 1.157-4.305），因可疑胆源性感染导致的先前入院（OR 4.409，95% CI 2.338-8.311）和由MDR-GNB引起的BSI发作（OR 2.857，95% CI 1.389-5.874）。利用这些变量，我们生成了一个得分，用于预测复发性胆源性BSI，接受者操作特性（ROC）曲线下面积为0.819。在23.8%的患者中使用了不合适的经验性抗生素治疗（IEAT），30天死亡率为19.5%。总结起来，肿瘤患者中的胆源性BSI主要由革兰氏阴性杆菌引起，其中多重耐药率高且逐渐增加，不合适的经验性抗生素治疗和高死亡率。复发性BSI发作频繁。我们开发了一个简单的得分来识别复发发作，以有可能制定预防策略。重要性：本研究显示肿瘤患者中的胆源性血流感染（BSIs）主要由革兰氏阴性杆菌（GNB）引起，多重耐药率高且逐渐增加。重要的是，复发性胆源性BSI发作非常频繁，并与化疗延误、不合适的经验性抗生素治疗率高以及30天死亡率（19.5%）相关。利用与复发BSI发作相关的变量，我们生成了一个高准确性的得分，用于预测复发性胆源性BSI发作。该得分可用于制定预防策略，降低复发发作和相关病死率与发病率的风险。

We aimed to describe the characteristics and outcomes of biliary source bloodstream infections (BSIs) in oncological patients. Secondarily, we analyzed risk factors for recurrent BSI episodes. All episodes of biliary source BSIs in oncological patients were prospectively collected (2008-2019) and retrospectively analyzed. Logistic regression analyses were performed. A rule to stratify patients into risk groups for recurrent biliary source BSI was conducted. Four hundred biliary source BSIs were documented in 291 oncological patients. The most frequent causative agents were Escherichia coli (42%) and Klebsiella spp. (27%), and 86 (21.5%) episodes were caused by multidrug-resistant Gram-negative bacilli (MDR-GNB). The rates of MDR-GNB increased over time. Overall, 73 patients developed 118 recurrent BSI episodes. Independent risk factors for recurrent BSI episodes were prior antibiotic therapy (OR 3.781, 95% CI 1.906-7.503), biliary prosthesis (OR 2.232, 95% CI 1.157-4.305), prior admission due to suspected biliary source infection (OR 4.409, 95% CI 2.338-8.311), and BSI episode caused by an MDR-GNB (OR 2.857, 95% CI 1.389-5.874). With these variables, a score was generated that predicted recurrent biliary source BSI with an area under the receiver operating characteristic (ROC) curve of 0.819. Inappropriate empirical antibiotic treatment (IEAT) was administered in 23.8% of patients, and 30-d mortality was 19.5%. As a conclusion, biliary source BSI in oncological patients is mainly caused by GNB, with high and increasing MDR rates, frequent IEAT, and high mortality. Recurrent BSI episodes are frequent. A simple score to identify recurrent episodes was developed to potentially establish prophylactic strategies. IMPORTANCE This study shows that biliary source bloodstream infections (BSIs) in oncological patients are mainly caused by Gram-negative bacilli (GNB), with high and increasing rates of multidrug resistance. Importantly, recurrent biliary source BSI episodes were very frequent and associated with delays in chemotherapy, high rates of inappropriate empirical antibiotic therapy, and high 30-d mortality (19.5%). Using the variable independently associated with recurrent BSI episodes, a score was generated that predicted recurrent biliary source BSI with high accuracy. This score could be used to establish prophylactic strategies and lower the risk of relapsing episodes and the associated morbidity and mortality.