Department of Hematology, Aarhus University Hospital, Aarhus, Denmark.
Department of Clinical Epidemiology, Aarhus University Hospital, Aarhus, Denmark.
Hematol Oncol. 2020 Aug;38(3):318-325. doi: 10.1002/hon.2734. Epub 2020 May 28.
Infections during first-line therapy for DLBCL are often associated with chemotherapy dose reductions and increased mortality. Systemic infections have also been suggested as beneficial promotors of immunological responses. However, whether there is an association between the timing of an infectious episode and outcome during treatment has not yet been clarified. We investigated how the occurrence and timing of infectious episodes during the first line of treatment for "de novo" DLBCL influenced patient outcome. We used data on DLBCL patients from the Danish Lymphoma Registry, the Danish National Patient Registry, and the Danish National Pathology Registry. Infections were categorized according to type (ICD-10) and time of occurrence after treatment start. "Early" infections were defined as occurring between days 7 and 42 and "late" infections between days 100 and 150 from treatment start. Patients experiencing both "early and late" infections were categorized separately. We used multivariable Cox regression and Kaplan-Meier estimates to assess the association between infections and survival adjusting for NCCN-IPI, sex, comorbidity, and rituximab treatment. We identified 3546 patients, median age 65 years (IQR 56,73). Infectious episodes occurred in 1171 (33%) patients, of which 666 had "early," 303 "late," and 202 both "early and late" events. Patients without registered infections had a 5-year overall survival (OS) rates of 74%. Those with "early," "late," or "early+late" had 5-year OS of 65%, 62%, and 53%, respectively. Compared with patients without any registered infections, hazard rate ratios (HR) were 1.24 (95% CI 1.05-1.47), 1.32 (95% CI 1.06-1.63), and 1.59 (95% CI 1.27-2.00), respectively, in the multivariable model. We observed that infectious episodes during first-line treatment for "de novo" DLBCL occurred in 44% of the patients. Irrespective of timing, patients with infectious episodes had an inferior outcome compared to those without. Outcome patterns were similar for patients registered with sepsis.
在弥漫性大 B 细胞淋巴瘤（DLBCL）的一线治疗期间发生的感染通常与化疗剂量减少和死亡率增加有关。全身性感染也被认为是免疫反应的有益促进因素。然而，感染事件的发生时间与治疗期间的结果之间是否存在关联尚未得到阐明。我们研究了“初发”DLBCL 一线治疗期间感染事件的发生和时间如何影响患者的预后。我们使用丹麦淋巴瘤登记处、丹麦国家患者登记处和丹麦国家病理登记处的 DLBCL 患者数据。感染根据类型（ICD-10）和治疗开始后发生的时间进行分类。“早期”感染定义为治疗开始后第 7 天至第 42 天发生，“晚期”感染定义为治疗开始后第 100 天至第 150 天发生。同时发生“早期和晚期”感染的患者单独分类。我们使用多变量 Cox 回归和 Kaplan-Meier 估计来评估感染与生存之间的关联，调整 NCCN-IPI、性别、合并症和利妥昔单抗治疗。我们确定了 3546 名患者，中位年龄 65 岁（IQR 56，73）。1171 名（33%）患者发生了感染事件，其中 666 名患者发生了“早期”感染，303 名患者发生了“晚期”感染，202 名患者同时发生了“早期和晚期”感染。未登记感染的患者 5 年总生存率（OS）为 74%。发生“早期”、“晚期”或“早期+晚期”感染的患者 5 年 OS 率分别为 65%、62%和 53%。与未发生任何感染登记的患者相比，多变量模型中的危险率比（HR）分别为 1.24（95%CI 1.05-1.47）、1.32（95%CI 1.06-1.63）和 1.59（95%CI 1.27-2.00）。我们观察到，“初发”DLBCL 一线治疗期间有 44%的患者发生了感染。无论时间如何，发生感染的患者的预后均劣于未发生感染的患者。发生感染的患者的结局模式与登记有败血症的患者相似。
