Intermountain Healthcare, Division of Epidemiology and Infectious Disease, Salt Lake City, Utah, USA.
Stanford University, Division of Infectious Diseases and Geographic Medicine, Palo Alto, California, USA.
Clin Infect Dis. 2020 Aug 14;71(4):960-967. doi: 10.1093/cid/ciz900.
Antibiotic stewardship is challenging in hematological malignancy patients.
We performed a quasiexperimental implementation study of 2 antimicrobial stewardship interventions in a hematological malignancy unit: monthly antibiotic cycling for febrile neutropenia that included cefepime (± metronidazole) and piperacillin-tazobactam and a clinical prediction rule to guide anti-vancomycin-resistant Enterococcus faecium (VRE) therapy. We used interrupted time-series analysis to compare antibiotic use and logistic regression in order to adjust observed unit-level changes in resistant infections by background community rates.
A total of 2434 admissions spanning 3 years pre- and 2 years postimplementation were included. Unadjusted carbapenem and daptomycin use decreased significantly. In interrupted time-series analysis, carbapenem use decreased by -230 days of therapy (DOT)/1000 patient-days (95% confidence interval [CI], -290 to -180; P < .001). Both VRE colonization (odds ratio [OR], 0.64; 95% CI, 0.51 to 0.81; P < .001) and infection (OR, 0.41; 95% CI, 0.2 to 0.9; P = .02) decreased after implementation. This shift may have had a greater effect on daptomycin prescribing (-160 DOT/1000 patient-days; 95% CI, -200 to -120; P < .001) than did the VRE clinical prediction score (-30 DOT/1000 patient-days; 95% CI, -50 to 0; P = .08). Also, 46.2% of Pseudomonas aeruginosa isolates were carbapenem-resistant preimplementation compared with 25.0% postimplementation (P = .32). Unit-level changes in methicillin-resistant Staphylococcus aureus and extended-spectrum beta lactamase (ESBL) incidence were explained by background community-level trends, while changes in AmpC ESBL and VRE appeared to be independent. The program was not associated with increased mortality.
An antibiotic cycling-based strategy for febrile neutropenia effectively reduced carbapenem use, which may have resulted in decreased VRE colonization and infection and perhaps, in turn, decreased daptomycin prescribing.
抗生素管理在血液恶性肿瘤患者中具有挑战性。
我们在血液恶性肿瘤病房进行了两项抗菌药物管理干预的准实验实施研究:每月发热性中性粒细胞减少症的抗生素循环,包括头孢吡肟(±甲硝唑)和哌拉西林他唑巴坦,以及指导抗万古霉素耐药粪肠球菌(VRE)治疗的临床预测规则。我们使用中断时间序列分析来比较抗生素的使用和逻辑回归,以便通过背景社区比率调整观察到的单位水平变化的耐药感染。
共纳入 3 年前和 2 年后实施的 2434 例住院患者。未调整的碳青霉烯类和达托霉素的使用明显减少。在中断时间序列分析中,碳青霉烯类药物的使用减少了 230 天的治疗剂量(DOT)/1000 患者天(95%置信区间[CI],-290 至-180;P<.001)。VRE 定植(比值比[OR],0.64;95%CI,0.51 至 0.81;P<.001)和感染(OR,0.41;95%CI,0.2 至 0.9;P=0.02)均减少。这一转变可能对达托霉素的处方产生了更大的影响(-160DOT/1000 患者天;95%CI,-200 至-120;P<.001),而 VRE 临床预测评分的影响较小(-30DOT/1000 患者天;95%CI,-50 至 0;P=0.08)。此外,实施前有 46.2%的铜绿假单胞菌分离株对碳青霉烯类药物耐药,而实施后为 25.0%(P=0.32)。耐甲氧西林金黄色葡萄球菌和超广谱β-内酰胺酶(ESBL)发生率的单位水平变化由背景社区水平趋势解释,而 AmpC ESBL 和 VRE 的变化似乎是独立的。该方案与死亡率的增加无关。
基于发热性中性粒细胞减少症的抗生素循环策略有效地减少了碳青霉烯类药物的使用,这可能导致 VRE 定植和感染减少,从而可能减少达托霉素的使用。