Office of Antimicrobial Products, Center for Drug Evaluation and Research, US Food and Drug Administration, Silver Spring, Maryland 20993, USA.
Clin Infect Dis. 2010 Aug 15;51(4):381-9. doi: 10.1086/655131.
On the basis of meta-analyses, concern has been raised regarding a possible signal of increased mortality associated with the use of cefepime versus other beta-lactam antibiotics. To further investigate this possible signal, we accessed findings and data from published and unpublished cefepime clinical trials.
We performed meta-analyses using trial- and patient-level data from comparative trials. Trial-level analyses were performed using summary data from all patients in the trials, and patient-level analyses were performed on trials for which patient-level data were available. Thirty-day, all-cause mortality was analyzed using the Mantel-Haenszel adjusted risk difference (ARD) method.
The trial-level meta-analysis was based on 88 trials (9467 cefepime patients and 8288 comparator patients). The 30-day, all-cause mortality rates were 6.21% (588/9467) for the cefepime patients and 6.00% (497/8288) for comparator patients (ARD per 1000 population, 5.38; 95% confidence interval [CI], -1.53 to 12.28). In the patient-level analysis (35 trials, 5058 cefepime patients, and 3976 comparator patients), 30-day, all-cause mortality rates were 5.63% (285/5058) for cefepime patients and 5.68% (226/3976) for comparator patients (ARD per 1000 population, 4.83; 95% CI, -4.72 to 14.38). A sensitivity analysis based solely on the 24 febrile neutropenia trials did not show a statistically significant increase in mortality with cefepime use (ARD per 1000 population, 9.67; 95% CI, -2.87 to 22.21).
In both trial-level and patient-level meta-analyses, we did not identify a statistically significant increase in mortality among cefepime-treated patients, compared with those treated with other antibacterials.
基于荟萃分析，人们对头孢吡肟与其他β-内酰胺类抗生素相比可能存在死亡率增加的信号表示担忧。为了进一步研究这一可能的信号，我们查阅了已发表和未发表的头孢吡肟临床试验的结果和数据。
我们使用来自对照试验的试验和患者水平数据进行荟萃分析。试验水平分析使用试验中所有患者的汇总数据进行，患者水平分析使用可获得患者水平数据的试验进行。使用 Mantel-Haenszel 调整风险差异 (ARD) 方法分析 30 天全因死亡率。
试验水平荟萃分析基于 88 项试验（头孢吡肟患者 9467 例，对照组患者 8288 例）。头孢吡肟患者的 30 天全因死亡率为 6.21%（588/9467），对照组患者为 6.00%（497/8288）（每 1000 人人群 ARD 为 5.38；95%置信区间 [CI]，-1.53 至 12.28）。在患者水平分析（35 项试验，头孢吡肟患者 5058 例，对照组患者 3976 例）中，头孢吡肟患者的 30 天全因死亡率为 5.63%（285/5058），对照组患者为 5.68%（226/3976）（每 1000 人人群 ARD 为 4.83；95%CI，-4.72 至 14.38）。基于仅 24 项发热性中性粒细胞减少症试验的敏感性分析并未显示头孢吡肟使用与死亡率增加之间存在统计学显著关联（每 1000 人人群 ARD 为 9.67；95%CI，-2.87 至 22.21）。
在试验水平和患者水平荟萃分析中，与接受其他抗菌药物治疗的患者相比，头孢吡肟治疗的患者死亡率均未出现统计学显著增加。
