贝叶斯网状Meta分析预测头孢菌素类抗菌药物联合甲硝唑治疗腹腔感染的疗效和安全性

崔瑞霞1; 王子1; 曲凯1; 苗润晨1; 张靖垚1; 任建安2; 刘昌1

doi:10.3760/cma.j.cn115610-20200914-00613

贝叶斯网状Meta分析预测头孢菌素类抗菌药物联合甲硝唑治疗腹腔感染的疗效和安全性

Efficacy and safety prediction of cephalosporins antibiotics combined with metronidazole for intra-abdominal infections: a Bayesian network meta analysis

摘要

摘要: 目的:应用贝叶斯网状Meta分析预测头孢菌素类抗菌药物联合甲硝唑经验性治疗腹腔感染的疗效和安全性。
方法:以(“intraabdominal infections”［MeSH Terms］)AND(“Cephalosporins*”［MeSH Terms］)AND(“randomized controlled trial”［MeSH Terms］)、腹腔感染、继发性腹膜炎、腹腔脓肿、头孢为检索词，检索PubMed、Embase、the Cochrane Library、中国知网、万方数据库、维普数据库。检索时间为1990年1月至2018年5月。纳入评价头孢菌素类抗菌药物联合甲硝唑对比其他类型抗菌药物治疗腹腔感染疗效和安全性的随机对照试验(RCT)。试验组患者给予头孢菌素类抗菌药物联合甲硝唑治疗腹腔感染，对照组患者给予其他类型抗菌药物治疗腹腔感染。结局指标：临床治愈率、微生物清除率和药物严重不良反应发生率。使用R 3.6.2软件随机贝叶斯模型进行网状Meta分析。采用马尔可夫链蒙特卡罗法直接评估和间接预测。采用轨迹图法、密度图法、杠杆图法对模型收敛性和稳定性进行评价。因各干预措施间没有闭环，不需要评估研究的一致性。
结果:(1)文献检索结果：最终纳入符合标准的相关文献18篇。18篇文献均为RCT，文献累积样本量为6 792例，其中试验组3 402例，对照组3 390例。(2)贝叶斯网状Meta分析结果：①第3代头孢菌素+甲硝唑、碳青霉烯类的临床治愈率显著低于第4代头孢菌素+甲硝唑，差异均有统计学意义［优势比(OR)=0.46，0.61，95%可信区间(CI)为0.26～0.81，0.38～0.97，P<0.05］。第5代头孢菌素+甲硝唑的临床治愈率与碳青霉烯类比较，差异无统计学意义(OR=1.03，95%CI为0.59～1.80，P>0.05)。②第 5代头孢菌素+甲硝唑、碳青霉烯类的微生物清除率显著低于第4代头孢菌素+甲硝唑，差异均有统计学意义(OR=0.84，0.41，95%CI为0.73～0.98，0.23～0.74，P<0.05)。第5代头孢菌素+甲硝唑的微生物清除率与碳青霉烯类比较，差异无统计学意义(OR=0.76，95%CI为0.27～1.80，P>0.05)。③第3代头孢菌素+甲硝唑、第4代头孢菌素+甲硝唑、第5代头孢菌素+甲硝唑、碳青霉烯类、喹诺酮+甲硝唑、替加环素的药物严重不良反应发生率均显著低于喹诺酮类，差异均有统计学意义(OR=0.13，0.13，0.14，0.13，0.15，0.13，95%CI为0.03～0.50，0.02～0.98，0.02～0.75，0.02～0.59，0.02～0.78，0.02～0.57，P<0.05)。第3代头孢菌素+甲硝唑、第4代头孢菌素+甲硝唑、第5代头孢菌素+甲硝唑的药物严重不良反应发生率与碳青霉烯类比较，差异均无统计学意义(OR=0.96，1.00，1.10，95%CI为0.52～1.60，0.31～3.50，0.49～2.30，P>0.05)。(3)疗效与安全性等级排名：①临床治愈率排名：各类用药方案的临床治愈率从高到低排名顺序为喹诺酮类+甲硝唑、第4代头孢菌素+甲硝唑、合成青霉素类、第2代头孢菌素+甲硝唑、第 5代头孢菌素+甲硝唑、碳青霉烯类、第3代头孢菌素+甲硝唑、替加环素、喹诺酮类。各排名相应的概率为51.73%、35.72%、22.57%、31.37%、24.98%、32.82%、34.69%、29.05%、72.36%。②微生物清除率排名：各类用药方案的微生物清除率从高到低排名顺序为喹诺酮类+甲硝唑、第4代头孢菌素+甲硝唑、第2代头孢菌素+甲硝唑、合成青霉素类、第5代头孢菌素+甲硝唑、碳青霉烯类、第3代头孢菌素+甲硝唑、替加环素、喹诺酮类。各排名相应的概率为：89.62%、77.01%、38.60%、20.94%、26.26%、26.39%、22.22%、20.19%、62.55%。③药物严重不良反应发生率排名：各类用药方案的药物严重不良反应发生率从高到低排名顺序为喹诺酮类、喹诺酮类+甲硝唑、第5代头孢菌素+甲硝唑、碳青霉烯类、第3代头孢菌素+甲硝唑、替加环素、第4代头孢菌素+甲硝唑。各排名相应的概率为：96.21%、30.46%、21.09%、25.27%、27.26%、19.45%、31.69%。
结论:中、低危腹腔感染时，推荐经验性使用头孢菌素+甲硝唑替代经验性使用碳青霉烯类。

Abstract: Objective:To predict the efficacy and safety of cephalosporins antibiotics combined with metronidazole for intra-abdominal infections using Bayesian network meta analysis.
Methods:Databases including PubMed, Embase, the Cochrane Library, CNKI, Wanfang database, VIP database were searched for literatures from January 1990 to May 2018 with the key words of (“intraabdominal infections” ［MeSH Terms］) AND (“Cephalosporins*” ［MeSH Terms］) AND (“randomized controlled trial” ［MeSH Terms］), 腹腔感染, 继发性腹膜炎, 腹腔脓肿, 头孢. The randomized controlled trials (RCTs) about comparison of efficacy and safety between cephalosporins antibiotics combined with metronidazole versus other antibiotics for intra-abdominal infections were received and included. Experimental group included patients who received cephalosporins antibiotics combined with metronidazole for intra-abdominal infections, and control group included patients who received other antibiotics for intra-abdominal infections. The primary outcomes were the clinical cure rates, microbial clearance rate and incidence of serious adverse drug reactions. R 3.6.2 software random Bayesian model was used for meta analysis. The Markov Chain Monte Carlo was used for direct evaluation and indirect prediction. The tracing method, density plotting and leverage figure method were used to evaluate the model convergence and stability. No closed loop formed between intervention measures, so there was no need to evaluate consistency.
Results:(1) Document retrieval: a total of 18 available RCTs were enrolled. There were 6 792 patients, including 3 402 in the experimental group, 3 390 in the control group. (2) Results of Bayesian network meta analysis. ① The clinical cure rates of the third generation cephalosporins+metronidazole, carbapenems were significantly lower than the fourth generation cephalosporins+metronidazole ［odds ratio (OR)=0.46, 0.61, 95% confidence interval(CI) as 0.26-0.81, 0.38-0.97, P<0.05］. There was no significant difference in the clinical cure rate between the fifth generation cephalosporins+metronidazole and carbapenems (OR=1.03, 95%CI as 0.59-1.80, P>0.05). ② The microbial clearance rates of the fifth generation cephalosporins+metronidazole, carbapenems were significantly lower than the fourth generation cephalosporins+metronidazole (OR=0.84, 0.41, 95%CI as 0.73-0.98, 0.23-0.74, P<0.05). There was no significant difference in the microbial clearance rate between the fifth generation cephalosporins+metronidazole and carbapenems (OR=0.76, 95%CI as 0.27-1.80, P>0.05). ③ The incidence of serious adverse drug reactions was significantly lower for the third generation cephalosporins+metronidazole, the fourth generation cephalosporins antibiotics+cetronidazole, the fifth generation cephalosporins+metronidazole, carbapenems, quinolones+metronidazole, and tigecycline than for quinolones (OR=0.13, 0.13, 0.14, 0.13, 0.15, 0.13, 95%CI as 0.03-0.50, 0.02-0.98, 0.02-0.75, 0.02-0.59, 0.02-0.78, 0.02-0.57, P<0.05). Compared with carbapenems, the third generation cephalosporins+metronidazole, the fourth generation cephalosporins+metronidazole, the fifth generation cephalosporins+metronidazole had no significant difference in the incidence of serious adverse drug reactions (OR=0.96, 1.00, 1.10, 95%CI as 0.52-1.60, 0.31-3.50, 0.49-2.30, P>0.05). (3) Ranking of the efficacy and safety. ① The ranking list for clinical cure rates of different therapeutic regimens showed from high to low as quinolones+metronidazole, the fourth generation cephalosporins+metronidazole, synthetic penicillins, the second generation cephalosporins+metronidazole, the fifth generation cephalosporins+metronidazole, carbapenems, the third generation cephalosporins+metronidazole, tigecycline, quinolones. The corresponding ranking probabilities of above regimens were 51.73%, 35.72%, 22.57%, 31.37%, 24.98%, 32.82%, 34.69%, 29.05%, 72.36%, respectively. ② The ranking list for microbial clearance rates of different therapeutic regimens showed from high to low as quinolones+metronidazole, the fourth generation cephalosporins+metronidazole, the second generation cephalosporins+metronidazole, synthetic penicillins, the fifth generation cephalosporins+metronidazole, carbapenems, the third generation cephalosporins+metronidazole, tigecycline, quinolones. The corresponding ranking probabilities of above regimens were 89.62%, 77.01%, 38.60%, 20.94%, 26.26%, 26.39%, 22.22%, 20.19%, 62.55%, respectively. ③ The ranking list for incidence of serious adverse drug reactions of different therapeutic regimens showed from high to low as quinolones, quinolones+metronidazole, the fifth generation cephalosporins+metronidazole, carbapenems, the third generation cephalosporins+metronidazole, tigecycline, the fourth generation cephalosporins+metronidazole. The corresponding ranking probabilities of above regimens were 96.21%, 30.46%, 21.09%, 25.27%, 27.26%, 19.45%, 31.69%, respectively.
Conclusion: In the treatment of middle-and low-risk intra-abdominal infections, it is recommended to empirically use cephalosporins+metronidazole instead of carbapenems.

HTML全文

参考文献(0)

施引文献

资源附件(0)