摘要目的 评价肿瘤坏死因子α(TNF-α)抑制剂单药疗法致感染的风险. 方法 检索PubMed、Embase、Cochrane图书馆和Web of Science数据库(建库至2014年11月),收集使用TNF-α抑制剂单药疗法(试验组)与安慰剂或阳性对照药物(对照组)比较的随机双盲对照试验英文文献.采用Jadad量表评价纳入研究的质量(评分＜3分为低质量,3～5分为高质量).采用Revman 5.2软件进行Meta分析,感染发生率和严重感染发生率分别以相对危险度(RR)和Peto比值比(Peto OR)为效应量,并计算其95％置信区间(CI). 结果 共32篇文献报道的33项研究纳入Meta分析,患者共11 819例,试验组(应用阿达木单抗、格列木单抗、英夫利西单抗或依那西普)7 408例,对照组(应用安慰剂、阳性对照药物甲氨蝶呤或柳氮磺吡啶)4 411例.33项研究Jadad评分均≥3分.Meta分析结果显示,试验组患者感染总发生率高于安慰剂对照组[33.03％(1 702/5 153)比29.53％(873/2 956),RR=1.17,95％ CI为1.09 ～ 1.25,P＜0.000 01],与阳性药物对照组比较差异无统计学意义[50.1％(362/723)比48.3％ (320/662),RR=1.10,95％ CI为0.90 ～ 1.34,P=0.36];严重感染发生率与安慰剂对照组和阳性药物对照组比较差异均无统计学意义[1.4％ (73/5 067)比1.7％ (48/2 902), Peto OR=0.90,95％ CI为0.61 ～ 1.32,P =0.46;2.4％ (34/1 410)比2.8％ (28/976) ,RR=1.10,95％ CI为0.90～1.34,P=0.36].亚组分析显示,应用阿达木单抗、格列木单抗与英夫利西单抗的试验组患者感染发生率均高于安慰剂对照组[41.4％ (568/1 373)比39.1％ (361/923),RR=1.11,95％ CI为1.01 ～1.23,P=0.04;25.5％ (397/1 558)比19.2％(120/625),RR=1.22,95％ CI为1.02 ～ 1.45,P=0.03;38.2％ (297/777)和27.7％(114/411),RR=1.35,95％CI为1.13 ～ 1.61,P=0.001],应用依那西普的试验组患者感染发生率与安慰剂对照组比较差异无统计学意义[30.4％(440/1 445)比27.9％ (278/997);RR=1.13,95％ CI为0.99～ 1.28,P=0.06];应用格列木单抗的试验组患者严重感染发生率低于安慰剂对照组[0.59％ (8/1 355)比2.12％ (11/520),Peto OR=0.21,95％ CI为0.08 ～0.59,P=0.003],应用其他3种药物的试验组患者严重感染发生率与安慰剂对照组比较差异均无统计学意义(均P ＞0.05). 结论 TNF-α抑制剂单药疗法可增加患者发生感染的风险,但未增加发生严重感染的风险,临床使用较为安全.

abstractsObjective To evaluate the risk of infection in monotherapy of tumor necrosis factor alpha (TNF-α) inhibitor.Methods The database of Pubed, Embase, Cochrane library, and Web with Science were searched from the inception to November, 2014.The literatures of randomized controlled trials in English which included reports that only used TNF-α inhibitor (the test group) and placebo or positive controlled drug (the control group) were selected.The methodological quality of the literatures which enrolled into the study were assessed by Jadad scale (inferior quality: ＜ 3 points, high quality: 3-5 points).The software RevMan 5.2 was used for Meta-analysis.The infection rate and the severe infection rate were expressed by relative risk (RR), Peto odds ratio (Peto OR)and 95 ％ confidence interval (CI).Results A total of 33 trials presented by 32 reports, and 11 819 patients (7 408 cases in the test group using adalimumab or golimumab or infliximab or etanercept, respectively, and 4 411 cases in the control group using placebo, positive control drug such as methotrexate or salazosulfapyridine, respectively) were enrolled into the Meta-analysis.The Jadad scores of the 33 trials were all ≥ 3 points.The results of the Meta-analysis showed that the overall total incidence of infection in patients who used TNF-α inhibitors only was higher than that in the patients who used placebo [33.03％ (1 702/5 153) vs.29.53％ (873/2 956), RR =1.17, 95％ CI: 1.09-1.25, P ＜0.000 01].There was no significant difference in the overall incidence of infection between the test group and the positive controlled drug group [50.1％ (362/723)vs.48.3％ (320/662), RR =1.10, 95％ CI: 0.90-1.34, P =0.36].There were no significant differences in the incidence of severe infection between the test group and the placebo control group, the positive controlled drug group [1.4％ (73/5 067)vs.1.7％ (48/2 902), Peto OR =0.90, 95％ CI: 0.61-1.32, P =0.46;2.4％ (34/1 410) vs.2.8％ (28/976), RR =1.10, 95％ CI: 0.90-1.34, P =0.36].The results of subgroup analysis showed that the incidence of infection in patients who used adalimumab, or golimumab or infliximab only were significantly higher than that in the patient who used placebo [41.4％ (568/1 373) vs.39.1％ (361/923), RR =1.11,95％ CI: 1.01-1.23, P=0.04;25.5％ (397/1 558) vs.19.2％ (120/625), RR: 1.22, 95％CI: 1.02-1.45, P=0.03;38.2％ (297/777) vs.27.7％ (114/411), RR=1.35, 95％ CI: 1.13-1.61, P =0.001].There was no significant difference in the incidence of infection between the patients who used etanercept and the patients who used placebo [30.4％ (440/1 445) vs.27.9％ (278/997), RR =1.13, 95％ CI: 0.99-1.28, P =0.06].The incidence of severe infection in patients who used golimumab was significantly lower than that in the patient who used placebo [0.59％ (8/ 1 355) vs.2.12％ (11/520), Peto OR =0.21, 95％CI: 0.08-0.59, P=0.003].There were no significant differences in the incidence of severe infection between the patients who used the other 3 kinds of drugs and the patients who used placebo (all P ＞ 0.05).Conclusions The monotherapy with TNF-α inhibitor may increase the overall incidence of infection, but will not increase the incidence of severe infection.The monotherapy with TNF-α inhibitor is relatively safety in clinical practice.