目的：对中药配合肝动脉栓塞化疗（Transcatheter Arterial Chemoembolization,TACE）辅助治疗原发性肝癌随机对照试验(Randomized controlled trial,RCT)文献进行评价。方法：对国内外公开发表的有关中药配合TACE辅助治疗原发性肝癌的随机对照试验进行检索，检索数据库包括Cochrane图书馆临床对照试验库，MEDLINE、CBM、CNKI和VIP电子数据库。手工检索中文相关期刊以及附加检索相关会议论文集。质量评价采用Jadad评分量表、CONSORT标准和其他自拟评价指标进行分析。结果：共纳入103篇RCTs，其中Jadad评分得5分有1篇文献，3分1篇，2分10篇，71篇1分，其余20篇均为0分，有3篇文献报告了分配隐藏。按CONSORT标准，仅1篇(0.97%)RCT描述了如何产生随机顺序，没有RCT报道如何执行随机，其中1篇为半随机，有1篇(0.97%)RCT采用安慰剂对照，有54篇(52.43%)报道了终点指标，分别各有2篇(1.94%)报道了双盲和单盲，14篇(1359%)进行了具体统计量计算，54篇(52.43%)提供了随访记录，7篇(6.80%)报道了阴性结果，16篇(15.53%)报道不良事件，仅有1篇(0.97%)进行样本含量计算与意向性分析(ITT),有2篇(1.94%)进行分层分析，属于多中心的仅有1篇(0.97%)，均未进行伦理审批和知情同意。描述了中医证型的有3篇(2.91%)。结论:目前中药配合TACE辅助治疗原发性肝癌临床研究的方法学和报告质量尚低，且多数RCT可能存在选择性偏倚和测量性偏倚，期待更多高质量的随机双盲对照试验为临床应用提供可靠的依据。
Objective To investigate the current situation of randomized controlled trials (RCTs) on compound salvia pellet (CSP) for angina pectoris and assess whether there is adequate evidence for clinical practice. Methods We collected all the published clinical studies on CSP for angina pectoris from 1994 to December 2005, and assessed each included report using the Jadad scale, the revised CONSORT statement and other self-edited items. Results We finally identified 115 RCTs. Among which, 1 scored 3 points, 6 scored 2 points, 106 scored 1 points and 2 socred 0 points. No RCT performed allocation concealment according to the CONSORT criteria, only 4 RCTs (3.5%) described the generation of the randomization sequence, among which 2 were quasi-randomized. No RCT provided randomization implementation,1 RCT (0.9%) carried out placebo control, 1 RCT (0.9%) reported endpoint, 9 RCTs (7.8%) adopted single blinding, 4 RCTs (3.5%) reported double blinding, 11 RCTs (9.6%) calculated statistical values, 2 RCTs (1.7%) provided the record of follow-up, 1 RCT (0.9%) reported negative outcome, 25 RCTs (21.8%) described adverse events, no RCT described how the sample size was estimated, and how an intent-to-treat (ITT) analysis and correlation analysis were reported, 1 RCT (0.9%) was multi-center, no RCT completed ethical approval and informed consent, 27 RCTs (23.5%) described syndrome type of TCM. Conclusion Currently, the methodology and reporting of studies on CSP for angina pectoris are not good enough to provide reliable evidence for clinical practice.
Objective To explore the quality of the reporting of randomized controlled trials (RCTs) published in the Chinese Journal of Integrated Traditional and Western Medicine from 1999 to 2004. Methods A manual search was performed and the 22 checklists of CONSORT statement and other self-established criteria were applied. Results Six volumes and 72 issues were checked. There were 1 874 clinical trials of which 1288 (68.73%) RCTs were identified in 2 765 articles. Randomization methods were described in 630 (48.92%) RCTs which showed there was more significant difference than the RCTs published in 1998 (Plt;0.001).In the 1288 RCTs, placebo control was reported only in 21 trials (1.63%) and endpoint measurements were reported in 114 trials (8.85%). Seven trials (0.55%) mentioned the estimation of sample size. None mentioned randomization concealment. Blinding was reported in 54 trials (4.2%) and P value was reported in 9 trials (0.70%).Flow chart was not mentioned in any trials. Compliance was reported in 2 trials (0.16%), and only one trial (0.08%) reported negative results. Ancillary analysis was used in 33 trials (2.57%). 10 trials are multi-centre RCTs. None of the trials reported the approval of ethics committee. Informed consent was reported in 3 trials (0.24%). Acknowledgements were mentioned in 3 trials (0.24%). Syndrome types defined by traditional Chinese medicine (TCM) were reported in 163 trials (12.66%). The criteria of intervention quality control was mentioned in 2 trials (0.15%). Double dummy was used in 8 trials (0.63%) and outcome measurement of TCM was reported in 258 trials (20.04%). All Items reported in RCTs were of low quality. Conclusions The quality of reporting of RCTs published in the Chinese Journal of Integrated Traditional and Western Medicine from 1999 to 2004 has been improved, but it does not meet the CONSORT statement.