【作者】 李先涛；

【导师】 梁伟雄；

【作者基本信息】 广州中医药大学 ， 中医内科学， 2005， 博士

【摘要】 研究背景 在一个临床试验正式完成之前的任何时间内,为了比较组间的有效性和/或安全性而进行的分析称为期中分析。如果在一个新药的临床试验中,出现了研究药物的有效性已清楚,或预期的组间效应差异不可能达到,或出现了无法耐受的药物不良反应,即可做出及早终止试验的决定。目前在新药开发中,期中分析被认为是一种有力的工具。 本研究以生脉胶囊治疗慢性充血性心力衰竭患者的临床试验作为研究平台,主要解决以下三个方面的问题: 一、期中分析评价生脉胶囊治疗心衰的有效性和安全性 (一)研究目的 期中分析评价生脉胶囊治疗慢性充血性心力衰竭的有效性和安全性。 (二)试验设计 采用双盲随机安慰剂对照期中分析设计。 (三)合格受试对象的选择 1.来源:广州中医药大学第二附属医院及其它多家三级甲等医院 2.研究对象:138例均符合诊断标准、纳入标准、排除标准 (四)样本量估算 样本量估算采用左室射血分数(LVEF)为效应指标,检验效能(1-β)=0.9,总Ⅰ型错误α=0.05,双侧检验,用成组序贯设计期中分析O’ Brien & Fleming法,PASS2002版软件计算,进行2次分析共需360例。 (五)病例分组、治疗方法及用法、剂量、疗程 1.研究用药 治疗组:予口服生脉胶囊3粒,每天三次,共6个月: 对照组:予口服安慰剂胶囊3粒,每天三次,共6个月。 2.心衰标准治疗:两组给予一致的心衰标准治疗,包括利尿剂、血管紧张素转换酶更多还原

【Abstract】 BackgroundAn interim analysis is any analysis intended to compare treatment arms with respect to efficacy or safety at any time prior to formal completion of a trial. The goal of an interim analysis is to stop the trial early if the superiority of the treatment under study is clearly established, or if the demonstration of a relevant treatment difference has become unlikely, or if unacceptable adverse effects are apparent. Interim analysis design has been most widely and successfully used in large, long-term trials, and it is seen as a powerful tool in new drug development.It was acted as a platform that Shengmai capsule treatment in patients with chronic congestive heart failure in this study. It will resolve three aspects of the problems:(一)To evaluate efficacy and safety of Shengmai capsule treatment in patients with chronic congestive heart failureObjective: To evaluate efficacy and safety of Shengmai capsule treatment in patients withchronic congestive heart failure (CHF) by interim analysis.Design: Double-blind, randomized, placebo-controlled interim analysis.Selection of patients1. Setting: Second affiliated hospital, Guangzhou university of TCM and other first grade of third class hospital.2. 138 patients accord with diagnose criterion, inclusion criterion, and exclusion criterion. Sample size calculationThe efficacy variable was based on the left ventricular ejection fraction (LVEF), a total alpha of 0.05 and power of 90% and two side test, O’Brien & Fleming of the group sequential analysis and the software of PASS 2002 were chosen as interim analysis, the total of sample size were 360 patients.Intervention1. Treatment group: Patients were given an oral dose of 3 grain of Shengmai capsule, which was repeated thrice daily for 6 months.2. Control group: Patients were given an oral dose of 3 grain of placebo capsule, which was repeated thrice daily for 6 months.All patients were given the CHF’ standard therapy: a diuretic, or/and an angiotensin converting enzyme inhibitor, or/and a Beta adrenoceptor antagonist, or/and a cardiac glycoside, and treatment complications also. Observation variable1. General items: Symptoms, signs, laboratory examinations, etc.2. Efficacy variable: The New York Heart Association (NYHA) classes, transthoracic Doppler echocardiography, the scores of integral symptoms of TCM, quality of life, apply of the drugs during the trial, etc.3. Mortality and serious clinical event rates.4. Safety monitoring.5. To inspect adverse events/adverse reactions, serious adverse events/serious adverse drug reactions.6. Follow-up: Patients were visit every 2 weeks, and with regular follow-up visits at before treatment, 12 weeks and 24 weeks respectively, for a planned average follow-up of 24 weeks.Interim analysis1. Method: The O’Brien & Fleming of the group sequential analysis.2. The stopping rules: The nominal alpha level and the O’Brien & Fleming decision boundary are: a1’=0.003051, Z1=2.96259; a2’=0.049002, Z2=1.96857 separately, controlling the overall type I error probability at the conventional 0.05 level of two side test.(l)p-value<a1’, the difference has statistical significance between two groups, todecide early termination of the trial in the interim analysis. (2)p-value< a 2’ or p-value≥ a 2’, to stop the trial in the final analysis. (3)Adverse reactions will increase in anyone of two groups during trial, and terminatethe trial.3. Schedule and times of interim analyses: To plan an interim analysis when 152 of 360 patients for double-blind, randomized enrollment, and a final analysis.4. Execution of the interim analysis: To establish Four organization: the executive committee, the independent data and safety monitoring board (DSMB), the clinical trial investigators and the statistic analyses groups, will be responsible for the execution of the interim analysis.5. To issue the effect of the interim analysis: Statisticians prepare the analysis report, andpresent it to the DSMB, and register the member list of the "need-to-know" people. Statistical analysisAll data will be examined by descriptive and explore analysis. The enumeration count data were compared by the chi-square test or Fisher exact probability, the measurement data were compared by Student’s Mest, or matched Mest, or Wilcoxon test, the ranked ordinal data were compared by Wilcoxon test, the repeat measurement data were compared by repeat measures analysis of variance. Data were presented as the mean ± S.D., X2, t, z and p value, etc. Results1. Effect of screen: 138 patients accord with diagnose criterion, inclusion criterion, and exclusion criterion. The composing of sex was male 67.4%, female 32.6% respectively, The form of illnesses were coronary heart disease 43.5%, cardiac muscle disease 30.4% separately, the NYHA Ⅱ、 Ⅲ、 Ⅳ classes were 21.0%, 53.6%, 25.4% respectively.2. The cases of A group and B group were 65 cases and 73 cases separately.3. The minority effects of explore analysis were normality, the majority were non-normality.4. The baseline characteristics of the two treatment groups were similar.5. Effect of efficacy: There were no significance differences between 2 treatment groups on the NYHA classes, transthoracic Doppler echocardiography, the scores of integral symptoms of TCM, quality of life, apply of the drugs during the trial, which indicated that treatment effects did no differ between two groups (p>0. 05).6. Event rates: There were no significance differences on mortality and serious clinical event rates between 2 treatment groups (p>0. 05).7. Safety: The drug-related adverse signs and symptoms occurred in A group 4.6153%(3/65). The abnormalities of the laboratory examinations occurred no significance differences between 2 treatment groups (p>0. 05). Overall, no adverse effects were serious enough to discontinue double-blind treatment during the maintenance period. All these patients were successfully treated and discharged.Conclusions: The first interim analysis did not indicated to have differences between two groups, and can not make decisions to terminate the trial early, and to recommend to proceed with this trail.(二) The methodoIogy for performing an interim analysis in the cIinical trial of traditional Chinese medicineThe interim analysis plan1. A standard operating procedure providing for(l)The possibility of an early termination of the trial as a result of the planned interimanalysis; (2)The introduction of unscheduled interim analyses into an ongoing trial withdescriptions of authority, request for the analysis, and what actions could result; (3) Accessibility and safeguarding of patient treatment assignment codes and identifyingwhich patient data will be unblinded; (4)Documentation of all procedures followed, actions, and decisions made that couldaffect interpretation of results or conduct of the trial.2. A prospective interim analysis plan within the protocol indicating timing, methods to control type I error probability, and the decision rule describing the expected consequences of various outcomes of the interim analysis. (l)The schedule and times of interim analyses.(2)The stopping rules reflect one of the following conditions: ①There is clear evidence of harm or harmful side-effects of the treatment; ②There is no likelihood of demonstrating treatment benefit; or ③There is overwhelming evidence of the benefit of the treatment.(3)The methods of interim analyses. The structure of interim analysisThe structure of interim analysis includes blinding and unblinding parts. Four key organizations are necessary for the successful conduct of an interim analysis. These are the executive committee, the DSMB, the clinical trial investigators and the statistic analyses groups. The execution of an interim analysisThe DSMB would be responsible for the execution of the interim analysis; communication of information would follow the "need-to-know" principle. Members of the DSMB were to be selected strictly by credentials and expertise. DSMB meetings were to consist of open and closed sessions. The DSMB was required to keep detailed meeting minutes. These minutes were to be kept in a confidential file inaccessible to the sponsors. Statistical and programming processThe analysis of the data must be performed on a static version of the database, which must be archived to provide a clear audit trail as well as to allow for future analyses. Terminate trialThe sponsor should launch the news of early stop the trial without delay.(三) Efficacy endpoint selection methods and technology issues for chronic congestive heart failure with shengmai capsule treatmentMethods and selection of endpointsThe selection of the best response variables for the assessment of the efficacy of atreatment in congestive heart failure patients is: ①Mortality as an endpoint; ②Rate of (re-)hospitalization as an endpoint; ③Functional status; ④Quality of life; ⑤Worsening heart failure; ⑥Objective evaluation of functional capacity; ⑦Composite variables, etc. Evaluation technology issues1. One primary efficacy endpointThe ideal situation is one that permits characterization of the disease and patients’ response by a single primary endpoint, but this is impossible in many situations,2. Multiple primary efficacy endpointsBesides the clinical desire to demonstrate therapeutic effect in several endpoints, disease complexity or lack of a clinical consensus on the most important clinical endpoint often necessitates the design of clinical trials with two or more primary endpoints.3. Composite primary efficacy endpointIn some therapeutic areas where the disease manifests itself in a multi-faceted manner, drug effectiveness is often characterized by the use of composite endpoints. Clinical and statistical expectation1. Clinical and statistical expectation under a single primary endpointFor a clinical trial with a single primary endpoint, there is a single decision-making process. Drug effectiveness is said to be demonstrated only if a clinically meaningful treatment effect size (as pre-specified in the protocol) is realized and statistical significance is shown at a prospectively specified nominal α -level with respect to the declared single primary endpoint. All therapeutic efficacy claims are limited to the efficacy evidence provided by the primary endpoint.2. Clinical and statistical expectation under multiple endpointsFor a clinical trial with two or more primary endpoints, there are at least three possible clinical decision-making scenarios regarding drug effectiveness. That is: ①statistical significance is demonstrated at the pre-specified nominal α -level for all primary endpoints, or ②statistical significance is demonstrated at the pre-specified nominal α -level for the majority (where what constitutes the majority is prospectively defined in the protocol) of the primary endpoints, or ③statistical significance is demonstrated at the pre-specified nominal α -level for one or more of the primary endpoints.3. Clinical and statistical expectations for composite endpointsThe statistical expectation of overall statistical significance is demonstrated at a pre-specified α -level for composite endpoints. Multiple endpoint adjustment strategies for controlling type I error rateIf the objective of the clinical trial is to demonstrate therapeutic effect using several independent, dissimilar, or lowly correlated (0≤p≤0.1) clinical primary endpoints and the endpoints of interest are not too many, endpoint specific conclusions can usually be made by the direct use of univariate adjustment methods. If therapeutic effect is to be更多还原