*For medical professionals only
“img class=”responsive ” sizes=”(min-width: 320px) 320px, 100vw” src=”https://mmbiz.qpic.cn/mmbiz_png/x5F5KAyDKw19I4VvcibrfNia7lD1fial5KribXqZxjxMxtoc3ichKKz6ib3w5kJias8QNRBYGn80MM0AxEgOvRLibqE2uw/640″ width=”6400″ >National Cardiovascular Disease Management Capability Assessment and Improvement Project——CDQI (Cardiovascular Disease Quality Initiative) , which is organized by the National Cardiovascular Disease Management It is jointly carried out by the Clinical Medical Research Center and the Cardiovascular Branch of the Chinese Medical Association. The goal is to launch a new model of heart disease treatment that is oriented to the whole country, information sharing, hierarchical diagnosis and treatment, and regional coordination, and becomes an important driving force for promoting “Healthy China” in an all-round way.
April 28, 14:00-15:00, the fourth phase of the clinical research training event invited Professor Du Xin from Beijing Anzhen Hospital Affiliated to Capital Medical University as a keynote speaker. Come to the wonderful science education content! The content of the conference is now organized as follows, for readers to learn and exchange!
Professor Du Xin made a wonderful interpretation of some basic concepts of non-inferiority trials.
Figure 1
Professor Du Xin He said that the current randomized controlled trials are divided into three categories: superiority studies, non-inferiority studies, and equivalence studies. Among them, a superiority study is usually used to prove that a drug treatment is better than a traditional treatment or a placebo in order to establish a new treatment. For example, β-blockers, SGLT2 inhibitors and other heart failure treatment drugs are compared with placebo to prove that new treatment methods can bring more benefits on the basis of standard treatment, that is, superiority. A non-inferiority study is the desire to demonstrate that one treatment is non-inferior to another in order to establish an alternative treatment regimen. In the cardiovascular field, a typical non-inferiority study design is an alternative drug study for warfarin anticoagulation. The standard treatment for stroke prevention in patients with atrial fibrillation is warfarin, but there are many inconveniences in the medication process. Through non-inferiority trials, drugs with similar anticoagulant effects and more convenient medication procedures can be found.
Figure 2
Professor Du Xin uses a more popular example to explain the concept of non-inferiority. As shown in Figure 2, intravenous infusion of antibiotics has good drug utilization and good efficacy, but the disadvantage is that the medication process is inconvenient, patients must be in the hospital to receive intravenous medication, and the price of intravenous medication is high. One of its alternative treatments is oral medication, which is more convenient, but the effect of oral medication is slower. Different treatment measures have their advantages and disadvantages. Compared with intravenous medication, the advantage of drug therapy is its convenience, and the disadvantage is the loss of drug effectiveness. The evaluation of a non-inferiority study is mainly based on whether the benefits brought by the novel treatment can compensate for its inferiority to the traditional treatment, which is also a premise of the design of non-inferiority trials.
Figure 3
Prof. Du Xin explained the concept of superiority research. A superiority study is done to demonstrate that one treatment is better than another or a placebo. When comparing two treatments, you may get different results:
The new treatment is better than For traditional treatment measures, as shown in Figure 3(1), the dotted line at “0” in the coordinate axis represents that the new treatment measures are indistinguishable from the traditional treatment measures, and the distance from the blue dot to 0 represents the superiority of the new treatment measures. The research results need to be judged by the same test multiple times. Suppose 10,000 tests are carried out, and the results of 9,500 times are within a range, which is called the 95% confidence interval. is the credible interval. The range of the first black short line is on the right side of the dotted line, and the farther from the dotted line, the better the efficacy of the new drug. Even the leftmost value is still on the right side of the dotted line, and the new drug can be considered effective, representing a statistically significant difference.
The therapeutic effect of new drugs is not completely better than that of traditional drugs. As shown in Figure 3(2), most of the credible intervals are on the right side of the dotted line, and some are on the left side of the dotted line, at this time p>0.05.
There is no difference between new drug treatment and traditional drug treatment, as shown in Figure 3(3).
Figure 4 p>
The design of a superiority study requires significant results to demonstrate superiority. As shown in Figure 4, this superiority study compares different treatment strategies for ablation of atrial fibrillation, with the black line representing the simple treatment strategy and the yellow and blue lines being the complex treatment strategy. This study attempted to demonstrate that complex treatment strategies are superior to simple treatment strategies, but the results did not meet the researchers’ prespecified values, indicating that these two complex treatment strategies were not superior to simple treatment strategies. The black line represents the simple treatment strategy. Treatment strategies show higher ablation rates for AF. But can it be considered that “simple ablation strategies for atrial fibrillation are non-inferior to complex treatment strategies”? the answer is negative. The original intention of this study was to demonstrate that the complex treatment strategy was superior to the simple treatment strategy, but did not obtain a significant difference, that is, did not demonstrate the superiority of complex ablation, but did not mean that the simple strategy was not inferior to the complex strategy, because the simple strategy The advantage is simplicity, which is acceptable as long as its effect is not inferior to complex strategies. It should be noted that although the research results do not prove that the complex strategy is better, it does not mean that the complex strategy is not necessarily better than the simple strategy, that is, “absence of evidence does not constitute evidence of absence”. Proving the superiority of complex strategies and the non-inferiority of simple strategies belong to two different concepts.
Proving that one treatment strategy is not inferior to another treatment strategy, Professor Du explained it with a classic study in the cardiovascular field. Professor Du said that this classic study is a clinical trial to prove that the new oral anticoagulant is not inferior to warfarin, and several preconditions are needed to design this trial. The first premise is to prove that warfarin is effective compared to placebo. To design a randomized controlled trial, it is necessary to ensure that the controlled treatment strategy is effective. The second premise is that the endpoints used for warfarin versus placebo must be consistent with the endpoints used for new drugs versus placebo. The third premise is to define a non-inferiority margin, that is, “margin”. It is necessary to set a range within which a new drug is not inferior to a traditional drug. The non-inferiority range of a new drug generally needs to be set according to the requirements of the Food and Drug Administration. , which plays a key role in whether the study can demonstrate non-inferiority. If the non-inferiority threshold is too large, it means that non-inferiority trials are easy to obtain significant differences; otherwise, it means that the trials are more difficult and difficult to obtain significant differences.
Figure 6
Professor Du believes that the non-inferiority margin is no different from the above-mentioned superiority test, that is, it is equal to 0, the solid line at “0” represents this point, and the dotted line represents the non-inferiority margin. The trial results to the right of the solid line represent the superiority of the new treatment over the conventional treatment, and for a non-inferiority study, multiple findings represented by the blue box may be obtained. As can be seen from Figure 6, the various top-down research results are:
Case 1, the 95% confidence interval is located on the right side of the vertical line, which not only represents a new The treatment is non-inferior to the traditional treatment, and represents that the new treatment is superior to the traditional treatment;
Cases two and three, the lower limit of the 95% confidence interval is on the left side of the vertical line, but The dotted line is not exceeded, indicating that the new treatment is not completely superior to the traditional treatment, but the new treatment is not inferior to the traditional treatment.
Case 4, the lower limit of the 95% confidence interval is on the left side of the dotted line, which has exceeded the non-inferiority margin, which means that this study has not proved that the new treatment is not inferior to traditional treatment. But the upper limit of the 95% confidence interval is to the right of the vertical line, indicating that the new treatment may be better than the traditional treatment. Therefore, the results of this study cannot clarify the advantages and disadvantages of the new treatment relative to the traditional treatment, and further research is needed to expand the sample size.
Case 5, the 95% confidence interval is between the dotted line and the solid line, which means that the effect of the new treatment must not be better than the traditional treatment, but the new treatment is not inferior to the traditional treatment Treatment, the noninferiority range is the range between the solid and dashed lines.
In the sixth case, the 95% confidence interval is located on the left side of the dotted line, which means that the effect of the new treatment is worse than that of the traditional treatment, which is lower than the non-inferiority margin.
Fig.7
Professor Du took the study in Figure 7 as an example and made a more accessible explanation. The study is a non-inferiority study of edoxaban and warfarin in the treatment of venous thrombosis. The results of the study showed that the incidence of the primary end point event was 3.2% in the edoxaban group and 3.5% in the warfarin group ( HR: 0.89; 95%CI [0.70-1.13]), p<0.001, the red line in Figure 7 represents HR=1, and there is no difference in the efficacy of edoxaban and warfarin. The pre-set non-inferiority margin is 1.5, which is the position of the blue dotted line, which means that the primary endpoint event caused by edoxaban is not 50% higher than that of warfarin, and it can be considered that edoxaban has achieved non-inferiority. inspection. shown in Figure 7From the results of the study, the upper limit of the 95% confidence interval was 1.13, which did not reach the pre-specified non-inferiority margin of 1.5, that is to say, this study achieved the test of the non-inferiority margin, proving that edoxaban Non-inferior to warfarin in the prevention of deep vein thrombosis. Part of the 95% confidence interval exceeds the red line, indicating that edoxaban has a tendency to be better than warfarin. If it is necessary to prove that edoxaban is better than warfarin, the 95% confidence interval needs to be narrowed for further research. .
Fig.8
The study in Figure 8 is a non-inferiority study comparing prasugrel and clopidogrel in preventing stroke events (HR: 1.05; 95% CI [0.76-1.44]) . The distribution of the 95% confidence interval spans HR=1 (i.e. the red line). The prespecified non-inferiority margin was 1.35 (ie, the blue dashed line), which means that the primary end point event rate of prasugrel was not 35% higher than that of clopidogrel, and prasugrel was considered non-inferior to clopidogrel. As can be seen from Figure 8, the upper limit of the 95% confidence interval is 1.44, which crosses the blue dotted line, representing that this study failed to prove that prasugrel was not inferior to clopidogrel.
Fig.9<
>Professor Du Xin emphasized that there are many limitations in designing a non-inferiority trial, including that the control treatment has been compared with a placebo, and the current research endpoints are comparable to the comparison of the control treatment and placebo. Endpoints are consistent, enrolment population, background therapy must also be similar. That is to say, the new trial must be similar in all respects to the original trial of the control treatment, because only under similar conditions can the clinical effects of the control group be replicated and the efficacy of the control group guaranteed to prove that the new study is /No was not inferior to the control group.
Fig.10
The design of a non-inferiority trial requires adequate execution of the trial and determination of the results, otherwise an erroneous conclusion may be drawn. One way to achieve clear noninferiority is to administer standard care in a suboptimal manner. In the superiority test, the ITT analysis is a conservative test, that is, once randomization is performed, no matter whether the patients receive intervention treatment or not, the analysis is only based on the research results obtained, and a non-significant result may be obtained. In the non-inferiority test (NI test), the per-protocol analysis is a conservative test.
The above is Professor Du Xin’s interpretation of non-inferiority studies. Professor Du used a very vivid example to simplify the complex and lead the audience to appreciate the complexity of non-inferiority trials. He pointed out that non-inferiority trials are difficult clinical trials with many restrictions and need enough samples. amount, and the researcher must have sufficient research experience. Designing a non-inferiority trial requires more learning and thinking for beginners!
Tg1w/640″ width img class=”content_title” data-backh=”298″ data-backw=”562″ data-fileid=”503413539″ data-ratio=”0.5296296296296297″ data-type=”jpeg” data-w=”1080″ layout= “responsive” sizes=”(min-width: 320px) 320px, 100vw” src=”300,640″ srcrc=”https://mmbiz.qpic.cn/mmbiz_jpg/Or5ialKAVHBiaLgMibHE85kOe9m3yO3SDPdBZG23Py4I2sDxDanC1LBzJI6oWiaXTmCHfetVDjsuoWKHiaSArPkIsJg/640?wx_fmt=jpeg”>