Indian Journal of Pharmacology Medknow Publications on behalf of Indian Pharmacological Society ISSN: 0253-7613 EISSN: 1998-3751 Vol. 43, Num. 4, 2011, pp. 371-374

Indian Journal of Pharmacology, Vol. 43, No. 4, July-August, 2011, pp. 371-374

Educational Forum

Non-inferiority clinical trials: Practical issues and current regulatory perspective

Sandeep K Gupta

Clinical Pharmacologist, J-1044, First Floor, Palam Vihar, Gurgaon, Haryana, India

Correspondence Address: Sandeep K Gupta Clinical Pharmacologist, J-1044, First Floor, Palam Vihar, Gurgaon, Haryana India drsandeep_gupta@rediffmail.com

Date of Submission: 12-Dec-2010
Date of Decision: 18-Jan-2011
Date of Acceptance: 25-Apr-2011

Code Number: ph11102

DOI: 10.4103/0253-7613.83103

Abstract

Keywords: Assay sensitivity, bio-creep, clinical trials, non-inferiority, sample size

Introduction

Now-a-days, active-control, non-inferiority trials are being performed with an increasing frequency because it helps in finding new treatment that have approximately the same efficacy, but may offer other benefits such as better safety profile. ^[1],[2] Non-inferiority clinical trial aim to demonstrate that the test product is no worse than the comparator by more than a pre-specified, small amount. This amount is known as the non-inferiority margin (M), or delta (Δ). There are several fundamental differences between non-inferiority and superiority trials. The null hypothesis of standard superiority trials states that there is no true difference between the interventions. A type I error is the error of rejecting a null hypothesis when it is actually true. A type II error is a failure to reject the null hypothesis when in fact the alternative is true (thus, erroneously missing an actual difference in treatments). In contrast, non-inferiority trials have a null hypothesis that the experimental treatment is inferior to the standard treatment by at least a certain pre-specified amount. The alternative hypothesis to be proven is that the experimental treatment is inferior to the standard treatment by less than that same amount. Thus, the definitions of the type I and type II errors are reversed for non-inferiority trial. When evidence is lacking that the new treatment truly is inferior to the control, but non-inferiority is concluded, then it is a type I error for a non-inferiority trial and a type II error for a non-inferiority trial is the error of failing to reject the null hypothesis when the new treatment is truly non-inferior. ^[3]

Basic Concepts of Non-Inferiority Analysis

Papers by Blackwelder WC and D′Agostino RB Sr have explained the statistical algorithms for assessing non-inferiority. According to their papers, the standard null and alternative hypotheses for proving non-inferiority are following:

H ₀ : C − T ≥ M (C is superior to T)

H ₁ : C −T < M (T is not inferior to C)

Here T is the new treatment and C is the active control. M or delta (Δ) is the non-inferiority margin, that is, how much C can exceed T with T still being considered non-inferior to C (M>0). Blackwelder WC and D′Agostino RB Sr et al have further explained that to prove the null hypothesis of non-inferiority, the active control C must exceed the experimental treatment T by at least M; if this cannot be rejected, then the active control is considered superior to the experimental treatment with respect to efficacy. The alternative hypothesis of non-inferiority states that the active control may indeed have better efficacy than the experimental treatment, but by no more than M. In such a scenario, it can be proved that the investigational product is not inferior to the active control. Rejection of the null hypothesis is needed to conclude non-inferiority. ^[4],[5]

If non-inferiority is to be proved, a non-inferiority margin (M or Δ) has to be specified in the protocol of the study. After the study is completed, the lower limit of the 95% confidence interval of the difference between the test and control must not exceed the limit of delta to be 95% sure that the test drug is not worse than the control by more than delta. ^[6]

A Simplified Example to Understand the Concept

Let us consider a non-inferiority clinical trial of anti-infective drug A (test) versus B (control). From past experiences with control drug B, anticipated eradication rate is assumed to be 85% and the minimum clinically relevant difference (M or Δ) is set at 15%. It is also supposed that the results will be expressed as an estimated percentage of treatment difference with a 95% confidence interval around it. Now, let us assume that after the completion of the study, the observed eradication rates of test drug A and control drug B is found to be around 86% and 85%, respectively and the 95% confidence interval for the treatment difference is found to be between -7.5% to +9.5%. Based on the above result, a difference of -15% can be clearly ruled out of consideration and non-inferiority of test drug A can be proved relative to control drug B. But if the observed eradication rates of test drug A and control drug B is found to be around 78% and 85%, respectively and the 95% confidence interval for the treatment difference is found to be between -16.2% to +3.2%, then non-inferiority of the test drug A cannot be proved relative to control drug B, because lower limit of the 95% confidence interval of the difference between the test and control exceeds the limit of delta. ^[7]

Practical Issues and Current Regulatory Perspective

Assay sensitivity

The International Conference on Harmonization (ICH) E10 defines assay sensitivity as the property of clinical trial to distinguish an effective treatment from a less effective or ineffective treatment. The guideline further states that if a trial is intended to demonstrate the efficacy by showing a test treatment to be non-inferior to an active control, but lacks assay sensitivity, the trial may find an ineffective treatment to be non-inferior and could lead to an erroneous conclusion of efficacy. ^[8],[9],[10]

As per ICH E10, the presence of assay sensitivity in a non-inferiority or equivalence trial may be deduced from below mentioned determinations: ^[8]

1. Historical evidence of sensitivity to drug effects (HESDE), i.e., that similarly designed trials in the past regularly distinguished effective treatments from less effective or ineffective treatments. ICH E10 further states that historical evidence of sensitivity to drug effects should be evaluated before beginning a non-inferiority trial. It should be demonstrated that in the specific therapeutic area under study and in appropriately designed and conducted trials, the active treatment intended for use as the active control was reliably found superior to placebo. ^[8]
2. Appropriate trial conduct, i.e., that the conduct of the trial did not undermine its ability to distinguish effective treatments from less effective or ineffective treatments. ICH E10 further states the trial should also be conducted with high quality (e.g., good compliance, few losses to follow-up). Together with historical evidence of sensitivity to drug effects, appropriate trial conduct provides assurance of assay sensitivity in the new active control trial. ^[8]

Choice of the Non-Inferiority Margin

M or Δ is the non-inferiority margin used in the non-inferiority study. This non-inferiority margin is selected prior to the trial. Non-inferiority margin is the degree of inferiority of the test treatments to the control and the trial will attempt to exclude it statistically. If the confidence interval for the difference between the test and control treatments can exclude that the degree of inferiority of the test treatment is not greater than the non-inferiority margin, then the test treatment can be declared non-inferior. ICH E10 guideline states that the margin chosen for a non-inferiority trial cannot be greater than the smallest effect size that the active drug would be reliably expected to have compared with placebo in the setting of the planned trial. ICH E10 further states that the determination of the margin in a non-inferiority trial should be based on both statistical reasoning and clinical judgement, and should reflect uncertainties in the evidence on which the choice is based, and should be suitably conservative. ^{[2],[5],[8],[12],[13]}

Sample Size

Sample size is inversely proportional to the size of the clinically relevant difference, M or Δ. Sample sizes increase with decreasing M or Δ. At the protocol planning stage itself, the magnitude of the non-inferiority margin should be specified. Non-inferiority margin is often much smaller than the treatment difference for which a placebo-controlled trial is powered. Moreover, the sample size of a non-inferiority trial is very sensitive to the assumed effect of the new treatment relative to the active control. Non-inferiority trials typically have considerably larger sample sizes than placebo-controlled trials. US FDA guideline states that if the test drug is somewhat more effective than the control, it will be easier to rule out any given non-inferiority margin and a smaller sample size could be used. But, if test drug is less effective, then larger sample size would be required. ^{[2],[7],[11],[14],[15],[16]}

Bio-Creep

Bio-creep is the phenomenon that can occur when a slightly inferior treatment becomes the active control for the next generation of non-inferiority trials and so on until the active controls become no better than a placebo. ^[5] Factors that may influence the rate at which bio-creep occurs are following: the distribution of the effects of the new agents being tested and how that changes over time; the choice of active comparator; the method used to model the variability of the estimate of the effect of the active comparator; and changes in the effect of the active comparator from one trial to the next (violations of the constancy assumption). ^[14],[17] It has been recommended that non-inferiority trials should generally choose the best available control. ^[5]

Analysis of Non-Inferiority Trials

The Committee on Proprietary Medical Products (CPMP) guideline states that in a superiority trial, the full analysis set based on the intention-to-treat (ITT) principle is the analysis set of choice, with appropriate support provided by the PP (per protocol) analysis set. ^[18] ITT approach gives the advantage of analyses being performed by including all of the data, irrespective of any protocol violations, patient dropouts, reduced number of patient visits, and crossovers across treatment options. ^{[2],[19],[20]}

It has been argued that protocol violations and poorly conducted trials may cause results obtained from two different treatment groups to appear similar, hence ITT analysis alone is not preferred for non-inferiority trial. A possible alternative is to conduct the PP analyses where only subjects meeting the inclusion criteria are considered. But, the conservative effect of the PP analysis on non-inferiority and equivalence trials has not been thoroughly explored. Therefore, it has been suggested that non-inferiority should be concluded only if both ITT and PP analyses permit that. ^{[5],[19],[20],[21],[22],[23],[24]} CPMP guideline states that in a non-inferiority trial, the full analysis set based on the ITT principle and the PP analysis set have equal importance and their use should lead to similar conclusions for a robust interpretation. ^[18]

Switching From Superiority to Non-Inferiority

As per CPMP guideline, switching the objective of a trial from superiority to non-inferiority may be feasible provided: ^[18],[25]

1. The non-inferiority margin with respect to the control treatment was predefined or can be justified (the latter is likely to prove difficult and to be limited to rare cases where there is a widely accepted value for non-inferiority margin).
2. Analysis according to the ITT principle and PP analysis showing confidence intervals and P values for the null hypothesis of inferiority give similar findings.
3. The trial was properly designed and carried out in accordance with the strict requirements of a non-inferiority trial (as described in ICH E9 and E10).
4. The sensitivity of the trial is high enough to ensure that it is capable of detecting relevant differences if they exist.
5. There is direct or indirect evidence that the control treatment is showing its usual level of efficacy.

Ethics and Non-Inferiority Trials

US FDA guidelines state that non-inferiority design for a clinical trial is chosen when it would not be ethical to use a placebo, or a no treatment control, or a very low dose of an active drug, because there is an effective treatment that provides an important benefit (e.g., life-saving or preventing irreversible injury). ^[11]

It has been asserted by Garattini S and Bertele V ^[26] that "non-inferiority trials are unethical because they disregard patients′ interest". But, the counter argument against this assertion is that it ignores the demonstrable and continuing value of non-inferiority trials. For example, anticancer drugs with comparable efficacy but lesser safety concerns have always been sought after. The same applies for other fatal diseases wherein a new drug with the same treatment efficacy, but without iatrogenic case-fatality, would offer a major clinical benefit. Non-inferiority trials have assessed most modern-day antibiotics. Another good example of the appropriate use of a non-inferiority trial is the current research programs to develop new drugs for treatment of tuberculosis. For antibiotics, equally effective options benefit patients in clinical practice, since physicians can select a therapy on the basis of local resistance, with confidence in the treatment outcome. It has been argued that apart from better safety profile, an alternative drug with similar efficacy might offer other advantages such as easier administration, lower cost, or lesser resistance. ^{[27],[28],[29],[30],[31]}

Recent Development

In March 2010, the US FDA issued a guidance document entitled "Guidance for Industry: Non-Inferiority Clinical Trials". It was followed by release of another guidance document entitled "Guidance for Industry: Antibacterial Drug Products: Use of Non-inferiority Trials to Support Approval" in November 2010. For serious infections, such as hospital-acquired bacterial pneumonia/ventilator-associated bacterial pneumonia, community-acquired bacterial pneumonia, and acute bacterial skin and skin structure infections, US FDA recommends that it is possible to define a reliable and consistent estimate of the efficacy of active treatment relative to placebo from available data, which can serve as the basis for defining a new inferiority margin for an active-controlled non-inferiority trial. But for some indications with a high rate of resolution without antibacterial drug therapy, such as acute bacterial sinusitis (ABS), acute bacterial exacerbation of chronic bronchitis (ABECB), and acute bacterial otitis media (ABOM), US FDA recommends that available data will not support the use of an non-inferiority design, and other trial designs (i.e., superiority designs) should be used to provide evidence of effectiveness in these three indications. ^[32],[33]

Acknowledgement

I gratefully acknowledge the help of anonymous reviewers for helpful comments and suggestions.

References

Copyright 2011 - Indian Journal of Pharmacology