Noninferiority Testing

  • Ton J. Cleophas
  • Aeilko H. Zwinderman
Chapter

Abstract

Traditionally, noninferiority studies have been designed to demonstrate that the efficacy of a new compound is not inferior to a standard compound documentedly efficacious. A major argument for performing noninferiority studies is, that a direct comparison versus placebo of the new compound is less ethical with an efficacious standard treatment already available, because half of the patients in such a trial is given an inferior treatment. The solution is given by a direct comparison of the new versus standard treatment. However, the comparison versus standard is at risk of establishing little difference, and, thus, a negative result. Non-inferiority studies are based on arbitrary margins of noninferiority. Generally, there are three possibilities (Fig. 63.1): (1) noninferiority is demonstrated, (2) it is uncertain, or (3) it is excluded, if the 95% confidence interval of a study is respectively (1) entirely on the right side of the margin of noninferiority, (2) crosses the margin, (3) is entirely left from the margin. The margin of noninferiority is of a rather subjective nature, and, usually, defined by expert investigators as the margin of undisputed clinical relevance (Mercola xxxx; Snapinn 2000). From Fig. 63.1 it can be easily perceived that investigators benefit from wide margins, increasing the chance of a positive study, and that, with very wide margins, it becomes virtually impossible to reject noninferiority. Scientists (Mercola xxxx; Snapinn 2000), statisticians (Hung and Wang 2004; Pocock 2003; Allen and Seaman 2007), and regulatory agencies (Anonymous 2010b) have expressed their worries about this practice. Recently, the EMEA (European Medicines Agency) has declared that noninferiority trials will not be accepted as proof of efficacy in Alzheimer’s and Parkinson’s trials, while the FDA (Food & Drug Administration) formally rejected the use of noninferiority trials in the development of antimicrobial drugs for chronic bronchitis (Anonymous 2010b).

Keywords

Standard Treatment Wide Margin Costly Repetition Noninferiority Trial Morning Peak Expiratory Flow 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

References

  1. Allen IE, Seaman CA (2007) Superiority, equivalence, and non-inferiority. Qual Prog 40(2):52–54. http://www.asq.org Accessed 15 Dec 2011Google Scholar
  2. Anonymous (2011a) European Medicines Agency guideline on the choice of the non-inferiority margin. www.emea.eu.int. Accessed 15 Dec 2011
  3. Anonymous (2011b) Scrip clinical research, new FDA and EMEA draft guidelines. www.scripnews.com. Accessed 15 Dec 2011
  4. Cleophas TJ (2000) Crossover trials should not be used to test one treatment against another treatment with a different chemical class/mode of action. J Clin Pharmacol 40:1503–1508PubMedGoogle Scholar
  5. Hansson L, Zanchetti A, Carruthers SG et al (1998) Effects of intensive blood-pressure lowering and low-dose treatment in patients with hypertension: principal results of the Hypertension Optimal Treatment (HOT) randomised trial. Lancet 351:1755–1762PubMedCrossRefGoogle Scholar
  6. Hung HMJ, Wang SJ (2004) Multiple testing of noninferiority hypotheses in active controlled trials. J Biopharm Stat 14:327–335PubMedCrossRefGoogle Scholar
  7. Kaul S, Diamond GA (2006) Good enough: a primer on the analysis and interpretation of noninferiority trials. Ann Intern Med 145:62–69PubMedCrossRefGoogle Scholar
  8. Kim JS (1997) Determining sample size for testing equivalence and noninferiority. Medical Device Link. www.devicelink.com. Accessed 15 Dec 2011
  9. Mercola J. Lying with statistics. http://randomjohn.wordpress.com. Accessed 15 Dec 2011
  10. Piaggio G, Elbourne DR, Altman DG, Pocock SJ, Evans SJ (2006) Reporting of noninferiority and equivalence randomized trials, an extension of the CONSORT Statement. JAMA 295:1152–1160PubMedCrossRefGoogle Scholar
  11. Pocock SJ (2003) The pros and cons of noninferiority trials. Fundam Clin Pharmacol 17:483–490PubMedCrossRefGoogle Scholar
  12. Snapinn SM (2000) Noninferiority trials. Curr Control Trials 1:19–21CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media B.V. 2012

Authors and Affiliations

  • Ton J. Cleophas
    • 1
    • 2
  • Aeilko H. Zwinderman
    • 1
    • 3
  1. 1.Applied to Clinical TrialsEuropean Interuniversity College of Pharmaceutical MedicineLyonFrance
  2. 2.Department of MedicineAlbert Schweitzer HospitalDordrechtNetherlands
  3. 3.Department of Biostatistics and EpidemiologyAcademic Medical CenterAmsterdamNetherlands

Personalised recommendations