What does a non-inferior result mean in a clinical trial?

Understanding Non-Inferiority Results in Clinical Trials

A non-inferiority result in a clinical trial means that a new treatment is not clinically meaningfully worse than a standard treatment, based on a pre-specified acceptable margin of difference (the non-inferiority margin). 1

Key Concepts in Non-Inferiority Trials

• Non-inferiority trials are designed to demonstrate that a new treatment is not worse than a standard treatment by more than a pre-specified margin, while potentially offering other advantages such as fewer side effects, easier administration, or lower cost 1

• The non-inferiority margin (often called "delta") is the smallest treatment difference that, if exceeded, would mean the new treatment is considered inferior 1

• The conclusion of non-inferiority depends on where the upper bound of the 95% confidence interval (CI) for the treatment difference falls in relation to this margin 1

• If the upper bound of the 95% CI is less than the pre-specified delta, non-inferiority can be claimed 1

Example of Non-Inferiority Assessment

• In the ACUITY trial comparing bivalirudin with standard heparin plus glycoprotein IIb/IIIa inhibitor in acute coronary syndrome patients:
  • The non-inferiority margin was set at a relative risk of 1.25 1
  • Results showed composite ischemia rates of 7.8% vs. 7.3% (relative risk: 1.08; 95% CI: 0.93-1.24) 1
  • Since the upper bound of the CI (1.24) was less than the pre-specified delta (1.25), bivalirudin was declared non-inferior 1
  • This was clinically important because bivalirudin also had lower bleeding risk 1

Common Misunderstandings

• A common error is assuming that lack of statistical significance between treatments (p > 0.05) automatically implies equivalence or non-inferiority 1

• For example, the INSIGHT trial comparing nifedipine with co-amilozide in hypertension found no significant difference (p = 0.35), but the 95% CI included up to a 34% excess risk with nifedipine, making it inappropriate to conclude non-inferiority 1

Interpretation Framework

• Non-inferiority results can be categorized into several scenarios:
  • Non-inferior and superior (when the entire CI is above zero and below delta)
  • Non-inferior but not superior (when the CI crosses zero but stays below delta)
  • Inconclusive (when the CI crosses both zero and delta)
  • Inferior (when the CI is entirely above delta) 1

Statistical Considerations

• Non-inferiority trials reverse the traditional null and alternative hypotheses 1

  • The null hypothesis is that the new treatment is inferior by at least the margin
  • The alternative hypothesis is typically that treatments are equivalent or that the difference is less than the margin

• Both intention-to-treat (ITT) and per-protocol (PP) analyses are recommended for non-inferiority trials, as ITT analysis alone may bias toward finding non-inferiority 1

Potential Pitfalls

• The choice of non-inferiority margin is subjective and can significantly impact trial conclusions 1

• Poor quality trials tend to favor non-inferior results 1

• Successive non-inferiority trials may introduce progressively less effective treatments while each is technically "non-inferior" to the previous one 1

• Post-hoc switching from superiority to non-inferiority testing without pre-specified margins can lead to biased conclusions 1

Clinical Implications

• When evaluating non-inferiority trial results, clinicians should:
  • Examine the choice and justification of the non-inferiority margin 1
  • Consider both relative and absolute differences between treatments 1
  • Focus on the effect size and confidence limits rather than just p-values 1
  • Evaluate all relevant outcomes, including safety and quality of life measures 1