Justification, margin values, and analysis populations for oncologic noninferiority and equivalence trials: a meta-epidemiological study

Abstract

Background: Noninferiority and equivalence trials evaluate whether an experimental therapy's effect on the primary endpoint is contained within an acceptable margin compared with standard of care. The reliability and impact of this conclusion, however, is largely dependent on the justification for this design, the choice of margin, and the analysis population used.

Methods: A meta-epidemiological study was performed of phase 3 randomized noninferiority and equivalence oncologic trials registered at ClinicalTrials.gov. Data were extracted from each trial's registration page and primary manuscript.

Results: We identified 65 noninferiority and 10 equivalence trials that collectively enrolled 61 632 patients. Of these, 61 (81%) trials demonstrated noninferiority or equivalence. A total of 65 (87%) trials were justified in the use of a noninferiority or equivalence design either because of an inherent advantage (53 trials), a statistically significant quality-of-life improvement (6 trials), or a statistically significant toxicity improvement (6 trials) of the interventional treatment relative to the control arm. Additionally, 69 (92.0%) trials reported a prespecified noninferiority or equivalence margin of which only 23 (33.3%) provided justification for this margin based on prior literature. For trials with time-to-event primary endpoints, the median noninferiority margin was a hazard ratio of 1.22 (range = 1.08-1.52). Investigators reported a per-protocol analysis for the primary endpoint in only 28 (37%) trials.

Conclusions: Although most published noninferiority and equivalence trials have clear justification for their design, few provide rationale for the chosen margin or report a per-protocol analysis. These findings underscore the need for rigorous standards in trial design and reporting.

Figure 1.

Identifying justification for the use of noninferiority or equivalence designs for 75 phase 3 oncology trials. For a noninferiority or equivalence trial design to be inherently justified, it must have at least 1 characteristic inherent to the treatment itself that provides it with an advantage over the control treatment. For trials that were not inherently justified, the respective publications were reviewed to determine whether the experimental arm demonstrated improved quality of life or toxicity compared with the control arm based on either statistical testing or a prespecified minimal clinically important difference threshold. Abbreviations: QOL = quality of life; NI = noninferiority.

Figure 2.

Margins and outcomes for noninferiority trials with time-to-event primary endpoints. The vertical lines indicate noninferiority margins. Each circle indicates the upper limit of the confidence interval for the primary endpoint of the trial. The filled and unfilled circles represent positive and negative outcomes lying below and above the noninferiority margin, respectively. Investigators in all trials reported margins and results in terms of hazard ratios. Trials are grouped based on type of primary endpoint and ordered by noninferiority margin value. The x-axis has been converted to the logarithmic scale. ^aInvestigators in these trials did not report a noninferiority margin and instead used a synthesis method for testing noninferiority. ^bInvestigators in this trial reported individual assessments of 2 experimental arms for noninferiority. NI = noninferiority; PEP = primary endpoint; CI = confidence interval.