Issues in Noninferiority Trials: The Evidence in Community‐Acquired Pneumonia

Abstract

When investigators hypothesize that experimental interventions provide advantages other than improved effectiveness, they can use noninferiority (NI) trials to determine whether one can rule out the possibility that those interventions have unacceptably worse effectiveness than the standard "active control" regimen. To conduct valid NI trials, there must be evidence from historical studies that provides reliable, reproducible, and precise estimates of the effect of the active control, compared with that of placebo, on the specific outcomes investigators plan to use in the NI trial; this effect should be of substantial magnitude, and the estimates of the active control's effect from historical studies must represent its effect in the planned NI trial had a placebo group been included. These conditions allow formulation of an NI margin such that, if the NI trial establishes that the effectiveness of the experimental intervention is not worse than the effectiveness of the active control by more than the NI margin, then one can conclude that the experimental regimen (1) preserves a substantial fraction of the effect of the active control and (2) will not result in a clinically meaningful loss of effectiveness. After general discussion of NI trial design issues, we consider the design of NI trials to evaluate antimicrobials in the treatment of community-acquired pneumonia. We present an extensive literature review, allowing estimation of the historical effect of active control regimens in community-acquired pneumonia primarily on the basis of evidence related to use of sulfonamides or penicillin. This review allows formulation of NI margins that are specific to age and bacteremia status of patients.When effective interventions are available to treat or prevent a disease, there are instances in which investigators are interested in evaluating new experimental agents because they hypothesize that these agents provide advantages other than enhancing effectiveness. For instance, in the treatment of invasive aspergillosis, voriconazole may provide a better adverseReprints or correspondence: Dr. Thomas R. Fleming, Dept. of Biostatistics, University of Washington, Box 357232, Seattle, WA 98195 (E-mail: tfleming@u.washington.edu).. The content of this publication does not necessarily reflect the views or policies of the US Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US government. Potential conflicts of interest. T.R.F. has received consulting fees from Boehringer-Ingelheim, Bristol-Myers Squibb, Eli Lilly, GlaxoSmithKline, Novartis, Pfizer, Roche, Schering-Plough, Theravance, and Wyeth; interactions with these companies were not on projects related to communityacquired pneumonia. J.H.P. has received consulting fees from Acureon, Astellas, Astra-Zeneca, Basilea, Centegen, Cerexa, Concert, Cubist, Destiny, Forest, Johnson & Johnson, Merck, Methylgene, MPEX, Octoplus, Takeda, Theravance, and Wyeth. NIH Public Access Author Manu...

Review of Recent Clinical Trials of Hospital‐Acquired Pneumonia and Ventilator‐Associated Pneumonia: A Perspective from Academia

“…The implication for trials of nosocomial pneumonia is that they will also be held to this stricter standard. This is particularly important and highlights the need to establish the clinical effectiveness of the control drugs used in nosocomial pneumonia trials over placebo, typically by use of historical data, to formulate appropriate noninferiority margins [41].…”

Some current issues in the design of HIV noninferiority trials

“…We have shown that, in the ACTG 5202 trial, the use of a 10% margin would have led to a demonstration of noninferiority. The protocolspecified noninferiority margin is intended to reflect the maximum clinically acceptable difference between the therapeutic alternatives, as well as to prevent successive erosion of efficacy as new agents become the old standards in future trials [30][31][32]. In reality, however, the margin is mostly driven by logistical considerations around sample size and power calculations [2,19,31,[33][34][35].…”

Cethromycin versus Clarithromycin for Community-Acquired Pneumonia: Comparative Efficacy and Safety Outcomes from Two Double-Blinded, Randomized, Parallel-Group, Multicenter, Multinational Noninferiority Studies

“…Demonstrating clinical utility of a ketolide agent in precisely the patient population most likely to benefit from this intervention remains a development challenge. The applicability and utility of noninferiority trials in determining the efficacy of new treatment regimens for mild to moderate CAP has come under scrutiny by regulatory agencies (16) and academic investigators since the initiation of this program (12). Specifically, the determination of an acceptable noninferiority margin in the absence of a supported treatment effect of the comparator agent has proven problematic.…”

“…This sliding delta method has since been suggested to be inadequate in assessing noninferiority by statisticians in this field. Indeed, the determination of noninferiority in this indication and in this patient population has become problematic (12).…”