Noninferiority Trial Designs and Issues: Community-Acquired Bacterial Pneumonia (CABP)

• Activity Number: 564
• Type: Topic Contributed
• Date/Time: Wednesday, August 3, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Thamban Valappil*+ and Daniel Rubin*+
• Companies: U.S. Food and Drug Administration
• Keywords: noninferiority ; Clinical trials ; CABP ; anti-infective

Abstract

Clinical trials evaluating antibacterial drugs for community-acquired bacterial pneumonia have used non-inferiority designs, and a primary endpoint of investigator-assessment of clinical response, defined as resolution of signs and symptoms of the disease 1-2 weeks after the end of therapy. However, historical data that reliably estimate the magnitude of the active control effect over placebo are not available for this endpoint, and a non-inferiority margin cannot be rigorously justified. There is stronger evidence for a non-inferiority margin when using an all-cause mortality endpoint, but recent registration trials have enrolled patients with low risk for mortality. We will discuss design and analysis considerations in current trials, including the use of endpoints defined early in the treatment course, the impact of prior antimicrobial therapy on non-inferiority results, and concerns about the use of biomarkers that do not capture patient feeling, function, or survival.