Clinical Trials

Clinical Trials: A Practical Guide ■❚❙❘Table 1. A summary of the key features of the bacterial community-acquired pneumonia (CAP) trial example [2].DesignObjectiveRandomized, double-blind, multicenter, parallel groupTo demonstrate that oral Augmentin SR 2000/125 mg twice daily for 7 days is at leastas effective clinically as oral Augmentin 875/125 mg twice daily for 7 days in thetreatment of CAP in adultsPrimary endpointClinical response (success/failure) at test of cure at 4 weeks posttreatmentTreatmentNew treatmentOral Augmentin SR 2000/125 mg twice dailyStandard treatment Oral Augmentin 875/125 mg twice dailyNoninferiority margin δ = –10%The prespecified maximum allowable limit of difference in success rates betweenthe new and standard treatmentsNull hypothesisH 0: π N– π S≤ δWhere π Nand π Srepresent the rates of success at test of cure with the newand standard treatments, respectivelyThe null hypothesis (H 0) implies that the new treatment is inferior to the standard oneAlternative hypothesisSample sizeAnalysis methodResultsConclusionH a: π N– π S> δThe new treatment is clinically noninferior to the standard treatment within thepredefined allowable range (δ) of clinical significanceApproximately 592 patients with CAP were required so as to provide 444 evaluablepatients (222 per treatment arm), with which the study will have 90% power toassess that Augmentin SR 2000/125 mg is noninferior to Augmentin 875/125 mgTwo-sided 95% confidence interval for the difference in the proportion of successesbetween the treatment groups, calculated using the normal approximation to thebinomial distributionEstimated difference in the proportion of successes between the treatment groupsand 95% confidence intervalPer-protocol: 2.7% (–3.0%, 8.3%)Intention-to-treat: 7.0% (0.9%, 13.0%)Augmentin SR 2000/125 mg, twice daily for 7 days, is at least as effective clinicallyas Augmentin 875/125 mg, twice daily for 7 days, in the treatment of CAP in adultsWhen are noninferiority trials used?A number of important factors contribute to choosing a noninferiority design.Primarily, noninferiority trials are employed in situations where efficacioustreatments already exist. Where this is the case, it will often be unethical to carryout a randomized, placebo-controlled clinical trial. The new treatment might betested to establish that it matches the efficacy of the standard, and at the sametime has secondary advantages (eg, in terms of safety, convenience to the patient,or cost-effectiveness). Alternatively, it might have potential as a second-linetherapy to the standard (in cases where the standard fails or is not tolerated).133