PEDIATRICS Vol. 116 No. 2 August 2005, pp. 493-495 (doi:10.1542/peds.2005-1231)
COMMENTARY |
When an Asthma Drug Has an Inferiority Complex: A Noninferiority Trial
Department of Pediatrics Center for the Evaluative Clinical Sciences
Dartmouth Medical School
Hanover, NH 03755
After a new drug ascends to the regulatory summit of approval, the immediate and appropriate response by clinicians is to wonder if it has risen to a greater height of efficacy than drugs already available. The answer can depend on the outcome of clinical interest and the relative balance of clinical efficacy, safety, patient acceptability, and cost. It is against this background of uncertainty that comparison studies, typically funded by pharmaceutical companies, triangulate the relative value of new drugs compared with old.
In this issue of Pediatrics, Garcia and colleagues1 report a multicenter, double-blinded, active-control, randomized trial to compare the efficacy of oral montelukast and inhaled fluticasone in children 6 to 14 years old with mild persistent asthma. The study's primary hypothesis was that montelukast was not inferior to fluticasone in the change in percent of asthma "rescue-free days" over the course of 1 year. Thus, this study is a noninferiority trial, a relatively recent study design that tests the equivalency of 1 drug to another. In the drug-approval process, an active control trial is used in place of a placebo-controlled study, when, for example, it would be unethical to randomize to a placebo. More recently, noninferiority designs have been used in postmarketing studies to establish a drug's equivalence to existing drugs in the clinical marketplace.2 The stakes in these studies are high for patients whose therapy may be modified as a result of the findings and for pharmaceutical companies selling the drugs. How should a reader evaluate noninferiority trials, and what should we learn from the Garcia et al study?
The concept of a noninferiority trial is simple: for a predefined outcome, the experimental drug (eg, montelukast) is tested against a control drug (eg, inhaled fluticasone). If the difference in the outcome is less than a preset margin, then the drug can be claimed to be "noninferior." In more formal terms, the null hypothesis states that the difference in outcomes exceeds the preset margin. This hypothesis is rejected if a 1-sided hypothesis test is less than the 
level of significance (eg, .05) or if the confidence interval of the difference in outcomes fails to exceed the value of the preset outcome margin. In other words, the drug is not inferior if there is a <5% chance that the difference in outcomes exceeds the margin.
Two important assumptions of noninferiority trials must be met before a claim of equivalency can be accepted.3,4 First, the comparison drug must be as active in the trial as in past studies that established its efficacy. If, for some reason, the comparison drug in the trial is not as efficacious as measured historically, then the test drug can have weak effects, or perhaps no effect at all, and still be declared comparable. This means that poor patient adherence to the comparison drug can give a lead to the conclusion of the test drug's noninferiority. Without a placebo arm, the reader needs to look closely at the study population, the treatment and active control interventions, and the patient adherence to the study protocol to determine if study conditions for the comparison drug were reasonably similar to past efficacy trials.
Did the fluticasone arm in Garcia's study achieve previously reported levels of efficacy? No previous pediatric clinical efficacy trial of inhaled fluticasone has lasted 52 weeks. More importantly, the readers should note that ". . . no efforts were made to maintain strict compliance with study procedures except for what would be performed during normal clinical practice, to approximate real-world use of the 2 treatments." Thus, the Garcia et al study is not a noninferiority trial of efficacy but of effectiveness. In the absence of a published randomized, placebo-controlled clinical trial of the effectiveness of fluticasone, the adequacy of the fluticasone arm of the study is difficult to assess.
The reader can evaluate the second assumption more easily; the predetermined outcome margin should not be clinically important. The problem with the concept of clinical importance is that it relies on one's point of view. Should the margin be decided by clinicians, the study investigators, or patients? For very small differences, the question is moot, but when the margin is greater, it may exceed a difference that, although judged to be small by investigators, is meaningful to the patient or family.
The primary outcome of the Garcia et al study was the percent change in asthma rescue-free days ("rescue free" was defined as no rescue medication or asthma-related health care utilization.). Both groups did very well, with the percent of rescue-free days at 84% in the montelukast group and 86.7% in the fluticasone group. The preset margin selected by the investigators was –7% or 
2 days per month. In fact, the adjusted difference was –2.8% (<1 day per month), which is a difference unlikely to be important to patients. The null hypothesis was rejected, and the investigators accepted montelukast as noninferior for change in rescue-free days in the study population.
The study also found that fluticasone was favored in several secondary outcomes: percent predicted forced expiratory volume in 1 second (FEV1), percent of days with ß-agonist use, use of other rescue medications, use of systemic steroids, proportion of patients with asthma attack, and in asthma-related quality of life. There were no differences in change in FEV1, blood eosinophils, or use of additional non–ß-agonist, non–systemic-steroids rescue medication.
This study is unlikely to change the landscape of medical opinion about the equivalency of montelukast and inhaled corticosteroids. The evidence with respect to equivalency is robust for adults and remains incomplete for children. The most recent Cochrane collaboration review (last substantive update, October 2003) concluded that for adults, "[i]nhaled steroids at a dose of 400 mcg/day of beclomethasone or equivalent are more effective than anti-leukotriene agents given in the usual licensed doses."5 Unfortunately, only 1 English-language pediatric trial with a clinical outcome was identified; the investigators used an open-label design and found similar outcomes in 124 children.6 The 2002 update of the National Institutes of Health guidelines for the diagnosis and management of asthma states concurs with the Cochrane collaboration: "Therefore, based on the available data comparing LTRAs (leukotriene receptor antagonists) to inhaled corticosteroids, the Expert Panel concludes that inhaled corticosteroids should be the preferred treatment option for mild persistent asthma in adults and, by extrapolation until published comparison data become available, for children."7 These are general recommendations, and an individual patient response can vary from low to high for both inhaled corticosteroids and leukotriene modifiers.8
The most important management principle is not the initial selection of the best available controller drug but the long-term and regular supervision of children with persistent asthma. When a child is having symptoms or functional impairment, controller-drug doses can be increased or a different drug can be substituted or added. With this type of consistent follow-up, the optimal regimen can be determined for each patient.
Across the larger panorama of drug trials, the dark cloud on the horizon is the observation that most studies have shown the source of funding to predict the study outcome.9–13 This relationship is also strong for asthma-drug–comparison trials (unpublished data, 2005). At the end of the day, we are left with a vexing question: How can we judge the highest summit of asthma controller-drug efficacy when the measurement tools are designed and paid for by investigators with a commercial interest in the results?
 |
FOOTNOTES |
|---|
Accepted May 24, 2005.
Address correspondence to David C. Goodman, MD, MS, Department of Pediatrics and the Center for the Evaluative Clinical Sciences, Dartmouth Medical School, 7251 Strasenburgh Hall, Hanover, NH 03755. E-mail: david.goodman{at}dartmouth.edu
Conflict of interest: Dr Goodman is a former consultant to Merck & Co, Inc.
|
REFERENCES |
|---|
 TOP REFERENCES |
|---|
- Garcia MLG, Wahn U, Gilles L, Swern A, Tozzi CA, Polos P. Montelukast, compared with fluticasone, for control of asthma among 6- to 14-year-old patients with mild asthma: the MOSAIC Study.
Pediatrics. 2005;116
:360
–369
[Abstract/Free Full Text] - Steinijans VW, Neuhauser M, Hummel T, Leichtl S, Rathgeb F, Keller A. Asthma management: the challenge of equivalence. Int J Clin Pharmacol Ther. 1998;36 :117 –125[Web of Science][Medline]
- James Hung HM, Wang SJ, Tsong Y, Lawrence J, O'Neil RT. Some fundamental issues with non-inferiority testing in active controlled trials. Stat Med. 2003;22 :213 –225[CrossRef][Web of Science][Medline]
- D'Agostino RB Sr, Massaro JM, Sullivan LM. Non-inferiority trials: design concepts and issues—the encounters of academic consultants in statistics. Stat Med. 2003;22 :169 –186[CrossRef][Web of Science][Medline]
- Ducharme F, Di Salvio F. Anti-leukotriene agents compared to inhaled corticosteroids in the management of recurrent and/or chronic asthma in adults and children. Cochrane Database Syst Rev. 2005;(2) :1 –146
- Maspero JF, Duenas-Meza E, Volovitz B, et al. Oral montelukast versus inhaled beclomethasone in 6- to 11-year-old children with asthma: results of an open-label extension study evaluating long-term safety, satisfaction, and adherence with therapy. Curr Med Res Opin. 2001;17 :96 –104[CrossRef][Web of Science][Medline]
- National Asthma Education and Prevention Program. Expert Panel Report: Guidelines for the Diagnosis and Management of Asthma—Update on Selected Topics 2002. Bethesda, MD: National Institutes of Health; 2003. NIH Publication No. 02-5074
- Knorr B, Matz J, Bernstein JA, et al. Montelukast for chronic asthma in 6- to 14-year-old children: a randomized, double-blind trial. Pediatric Montelukast Study Group.
JAMA. 1998;279
:1181
–1186
[Abstract/Free Full Text] - Yaphe J, Edman R, Knishkowy B, Herman J. The association between funding by commercial interests and study outcome in randomized controlled drug trials.
Fam Pract. 2001;18
:565
–568
[Abstract/Free Full Text] - Djulbegovic B, Lacevic M, Cantor A, et al. The uncertainty principle and industry-sponsored research. Lancet. 2000;356 :635 –638[CrossRef][Web of Science][Medline]
- Melander H, Ahlqvist-Rastad J, Meijer G, Beermann B. Evidence b(i)ased medicine—selective reporting from studies sponsored by pharmaceutical industry: review of studies in new drug applications.
BMJ. 2003;326
:1171
–1173
[Abstract/Free Full Text] - Bhandari M, Busse JW, Jackowski D, et al. Association between industry funding and statistically significant pro-industry findings in medical and surgical randomized trials.
CMAJ. 2004;170
:477
–480
[Abstract/Free Full Text] - Montgomery JH, Byerly M, Carmody T, et al. An analysis of the effect of funding source in randomized clinical trials of second generation antipsychotics for the treatment of schizophrenia. Control Clin Trials. 2004;25 :598 –612[CrossRef][Web of Science][Medline]
PEDIATRICS (ISSN 1098-4275). ©2005 by the American Academy of Pediatrics
CiteULike 
Connotea 
Del.icio.us 
Digg 
Facebook 
Reddit 
Technorati 
Twitter What's this?
eLetters:
Read all eLetters
- EVALUATING ASTHMA CONTROL WITH MONTELUKAST IN CHILDREN WITH MILD ASTHMA IN THE MOSAIC STUDY
- Maria L. Garcia
- Pediatrics Online, 3 Oct 2005 [Full text]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
