The article was typical of the commercially sponsored clinical trials that appear in many journals. The Annals, however, made a publishing decision that was atypical of how many journals handle such articles. It included an accompanying editorial that was appropriately critical of the study design and methods, and appropriately attuned to how commercial research sponsors, such as pharmaceutical companies, biotechnology companies, and device manufacturers, may manipulate how research is designed, carried out, analyzed, and reported to serve their vested interests.
We have blogged about such research manipulation before. See our posts about Richard Smith's catalog of manipulation strategies, about specific strategies used in a study of muraglitazar, about David Healy's description about manipulation strategies used for studies of SSRIs (selective serotonin reuptake inhibitors), and Wally Smith's summary of the creation of pseudoevidence. But commercially sponsored articles whose design, execution, and analysis have been manipulated to serve their sponsors interests slip into print regularly, unaccompanied by any notice of how they serve commercial purposes.
The Clinical Trial of Exenatide
In summary, the trial by Zinman et al compared exenatide to placebo in patients already received either rosiglitazone (Avandia, by GlaxoSmithKline) or pioglitazone (Actos, by Takeda Pharmaceuticals), sometimes also metformin, but who were not necessarily trying to modify their lifestyles (either their diet or exercise schedule). A total of 233 patients were randomized, and followed for 16 weeks. Patients in the exenatide group had a statistically significantly lower hemoglobin A1c level, a measure of overall control of blood sugar, at the end of the trial. They also had statistically significantly greater rates of nausea and vomiting, and greater rates of dropping out due to adverse events.
Accompanying the trial was an editorial [Malazowski S. Exenatide in combination therapy: small study, big market, and many unanswered questions. Ann Intern Med 2007; 146: 527-528.] The editorial amounted to a rigorous critical review of the trial. It pointed out most of the trial's many design and methodologic problems, and explained why its results were less than earth-shaking.
These problems included:
- Difficulty determining to whom the results would apply - The patient population consisted of people who did not have their diabetes under excellent control, but were not on optimal medication regimens, and were not trying to improve their diets or increase their exercise. Thus, "we simply don't know whether patients optimally treated with diabetes education, diet, TZDs, [thiazolidinediones] and metformin will receive as much benefitfrom exenatide as the paper reports," since such patients were not included in the study. Nor could the study predict whether adding exenatide would work better than simply increasing the doses of conventional medication, or starting a diet modification or exercise program.
- Short duration - Although diabetes is a chronic problem, the study did not assess the new medication in long-term use. Thus, it was not designed to tell whether it would work in the long run, whether it would affect the patients' longevity, development of complications, symptoms or health status in the long run, and whether it would lead to adverse effects in the long run that did not occur in the short run.
- Small sample size - "Small and short studies provide a false sense of safety, because common severe drug reactions may not occur in the condensed timeline and the limited number of patients." Even so, patients who received exenatide had a relatively high rate of nausea and vomiting, often leading to discontinuation of the drug. So the study did not suggest that any benefits of the drug clearly outweighed its harms. [Note that the editorial did not mention that patients on exenatide also had higher (but not statistically significantly higher) rates of hypoglycemia, low blood sugar, than did patients given placebo. Hypoglycemia, which can be serious, is the main hazard of aggressive treatment of blood sugar. One patient on exenatide also developed an unusual, possibly allergic reaction, allergic alveolitis. The study was clearly too small to predict the rates of these effects that would occur if the drug were used in large numbers of patients. Thus, there is suspicion that the benefit/ harm ratio of this drug is even less favorable.]
Furthermore, the editorial pointed out the relationship between the study's commercial sponsorship and its flawed design. Malozowski stated, "the study was designed, conducted, and analyzed by employees of the manufacturer in collaboration with academicians...." (Note also that the three academicians who contributed to the study all worked part-time for the manufacturers as consultants and/or speakers.) So, "the design and reporting of Zinman and colleagues' study reminds us that the manufacturers control the flow of information about its product. By virtue of FDA approval for the combination of exenatide and TZDs, the data obtained in the study can lead to enormous financial benefits to the sponsor. Millions of patients received TZDs and metformin - now physicians may consider adding exenatide. Great power requires great responsibility. Physicians and patients need answers to the many questions raised by this small study."
In my experience, rarely have I seen a commercially sponsored study whose flaws all seemed to increase the likelihood of finding results favoring the sponsor's vested interests published with an accompanying commentary that pointed out these flaws and their relationships to these vested interests.
Given the great efforts commercial sponsors make to get studies with results favorable to their products published, it would be nice if all journals accompanied such articles with appropriately critical accompanying commentaries.
Kudos to the Annals of Internal Medicine for showing the way forward.