Healthy Skepticism AdWatch
AdWatch illuminates the logical, psychological and pharmacological techniques used in drug advertisements.
August 2007, Australia
Criticism of Adwatch on Avandia (rosiglitazone)
This issue has been written in response to a subscriber's comments regarding the March AdWatch about GlaxoSmithKline's advertising of rosiglitazone (Avandia) in Australia.
This issue has been written in response to a subscriber’s comments regarding the March AdWatch about GlaxoSmithKline’s advertising of rosiglitazone (Avandia) in Australia. We believe that his concerns may be shared by others. Therefore they deserve a public response. As well, the intervening period has seen the publication and wide reporting of a meta-analysis which has revealed another safety problem, so questions about the reliability of the company’s communications remain topical. The subscriber’s comments are in blue and our responses are in black.
I visited the AdWatch site, read the Avandia ad and re-read the original article in the New England Journal of Medicine. I must say that I derived a different view of the ad and thought that you might like to read my version. I had better say, from the outset, these things: I have no connections with GlaxoSmithKline and have received no funds from them for over 25 years, I am not a diabetologist, I don’t start patients on rosiglitazone monotherapy and I have only a small number of GlaxoSmithKline shares in the UK.
1) The ad refers to an article in the New England Journal of Medicine.
Yes. It also refers to a GSK website, which puts a more positive spin on the study than the NEJM article does. For example, according to the article, ‘Whether the statistically significant differences between rosiglitazone and metformin would translate into longer-term effects on disease progression or on microvascular or macrovascular outcomes needs to be determined’ (p. 2440).[1] On the website, Professor Paul Zimmet is quoted as saying ‘ADOPT provides important information on how the medical and diabetes communities can gain control of type 2 diabetes and impact on the progression of the disease [emphasis added]’.[2]
2) The ad summarizes the conclusions of that article accurately.
No, it does not accurately summarize the conclusions of the article, which stress that risks, adverse events, and costs should be considered as well as benefits: ‘The potential risks and benefits, the profile of adverse events, and the costs of these three drugs should all be considered to help inform the choice of pharmacotherapy for patients with type 2 diabetes’ (p. 2427) [emphases added].
Furthermore, an advertisement is not just required to summarise the conclusions of a selected article appropriately. Medicines Australia’s code of conduct states that companies should ensure that they are not using papers selectively to support their claims.
3) Maybe, if the editor of AdWatch has quibbles with the methodology described in the article, he/she should take them up with the editor of the New England Journal of Medicine rather than GSK. They only sponsored the trial.
The AdWatch did not primarily criticise the methodology of the trial per se. Instead, we focused primarily on the problems with the advertisement including selective reporting and interpretation of the trial, and inappropriate generalisation.
Our main criticism of the methodology of the article is the choice of a surrogate outcome (glycaemic control) over a more meaningful outcome, glycated haemoglobin, which correlates best with the risk of diabetes complications. While the publication of this kind of article in medical journals seems acceptable, reliance on this article in this advertisement may implicitly promote the view that rosiglitazone is a better first-line choice than metformin or sulfonylureas for patients with type 2 diabetes. This is in contradiction to all clinical guidelines for type 2 diabetes, as rosiglitazone has not been shown to be as effective as metformin and sulfonylureas in reducing cardiovascular risk, which is the main issue in these patients.
As stated in the AdWatch, industry-funded trials are significantly more likely to have positive results than independently funded trials .[3] Similarly, Ridker and Torres (2006) found that industry-funded trials and trials using surrogate outcomes are significantly more likely to have positive results.[4]
Kjaergard and Als-Nielsen (2002) found that authors’ conclusions about randomised clinical trials significantly favoured pharmacological and non-pharmacological interventions if financial competing interests were declared,[5] as was the case with the ADOPT trial - all eleven of the authors had financial links to GSK.
Editors of medical journals have a clear responsibility to ensure that the methodology of clinical trials published is appropriate, but the primary responsibility lies with the sponsor of the clinical trial.
4) In fact the ad doesn’t say to use any of these drugs as monotherapy; it only describes what was done in the trial. The trial was set up in the way described because the folks at the FDA ask: that new drugs be used singly and be shown to be better than placebo. By contrast, MDs ask whether a new drug is better than the existing treatments. The only way to find out is to try them head-to-head. So AdWatch’s quibble is with the FDA and our colleagues in clinical pharmacology.
Monotherapy is referred to twice in the ad. Firstly, the second figure focuses on ‘Reduction in risk of monotherapy failure at five years’. Secondly, the indications are listed as: ‘Treatment of Type 2 diabetes mellitus. Monotherapy, dual therapy…’.
The GSK website refers to monotherapy eight times. It also says that oral therapy is usually a single anti-hyperglycaemic agent.
It is correct that the FDA requires monotherapy placebo trials, but this trial was not conducted to obtain FDA approval. It was a head-to-head trial funded by GSK to increase market share.
5) As a matter of fact I don’t like relative risk either but that seems to be an acceptable and standard way of expressing results that has lasted for 20 years. So AdWatch needs to convince the statisticians first that it is not correct. What can an advertiser do but take accepted methods and use them?
Clinical judgements should be based on both relative risk and absolute risk. According to Schechtman (2002, p. 431), ‘It is recommended that researchers report both a relative and an absolute measure and present these with appropriate confidence intervals.’[6]
6) It is a 5-year trial, but the critic seems not to know how difficult it is to run a trial of daily treatment for 5 years in a group of 6000 who start by saying they will comply but then have second thoughts. Have they conducted a clinical trial like this themselves? Can you refer me to other trials with thousands of patients on daily medications who have been followed up for 5 years? What is the usual compliance rate?
The NEJM article acknowledged that the dropout rate was higher than expected, necessitating an increase in the number of participants. The advertisement was potentially misleading by stating that there were 4,360 participants, without mentioning that 1,733 dropped out.
7) Your editor complains about lack of mention of side-effects in the ads, but it is not so. The side-effects are given in small print at the end. If doctors have not learned by now to read the whole ad, they have become a remarkably naive and unsuspecting lot. We were taught, as medical students 40 years ago, to read the whole ad including the side effects. Indeed, my first consultant told me that I could use any drug I liked, but I had to be able to tell him all about it and that included telling him about the side effects. Those were good lessons. We should all do it. But if we don’t read the whole ad, can we blame the advertiser?
The AdWatch does not claim that side-effects in general are not mentioned. It criticises the omission of the fact that the rosiglitazone group had significantly more heart failure and oedema than the glibenclamide group, and higher weight gain than the metformin group.
Furthermore, the advertisement lists only some of the potential adverse events: ‘Headache, back pain, hyperglycaemia, fatigue, diarrhoea, hypoglycaemia, oedema, anaemia, hypercholesterolaemia, weight gain, hepatic dysfunction, macular oedema*, cardiac ischaemia* - This is not a full list, for more details, refer to full PI’.
Notably, it does not mention the significantly increased risk of myocardial infarction associated with Avandia, which was revealed by a recent meta-analysis of the company’s studies.(Nissen & Wolski 2007).[7]
8) There is a bitter complaint about the graphs without a zero axis, but this too is a well recognized way of displaying results, in the whole of society, that we learned about more than 45 years ago at school. So maybe the problem your critic has is more with the education system in Australia, than with advertising. I did go back to the New England Journal of Medicine’s web-site and re-read the full text, and found that the graphics were so small that I had to enlarge them to a scale very similar to that depicted in the ad before I could read them. I did that before I had read past page 1 of the ad.
The AdWatch made the point that the split y-axis makes the differences look larger, as the full-scale graph shows. However, we acknowledge that the NEJM article also used a split y-axis.
9) Again I say Healthy Skepticism should demonstrate healthy scepticism, but unhealthy and unjustified criticism just weakens the case.
We agree that we should avoid unjustified criticism, but we do not believe that the criticisms in the AdWatch are unjustified.
10) If I were asked to write an ad or pass one for publication, even as an academic, I would have no difficulty with this one given the data in the article. Perhaps Healthy Skepticism should ask the NEJM, the FDA and others involved to re-examine their methods and requirements for demonstrating the efficacy of new drugs. You are quite right in your premise that I think I hear coming through that data like these are extremely influential and therefore it is important that the correct questions should be asked in the first place, so that the correct answers may be obtained and the correct inferences drawn, but if the drug manufacturers are asked to do something to comply with regulations for registration of a new drug, then it is unfair to complain about the manufacturer when they report what they were asked to do exactly as they did it. It should also be remembered that the longer we make the registration process and the shorter the unique time the manufacturer has to sell the product before the patent expiry date, the more likely we are to see high prices, questionable advertising, or both.
Yes, we are aware of the shortcomings of the regulatory requirements for registration of new drugs such as short-term placebo-controlled clinical trials with inadequate surrogate endpoints without meaningful long-term health outcomes, testing in highly selected populations, etc. The most significant institutional entity involved in the harmonisation of medicine regulatory standards worldwide is the International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH), which comprises the three pharmaceutical industry associations and regulatory agencies of the EU, US and Japan. The drug industry has thus a strong influence on the quality of regulatory standards and has an interest in keeping it low to facilitate speedy approval of new drugs. Furthermore, the ADOPT trial was a post-marketing study where the drug company was free to use any methodology approved by an ethics committee.
So please be careful that criticism like this is not unfounded and does not complain about the wrong people.
AdWatch is written with great care to ensure that our criticisms are well founded and are addressed to the people responsible. In this case, no other group of people compelled the drug company to produce misleading advertising.
1. Kahn SE, Haffner SM, Heise MA, Herman WH, Holman RR, Jones NP, Kravitz BG, Lachin JM, O’Neill MC, Zinman B, Viberti G; ADOPT Study Group. Glycemic durability of rosiglitazone, metformin, or glyburide monotherapy. N Engl J Med. 2006 Dec 7;355(23):2427-43.
2. GlaxoSmithKline. ADOPT. Boronia, Victoria: GlaxoSmithKline [2006; cited 2007 Jun 25]. Available from: www.gsk.com.au/adopt
3. Bhandari M, Busse JW, Jackowski D, Montori VM, Schünemann H, Sprague S, Mears D, Schemitsch EH, Heels-Ansdell D, Devereaux PJ. Association between industry funding and statistically significant pro-industry findings in medical and surgical randomized trials. Can Med Assoc J. 2004;170(4):477-80.
4. Ridker PM, Torres J. Reported outcomes in major cardiovascular clinical trials funded by for-profit and not-for-profit organizations: 2000-2005. JAMA. 2006 May 17;295(19):2270-4.
5. Kjaergard LL, Als-Nielsen B. Association between competing interests and authors’ conclusions: epidemiological study of randomised clinical trials published in the BMJ. BMJ. 2002 Aug 3;325(7358):249-52.
6. Schechtman, Edna (2002). Odds ratio, relative risk, absolute risk reduction, and the number needed to treat - which of these should we use? Value in Health, 5(5), 431-436.
7. Nissen SE, Wolski K. Effect of rosiglitazone on the risk of myocardial infarction and death from cardiovascular causes. N Engl J Med. 2007 Jun 14;356(24):2457-71.
Indexes:
AdWatch (Australia)
Page views since 15 March 2010: 10523
Comments
Our members can see and make comments on this page.
