Trials in Alzheimer’s disease were similarly revealing.41 Internal Merck analyses in April 2001 showed that Vioxx increased total mortality significantly by a factor of three, but these analyses were not submitted to the FDA until 2 years later and they were not made public. Merck continued to recruit patients in one of the trials for an additional 2 years after it knew that Vioxx was deadly. Despite the deaths, the two published papers stated that Vioxx was ‘well tolerated’. That must be the most obscene interpretation that exists of a drug being ‘well tolerated’, but I accept that dead patients cannot complain about lack of tolerance. What Merck did was to discard all deaths that occurred more than 2 weeks after the patients got off the drug, e.g. because of adverse effects, in violation of Merck’s own protocol that stated that such deaths should be included in the results.42 In fact, the risk of thrombosis may be increased a whole year after patients come off the drug. Merck spokespeople lied to the FDA and Congress about what and when the company knew that Vioxx is deadly.
There were lies all over the place. Two months after the withdrawal of Vioxx, the medical director of Merck in Sweden wrote in the Swedish Medical Journal that none of the trials before one on adenomas from 2005 had shown an increased risk of Vioxx compared with placebo.43
The same year Merck pulled rofecoxib off the market, its CEO received performance-based bonuses worth over $36 million in addition to his base salary44 and he was never indicted. Merck pleaded guilty in 2012 to a criminal violation of federal law related to its promotion and marketing of Vioxx and was to pay nearly a billion dollars in a criminal fine and civil damages.45 In 2007, the company announced a settlement worth $4.85 billion.46 At that time, the company had already spent more than $1.2 billion on legal fees.47 The crimes involved off-label marketing of Vioxx and false statements about the drug’s cardiovascular safety. In relation to our study of Vioxx abstracts, we registered the conditions (apart from arthritis) that rofecoxib was proposed for, in 852 abstracts. Although almost half of the abstracts were published after the withdrawal of rofecoxib, where there was no longer any interest in suggesting new indications for the drug, the number and variety of conditions for which an effect of rofecoxib was proposed was astounding, no less than 30.28 It was as if the drug – like Neurontin – could be used for everything, e.g. schizophrenia, sclerosis, eight different cancers and premenstrual acne (see Table 13.1), and yet we only studied abstracts. There were likely more conditions mentioned in the main text of the papers.
Table 13.1 Conditions for which an effect of rofecoxib was mentioned in 852 abstracts
Neurological disorders
Hemicrania continua
Schizophrenia
Sclerosis
Alzheimer’s dementia
Migraine
Premenstrual migraine
Surgery
Prevention of urethral strictures after TURP
Pre-medication for tonsillectomy
Pre-medication for uterine curettage
Hernia operations
Post CABG
Pre-medication for ear-nose-throat surgery in general
Minor dental surgery (e.g. removal of molars)
Minor orthopaedic surgery
Cancer
Treatment for glioblastoma multiforme
Protection against colorectal neoplasia in familiar polyposis
Treatment of malignant melanoma and sarcomas
Treatment of prostate cancer
Treatment of bone cancer
Treatment of breast cancer
Treatment of lung cancer
Other
Reduction of atherosclerosis among ACS-patients post-infarction
Congenial nephrogenous diabetes insipidus
Menstrual pain
Endometriosis
Non-bacterial prostatitis
Haemophilic arthropathy
Premenstrual acne
Prevention of ectopic ossification in arthroplasty
How many patients did Merck kill with Vioxx because of thrombosis? In its trial of colorectal adenomas, Merck assessed thrombotic events and there were 1.5 more cases of myocardial infarction, sudden cardiac death or stroke on rofecoxib than on placebo per 100 patients treated.37 More than 80 million patients have been treated with rofecoxib,4 and since about 10% of such events are fatal, a crude estimate is that rofecoxib has killed about 120 000 people. The patients were treated for 2.4 years, on average, and as many patients in clinical practice are treated for shorter periods, this could be an overestimate. However, other factors tend to lead to underestimation: only events that occurred within 2 weeks after the patients stopped their drug were recorded and the patients were only 59 years of age, on average, and at low risk for thrombotic events.37 This is a general problem with Merck’s trials. Merck only included patients that had an unusually low risk of thrombosis, e.g. Medicare patients in Tennessee treated with rofecoxib in clinical practice had a baseline risk of getting a myocardial infarction that was eight times higher than that for the patients in the trials.6 I therefore believe, also considering that patients with arthritis are usually treated for years with NSAIDs, that my estimate of 120 000 deaths because of thrombosis is realistic. In addition, Vioxx has killed many thousands of patients because of ulcer complications.
In 2006, I saw a TV commercial in the United States on CNN that ended with a very deep voice saying, ‘Merck, where the patients come first.’ I couldn’t help thinking, ‘Merck, where the patients die first.’
References
1 Tanne JH. Merck appeals rofecoxib verdict. BMJ. 2007; 334: 607.
2 Lenzer J. FDA is incapable of protecting US ‘against another Vioxx’. BMJ. 2004; 329: 1253.
3 Krumholz HM, Ross JS, Presler AH, et al. What have we learned from Vioxx? BMJ. 2007; 334: 120–3.
4 Topol EJ. Failing the public health – rofecoxib, Merck, and the FDA. N Engl J Med. 2004; 351: 1707–9.
5 Petersen M. Our Daily Meds. New York: Sarah Crichton Books; 2008.
6 Jüni P, Nartey L, Reichenbach S, et al. Risk of cardiovascular events and rofecoxib: cumulative meta-analysis. Lancet. 2004; 364: 2021–9.
7 Graham DJ. COX-2 inhibitors, other NSAIDs, and cardiovascular risk: the seduction of common sense. JAMA. 2006; 296: 1653–6.
8 Topol E. Arthritis medicines and cardiovascular events – ‘house of coxibs’. JAMA. 2005; 293: 366–8.
9 Psaty BM, Furberg CD. COX-2 inhibitors – lessons in drug safety. N Engl J Med. 2005; 352: 1133–5.
10 US Senate Finance Committee. Testimony of David J Graham, MD, MPH. 2004 Nov 18. Available online at: www.finance.senate.gov/imo/media/doc/111804dgtest.pdf (accessed 21 February 2013).
11 US Food and Drug Administration. Memorandum. 2001. Available online at: www.fda.gov/ohrms/dockets/ac/01/briefing/3677b2_06_cardio.pdf (accessed 23 June 2009).
12 Weaver AL, Messner RP, Storms WW, et al. Treatment of patients with osteoarthritis with rofecoxib compared with nabumetone. J Clin Rheumatol. 2006; 12: 17–25.
13 Bombardier C, Laine L, Reicin A, et al. Comparison of upper gastrointestinal toxicity of rofecoxib and naproxen in patients with rheumatoid arthritis. N Engl J Med. 2000; 343: 1520–8.
14 Armstrong D. The New England Journal missed Vioxx warning signs. 2006 May 15. Available online at: www.post-gazette.com/pg/06135/690336-114.stm (accessed 27 November 2012).
15 Curfman GD, Morrissey S, Drazen JM. Expression of concern reaffirmed. N Engl J Med. 2006. 10.1056/NEJMe068054. Accessed 23 Feb 2006.
16 Curfman GD, Morrissey S, Drazen JM. Expression of concern: Bombardier et al., ‘Comparison of upper gastrointestinal toxicity of rofecoxib and naproxen in patients with rheumatoid arthritis,’ N Engl J Med 2000;343:1520–8. N Engl J Med. 2005; 353: 2813–14.
17 Mukherjee D, Nissen SE, Topol EJ. Risk of cardiovascular events associated with selective COX-2 inhibitors. JAMA. 2001; 286: 954–9.
18 Liévre M, Abadie E, on behalf of the French Marketing Authorization Committee. Discontinuation of Vioxx. Lancet. 2005; 36
5: 23–4.
19 Konstam MA, Weir MR, Reicin A. Cardiovascular thrombotic events in controlled, clinical trials of rofecoxib. Circulation. 2001; 104: 2280–8.
20 Reicin AS, Shapiro D, Sperling RS, et al. Comparison of cardiovascular thrombotic events in patients with osteoarthritis treated with rofecoxib versus nonselective nonsteroidal anti-inflammatory drugs (ibuprofen, diclofenac, and nabumetone). Am J Cardiol. 2002; 89: 204–9.
21 Corporate sponsorship. American Heart Association. Updated 2012 Oct 25. Available online at: www.heart.org/HEARTORG/Giving/ForCompanies/SponsorshipOpportunities/Corporate-Sponsorship_UCM_321431_Article.jsp (accessed 31 October 2012).
22 Kassirer JP. On the Take: how medicine’s complicity with big business can endanger your health. Oxford: Oxford University Press; 2005.
23 Sanon S, Patel R, Eshelbrenner C, et al. Acute coronary syndrome in patients with diabetes mellitus: perspectives of an interventional cardiologist. Am J Cardiol. 2012; 110 supplement: 13B–23B.
24 Hill KP, Ross JS, Egilman DS, et al. The ADVANTAGE seeding trial: a review of internal documents. Ann Intern Med. 2008; 149: 251–8.
25 Lisse JR, Perlman M, Johansson G, et al. Gastrointestinal tolerability and effectiveness of rofecoxib versus naproxen in the treatment of osteoarthritis: a randomized, controlled trial. Ann Intern Med. 2003; 139: 539–46.
26 Berenson A. Evidence in Vioxx suits shows intervention by Merck officials. New York Times. 2005 Apr 24.
27 Ross JS, Hill KP, Egilman DS, et al. Guest authorship and ghostwriting in publications related to rofecoxib: a case study of industry documents from rofecoxib litigation. JAMA. 2008; 299: 1800–2.
28 Jørgensen AW, Jørgensen KJ, Gøtzsche PC. Unbalanced reporting of benefits and harms in abstracts on rofecoxib. Eur J Clin Pharmacol. 2010; 66: 341–7.
29 Grant B. Merck published fake journal. The Scientist. 2009. Available online at: www.the-scientist.com/blog/display/55671 (accessed 23 June 2009).
30 Day M. Don’t blame it all on the bogey. BMJ. 2007; 334: 1250–1.
31 Psaty BM, Furberg CD. COX-2 inhibitors – lessons in drug safety. N Engl J Med. 2005; 352: 1133–5.
32 Waxman HA. The lessons of Vioxx – drug safety and sales. N Engl J Med. 2005; 352: 2576–8.
33 Waxman HA. The marketing of Vioxx to physicians. Memorandum. Congress of the United States. 2005 May 5.
34 Frazier KC. The lessons of Vioxx. N Engl J Med. 2005; 353: 1420.
35 Waxman HA. The lessons of Vioxx. N Engl J Med. 2005; 353: 1420–1.
36 Kim PS, Reicin AS. Rofecoxib, Merck, and the FDA. N Engl J Med. 2004; 351: 2875–6.
37 Bresalier RS, Sandler RS, Quan H, et al. Cardiovascular events associated with rofecoxib in a colorectal adenoma chemoprevention trial. N Engl J Med. 2005; 352: 1092–102.
38 Nissen SE. Adverse cardiovascular effects of rofecoxib. N Engl J Med. 2006; 355: 203–4.
39 Correction. N Engl J Med. 2006; 355: 221.
40 Topol E. Rofecoxib, Merck, and the FDA. N Engl J Med. 2004; 351: 2877–8.
41 Psaty BM, Kronmal RA. Reporting mortality findings in trials of rofecoxib for Alzheimer disease or cognitive impairment: a case study based on documents from rofecoxib litigation. JAMA. 2008; 299: 1813–17.
42 Madigan D, Sigelman DW, Mayer JW, et al. Under-reporting of cardiovascular events in the rofecoxib Alzheimer disease studies. Am Heart J. 2012; 164: 186–93.
43 Juhlin R. [MSD about Vioxx]. Läkartidningen. 2004; 46: 3720–1.
44 Whelton RS. Effects of excessive CEO pay on U.S. society. Available online at: www.svsu.edu/emplibrary/Whelton%20article.pdf (accessed 6 November 2007).
45 Department of Justice. U.S. pharmaceutical company Merck Sharp & Dohme sentenced in connection with unlawful promotion of Vioxx. 2012 April 19.
46 Charatan F. 94% of patients suing Merck over rofecoxib agree to terms. BMJ. 2008; 336: 580–1.
47 Berenson A. Merck agrees to settle Vioxx suits for $4.85 billion. New York Times. 2007 Nov 9.
14
Fraudulent celecoxib trial and other lies
Companies cannot be expected to play the role of judge and jury, and there is a real risk of fraud.
Prescrire International, about the CLASS study1
Pharmacia, later bought by Pfizer, published a large trial, the CLASS trial,2 of celecoxib (Celebrex) in JAMA in 2000, which was fraudulent. All study authors were employees or paid consultants to the company, and eight US medical schools had contributed authors.3
According to the paper, celecoxib resulted in fewer stomach ulcers than the two comparators, diclofenac and ibuprofen, and two clinical experts wrote a favourable editorial in JAMA.4 One of the editorialists was furious when he learned later – because of his membership of an FDA advisory committee – that it was not one trial but two trials bundled together to look like one, and that the trials ran for 12 and 15 months, not for 6 months as stated in JAMA.
The protocols for the two trials differed markedly from the published paper in design, outcomes, duration of follow-up and analysis, and the advantage of celecoxib disappeared when the protocol-specified analyses were performed by independent researchers.5
People in the company knew perfectly well what they were doing. In one email, an associate medical director at Pharmacia disparaged the way the study was being presented as ‘data massage’, for ‘no other reason than it happens to look better’.6 In another email, a medical director at Pfizer described it as ‘cherrypicking the data’ even as officials were publicly boasting of the study’s success. Internal documents show a game plan on how the company might present unwelcome findings: ‘Worse case: we have to attack the trial design if we do not see the results we want … If other endpoints do not deliver, we will also need to strategize on how we provide the data.’ A slide proposed explaining poor results through ‘statistical glitches’.
The FDA’s advisory committee concluded that, based on the full data, celecoxib exhibited no advantage in reducing ulcer complications over the two old, much cheaper drugs. The FDA’s statistical reviewer explained why the company’s arguments for the 6-month analysis were obviously invalid.7 A committee meeting in 2005 was also illuminating. All 32 participants considered that celecoxib, rofecoxib and valdecoxib increase the risk of cardiovascular events.8
However, the drug agencies continued to drag their feet and downplay the facts. For example, the Danish drug agency changed its product information for etoricoxib (Arcoxia, a Merck product) a week after the FDA meeting, so that it now said that ‘Clinical studies suggest that the group of selective COX-2 inhibitors may be associated with a risk of thromboembolic events.’ No way! The terms suggest, may be, and associated with document just how difficult it is for drug agencies to acknowledge the harms of drugs they have approved. Here is an honest version: clinical studies have shown that the group of selective COX-2 inhibitors increase thromboembolic events. Note that I left out a risk of. When a harm has been shown to occur in randomised trials, it is not a risk of harm, it is a real harm. We don’t talk about the chance of benefit, we talk about benefit and therefore also need to talk about harms. It is so typical for regulators and industry to use different language when they see what they like to see from what they use when they prefer to close their eyes.
Merck Denmark must have welcomed this wool-in-mouth statement, as its letter to Danish doctors 5 days later said that ‘selective COX-2 inhibitors can possibly be associated with a risk of thromboembolic events’. Oh dear. Merck had just killed more than one hundred thousand patients with Vioxx but didn’t even on this occasion admit the proven cause–effect relationship but said that such drugs are possibly associated with a risk. Downgrading the unwelcome facts three times in just five words is something of an achievement.
In 2002, a Pfizer sponsored meta-analysis was published in the BMJ,9 which shows how risky it is to collaborate with the industry, even for a skilled statistician who has done a lot of good work for the Cochrane Collaboration. The paper surprised many of his Cochran
e colleagues when it came out. It claimed that celecoxib leads to fewer serious gastrointestinal events, and the abstract only mentioned relative benefit, not absolute benefit, which was far more modest. The authors only included the misleading 6 months data for the CLASS trial, which was by far the biggest one. What was most strange, however, was that, although the gastrointestinal events were described in detail over several pages, including many graphs, there were no data on thromboses, which makes the review completely worthless.
The authors, one of which was from Pfizer, explained that the review was limited to assessing only upper gastrointestinal safety, with the excuse that the trials did not report on thromboses. This excuse is pathetic. It is irresponsible not to report the number of thromboses, given that it is the most important harm of COX-2 inhibitors. Furthermore, the clinicians are obliged to report all serious adverse events immediately to the company, which means that the company must have had data on thromboses, whether or not they preferred to forget about them. In fact, thromboses were reported in the CLASS trial, and even using only the misleading 6 months data, there were 4.3% serious adverse events with celecoxib and 4.2% with the other two drugs, i.e. no advantage at all for celecoxib.2
The manipulations paid off, as they always do. About 30 000 reprints were bought from the publisher and less than 2 years after its publication, the CLASS trial had already been cited 169 times, and sales increased from $2.6 billion to $3.1 billion in just 1 year.5 The fraud in JAMA, which has been propagated in many meta-analyses, must have been worth billions of dollars for the company.
The decision to report only data for the first 6 months was taken post hoc while the trial was running. The company might therefore have known beforehand that it would benefit the drug not to report the full data. In reply to the criticisms, the authors wrote that their decision ‘was made before the trial analysis was completed’.10 This explanation stinks. I’m sure that if Pharmacia/Pfizer (the other two authors of the meta-analysis weren’t involved with the trial, although they say ‘we’ in their reply) had made this decision earlier, during the trial’s execution, the company would have told us. And if they had made the decision without looking at the data, they would have told us, as it would have strengthened their credibility. Another problem was that, in the CLASS trial, adverse events were only recorded if they occurred within 48 hours of stopping the drug in case a patient dropped out because of harms. This appallingly bad trial conduct means the company might have missed many cases of myocardial infarction and other thromboses on celecoxib.
Deadly Medicines and Organised Crime Page 29