Deadly Medicines and Organised Crime

by Peter Gotzsche

  Access to data at other drug agencies

  In 2010, we contacted chief statistician Hans Melander at the Swedish drug agency and asked for access to the placebo-controlled trials and protocols for three SSRIs (citalopram, escitalopram and venlafaxine) submitted to the agency.

  We could get everything we wanted, but there was a problem. The reports had been filed in a mountain cave somewhere in Sweden where they took up 70 metres. It would cost about €50 000 to retrieve all this and move it back to the agency in Uppsala, but the agency generously offered to cover this cost. We could then work with the material at the agency, or have it all copied for €0.13 per page, or copy it ourselves at no cost, and take it to Denmark. I estimated that 70 metres in binders was about 500 000 pages, or around €70 000 to get the material copied. In order to work with all this, we needed to scan it, using special software that could also handle tables and convert it to searchable text.

  I told Melander to hold the horses and wait while we worked on our pilot study of duloxetine. Over more than a year, we had received documents from the EMA, also on other SSRIs, and they were still coming in. These documents were pdf files, which we converted to searchable text, but even so, it took two of our researchers more than a year before they had extracted the data we needed.

  The Dutch regulator was also very forthcoming, but they redacted the adverse effects before they sent us the files, which they were obliged to do according to a court verdict, so the material wasn’t very useful.

  In 1993, a bill was put before the British Parliament that would lead to greater access to regulatory information about the efficacy and safety of drugs, but it was immediately shot down by the industry aided by its apologists in government, ironically in the same year as the government published its white paper on Open Government.31

  Contacting the UK regulator to get data on fluoxetine, which the EMA didn’t have, was like contacting the MI5. The reply we got was anonymous and we were told that the agency had destroyed the files! The Medicines and Healthcare Products Regulatory Agency (MHRA) destroys the files after 15 years, ‘unless there is a legal, regulatory, or business need to keep them, or unless they are considered to be of lasting historic interest’.32 No legal or historic interest for unpublished drug trials for drugs that are still on the market? Can the irony be deeper than this?

  There were also bureaucratic obstacles: ‘Each individual document should be requested through a separate request and will be reviewed and assessed in terms of its suitability for release.’ We limited our request to very little, which the agency had told us was in their possession, but were then told that ‘public authorities are not required to comply with requests that they judge to constitute an excessive use of their resources. The time taken to complete an FOI [Freedom of Information] request should take no longer than 24 working hours, otherwise it is deemed an excessive use of resource. Your request falls into this category …’.

  Undeterred by this, I wrote that members of the European Commission and Parliament had been shocked when I told them that the MHRA destroys its files after 15 years. I suggested that, since the UK was the EU Reference Member State for fluoxetine and only the marketing authorisation holder, Eli Lilly, had the files, the agency should ask Lilly to resubmit the files to the agency, as companies are obliged by law to retain them. Finally, I noted that, based on our collaboration with other drug agencies, what we had requested couldn’t come anywhere near 24 hours, and I reminded the MHRA about the basic principles about citizens’ access to EU documents and that the UK was in fact a member state in the EU.

  New obstacles appeared: ‘From my preliminary assessment of your request I will not be able to answer it within the 20 days specified in the Act. It is my initial view that section 43 (commercial interests) of the Act may apply to at least some of the information you have requested.’

  Oh boy. This message came a year after the ombudsman’s press release accusing the EMA of maladministration and saying there were no commercial interests to protect! In its next letter, the MHRA said it had consulted with Lilly, which had refused to release the documents to us, as such release would harm the company. How could they know? Did they cover up something? Very likely they did (see Chapters 17 and 18).

  I changed tactics and asked whether the MHRA had thought about what its attitude might mean for its image. And complained that the MHRA hadn’t realised that it needed to update its policies and routines and bring them at par with the recent openness at the EMA.

  It worked. After three additional months, and 7 months after our initial request, the MHRA informed us that they would send us the documents. But the MHRA was still the lap dog of big pharma:

  ‘Please bear in mind that the volume of information you have asked for is large, and has taken time to redact and to liaise with the marketing authorisation holder, to ensure that they were fully informed of what we were intending to release to you.’

  A good thing was that, in contrast to the files we received from the Dutch agency, adverse effects had not been redacted. Only signatures, names, addresses, investigator CVs, ethics committee information and consent forms were redacted. Why weren’t we allowed to see the bits about ethics? Was Lilly afraid we might find out that some of their trials were unethical? We already know that consent forms routinely lie to the patients as they are told they contribute to science when in reality many results are shelved.11 It doesn’t make sense to delete this information, as it isn’t commercially confidential information, but it illustrates how arbitrary regulatory decisions are.

  The FDA isn’t forthcoming.11 Requests for data need to be very specific, which is difficult when you don’t know what is available. And searches on its website to find information may turn up hundreds of documents that are not clearly named, not indexed, don’t have a title page, and which exist only as non-searchable scanned images. The documents may not tell you what they are about until you get to page 19.11 That results in many people giving up, as we did when we tried. Furthermore, data are missing or arbitrarily deleted, e.g. only 16 out of at least 27 trials of celecoxib (Celebrex) were included in the FDA reports requested by researchers according to the Freedom of Information Act.33 Independent researchers who had access to FDA data nevertheless confirmed the cardiovascular harms of the drug.34

  For another COX-2 inhibitor, valdecoxib (Bextra, from Searle), 28 consecutive pages had been deleted by the FDA before they were sent to independent researchers, as they contained ‘trade secrets and/or confidential information that is not disclosable’.33 This is totally absurd, as these pages came from an FDA statistical review and evaluation of valdecoxib. There are absolutely no trade secrets or confidential information that is not disclosable in such reports.

  Deadly slimming pills

  The history of the slimming pills is a dire one that confirms that drug regulators aren’t willing to learn from history. Phentermine was approved in the United States in 1959 and is still on the market, although it’s similar to amphetamine, both chemically and in its effects. In the 1960s, another appetite suppressant with amphetamine effects, aminoxaphen (Aminorex), was very popular in Europe,35 but it causes pulmonary hypertension and was withdrawn after 7 years when hundreds of patients had died under terrible conditions.

  In 1973, fenfluramine (Pondimin), yet another amphetamine-like drug, was introduced on the US market. It increases the neurotransmitter serotonin, which the SSRIs also do (see Chapter 17). The drug was withdrawn in 1997, as it causes pulmonary hypertension and a serious form of fibrosis of the heart valves that also kills people. Pondimin was close to never making it to the market, but the FDA scientist who had written a disapproval letter was removed to another job. This led to a congressional investigation of misconduct at the FDA that concluded that a leading FDA officer had misled congress. This officer left the FDA to become an ‘expert witness’ for drug companies. Of course he did. History surely repeats itself in drug regulation.

  In the 1990s, many scientific articles in
Europe described Pondimin’s detrimental effects, but Wyeth, the maker of the drug, didn’t send these reports to the FDA.35 An obvious reason why Wyeth didn’t draw attention to the dangers with Pondimin was that the company was trying to get approval for a similarly deadly drug, dexfenfluramine (Redux), which was simply the d-enantiomer of fenfluramine (that consists of two enantiomers, which are mirror images of each other). A researcher who had worked with the drug while being employed by Servier went privately to the FDA with his findings that fenfluramine and dexfenfluramine lead to brain damage in apes and baboons, but he was immediately fired and nothing happened that might have protected the patients.

  Everything goes in drug regulation and dexfenfluramine came on the market in Europe. However, its use was severely restricted in 1995 after French researchers had shown that both Pondimin and Redux increase the risk of pulmonary hypertension 10 times. These findings were arrogantly dismissed by the FDA and the industry complained over a critical FDA officer. Nonetheless, the FDA advisory committee rejected the drug because of safety concerns. Wyeth complained and a new meeting was held just 2 months later, which is highly unusual. The committee now included more Redux supporters and the drug was narrowly approved in November 1995 with the votes six to five.36 When numbers of cases with pulmonary hypertension increased rapidly, FDA doctors tried to convince Wyeth/Interneuron that they should add a black box warning on the label. Instead, they graciously added a notice that Redux could cause hair loss, which had been reported more rarely than pulmonary hypertension!35

  This story of unbelievable crimes towards the patients just continued. Four months after approval of Redux, the damning French results were published in the New England Journal of Medicine but with an editorial that praised the drug and said that the risk of pulmonary hypertension was small and outweighed by the benefits of the drug. There wasn’t a trace in the editorial that its two authors were paid by the industry, which fact infuriated the editors of the journal when it was revealed by the Wall Street Journal. The benefit was a mere 3% weight loss, as stated by the company, e.g. from 100 kg to 97 kg. However, many patients drop out of the trials and the conventional statistical adjustment for this is flawed. Companies use the last recorded weight and carry it forward till the end of the trial. However, much of the weight people lose in the beginning comes back later, and even more important: if people cannot tolerate a drug, they cannot benefit from it. It would therefore be more sensible to carry forward the weight at baseline. In one of our studies, of rimonabant, the last observation carried forward showed a weight loss of 6.4 kg above placebo while baseline carried forward showed a benefit of only 1.5 kg.37

  While the patients continued to die because of the slimming pills they took, an academic researcher, Mike Weintraub, touted treatment with a combination of two amphetamine-like products, fenfluramine (Pondimin) and the old drug phentermine, on TV programmes and elsewhere, although this off-label use had not been approved by the FDA. The combination pill was called Fen-Phen. It became extremely popular even though an article flagged problems with memory loss. In 1996, the total number of prescriptions exceeded 18 million.38 However, in the summer of 1997, a series of 24 women who had developed valve disease while on Fen-Phen was published in the New England Journal of Medicine38 accompanied by an editorial written by the editor that this time left no doubt that the drugs are dangerous. Based on that paper, the FDA pressed Wyeth/Interneuron to withdraw Redux and Pondimin from the market.35

  But Wyeth didn’t give up. It had plans for ‘neutralising’ critical doctors and its doubt industry came into gear: hired guns among physicians lent their names to flawed results, and specialist journals lent their pages to the dirty work, above all the Journal of the American College of Cardiology, although the cardiologists should have been the most concerned doctors because the patients died from diseases in their specialty. One such cardiologist, Neil Weissman, published a paper in this journal in 1999 and similar papers in other journals purporting there was no problem; he received a total of almost $18 million from Wyeth for his studies. Richard Atkinson, the president of the American Obesity Association, which received money from Wyeth/Interneuron, strongly defended the drugs and aired that the study in the New England Journal of Medicine was inadequate. The American College of Cardiology issued a press release declaring that the heart problems disappeared once patients stop taking the pills. This was a blatant lie.

  Hired moles asked their colleagues for the medical data that had showed valve disease without revealing they were sent by Wyeth, in one case even indicating they worked for the FDA. The company also launched campaigns trivialising the harms in a hope of getting the drugs back on the market. A famous obesity specialist, George Blackburn, gave many pep talks but filed a sworn affidavit in Boston’s court that he had not given talks and had not received any money from the companies. When confronted with his lies and whereabouts, he didn’t remember anything.

  Wyeth could have warned the public years before independent researchers found out about the harms. Another big company, American Home Products, which marketed Pondimin, behaved similarly badly. It had 160 cases of pulmonary hypertension buried in-house while patients were still being prescribed Pondimin. Starting in April 1996, American Home Products even circulated an internal monthly memo called the Pondimin Monthly Death List. The company obstructed justice by destroying thousands of documents and emails after an order was issued by court not to do this. American Home Products denied it had done anything wrong, denied that it had known earlier that its drugs might be dangerous and said that ‘We never even promoted Pondimin.’

  The only thing that is missing in this soap opera is a denial that the dead patients had ever existed. Perhaps even the company didn’t exist but was just a figment of our imagination in line with social constructivism?

  When the plaintiffs’ lawyers got access to Wyeth’s archives, they had close to three million pages copied into computers so that they became searchable. This was an amazing feat. If we stack so many pages, the height of the stack will be about 300 metres! The lawyers found 101 reports of pulmonary hypertension and more than 50 cases of valve disease, which Wyeth had marked as something else. After the FDA had refused to approve Redux at its first meeting, Wyeth sent a document to a different bureau in the FDA, where 52 cases of pulmonary hypertension were well hidden on a little graph in a 40-page document. This Wyeth had the nerve to call ‘disclosure’.

  An FDA investigation at Wyeth’s headquarters uncovered that Wyeth’s safety officer had written over the first 13 reports of valve disease the company had received from the Mayo Clinic about Fen-Phen and used the same log numbers for other drugs and less serious adverse effects.35 However, instead of a criminal investigation the FDA wrote to Wyeth that its reporting system failed to assure that all reports are accurate. This was to put it mildly, but Wyeth’s lawyers protested, and a second FDA letter was apologetic about the first letter and politely urged Wyeth to clean up its act. I wonder what kind of society we would have if this was how the police was supposed to address a murderer: ‘Dear little thing, we would be so happy if you wouldn’t do it again. Please accept our profuse apologies that one of our officers accused you of murder and have a nice day.’

  There were other revelations. When an FDA officer had threatened Wyeth that if they didn’t warn the physicians about neurotoxicity the FDA would, Wyeth went to the top at the FDA and no warning letter was ever sent. It seems that the top of the FDA is capable of almost anything that benefits the drug companies. In 1994, the FDA decided at a meeting that a black box warning was needed for Pondimin telling about 50 cases of pulmonal hypertension, but an addendum to the minutes said that nothing would be done anyhow, without any explanation. An FDA scientist produced a report in 1999 that showed exactly what information the companies had given the FDA about valve disease and when, but FDA lawyers made sure the FDA could not be incriminated by locking it away in a drawer.

  As for reports of adverse events, the FDA had
left it to the companies to decide themselves whether an event was serious and what to mention first, which had the effect that many cases of reported valve problems were overlooked by the seriously understaffed FDA safety division, as they weren’t mentioned on the first page. During court proceedings, 52 cases of left valvular disease (which cannot be caused by pulmonary hypertension, as this affects the right valves) were discussed, and none of them had been coded as valvular disease. The company had also deceived the FDA originally about its animal studies. The valves of the rats’ hearts had thickened dramatically and stiffened, but this was hidden under the rather innocent term ‘focal fibrosis’, which was camouflage for the real thing. What the company told the FDA was only the good news: the rats didn’t develop cancer. Marion Finkel, the FDA official who had originally approved Pondimin, but now consulted for drug companies, tried her best to put her client in good light.

  Pulmonary hypertension is a terrible disease, and symptoms can start already after a week on the drug. It’s uniformly fatal, with a mean survival less than many cancers, and the symptoms feel like strangulation or drowning. The valve disease is similarly devastating. At the time of the mass tort lawsuits, an estimated 45 000 US women were believed to have developed one or both of the two diseases,35 with an expected death toll of the same size.