Complaint to the European Medicines Agency (EMA) over maladministration at the EMA 

 

Updated: 1 June 2016

 

  

According to Article 6 of the EU Treaty and the Charter of Fundamental Rights of the European Union (1), “Openness enables citizens to participate more closely in the decision-making process and guarantees that the administration enjoys greater legitimacy and is more effective and more accountable to the citizen in a democratic system. Openness contributes to strengthening the principles of democracy and respect for fundamental rights.” 

​

On 26 November 2015, the European Medicines Agency (EMA) released a 40-page Assessment Report dated 11 November (2) on the safety of vaccines against human papilloma virus (HPV), which is supposed to decrease deaths from cervical cancer. 

​

We are concerned about the EMA’s handling of this issue as reflected in its official report (2) and ask the EMA to assess: 

1. Whether the EMA has been open and accountable to the citizens and has respected their rights to know about the uncertainties related to the safety of the HPV vaccines. 

2. Whether the EMA has lived up to the professional and scientific standards that must be expected of the agency to guarantee that the administration enjoys legitimacy when evaluating the science and the data related to the safety of the HPV vaccines. 

3. Whether the EMA has treated fairly - in a manner that guarantees that the administration enjoys legitimacy - a Danish whistleblower, PhD Louise Brinth, when she raised concerns about possible serious harms of the HPV vaccines. 

4. Whether the EMA has treated fairly - in a manner that guarantees that the administration enjoys legitimacy - the observations and concerns the Danish Health and Medicines Authorities and the Uppsala Monitoring Centre had raised about possible serious harms of the HPV vaccines. 

5. Whether the EMA’s procedures for evaluating the safety of medical interventions guarantee that the administration enjoys legitimacy. The EMA asked the manufacturers of the vaccines to assess potential harms of their own products in which they have huge financial interests. 

6. Whether the extreme secrecy, with life-long confidentiality agreements, which the EMA imposed on its working group members and scientific experts, is needed; is legitimate; is in the public interest; and guarantees that the administration enjoys legitimacy. 

7. Whether the redactions the EMA imposed on documents it delivered to the citizens according to Freedom of Information requests were needed; were legitimate; are in the public interest; and guarantees that the administration enjoys legitimacy. 

8. Whether the EMA has behaved in a manner that guarantees that the administration enjoys legitimacy in relation to declaring conflicts of interest. We noticed a Guido Rasi’s name associated with patents for inventions and wonder whether this is the same person who is the EMA’s director. If so, we believe Rasi has failed to declare his conflicts of interest. We also believe that the rapporteur for the EMA’s report, Julie Williams (2), has failed to declare her conflicts of interest. 

9. Whether the EMA behaves in a manner that guarantees that the administration enjoys legitimacy when the agency use experts with financial ties to the manufacturers, in particular considering that it is always possible to find experts without such conflicts. 

10. In the interest of transparency, we urge the EMA to ensure that the names of all the experts consulted are disclosed together with their conflict of interest declarations. We also urge the EMA to ensure that the conflicts of interest statements from the rapporteur, the co-rapporteurs (Jean-Michel Dogne (BE) and Qun-Ying Vue (SE)), their contact persons at the EMA and everyone else who has given statements to the EMA are brought out in the open. Finally, we urge the EMA to ensure that Declarations of Interests for officials at the EMA are honest..

​

Outline of the case 

After Denmark had contacted the European Commission, and the Commission requested the EMA to give its opinion, the Danish Health and Medicines Authorities (DHMA) asked the EMA in July 2015 to assess the research linking HPV vaccines to serious harms, which included peer reviewed articles published by PhD and physician Louise Brinth from the Danish Syncope Unit at Frederiksberg Hospital in Copenhagen describing possible serious neurological harms of the vaccine. In its submission to the EMA, the DHMA included a review of the global data made by the Uppsala Monitoring Centre (UMC, a WHO collaborating centre). 

​

The prominent symptoms, which are suspected of being caused by the vaccine, are similar to those seen in so-called functional disorders such as chronic fatigue syndrome (CFS) and include postural orthostatic tachycardia syndrome (POTS) and chronic regional pain syndrome (CRPS). The hypothetical mechanism is an autoimmune reaction triggered by either the active component of the vaccine or the adjuvant in the vaccine. These syndromes are difficult to diagnose; their causes are poorly understood; and they are likely to be underreported. This complicates studies of a causal link. 

​

The EMA’s official report (2) gives the impression of a unanimous rejection of the suspected harms. However, only seven months earlier, the EMA had resolved that “A causal relationship between the dizziness and fatigue syndrome, Postural Orthostatic Tachycardia Syndrome (POTS) and Gardasil [one of the HPV vaccines] can neither be confirmed nor denied” (3). Moreover, the EMA’s internal report of 256 pages (4), which provided the draft for its 40-page official report, tells a very different story. The internal report is confidential but has been leaked. 

​

There are many problems with the EMA’s official report, and Louise Brinth has drawn attention to them in a 63-page document from 17 December 2015 (5). Brinth did not have access to the confidential internal report (4) when she wrote the document, only to the official one (2).

​

Brinth’s observations 

1. The EMA asserts that: “Overall, the case series reported by Brinth and colleagues (2015) is considered to represent a highly selected sample of patients, apparently chosen to fit a pre-specified hypothesis of vaccine-induced injury” (2, p24). 

Brinth states in her report that she and her co-workers included all consecutively referred patients, with the exception of the ones excluded because they met the exclusion criteria (5, p17). 

We find that the EMA’s allegations constitute guesswork (“apparently”), are pejorative and come close to an accusation of scientific misconduct. The EMA’s comments are totally inappropriate for an EU authority to make on honest researchers and the criticism is furthermore totally unfounded. 

 

2. EMA: “many of these symptoms would require some sort of objective clinical evaluation, yet there is no information on how this was done or what other clinical assessment may have been undertaken to exclude other causes of the symptoms” (2, p23). 

Brinth: Nausea, headache, abdominal pain, fatigue etc. are in essence subjective symptoms and as such they escape “objective clinical evaluation”. Orthostatic intolerance was quantified through tilt-testing, which constitutes more than “some sort of objective clinical evaluation” ... “altogether I find the criticism put forth with regard to my work and my scientific approach unjust. I find that the evaluation of my work presented by EMA is based on many assumptions. Most of these assumptions are to my best knowledge wrong. I find this approach strange and unscientific … In the EMA report, it is concerning that the complete data set was not included. Only specific parts of the report have been reproduced, and it is not defined how these parts were chosen. As a result, there are a number of examples of inconsistency between the representation of UMC data in the original DHMA report and the EMA report” (5, p25-26). Brinth furthermore writes that her observations and hypotheses were supported by the independent review of the global data made by the Uppsala Monitoring Centre (5, p26). 

We find that the EMA’s comments are unprofessional, misleading, inappropriate and pejorative, and that the EMA’s approach involves cherry-picking, which is unscientific. Furthermore, the Uppsala centre compared reports of potential harms from the HPV vaccines with reports of similar harms from all other vaccines offered to women and found that POTS was reported 82 times for HPV vaccines versus once; CRPS 69 times versus 16; autonomic nervous system imbalance 77 times versus 16; and fibromyalgia 62 times versus 39. Several other comparisons were made but did not show differences. In the EMA’s opinion, this does not allow any conclusions to be made (4, p230 in the pdf, or 51/77 in the subdocument). The rapporteur noted that the analyses were not corrected for multiple comparisons (although that would not have made a ratio of 82 to 1 disappear) and that the higher frequency of harms could be due to media attention (no evidence for this was presented). The finding by the Uppsala centre that a substantially higher proportion of case reports were serious for the HPV vaccines compared to other vaccines was mentioned but not paid attention to in either of the EMA reports (2, 4). 

​

3. EMA: Chronic fatigue syndrome has “been reported relatively constantly since 2009” (2, p27). 

Brinth: According to the UMC report, chronic fatigue syndrome has “been increasing since 2012 with a marked increase between 2012 and 2013” (5, p27). 

We believe that the EMA misrepresents the UMC report with regard to possible serious harms of the vaccine (see more examples below). 

 

4. The EMA failed to mention that the Danish reports were more often classified as “serious” and had more often included a large amount of clinical relevant information. Instead, the EMA mentioned that “the terms POTS, orthostatic intolerance and autonomic nervous system imbalance are reported disproportionately more in HPV reports from Denmark vs HPV reports in other countries” (2, p27). 

Brinth: The UMC noted that “a significantly greater proportion of the reports from Denmark were considered “good reports”; were classified as “serious”; and were received from either a physician, consumer or a lawyer (5, p28). There were a number of potential side effects, which were statistically significantly over-reported with the HPV vaccine, but which the EMA failed to mention: disturbances in consciousness, muscular weakness, disability issues, and neurological signs and symptoms (5, p29). 

We find that the EMA, rather than praising the Danish diligence, casts doubts on whether the Danish peer reviewed research should be believed. This is an inappropriate attitude to safety for a drug agency. We agree with Brinth that the EMA misrepresented the UMC review in its own report, thereby creating the impression that Brinth should have published substandard research. It seems to us that it is the EMA that uses and accepts substandard methods, not Brinth (see more examples below). The EMA omitted some of the important suspected harms of the vaccine, which is another example of cherry-picking. 

 

5. Brinth: “As far as I can see, the minutes of the [EMA] meetings are not released. What data did they get to review? Did they write a report for the PRAC [Pharmacovigilance Risk Assessment Committee] to use? And if they did where is the report – can we see it? To my best knowledge the conclusion of the meetings and the report put forth by EMA is presented as if it is the result of an unanimous discussion and opinion of the SAG [Scientific Advisory Group] and these experts. This makes us unable to judge if all the experts agreed on the conclusion of the final report. I find this a very strange, unscientific and undemocratic approach. In a matter as complex as this I believe that not two single scientists will share exactly the same viewpoints … Creating a sort of false consensus based on the participants being bound to secrecy will inevitably lead to the impression that “all these many experts agreed on the conclusion set forth by EMA”” (5, p33). 

 

We agree with Brinth. The EMA’s official 40-page report (2) is misleading, as it gives the citizens the impression that there is nothing to worry about in relation to vaccine safety and that the experts consulted by the EMA agreed on this. However, the EMA’s confidential, internal report (4) reveals that several experts had the opinion that the vaccine might not be safe and called for further research, but there was nothing about this in the official report. We find that the extreme level of secrecy imposed by the EMA on its working group members and other experts is inappropriate and goes against the public interest in openness and transparency about possible serious harms of the vaccine. In the EMA’s internal report (4), there is a “Confidentiality Reminder” on page 2: 

“As an EMA expert you are bound to life-long duty of confidentiality. The duty of confidentiality applies to all information of the kind covered by the obligation of professional secrecy. This includes, for example, the fact that there is a meeting, that you have been nominated to participate, the agenda of the meeting, the product or company concerned, the participants, any part of the discussions and outcome. All documentation (correspondence, reports, minutes, etc.) must be kept as confidential and stored in a secure place or destroyed. The duty of confidentiality stops only if information has been made public and only to the extent that has been released into the public domain.” 

 

6. Brinth: “Reading EMA’s report it seems to me that the review of the safety data which includes available data from the clinical trials and post-marketing surveillance and the literature review was all collected by the marketing authorization holders (MAHs) [GlaxoSmithKline Biologicals, Merck Sharp & Dohme Limited, and Sanofi Pasteur MSD]? This must have been in the form of a report written by the MAHs to the EMA?. Is this report freely available?” (5, p34). 

We find it unacceptable that the EMA in its official report (2) did not make it clear that it allowed the drug companies to be their own judges (4) when evaluating whether the vaccine is safe, particularly since there is a huge amount of money at stake: The global expenditure so far on HPV vaccines can be roughly estimated at €25 billion. The EMA asked the companies to search for side effects of the vaccine in their own databases and did not check the companies’ work for accuracy. This is not an acceptable procedure. There are countless examples of drug companies hiding serious - even lethal - harms from the authorities (6) (see also under “8. Brinth” below). 

 

7. Brinth: “Were there any opportunities for PRAC-members, SAG-members or other assessors to ask additional questions or for clarifications? I find it highly problematic that we as readers of the EMA reports – and I as medical professional having diagnosed and evaluated a substantial part of the Danish Cases – my diagnoses and work altogether downed and overruled by MAH/ EMA – have not access to their case detection methods ... I know that a substantial proportion of the POTS-cases reported as suspected side effects from Denmark have been done by me. I therefore find it troubling and strange to see that POTS-cases have been overruled and judged as not meeting or only partially meeting the diagnostic criteria in 50 out of 83 cases given the diagnosis POTS in the AER [Adverse Events Reports] … I interpreted the report as if it is the MAH who has evaluated the AER – and have found that most of the POTS cases do not meet the proper diagnostic criteria? The EMA report mentions the diagnostic criteria put forth by Raj and Sheldon … We use the exact same criteria and have experience in diagnosing and treating POTS – and are to some extent quite restrictive in our diagnostic practice. It can be discussed - as we do in our papers – whether POTS is a relevant diagnosis or not. However, that is a whole other issue. Therefore, we need to know: Has the MAH based their evaluation on the AER alone – or have they been through the whole medical record of these patients? It is well known that an AER will not include all the details of the clinical history and therefore it is rare that any spontaneous report will meet diagnostic criteria. Evaluating the diagnosis given in AER will be very difficult bordering on impossible if based only on the AER. But – if they did obtain full medical record – then the discrepancy between the diagnoses given by the clinicians seeing the patient and filing the report and the MAH is a more relevant discussion. But then this discussion should be out in the open. We need to close the gap between the reality as it looks like from the clinicians (and the patients) point of view and the MAH and EMA. Either I have misunderstood the whole diagnostic approach to POTS – and then I need to know. Or – the reports have been judged on a too lose background. Both scenarios represent - as I see it - a quite serious problem. We have a common interest in reaching consensus on how to use this POTS- diagnosis as it is now very much on the agenda … Our thoroughness underlined by the fact that WHO has stated that the Danish AER did not in the essense [sic] differ from reports from the rest of world – but were of a higher quality” (5, p34-35). 

​

We find it totally unacceptable that no one - not even the person whose research is so strongly and unfairly criticised in the EMA’s official report (2) - can judge on what grounds and by whom 50 out of 83 cases of POTS were dismissed by the EMA. Furthermore, the secret internal EMA report (4) reveals that the POTS cases were dismissed without having access to the full medical records. Clearly, the “research” carried out by the companies did not live up to the accepted standards for such research and the EMA did not ensure that it did. This is very serious, as the EMA used the companies’ assessments to overrule the research done by Brinth, which is of a much higher standard. The Danish Health and Medicines Authorities (DHMA) also criticised pretty strongly the EMA’s approach (7, p9-10): 

“The main conclusion in the Danish report is not, as described in the assessment, to change focus to 

CFS. Rather the review highlights the necessity to evaluate combinations of symptoms rather than only performing separate evaluations of individual diagnoses. It shows that although the number of POTS cases is very high in Denmark, compared to the rest of the world, the symptom pattern seen in the Danish dataset is similar to reports submitted from other countries. Even though it cannot be shown for certain at this point, it is likely based on these data, that patients with the same symptomatology would receive different diagnoses in different member states e.g. POTS in DK and CFS/ME in others. This consideration is important for the discussion of consistency regarding the POTS signal, where it is stated that the finding of the majority of POTS cases in Denmark does not support a causal relationship. We do not agree with this conclusion based on the data.” 

​

Immediately after this, the DHMA states: 

“Risk Management Plan/ Post-authorisation Safety Studies/ Conditions 

Need for further studies regarding the signal for POTS: 

We agree with the conclusion from the rapporteurs and also state in the Danish report [this shows that these comments come from the DHMA], that the data from spontaneous reports cannot be used to provide evidence for a causal relationship between symptoms and vaccination. 

However in view of the methodological limitations of the data available and the fact that the observed cases did exceed the expected cases, especially in Japan and Denmark, the conclusions should be cautious and the signal cannot be dismissed either based on the current evidence. 

We recommend that the vaccine SAG and expert meeting include a discussion of the need and 

possibilities to design appropriate PASS studies to explore POTS further. Similar question as Q3 

regarding CRPS.” 

​

8. Brinth writes that the EMA’s official report notes on page 12 that PRAC requested the MAHs to search not only “for reports specifically containing the terms POTS and CRPS” but also to use 

“common search strategies” “to identify possible cases of undiagnosed CRPS and POTS.” Brinth says it would have been relevant to have these search strategies clearly defined and given because, to the best of her knowledge, there are no common search strategies which have been previously defined for either POTS or CRPS (5, p35-36). 

​

We agree with Brinth. It is notoriously difficult to identify the harms in question and it is unacceptably poor standards for such research not to define precisely what the search strategies were. It is a fundamental requirement for systematic searches that the combination of search terms is clearly described (8), but not even in the internal 256-page report are these search strategies defined (4). This is extraordinary, as the companies have a huge vested interest in not finding these possible harms in their databases. When the DHMA in 2014 asked Sanofi Pasteur MSD to review its database for potential side effects of its HPV vaccine, the company searched for POTS in a way that made the number of cases retrieved very low (9). This was discovered by the Danish National Board of Health, partly because only 3 of 26 Danish reports of POTS showed up in the company’s searches. Sanofi Pasteur MSD had been asked to search on a number of specific symptoms including dizziness, palpitations, rapid heart rate, tremor, fatigue and fainting, but the company ignored the orders. Instead, the company searched on three symptoms: “postural dizziness”, “orthostatic intolerance” and “palpitations and dizziness." As terms used in reports of harms are the ones used by the doctors reporting them, unusual search terms will yield few results. 

​

9. Brinth: The EMA’s question 2 to the companies ”does not make any sense to me: If both the “vaccine group” and the “control group” received aluminium adjuvanted “placebo” or another aluminium adjuvanted vaccine as “placebo” – how does it make sense to ask the company to only discuss potential explanations including risk factors for the development if a difference is observed? By asking this question – I find that EMA actually states and take for granted that we know with a very high degree of certainty that we will not see side effects due to the adjuvant?” (2, p37). 

​

We agree with Brinth. In all the vaccine trials apart from a small one, the “placebo” contained aluminium adjuvant, which is suspected of being neurotoxic. It is therefore difficult to find a difference between harms of the vaccine and the “placebo,” but the EMA failed to address this fundamental problem in its official report (2). It is clear from the EMA’s internal report (4) that the MAHs simply lumped the results from trials with a genuine placebo with those that had a potentially neurotoxic “placebo": “Clinical safety data. For the purpose of the referral, the MAH was requested to provide an in depth review of the CRPS and POTS cases observed within all clinical studies. To respond to this request, the MAH has pooled the safety data from 18 completed and unblinded studies designed with an active comparator group (either placebo or another vaccine other than an HPV vaccine, i.e. Hepatitis B, Hepatitis A) which includes a total of 42,047 vaccinees (21,268 in HPV group and 20,779 in comparator groups)” (4, p119 in the pdf, or 7/67 in the subdocument). We believe this constitutes scientific misconduct, but the EMA accepted it nonetheless, without reservations: “Strength of the potential association. The few cases reported from RCTs [randomised clinical trials] are evenly distributed between the qHPV and placebo groups which does not suggest an association” (4, p20 in the pdf, or p11 in the subdocument). 

 

10. Brinth: The EMA asked the MAHs to provide an analysis of the observed number of post-marketing cases of CRPS and POTS in association with their HPV vaccine in comparison to those expected in the target population. The analysis should discuss the assumptions made with respect to the background incidence in the target population and also the influence of potential under-

reporting of cases in association with HPV vaccines. “I think the whole assumption underlying this question – that it is even possible to establish a reasonable estimate of the background incidence in the target population is a key issue. It is not possible for the time being to give a reasonable estimate of the incidence of these very underrecognized, underdiagnosed and poorly understood disease entities with very different diagnostic practices applied depending on nationality, medical specialty etc” (5, p39). 

We agree with Brinth. One of the key arguments in the EMA’s report (2) was that there was no difference between what was observed and the expected background incidence. However, the underlying research is of very poor quality, which renders this observation meaningless. For example, for some of the analyses, the observed incidence of chronic fatigue syndrome was used to estimate the expected incidence of POTS (4, p96, or 87 in subdocument). Furthermore, the EMA writes in its report that for POTS with the Gardasil/Silgard vaccine, the observed number of cases was generally lower than expected under almost all assumptions for all regions and countries except for Denmark (2, p17). This observation should have alerted the EMA to the fact that analyses based on expected incidence are grossly unreliable. We find it curious and scientifically unacceptable that the official EMA report (2) puts more weight on the “observed versus expected” analyses produced by the companies than the much more reliable Uppsala data. 

​

11. Brinth: The EMA asked the MAHs to provide a critical appraisal of the strength of evidence for a causal association between the HPV vaccine and CRPS and POTS considering the published literature (including epidemiological studies) and the possible causes and pathophysiology of CRPS and POTS, and to discuss whether there is a biological basis for a possible causal association. “I think it would be relevant to know if it was performed by the MAH only or supplemented by the EMA? I think it would be highly relevant to gain insight into the search strategies applied in the literature research. I have an interest in POTS, and have had for many years. I know that POTS is probably associated to autoimmunity, because this is the current perception in the scientific field of POTS. Publications are starting to emerge describing the findings of autoantibodies in patients diagnosed with POTS. The same autoantibodies are also found in CRPS and ME/CFS [myalgic encephalomyelitis/chronic fatigue syndrome]. I know that these studies are small studies, – but I think that they represent some very important information and a hypothesis generating body of evidence” (5, p39). 

 

We find it totally unacceptable to perform a literature review without giving the readers details of its methods, in particular the search strategies used. This is clear from all manuals about systematic literature reviews, e.g. the Cochrane Handbook (8). We also find that the EMA has ridiculed and dismissed the research performed at the Danish Syncope Unit in a way that is unfair, misleading, partly erroneous and pejorative. If drug agencies behave like this when doctors report their observations about possible serious harms of approved products, doctors will be unlikely to alert the public to their observations in future. This would be a tragedy for public health. 

 

Many redactions by the EMA in its documents are not legitimate 

 

Various people have obtained redacted documents from the EMA through Freedom of Information requests. As we have access to the unredacted version of the EMA’s confidential 256-page internal document (4), we can see which bits the EMA has redacted. We give some examples of redactions that we find unreasonable (our comments are in brackets): 

 

1) Names of contact people at the EMA for the rapporteur and co-rapporteurs. (This does not make sense, particularly since the names for the rapporteur and co-rapporteurs were not redacted).

2) Case numbers of patients for which harms were reported. (This is not necessary, as it is not possible to identify individual people from a case number, but it can make it difficult to assess a scientific report when such numbers are missing). 

3) Country names for individual cases. (This does not make sense unless the idea is to obscure for outsiders whether or not the EMA’s assessments are trustworthy, particularly since the EMA considers the Danish cases to be dubious).

4) Numbers of reported harms for individual countries, names of countries where there is an excess incidence of reported harms, and number of doses of the vaccine used in individual countries. (This is even more difficult to understand unless the idea is to obscure for outsiders whether or not the EMA’s report is trustworthy and whether or not the EMA’s criticism of the Danish Syncope Centre is warranted).

5) While it is indicated that some harms reports come from the Danish Syncope Centre, it is redacted that the centre is located at Frederiksberg Hospital. (This does not make any sense, particularly not when the name of the hospital was not redacted elsewhere).

6) The publication identifier for an article in press. (This is pretty meaningless unless the idea is to keep readers of the report in the dark. Scientists routinely refer to papers in press and it should not be hidden where such papers will appear). 

Here is an example of an absurd redaction

(7). It is clear to anyone familiar with this area that the word left out must be “Denmark.” The example also illustrates that there were far more uncertainty and disagreement than the EMA’s official report reveals. The co-rapporteur is obviously not happy with the approach taken. 

​

​