Narcolepsy and Pandemrix Vaccine – an association in adults as well as childrenArchived

Vaccine Adverse Event Surveillance and Communication (VAESCO) Consortium & ECDC Narcolepsy in association with pandemic influenza vaccination – a multi-country European epidemiological investigation Full Report - ECDC, 20 September 2012

L'Agence nationale de sécurité du médicament et des produits de santé (ANSM) Service de pharmacologie (INSERM CIC-P 0005 Pharmaco-Epidémiologie), Université Bordeaux Segalen – CHU de Bordeaux aussi Etude NarcoFlu-VF (NarcoFlu VAESCO-France) : Grippe, vaccination antigrippale et narcolepsie : contribution française à l’étude cas-témoins européenne. Août 2012 (published 20 September 2012)

In September 2010 Sweden and Finland noted a number of children had developed narcolepsy seemingly in association with having received the pandemic vaccine used almost exclusively in those countries (Pandemrix) - see the ECDC timeline. As the pandemic had been declared over very little of the vaccine has been used since then. The European Medicines Agency took measured regulatory action and asked for careful studies to be undertaken to be undertaken. The results of these are now becoming available including both the ECDC-led VASECO 8-country study with the Brighton Collaboration and various national investigations. Last week the first results of VAESCO became available and in parallel with that a related French study NarcoFlu VAESCO-France.

The VAESCO study:

On 20 September the VAESCO Consortium issued the first major report of its epidemiological studies of narcolepsy in Europe and its possible relationship to pandemic influenza vaccines. These comprise two distinct studies, one on narcolepsy background incidence rates and changes in incidence rates over time and one case-control study to determine risk factors. Both studies concern the association between narcolepsy and influenza A(H1N1)pdm09 vaccines while at the same time considering other likely causes. Collaborators and data come from eight countries: Denmark, Finland, France, Italy, the Netherlands, Norway, Sweden and the United Kingdom. These studies are an extension of planned VAESCO post-authorisation studies of potential adverse events following immunisation (AEFIs) such as Guillain Barre syndrome, convulsions etc which were planned following the declaration of the 2009 pandemic ahead of the introduction of the pandemic vaccines. [1]

Following the surprising discovery of a rise in diagnoses of narcolepsy and cataplexy in children and adolescents (5 to 19 years of age) in Sweden and Finland in association with the vaccine Pandemrix (the only pandemic vaccine used in those countries) the studies were rapidly extended and adapted to include narcolepsy and cataplexy.[1] There are many findings in the 160 page preliminary report which had already been shared with the EU Member States, the European Medicines Agency (EMA) and the World Health Organization (WHO).[2] One of the most important was confirming with different epidemiological methods the association between vaccination with Pandemrix and an increased risk of narcolepsy in children and adolescents (5 to 19 years of age) in Sweden and Finland (signalling countries). New findings arise from pooling data and from performing sensitivity analyses. According to the strictest primary analysis (using the date for referral to a sleep-laboratory as date of onset of disease) that aims at avoiding most regulatory, media and diagnostic awareness biases, no significant risk was found in data concerning children and adolescents from the other six countries included in the study – Denmark, Italy, France, the Netherlands, Norway and the United Kingdom (non-signalling countries). Equally no association was found with adult narcolepsy in either signalling or non-signalling countries.[2] However the sensitivity analyses, that assess the robustness of the results obtained in the primary analysis have shown the importance of time-related factors to confirm the strength of association between influenza pandemic vaccine and narcolepsy. One such sensitivity analysis was based on using the date of onset of excessive day time sleepiness as date of start of disease, and only used cases with an onset before media attention. This identified an increased risk for narcolepsy for children and adolescents following influenza A(H1N1)pdm09 vaccination in both signalling and non-signalling countries. In addition, a similar sensitivity analysis showed an association in adults in the non-signalling countries – also only including cases falling ill before increased awareness.[2]

NarcoFlu VAESCO-France (ANSM) study:

The French study is a national variant based on the VAESCO study protocol. It was undertaken by the new French regulatory agency for medicines L’Agence nationale de sécurité du médicament et des produits de santé (ANSM) with an academic clinical and epidemiological consortium.[3] The French study used intensive case-finding for narcolepsy through fifteen sleep centres and applied the case-control analytic approach. It is important to note that in France two pandemic vaccines were used, the adjuvanted Pandemrix (GSK - around 4.1 million vaccinations) and the unadjuvanted Panenza (Sanofi-Pasteur – around 1.6 million) and the intensive case finding for narcolepsy with clinical validation was applied to adults as well as children according to the VAESCO study protocol. The report details how the conventional but intensive pharmacovigilance work found 50 cases of narcolepsy AEFI where the brand of pandemic vaccine was known, 22 were aged 16 years and over and another 28 cases were in children aged 8 to 15 years. Most exposed cases had received Pandemrix®-GSK (1.15/105 vaccinations)[5] while only 4 had received Panenza®-Sanofi Pasteur (0.19/105)[5]. All cases met the newly developed narcolepsy Brighton collaboration case definition for the VAESCO study protocol that was produced in collaboration with European narcolepsy experts. In thecase-control study the investigators looked at patients attending sleep centres with validated narcolepsy and compared their risk factors with up to 4 controls per case. Controls were selected from the population of patients managed in the same centre as a case, matched for age and gender but not having a disease related to narcolepsy or a condition indicated for influenza immunisation. A multivariate case-control analysis found an association with pandemic vaccination (either Pandemrix or Panenza, but almost all were Pandemrix) in both those 19 years and younger (odds ratio 5.09 [95% CI 2.1 to 12.26]) and in older patients OR=(3.88 [95%CI 1.37-10.95]). The narcolepsy cases associated with pandemic vaccination were younger at onset than cases without vaccination (median 14.5 vs 20.2 years) and cataplexy was more common in those vaccinated than unvaccinated (91% vs 57%). There was no association of narcolepsy with receipt of seasonal influenza vaccination. Indeed the proportion receiving seasonal vaccine was significantly lower in the cases than in the controls (8.2 % vs. 18.8 %, p=0.01). Looking at time periods there seems to have been no strong effect of media attention.[3]

 ECDC Comment, 21 September 2012:

The history of this work and event goes back almost exactly two years (see the detailed timeline on monitoring of vaccine safety for 2009 pandemic vaccines) when in August 2010, Sweden and Finland alerted the rest of Europe to cases of narcolepsy in children and adolescents after vaccination with Pandemrix.

Narcolepsy is a particularly difficult condition for study of this kind of association. The clinical and epidemiological reports from Finland, Sweden, Ireland and now France, as well as the VAESCO results are highly convincing of a new phenomenon in children and adolescents that appeared abruptly in children following the start of immunisation and before media attention.[4,5,6,7] The relative risks and odds ratios indicate a strong relationship even though the absolute risk is small (about 1 in 20,000 injections), a rate that would make this phenomenon impossible to pick up in initial trials.[4,5,6,7] The clinical descriptions are of a distressing and handicapping condition that would not be missed.[4] Beyond that its one of the more difficult conditions for this kind of study. Its incidence is low in adults and very low in children, though the peak age of onset is in late adolescence early adulthood (15-30 years determined in the VAESCO background incidence rate study). It is not always suspected by clinicians and time to diagnosis may be years. Diagnostic tools available have improved significantly over the last years and today multiple sleep latency test and determination of hypocretin-1 in cerebrospinal fluid is available in many countries/regions. However, accessibility to the laboratory tests may be an issue. Narcolepsy has never before been linked to vaccination. Narcolepsy can have a gradual onset and public and professional awareness are important factors in cases coming to light. Unfortunately initial underreporting to the European Medicines Agency (EMA) EudraVigilance database of the suspected AEFIs complicated and delayed the regulatory evaluation. Only one case of narcolepsy associated with pandemic influenza vaccines had been reported to the EMA by the time the possible safety signal was announced publically by Finland and Sweden where many cases had already been diagnosed. Consequently, it is understandable that a surge of reports occurred following the professional awareness and subsequent media attention in both signalling and non-signalling countries, however with a delay with reports from the non-signalling countries. Due to consequent delays in diagnosis and reporting, new cases associated with Pandemrix are still being reported to EudraVigilance in 2012, some of them with first symptoms dating back as far as 2010. Sensitivity and specificity are also issues. Though there is now a strong case definition it is a condition that requires specialist diagnosis often made in out-patient clinics. Primary care datasets may therefore miss it and case validation is crucial. Sometimes cases coded as ‘narcolepsy’ in data-sets turn out not to have that diagnosis confirmed when validation is attempted.

The findings of paediatric cases associated with Pandemrix in France [3] and Ireland [6] emphasises that this is not just a Nordic issue. It will be important that these new national studies evolve into peer-reviewed articles (that is also the case for Sweden[5]) but this should be expected given the strong national groups behind the studies and in the VAESCO core team. Indeed there is an impression growing that where there is intensive case finding with case validation the Pandemrix – narcolepsy relationship can be found in many parts of Europe in children at least, “seek and you shall find”. The French finding of something similar for adults, and weaker hints of this in the main VAESCO sensitivity analyses exposes how much of the surveillance has focused mostly on children appearing in sleep clinics. Countries will now have to investigate whether there is also an association for narcolepsy in adults such as Finland is doing. Though this may be even more challenging than for narcolepsy in children because of the long time from onset to diagnosis often observed in adults, with the bias this can bring.[8] Age specific risks will need to be calculated since a plausible hypothesis is that there is not an abrupt cut-off in risk at age 20 but a more gradual decline in risk with age. The phenomenon is not one that can be explained by genetics since though there is a strong relationship with a particular HLA type (DBQ1*0602) that type is widely spread across Europe. An especially important and reassuring French finding is the lack of association with seasonal influenza vaccination.[3]

The lack of association with unadjuvanted Panenza in France contrasts with the strong relationship with adjuvanted Pandemrix but it must not be over-interpreted. It is not necessarily the case that the adjuvant is the cause here. There are oral reports from a recent scientific conference held in Paris that narcolepsy cases are not being seen in a study in Canada where Arepanrix, a very similar AS03-adjuvanted influenza A(H1N1)pdm09 vaccine(GSK) was used.

So the narcolepsy and pandemic vaccine story is not over. Indeed it has a long way to run. The studies need to continue and the VAESCO report makes a series of logical recommendations for this including

• more cases collected from the period before awareness increased and including more European countries that had significant vaccine coverage,
• pooling national studies that weren't included in report, including those from Ireland, the United Kingdom, and Germany.
• expanding the investigation to countries outside Europe—such as Canada and Brazil—that used pandemic vaccine with the AS03 adjuvant but didn't have as much media attention about the narcolepsy link.
• mechanistic studies including animal work to try to understand causality

Since September 2010 hardly any pandemic vaccines have been used in Europe. There was no need to, the pandemic was declared over in August 2010, and the new seasonal trivalent vaccines became available in September 2010. All the Pandemrix vaccines have now expired. However the question-mark over the specific adjuvant and its implications for other new adjuvants is an important public health issue that still has to be resolved. One of the global pandemic influenza assumptions is that adjuvants will increase the global pandemic vaccine capacity three or four fold. New pandemic vaccine contracts are being developed and it will be essential to know whether to include adjuvants in these or not. The 1976 association of a swine flu vaccine with Guillain Barre syndrome cast a long shadow over pandemic influenza vaccines and surveillance for GBS was the prime objective of AEFI investigations in the 2009 pandemic [9,10,11]. The 2010 narcolepsy and pandemic vaccine association may do the same.