Association of childhood narcolepsy and cataplexy and receipt of a pandemic influenza vaccine in Finland in 2009-10 – First papers in a peer review journalArchived

A pair of linked papers from Finland concerning the above topic, one epidemiological and the second combining both clinical and epidemiological were published on March 28th in the open access journal, PLoS One.(1,2). Narcolepsy is a rare chronic, genetically-associated sleep disorder featuring excessive daytime sleepiness frequently including cataplexy, a sudden and transient loss of muscle tone triggered by emotions. The diagnosis of narcolepsy is confirmed by typical sleep polygraphy findings and low CSF-orexin/hypocretin concentration. Narcolepsy is of gradual onset, and often difficult to recognise. Its diagnosis is unusual in children.(2,3)

Cases of narcolepsy in children and adolescents started to be observed by paediatricians and neurologists working in the 5 sleep clinics in Finland in 2010. This was soon after the vaccination with a specific pandemic vaccine in late 2009, and the pandemic influenza epidemic of the autumn of 2009. No increase in narcolepsy was noticed in other age groups.(1,2,4)  Finland was one of the European countries where a substantial percentage of the population was immunised, including children and adolescents among whom coverage was highest (around 75%). Almost all vaccination was undertaken in weeks 44 to 52 of 2009.(1,5) The only vaccine used in Finland was Pandemrix®, a split virion, inactivated influenza vaccine with an AS03 adjuvant. In August 2010 Finland formally recommended stopped use of Pandemrix and the European Medicine Agency (EMA) started an investigation under EU legislation.(4)   

The Epidemiological Study used a retrospective cohort study design.(1) Data for the period between January 2009 to December 2010 inclusive were collected from the cohort of all children living in Finland and born in the years 1991 to 2005. Vaccination data of the whole population was obtained from primary health care databases. All new cases with assigned ICD-10 code of narcolepsy were identified and the medical records independently reviewed by two of the authors who were experts in the field of child sleeping disorders. This was to validate the diagnoses of cases and classify them according to the three point scale of the Brighton collaboration (Table here). Onset of narcolepsy was defined as the first documented contact with health care due to excessive daytime sleepiness. The primary analysis period for contact with the health services was restricted to end in mid-August 2010 as after that media attention to post-vaccination narcolepsy began in Finland following reports in neighbouring Sweden.(4,5)  About 75% of the child cohort had been vaccinated and in the epidemiologic study after case validation 46 Pandemrix® vaccinated and 7 unvaccinated children were included.  The incidence of narcolepsy was 9.0 / 100,000 in the vaccinated as compared to 0.7 /100,000 person years in the unvaccinated individuals. This gave a rate ratio of 12.7 with 95% confidence intervals from 6.1 to 30.8. The vaccine-attributable risk of developing narcolepsy was 1:16,000 vaccinated 4 to 19-year-olds (95% CI 1:13,000 - 1:21,000).  A series of sensitivity analyses were undertaken, especially concerning possible alternative dates of onset of narcolepsy (as recalled by the family, first contact with health service, referral to a specialist and diagnosis) and these essentially made little difference to the findings.

The Clinical Study was based on diagnoses with narcolepsy in people in Finland of all ages over the period 2002-2010.(2) It found a substantial increase in narcolepsy diagnosis in those under 19 years in 2010, but no increase in those over 19 years (See Figure). There were 54 children (under age 17 years) diagnosed with narcolepsy in 2010 and 50 of these had been vaccinated with Pandemrix®. What follows focuses on those fifty immunised children and adolescents.  There was no marked sex difference. All children were of Caucasian origin. The onset date of narcolepsy is always difficult to determine because of its insidious onset but the mean age at reported onset of narcolepsy was 11.0 years with a standard deviation of 3 years. The range of age in affected individuals at the time for pandemic vaccination was between 4.5 and 16 years.  None of the fifty had a reported prior history of sleep disorder. Thirty four children of the children were HLA-typed and all were found to be positive for the known narcolepsy risk allele which is considered to be present in about 28% of the Finnish population, which is similar to figures in much of Northern Europe (it is somewhat lower in Southern Europe). Forty seven (94%) of the children developed cataplexy and 26 out of 49 experienced hypnagogic hallucinations (visual, auditory or sensory hallucinations such as seeing colours and shapes at the transition from sleep to wakefulness or wakefulness to sleep). Where weight was recorded 26 out of 41 showed a rapid gain in weight (defined as more than 5 kg gain) in the beginning of their illness.  Only four of the fifty children had experienced problems with hyperactivity before onset of their narcolepsy. However, twenty four (48%) of the children showed behavioural change which needed psychiatric treatment after development of narcolepsy. The onset of narcolepsy was at a younger age in the vaccinated children compared to children with narcolepsy not associated with vaccination: 12.1 years (95%CI 11.2-13.0) versus 15.3 years (95%CI 11.8-18.7), and the progression to cataplexy was faster: mean 35.9 days (95%CI 23.3-48.5) in the vaccinated children versus mean 546.2 days (95%CI 48.7-1043.5). The time from onset of symptoms to diagnosis was much longer in the in the pre-vaccination era.

ECDC Comment (March 29th 2012)

These two papers make a strong case for an association between pandemic vaccination and onset of narcolepsy with cataplexy in genetically susceptible children in Finland. Though the vaccination attributable risk is low (only 1 per 16,000 childhood vaccinations) the risk ratio is very high and highly statistically significant. It is important to note that though this paper describes fifty children there are now over 100 children and adolescents and their families in Finland who are suffering from a distressing and handicapping chronic physical condition seemingly in association with pandemic vaccination. A proportion of the children also have significant psychological and behaviour difficulties.

Equally it should be noted that the investigators have delivered high quality analytic studies in a difficult political climate following the appreciation of the phenomenon in Finland. The researchers have done as much as they can in the context of the setting to exclude biases that might reasonably have been expected to result in a spurious association, notably referral bias following the extensive coverage of the issue in the Finnish media in August 2010.  Also, similar findings have been reported in Sweden, although these have yet to undergo the same kind of rigorous analysis and peer review process.(5)

However, many scientific and clinical questions remain. A difficulty here is that only one pandemic vaccine was used in Finland so it is unclear if the association should be stated to be with pandemic vaccination or the specific vaccine Narcolepsy has never been reported as an adverse event following vaccination. What could the mechanism be behind development of narcolepsy following vaccination  Is it important that an adjuvant was in this vaccine The authors in the clinical study speculate about an auto-immune process underlying this phenomenon. This is not a new idea and has been pursued as a possible biological mechanism for narcolepsy and cataplexy for some years.(6)  What will be the natural history of the narcolepsy and the associated behavioural difficulties and would these children/adolescents have developed narcolepsy anyway at a later age Why has the association been observed in Finland and Sweden but not confirmed so far in other countries where Pandemrix and other vaccines were used There have been no reports of narcolepsy linked to vaccination from Canada, where a similar vaccine (Arepanrix®) containing the AS03 adjuvant was used.  Could there have been another infection circulating at the time of vaccination – or even the pandemic virus (H1N1)pdm – and was it the combination that was dangerous, and not the vaccination alone  Some of these questions can be answered in Finland or Sweden but others will need a pan-European or global approach. Fortunately there is now time for studies that will answer these important questions as use of Pandemrix had almost ceased even before the prompt announcement of the phenomenon by Finland in August 2010. Immunisation was moving back to use of seasonal unadjuvanted triple vaccines in Europe (mostly in adults) and few pandemic vaccines – including Pandemrix – were used after the pandemic was declared over in August 2010.(4,7) 

After appreciation of the phenomenon, ECDC and the Brighton Collaboration used ECDC emergency funding to rapidly expand their already on-going VAESCO pandemic vaccine-safety programme (4) to include narcolepsy, and started working with 8 countries including Finland and Sweden. At the same time national studies began in some of those countries. A feature of all studies, including the one now reported from Finland, is that they are publicly funded and considered together can look at multiple hypotheses, though none of them are easy because of subtle onset of narcolepsy and the need for careful case finding and validation.

All this demonstrates the need for rigorous post-marketing surveillance of vaccines in Europe and it is important to note the landmark that as of July 2012 the European legislation comes into force whereby the EU Regulatory Body the European Medicines Agency will start receiving notification of all adverse events following immunisation, not just those associated with centrally authorised products