Saturday, October 20, 2012

European Report on Narcolepsy and Pandemrix gets thumbs down from Finland, France and Norway

From the September 2012 European ECDC report titled, "Narcolepsy in association with pandemic influenza vaccination" commissioned by the European Centre for Disease Prevention and Control, but produced by the Brighton Collaboration:

(page 149) Finland's Squalene -- ASO3 blockbuster:
"Contrary to previous studies, studies at THL suggest that squalene (AS03) produces an antibody response. Preliminary studies suggest that repeated immunisations in healthy subjects produce AS03 antibodies. Roughly 25% of narcolepsy cases (children) have antibodies against AS03.
Furthermore, Finland disagreed with the conclusions of the report, stating in Annex 1 (page 155):
our position is that fundamental concerns related to pooling heterogeneous data are not reflected in the key findings nor in the overall conclusions of the study. In the following, we present our concerns related to standardisation and pooling. Our position is that unsubstantiated pooling of data may lead to misinformed interpretations regarding e.g. diagnostic bias and, subsequently, erroneous overall conclusions. 
... The study combined data from areas with well- and less well-defined source populations which e.g. resulted in highly variable selection methods for controls in e.g. Italy, Netherlands and France. This fundamental problem is highlighted by the disclaimer provided by the French investigators. After excluding countries with low or zero exposure, and the potentially biased data from France (based on the disclaimer), the only country with exposure similar to the ‘signalling countries’ (Finland and Sweden) is Norway. The key findings should more clearly state these inherent deficiencies in power and lack of data comparability from all other countries except Norway.
Based on Finnish data alone, there is clear distinction between children and adults in onset definitions: in adults, confirmed cases diagnosed in 2009-2010 had a median lag of 5 years between EDS onset and referral to specialist. However, in children the median lag from onset to referral was 5 months."    
Here is the explanation for the low number of cases occurring in adults:  they take 12 times as long to present for diagnosis as children and adolescents, so most may not have been identified, yet.  Similarly, the Finns point out that cases identified early had an accelerated diagnosis compared to other, unvaccinated cases, but that this resulted in a spurious methodology:  later diagnosed cases were excluded from the analysis.  Yet in Finland, an additional 50 cases were diagnosed with narcolepsy in 2011, 47 of whom received Pandemrix, and these are almost certainly vaccine-related cases which have been excluded from statistical consideration in the ECDC report--Nass)

(page 149) France:
"The total number of observed cases is less than expected, but in the age group 10–15 years of age, it was higher: nine observed whereas 2.1 were expected. France is investigating the signal by participating in the VAESCO study and with an extension afterwards." 
Yet the French said elsewhere that cases were increased in both children and adults. France also has comments in Annex 1 (page 156) criticizing the ECDC report:
At the time of the VAESCO study data lock, the recruitment of case and control was still ongoing in France. The results herein presented should therefore only be considered as preliminary.
(page 150) Ireland:
"The rate was 5.8 per 100 000 PY in the vaccinated and 0.5/100 000 PY in the non-vaccinated, resulting in a 13-fold higher risk of narcolepsy in vaccinated compared to non-vaccinated in children/adolescents."
(page 150) Germany:
"18 cases reporting of narcolepsy following Pandemrix vaccination were reported to the Paul-Ehrlich Institute between October 2010 and June 2011. It concerned 13 children/adolescents and five adults [38]. The Paul Ehrlich Institute is conducting a case control study based on the VAESCO protocol together with Dr. Mayer.  (Still looking--Nass)
(page 150) UK:
" ... currently the Health Protection Agency is investigating the association independently by using a self controlled case series design."  (Still looking--Nass)
(page 156) Norway:
The Norwegian position is that the VAESCO primary analysis is hampered by too strict exclusion criteria and hence a lack of power. According to the Norwegian investigators the best way to understand the Norwegian data is to look at the total (secondary) study period.
So the report appears to have obscured the possible effect of Pandemrix in multiple countries:
  1. by using data before countries had completed their data collection, 
  2. by making an assumption without justification that all cases appeared soon after vaccination, ignoring the previous known average 5 year delay between symptom onset and diagnosis in adults, 
  3. by too-strict exclusion criteria 
--and all the above furthermore lowered the power of the study.

Should the Brighton Collaboration, an ad hoc, unofficial vaccine safety collaborative, have been given the contract for this research? Brighton was originally funded by Swiss charities and the US CDC.  It is uncertain who all its donors are today.  A disclaimer:  I am a Brighton partner.

No comments: