Lancet: Febrile convulsions after seasonal influenza immunisation in UK

[tetano]

The Lancet, Volume 378, Issue 9789, Pages 399 - 400, 30 July 2011

doi:10.1016/S0140-6736(11)61207-5Cite or Link Using DOI


Febrile convulsions after seasonal influenza immunisation in UK
Sheila M Bird aEmail Address
Philip Bryan and Suzie Seabroke (March 12, p 904)1 from the Medicines and Healthcare products Regulatory Agency (MHRA) report that at least 72 000 children younger than 5 years had received the 2010?11 seasonal influenza immunisation by Feb 15, 2011, and that there were only two reports of febrile convulsions. However, key statistical issues were below the radar.
First, within practices taking part in the General Practice Research Database (GPRD), 20 times as many under 5s received the H1N1 vaccine in 2009?10 as had received immunisation against seasonal influenza in a typical year in the decade to 2009?10 (table). The risk profile of GPRD children who received immunisation against seasonal influenza and against H1N1 might therefore be quite different. In 2010?11, the UK's risk-based advice on seasonal influenza immunisation of the under 5s differed from the previous year's advice on H1N1 vaccination and, accordingly, vaccination rate in the under 5s fell back to around 2% UK-wide. Bryan and Seabroke's ?expectation? of ten cases of febrile convulsion after 2010?11 seasonal influenza immunisation, extrapolated from rates of febrile convulsions after H1N1 immunisation in 2009?10, might therefore be an overestimate.

Second, uncertainty in the estimation of either GPRD-applicable rate (seasonal or H1N1) rests on just eight cases. This uncertainty was not taken into account in calculating Bryan and Seabroke's observed-to-expected ratio of 0?19 (95% CI 0?02?0?67). ?Observed? is MHRA's two yellow cards, and different from actual cases because it has not taken into account the ?possibility of under-reporting??ie, failure to submit yellow cards. A 20% yellow-card submission rate would transform the above 95% CI to 0?10?3?35, which straddles unity, and does not exclude a three times higher rate in 2010?11 of febrile convulsions than after H1N1 vaccination. The upper limit would be higher still if the yellow-card rate was at best 10%, as for H1N1 in 2009?10. However, for seasonal influenza in 2000?01 to 2009?10, MHRA received only one yellow card when an estimated 344 000 under 5s were vaccinated, for whom the febrile convulsion rate within 72 h was 2?9 per 10 000 and so 100 could have been received.
Although laudable to offer reassurance, much more detailed data and explanation were needed. Even in Correspondence, key rates or percentages should not be published without numerators and denominators.
I write in a personal capacity, hold shares in GlaxoSmithKline, and serve on the UK's Scientific Pandemic Influenza Advisory Committee.


http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(11)61207-5/fulltext?rss=yes

 

[tetano]

Re: Lancet: Febrile convulsions after seasonal influenza immunisation in UK

The Lancet, Volume 378, Issue 9789, Page 400, 30 July 2011

doi:10.1016/S0140-6736(11)61208-7Cite or Link Using DOI
Febrile convulsions after seasonal influenza immunisation in UK ? Authors' reply
Philip Bryan a, Suzie Seabroke aEmail Address
Our strategy was to identify rapidly whether influenza vaccines used in children in the UK from September, 2010, could potentially cause febrile convulsions at a frequency of one case per 100 doses. This was the risk with CSL influenza vaccine seen earlier in Australia.1 Although Sheila Bird raises issues around how data from the General Practice Research Database (GPRD) are described in our analysis, these points do not detract from the conclusion we have drawn.
Variable under-reporting of suspected adverse reactions is an inherent limitation of all passive surveillance systems. We attempted to reduce this under-reporting by proactively writing to all health-care professionals to request reporting of any case of convulsions in temporal association with influenza vaccine.2 With around 70 000 children vaccinated, if a 1% risk of convulsions with other influenza vaccines existed, two spontaneous reports would suggest an under-reporting rate of more than 99%. We consider this highly unlikely in this instance. In the context of the methods and data used, we fully agree that a small increased risk cannot be excluded. However, this strategy was for rapid signal detection, not risk quantification. On the basis of enhanced passive surveillance, we believe that our conclusion that ?there remains no indication that other influenza vaccines are associated with a large increase in risk of febrile convulsions as seen in Australia? is entirely valid.
Owing to historical variability in passive reporting rates, comparison of absolute numbers of passive reports between years is not a robust approach to signal detection. The background incidence data from GPRD were included in our letter to provide some context for the ?expected? rate. However, within the text constraints of Correspondence, a full description of the GPRD data was not possible.
We calculated the background rate of febrile convulsions in children younger than 5 years within 72 h of vaccination to be 2?91 per 10 000 and 1?45 per 10 000 for seasonal and H1N1 influenza vaccines, respectively. We accept that children who received monovalent H1N1 vaccine might be different from those who received the seasonal influenza vaccine. However, since the same H1N1 vaccine strain was included in the seasonal vaccines, and to take a more conservative approach to the ?expected?, we used the latter background rate. Although the observed-to-expected ratio we presented does not account for under-reporting of spontaneous reports, this is acknowledged in our letter and does not affect our conclusion.
Despite the known limitations of the Yellow Card scheme, it remains a vital early warning system and the health professionals who promptly report their suspicions to us deserve early feedback on the important role their reports serve in protecting public health.
We declare that we have no conflicts of interest.

http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(11)61208-7/fulltext?rss=yes

 

[tetano]

Re: Lancet: Febrile convulsions after seasonal influenza immunisation in UK

The original article.

No increased risk of febrile convulsions after seasonal influenza immunisation in UK

The Lancet, Volume 377, Issue 9769, Page 904, 12 March 2011

doi:10.1016/S0140-6736(11)60352-8Cite or Link Using DOI
No increased risk of febrile convulsions after seasonal influenza immunisation in UK
Philip Bryan a, Suzie Seabroke aEmail Address
Seasonal influenza immunisation for healthy children was suspended in Australia in April, 2010, owing to an increased risk of febrile convulsions associated with a vaccine manufactured by CSL (Fluvax). The risk was estimated to be as high as one case per 100 doses. A thorough assessment1 has yet to establish a reason for this risk, which seemed to be product-specific.
A similar CSL vaccine is being used in the current UK immunisation campaign (Enzira/CSL Biotherapies generic brand) and the Department of Health has advised that it should not be used in children younger than 5 years. Several alternative brands of influenza vaccine are available in the UK and these should be used in children in clinical risk groups2 as recommended.
Although there was no reason to suspect that other influenza vaccines might be associated with an increased risk of febrile convulsions, the Medicines and Healthcare products Regulatory Agency (MHRA) implemented enhanced surveillance, based around the Yellow Card Scheme. We issued a letter to health professionals to encourage reporting of febrile convulsions after influenza immunisation, emphasising the importance of including in the report the vaccine brand name.3
We calculated background rates of febrile convulsion after influenza vaccines from 2000?09 and 2009?10 with data from the General Practice Research Database. Using these rates and data on the number of children immunised across the UK, we calculated an expected number of cases of febrile convulsions within 72 h of immunisation every week. Comparing this number to Yellow Card reporting, we looked for any indication of excess reporting.
As of Feb 15, 2011, MHRA had received only two reports of febrile convulsions in children younger than 5 years after influenza immunisation. At least 72 000 children had been vaccinated in this age-group at this time, among whom we expected up to ten cases to have occurred. The observed-to-expected ratio for the age-group was 0?19 (95% CI 0?02?0?67).
Taking into account the possibility of under-reporting, there remains no indication that other influenza vaccines are associated with a large increase in risk of febrile convulsions as seen in Australia. This enhanced surveillance supports the decision that children in clinical risk groups should continue to receive the alternative influenza vaccines as recommended, since the balance of benefits and risks is clearly favourable.

http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(11)60352-8/fulltext