Sudden deaths following influenza vaccination: Can this be expected?

• Published in: Vaccine 2008-01, Vol.26 (3), p.379-382
• Main Authors: van der Sande, M.A.B, van Asten, L, Straus, S.M.J.M, Schim van der Loeff, M.F, Wallinga, J, Conyn-van Spaendonck, M.A.E
• Format: Article
• Language: English
• Subjects: Age; Aged; Aged, 80 and over; Allergy and Immunology; Binomial distribution; Biological and medical sciences; Cardiovascular Diseases - epidemiology; Cardiovascular Diseases - mortality; Death, Sudden - epidemiology; Family medical history; Fundamental and applied biological sciences. Psychology; Humans; Immunization; Influenza vaccination; Influenza Vaccines - administration & dosage; Influenza, Human - prevention & control; Likelihood Functions; Likelihood ratio; Microbiology; Middle Aged; Netherlands - epidemiology; Population; Population number; Probability; Sudden deaths; The Netherlands; Vaccination; Vaccines; Vaccines, antisera, therapeutical immunoglobulins and monoclonal antibodies; Virology
• ISSN: 0264-410X; 1873-2518
• DOI: 10.1016/j.vaccine.2007.11.006

Summary Background In November 2006, four Dutch people, aged 53, 58, 80 and 88, died unexpectedly on the day they had received their influenza vaccination. A rapid epidemiological assessment was needed to quantify the risk of a causal association. Methods Using routinely available data on age-stratified population size, cardiovascular mortality, and vaccination coverage, a daily rate and daily risk of sudden death per 5-year age-group was calculated. A cumulative probability that at least one person in four specific age-groups would die on the day of vaccination was calculated using a binomial distribution. No assumptions on deaths in other age groups were included. Results The overall likelihood that at least one person in each of the four age categories 50–54, 55–59, 80–84 and 85–89 would die suddenly on the day of influenza vaccination in the Netherlands was calculated to be 0.016. This was 330 times more likely than nobody dying in each of these categories, and 45 times less likely than the most probable outcome. Conclusion We concluded that there was a small but real chance of the four deaths occurring without a causal link to the vaccination. Policy decisions regarding unexpected deaths following vaccination can benefit from a rapid epidemiological evaluation.