In their meta-analysis of the efficacy and effectiveness of influenza vaccines licensed in the USA, Michael Osterholm and colleagues1 stated that “evidence for protection in adults aged 65 years or older is lacking”, although chronic disorders such as cardiac and pulmonary diseases could require influenza vaccination. The cost-effectiveness of influenza vaccination policy requires consideration of adverse effects such as autoimmune disorders, which are rarely reported. Possible reasons for disregarding these events include the subacute presentation in some cases and the variable latency period (from days to years), which makes ascertainment of the causality link difficult.2 The 1976 national influenza immunisation programme against swine flu subtype A/NJ/76 in the USA was stopped because of the emergence of Guillain-Barré syndrome in some vaccine recipients. The USA, the UK, and Germany have initiated active surveillance studies to detect potential rare adverse events, and “the overall risk–benefit assessments must be interpreted very carefully, since the composition of annual influenza vaccines varies each season“3 and the risk of side-effects might depend on the subtype, or adjuvants, or both. For the past few years, all post-vaccination events have been included in the spectrum of autoimmune or inflammatory syndrome induced by adjuvants, because of previous exposure to immune adjuvants, including those used in vaccines to boost an immune response.4 In our Periodic Fevers Research Centre, which admits patients with fever of unknown origin, over 6 years we identified ten cases of giant cell arteritis or polymyalgia rheumatica, occurring within 3 months of influenza vaccination. A Medline search from 1978 to 2011 showed 11 isolated cases of the same disorders happening after influenza vaccination.5 Such disorders are common in people older than 65 years, who are a major target of influenza vaccination policy. In the Comment accompanying Osterholm and colleagues' meta-analysis, Heath Kelly and Marta Valenciano6 state that “now might also be an appropriate time to use revised estimates of the most probable effectiveness of influenza vaccines to re-examine the effectiveness and cost-effectiveness of some policy options”. We support this statement and suggest that evaluation of the costs of influenza vaccination should include autoimmune reactions as adverse effects, with both a central worldwide registration system and, where they exist, local active surveillance registries to monitor vaccine safety. This strategy could be useful to identify genetic markers of populations at high risk for after-vaccination autoimmune events and to develop more effective, safe, and personalised vaccines.