Safety of vaccines must be excellent to make vaccine's strategy acceptable, since it usually has a deferred individual benefit but immediate adverse drug reactions (ADRs). Pharmacovigilance of vaccines after their marketing is crucial because, prior to its availability on the market, the size of clinical trials is insufficient to identify rare or deferred adverse effects. The Pharmacovigilance is based on "spontaneous reporting" of ADRs to the Pharmacovigilance Regional Centre (PVRC) which establishes a relationship between each drug taken by the patient and the ADRs occurrence (imputability). This method is crucial to generate alerts, but under-estimates the real frequency of ADRs (1 to 10% of severe ADRs are reported). Thus pharmacoepidemiology studies are necessary to confirm the alerts identified by spontaneous reporting. ADRs can be specific, related to the antigen of an attenuated alive virus vaccine (lymphocyte meningitis after anti-mumps vaccine) or non-specific, related to a component different from the antigen (aluminium hydroxide involved in the "macrophagic myofasciitis", allergic reactions to neomycin, latex, egg or gelatine). Importance of Pharmacovigilance of vaccines is illustrated. Data, especially case-control studies, about the relationship between multiple sclerosis and hepatitis B vaccine are summarised. Data about the relationship between Crohn's disease or autism and MMR vaccine are analysed. As vaccines are used in healthy people, their safety must be excellent to be accepted. To monitor them after their marketing is the unique way to detect rare ADRs. This surveillance is made through reporting of ADRs to the PVRC. However, an active and intensive surveillance of ADRs as the one set up from the marketing of Prevenar should be systematic.