(1) Influenza is a common acute respiratory disease due to a virus that causes annual seasonal epidemics. Three major pandemics occurred in the 20th century, in 1918-1919, 1957 and 1968, mainly due to genetic variants of type A influenza virus. (2) In temperate regions the incidence of hospitalisation increases during annual influenza epidemics. More than 90% of deaths linked to influenza involve people over 65 years of age. (3) The clinical manifestations of influenza virus infection are non specific. The main complications are secondary bacterial respiratory tract infections (especially pneumonia); those most at risk are people over 65, infants less than one year old, and people with underlying chronic disorders (pulmonary, cardiac, renal or metabolic) or immune deficiencies. (4) Vaccination is the main preventive measure. During most years the vaccine strain closely matches the epidemic strain. In relative terms, vaccination of people over 65 reduces the number of deaths linked to influenza by about 80%, hospitalisation and pneumonia by about 50%, and symptomatic influenza by about 30%. Yearly vaccination is recommended for younger people with serious chronic disease. (5) Three antiviral drugs are currently approved in France for prevention or treatment of influenza: amantadine and the neuraminidase inhibitors zanamivir and oseltamivir. (6) Efficacy of antiviral drugs has not been evaluated in comparative randomised trials in which death and influenza complications were the primary outcome measures. (7) A systematic review of 20 comparative randomised trials involving about 2500 healthy people showed that amantadine reduced the frequency of flu-like syndromes by about 7% in absolute terms (26.3% versus 33.1% with placebo). Zanamivir and oseltamivir have only been shown to reduce the frequency of serologically confirmed episodes of influenza (0.4% to 2.5%, compared to 4.4% to 14.9% with placebo). (8) In a randomised placebo-controlled trial of oseltamivir, involving 548 institutionalised subjects over 65 years of age, more than 80% of whom had been vaccinated, respiratory tract infections were less frequent in the oseltamivir group, but the relevance of this result is undermined by the small number of observed cases. (9) Efficacy of antiviral drugs on avian influenza (bird flu) was studied during a 2003 Dutch outbreak due to a type A/H7N7 virus. Among the 38 exposed persons who were treated, about 3% developed symptoms, compared with about 10% of 52 exposed persons who refused treatment (p = 0.38). The low statistical power and the lack of randomisation rule out any firm conclusions on preventive effects. (10) The three antiviral drugs have different profiles of adverse effects and drug interactions. Amantadine carries a risk of neuropsychological, atropinic and dopaminergic adverse effects, and can interact with drugs that have similar effects. Zanamivir carries a risk of life-threatening bronchospasm. Oseltamivir was approved relatively recently and its full spectrum of adverse effects is not yet known; its main adverse effects appear to be mild gastrointestinal disturbances, although a few cases of serious cutaneous reactions have been reported. (11) In vitro resistance to the three drugs has been demonstrated, but the possible clinical and epidemiological consequences are unclear. (12) In situations warranting antiviral therapy for the prevention of influenza, oseltamivir, at a dose of 75 mg/day for 10 days, is the drug with the best risk-benefit balance. Its use should be limited to situations where a major potential benefit exists in order to avoid selection for resistant strains. (13) Testing of oseltamivir in children is limited. Oseltamivir should be avoided during pregnancy, because of evidence that it may harm the unborn child. (14) In practice, the use of antiviral drugs in otherwise healthy adults and children is not generally recommended. (15) Despite the lack of convincing data regarding the efficacy of oseltamivir in preventing complications of influenza, its effect on documented infections suggests it may be useful for unvaccinated individuals who are at high risk of infection and severe complications. Under these conditions, treatment should be started within 48 hours after contact with a person who has flu-like symptoms during a seasonal epidemic; residents in institutions in which influenza cases occur may also qualify for preventive treatment. Other preventive measures should also be used, including immediate vaccination, case isolation, use of face masks, and more frequent hand washing. (16) During seasonal influenza epidemics due to viral strains against which the current vaccine is of limited effectiveness, the utility, target populations and optimal duration of preventive antiviral treatment must be determined by examining the groups most at risk and the severity of complications. (17) Most flu-like syndromes are not due to the influenza virus, and the preventive effect of antiviral drugs on complications in persons at risk has not yet been demonstrated. (18) In practice, antiviral drugs are not an alternative to influenza vaccination, but may be a useful adjunct in some situations. It is best to limit their use to short-term prophylaxis of vulnerable persons in situations where the risk of contracting influenza virus infection is high.