OBJECTIVES

Influenza is an acute respiratory infection associated with raised temperature, headache, muscle ache and cough. The objective of this review was to assess the effects of neuraminidase inhibitors (NIs) in preventing cases of influenza and shortening or reducing the severity of influenza in healthy adults. A further objective was to estimate the frequency of adverse effects associated with NI administration.

SEARCH STRATEGY

We searched Medline, the Cochrane Acute Respiratory Infections Group trials register, the Cochrane Controlled Trials Register (CCTR), manufacturers' databases, Embase (1991 to 1998) and reference lists of articles in May 1999. We also contacted manufacturers, researchers in the field, and authors of studies evaluated in the review.

SELECTION CRITERIA

Randomised or quasi-randomised placebo-controlled studies of NIs in healthy adults. Studies assessing protection or treatment from exposure to naturally occurring and experimental influenza were considered. The main outcomes were numbers and/or severity of influenza cases and the number and seriousness of adverse effects.

DATA COLLECTION AND ANALYSIS

Two reviewers applied the inclusion criteria to the retrieved studies, assessed trial quality and extracted data.

MAIN RESULTS

Eight trials with 1180 adults were included. Overall the methodological quality of the studies appeared to be good. As a preventive measure, NIs when compared to placebo were 74% effective (95% confidence interval 50% to 87%) in preventing naturally occurring cases of clinically defined influenza, and 60% effective (95% confidence interval 76% to 33%) in preventing cases of laboratory confirmed influenza. As a treatment, NIs shorten the duration of symptoms by one day - weighted mean difference 1 (95% confidence interval -1.3 to -0.6). The time gained in returning to normal activities is half a day - weighted mean difference -0.5 ( 95% confidence interval -1.1 to -0.1) for laboratory cases of influenza. The adverse event profile (local nasal irritation) of Zanamivir appears no better than placebo - odds ratio 1.19 ( 95% confidence interval 0.39 to 3.62). Compared with rimantadine in a preventive role, Oseltamivir has a significantly lower incidence of adverse effects and significantly higher incidence of nausea. For treatment, the adverse event profile shows that gastrointestinal symptoms are significantly worse in NIs than placebo - Peto odds ratio 2.32 ( 95% confidence interval 1.55 to 3.47).

REVIEWER'S CONCLUSIONS: NIs are effective for the prevention and treatment of influenza. Overall NIs are safe, although Oseltamivir causes significant nausea.