- Nicola J Cooper (), research fellow1,
- Alexander J Sutton, lecturer in medical statistics1,
- Keith R Abrams, professor of medical statistics1,
- Allan Wailoo, research fellow2,
- David Turner, research associate in health economics1,
- Karl G Nicholson, professor of infectious diseases3
- 1 Department of Epidemiology and Public Health, University of Leicester, Leicester LE1 6TP,
- 2 Sheffield Health Economics Group, School of Health and Related Research, University of Sheffield, Sheffield S1 4DA,
- 3 Infectious Diseases Unit, Leicester Royal Infirmary, Leicester LE1 5WW
- Correspondence to: N J Cooper
- Accepted 10 April 2003
Objective To review the clinical effectiveness of oseltamivir and zanamivir for the treatment and prevention of influenza A and B.
Design Systematic review and meta-analyses of randomised controlled trials.
Data sources Published studies were retrieved from electronic bibliographic databases; supplementary data were obtained from the manufacturers.
Selection of studies Randomised controlled, double blind trials that were published in English, had data available before 31 December 2001, evaluated treatment or prevention of naturally occurring influenza with zanamivir or oseltamivir (if given using the formulation and dosage licensed for clinical use), and reported at least one end point of relevance.
Review methods The main outcome measures were the median time to the alleviation of symptoms (for treatment trials) and number of flu episodes avoided (for prevention trials). Three population groups were defined: children aged 12 years and under; otherwise healthy individuals aged 12 to 65 years; and “high risk” individuals (those with certain chronic medical conditions or aged 65 years and older).
Results Seventeen treatment trials and seven prevention trials identified met the inclusion criteria. All trials included compared one of the drugs against placebo or standard care. Treatment of children, otherwise healthy individuals, and high risk populations with zanamivir reduced the median duration of symptoms in days respectively by 1.0 (95% confidence interval 0.5 to 1.5), 0.8 (0.3 to 1.3), and 0.9 (−0.1 to 1.9) for the intention to treat population. The corresponding results, in days, for oseltamivir were 0.9 (0.3 to 1.5), 0.9 (0.3 to 1.4), and 0.4 (−0.7 to 1.4). The effect of giving zanamivir and oseltamivir prophylactically resulted in a relative reduction of 70-90% in the odds of developing flu, depending on the strategy adopted and the population studied.
Conclusions Evidence from randomised controlled trials consistently supports the view that both oseltamivir and zanamivir are clinically effective for treating and preventing flu. However, evidence is limited for the treatment of certain populations and for all prevention strategies.
Influenza epidemics occur almost every winter and are associated with considerable morbidity and mortality.1 All age groups are susceptible, but increasing age, certain chronic medical conditions, and residential care increase the risk of complications and death. Two interventions can lessen the impact of flu: immunisation with inactivated vaccines and treatment and prophylaxis with antivirals. Current policy in the United Kingdom recommends that individuals at high risk of serious illness or death from flu should be vaccinated against the virus annually (see www.doh.gov.uk/flu.htm). Antivirals represent a rational approach to flu management to complement vaccination, particularly in “high risk” individuals, but until recently only the M2 inhibitors, amantadine and rimantadine, were available. Limitations of these drugs include rapid emergence of resistance,2 3 lack of antiviral activity against influenza B, and frequent adverse central nervous system events, particularly in elderly people.4 5
The goal of providing activity against influenza A and B with few adverse events came with the development of the neuraminidase inhibitors for flu, zanamivir (Relenza, GlaxoSmithKline) and oseltamivir (Tamiflu, Roche). In this systematic review, commissioned by the National Institute for Clinical Excellence (NICE), we examined randomised controlled trials of zanamivir and oseltamivir, both for treatment and prophylaxis, in three populations—children, high risk adults, and otherwise healthy adults—to assess the evidence for the clinical effectiveness of these two drugs. The results of this systematic review were incorporated into an economic decision model to produce the NICE guidance on zanamivir and oseltamivir, which was issued in February 2003.6
We searched Medline (1966 to December 2001), Embase (1980 to December 2001), Integrated Science Citation Index (1981 to December 2001), and the National Library of Medicine (PubMed). In addition, we searched cited literature in retrieved articles, previous systematic reviews and meta-analyses of neuraminidase inhibitors,7–10 and manufacturers' trial databases. We contacted drug companies for information on unpublished trials.
We selected randomised controlled, double blind trials that met all the following criteria: were published in English, had data available before 31 December 2001, evaluated treatment or prevention of naturally occurring influenza with zanamivir or oseltamivir (if these were given using the formulation and dosage licensed for clinical use), and reported at least one end point of relevance (see below).
We used a validated instrument developed previously11 to assess the methodological quality of the treatment and prevention trials according to the method of randomisation, concealment of allocation, blinding of trial investigators and patients, and completeness of follow up.
Summary outcome data were initially extracted from trial publications and final trial reports. Additional data were also requested from the drug companies—for example, separate data on individual population groups when publications reported combined results for otherwise healthy and high risk individuals, standard errors of the median “time to an event,” and, where appropriate, re-analysis of the data to allow for censored observations (whereby the event has not occurred by the end of the trial).
We considered three populations: children aged 12 years and under; otherwise healthy individuals aged 12 to 65 years; and high risk individuals. We defined high risk individuals as those aged 65 years and older or those having certain chronic medical conditions, such as respiratory disease, heart disease, and pulmonary disorders.
The primary treatment end points were “time to the alleviation of symptoms” and “complications requiring antibiotics.” We also considered “time to return to normal activities” and “admissions to hospital”; these are reported elsewhere.12 In this paper, we present results for both the intention to treat (ITT) populations and the populations with a laboratory confirmed “flu positive” diagnosis.
The primary prevention end point was the number of individuals with laboratory confirmed symptomatic flu at the end of the trial. We reported our assessment of adverse events elsewhere.12
We present results of trials comparing each neuraminidase inhibitor with placebo or standard care for treatment and prevention. We identified no “head to head” (zanamivir v oseltamivir) trials. Varying proportions of randomised individuals were vaccinated before entry into the trials (table 1 and 2). The full list of trials included in this review is available on bmj.com (as “web extra”).
We identified 44 studies evaluating zanamivir for the treatment of flu and 18 for oseltamivir (fig 1). Of these, eight randomised controlled trials of zanamivir and nine of oseltamivir met our eligibility criteria (table 1).
Time to alleviation of symptoms—In the ITT population, the reduction in the median time to alleviation of symptoms in the treatment groups, when compared with placebo, ranged from 0.8 (95% confidence interval 0.3 to 1.3) days for otherwise healthy adults to 1.0 (0.5 to 1.5) days for children (fig 2). In the flu positive population, this reduction ranged from 1.0 (0.4 to 1.6) days for children to 2.0 (0.9 to 3.1) days for high risk adults (fig 3).
Complications requiring antibiotics—Few trial level data on complications requiring antibiotics were obtained from the literature; however, two published, pooled marginal analyses were identified.15 16 When considering all three ITT populations combined, Monto and colleagues observed a 29% (10% to 44%) relative reduction (zanamivir v placebo) in the odds of complications requiring antibiotics.15 However, among the high risk flu positive population, they found a non-significant relative reduction (45%) in the odds of antibiotic use. Lalezari and colleagues,16 who focused on high risk adults and on children, obtained similar results.
Time to symptom alleviation—In the ITT population, the reduction in median time to alleviation of symptoms, when oseltamivir was compared with placebo, ranged from 0.4 (−0.7 to 1.4) days for high risk adults to 0.9 (0.3 to 1.5) days for children (fig 4). In the flu positive population, the reduction in time to alleviation of symptoms ranged from 0.4 (1.0 to 1.9) days for high risk adults to 1.5 (0.8 to 2.2) days for children (fig 5).
Complications requiring antibiotics—Only one study (WV15670) in otherwise healthy adults reported a non-significant relative reduction (oseltamivir v placebo, 43%) in the odds of complications requiring antibiotics in the ITT population and a significant relative reduction (87%) in the flu positive population.17 Among children, a 35% relative reduction in the odds of complications requiring antibiotics was observed in one study (WV15758).18
We identified 11 randomised controlled trials that evaluated zanamivir for the prevention of flu and seven for oseltamivir (fig 1). Of these, three trials of zanamivir and four of oseltamivir met our eligibility criteria (table 2).
Seasonal prophylaxis of a healthy population (NAIA3005)—A 69% (36% to 86%) relative reduction (zanamivir v placebo) in the odds of laboratory confirmed symptomatic flu was observed (table 2).
Post-exposure prophylaxis in households (NAIA/B2009, NAI30010)—A meta-analysis of these two trials showed an 81% (62% to 91%) relative reduction (zanamivir v placebo) in the odds of laboratory confirmed symptomatic flu.
Seasonal prophylaxis of a healthy population (WV15673, WV15697)—A meta-analysis of these two trials showed a 74% (16% to 92%) relative reduction (oseltamivir v placebo) in the odds of laboratory confirmed symptomatic flu.
Post-exposure prophylaxis in households (WV15799)— A 90% (71% to 96%) relative reduction (oseltamivir v placebo) in the odds of laboratory confirmed symptomatic flu was observed (table 2).
Seasonal prophylaxis in residential care (WV15825)— In a mostly vaccinated elderly population receiving residential care there was a 92% (39% to 99%) relative reduction (oseltamivir v placebo) in the odds of laboratory confirmed symptomatic flu (table 2). Similar benefits were observed in those previously vaccinated.
The results of our systematic review show that treating otherwise healthy adults and children with zanamivir and oseltamivir reduces the duration of symptoms in the intention to treat population by between 0.4 and 1.0 days and provides 29% to 43% relative reduction in the odds of complications requiring antibiotics when these are given within 48 hours of onset of symptoms. The results were less conclusive in the high risk population (as defined in the methods) though these were based on fewer patients. Caution is required when comparing the results because the definition of symptoms assessed for alleviation in the treatment trials varied among trials of the two compounds, and between adults and children for each compound. Moreover, the time to event outcomes were measured on different scales (days and hours). Also, the rates of flu positive (≥49%) individuals who were enrolled in the trials may be higher than the rates identified routinely in clinical practice. Thus, the treatment effects estimated for the ITT trial populations may not be achievable in routine practice.
The data on complications reported above were not ideal because they relied primarily on pooled marginal analyses and thus did not take into account any heterogeneity between trials.19 20 It is not clear how well complications requiring antibiotics correlate with the incidence of more serious complications of flu. Little evidence exists either on serious complications requiring admission to hospital or causing death or on adverse events. Both of these are evidently rare (at least in otherwise healthy individuals) but are potentially important in the evaluation of treatments; the trials were underpowered in terms of such outcomes. Insufficient data are available from clinical trials to assess adequately the risk of emergence of resistance to neuraminidase inhibitors.
A lack of evidence exists for use of neuraminidase inhibitors for preventing flu in children and in frail elderly people in residential care. We found that neuraminidase inhibitors given for flu prevention led to a relative reduction of 70% to 90% in the odds of developing flu, depending on the strategy adopted and the population studied.
In conclusion, although evidence from randomised controlled trials consistently supports the clinical effectiveness of both oseltamivir and zanamivir for the treatment and prevention of flu, evidence is limited for the treatment of high risk populations and for all prevention strategies. Research is needed into the comparative effectiveness of neuraminidase inhibitors with one another and the potential “added value” of these drugs compared with or in combination with flu vaccine.
What is already known on this topic
Neuraminidase inhibitors (zanamivir and oseltamivir) may be useful in treating and preventing flu
Data are limited, however, for certain population groups and for prophylactic use
What this study adds
Unlike previous systematic reviews, this review considered three distinct populations (children, high risk adults, and otherwise healthy adults)
For the use of neuraminidase inhibitors for flu prevention, different strategies evaluated by randomised controlled trials were reviewed
Oseltamivir and zanamivir are clinically effective for treating and preventing flu, but evidence is limited for certain population groups and for all prevention strategies
The full list of trials included in this review is available on bmj.com Editorial by Stöhr
We thank GlaxoSmithKline (in particular, Stephen Sharp) and Roche (in particular, Paul Mahoney) for providing additional trial data.
Contributors All authors participated in designing the review, checking the data, and revising the manuscript, which was drafted by NJC. NJC, AJS, AW, and DAT decided on trial inclusion or exclusion, extracted data, and assessed study quality. NJC, AJS, and KRA performed the statistical analyses. KRA and KGN were the principal advisers, guiding and interpreting the review. NJC is the guarantor for the paper.
Funding This work was commissioned by the NHS Health Technology Assessment programme and the National Institute for Clinical Excellence. NJC is funded by University Hospitals of Leicester NHS Trust. DAT and AW are funded by the Trent Institute for Health Services Research. The guarantor accepts full responsibility for the conduct of the study, had access to the data, and controlled the decision to publish.
Competing interests KGN has received travel sponsorship and honorariums from GlaxoSmithKline, the manufacturer of zanamivir, and Roche, which makes oseltamivir, for consultancy and speaking at international respiratory and infectious diseases symposiums. His research group has received research funding from GlaxoSmithKline and Roche to participate in multicentre trials of neuraminidase inhibitors; Berna Biotech and Chiron for trials of flu vaccines; and Wyeth for work on the epidemiology of flu in young children. KGN was a founder member and vice chairman of the European Scientific Working Group on Influenza (ESWI), a group of European scientists promoting the study of flu. ESWI is supported by the vaccine manufacturers, Roche and GlaxoSmithKline, but is scientifically independent.