Oseltamivir effective in reducing risk of some complications of fluArchived

Oseltamivir and Risk of Lower Respiratory Tract Complications in Patients With Flu Symptoms: A Meta-analysis of Eleven Randomized Clinical Trials Hernán MA and Lipsitch M Clin. Infect. Diseases; first published online: June 15, 2011.

In recent years a controversial topic has been whether or not antivirals, specifically the neuraminidase inhibitors, oseltamivir and zanamivir, reduce the likelihood of severe complications of seasonal or pandemic influenza if given soon after infection with influenza viruses. Observational data repeatedly suggests they are  effective but the optimal answer to such questions would come from randomised placebo controlled trials.(1-3)  Typically such trials are conducted with healthy adults, among whom the complications of influenza are less common. Hence meta-analyses have been used, combining data from multiple trials. Otherwise one simply ends up concluding there is no evidence of effectiveness.(3) An important paper using this approach for oseltamivir, the most widely used antiviral in Europe, was published in 2003 by Kaiser et al.(4) that used data from 10 trials supported by the manufacturer of the drug (Roche). Only two of the trials had been published. Kaiser et al analysis found that early use of oseltamivir was protective.(4) This approach proved controversial as a later publication by a group led by Jefferson, publishing under the Cochrane label, was unhappy to accept data from the unpublished trials and analysed only the data from the two published trials. They concluded in 2009 that there was no evidence demonstrating an effect.(3) This publication which appeared at the height of the 2009 pandemic might have contributed to doubts among practitioners in official advice promoting early use of antivirals in Europe.  The Jefferson group had asked Roche for the original trial data which they were eventually supplied with, but not in time for the 2009 publication. Roche later asked a respected independent researcher (Professor Lipsitch of Harvard University) to conduct an analysis with the same data. Despite the surrounding controversy (5) Lipsitch and a colleague Hernan agreed to do so because of the importance of the topic.(6) After negotiation to ensure that their work was independent and obtaining data from another unpublished Roche-supported trial the Harvard researchers (Hernan and Lipsitch) reached agreement with the company to receive full access to efficacy and safety data from the trials, assistance from Roche statisticians in answering data-related questions and complete freedom to publish any results.

Hernán and Lipsitch have now published their analysis following a peer-review process. The new analyses are an improvement methodologically on the original Kaiser publication and addresses some of the concerns of the Jefferson group.(3,4,6) Hernán and Lipsitch combined data from 11 randomized clinical trials on oseltamivir effectiveness in reducing the risk of complications of influenza. The trials included adults and adolescents with influenza symptoms during the 1997–2001 influenza seasons. The primary outcome of their analysis was looking at any lower respiratory tract complications (LRTC) requiring treatment with antibiotics, whether or not hospitalised. Because of its low frequency in the trials this complication was neither the primary nor secondary outcome in the original studies but was reconstructed retrospectively. Other outcomes taken into account were a number of adverse events following treatment: gastrointestinal disorders (nausea, vomiting and diarrhoea), neuropsychiatric disorders (other than headache) and headache itself. Neuropsychiatric disorders  were included because of observations in Japan, where oseltamivir is used especially frequently, associating oseltamivir with acute neuropsychiatric disorders in adolescents.(7) Hernan & Lipsitch analyses excluded participants taking antibiotics at baseline. Three features made the analyses an improvement from the two previously mentioned studies:- Computed study-specific risk ratios of LRTC treated with antibiotics within the first 24 days of follow-up were derived for oseltamivir versus placebo, and then the study-specific risk ratios were pooled using meta-analysis; - Endpoints diagnosed during the first 2 days after randomization were included. In addition, and taking into account the fact that a true intention-to-treat analysis would include the complete follow-up of all randomized patients, the authors explored the sensitivity of the estimates to the exclusions performed in the individual trials;- Subset analyses by influenza infection status was included, for comparability with the meta-analysis by Jefferson et al. (3) and the subset analysis by Kaiser et al. (4) Hernán and Lipsitch found that oseltamivir treatment was associated with significantly reduced  risk of LRTC requiring antibiotic treatment by 28% overall (95% confidence interval [CI]; 11%–42%) and by 37% among patients with confirmed influenza infections (95% CI; 18%–52%). No reduction was observed in patients without influenza and the effect estimates changed little under a series of sensitivity analyses.(6)  The study confirmed previous reports of increased risk of nausea and vomiting , but found no evidence of increased risk of neuropsychiatric disorders (apart from headache) - among those assigned to oseltamivir. The authors point out that, as for any meta-analysis, the quality of their estimates depends on the quality of the individual studies.

ECDC Comment (26th June 2011):The important result here is the confirmation of the original Kaiser meta-analysis result namely that early oseltamivir use was associated with reductions in the risk of some complications of influenza in previously well adults and adolescents in the period 1997-2001. It is important to note these are not new data but new analyses of prior data. The Jefferson group are not known to have published a re-analysis.(6,8)  Certainly there are good arguments for making trial data ever more public in Europe.(9)  However the practicalities of making all data, especially primary data publically available in timely manner makes this questionable  and exclusion of such data may lead to biased results. 

A second issue is that of using observational data.  The Jefferson Group has argued for ever larger publicly funded randomised-controlled trials of influenza antivirals (and vaccines).(3,10)  These are probably not going to take place, apart from occasional ‘probe’ studies (community randomised trials introducing pharmaceutical interventions).(11)  It seems wasteful not to allow cautious use of observational data, especially as in this case where observational and trial data concur. There is also the issue over how to deal with the changing nature of influenza viruses. The trials that have been reanalysed here were conducted against the influenza of over a decade ago and the viruses do not stand still.  Notably the old results may not apply against those the new pandemic influenza A(H1N1)2009 viruses.(12)  Fortunately the observational data  to date are reassuring and in the same direction as the previous trial data, and mechanisms exist now for incorporating observational data like the GRADE system .(2,13,14)

A final consideration is the perilously limited array of influenza antivirals available to clinicians and public health.  Most circulating human influenza viruses are already resistant to the adamantenes.  In 2007/8 a variant of the previous human seasonal A(H1N1) viruses emerged with oseltamivir resistance (but not zanamivir resistance) and out-competed their susceptible counterpart viruses,  without treatment pressure.(15,16)  Most of the new A(H1N1)2009 viruses remain sensitive, at present. But clinicians and public health alike have a narrow base for treatment and few novel drugs coming through.  That is why is the European Commission Directorate of Research and Innovation (DG-RESEARCH) is  supporting the large research programmes FLUPHARM and FLUCURE working with industry to carrying forward previous EU research initiatives.