Oseltamivir for influenza in adults and children: systematic review of clinical study reports and summary of regulatory commentsBMJ 2014; 348 doi: https://doi.org/10.1136/bmj.g2545 (Published 09 April 2014) Cite this as: BMJ 2014;348:g2545
All rapid responses
We have recently published a study that used data from clinical study reports to investigate the possible causal effect of Tamiflu on psychiatric adverse events. Our study used data from four prophylaxis trials, incorporated duration and intensity of events in the analysis, and analysed the proportion of days patients suffered from AEs during the on-treatment phase using single stage individual patient data meta-analysis. Results showed evidence of a causal effect of Tamiflu with odds ratio 3.46 (95% CI: 1.28, 9.32) overall and 34.5 (95% CI: 3.66, 325) for severe intensity events. However, the absolute risk increase due to Tamiflu was small (around a third of 1% overall). The study report can be accessed here: https://onlinelibrary.wiley.com/doi/10.1002/pds.4651
Competing interests: MJ and CDM were co-recipients of a UK National Institute for Health Research grant for the systematic review discussed in this article (HTA – 10/80/01 Update and amalgamation of two Cochrane Reviews: neuraminidase inhibitors for preventing and treating influenza in healthy adults and children—http://www.nets.nihr.ac.uk/projects/hta/108001). MJ is also a co-recipient of a Cochrane Methods Innovations Fund grant to develop guidance on the use of regulatory data in Cochrane reviews and reports personal fees from Laura and John Arnold Foundation, outside the submitted work. CDM receives grant funding for the Acute Respiratory Infections Cochrane Group, and for other research funding from the Australian NHMRC. He receives book royalties from Elsevier and Wiley, and has been paid consultancy funding together with colleagues for work on shared decision making from BUPA (UK) and the Australian Commission for Safety and Quality in Healthcare. SET declares no competing interests.
In our Cochrane review , we reported a non-significant increase in renal events in oseltamivir prophylaxis trials on-treatment (RR 3.17, 95% CI 0.96 to 10.49) by random effect model. However there was a statistically significant increase (P=0.02) in sensitivity analysis using Peto’s method. We also noticed two patients in oseltamivir groups who experienced renal failure prior to death : one in trial WV15825 and on in trial WV15708. While assessing the evidence in individual patient data in prophylaxis trials, we found imbalance of baseline renal disorders between the oseltamivir and placebo groups. We independently extracted the data on baseline creatinine level and concomitant renal and urinary tract diseases (renal/UTD) for five studies in total: two in adults (WV15673/WV15697), two in the elderly (WV15708; WV15825) and one in households (WV15799) .
Participants with high baseline creatinine level (154μmol/L or more) were significantly less reported in the oseltamivir group than in the placebo group: pooled odds ratio was 0.29 (95% CI = 0.12 to 0.69, P = 0.005, I2= 0%: Figure). In trial WV15708 which was never published, participants with high baseline creatinine level (154μmol/L or more) were significantly less reported (OR 0.23; 95％CI: 0.06, 0.82、P = 0.0168). Participants with concomitant renal/UTD were less reported (OR 0.38; 95%CI: 0.14, 0.93, p=0.0269). In trial WV15708, Proportions of participants with high baseline creatinine level and/or concomitant renal/UTD were 5.8% (11/190) and 16.5% (30/182) for oseltamivir group and placebo group respectively: odds ratio 0.31 (95%CI = 0.14 to 0.67, P = 0.0014).
In an experiment using beagle dogs to test the effects on cardiac functions such as especially QT time , mean baseline QTc intervals (msec±SE) were 417±16 in the control (vehicle) group (n=4) and 374±2 in the oseltamivir carboxylate (OC) group (n=4). Difference was significant (p=0.0372) by the summary data t-test. There was also evidence that the variation was higher in the control group (P=0.005; Bartlett’s test).
We don’t know the reason why such large and systemic imbalances occurred, but they are unlikely to have occurred by chance.
1. Jefferson T, Jones MA, Doshi P, Del Mar CB, Hama R, et al. Neuraminidase inhibitors for preventing and treating influenza in adults and children. Cochrane Database of Systematic Reviews 2014, Issue 4. Art. No.: CD008965. DOI: 10.1002/14651858.CD008965.pub4
2. The Dryad Digital Repository. http://datadryad.org/resource/doi:10.5061/dryad.77471
3. Marketing Authorization Application, Table of Contents – Volume 81 1007- Brewster M. Ro 64-0802/002 (GS-4071) cardiovascular and respiratory evaluation in the anaesthetized dog following intravenous administrations (DHB08601). RR W-142974. 1999 (disclosed document from EMA)
Competing interests: All authors were co-recipients of a UK National Institute for Health Research grant (HTA–10/80/01, Update and amalgamation of two Cochrane reviews: neuraminidase inhibitors for preventing and treating influenza in healthy adults and children—www.nets.nihr.ac.uk/projects/hta/108001). RH wrote two books on the harm of Tamiflu and/or antipyretics and provided scientific opinions and expert testimony on 14 adverse reaction cases related to oseltamivir in the law suits. MJ has no additional conflict of interest to be declared. TJ receives royalties from his books published by Blackwells and Il Pensiero Scientifico Editore, Rome. TJ is occasionally interviewed by market research companies for anonymous interviews about phase I or II pharmaceutical products. In 2011-13, TJ acted as an expert witness in a litigation case related to oseltamivir and in a labour case on influenza vaccines in healthcare workers in Canada. TJ was a consultant for IMS Health in 2013, and in 2014 was retained as a scientific adviser to a legal team acting on oseltamivir. In the next 12 months, TJ anticipates reimbursement for travel and accommodation in a potential lawsuit related to oseltamivir, and expert witness fees for cases involving alleged harm from vaccination. TJ is a recipient of a travel grant from the Cochrane Collaboration to attend the bi-annnual editors’ meeting in 2015.
Dear Barry Clinch, James Smith and Andy Kenwright,
Thank you for your response to our Cochrane review.  While we always appreciate feedback on our methodology, it is hard to understand why you waited until now to comment. Our initial protocol was published in December 2010 and you were specifically invited to comment. (See Professor Chris Del Mar’s letter to James Smith dated 11th of December 2010 (http://www.bmj.com/tamiflu/roche?page=3). By contrast, GlaxoSmithKline gave us much useful feedback at the planning stages which was incorporated into our methodology.
Below we provide a response to each of your points.
Roche criticism: Cochrane only included a proportion of the CSRs that Roche provided.
Your suggestion of bias in our selection process is not consistent with the facts. First, we pre-specified the criteria for selection, a standard practice for Cochrane systematic reviews. Second, our methodology was peer-reviewed. Third, we sought feedback from interested parties including Roche and GlaxoSmithKline. At no time in the last 3 years since publication of our protocol did anyone suggest a change in our selection criteria.
As you must know, it is standard practice for high quality systematic reviews to only include studies relevant to the pre-specified research question. The studies which by their characteristics (population, intervention, comparator, outcomes, timing and setting) are likely to answer the study question are clearly indicated in our review’s inclusion criteria. These were defined in December 2010 and subsequent amendments, all of which were published years before the final review. This is why we only included randomized controlled trials and did not include, for example, non-randomised studies, non-blinded studies, non-controlled studies, studies of experimental influenza and studies which did not answer the question (e.g., taste studies). Many of the CSRs that you provided, but which we subsequently excluded, were of these types. For every trial that was excluded, we gave our reasons for excluding the trial, which are documented in an appendix to the review. (http://www.bmj.com/content/bmj/suppl/2014/04/09/bmj.g2545.DC1/jeft017746...)
If you think we were biased in our selection, then the onus is on you to explain which of the trials we excluded you would include, and how you would incorporate trials with such widely varying design into a single analysis.
If you are confused as to why we asked for more trials than met our inclusion criteria, the answer is that asking for the complete oseltamivir trial programme was necessary to avoid guess work in identifying eligible studies to include. We could not rely on the public record because Roche’s publications of its oseltamivir trials are incomplete (there are many unpublished trials a decade after the trials completed) and inaccurate (published studies misreport what occurred in the studies according to the CSRs).
Roche criticism: Time to symptom alleviation should be reported as medians not means.
You suggest that we have electronic individual patient data (IPD). This is incorrect. We asked Roche whether it would be willing to provide us with IPD similar to Roche’s provision of IPD to the Harvard researchers Marc Lipsitch and Miguel Hernan. Unfortunately, we never received an answer despite follow-up (see letter by Professor Chris Del Mar to Dr Donald MacLean dated 22 January 2011 http://www.bmj.com/tamiflu/roche?page=3.)
You provided full CSRs, and the appendices of CSRs include printed (i.e. PDF) tables of patient level data. However for large analyses with many observations per participant, this “printed” format is not a realistic user option. Our conclusion on time to first alleviation of symptoms (16.8 hours) is not dramatically different than the FDA’s conclusion (1.3 days), and does not make a difference from a public health perspective because the drugs were stockpiled with an eye to the purported effect against complications, not symptoms.
However given your concerns regarding our analyses, are you now willing to provide IPD to us without restriction in a more usable format e.g. SAS datasets?
To remove any confusion, our use of means was facilitated by Roche’s use of means in the CSRs Roche provided. Without IPD, we had to use the summary information provided in the CSRs which include medians and means. If means are a poor summary of the outcome, why did you provide them in the CSRs? In fact we believe means may provide a better summary than medians as medians only reflect the middle of the distribution whereas means reflect the entire distribution of patients who recovered. Also we have observed in the survival curves that a significant proportion of treated patients tended to take a long time to recover. Again, this is not reflected when using medians.
There are limitations when censored patients are not included, however these were usually a small proportion of patients and was of similar magnitude in each group. Hence we do not believe this has introduced a bias. And we could potentially correct this limitation if we had access to IPD. In particular, we would also be interested to receive information on relapse following initial alleviation of symptoms, as this information is not provided in the CSRs. Indeed, the FDA noted that a high proportion of patients in the zanamivir trials relapsed (more in the treatment group) therefore we believe this is important information that should be publically available. In addition, IPD may allow us to analyze the effects in those on and off relief medication - this analysis was provided by some zanamivir CSRs, but not by oseltamivir CSRs. This analysis is important to provide information on whether the symptom alleviating effect is any better than common over the counter symptom relieving medications.
Also for oseltamivir the FDA could not reach any conclusion on symptom relapse because of the mix up with diary cards after day 8 of follow up in the two "pivotal" treatment trials WV 15670 and WV 15671. In addition, the numbers of patients with missing symptoms data is not reported in the CSRs, but it is mentioned that the amount of missing data is significant.
Roche criticism: Other sources should be used for evaluating harms
High quality systematic reviews tend to only include the most methodologically robust evidence. Non-randomised studies and spontaneous safety reports are not included. We agree this information is important but it is publically available and causation cannot be inferred as there is no control group for spontaneous reports and non-randomised studies are subject to unknown biases that make robust inferences difficult. Another problem that we point out in the review is that of “compliharms”. The Roche treatment trials follow a contorted logic for events (such as pneumonia or vomiting) in that they either classified them as a complication of the syndrome observed (influenza like illness and influenza) or as a harm and it is impossible to review that decision in the absence of case report forms. Observational studies and pharmacovigilance signals suffer from similar problems. Only in randomised comparisons in healthy people (as in the oseltamivir prophylaxis trials) is it possible to eliminate the alternative explanation (e.g. that vomiting is associated with exposure to oseltamivir and not to influenza-like illness).
From a methodological perspective, it is worth restating that at no time did a peer reviewer suggest we include uncontrolled studies in our review. Your criticism of our methods again only comes after the publication of our complete review, and raises the question of whether your disagreement is truly with our methods and not our results.
Our evaluation of harms reported in the CSRs is the first time these data have been made public. It provides, for the first time, evidence of statistically significant increases in important harms that were previously either unknown (renal events) or of unclear causality (psychiatric events).
Roche criticism: The Cochrane group looked for short-cuts (e.g. word clouds as a method of identifying issues of significance in regulatory documents)
Had you read the narrative of the review you would have known that word clouds were used at the beginning of the four year process when we had identified substantial reporting bias in the published treatment trial datasets but were faced with Roche’s refusal to provide us with the relevant full clinical study reports. This ‘culture of secrecy’ meant that we lacked information about what was below the publication iceberg, where the published papers represented only a proportion (potentially biased) of the entirety of the trial programme. Word clouds were one way we explored the FDA’s Summary Basis of Approval documents in addition to reading them. We think the onus is on you to show what we got wrong as a result of what we did.
Roche criticism: Human trials are the wrong place to look for mechanism of action
You say that we “attempted to infer how oseltamivir works from only the observed clinical effects.” This is misleading. What we did is to discuss the implications of observations from clinical studies to the drug’s mode of action. We found in the oseltamivir treatment trials that oseltamivir suppressed antibody production leading to misclassification of influenza infection and also cite evidence from various sources that underpins these clinical observations.
For example we wrote in the original Cochrane review: “There is decisive evidence that administration of oseltamivir in animals challenged by respiratory syncytial virus (RSV) that lacks a neuraminidase gene showed a symptom-relieving effect (decreased weight loss) and inhibition of viral clearance (Moore 2007).”[1,2] Endogenous neuraminidase suppression by oseltamivir is implicated in its mechanism.
You say that there is no evidence to show meaningful effects of oseltamivir on patients without influenza infection, however we have found in the oseltamivir treatment trials that influenza infection has been misclassified due to oseltamivir’s suppression of antibody response. Therefore it is difficult to investigate effects stratified by influenza infection given there appears to be no valid way to determine who had true influenza. In addition, we are not aware of any published analysis of just the patients classified as not having influenza infection?
We also discussed the reason why clinical usual dose of zanamivir did not suppress antibody production but may have this effect in higher dose in the Cochrane systematic review.  In fact, in the zanamivir treatment trials we could compare effects on time to first symptom relief and there was no evidence of difference between infected and non-infected patients.
Your response does not comment on these findings. We published a 4 page web appendix of apparently contradictory statements on the mode of action of oseltamivir made by Roche over the years. We request you respond and provide a complete and coherent explanation of mode of action consistent with the effects observed in humans.
Roche criticism: Given Cochrane themselves admit they are novices in dealing with regulatory information, how can they simultaneously express certainty?
We don’t express certainty; we report confidence intervals and p-values. We were necessarily novices when we started. Given that CSRs have historically been secret documents, a learning curve is inevitable. But we have learnt a lot in a short space of time and feel confident in our analysis of the CSRs.
We agree that our methods are novel, and as far as we are aware, there have not been any other systematic reviews of full clinical study reports of trials of a family of drugs. If such publications (and methodologies) exist, we would value Roche pointing these out to us.
Finally, we note with interest that Roche has not contested our review’s conclusion that there is no compelling evidence that oseltamivir reduces the risk of complications from influenza. This is a surprising omission. We assume Roche still stands by its 2003 publication of a pooled analysis of randomised trials in which Roche concludes that “Oseltamivir treatment of influenza illness reduces LRTCs, antibiotic use, and hospitalization in both healthy and ‘at-risk’ adults.” We have reviewed the same trials on which this statement is made and suggest the medical profession has been misled. We therefore look forward to a detailed response from Roche addressing this critical issue.
Jefferson T, Jones MA, Doshi P, Del Mar CB, Hama R, Thompson MJ, Onakpoya I, Heneghan CJ
1. Jefferson T, Jones MA, Doshi P, Del Mar CB, Hama R, Thompson MJ, Spencer EA, Onakpoya I, Mahtani KR, Nunan D, Howick J, Heneghan CJ. Neuraminidase inhibitors for preventing and treating influenza in healthy adults and children. Cochrane Database of Systematic Reviews 2014, Issue 4. Art. No.: CD008965. DOI: 10.1002/14651858.CD008965.pub4.
2. Moore ML, Chi MH, Zhou W, Goleniewska K, O’Neal JF, Higginbotham JN, et al.Cutting edge: oseltamivir decreases T cell GM1 expression and inhibits clearance of respiratory syncytial virus: potential role of endogenous sialidase in antiviral immunity. Journal of Immunology 2007;178(5): 2651–4.
3. Kaiser L, Wat C, Mills T, Mahoney P, Ward P, Hayden F. Impact of oseltamivir treatment on influenza-related lower respiratory tract complications and hospitalizations. Arch Intern Med 2003;63:1667-72.
Competing interests: TJ, MJ, MT, CDM, RH, CH and PD are co-recipients of a UK National Institute for Health Research grant (HTA – 10/80/01 Update and amalgamation of two Cochrane Reviews: neuraminidase inhibitors for preventing and treating influenza in healthy adults and children—http://www.nets.nihr.ac.uk/projects/hta/108001). In addition: PD received €1500 from the European Respiratory Society in support of his travel to the society’s September 2012 annual congress in Vienna, where he gave an invited talk on oseltamivir. Dr Doshi is also an associate editor of The BMJ. TJ receives royalties from his books published by Blackwells and Il Pensiero Scientifico Editore, Rome. TJ is occasionally interviewed by market research companies for anonymous interviews about Phase 1 or 2 pharmaceutical products. In 2011-2013, TJ acted as an expert witness in a litigation case related to oseltamivir phosphate; Tamiflu [Roche] and in a labour case on influenza vaccines in health care workers in Canada. In 1997-99 TJ acted as consultant for Roche, in 2001-2 for GSK, and in 2003 for Sanofi-Synthelabo for pleconaril (an anti-rhinoviral which did not get approval from FDA). TJ was a consultant for IMS Health in 2013, and is currently retained as a scientific advisor to a legal team acting on the drug Tamiflu (oseltamivir, Roche). CDM was a Board member of two companies to commercialise research at Bond University, part of his responsibilities as Pro-Vice Chancellor (Research) until 2010 and receives fees for editorial and guideline developmental work and royalties from books and in receipt of institutional grants from NHMRC (Aus), NIHR (UK) and HTA (UK) and from a private donor (for support of the editorial base of the Cochrane ARI Group). RH receives royalties from two books published in 2008 titled “Tamiflu: harmful as was afraid” and “In order to escape from drug-induced encephalopathy”. Dr Hama provided scientific opinions and expert testimony on 11 adverse reaction cases related to oseltamivir and gefitinib. CH receives payment for running educational courses at the University of Oxford and University of Oxford ISIS consulting services for external teaching and training. He also receives royalties for books (Evidence Based Toolkit series by Blackwell BMJ Books).
Re: Oseltamivir for influenza in adults and children: systematic review of clinical study reports and summary of regulatory comments
Roche welcomes all third-party research and is dedicated to sharing clinical data for our medicines in the interest of advancing science. We provided clinical study reports (CSRs) for all completed Roche-sponsored studies investigating oseltamivir (77 CSRs) to the Cochrane Acute Respiratory Infection group. Despite insisting that they required full CSRs, and in particular, individual patient data (IPD), the authors selected only 15 CSRs (20 studies) and their analyses are based on study level meta-analysis rather than IPD. Roche has identified many issues with the approach taken to analyse these 20 studies, including the methodology used and assumptions made by the authors. We have summarised some key concerns based on the BMJ summary article “Oseltamivir for influenza in adults and children: systematic review of clinical study reports and summary of regulatory comments” here and will provide a more detailed response directly to the Cochrane group.
We consider the analysis of the time to alleviation of symptoms used by Cochrane to be seriously flawed. For the majority of the studies, the authors have used SAS generated Kaplan-Meier mean estimates reported in the CSRs, but have evaluated the time to alleviation of symptoms as a continuous endpoint, ignoring the distribution of the data (see Figure; example from study M76001) and presence of censoring. For studies WV15707, WV15730, and WV16277, it is unclear how these summary mean estimates were derived as we have been unable to reproduce them. We would consider the most accurate scientific approach to be an IPD meta-analysis using an appropriate survival model. However, if using a two-stage or aggregated data meta-analysis, as per the Cochrane analysis, then the medians and associated 95% confidence intervals quoted in the CSR summary tables (derived from the Mantel-Haenszel test) should have been used, as these are a better representation of the skewed distribution of the time-to-event data.1 The consequence of ignoring the underlying distribution and the censored nature of this data is that it leads to an inaccurate estimate of the true treatment effect and undermines the author’s conclusions.
In arriving at their safety assessment, the authors used data from only 20 randomised clinical trials, which is not an acceptable approach for a product that has been licensed for over 15 years and taken by approximately 130 million patients. In order to properly make a safety assessment, all data must be evaluated, including all relevant CSRs, published literature, and spontaneous safety reports. There were 35,358 adverse events reported to Roche between 21 September 1999 and 20 September 2013, which is almost triple the number of events considered by the authors from their analysis of only 20 randomised clinical trials. It is this total dataset that Roche and regulators worldwide review and monitor (according to standard pharmacovigilance practice) on a continuous basis to arrive at a thorough and reliable assessment of the safety of a medicine. As a result, no new safety information has been identified by the authors that Roche and the regulators have not considered through the extensive and continuous pharmacovigilance process. Product labelling is updated as part of the ongoing benefit/risk assessment of oseltamivir performed by Roche and regulators worldwide, based on all data sources mentioned above.
The authors state that “the influenza virus-specific mechanism of action proposed by the producers does not fit the clinical evidence”, but in making this assertion they have attempted to infer how oseltamivir works from only the observed clinical effects. This is neither necessary nor scientifically valid given the basic science that underpins neuraminidase inhibitor development. Oseltamivir is the product of rational drug design using data from high resolution X-ray crystal structures of the influenza receptor (sialic acid) bound to the viral neuraminidase.2 The result of this deliberate design is that oseltamivir carboxylate is a potent and specific inhibitor of influenza A and B neuraminidases with a high degree of selectivity and antiviral potency. Given the intentional design of oseltamivir and the existence of compelling data to support its specific anti-influenza mode of action we do not understand how the authors make their claims while offering no explanation for the development of treatment emergent oseltamivir-resistant viruses or the absence of any meaningful effect of oseltamivir on symptoms in patients who have no evidence of influenza infection.
The Cochrane authors have drawn many conclusions that would not be supported by a methodologically robust and comprehensive analysis of all relevant oseltamivir data. Equally important, they have not exhibited the diligence necessary when working with large clinical trial datasets and extensive regulatory documents. They claim these materials are necessary to conduct their review, but either do not use the information or look for short-cuts (e.g. word clouds as a method of identifying issues of significance in regulatory documents). While we note their honesty in stating “none of the authors (all experienced systematic reviewers) had any experience of reviewing regulatory information” and that their “novel methodology remains a work in progress”, it is hard to understand how they can then be so certain of their conclusions.
In summary, the Cochrane authors made basic errors in their assessment of efficacy data, over-interpreted the limited safety data they evaluated, and wholly ignored certain types of data. This has resulted in inappropriate statements about the mode of action, efficacy, and safety of an important and long established medicine. Third party analyses can be valuable, but in our view this work adds nothing to the abundance of data on oseltamivir; the conclusions drawn risk unnecessarily confusing both patients and physicians and could lead to public health risks due to patients not complying with their prescribed medicine.
1. Parmar M, Machin D. Survival Analysis - A Practical Approach. John Wiley & Sons, 1995.
2. Lew W, Chen X, Kim CU. Discovery and development of GS 4104 (oseltamivir): an orally active influenza neuraminidase inhibitor. Curr Med Chem 2000;7:663-72.
Competing interests: BC, JS and AK are employees of Roche
Re: Oseltamivir for influenza in adults and children: systematic review of clinical study reports and summary of regulatory comments
The comments colleagues and I made previously still stand: http://www.bmj.com/content/339/bmj.b2728
Early in the "containment phase" of the pandemic the health secretary promised that everybody with flu-like symptoms would be given antivirals.
The effect of this was that we were obliged to honour this promise. There were not the resources to do this in a timely fashion: most people, while I was involved, were delivered the antivirals a week after the onset of symptoms. We were unable to prioritise high-risk patients.
The government has claimed that the "containment phase" was effective; that the reduction in viral shedding that may plausibly have been a consequence of antiviral treatment slowed the spread of the pandemic.
There is no reason to believe that antivirals given more than 48 hours after the onset of symptoms had any other benefit; although respiratory physicians tell me they believe it may have had some benefit in the most seriously ill patients.
Competing interests: I was seconded to work in a flu response centre early in the 2009 pandemic
Re: Oseltamivir for influenza in adults and children: systematic review of clinical study reports and summary of regulatory comments
The BMJ and Cochrane Collaboration have revealed serious generic failings in our system of publishing full trial evidence. The lessons in relation to antiviral drugs and future pandemics still need to be properly debated. Service insights from the last pandemic have been little in evidence in the current debate.
Oseltamivir was considered, even in 2009, to be a drug of hope, of weak benefit if given within 48 hours of symptoms, ‘free of side effects’ and ‘ the best we have got’. The limited evidence base around seasonal flu use was the basis for pandemic planning. But even this was sacrificed by Governments, their Departments of Health and the WHO when their ‘expert committees’ encouraged use of the drug, once the pandemic was declared, in people who may have been ill for a week. This led us into massively expensive over reliance on a drug of marginal, if any, benefit during the 2009/10 H1N1 pandemic.
Apologists for Roche are already mounting a rearguard argument that the Cochrane analysis relates to trials undertaken on seasonal flu, not pandemic flu. But this is a bizarre reinterpretation of history given it was precisely the evidence about seasonal flu trials which was used to determine national policy for pandemic stockpiling. The evidence was extrapolated by Departments of Health and governments, advised by experts, paid for by the industry, that said ‘antiviral distribution may be the best we can do’, ‘we don’t have the evidence for these circumstances but we have to make best guesses’, ‘we have to take action in case it becomes a serious infection’. Like Cochrane, like the ALLTrials campaigners we conclude that there would have been even less justification for use in the pandemic, had the limited evidence for its use in seasonal flu been known from the start.
We remain concerned however, that the insights of service providers on the ground during the pandemic have not been given the same level of public consideration. It is well documented that even drugs which may be effective in clinical trials can be inefficient, or fail to deliver the patient benefit predicted by trial results, when subject to limitations of general service use. The lessons from service insights in the pandemic are: the wanton abandonment of first principles such as isolation, basic control of infection measures and clinical assessment in favour of the stubborn insistence on managing ‘England as a single epidemiological unit’ ; and the irrational maintenance of the ‘containment’ phase which led directly to perverse and damaging interventions and over-reliance on antivirals in mass prophylaxis exercises particularly in schools. Anti-viral collection centres became loci for the spread of infection as thousands of symptomatic and sub-clinical cases (there is good evidence that flu can be spread by asymptomatic patients [2, 3]) and unaffected contacts convened for a wonder drug with serious potential side effects ,and which would now appear to be no more effective in pandemic management than paracetamol .
It would be irresponsible for these lessons not to underpin current planning for pandemics and any subsequent responses. We believe there is no place for antiviral distribution in a pandemic based on the current evidence of the effectiveness of the drugs, their ineffectiveness for mass prophylaxis and the likely spread of infection brought about by bringing people to a centre for the drugs.
 Chambers J, Barker K, Rouse A. Reflections on the UK’s approach to the 2009 swine flu pandemic: Conflicts between national government and the local management of the public health response. Health Place 2012; (18): 737–745.
 Centers for Disease control. Additional Information about Vaccination of Specific Populations. Influenza Prevention and Control Recommendations. Published for the 2010-11 Influenza Season; Adapted for the 2012-13 Influenza Season http://www.cdc.gov/flu/professionals/acip/specificpopulations.htm last accessed 13th April 2014.
 Chao D-Y, Cheng K-F, Li T-C, Wu T-N, Chen C-Y, et al. Serological Evidence of Subclinical Transmission of the 2009 Pandemic H1N1 Influenza Virus Outside of Mexico. PLoS ONE 2012; 6(1): e14555. doi:10.1371/journal.pone.0014555
 Jefferson T, Jones M, Doshi P, Spencer E, Onakpoya I, Heneghan C. Oseltamivir for influenza in adults and children: systematic review of clinical study reports and summary of regulatory comments. BMJ 2014; 348:g2545
 Tamiflu: Millions wasted on flu drug, claims major report. BBC News. http://www.bbc.co.uk/news/health-26954482. Last accessed 13th April 2014.
Competing interests: No competing interests
In the BNF it clearly states that Oseltamivir is licensed for use in the first 48 hours following the onset of symptoms, and that it is a prescription only medicine.
If Oseltamivir had been effective there is no way that the public could have accessed it rapidly enough.
If people had died in large numbers then there would have been a panic at every GP surgery in the country with the possibility of people dying in the crush.
Even if an effective drug comes on the market it would be useless unless it could be distributed and taken rapidly in accordance with its instructions for use.
Competing interests: No competing interests
With the publication of our systematic reviews on oseltamivir and zanamivir for influenza in adults and children, [1,2] we are making all 107 full clinical study reports publicly available.
Despite the worldwide stockpiling of antivirals, these reports have never been reviewed by the World Health Organization, the US Centers for Disease Control and Prevention (CDC) and its European counterpart, the ECDC.
It took us four years to obtain the full set of reports. The story relating to the acquisition has been documented at the BMJ's open data campaign (http://www.bmj.com/tamiflu).
If you disagree with our findings, or if you want to carry out your own analysis or just want to see what around 150,000 pages of data look like, they are one click away (http://dx.doi.org/10.5061/dryad.77471).
Carl Heneghan, Tom Jefferson, Peter Doshi
 Tom Jefferson, Mark Jones, Peter Dosh, Elizabeth A Spencer, Igho Onakpoya, Carl J Heneghan. Oseltamivir for influenza in adults and children: systematic review of clinical study reports and summary of regulatory comments
 Carl J Heneghan, Igho Onakpoya, Matthew Thompson, Elizabeth A Spencer, Mark Jones, Tom Jefferson. Zanamivir for influenza in adults and children: systematic review of clinical study reports and summary of regulatory comments
Competing interests: We are authors of the Systematic reviews and our competing interests are declared in the manuscripts.