Source: Jefferson T, Jones MA, Doshi P, et al. Neuraminidase inhibitors for preventing and treating influenza in healthy adults and children. Cochrane Database Syst Rev. 2014;(4):CD008965.
Study Population: Treatment: Adults and children with confirmed or suspected to have influenza and also those with confirmed or possible exposure to influenza
Efficacy Endpoints: Hospitalization and time to symptom alleviation
Harm Endpoints: Adverse events (e.g. nausea, vomiting, and neuropsychiatric events)
Narrative: Neuraminidase inhibitors (NAIs) are commonly used in the prevention and treatment of influenza. Previous studies and reviews have demonstrated a questionable and modest benefit of their use while demonstrating potential adverse effects.1, 2
The most recent Cochrane review discussed here3 analyzes the data from randomized controlled trials evaluating the effects of NAIs in children and adults with confirmed or suspected exposure to influenza. The systematic review included trials testing the effectiveness and safety of two commonly used NAIs, oseltamivir (Tamiflu) and zanamivir (Relenza). Twenty-three oseltamivir trials were included, with a total of 9623 subjects (6574 in treatment trials and 3049 in prophylaxis trials) with ages ranging from 1 to 82 years. Twenty-eight zanamivir trials were included, with a total of 14,628 subjects (7678 in treatment trials and 6950 in prophylaxis trials) with ages ranging from 5 to “over 65.” The authors’ novel methodology for study selection was complicated. All of the included trials were industry-supported randomized control trials (RCT) comparing oseltamivir or zanamivir versus placebo, some of which were unpublished reports from manufacturers (hence not peer-reviewed). Studies were excluded if they were not placebo RCT’s, were pharmacokinetic studies, were dose comparison studies, or were ongoing trials.
The authors of the Cochrane systematic review state that they are not convinced they had access to all existing manufacturer data. They report higher risk of bias with frequent lack of reporting of random sequence generation methods, incomplete data (symptoms, complications, and safety data), and concern for lack of full blinding of participants and personnel.3 The Cochrane review included trials that enrolled previously healthy children or adults diagnosed with influenza (with or without symptoms) for the treatment analysis, and similar populations who took NAIs for prophylaxis (with or without exposure).
The primary outcome measures for treatment analysis were: 1. symptom relief, 2. hospitalization, and 3. harms. The primary outcome measures for prophylaxis analysis were: 1. influenza (symptomatic and asymptomatic, with laboratory confirmation) and influenza-like illness (ILI), and 2. hospitalization.
Treatment: In adults, oseltamivir reduced time to first alleviation of symptoms by 16.8 hours (95% CI 8.4 to 25.1). Zanamivir reduced the time to first alleviation of symptoms in adults by 14.4 hours (95% CI, 9.4 to 19.2 hours). In healthy children, based on a single study, oseltamivir reduced the time to first alleviation of symptoms by 29 hours. However, in asthmatic children, there was an increase in time to first alleviation of symptoms by 5.2 hours (95% CI 11.1 hours lower to 21.4 hours higher). Standard asthma medical care and close follow-up may overshadow the incremental benefit of oseltamivir in this specific population. The effect of zanamivir on time to first alleviation of symptoms in children was not statistically significant. Data for asthmatic adults was not reported. Oseltamivir treatment did not reduce the risk of hospitalization in children or adults.
Serious adverse events or those leading to withdrawal from the study were higher in treatment groups. Oseltamivir treatment in adults was associated with nausea (Relative risk [RR]: 1.57, 95% CI 1.14 to 2.15; Absolute risk difference [ARD] 3.7%; Number needed to harm [NNH]: 28) and vomiting (RR: 2.43, 95% CI 1.75 to 3.38%; ARD: 4.56%; NNH: 22). Vomiting was also seen in pediatric studies (RR: 1.70, 95% CI 1.23 to 2.35; ARD: 5.3%; NNH: 19). There was no significant increase in the risk of neuropsychiatric events during treatment.
Oseltamivir reduced self-reported, unverified pneumonia (RR: 0.55, 95% CL 0.33-0.90). The lack of clear pneumonia definitions across studies makes this outcome unreliable. There was no reduction in pneumonia with zanamivir. Neither oseltamivir nor zanamivir reduced incidence of sinusitis, bronchitis, or otitis media.
Prophylaxis: Oseltamivir prophylaxis was initiated on the basis of “local outbreaks” which were not well elucidated. Patients generally took the medication for 6 weeks. Zanamivir prophylaxis was initiated based on population characteristics (e.g. patients in nursing homes) without an indication based on direct exposure or local outbreaks. Patients were treated for 28 days. There were no pediatric prophylaxis studies.
The studies reporting the benefits of NAIs for prophylaxis were of low quality (significant biases), had small sample sizes, and did not clearly define the indications for giving prophylaxis. Therefore, we are refraining from reporting the data for this particular indication here.
Caveats: The use of NAIs for the treatment of influenza in adults confers a small decrease in time to symptom alleviation (17 hours). The statistical heterogeneity for the treatment analysis was reported to be low. The Cochrane review employed the use of clinical study reports from national drug regulators, in addition to studies published in biomedical journals. This resulted in a comprehensive study that better represents the effects of these medications. Although the meta-analysis only included RCT’s, limitations such as selective publication, inclusion of non-peer-reviewed data, and high attrition rates negatively impacted the quality of the evidence. A subgroup analysis on the treatment effect of zanamivir between influenza-positive and influenza-negative patients revealed that both populations had an identical duration of reduction in symptoms. A recent open-label RCT published in January 2020 (the ALIC4E trial4) also showed that patients with symptomatic influenza treated with oseltamivir plus usual care recovered 1 day (95% CI 0.74 to 1.31) sooner than those who received usual care alone. As demonstrated in the Cochrane review, this reduction occurred equally in influenza positive and influenza negative groups, irrespective of laboratory confirmed influenza. This suggests that NAIs most likely do not have an influenza-specific effect in symptom alleviation, or as likely that this was a placebo effect given that the study was unblinded and patients knew if they were receiving an NAI. The use of NAIs in healthy children may reduce time to symptom alleviation by about a day, but carries the risk of adverse events.
The systematic review did not report outcomes where incidence was less than 0.5%, and this included mortality. Lack of clarity regarding mortality is unfortunate, given the estimated 200,000-600,000 estimated annual deaths worldwide due to influenza.5 It is important to note that the Infectious Diseases Society of America (IDSA)6 and the Centers for Disease Control and Prevention (CDC)7 endorse the consideration of NAIs for all people with influenza-like illness within 48 hours of onset. These groups recommend NAIs for all persons at higher risk of dying from influenza (i.e. immunocompromised status, advanced age, multiple comorbidities) and for all persons hospitalized with influenza. These recommendations are based on observational data and expert opinion. Unfortunately, we do not have strong data to support these recommendations, and given their widespread use, it is hard to imagine we will see a randomized control trial in the future, answering the questions “Do NAIs save lives?” or “Is there a benefit with NAIs in people who are very sick (hospitalized)?”
The data for NAIs are further confounded by the financial conflicts of interest that are present in many of the studies on NAI use. Dunn et al8 demonstrated that among NAI studies associated with a FCOI, 88% of the studies were classified as favorable. For those without FCOI (among which was this Cochrane review), 17% of the studies were classified as favorable. Thus, much of the positive data regarding NAI is heavily biased.
In summary, the existing data indicate that NAIs reduce the duration of symptoms by less than a day in patients with confirmed or suspected influenza. The use of NAIs to treat influenza does not prevent hospitalization and is associated with adverse events. Therefore, we have assigned a color recommendation of yellow (unclear if it provides benefit, more data needed) to this treatment.
The original manuscript was published in Academic Emergency Medicine as part of the partnership between TheNNT.com and AEM.
See theNNT.com's previous reviews of this topic:
Neuraminidase Inhibitors Given for Influenza, January 26, 2010
Author: Marc Zwillenberg, MD; Eric Tang, MD; Joshua Quaas, MD
Supervising Editors: Kabir Yadav, MD
Published/Updated: April 1, 2021
The title bar is color-coded with our overall recommendation.
If you have suggestions, requests, or questions about a particular NNT review, please send us a message and we’ll try to address it as soon as possible.