A major international effort to assess the impact of neuraminidase inhibitors during the 2009 H1N1 pandemic found that adults who were hospitalized were 25% less likely to die if they were given the drugs, with even better odds if they received it within 2 days of getting sick.
The research collaboration, formed in October 2011 and coordinated by a group at the University of Nottingham, included data from more than 29,000 patients in 38 countries. The team published its findings today in an early online edition of The Lancet Respiratory Medicine.
Use of neuraminidase inhibitors in treating severe flu infections has been based largely on the apparent beneits reported in observational studies. Because placebo-controlled studies during a pandemic wouldn't be ethical, the team believed that a major effort to pool data and uniformly adjust for treatment confounders would be the next best way to scientifically flesh out the impact of the drug.
The study was funded by Roche, the maker of Tamiflu. However, the team noted that the funder had no role in the study design, data collection, data analysis, data interpretation, or writing of the report. They also said Roche has not had and will never have access to the study data.
Jonathan Nguyen-Van-Tam, MD, PhD, lead author of the study and leader of the Health Protection Research Group at the University of Nottingham, said in a Lancet press release that many governments have Tamiflu stockpiles that are close to expiring and have no data on whether the drugs helped save lives during the pandemic, leaving them with little to guide decisions on replenishing their supplies.
He added that in a pandemic setting, a vaccine arrives 6 months too late, even in a best-case scenario. "Thus we are left with antivirals like Tamiflu and public health measures like hand washing and social distancing as the only defenses we have for the first 6 months of a pandemic."
78 studies examined
The team analyzed data from 78 studies involving 29,234 patients of all ages. Though the study reflected a broad international cross-section, there were relatively fewer cases from the African and Southeast Asian regions. They looked at deaths in hospitalized patients with confirmed or suspected 2009 H1N1 infections between Jan 2, 2009, and Mar 14, 2011. Modeling methods were used to control for the likelihood of treatment bias and other confounders such as corticosteroid or antibiotic use.
Overall, treatment with neuraminidase inhibitors, primarily oseltamivir, reduced the risk of death by 19% compared with no treatment. However, the risk of death was halved when patients were treated within 48 hours of symptom onset, compared with later treatment or no treatment. For each day (up to 5 days) that antiviral treatment was delayed after the 2-day period, researchers saw a 20% increase in the risk of death.
In all adults, treatment reduced the risk of death by 25%, the authors found. Improved survival rates were also seen in pregnant women and in adults in intensive care—even with late treatment.
On the other hand, though the team found reduced mortality in children ages 0 to 15 years who were hospitalized and received neuraminidase inhibitors, the effect didn't reach statistical significance.
The researchers wrote that the findings are in line with other studies, but are more precise because of the adjustment for confounders. They added that the lack of mortality reduction in kids could result from several factors, such as a lower case-fatality rate leading to reduced statistical power, or higher viral loads that might affect drug effectiveness and dosing.
The team pointed out that one of the study limitations was their inability to adjust for disease severity, because severity measures varied across the different data sources, and they warned that there might be residual confounding.
However, they concluded that the study is the most rigorous one so far to assess the mortality benefits of neuraminidase treatment during the 2009 H1N1 pandemic, and the findings show that the biggest benefit is achieved when treatment is started within 2 days of symptom onset. Also, the results seem to vindicate government decisions to stockpile the drugs, the group wrote.
Influenza clinical guidance should put even more emphasis on early empirical treatment for hospitalized patients, as well as outpatients who have suspected or confirmed flu and are at increased risk of complications, the authors added.
Support for previous findings
In an accompanying editorial, Alicia Fry, MD, MPH, an epidemic intelligence officer with the US Center for Disease Control and Prevention's (CDC's) National Center for Immunization and Respiratory Diseases, wrote that the study adds to evidence provided by other observational studies but addresses several biases that can occur when treatment isn't randomly assigned and shows how complex the analyses can be.
She wrote that when the researchers controlled for treatment timing bias, they found a significant reduction in death when compared to no treatment, and that the results bolstered what's already known about early versus late treatment.
The many possible reasons for a lack of significant benefit in kids need to be reassessed and explored in clinical studies and animal models, Fry wrote. However, in the absence of any other flu-specific treatment, the potential benefits of neuraminidase inhibitors for kids are still substantial and outweigh the risks, she added.
Fry wrote that ongoing efforts to find other flu drugs that have other virus targets or mechanisms of action might help improve the treatment of patients with severe illness and cut death rates even further.
Muthuri SG, Venkatesan S, Myles PR, et al. Effectivness of neuraminidase inhibitors in reducing mortality in patients admitted to hospital with influenza A H1N1pdm09 virus infection: a meta-analysis of individual participant data. Lancet Respir Med 2014 Mar 19 [Abstract]
Fry A. Effectiveness of neuraminidase inhibitors for severe influenza. (Editorial) Lancet Respir Med 2014 Mar 19 [Extract]
Mar 19 Lancet press release