A network meta-analysis of 73 trials comparing the effectiveness of flu antivirals suggests that only baloxavir (Xofluza) may be better than standard care or a placebo in reducing the risk of hospitalization for high-risk patients with non-severe flu and likely speeds symptom relief with few or no adverse effects but may increase antimicrobial resistance in 10% of recipients. A
Flu antiviral experts, however, tell CIDRAP News that they take issue with its conclusions that oseltamivir (Tamiflu) had little to no effect on death or hospitalization for non-severe flu, likely didn't shorten time to symptom relief, and probably increased treatment-related adverse events. And they say it won't change current US clinical practice, in which both baloxivir and oseltamivir have a role.
"We have very, very strong evidence that oseltamivir decreases duration of symptoms in adults and children and that both of those medications can decrease complications of influenza, such as pneumonia, sinusitis, and ear infections," said James Antoon, MD, PhD, MPH, assistant professor of pediatrics at Vanderbilt University Medical Center.
"We know that subsequent antibiotic use in those with influenza is also reduced with those antivirals, because we're preventing those secondary bacterial infections," he said. "I think it's still somewhat unclear the benefit of these antivirals for some rare outcomes [death and hospitalization], although we have evidence from observational studies that are supportive for some of those rare outcomes."
For the network meta-analysis, published last week in JAMA Internal Medicine, researchers in China and Canada analyzed pooled data from 73 randomized clinical trials published from 1971 to 2023 involving 34,332 participants. Their aim was to compare direct-acting flu antiviral drugs to placebo, standard care, or another antiviral drugs for patients with mild or moderate flu.
In addition to baloxavir and oseltamivir, the study also included laninamivir, zanamivir, peramivir, umifenovir, favipiravir, amantadine, and rimantadine.
"Compared with standard care or placebo, all antiviral drugs had little or no effect on mortality for low-risk patients and high-risk patients (all high certainty)," the study authors wrote. "All antiviral drugs (no data for peramivir and amantadine) had little or no effect on hospital admission for low-risk patients (high certainty)."
Oseltamivir (risk difference [RD], -0.4%; 95% confidence interval [CI], -1.0 to 0.4; high certainty) had little or no effect on hospitalization in high-risk flu patients, but baloxavir may have reduced the risk (RD, -1.6%; 95% CI, -2.0 to 0.4; low certainty). All other drugs suggested few or uncertain effects.
Baloxavir probably shortened symptom duration (mean difference [MD], -1.02 days; 95% CI, -1.41 to -0.63; moderate certainty), umifenovir (not sold in the United States) may have reduced it (MD, -1.10 days; 95% CI, -1.57 to -0.63; low certainty), and oseltamivir probably had no important effect (MD, -0.75 days; 95% CI, -0.93 to -0.57; moderate certainty).
Baloxavir was tied to few or no adverse events (RD, -3.2%; 95% CI, -5.2 to -0.6; high certainty), while oseltamivir likely increased them (RD, 2.8%; 95% CI, 1.2 to 4.8; moderate certainty).
"Our findings support the use of baloxavir for treatment of high-risk patients with nonsevere influenza," the researchers concluded. "However, drug resistance monitoring may be needed because baloxavir may increase emergence of drug resistance."
"Future studies could focus on patient-important outcomes (eg, admission to ICU [intensive care unit], duration of hospitalization) and test important subgroup analyses such as the influenza virus type, age, and time from onset of symptoms to treatment," they added.
Andrew Pavia, MD, of University of Utah Health, said that the meta-analysis's interpretation is problematic. "The interpretation is that the treatment effect for all [antivirals] falls below the pre-specified, minimally important difference that WHO [World Health Organization] asked for," he said. "Therefore, rather than saying it's a modest effect, they have a dichotomous answer which says no. And that's simply clinically not a useful answer."
The WHO, which commissioned the meta-analysis, used different criteria and lower setpoints than those used in higher-income countries, because it is concerned about treatment costs in poorer countries, he said.
Oseltamivir was estimated to reduce symptom duration by 0.73 days, below the WHO threshold for clinical meaningfulness of 1.0 day, but other studies have estimated the benefit to be 1 day or slightly more, and the confidence intervals in the network meta-analysis and those studies had overlapping confidence values, Pavia noted. "So both estimates are within a reasonable confidence band of being identical, but you come up with opposite conclusions? That makes little sense," he said.
Antoon pointed out that a network meta-analysis differs from a traditional meta-analysis, which focuses on direct comparisons. "Studies that have looked at direct comparisons between oseltamivir and baloxavir have found very similar effectiveness as far as duration of symptoms, and those were large randomized trials," he said. "Those effects were not captured in this study, probably because the effect is watered down by those indirect comparisons."
Likewise, Pavia said that the network meta-analysis's inclusion of trials in which a quarter or third of participants didn't have the flu also likely diluted the effect. This group, if excluded, would have raised oseltamivir's estimated effectiveness in speeding symptom relief to 0.83 days.
Pointing out that baloxivir is much more expensive than oseltamivir, which costs $25, he said he will continue his current protocol for treating flu patients. "We use oseltamivir as our first-line drug," he said. "In rare cases of critically ill patients, like transplant patients who can't control viral replication, I am still using oseltamivir and baloxivir together."
"In those rare circumstances of someone like a bone marrow patient who's critically ill with influenza, until I have better evidence that it doesn't help, I'm using combination therapy," Pavia added. "The other setting in which I definitely use baloxivir first is when influenza B is circulating, because it's clearly a superior drug in that setting."