Resistant on Day Four: Switching Antivirals Mid-Admission for Influenza
A single patient, worsening on hospital day five despite oseltamivir, with genotypic confirmation that the virus itself has already answered the question everyone is still debating. The disagreement is about what to do with a drug that's still in his system doing nothing.
D.W., a 58-year-old long-haul truck driver with poorly controlled type 2 diabetes, was admitted four days ago with influenza A pneumonia and started on oral oseltamivir within the first 24 hours — early, by any standard, and exactly the scenario the drug is supposed to work in. He hasn't gotten better. His oxygen requirement has climbed from 2L nasal cannula on admission to 6L today, his repeat chest imaging shows new bilateral infiltrates rather than resolving ones, and a nasopharyngeal sample sent yesterday, on hospital day four, for genotypic resistance testing — ordered specifically because four full days of appropriately-dosed oseltamivir with clinical worsening is itself the textbook trigger for that test — has come back this morning positive for the H275Y neuraminidase substitution.
H275Y is not a partial or borderline finding; CDC's own influenza antiviral guidance treats it as the marker conferring high-level oseltamivir resistance in A(H1N1)pdm09 virus, and because peramivir binds the same neuraminidase active site, cross-resistance to peramivir travels with it. Every hour he has continued taking oseltamivir since that mutation was almost certainly already present — his day-four sample, not a fresh event — has been treating him with a drug his own virus doesn't answer to. Baloxavir marboxil works through an entirely different step in the viral life cycle, inhibiting the cap-dependent endonuclease that initiates viral mRNA synthesis rather than the neuraminidase that releases new virions, which is exactly why the XOFLUZA label's own microbiology section records that it retains activity against neuraminidase-inhibitor-resistant strains, H275Y included. That mechanistic independence is the fact the whole bedside conversation now turns on: does switching to a drug with a genuinely different target, in a patient still worsening and still diabetic, mean simply switching, or does his severity argue for hitting two targets at once? The genotyping lab flagged one further wrinkle worth naming plainly: H275Y was identified on a bulk-sequencing assay of his day-four sample, a method that only detects a resistant variant once it represents a meaningful share of the total viral population — meaning the mutation was almost certainly already dominant well before today's report returned, not something that emerged fresh under four days of selective drug pressure.
Bedside, hospital day five
Stop the oseltamivir and start baloxavir marboxil, a single dose today. H275Y confers high-level resistance at the neuraminidase active site — this isn't a partial or borderline reduction in susceptibility, it's the mutation the whole resistance-testing protocol exists to catch. Every dose of oseltamivir he takes from here is doing nothing antiviral while still costing him GI tolerability in a patient whose glucose control is already fragile.
If this were a susceptible strain still worsening on appropriate dosing, I'd be looking at absorption or severity, not the drug itself — the genotype is what changes the whole conversation here.
I'd add baloxavir rather than stop the oseltamivir outright. He's gone from 2L to 6L of oxygen in four days with new bilateral infiltrates — that's a patient actively losing ground, not a stable one where I'm comfortable narrowing to a single mechanism on the strength of one resistance sample while we wait to confirm it on a repeat specimen. Baloxavir's activity doesn't depend on oseltamivir being present or absent; there's no pharmacologic reason continuing it alongside baloxavir does harm.
I'm not disputing that H275Y confers real resistance — I'm saying "resistant to oseltamivir" and "harmful to keep giving oseltamivir" aren't the same claim, and in a patient this sick I'd rather hedge one extra day than narrow therapy on a single sample.
I'd stop the oseltamivir — a confirmed high-level resistance mutation isn't the same evidentiary situation as an unconfirmed one, and there's a real cost to leaving him on it that's been undersold in this conversation: his fingersticks have run 220 to 310 despite sliding-scale coverage, and hyperglycemia independently impairs neutrophil function and viral clearance. Whatever we decide about the antiviral, his insulin regimen needs to stop being background noise to this discussion.
That's not a smaller point than the antiviral switch — it may be the more modifiable one, and it's been getting less attention than a drug decision that, on the genotype in hand, is actually the more straightforward call of the two.
Agreed: baloxavir marboxil given as a single dose, oseltamivir discontinued, and his insulin regimen escalated to basal-bolus dosing with a target range reset for the admission. A repeat resistance genotype was not pursued — the intensivist's request to confirm on a second sample was set aside once the group agreed the first result was unambiguous and re-testing would only delay a switch already judged necessary.
Not separately debated but explicitly acknowledged: the critical care physician's preference for combination therapy in a deteriorating patient wasn't wrong in principle — it simply didn't survive contact with a confirmed, target-level resistance result. His oxygen requirement and glycemic control were both flagged for reassessment at 48 hours, with the door left open to escalate respiratory support independent of whichever antiviral turns out to have mattered.