Empiric Acyclovir in Suspected HSV Encephalitis: Treating Before the Test Returns
A clinical picture concerning for herpes encephalitis, an HSV PCR still a day away, and a mild volume-depletion complication that turns urgent treatment into a question of how to start it, not whether.
M.A., a 52-year-old man, teaches shop class at the same high school he graduated from and spends most weekends in his garage building furniture for a small side business he started two years ago. He has no chronic medical conditions and takes no regular medications; by his wife's account, his only real standing habit is that he “never remembers to drink water once he's cutting wood.” Four days ago he developed a dull, worsening headache and a low-grade fever he assumed was a cold. Yesterday his wife noticed he was repeating himself and struggling for ordinary words mid-sentence, and this morning, moments after waking, his right arm and face began twitching rhythmically for close to a minute before he came back to baseline, frightened and confused. In the emergency department his temperature is 38.6°C, he is oriented to person only, and he has mild word-finding difficulty but no focal weakness. His mucous membranes are dry and his skin tents slightly at the forearm; by his and his wife's account he has taken in almost nothing by mouth for two days between the headache and intermittent nausea.
MRI shows asymmetric T2/FLAIR hyperintensity centered in the right temporal lobe without significant mass effect — a distribution that, on its own, is exactly where herpes simplex encephalitis is expected to show up first, though it is not by itself proof of it. Lumbar puncture shows a mildly elevated opening pressure, a modest lymphocytic pleocytosis, and mildly elevated protein with normal glucose — a nonspecific viral pattern compatible with, but not diagnostic of, HSV. HSV PCR has been sent and the result will not be back for roughly a day. That gap matters on its own terms: a negative result drawn this early in a genuinely evolving case has never reliably ruled the diagnosis out. His level of consciousness is the variable the outcome data actually turn on. In the NIAID Collaborative Antiviral Study Group's acyclovir-versus-vidarabine trial (Whitley et al., 1986), six-month mortality among acyclovir-treated patients ran from none in those merely disoriented when therapy began to roughly a quarter of those already semicomatose or comatose. Oriented to person only, he sits tonight in the better of those groups, which is exactly what makes the next few hours an argument rather than a formality. The other number in the room is his fluid status — clinically dry after two days of poor intake — sitting next to a drug whose main serious toxicity is a crystalline nephropathy that gets meaningfully worse exactly when a patient is volume-depleted and the infusion runs in too fast.
In the emergency department, before the PCR is back
Start full-dose IV acyclovir now — ten milligrams per kilogram, every eight hours — and don't let anything else in this room become the reason it's delayed. Two separate NIAID Collaborative Antiviral Study Group trials are why we treat presumptively at all, and they're worth keeping straight. The seventy-percent untreated mortality figure comes from the 1977 placebo-controlled vidarabine trial. The 1986 acyclovir study (Whitley et al.) had no placebo arm at all; it put acyclovir against vidarabine, and acyclovir won at twenty-eight percent six-month mortality against fifty-four. Skoldenberg's Swedish trial two years earlier came in lower still, at nineteen percent. Take whichever of those you trust — not one of them is the number for a patient we watch instead of treat. And Whitley's own follow-up analysis found the driver of a good outcome wasn't simply getting the drug, it was getting it before consciousness had deteriorated further than his already has. He seized this morning. Every hour we spend correcting anything else first is an hour his temporal lobe doesn't get back.
I'm not arguing against starting today — I'm arguing against how fast we push it into a patient who's clinically dry. Acyclovir's real toxicity here is a crystalline nephropathy: the drug is poorly soluble in urine, and it precipitates in the renal tubules fastest exactly when urine flow is low and the infusion runs in over less than an hour. He's tented, his mucous membranes are dry, and he hasn't taken in meaningful fluid in two days — running his first dose into him the way we'd run it into someone with normal intake risks acute kidney injury by tomorrow, and once that happens we're the ones cutting his dose or holding doses at the exact point his treatment needs to be uninterrupted.
The urgency argument doesn't actually conflict with mine — a rushed dose that causes AKI costs him more treatment time than the twenty minutes it takes to run a liter of normal saline alongside it.
Then do both at once, not one after the other — start an isotonic fluid bolus in the same hour as the first acyclovir infusion, run the drug over the full hour it's supposed to take regardless, and nobody has to choose. The other piece worth deciding now, before it becomes an argument tomorrow: his PCR won't be back for a day, and if it comes back negative, that result doesn't rule anything out at this stage of illness. CSF HSV PCR sensitivity is very high once a case is a few days established, but a negative result drawn this early in a genuinely evolving picture has never reliably meant “not HSV.” Commit now to a second lumbar puncture and repeat PCR in several days if he isn't clearly better and the first result comes back negative, so nobody is tempted to stop treatment on one early negative test.
Agreed within the hour: an isotonic fluid bolus and the first dose of IV acyclovir started together, the acyclovir infused over a full hour as protection against the drug's own main toxicity. A second lumbar puncture with repeat HSV PCR is planned in several days regardless of what the first PCR shows, if he is not clearly improving by then.
Not agreed, and left open rather than smoothed over: how long to continue empiric acyclovir if the PCR eventually comes back negative twice while he is improving on his own. The pharmacologist wanted a defined stopping point tied to two negative results plus clinical improvement, to avoid unnecessarily prolonged exposure to a nephrotoxic drug. The neurologist was reluctant to commit to a firm number of days in advance, preferring to make that call once two negative results are actually in hand, given how much about his trajectory is still unknown tonight.