Death Rattle: Treating the Sound or Treating the Patient
An actively dying man's audible secretions are distressing everyone in the room except, as far as anyone can tell, him — and the one randomized trial supporting the drug his family is asking for tested it as prevention in patients who had not yet started rattling — which is not the situation in front of them.
B.L. has been unresponsive for roughly fourteen hours, breathing has slowed and grown irregular, and for the last two hours a coarse, wet, rattling sound has accompanied each breath — pooled upper-airway secretions he can no longer clear or swallow. His three adult children, gathered at the bedside, keep glancing toward the door each time the sound rises, and one of them has asked twice now whether there's a medication that can stop it. By every observable measure the team has, including his stable heart rate and absence of any grimace or withdrawal response, B.L. himself shows no sign the sound is distressing him.
The evidence divides along a line most bedside conversations skip over: whether the drug is given to treat a rattle already present, or before one starts. For treating an established rattle the record is poor — Wildiers found no meaningful difference between atropine, hyoscine butylbromide, and scopolamine, and Heisler found sublingual atropine no better than placebo. Prophylaxis is a different question with a different answer: van Esch and colleagues' SILENCE trial randomized 157 dying hospice patients to subcutaneous scopolamine butylbromide or placebo and found a rattle in 13% of the treated group against 27% on placebo. B.L. sits on the wrong side of that line for the positive result to reach him. His rattle began two hours ago; SILENCE enrolled patients who did not yet have one. The safety comparison turns on a distinction the word "scopolamine" hides. Scopolamine hydrobromide is a tertiary amine that crosses the blood-brain barrier and has been linked to paradoxical agitation and delirium in elderly patients. Scopolamine butylbromide — the SILENCE drug — is a quaternary compound that does not, and neither does glycopyrrolate. Atropine does. So tonight's choice isn't central against peripheral: it's which of two peripherally-restricted agents to reach for, at the one moment none of them has been shown to work.
One of B.L.'s daughters, a hospital pharmacist herself, has already read the same trial abstracts the team is now discussing outside his door, and arrived tonight expecting to be told there was nothing worth trying — which is part of why the conversation about to happen inside the room needs to be more than a drug order. Her professional familiarity with the evidence doesn't make the sound any easier for her siblings to sit with, and it leaves the team with a question the trials were never designed to answer: whether a drug that has not been shown to change what the patient experiences can still be the right thing to give, when the person it would actually settle is standing in the doorway rather than lying in the bed.
Just outside the room, so his family wouldn't overhear the disagreement
I don't think we should start an antimuscarinic. Wildiers and Heisler both studied treating a rattle already established, and neither found a drug that beat its comparator. The one clearly positive randomized result in this space is van Esch's SILENCE trial, and that's prophylaxis — butylbromide started before any rattle appeared, 13% against 27% on placebo. B.L. started rattling two hours ago. We'd be asking the drug to do the thing it has never been shown to do rather than the thing it has. I don't want to treat that as an established intervention just because his family is asking for something to do.
I hear that, and I'm not disputing the efficacy data. But I'd be careful how we say "scopolamine" here, because our cabinet stocks the hydrobromide and the trial you just cited used the butylbromide. Those aren't interchangeable at the blood-brain barrier — the hydrobromide crosses it, and I've personally watched it cause real agitation and delirium in dying patients; the butylbromide and glycopyrrolate don't. So if the answer ends up being a drug, it should be glycopyrrolate, not the bottle upstairs that shares its first word. If it doesn't reduce the sound, we've lost very little.
I want to name something neither of you has said directly: by every sign we can observe, B.L. isn't distressed by this sound. No grimace, no withdrawal, stable heart rate through fourteen hours of decline. The patient actually suffering in that room right now is his family. I don't think the right first intervention is a drug aimed at a sound he isn't reacting to — it's going back in and explaining plainly what the sound is, that it doesn't appear to be causing him pain, and being honest that the medications sometimes offered for it haven't reliably been shown to help. If they still want to try glycopyrrolate after hearing that, I won't stand in the way — but I don't want us to reach for the prescription pad before we've had that conversation.
Agreed: the Palliative Care Physician went back in first, alone, and explained directly to B.L.'s children what the sound was, that their father showed no observable sign of distress from it, and that the one randomized trial showing a real benefit tested prevention in patients who had not yet begun to rattle, not rescue of one already underway. Glycopyrrolate was offered as an option they could still choose, honestly framed as low-risk rather than proven-effective.
Genuinely unresolved, and recorded as such rather than smoothed into a tidy consensus: the family, after hearing all of that, asked for glycopyrrolate anyway, which the team gave. Whether that was the right call — treating family distress with a drug of unproven efficacy once they've been fully informed — was not something the three clinicians agreed on afterward; the Clinical Pharmacologist remained uncomfortable with it even as informed choice, while the other two treated an informed family's request, once honestly counseled, as sufficient grounds on its own.