Terminal agitation and restlessness in the last days

The picture is fairly consistent. Someone who has been settled becomes restless: plucking or picking at the sheets and their own clothes, reaching for things that are not there, trying to get out of bed, calling out or moaning, a face that looks worried or cross. There may be confusion about where they are, brief hallucinations, or a flash of anger or suspicion that is nothing like them. It can come and go across a single day.

The question every family asks is whether the person is suffering, and the honest answer is that it depends, and that we cannot always know. Some people, when they recover from a delirium and can describe it, remember it, and remember it as frightening. In one study of patients who came through it, a little over half recalled the episode, and on average they rated the distress high. So this is not something to wave away as peaceful. But many people at the very end are too deeply unrousable to be reached at all, and for them the distress you are watching may be more in the body's motion than in an experience they are having.

The useful stance is the honest middle. Do not assume they feel nothing, and do not assume the worst either. Watch for the real signs of distress, a face straining, a body working, and treat those. And know that the quiet, still version of this is not proof of comfort. Patients with the withdrawn form of delirium, in that same research, were just as distressed as the agitated ones when they could report it. Calm is not the same as fine.

Here is the section you will not find on most pages, and it may be the most important one. Not everything that looks like agitation is a medical problem to be calmed. Some of it is what hospice workers have long called nearing-death awareness. A dying person starts to talk about a journey, about needing to go home when they are already home, about getting ready, about packing. They speak to a parent or a spouse who died years ago, and they seem comforted by it. They reach toward a corner of the room.

For a long time this was dismissed as confusion, or written up as a symptom to treat. It is not. In a study that followed 59 hospice patients through more than 450 interviews, the large majority described dreams and visions of people who had died, and far more found them comforting than found them frightening, roughly six in ten comforting against about two in ten distressing. The visions of the dead were the most comforting of all, and they grew more frequent as death came closer. The patients in that study were screened to make sure they were not delirious. This is a different thing.

The difference matters because the two call for opposite responses. Delirium is disorganized and it fluctuates: the thinking scatters, attention will not hold, the person is often distressed, and there is frequently a physical cause underneath that can be found. Nearing-death awareness tends to be the reverse, coherent and clear, built out of real people and real relationships, carrying meaning, and leaving the person calmer rather than more upset. One is a symptom to assess and often to treat. The other is not something to medicate away. If you sedate a peaceful goodbye because a page told you that talking to the dead is a warning sign, you have taken something from them that they will not get back.

It is not always a clean line, and sometimes it is both at once. If the person seems frightened, tormented, or unreachable, treat it as distress and call the team. If they seem somewhere else but at peace, you can let them be there. Sitting with the uncertainty is allowed. You do not have to diagnose it to be gentle with it.

Before anyone reaches for a sedative, there is a short list worth running, because a real fraction of this is fixable. Across studies of delirium in advanced illness, roughly half of episodes can be reversed when someone looks for the cause. That number falls the closer a person is to death, and in the last days most of it is not reversible. But roughly half is not nothing, and the quiet tragedy is that the reversible cases are the ones most often missed.

The causes worth checking are ordinary and physical. A full bladder that cannot be emptied is a genuinely common culprit and one of the easiest to check. So is constipation, days of it, backed up and unnoticed. Uncontrolled pain can surface as agitation in someone who can no longer tell you where it hurts. So can breathlessness. And so can the medicines themselves: a steroid started or raised this week, an opioid rotated to a new drug, a dose that has quietly climbed. Infection, dehydration and a high calcium level round out the list. None of these require you to be a clinician. They require someone to ask.

This is exactly where a family gets stuck, and where knowing your own information changes what happens. When the nurse asks what medications changed and when, or whether there has been a bowel movement in the last three days, the answer is often a shrug and a guess, because the person who would know is a sibling three states away. A single current medication list, with the recent changes visible, is the difference between a fixable cause being found and being missed. The opioid rotated on Tuesday, the steroid that went up on Monday: those are the clues, and they only help if someone can see them.

There is one error so common it is almost the default, and it is worth naming plainly. Agitation gets read as pain, and the answer given is more opioid. Sometimes that is right. Often it is not, and it can deepen the very confusion it was meant to relieve.

The evidence for this is pointed. In a small study, nurses rated patients' pain as higher during agitated episodes than the patients themselves had rated it before or after, and those patients received roughly five extra doses of opioid a day compared with about two for everyone else. When they recovered, not one of them remembered any pain during the episode. In a larger study, pain was the stated reason for referral in half of the delirium cases that had been missed entirely, and being referred for pain made a missed delirium more than twice as likely.

The reason this is more than an academic point is that opioids are themselves one of the leading fixable causes of delirium. So the well-meant response, more morphine for a person who looks like they are hurting, can be the thing tipping them further into confusion. None of this means pain should go untreated, or that a dying person in genuine pain should wait. It means that when agitation appears, more opioid is a question, not an automatic answer, and it is fair to ask the nurse whether pain is really what they are seeing, or whether something else is driving it. The plainest version of that question, and a fair one to ask at any hour, is this: are we treating pain here, or are we treating agitation, and how do we know the difference.

When a medicine is given for the agitation itself, it is usually an antipsychotic, most often haloperidol, and it remains the standard first choice. But the evidence for it is more humbling than the confident tone of most pages. In a good randomized trial, patients who got an antipsychotic on top of careful supportive care actually scored slightly worse on their distressing symptoms than patients who got the same care plus a placebo, and they had more side effects. The group that did best was the one where the team treated the causes and the environment and did not add the drug.

That trial has an important limit, and honesty requires stating it. It studied milder delirium in people who could still take a pill, not the severe, thrashing agitation of the final hours, where a medicine to settle someone is sometimes genuinely the kindest thing available. So it does not mean antipsychotics are useless at the very end. It means they are not the magic the word treatment implies, and that looking for a cause should come first.

Then there is the drug worth understanding before it is used, because it cuts both ways. Benzodiazepines, lorazepam and midazolam, are sedatives. Given on their own for delirium, they can make it worse, and one trial of a benzodiazepine alone was stopped early because every patient on it developed limiting side effects and none improved. But in the last days, when someone is agitated and an antipsychotic alone is not holding it, adding a benzodiazepine can genuinely calm them. In a trial of exactly that situation, lorazepam added to haloperidol reduced agitation where haloperidol alone had not, and the families at the bedside saw the person as more comfortable.

The catch, and it should be said out loud, is that the calm comes partly from sedation. The medicine settles the agitation in part by making the person sleepier. That can be exactly the right trade in the final hours, and it is worth knowing that it is the trade being made. A benzodiazepine alone, early, for confusion is usually the wrong move. A benzodiazepine added for agitation that will not settle at the very end is often a mercy. The two are not a contradiction. They are answers to different questions.

The things that help are mostly quiet and unglamorous, and they are things you can do. Lower the stimulation: dim the lights, turn off the television, keep the voices soft, and do not have three people talking at once. Keep familiar things in reach and familiar faces near. Speak slowly and gently, and say who you are, because a person who cannot place the room can often still be steadied by a known voice.

Do not argue with the confusion, and do not try to correct it. If they think it is 1975, or that they need to catch a train, contradicting them tends to agitate them more, and it costs them nothing to be met where they are. Answer the feeling rather than the fact. "You are safe, I am right here" does more than "no, there is no train." Reorientation, the gentle reminding of the day and place, genuinely helps earlier in an illness, but in the final hours it mostly does not, and pushing it can make things worse.

Keep them safe without holding them down. The instinct, when someone keeps trying to get out of bed, is to restrain them, and it is almost always the wrong one. Physical restraints, and even the tug of a catheter or an IV line, tend to feed the very distress they are meant to contain. Lower the bed, pad what is hard, move what is sharp, put a mattress on the floor if you have to, and have someone sit with them. The goal is a soft, safe space to be restless in, not a still body.

And do the small comforting things even when they change nothing about the agitation, because they change something for you. Mouth care, a cool cloth, lotion worked into the hands. When the helplessness is its own kind of pain, having something useful to do with your hands matters more than it should have to.

Sometimes, despite everything, the agitation cannot be settled, and a team will raise the possibility of palliative sedation. This is medicine used deliberately to lower a person's awareness enough to relieve suffering that nothing else has touched. Agitated delirium in the last days is, in fact, the single most common reason it is used.

Families hear it and think one thing, even when they do not say it: are we ending his life. It is the right question to ask out loud, and the answer is no. Palliative sedation is aimed at the suffering, not at the person. The dose is titrated to comfort, kept no deeper than it needs to be, and the intent is relief. The best evidence we have, which comes from watching many patients rather than from the strictest kind of trial, is that it does not shorten life. People sedated for symptoms that will not yield tend to live about as long as those who are not.

There is an old and honest principle underneath this. An action taken to do good, to relieve unbearable suffering, is right even if it carries a risk you did not intend and did not seek. That is not a loophole. It is the difference between a death caused and a death allowed while the dying is made bearable. It is worth adding that the evidence base for sedation is thinner than anyone would like, and that it is not a decision to rush or to make alone. But if it is offered because someone you love is suffering in a way nothing else has reached, it is an act of care, and choosing it is not choosing to let go early.

The arrival of delirium usually does mean that time is shortening. Across studies, people who develop it live meaningfully less long than those who do not. The counterintuitive part, worth holding, is that the quiet, withdrawn form tends to signal a shorter time than the dramatic, agitated one. So the very thing that frightens you most, the thrashing and the calling out, is not necessarily the sign that the end is closest.

Whatever they say in it, it is not their true self speaking, and it is not a verdict on your care. People say things in delirium that are cruel, or frightened, or nothing like the person you know. It is the illness moving their mouth. If they are angry at you, or accuse you, or do not know you, none of that is a message. And if it happened on your watch, you did not cause it. It came from a body that is dying, on a timeline that was never yours to set.

Some people surface. After hours or days of agitation or absence, a person will sometimes turn briefly, startlingly clear, know everyone, say what they needed to say, and then slip back. It is real, families describe it vividly, and it is rare enough that no one can promise it or predict it. If it comes, take it. And when you are unsure whether what you are seeing is normal, or the medicines are not holding, or you simply cannot tell distress from peace, call the hospice number, any hour. It is what it is for. Knowing which questions get you a straight answer helps, and if you are still choosing a hospice, the questions worth asking one cover what the brochures leave out.