Eating, Drinking and Tube Feeding at the End of Life

Eating and drinking at the end of life: the hardest conversation, and what the evidence actually says

Of everything that happens as somebody dies, this is the one families cannot bear.

Food is care. Feeding somebody is the most basic act of love there is, and it is the thing a parent has done for their child every day for forty years. When a dying person stops eating, the instinct to feed them is overwhelming, and the suggestion that feeding might not help can land as a suggestion that you stop caring — or worse, that somebody is being starved.

So this guide has to do two things at once. It has to set out what the evidence actually shows, which is uncomfortable. And it has to draw a distinction that most writing in this area fails to draw, and which matters enormously for people with disability.

The distinction is this: the evidence about tube feeding at the end of life is about dying. It is not about disability.

A gastrostomy placed for a seven-year-old with cerebral palsy and unsafe swallowing is one of the great quality-of-life interventions in paediatric disability medicine. It allows nutrition, growth, medication and hydration, and people live well with one for decades. Nothing in this guide argues against that, and any family reading it and concluding that their child’s PEG was a mistake would be drawing exactly the wrong conclusion.

The question is never “is tube feeding good or bad.” It is: what is this intervention for, in this person, at this point in their life? The answer changes, and the point at which it changes is what this guide is about.

We support NDIS participants across New South Wales, including people who are tube fed. We are not a clinical service and nothing here is medical advice. Decisions about nutrition and hydration at the end of life belong to the person, their substitute decision-maker where they have one, and their treating team — never to a provider.

What happens naturally as somebody dies

Worth establishing first, because it reframes everything that follows.

Loss of appetite and reduced thirst are part of the dying process, not a cause of it. As the body’s systems slow, it stops being able to use nutrition, and the drive to eat and drink reduces accordingly. This is not the person giving up and it is not a failure of care.

It also means the experience is different from what a well person imagines. A healthy person who goes without food and water for days is ravenous and desperately thirsty, and the natural assumption is that a dying person must feel the same. The evidence does not support that. People at the end of life commonly report little hunger, and thirst at that stage is largely experienced as dry mouth — which is a symptom that responds extremely well to mouth care and not at all to intravenous fluids.

Which produces one of the most useful facts in palliative care: the treatment for thirst at the end of life is usually moisture in the mouth, not fluid in a vein. Intravenous fluids do not reliably relieve thirst. A sponge of water every hour does.

The evidence on tube feeding in advanced dementia

The clearest evidence base concerns advanced dementia, and it is consistent enough that it has become a formal recommendation.

The findings, drawn from reviews of artificial nutrition and hydration in advanced dementia:

Survival. There is no evidence that tube feeding prolongs life in advanced dementia.

Nutritional status. Tube feeding does not improve it. In advanced disease, enteral feeding may actually worsen the perception of hunger.

Thirst. Intravenous fluids do not reliably relieve thirst.

Aspiration — and this is the one that surprises people most. Aspiration risk increases with tube feeding. Clinicians sometimes cite aspiration pneumonia as a reason to insert a tube, which is a reversal of the evidence. A tube does not stop somebody aspirating their own saliva and refluxed feed, and in some ways makes it more likely.

Pressure injuries. Also counterintuitive. Pressure ulcer risk increases after tube feeding starts, contrary to the common assumption that improved nutrition aids wound healing. One proposed mechanism is increased stool output, often diarrhoea, creating a moist environment that promotes skin breakdown.

The documented harms include pain, infection and bleeding from tube placement; increased aspiration risk; pressure ulcers; diarrhoea, constipation and reflux; fluid overload producing oedema and increased upper airway secretions; possibly increased perception of hunger; and — the one that belongs in a disability publication — the use of physical or chemical restraint to prevent the person removing the tube.

That last item deserves naming plainly. If somebody keeps pulling at a feeding tube and the response is mittens, a restraining device or sedation, that is a restrictive practice. Under the NDIS framework, mechanical restraint and chemical restraint are regulated restrictive practices requiring authorisation, a behaviour support plan and reporting. A tube that can only be maintained by restraining the person is a tube whose benefit needs re-examining.

What is recommended instead. The American Geriatrics Society recommends careful hand feeding, where time and staffing allow. Choosing Wisely puts it directly: do not insert percutaneous feeding tubes in people with advanced dementia; offer oral assisted feeding instead. Small amounts of food and fluid, artificial saliva and good mouth care are recommended for hunger and thirst near the end of life. A limited, case-by-case trial of parenteral hydration may help where there is a reversible cause of delirium, with its own risks of oedema and secretions. And early, preventive discussions, while the person can still express their wishes.

Note what that last point implies. The recommendation is to have this conversation before swallowing problems lead to a crisis, a hospital admission and a decision made at speed by people who do not know the person. For somebody with progressive disease, that conversation should happen years early.

The question this sector actually faces

Most writing on this topic addresses whether to insert a feeding tube. In disability services, that is frequently not the question, because the tube has been there since 1998.

The real question is: this person has had a PEG for twenty years, lived well with it, and is now dying. What do we do about the feeds?

Almost nothing written about end-of-life nutrition addresses this scenario, and it is the common one here.

The relevant considerations:

Continuing full-rate feeds in a dying person can cause harm. As the body shuts down it cannot process the volume. The consequences are the ones listed above — oedema, fluid overload, and increased respiratory secretions, which means more of the noisy breathing that distresses families, and sometimes breathlessness. A person being fed at their usual rate in their final days may be visibly more uncomfortable because of it.

Reducing the rate, or stopping, is a legitimate clinical decision. It is not starvation and it is not withdrawal of care. It is the recognition that an intervention which was helping has stopped helping and started causing harm.

It is also a decision, with a decision-maker. It cannot be made by a provider, a house manager or a worker. It belongs to the person if they have capacity, to their substitute decision-maker if they do not, and to the treating team. It should be documented, with reasons.

And it is frequently done in stages, because that is easier for everybody. Reducing the rate, then reducing further, then stopping — with the person observed at each step — is both clinically reasonable and considerably kinder to a family than a single abrupt decision.

Hydration via the tube may be continued when feed is stopped, or may not, and that is a separate decision with its own considerations. Ask about it specifically rather than assuming the two go together.

What matters most practically: if somebody with a long-standing PEG is entering the final weeks, this needs discussing before the final days, with the palliative care team and the person responsible, and the outcome needs writing down. Otherwise a worker on a Sunday night is left looking at a feed pump and a dying person and nobody has told them what to do.

Dysphagia, and the shift to comfort feeding

The other situation specific to this population.

A great many people we support have dysphagia and a speech pathologist’s recommendations: modified textures, thickened fluids, specific positioning, sometimes nil by mouth. Those recommendations exist to prevent aspiration, and they are correct for most of somebody’s life.

At the end of life the calculation changes, and there is a formal framework for it: eating and drinking with acknowledged risk, sometimes called comfort feeding or risk feeding.

The reasoning is this. The purpose of a nil-by-mouth restriction is to prevent aspiration pneumonia, which could shorten life. In somebody who is dying, that risk is no longer the dominant consideration — and the pleasure of tasting something they love may matter more than avoiding a pneumonia that is no longer the thing that will determine when they die.

So a person who has been nil by mouth for years may, with the speech pathologist and the palliative care team involved, be supported to have small amounts of something they enjoy, with the risk explicitly acknowledged and documented.

This is a real and legitimate approach and it must be done properly:

• Assessed and agreed by the speech pathologist and the treating team, not improvised
• The risk explicitly documented, along with the reasoning and who agreed to it
• The person’s own wishes central, where they can express them
• The substitute decision-maker involved where they cannot
• Clear written guidance for workers: what, how much, what consistency, in what position, who may offer it
• Family informed, so nobody panics when they see it

What it must not be: a worker deciding on their own that it would be nice to give somebody a biscuit. That is not comfort feeding; it is an unassessed risk taken on somebody else’s behalf. The difference is documentation and agreement.

And it goes both ways. Sometimes the right answer at the end of life is still nil by mouth, because the person chokes distressingly and the experience is frightening rather than pleasant. The framework is about deciding deliberately, not about always permitting.

What to offer, and what not to push

For somebody in the final weeks who is still taking something by mouth:

Offer, do not push. The most important sentence here. Offering food and declining to force it is good care. Pushing food on somebody whose body cannot use it causes nausea, distress and sometimes aspiration, and it damages the relationship at a point when the relationship is most of what is left.

Small amounts, frequently, of whatever they actually want. Nutritional balance has stopped being the goal. If the only thing they want is ice cream, the answer is ice cream.

Taste without volume. A teaspoon of something strongly flavoured, a sip of beer, a lick of a lemon, the smell of coffee. Pleasure does not require quantity, and this is frequently what people most appreciate.

Ice chips, where safe, which address dry mouth and give something to do.

Temperature and texture matter more than they did — cold things are often better tolerated.

Position properly, and never offer anything to somebody who is drowsy or not alert enough to manage it.

Mouth care instead of food, when food is no longer possible. This is the substitution that matters and it is covered in the symptom management guide here. Hourly moisture and lip care does more for comfort than any feed.

Let the family do it. For a family who have fed this person for decades, being given the sponge and shown how to do mouth care gives them something real to do with their hands at a time when they feel useless. That is not a small thing.

Talking to families about it

This is the conversation that goes wrong most, and a few things help.

Lead with what dying does to appetite, not with a decision. “As people come to the end of their life, their body stops being able to use food, and the hunger goes away. It is part of what is happening, not a cause of it.” That sentence does a lot of work.

Name the fear directly. Most families are afraid of one specific thing: that the person will starve, and that it will hurt. Say out loud that this is what people worry about, and let the clinical team answer it. Unspoken, it festers into a conviction that somebody is being neglected.

Separate “not feeding” from “not caring”. Offer the alternatives explicitly — mouth care, taste, presence, being there. Give the family a job.

Do not promise comfort you cannot guarantee, and do not pretend certainty the evidence does not support. Families can tell.

Give it time and expect to repeat it. This is not a conversation that happens once. It happens several times, and people need to arrive at it rather than be told.

Involve the palliative care team. This is their expertise and they are better at it than a provider. Ask them to have the conversation, with the family and the support team in the room together.

And accept that some families will want feeding continued, and that this is their decision to make or their relative’s, not yours to overrule. Where a substitute decision-maker wants feeding to continue, the clinical team manages the consequences and the role of a provider is to support the person, not to argue.

Consent and who decides

If the person has capacity, it is their decision. Including the decision to accept risk, and including the decision to stop. Supported decision-making applies: accessible information, time, a trusted person present, teach-back to check understanding.

If they have made an advance care directive, it may address artificial nutrition and hydration specifically — and this is one of the most useful things such a document can cover, precisely because the decision so often arrives when the person can no longer participate. There is a full guide on advance care planning in New South Wales here.

If they lack capacity and have no directive, the substitute decision-maker decides. In New South Wales medical treatment consent for adults who cannot consent is governed by the Guardianship Act 1987, with a person responsible hierarchy — broadly a guardian with a health function, then a spouse or de facto partner, then the unpaid carer, then a close relative or friend with a continuing relationship. A paid carer, including a support worker or provider, is not the person responsible and cannot consent.

Withholding or withdrawing a treatment that is no longer benefiting somebody is legally and ethically distinct from causing death. It is a recognised part of medical practice. Families frequently fear it is not, and that fear deserves a direct answer from the clinical team rather than reassurance from a provider.

Where there is disagreement between family members, or between family and the treating team, NSW Civil and Administrative Tribunal’s Guardianship Division is the forum, and a hospital or community palliative care social worker can frequently resolve it faster than any formal process.

What a support team should do

• Raise it early. If somebody with progressive disease or a long-standing PEG is deteriorating, ask the palliative care team and the GP what the plan is for nutrition and hydration, before the final days.
• Get the decision in writing, including who made it and on what reasoning.
• Make sure every worker knows the plan — including whether feeds are continuing, at what rate, and what to do if the person appears uncomfortable.
• Record intake factually, without judgement, and report decline rather than trying to correct it.
• Report distress during feeding — coughing, wet voice, grimacing, pulling away, breathlessness. That is clinical information.
• Do the mouth care. Hourly or two-hourly. This is the single most valuable thing on this list.
• Do not improvise. No food or fluid outside the documented plan, however kindly meant.
• Do not physically prevent somebody removing a tube without this being properly assessed, authorised and reported as a restrictive practice.
• Support the family rather than taking a position in their decision.