How Do You Know When a Dialysis Patient Is Dying?

Recognizing that a dialysis patient is nearing the end of life involves reading a pattern of overlapping signals rather than spotting one definitive sign. Declining lab values, increasing frailty, episodes of confusion, and a growing inability to tolerate dialysis sessions all contribute to the picture. Clinicians also rely on a deceptively simple tool called the “surprise question,” which has been validated across multiple studies as a rough but useful mortality predictor. For families, understanding these signs can open the door to honest conversations about goals of care before a crisis forces the decision.

The Surprise Question and Clinician Intuition

One of the most studied tools for identifying dialysis patients at high risk of dying is not a blood test or a scoring algorithm. It is a single question posed to the treating clinician: “Would I be surprised if this patient died in the next year?” When the answer is “no,” the patient is flagged as high risk. It sounds almost too informal to be useful, but the data supporting it are solid. In one study of dialysis patients followed for 12 months, the mortality rate among those whose clinicians answered “no” was about 29%, compared with roughly 11% in the “yes, I would be surprised” group. The odds of dying within a year were about 3.5 times higher for patients in the “no” group.

A separate study in peritoneal dialysis patients found a similar pattern. Mortality was about 25% in the “not surprised” group versus roughly 7% in the group whose clinicians expected them to survive. The negative predictive value was over 93%, meaning that when a clinician said “yes, I’d be surprised if they died,” they were right the vast majority of the time.

The surprise question works because experienced clinicians unconsciously integrate dozens of data points: how a patient looks during treatment, whether they have been losing weight, how often they have been hospitalized recently, whether their conversation has become scattered. The catch is that while the question is good at identifying patients who will probably survive, it is less reliable at catching everyone who will die. One prospective study comparing the surprise question to formal prediction models found that the question’s sensitivity for 12-month mortality was only about 35%, meaning it missed roughly two-thirds of patients who actually died within the year. Statistical models performed better, with sensitivities around 75%.

For families, the takeaway is this: if a nephrologist or dialysis nurse expresses concern that a patient may not have much time left, that clinical impression is worth taking seriously, even if no single test result looks alarming on its own. But the absence of that concern does not guarantee safety. Formal risk tools catch more cases, and asking the care team whether any formal prognostic assessment has been done is a reasonable step.

What the Lab Values Tell You

Families often look to lab reports for reassurance or alarm, and a few markers do reliably track with worsening prognosis. Serum albumin is one of the most consistently studied. Albumin is a protein made by the liver, and low levels in dialysis patients have long been linked to higher mortality. In hemodialysis patients, each one-unit drop in serum albumin (measured in grams per deciliter) was associated with about a 47% increase in the risk of death. Interestingly, much of that risk turned out to be driven by underlying inflammation rather than poor nutrition alone. When researchers adjusted for markers of inflammation, the link between low albumin and death weakened considerably.

⁴

This matters because a falling albumin level does not simply mean a patient is not eating enough. It often signals that the body is in a state of chronic inflammation, fighting infections or dealing with heart failure or other organ stress. When you see albumin trending downward over several months despite adequate nutrition, it suggests the body’s reserves are being overwhelmed. In combination with other signs of wasting, such as unintentional weight loss, muscle shrinkage, and fatigue, declining albumin points toward a condition called protein-energy wasting, which becomes common in advanced kidney disease and is closely linked to cardiovascular death.

⁵

Other lab trends families and clinicians watch include rising phosphorus that becomes harder to control, persistently elevated potassium between treatments, and worsening anemia despite medication adjustments. None of these is a death sentence on its own. But when multiple lab values start drifting in the wrong direction simultaneously and resist the usual interventions, the trajectory is telling you something that individual numbers cannot.

Delirium and Cognitive Decline

Changes in a patient’s mental state are among the most distressing signs for families, and they carry real prognostic weight. Delirium, a state of acute confusion marked by fluctuating awareness, disorientation, and sometimes agitation or hallucinations, is a strong predictor of death in hemodialysis patients. One long-term observational study found that 80% of dialysis patients who developed delirium died within a year of starting hemodialysis, compared with 22% of those who did not experience delirium.

⁶

The association held up even after accounting for other risk factors. Delirium was independently linked to death from cardiovascular disease and from infection, suggesting that it serves as a marker of overall physiological fragility rather than just a brain problem. The infection-related mortality risk was particularly striking, with more than a threefold increase among patients who had experienced delirium.

⁶

More gradual cognitive decline, including increasing forgetfulness, difficulty following conversations, and loss of interest in activities, can also signal that the overall disease burden is advancing. Dementia itself is an independent predictor of early death in dialysis patients, with one prognostic model assigning it a hazard ratio of about 2.2, meaning roughly double the mortality risk compared with patients who are cognitively intact.

⁷

For families, a new episode of confusion during or after dialysis is worth reporting immediately to the care team. It could be caused by something treatable, like an infection or a medication interaction. But when confusion becomes recurrent or progressive, it often means the body is struggling at a systemic level, and it should prompt a conversation about long-term goals.

When Dialysis Sessions Get Harder to Tolerate

Dialysis works by removing fluid and waste products from the blood, and that process puts real stress on the cardiovascular system. As patients become sicker, their bodies lose the ability to compensate. One common signal is recurrent intradialytic hypotension, meaning blood pressure drops sharply during treatment. Patients may feel dizzy, nauseated, or pass out. Frequent episodes of this kind force the dialysis team to shorten sessions or reduce the amount of fluid removed, which means toxins and fluid build up between treatments.

Research has shown that high ultrafiltration rates, the speed at which fluid is pulled from the body during dialysis, can drive structural changes in the heart over time. A five-year study found that patients who experienced repeated drops in blood pressure during dialysis developed measurably worse heart function, including a larger, weaker left ventricle and increased heart mass.

⁸

This creates a vicious cycle: the heart weakens, which makes it harder to tolerate dialysis, which leads to worse fluid control, which further damages the heart. When a patient who previously handled four-hour sessions without trouble starts crashing during every treatment, it can be a sign that their cardiovascular reserve is nearly gone. Families notice this as increased exhaustion after sessions, longer recovery times, and sometimes a refusal to go to dialysis at all.

Cardiac causes dominate mortality statistics in dialysis patients generally. Sudden cardiac death and other cardiac events account for a large share of deaths, with sudden death frequently occurring from cardiac arrest or arrhythmia.

⁹

Formal Prognostic Risk Scores

Beyond clinical intuition, researchers have developed scoring systems that combine measurable risk factors into a single estimate. One model designed for older adults starting dialysis uses a 19-point scale that accounts for age, kidney function at the time dialysis begins, heart failure, atrial fibrillation, recent hospitalizations, and certain cancers. Patients scoring below 5 had less than a 25% chance of dying within six months, while those scoring above 12 had more than a 50% chance.

¹⁰

A different model focused on patients already on maintenance hemodialysis identified five independent predictors of death within six months: older age, dementia, peripheral vascular disease, low serum albumin, and a “no” answer to the surprise question. When those factors were combined, the model’s accuracy was strong, with an area under the curve of 0.87 in the group it was built on and 0.80 when tested on new patients.

⁷

These scores are not crystal balls, and they are rarely presented to families as a single number. But they shape the conversations that happen behind the scenes. When a patient’s risk score climbs, the nephrology team may begin discussing whether continued aggressive treatment is aligned with what the patient actually wants. Families who are aware that these tools exist can ask directly: “Based on what you’re seeing, what does the next six months look like?”

Deciding to Stop Dialysis

About a quarter of dialysis deaths in North America happen after the patient or their family decides to stop treatment. This is not considered a radical choice by the medical community. Broad ethical and legal consensus supports discontinuation when dialysis is no longer achieving a meaningful goal for the patient.

¹¹

Certain patient characteristics are more commonly associated with the decision to withdraw. Older patients, those who have been on dialysis for a long time, those living alone, and those with severe pain or multiple other chronic illnesses are more likely to choose to stop. Cognitive status plays a major role: estimates of how many patients still have full decision-making capacity at the time of withdrawal range widely, from roughly 37% to 80%, suggesting that in many cases, cognitive decline itself is a driving factor. Ethnic differences also exist, with African American and Hispanic patients being less likely to discontinue treatment than white patients.

¹¹

Nephrologists themselves are generally supportive of honoring these decisions. Surveys indicate that the vast majority of North American nephrologists are willing to comply with a patient’s request to stop dialysis, even when they personally might prefer to continue. This reflects a professional culture that, over time, has moved away from viewing dialysis withdrawal as a form of failure or self-harm and toward understanding it as a legitimate end-of-life choice.

What Happens After Dialysis Stops

One of the most common questions families ask is: “How long will they live once we stop?” The answer varies, but the timeframe is typically short. In one early study, patients who discontinued hemodialysis after an average of about 3.5 years of treatment survived a mean of roughly 10 days afterward.

¹²

That average conceals real variation. Patients with some residual kidney function, meaning their kidneys still produce a small amount of urine, may survive somewhat longer. Those who are already very debilitated may die within days. During this time, the body gradually accumulates fluid and toxins. Patients often become increasingly drowsy and eventually slip into a coma-like state. The process is generally described by palliative care specialists as relatively peaceful when properly managed with symptom-focused medications.

Families should expect that the medical focus shifts entirely to comfort during this period. Pain management, anti-nausea medication, managing restlessness or agitation, and keeping the patient’s mouth moist become the priorities. Hospice teams, when involved, are experienced with this trajectory and can guide families through what to expect hour by hour.

When Palliative Care Enters the Picture

One of the persistent problems in dialysis care is that palliative and hospice services are brought in too late. An Irish study examining end-stage kidney disease patients found that among those who were referred to specialist palliative care, the median time between referral and death was just 12 days.

¹³

Twelve days is barely enough time to establish a relationship with a new care team, let alone address the complex emotional and practical needs of a dying patient and their family. Part of the problem is structural: in some healthcare systems, hospice enrollment requires that a patient stop curative or life-prolonging treatment, which for dialysis patients means stopping dialysis itself. This creates an all-or-nothing choice that many patients and families are not ready to make until the very end.

Programs that integrate palliative care principles earlier in the dialysis journey have shown promise. One initiative that introduced structured end-of-life care planning discussions for dialysis patients found that among enrolled participants who died during the study period, nearly half had stopped dialysis before death and over 40% received hospice services. A large majority of participants had also taken practical steps like designating a healthcare proxy.

¹⁴

The evidence points in one direction: earlier conversations lead to better-prepared families and more comfortable deaths. Waiting until someone is actively dying to discuss goals of care robs everyone, patient, family, and clinical team, of the chance to plan meaningfully.

What Families Can Watch For at Home

Not every sign of decline happens in the dialysis unit or shows up on a lab report. Families are often the first to notice changes that signal a shift toward the end of life. Some of the patterns that matter most are gradual and easy to dismiss individually but become unmistakable in combination:

Appetite loss: A slow but persistent disinterest in food, especially protein-rich foods, that does not respond to changes in preparation or timing. This often reflects the protein-energy wasting process described earlier and can precede measurable drops in albumin.
Increasing sleep: The patient sleeps more during the day, naps between dialysis sessions grow longer, and it becomes harder to rouse them for activities they previously enjoyed.
Social withdrawal: Less interest in visitors, phone calls, or conversation. This can overlap with depression, which is common in dialysis patients, but when it occurs alongside physical decline it carries a different significance.
Functional loss: Tasks the patient used to manage independently, like bathing, dressing, or walking to the car for dialysis appointments, now require help. Nephrologists consistently rank declining functional status among the most important factors in their own prognostic assessments.
Recurrent infections: Frequent hospitalizations for pneumonia, access-site infections, or urinary tract infections suggest the immune system is faltering.

Any one of these alone could be a temporary setback. When three or four appear together and persist over weeks, they suggest the patient’s overall trajectory has shifted. Families who notice this pattern should feel empowered to initiate a conversation with the nephrology team rather than waiting for the team to bring it up. Asking “Are we at a point where we should be talking about what comes next?” is never premature. In a disease where the evidence consistently shows that planning conversations happen too late, the family that asks early is doing everyone a favor.

Sudden Death Without Warning Signs

It would be misleading to suggest that every dialysis death follows a recognizable decline. A significant proportion of dialysis patients die suddenly, often from cardiac arrest or arrhythmia, without an extended period of observable deterioration. These events tend to cluster around certain high-risk windows: the long gap between the last dialysis session of the week and the first session of the next week, when fluid and potassium levels peak, is a well-known danger zone.

⁹

For families, this means that even when a patient seems stable, dialysis itself carries ongoing cardiac risk. Advance care planning is valuable not just for patients who are visibly declining but for anyone on long-term dialysis. Having a healthcare proxy designated, treatment preferences documented, and a general understanding of what the patient would want in an emergency removes the burden of making agonizing decisions under crisis conditions. The research consistently shows that patients who engage in these conversations earlier, while they still have the cognitive capacity to participate, end up with care that better matches their wishes.

1
PubMed Central. Utility of the “surprise” question to identify dialysis patients with high mortality
2
PubMed Central. Predicting 12-month mortality for peritoneal dialysis patients using the “surprise” question
3
PubMed. Performance of the Surprise Question Compared to Prediction Models in Hemodialysis Patients: A Prospective Study
4
PubMed Central. Association between serum albumin and mortality in dialysis patients is partly explained by inflammation, and not by malnutrition
5
PubMed Central. Protein-energy wasting and mortality in chronic kidney disease
6
PubMed Central. Delirium in hemodialysis predicts mortality: a single-center, long-term observational study
7
PubMed Central. Predicting six-month mortality for patients who are on maintenance hemodialysis
8
PubMed Central. High ultrafiltration rate induced intradialytic hypotension is a predictor for cardiac remodeling: a 5-year cohort study
9
Kidney International. Distribution and causes of sudden and cardiac death in hemodialysis patients
10
PubMed. A Clinical Risk Prediction Tool for 6-Month Mortality After Dialysis Initiation Among Older Adults
11
Palliative Care Network of Wisconsin. Withdrawal of Dialysis: Decision Making
12
PubMed. Dialysis discontinuation. A ‘good’ death?
13
PubMed Central. The use of palliative care services amongst end-stage kidney disease patients in an Irish tertiary referral centre
14
Patient-Centered Outcomes Research Institute (PCORI). Effect of End-of-Life Care Planning Discussions on Hospice Use Among Patients Receiving Dialysis for Kidney Failure