Withholding or Withdrawing Dialysis in the ICU: What's Your Policy?

Lafayette, Richard A. MD

doi: 10.1097/01.NEP.0000395399.36210.59
In Practice

Case 1: A 91-year-old retired dentist has mild dementia, severe cardiomyopathy, and progressive congestive heart failure, chronic kidney disease and valvular heart disease. Two days ago, he underwent reoperative coronary artery bypass grafting, and aortic valve replacement. Unfortunately, he had prolonged bypass time and recurrent postoperative episodes of ventricular tachycardia requiring CPR, and he became anuric. He remains unresponsive, requires aggressive ventilatory support for acute respiratory distress syndrome (but the central venous pressure is climbing to 22 mmHg), and has a climbing bilirubin with substantial elevations of his liver enzymes. He is on three pressors and dobutamine for a cardiac index of 1.8 L/min/m2. The potassium is 5.9 mmol/L and the serum bicarbonate is 10 mEq/L, with a lactate level of 8 mmol/L. You are called to provide volume removal and correct the solute abnormalities.

Case 2: A 28-year-old woman had a bone marrow transplant for relapsed acute myelogenous leukemia three weeks ago. She has had some fevers and chills and persistent pancytopenia while on multiple antibiotics. She complains of right-upper quadrant pain and has had more than 10 L of volume retention. One week ago, imaging suggested reversal of flow in her portal veins, likely representing veno-occlusive disease. She was transferred to the intensive care unit (ICU) yesterday and intubated for progressive hypoxemia with large pleural effusions and questionable bilateral infiltrates. The bilirubin has climbed to 8.5 mg/dL, and she is increasingly confused. She has clonus and asterixis. Her creatinine has been rising by 1.0 mg/dL a day for the past four days, and her blood urea nitrogen is 120 mg/dL. The hematologist feels she might still engraft, and you have been called to evaluate dialytic options for her altered mentation and azotemia.

These two cases demonstrate scenarios in which acute kidney injury (AKI) has occurred in patients who are expected to have a poor outcome with a very low likelihood of either a meaningful recovery or short-term survival. What approach to their medical care is reasonable? Is the implementation of dialysis futile? When would it be reasonable to refuse the wishes of the consulting team or the patient's family to provide dialysis?

Back to Top | Article Outline

Predicting Mortality in AKI

Acute kidney injury is an increasingly frequent complication in hospitalized patients, and it is associated with dramatic increases in morbidity, mortality, length of stay, and hospitalization costs.1 The population experiencing AKI, like the wider hospital population, has become older and more complex.1

Outcomes over the past two decades have been reported as stable to slightly improving, but they still are disappointing, with mortality rates of approximately 50% overall. Mortality is even higher among ICU patients requiring renal replacement therapy, with reported rates approaching and sometimes exceeding 70%, depending on the definition used.2

Over the past two decades, efforts have been made to predict the outcome of hospital-acquired AKI based on demographic, laboratory, and clinical features.2-4 Many of these scoring systems identify populations of patients who have predicted mortality rates of 100% or very close to that. Most of these scoring systems identify older age, malignancy, pressor dependence, organ failures (especially respiratory or liver failure), central nervous system changes, and low platelets as dominant factors in predicting mortality.

These scoring systems were developed to help evaluate outcomes of interventions and to adjust for potential risk differences in treatment groups. They were not designed to evaluate individual patient risk. However, individual patients frequently present with high scores predicting mortality, and it becomes a challenge to determine how, or whether, to utilize this information in clinical care.

None of these scoring systems are perfect, and sometimes patients with the very highest score survive. But validation of some of the scoring systems, especially those looking at status at dialysis initiation or evaluating response over the first days, does suggest the ability to identify cohorts of patients who could be assigned a survival probability below 1% to 5%.3,5

Cases 1 and 2 demonstrate situations in which acute kidney injury in the ICU is further complicated (and caused) by profound comorbidities. In the first case, there is irremediable heart disease, compounded by concern for brain, liver, and lung injury, with an additional poor prognostic marker of an elevated lactate level. In terms of the second case, posttransplant veno-occlusive disease complicated by renal failure has a terrible prognosis, especially in the setting of febrile neutropenia and multiorgan failure including the need for ventilator support.

Back to Top | Article Outline

The Nephrologist's Role

The nephrologist seems to be the medical professional best suited to evaluate the risk-to-benefit ratio of renal replacement therapy in the patient with acute kidney injury. Familiarity with dialysis techniques, personal experience, and knowledge of the published data on mortality predictors for AKI in the intensive care unit should equip the nephrologist with the tools to assess the appropriateness of dialysis for an individual patient.

As has been determined in the chronic dialysis population, the nephrologist's or nurse practitioner's response to the question “would you be surprised if this patient died?” is likely to be a powerful predictive instrument.6,7 In settings where intensivists determine and manage renal replacement therapy in the ICU, those physicians may gather these skills and be the primary decision makers.

Policy makers have defended the right of physicians to refuse care in the setting of “futility,” especially when it comes to CPR.8 If an intervention can never improve the outcome of a patient, it is certainly futile.

Some definitions loosen this rule, calling an intervention futile if it will not prevent death in the overwhelming majority of patients, approximating a less-than-2% chance of misclassification.8,9 For example, it has generally been held that dialysis treatment for cirrhotic patients with hepatorenal syndrome and no option of transplantation is futile, and acute dialysis frequently is not offered or is refused these patients and their families.

There are policies regarding the refusal of chronic hemodialysis to patients with end-stage renal disease when such treatment is considered futile. The Renal Physicians Association (RPA)'s clinical practice guideline on Shared Decision-Making in the Appropriate Initiation of and Withdrawal from Dialysis has been reviewed and recently updated, and it contains suggestions on how to interact with patients and their advocates in sharing decision-making when possible.10

While these guidelines also discuss patients with acute kidney injury, an exact definition of when care is futile is not offered but is left to the nephrologist (and to the patient and medical team).

These guidelines do help outline an approach to accomplish shared decision-making and limit confusion, friction, and potential legal/professional complications. There certainly are times when the nephrologist must present dialysis as an undesirable or futile option.

Back to Top | Article Outline

Team Approach

Carrying out a decision not to offer or to withdraw dialysis clearly is the most difficult part of limiting potentially futile care.

It has been suggested that physicians should not act alone in determining futility and a second expert should be recruited to assure there is no bias in making the determination. Bias has been found in studies examining patient perception of their choices, suggesting greater physician willingness to discuss limiting care in the elderly and in women.11

Ideally, clear and informative dialogue with the patient or his decision-maker, as well as with the primary medical team, is essential in helping decisions to be accepted with minimal risk of controversy, miscommunication, and misunderstanding or backlash.12

When controversy does arise or differences in opinion persist, assistance from an ethics committee or palliative care specialist may be helpful in determining whether adequate information is available to classify dialysis care as futile and whether individuals have the full support of the institution. The RPA guidelines can be very helpful in trying to establish a method of approaching consultants and families regarding withholding dialysis care. It is always best when decisions are shared with all concerned.

In Cases 1 and 2, it would be best to discuss the poor prognosis of the patients with the primary caregiver and elect someone who has a close relationship with the family to discuss the option of comfort care only. Should there exist a prior living will, it should be honored as best possible.

Generally, families will agree with the decisions of the team in regard to end-of-life care. Should there not be consensus to withhold dialysis when the nephrologist believes this is the responsible choice, a short trial of dialysis with withdrawal if meaningful gains are not seen in two to three days is a good tool to assess the chances of changing status.

Back to Top | Article Outline

Cost-Effective Care

On one hand, the patient-physician relationship should be guided by risk-benefit ratios and not by cost-benefit concerns. However, physicians have an obligation to protect society as well and to limit health care costs that are unnecessary.

Often, patients (or their families) share the burden of expenses in ICU care, and it should be reasonable to include costs as a determinant of treatment, especially when chances of a good outcome are very slim (and thus the cost-effectiveness poor). Having said this, true policy on cost-effective care should come from society and its arms (government, administration, insurance, etc.) and likely should not be the key factor in determining the use of resources for an individual.

Back to Top | Article Outline

A Devastating Event

Acute kidney injury in the ICU is a devastating event leading to prolonged and costly stays and high mortality rates. Predictive indices can identify patients at exceedingly high risk for mortality, although rarely are these predictions absolute. Nephrologists should share this data with treatment teams and families early and often to help with decision-making, prepare for outcomes, and, in some cases, limit futile care.

Ideally, the nephrologist's professional and ethical obligations can be met in this fashion. In extreme cases, it may be reasonable to try unilaterally to withhold or withdraw dialysis without the support of the treatment team or family, but one should allow time for second opinions and be prepared for challenges and potential litigation if this occurs.

Back to Top | Article Outline


1. Hoste EA, Schurgers M. Epidemiology of acute kidney injury: how big is the problem? Crit Care Med 2008;36(suppl 4):S146-S151.
2. Chertow GM, Soroko SH, Paganini EP, et al. Mortality after acute renal failure: models for prognostic stratification and risk adjustment. Kidney Int 2006;70:1120-1126.
3. Costa e Silva VT, de Castro I, Liaño F, Muriel A, Rodríguez-Palomares JR, Yu L. Sequential evaluation of prognostic models in the early diagnosis of acute kidney injury in the intensive care unit. Kidney Int 2009;75:982-986.
4. Maccariello E, Valente C, Nogueira L, et al. SAPS 3 scores at the start of renal replacement therapy predict mortality in critically ill patients with acute kidney injury. Kidney Int 2010;77:51-56.
5. Leacche M, Winkelmayer WC, Paul S, et al. Predicting survival in patients requiring renal replacement therapy after cardiac surgery. Ann Thorac Surg 2006;81:1385-1392.
6. Moss AH, Ganjoo J, Sharma S, et al. Utility of the “surprise” question to identify dialysis patients with high mortality. Clin J Am Soc Nephrol 2008;3:1379-1384.
7. Cohen LM, Ruthazer R, Moss AH, Germain MJ. Predicting six-month mortality for patients who are on maintenance hemodialysis. Clin J Am Soc Nephrol 2010;5:72-79.
8. Tomlinson T. Futility beyond CPR: the case of dialysis. HEC Forum 2007;19:33-43.
9. Patel SS, Holley JL. Withholding and withdrawing dialysis in the intensive care unit: benefits derived from consulting the Renal Physicians Association/-American Society of Nephrology clinical practice guideline, Shared Decision-Making in the Appropriate Initiation of and Withdrawal from Dialysis. Clin J Am Soc Nephrol 2008; 3:587-593.
10. Renal Physicians Association. Shared Decision-Making in the Appropriate Initiation of and Withdrawal from Dialysis: Clinical Practice Guideline. 2nd ed. Rockville, MD: Renal Physicians Association; 2010.
11. Johnson MF, Lin M, Mangalik S, Murphy DJ, Kramer AM. Patients' perceptions of physicians' recommendations for comfort care differ by patient age and gender. J Gen Intern Med 2000;15:248-255.
12. Gabbay E, Meyer KB. Indentifying critically ill patients with acute kidney injury for whom renal replacement therapy is inappropriate: an exercise in futility? NDT Plus 2009;2:97-103. •
© 2011 Lippincott Williams & Wilkins, Inc.