The Frequency of Withdrawal from Acute Care Is Impacted by Severe Acute Renal Failure

Abstract

Introduction: In the general intensive care setting, decisions to withdraw life support when patients deteriorate despite aggressive treatment are estimated to occur in 10% of all patients and in 40% of the patients who die. Acute renal failure (ARF) severe enough to necessitate renal replacement therapy (RRT) is associated with in-hospital mortality approximating 50%. Yet the impact of severe ARF on decisions to withdraw treatment has not been previously described. In chronic renal failure patients, voluntary withdrawal from maintenance dialysis occurs in 10%–20% of patients when increasing complications and poor quality of life ensue, and knowing these data facilitates discussions with patients and families. Having similar data for complicated ARF would facilitate decision making for families and caregivers when these difficult situations arise. Methods: All cases of ARF requiring RRT during 2000–2001 at University of Michigan Hospital (n = 383) were entered prospectively into an outcome study at the time RRT was initiated. Comprehensive data collection included demographic and clinical characteristics, outcome and complications, and severity of illness. Additional information for patients who died included cause of death, life-support withdrawal decisions, and the presence of prior advance directives. Results: Overall mortality in severe ARF (i.e., severe enough to require RRT) was 53%. Lifesupport withdrawal occurred in 72% of deaths (compared to 40%–50% reported among general intensive care cases) and was associated with intensive care stay well beyond 2 weeks. Severity of illness, as indicated by modified APACHE III scores, was higher in patients who died than in survivors, but severity of illness was not higher for withdrawal from treatment than death without withdrawal decisions. Life-support withdrawal was not associated with other demographic or clinical characteristics (hospital service, primary admitting diagnosis, ventilator or pressor dependence, sepsis, or initial type of RRT chosen). Prior advance directives were available in 29% of patients overall, but having advance directives did not predict withdrawal from acute treatment. Death occurred within 2 days of withdrawal in more than 90% of cases, emphasizing the severity of underlying illness. Conclusions: Severe ARF reflects the severity of underlying illness, impacts overall survival, and is associated with more frequent withdrawal from aggressive treatment. High severity of illness and prolonged intensive care without improvement beyond 2 weeks presage decisions to withdraw treatment and signal patients and caregivers that death is imminent and that further aggressive care should be reconsidered or limited.