PURPOSE: Terminally ill patients often receive burdensome care during their final hospitalization, and for some, the withdrawal of mechanical ventilation (MV) is necessary to permit death. We aimed to identify the characteristics that are associated with terminal MV withdrawal and the factors associated with time to death after MV withdrawal.
METHODS: This is a secondary analysis of data that was originally collected to evaluate the quality of care at the end of life. We retrospectively evaluated 317 patients who died on MV or after MV withdrawal during a 1-year period at single tertiary care center.
RESULTS: Of 317 ventilated deaths, 156 patients had MV withdrawn and 161 patients died while receiving MV. Length of stay and days on MV did not differ between the two groups. Patients in the MV withdrawal group were older (63 v. 58, p=0.009) and less likely to have end-stage disease on admission (45% v. 60%, p=0.006). Patients on surgical services, when compared to medical services (OR 3.4) and neurology/neurosurgical services (OR 11.4), are less likely to have MV withdrawn prior to death. The median time to death was 1.0 hours (IQR 0.2-7.3h, min 0, max 115 hours) after MV withdrawal. In a multivariate regression model, the use of vasopressors was significantly associated with a shorter time to death (by 9.7 hours, p=0.02). Time to death was not associated with ventilator settings, dialysis, age, or presence of end-stage disease; there was a trend toward neurology/neurosurgical patients having a longer time to death (by 8.4 hours, p=0.08).
CONCLUSIONS: Withdrawal of MV to permit death was performed in approximately half of mechanically ventilated patients who were terminally ill. Patients on vasopressors have a shorter time to death after MV withdrawal.
CLINICAL IMPLICATIONS: Targeted interventions, especially for surgical services, may improve the quality of care at the end of life. The time from MV withdrawal to death is varied; however, vasopressor requirement predicted a shorter time to death and families should be counseled accordingly.
DISCLOSURE: The following authors have nothing to disclose: Thanh Huynh, Anne Walling, Thuy Le, Eric Kleerup, Neil Wenger
No Product/Research Disclosure Information