Abstract
Dear Editor:
A 71-year-old woman with stage IV colon cancer with metastases to liver was admitted from home after hours to a 20-bed free-standing inpatient hospice unit for agitation. Her home medication regimen consisted of as needed oral morphine 10 mg and oral lorazepam. There was no clinical response to four doses of 1 mg of oral lorazepam administered hourly. Nursing report indicated the patient was very confused, picking at her hair and clothes, and saying, “Something isn't right.” The on-call physician (fellow) ordered, by phone, midazolam 10 mg parenterally every 4 hours on schedule and every 30 minutes as needed for hyperactive delirium. Since uncontrolled pain can be an etiology for delirium, morphine 10 mg parenterally was ordered every 4 hours on schedule.
At 08:00 h morning report, the night nursing staff reported the patient's behavior responded to treatment and she was resting comfortably. She was arousable at 06:00 h medication time and at 07:15 h was noted to be breathing comfortably. The fellow and attending physician started morning rounds at 08:30 h. The patient was unresponsive, breathing once every 45 seconds, pupils fixed and 2 mm bilaterally, no gag reflex, and irregular heart rhythm at 130/min. A differential diagnosis of iatrogenic sedation and respiratory depression versus natural dying was entertained. Our clinical dilemma was whether to administer reversing agents.
Discussion
Specialty hospice units are not typically staffed with physicians on a 24-hour basis. Nor do physicians come in to see out-of-hours admissions. Patients are admitted based on reports from nursing staff, orders are given verbally over the telephone, and the patient is seen by a physician the next day.
We notice many units such as ours use tacit assumptions about responding to such events. “She has a Do Not Resuscitate Order” or “She's enrolled in the home hospice program” are interpreted to mean that, even if there were an iatrogenic complication, no attempt at resuscitation should be attempted. Standard written advance care directives do not typically address such events. Others opine that if iatrogenic harm is suspected, the patient should be sent to a properly staffed hospital emergency department where standard resuscitation can be attempted. Partial resuscitation attempts with attention to symptom control are not a part of standard emergency department approaches. We are not aware of this being discussed in the standard hospice and palliative medicine literature.
What We Did
We telephoned the daughter, our only contact, and left a message on her voicemail. We noted that the patient had been tolerating these medications overnight and had been responding well. This argued against iatrogenic sedation. Since the patient had been receiving these medication classes at lower doses at home, we inferred pharmacological tolerance would have developed to both opioids and benzodiazepines. However, we were concerned about whether the patient and family would be at peace with her dying under these circumstances. We decided to err on the side of preventing iatrogenic premature death. We held present medication orders, started 8 L/min high flow nasal cannula oxygen and gave naloxone 0.2 mg subcutaneously, by mixing 0.4 mg (1 mL) naloxone with 9 mL normal saline then gave 2.5 mL bilaterally in the deltoid areas. There was no response to the interventions and the patient continued to breathe once every 45 seconds until her death 36 hours later. We did not administer flumazenil.
Within an hour we were able to reach the patient's daughter, who reported that the patient had been expressing she was ready to die for several weeks and her main hope was that she would not die at home. For the past two weeks, she had become increasingly confused. She would repeat herself frequently, was having difficulty accomplishing simple tasks, was not sleeping, and would want to call and talk to people that she had just talked with. In addition, she had complained of blurry vision and foreign body sensation in her left eye, along with an unpleasant tingling on the left side of her face.
We were also able to clarify the patient's cancer history. She was diagnosed three years prior with metastases to liver at diagnosis. She had received chemotherapy, seed radiation to liver and immunotherapy. The family also mentioned that she had a history of atrial fibrillation but had stopped apixaban approximately a month before admission, being told it was no longer indicated in her advanced disease.
In the calm of retrospection, we suspect that the patient's course for the two weeks before admission had been the usual course of advancing cancer. The symptoms also suggest the possibility that she experienced an embolic event from atrial fibrillation with or without hemorrhage.
We wonder what others would have done in a similar circumstance. A hospice physician's goal is never to hasten death, but to treat suffering and preserve dignity, allowing terminally ill patients to achieve self-actualization for peace and transcendence. We frequently do not have access to a complete medical record. We are frequently expected to respond to acute symptom issues, such as in this case. Did we do the right thing? How would a generalizable ethical principle be articulated?