Visualizing How to Use Antipsychotics for Agitated Delirium in the Last Days of Life

How physicians use antipsychotics for agitated delirium in the last days of life varies markedly, which could hamper the quality of care.

To examine adherence to an algorithm-based treatment for terminal agitated delirium, and explore its effectiveness and safety.

A single-center, prospective, observational study was conducted in a 27-bed palliative care unit in Japan. All adult cancer patients who developed agitated delirium with a modified Richmond Agitation-Sedation Scale (RASS) of +1 or more were included; the palliative care specialists determined that the etiology was irreversible, the estimated survival was three weeks or less, and the Eastern Cooperative Oncology Group (ECOG) performance status was three or four. Patients were treated with an algorithm to visualize how to use antipsychotics, with the treatment goal defined as no agitation (RASS≤0) or acceptable agitation for patients and families. We provided all patients nonpharmacological management to alleviate the symptoms of delirium and administered antipsychotic medications when the nonpharmacological approach was insufficient. We measured the adherence rate, RASS, Nursing Delirium Screening Scale items 2, 3, 4 (Nu-DESC), and Agitation Distress Scale item 2 (ADS) on days 0, 1, 3, 7, 14, 21, and 24 hours before death.

A total of 164 patients were enrolled. Adherence rates were 99, 94, and 89%, and treatment goals were achieved in 66, 83, and 93% on days one, three, and seven, respectively. The mean RASS decreased from +1.41 to -0.84 on day three; Nu-DESC decreased from 4.19 to 1.83, and ADS decreased from 1.54 to 0.38. There were seven severe adverse events (Common Terminology Criteria for Adverse Events (CTCAE) of 3), including aspiration (n = 3), apnea (n = 2), tremor (n = 1), and muscle rigidity (n = 1) on day three.

The algorithm-based treatment could be feasible, effective, and safe. Visualizing how palliative care specialists provide pharmacological management could be beneficial for nonspecialist clinicians, and clinical, educational, and research implications warrant further empirical testing.