Morphine for End-of-Life Dyspnea: Underuse in the Emergency Department Versus Inpatient Care

Dyspnea is a common and distressing symptom at the end of life, particularly among patients with advanced cancer, heart failure, or chronic lung disease. Although morphine is well established for relieving dyspnea and supported by clinical guidelines, it remains underutilized in the Emergency Department (ED). This study compared morphine administration for dyspnea at the end of life within the same cohort of patients across two settings, the ED and subsequent inpatient care, and examined the clinical determinants of administration in the ED.

A retrospective cohort analysis was conducted on adult patients presenting with dyspnea at the end of life admitted to the ED at Shaare Zedek Medical Center, Israel, between 2023 and 2024, who died during the same hospitalization. Clinical characteristics, morphine administration, and survival outcomes were analyzed.

A total of 250 patients were included (mean age 83.2 ± 11.5 years; 51.2% female). Only 15.2% received morphine in the ED, whereas 68% of the same patients received morphine during subsequent hospitalization; the mean time to administration among ED recipients was 5.8 hours. Multivariate logistic regression identified lung cancer (OR = 3.41, p = 0.022), DNR status (OR = 3.17, p = 0.010), and pulse rate on arrival (OR = 1.02, p = 0.008) as significant predictors of ED administration. Morphine administration was not associated with shortened survival: no difference in time to death was observed for ED morphine use (p = 0.165). Observed survival was longer among patients who received morphine during hospitalization (282.20 vs 152.35 hours, p < 0.001); this comparison is subject to immortal time bias, as morphine could be administered at any point after admission, and is not interpreted as a treatment effect. This direction was also preserved in a multivariable Cox model adjusting for respiratory support, age, pulse rate, and blood pressure (HR = 1.51, 95% CI: 1.14-2.02), though that estimate carries the same immortal-time-bias caveat.

The same patients who did not receive morphine in the ED frequently received it within hours of admission, indicating that underutilization reflects the setting and structure of emergency care rather than patient characteristics or symptom severity. Lung cancer and DNR status, rather than symptom burden, predicted administration. Morphine administered in the emergency department, where timing is fixed and not subject to immortal time bias, showed no association with time to death, consistent with prior observational literature.