Variations in end-of-life practices in intensive care units worldwide (Ethicus-2): a prospective observational study

Background End-of-life practices vary among intensive care units (ICUs) worldwide. Differences can result in variable use of disproportionate or non-beneficial life-sustaining interventions across diverse world regions. This study investigated global disparities in end-of-life practices. Methods In this prospective, multinational, observational study, consecutive adult ICU patients who died or had a limitation of life-sustaining treatment (withholding or withdrawing life-sustaining therapy and active shortening of the dying process) during a 6-month period between Sept 1, 2015, and Sept 30, 2016, were recruited from 199 ICUs in 36 countries. The primary outcome was the end-of-life practice as defined by the end-of-life categories: withholding or withdrawing life-sustaining therapy, active shortening of the dying process, or failed cardiopulmonary resuscitation (CPR). Patients with brain death were included in a separate predefined end-of-life category. Data collection included patient characteristics, diagnoses, end-of-life decisions and their timing related to admission and discharge, or death, with comparisons across different regions. Patients were studied until death or 2 months from the first limitation decision. Findings Of 87 951 patients admitted to ICU, 12 850 (14middot6%) were included in the study population. The number of patients categorised into each of the different end-of-life categories were significantly different for each region (p<0middot001). Limitation of life-sustaining treatment occurred in 10 401 patients (11middot8% of 87 951 ICU admissions and 80middot9% of 12 850 in the study population). The most common limitation was withholding life-sustaining treatment (5661 [44middot1%]), followed by withdrawing life-sustaining treatment (4680 [36middot4%]). More treatment withdrawing was observed in Northern Europe (1217 [52middot8%] of 2305) and Australia/New Zealand (247 [45middot7%] of 541) than in Latin America (33 [5middot8%] of 571) and Africa (21 [13middot0%] of 162). Shortening of the dying process was uncommon across all regions (60 [0middot5%]). One in five patients with treatment limitations survived hospitalisation. Death due to failed CPR occurred in 1799 (14%) of the study population, and brain death occurred in 650 (5middot1%). Failure of CPR occurred less frequently in Northern Europe (85 [3middot7%] of 2305), Australia/New Zealand (23 [4middot3%] of 541), and North America (78 [8middot5%] of 918) than in Africa (106 [65middot4%] of 162), Latin America (160 [28middot0%] of 571), and Southern Europe (590 [22middot5%] of 2622). Factors associated with treatment limitations were region, age, and diagnoses (acute and chronic), and country end-of-life legislation. Interpretation Limitation of life-sustaining therapies is common worldwide with regional variability. Withholding treatment is more common than withdrawing treatment. Variations in type, frequency, and timing of end-of-life decisions were observed. Recognising regional differences and the reasons behind these differences might help improve end-of-life care worldwide. Funding None.