Global variability in withholding and withdrawal of life-sustaining treatment in the intensive care unit: a systematic review

Prior studies identified high variability in prevalence of withdrawal of life-sustaining treatment in the ICU. Variability in end-of-life decision-making has been reported at many levels: between countries, ICUs, and individual intensivists. We performed a systematic review examining regional, national, inter-hospital, and inter-physician variability in withdrawal of life-sustaining treatment in the ICU. Using a predefined search strategy, we queried three electronic databases for peer-reviewed articles addressing withdrawal of life-sustaining treatment in adult patients in the ICU. Data were analyzed for variability in prevalence of withdrawal of life-sustaining treatment. Withholding of life-sustaining treatment was also examined where information was provided. An assessment tool was developed to quantify the risk of bias in the included articles. We identified 1284 studies, with 56 included after review. Most studies had unclear or high risk of bias, primarily due to unclear case definitions or potential confounding. The mean prevalence of withdrawal of life-sustaining treatment for patients who died varied from 0 to 84.1 % between studies, with standard deviation of 23.7 %. Sensitivity analysis of general ICU patients yielded similar results. Withholding also varied between 5.3 and 67.3 % (mean 27.3, SD 18.5 %). Substantial variability was found between world regions, countries, individual ICUs within a country, and individual intensivists within one ICU. We identified substantial variability in the withdrawal of life-sustaining treatment across world regions and countries. Similar variability existed between ICUs within countries and even between providers within the same ICU. Further study is necessary, and could lead to interventions to improve end-of-life care in the ICU.