Mortality rates of patients with COVID-19 in the intensive care unit: a systematic review of the emerging literature

Abstract

The understanding of outcomes in the intensive care unit (ICU) for the coronavirus disease 2019 (COVID-19) remains poor. Studies have reported close to 100% mortality amongst patients requiring mechanical ventilation [1], and this together with the hypothesis that COVID19 may not cause classic acute respiratory distress syndrome (ARDS) has led to concerns regarding the use of mechanical ventilation [2, 3]. We thus aimed to review the outcomes of ICU patients with COVID-19 from the existing literature. We searched PubMed for studies published between Dec 1, 2019, and May 8, 2020, with at least ten ICU patients with COVID-19 and reported ICU mortality data. We excluded studies that had duplicate patients from other reports, did not provide data on ICU survival, enrolled only decedents, and excluded patients who were still hospitalised (Fig. 1 and Electronic Supplementary Material). Several lessons can be surmised from Table 1, which outlines the 15 included studies conducted largely in countries worst hit by the pandemic. First, 56.1% of patients were still in the ICU at the time of study publication, and attempts to calculate mortality based on a sample of only deceased or discharged patients risk painting a skewed picture of reality [4]. Second, with the prior limitation in mind, the overall ICU mortality rate was 25.7%. In China, with 14.1% of patients still in the ICU, the mortality rate was 37.7%. These figures are not higher than the mortality rates of 35 to 45% seen in ARDS. Third, 29% of the ICU patients who died in the Chinese studies did not receive mechanical ventilation, and where systems experienced a surge of critically ill patients, up to 53.2% of patients who required ICU care were unable to receive it because of resource constraints [5]. In New York, 262 deaths occurred in hospital wards and outside the ICU, compared to 291 deaths in the ICU [4]. We hypothesise that rationing of ventilators and ICU beds in overwhelmed health systems may have resulted in attempts at postponing intubation, with a significant minority of patients received high-flow nasal cannula (13.7%) and noninvasive ventilation (11.3%) based on available data, despite uncertainty surrounding their roles. We conclude that while there is a need for further studies which capture patients’ final dispositions, the current preliminary data does not suggest unusually high ICU mortality rates for COVID-19. The poor outcomes seen in various studies may be related to rationing of resources in overwhelmed ICUs.