Critically ill patients with COVID-19 in New York City.

With more than 4·2 million people infected and over 288 000 deaths as of May 12, 2020, COVID-19 is threatening the entire world, with dramatic consequences for global health and the economy. After the initial outbreaks in China and Italy, the virus rapidly spread worldwide and today the USA is one of the most affected country, with more than 1·3 million cases.

In the absence of effective containment measures, COVID-19 outbreaks are characterised by an overwhelming number of patients requiring hospital admission, which could exceed the capacity even of well organised health-care systems.

Current knowledge of the clinical course of severe forms of COVID-19 is mainly based on retrospective analyses of Chinese, Italian, and US case series.3, 4, 5, 6, 7, 8, 9 These studies show that critically ill patients with COVID-19 are mainly older, mostly men, and have at least one comorbidity (primarily hypertension), and mortality is very high (ranging from 56% to 97%) when invasive mechanical ventilation is required. However, limited data are available on factors independently associated with mortality. In The Lancet, Matthew Cummings and colleagues add to our knowledge by reporting the results of a prospective cohort study describing clinical characteristics, management, and outcomes of 257 critically ill patients (86 [33%] women and 171 [67%] men, median age 62 years) admitted to high-dependency and intensive care units in two hospitals in New York City over a period of 4 weeks. Notably, critically ill patients with COVID-19 represented 22% of all patients admitted.

Compared with other reports, in the cohort described by Cummings and colleagues the proportion of patients younger than 50 years of age was relatively high (55 [21%] of 257 patients). This could be explained by the high incidence of obesity, which affected 39 (71%) of those 55 younger patients. Other common comorbidities were hypertension (162 [63%]), diabetes (92 [36%]), chronic cardiovascular disease (49 [19%]), chronic kidney disease (37 [14%]), and chronic pulmonary disease (24 [9%]). 203 (79%) patients required invasive mechanical ventilation, median respiratory system compliance was low (27 mL/cm water), and all patients required a high fraction of inspired oxygen despite having relatively high levels of positive end-expiratory pressure (median 15 cm water). Although obesity might have affected the mechanical properties of the respiratory system, these findings underline that the optimal management of mechanical ventilation in patients with COVID-19 and acute respiratory failure remains poorly understood.

The study by Cummings and colleagues confirms that COVID-19 is characterised by a high incidence of multiple organ dysfunction, as shown by the proportion of patients requiring vasopressors (170 [66%]) and renal replacement therapy (79 [31%]). Regarding pharmacological treatments, antibacterial agents were administered empirically to nearly all critically ill patients (229 [89%]) and hydroxychloroquine was administered to 185 (72%), while corticosteroids and interleukin-6 (IL-6) receptor antagonists were administered to fewer patients (68 [26%] received corticosteroids and 44 [17%] received IL-6 receptor antagonists). No data are available on the temporal changes of inflammatory markers in patients receiving immunomodulating treatments. Furthermore, no information is provided about strategies of anticoagulant therapies, which are particularly interesting given the high incidence of thromboembolic complications associated with COVID-19.

With regards to patient outcome, the study conveys important messages. In particular, it shows that the disease is characterised by a high mortality (101 [39%] after a minimum follow-up of 28 days) and prolonged clinical course, as shown by the high percentage of patients still in the hospital (94 [37%]) at the end of follow-up. The multivariable Cox model analysis showed that history of chronic pulmonary disease had the highest adjusted hazard ratio (aHR) for mortality (aHR 2·94 [95% CI 1·48–5·84]). Other independent predictors of death were history of chronic cardiovascular disease (aHR 1·76 [95% CI 1·08–2·86]), older age (aHR 1·31 [95% CI 1·09–1·57]), higher concentrations of IL-6 (aHR 1·11 [95% CI 1·02–1·20] per decile increase), and higher concentrations of D-dimer (aHR 1·10 [95% CI 1·01–1·19] per decile increase) on admission. There appear to be no differences by sex. The association of mortality with higher concentrations of IL-6 and D-dimer is particularly relevant for two reasons. First, it confirms the key pathogenic role played by the activation of systemic inflammation and endothelial-vascular damage in the development of organ dysfunction. Second, it provides the rationale for the design of clinical trials for measuring the efficacy of treatment with immunomodulating and anticoagulant drugs.

The study by Cummings and colleagues shows that clinicians can produce high-quality research even when facing an overwhelming clinical workload. However, despite providing important insights, this work leaves us with some unanswered questions. While waiting for the availability of a COVID-19 vaccine, further studies are required to improve and personalise patient treatment, with particular attention to the role of initial non-invasive respiratory support strategies, timing of intubation, optimal setting of mechanical ventilation, and efficacy and safety of immunomodulating agents and anticoagulation strategies.

This online publication has been corrected. The corrected version first appeared at thelancet.com on June 4, 2020