key: cord-0755260-io1lf4n6
authors: Qian, Zhicheng; Lu, Shuya; Luo, Xufei; Chen, Yaolong; Liu, Ling
title: Mortality and Clinical Interventions in Critically ill Patient With Coronavirus Disease 2019: A Systematic Review and Meta-Analysis
date: 2021-07-23
journal: Front Med (Lausanne)
DOI: 10.3389/fmed.2021.635560
sha: d4a088bebfd7713a9e094deeca18591930fc18fe
doc_id: 755260
cord_uid: io1lf4n6

Objective: The aims of this systematic review and meta-analysis were to summarize the current existing evidence on the outcome of critically ill patients with COVID-19 as well as to evaluate the effectiveness of clinical interventions. Data Sources: We searched MEDLINE, the Cochrane library, Web of Science, the China Biology Medicine disc, China National Knowledge Infrastructure, and Wanfang Data from their inception to May 15, 2021. The search strings consisted of various search terms related to the concepts of mortality of critically ill patients and clinical interventions. Study Selection: After eliminating duplicates, two reviewers independently screened all titles and abstracts first, and then the full texts of potentially relevant articles were reviewed to identify cohort studies and case series that focus on the mortality of critically ill patients and clinical interventions. Main Outcomes and Measures: The primary outcome was the mortality of critically ill patients with COVID-19. The secondary outcomes included all sorts of supportive care. Results: There were 27 cohort studies and six case series involving 42,219 participants that met our inclusion criteria. All-cause mortality in the intensive care unit (ICU) was 35% and mortality in hospital was 32% in critically ill patients with COVID-19 for the year 2020, with very high between-study heterogeneity (I(2) = 97%; p < 0.01). In a subgroup analysis, the mortality during ICU hospitalization in China was 39%, in Asia—except for China—it was 48%, in Europe it was 34%, in America it was 15%, and in the Middle East it was 39%. Non-surviving patients who had an older age [−8.10, 95% CI (−9.31 to −6.90)], a higher APACHE II score [−4.90, 95% CI (−6.54 to −3.27)], a higher SOFA score [−2.27, 95% CI (−2.95 to −1.59)], and a lower PaO(2)/FiO(2) ratio [34.77, 95% CI (14.68 to 54.85)] than those who survived. Among clinical interventions, invasive mechanical ventilation [risk ratio (RR) 0.49, 95% CI (0.39–0.61)], kidney replacement therapy [RR 0.34, 95% CI (0.26–0.43)], and vasopressor [RR 0.54, 95% CI (0.34–0.88)] were used more in surviving patients. Conclusions: Mortality was high in critically ill patients with COVID-19 based on low-quality evidence and regional difference that existed. The early identification of critical characteristics and the use of support care help to indicate the outcome of critically ill patients.

With the rapid spread of coronavirus disease 2019 (COVID- 19) globally, as of June 2, 2021, a total of 171,222,477 confirmed cases had been reported in 215 countries, areas, or territories, and COVID-19 has been responsiblefor at least 3,686,142 deaths (1) . Critically ill patients are always companied by a high risk of lives, which may be complicated by an uncontrolled systemic inflammatory response leading to acute respiratory distress syndrome (ARDS) and multiple organ dysfunction. Patients with ARDS and requirement for respiratory support need urgently to be transferred to the intensive care unit (ICU). It is reported that nasal cannula or mask, high-flow nasal cannula, non-invasive ventilation (NIV), invasive mechanical ventilation (IMV), and veno-venous extracorporeal membrane oxygenation (VV-ECMO) were widely used in COVID-19 according to the severity of respiratory dysfunction (2) (3) (4) . Cardiac injury is common in COVID-19, with an incidence of 36% and closely related to a higher risk of mortality (5) . It is reported that, in a systematic review and meta-analysis, the pooled incidence of acute kidney injury (AKI) was 28.6% among hospitalized COVID-19 patients from the USA and Europe and 5.5% among patients from China. Kidney replacement therapy (KRT) was used in 20.6% of patients admitted to the intensive care unit (6) .

As is universally known, the mortality of critically ill patients is higher than that of ordinary patients. A systematic review reported that the summary estimate for all-cause mortality was 10% for adult patients with COVID-19 and 34% for critically ill patients within minor countries (7) . In order to gain a clearer picture of the mortality of critically ill patients within major countries and clinical interventions or supportive care for organ dysfunction in the ICU, we meta-analyzed the relevant literature. The results may provide a narrative for the mortality of critically ill patients with COVID-19 as well as the effect of clinical characteristics and interventions between surviving and non-surviving patient groups.

This systematic review was performed in compliance with the Centre of Reviews and Dissemination guidelines (8) and reported according to the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) statement (9) . In order to complete the systematic review and provide some references for clinical intervention during COVID-19 as soon as possible, this review was not registered.

We included studies that focused on the mortality of critically ill patients with laboratory-confirmed COVID-19, clinical characteristics, and interventions or supportive care of organ dysfunction.

We included original studies that fulfill the following criteria: (1) the type of study was cohort, case-control, or case-series designs, (2) the study topic was related to the mortality, clinical characteristics, and interventions or supportive care of critically ill patients with COVID-19, which is defined as a positive result of a real-time reverse transcriptase-polymerase chain reaction (RT-PCR) assay of nasal and pharyngeal swabs (10), and (3) the study was published or posted in English or Chinese. We excluded duplicates, conference abstracts, letters, and studies for which we could not access the full text and missing data of outcomes. In order to avoid a small size, only studies of more than 50 patients were included. If there were two or more studies that included the same population, only the study with the largest sample size was chosen.

In this review, the primary outcome was the mortality of critically ill patients with COVID-19. The secondary outcomes included all sorts of supportive care, including non-invasive respiratory support, IMV, KRT, and vasopressor. Critically or severely ill patients were defined as those patients who were admitted to the ICU or required respiratory support. Surviving patients were defined as those discharged from the ICU or hospital or who remained hospitalized. Non-surviving patients were defined as those who died in the ICU or hospital. Immunoregulation therapy includes corticosteroids, interferon, and intravenous immunoglobulin G.

Two reviewers (ZQ and SL) carried out the search independently in the following six electronic databases from their inception to May 15, 2021: MEDLINE (via PubMed), the Cochrane library, Web of Science, China Biology Medicine disc, China National Knowledge Infrastructure, and Wanfang Data. The main terms were "mortality, " "critically ill patient, " "severely ill patient, " "novel coronavirus, " "2019-novel coronavirus, " "Novel CoV, " "SARS-CoV-2, " "COVID-19, " "2019-CoV, " "invasive mechanical ventilation, " "high flow nasal cannula, " "noninvasive ventilation, " "extracorporeal membrane oxygenation, " "renal replacement therapy, " "kidney replacement therapy, " "vasopressor, " and so on (the details of the search strategy can be found in Supplementary File 1) . Moreover, we also searched the clinical trial registry platforms, the Google Scholar, the reference lists of the identified reviews, and the preprint platforms [including SSRN (https://www. ssrn.com/index.cfm/en/), medRxiv (https://www.medrxiv. org/), and bioRxiv (https://www.biorxiv.org/)] for further potential studies.

After eliminating duplicates by using EndNote X9.3.2 software, two reviewers independently screened all titles and abstracts first, and then the full texts of potentially relevant articles were Frontiers in Medicine | www.frontiersin.org reviewed to identify the final inclusion. Discrepancies were settled by discussion or consultation with a third reviewer. All reasons for exclusion of ineligible studies were recorded, and the process of study selection was documented using a PRISMA flow diagram (11) .

Two reviewers (ZQ and SL) extracted data independently with a standard data collection form. Any disagreements were resolved by consensus, and a third reviewer (XL) checked the consistency and accuracy of all data. The following data and information were extracted for each included study: basic information (title, first author, publication year, funding, and study design), information on the participants (sample size, age, and inclusion/exclusion criteria of participants), details of the intervention and control conditions, outcome information [for dichotomous data, we abstracted the number of events and total participants per group; for continuous data, we abstracted the means, standard deviations (SD), and number of total participants per group].

Two reviewers (ZQ and SL) assessed the potential risk of bias of each included study independently. Discrepancies were resolved by discussion and consensus with a third researcher (XL). We assessed the risk of bias in cohort studies using Newcastle-Ottawa Scale (12) , which contains eight domains: representativeness of exposure cohorts, selection of nonexposure cohorts, determination of exposure, outcome events that did not occur before study initiation, comparability of cohort based on design or analysis, assessment of outcome events, adequacy of follow-up time, and completeness of follow-up. For case series, we used the Joanna Briggs Institute critical appraisal checklist for case series (13) , which consists of 10 domains. Each domain was graded as one sore if reported. 

All statistical analyses were performed using RStudio, version 1.3.1056. Comparable data from studies with one outcome were pooled using forest plots according to the Cochrane Handbook by using random-effects model separately (14) . Mortality in the ICU and in hospital was used for a detailed description. A subgroup analysis was performed according to different regions. For dichotomous outcomes, we calculated the risk ratios (RR) and the corresponding 95% confidence intervals (CI) and Pvalues. For continuous outcomes, we calculated the standardized mean difference and its corresponding 95% CI if means and SD were reported. Furthermore, 95% prediction interval (PI) was used to evaluate the range that, we assert with 95% certainty, will fall into during a future validation test. We reported the effect size with 95% CI by using random-effects models. Two-sided P < 0.05 were considered statistically significant. Heterogeneity was defined as P < 0.10 and I 2 >50%.When effect sizes could not be pooled due to only one study for a comparison, we reported the study findings narratively. We used sensitivity analyses to evaluate the stability of mortality outcomes of the included studies. For a result that included more than 10 studies, publication bias was tested by visual funnel plots.

The quality of evidence for each outcome was assessed by using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach. The judgments of quality for specific outcomes were based on five main factors: study design and execution limitations, inconsistency, indirectness, imprecision of results (randomeffects model), and publication bias across all studies (15, 16) . The quality of evidence for each outcome was graded as high, moderate, low, or very low (17) and presented in "GRADE Evidence Profiles" (18) .

The literature search retrieved 9,362 records through database searching and 51 additional records through other sources, which included 36 from the Google Scholar and 15 from preprint platforms. After removing duplicates, we screened the titles and abstracts of 5,138 records and reviewed the full text of 101 articles. Finally, we included 33 studies (cohort studies and caseseries) that reported either the mortality of critically ill patients or the clinical interventions between surviving and nonsurviving patients with COVID-19 ( Figure 1 ). All of them were published in English.

The basic characteristics of the included studies of the mortality of critically ill patients are summarized in File 3) . A visual analysis of the funnel plot indicated that no publication bias was suspected in the results of age and mortality in the ICU. The results of IMV, PaO 2 /FiO 2 ratio, and SOFA source were suggestive of publication bias (Supplementary File 4) . 

We evaluated the quality of evidence for 11 outcomes. Among them, two outcomes (18%) were graded as of moderate quality, four outcomes (36%) were graded as of low quality, and five (45%) outcomes were graded as of very low quality. We produced "GRADE evidence profiles, " and the details of GRADE can be found in Supplementary File 5.

We conducted a sensitivity analysis on each result by omitting one study at a time. No study had a significant impact on the results of the meta-analysis (Supplementary File 6) . A sensitivity analysis showed that all studies had little or acceptable effect on the total combined effect and that the results were stable.

The epidemic of COVID-19 is not stopping yet, especially in western countries. In previous reports, the incidence of mortality associated with critically ill patients remains poorly characterized. The novel findings in this study include the mortality of critically ill patients with laboratory-confirmed COVID-19 worldwide and the clinical interventions between surviving and non-surviving patients. The results show that allcause mortality in ICU was 35% and mortality in hospital was 32% around the world for the year 2020. Differences were distinct between regions. The incidence of mortality that occurred in Southeast Asia was as high as 48%, followed by 39% in China and the Middle East. The lowest incidence occurred in America, which is 15%. The plausible explanations for the high mortality in China and other Asia countries are that the arrival and peak of the COVID-19 pandemic in Asia were earlier than in any region, and there was a shortage of ICU resources and experience. Moreover, data may be subject to patient selection for ICU admission, and some nations adopted a stringent strategy (19) . In addition, mortality also relates to the time of follow-up. Some of the participants remained in the hospital in mechanical ventilation even at the end of follow-up. A recent meta-analysis reported that all-cause mortality associated with COVID-19 was 10% overall and 34% in patients admitted to the ICU (7), but most of their participants were from China; in this part, we had a close result. This new meta-analysis included more participants and covered much wider regions. Early identification and prompt organ function support care would provide relief in critical cases (53) . Among the included studies, five identified independent risk factors were associated with ICU mortality from laboratory parameters to clinical intervention, but the results are not the same (22, 25, 38, 50, 51, 54) . We compared the baseline clinical characteristics between surviving and non-surviving patients. What we found based on the univariate analysis was that old age, APACHEII score, and SOFA score displayed consistency with multivariate Cox regression analysis in these five studies. Besides these, the PaO 2 /FiO 2 ratio is an important index to reflect the severity of respiratory failure. Our results also showed that the PaO 2 /FiO 2 ratio is helpful to predict the outcome.

With regard to the outcome of the clinical interventions of this meta-analysis, respiratory support is the most important part of life sustaining treatments. According to this study, HFNO during ICU hospitalization was more often used in non-surviving patients, and IMV was more often used in surviving patients.

In previous studies, Auld and Capone (22, 54) reported that receipt of IMV was associated with a decreased likelihood of survival. When we discuss the difference of respiratory support, respiratory support as rescue therapy and the different severity levels of the two groups should not be ignored. HFNO and NIV can be safely used in COVID-19-related mild-moderate ARDS. In the study of non-COVID-19, HFNO has been associated with lower mortality in hypoxemic respiratory failure (55) , but in some moderate-severe ARDS patients, HFNO or NIV should be used cautiously due to rapid progression to severe type and a high risk of treatment failure. According to Mukhtar et al. (56) , the use of NIV with a predefined algorithm in subjects with moderatesevere COVID-19 ARDS was successful in 77% of the subjects. IMV is the most widely used therapy of severe hypoxemia. The population with IMV was larger than with non-invasive support in this study. The need of endotracheal intubation and invasive mechanical ventilation was eight times that of noninvasive ventilation in a previous study (30) . Although the timing of IMV is disputed, as evidenced in a recent publication, a metaanalysis reported that early intubation was not associated with improved survival (57) . A latest meta-analysis (42) reported that the timing of intubation may have not influenced the mortality of critically ill patients with COVID-19. ECMO can be taken into consideration if the respiratory dysfunction of patients develop into severe ARDS, which cannot sustain with IMV, but this salvage treatment did not have a statistically significant difference between the two groups. In a study with a small sample (3), two of five patients survived by the support of ECMO. The appropriate time and eligible patients need to be evaluated.

In a previous research, as high as 31% of patients in a cohort developed severe acute kidney injury requiring renal replacement therapy during hospitalization (25) . High creatinine level, AKI, and receipt of RRT were independent risk factors for the inhospital mortality of patients (22, 51, 58) . Similarly, high highsensitivity cardiac troponin I level, ischemic heart disease, cardiac injury, and vasopressor support were associated with death in patients (22, 38, 50, 51, 54) . In the present study, the result shows that vasopressors and RRT were more often used in the surviving group.

There were some limitations in the current study that must be acknowledged. First is the high level of heterogeneity in the study. Plausible explanations for the heterogeneous risks of mortality include differences in age, nation and race, disease severity, and insufficient length of follow-up. It was difficult for us to control for the effects of these confounding factors. The heterogeneity in the component studies was addressed with random-effects models. Second, as for the secondary outcomes, is that this part of the clinical interventions was derived from an observational cohort, not a randomized controlled trial, so these results should be treated cautiously. The key purpose of this study is to describe the effect of the actual use of various clinical interventions in the surviving group and non-surviving group rather than the impact of individual measures on the prognosis. Third is that most studies were retrospective and recall bias might have occurred.

Mortality was high in critically ill patients with COVID-19 based on low-quality evidence, and intercontinental differences existed. The early identification of critical characteristics and the use of support care help to indicate the outcome of critically ill patients.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/ Supplementary Material.

WHO Coronavirus Disease (COVID19) Dashboard. Available online at

Feasibility and clinical impact of out-of-ICU non-invasive respiratory support in patients with COVID-19 related pneumonia

Clinical characteristics and outcomes of five critical COVID-19 patients treated with extracorporeal membrane oxygenation in Leishenshan Hospital in Wuhan

Timing of invasive mechanic ventilation in critically ill patients with coronavirus disease 2019

Prevalence and impact of myocardial injury in patients hospitalized with COVID-19 infection

Acute kidney injury and kidney replacement therapy in COVID-19: a systematic review and meta-analysis

Acute complications and mortality in hospitalized patients with coronavirus disease 2019: a systematic review and meta-analysis

CRD's guidance for undertaking reviews in health care

Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement

Clinical Management of Severe Acute Respiratory Infection When Novel Coronavirus (nCoV) Infection Is Suspected: Interim Guidance

The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration

Critical evaluation of the Newcastle-Ottawa scale for the assessment of the quality of nonrandomized studies in meta-analyses

Methodological quality of case series studies: an introduction to the JBI critical appraisal tool

Meta-analysis in clinical trials

GRADE: an emerging consensus on rating quality of evidence and strength of recommendations

GRADE Handbook for Grading Quality of Evidence and Strength of Recommendation

The skills and experience of GRADE methodologists can be assessed with a simple tool

GRADE guidelines: 1. Introduction-GRADE evidence profiles and summary of findings tables

COVID-19 in critically ill patients in North Brabant, the Netherlands: patient characteristics and outcomes

Clinical characteristics and predictors of 28-day mortality in 352 critically ill patients with COVID-19: a retrospective study

Multisystem outcomes and predictors of mortality in critically ill patients with COVID-19: demographics and disease acuity matter more than comorbidities or treatment modalities

ICU and ventilator mortality among critically ill adults with COVID-19. medRxiv

Bilevel and continuous positive airway pressure and factors linked to all-cause mortality in COVID-19 patients in an intermediate respiratory intensive care unit in Italy

Risk factors for mortality in critically ill patients with COVID-19 in Huanggang, China: a single-center multivariate pattern analysis

Epidemiology, clinical course, and outcomes of critically ill adults with COVID-19 in New York City: a prospective cohort study

Epidemiology, outcomes, and utilization of intensive care unit resources for critically ill COVID-19 patients in Libya: a prospective multi-center cohort study

Disparities in intensive care unit admission and mortality among patients with schizophrenia and COVID-19: a national cohort study

Factors influencing liberation from mechanical ventilation in coronavirus disease 2019: multicenter observational study in fifteen Italian ICUs

The association between blood pressure levels and mortality in critically ill patients with COVID-19 in Wuhan, China: a case-series report

Baseline Characteristics and Outcomes of 1591. Patients Infected With SARS-CoV-2 Admitted to ICUs of the Lombardy Region

CHA2DS2-VASc score and modified CHA2DS2-VASc score can predict mortality and intensive care unit hospitalization in COVID-19 patients

Association between early treatment with tocilizumab and mortality among critically ill patients with COVID-19

Respiratory support in severely or critically ill ICU patients with COVID-19 in Wuhan, China

The impact of viremia on organ failure, biomarkers and mortality in a Swedish cohort of critically ill COVID-19 patients

Dosing of thromboprophylaxis and mortality in critically ill COVID-19 patients

Evolving changes in mortality of 13,301 critically ill adult patients with COVID-19 over 8 months

Admission high-sensitive cardiac troponin t level increase is independently associated with higher mortality in critically ill patients with COVID-19: a multicenter study

Risk factors for severity and mortality in adult COVID-19 inpatients in Wuhan

Predictors of mortality in patients with severe COVID-19 pneumonia -a retrospective study

Baseline characteristics and outcomes of patients with COVID-19 admitted to intensive care units in Vancouver, Canada: a case series

Prevalence and clinical significance of antiphospholipid antibodies in patients with coronavirus disease 2019 admitted to intensive care units: a prospective observational study

Effect of timing of intubation on clinical outcomes of critically ill patients with COVID-19: a systematic review and meta-analysis of non-randomized cohort studies

Predictors of mortality in critically ill COVID-19 patients demanding high oxygen flow: a thin line between inflammation, cytokine storm, and coagulopathy

Mortality of patients with severe COVID-19 in the intensive care unit: an observational study from a major COVID-19 receiving hospital

Deploying unsupervised clustering analysis to derive clinical phenotypes and risk factors associated with mortality risk in 2022 critically ill patients with COVID-19 in Spain

Declining intensive care unit mortality of COVID-19: a multi-center study

Hospital resources may be an important aspect of mortality rate among critically ill patients with COVID-19: the paradigm of Greece

Clinical features and outcomes of ICU patients with COVID-19 Infection in Tehran, Iran: a single-centered retrospective cohort study

Characteristics and risk factors associated with mortality in critically ill patients with COVID-19

Prognostic factors associated with mortality risk and disease progression in 639 critically ill patients with COVID-19 in Europe: initial report of the international RISC-19-ICU prospective observational cohort

Clinical characteristics and outcomes of critically ill patients with novel coronavirus infectious disease (COVID-19) in China: a retrospective multicenter study

High Mobility Group Box 1 and Interleukin 6 at Intensive Care Unit Admission as Biomarkers in Critically Ill COVID-19 Patients

Novel coronavirus infection during the 2019-2020 epidemic: preparing intensive care units-the experience in Sichuan Province

Characterization of critically ill COVID-19 patients at a brooklyn safety-net hospital

High-flow oxygen through nasal cannula in acute hypoxemic respiratory failure

Outcome of non-invasive ventilation in COVID-19 critically ill patients: a retrospective observational Study

Clinical significance of timing of intubation in critically ill patients with COVID-19: a multi-center retrospective study

Kidney impairment is associated with in-hospital death of COVID-19 patients

LL, YC, and ZQ: concept and design. LL: administrative, technical, or material support. ZQ and SL: statistical analysis and drafting of the manuscript. LL and YC: supervision. All authors critical revision of the manuscript for important intellectual content, acquisition, analysis, or interpretation of data.

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fmed.

Conflict of Interest: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.