Management of critically ill patients with COVID-19 in ICU: statement from front-line intensive care experts in Wuhan, China

As the front-line of the COVID-19 outbreak response, health care workers are exposed to a huge risk of infection. Therefore, health care workers must follow the standard precautionary principles and try their best to ensure the personal protection, hand hygiene, ward management, environmental ventilation, and sanitization of the object surface, so as to avoid nosocomial cross-infection.

As a high-risk environment, tertiary class protection is suggested for health care workers in intensive care unit (ICU). Personal protective equipment (PPE) includes disposable surgical cap, N95 mask, work uniform, disposable medical uniforms, disposable latex gloves, goggles, and full-face shields. Full-face respiratory protective devices or powered air-purifying respirators are required when performing aerosol-generating procedures. Destroying and disposing of masks properly, putting on and removing PPE, and practicing hand hygiene are necessary to avoid self-contamination. Special attention should be paid to details such as the side exposure of the eyes and wrists with glove slippage, as well as the risks of infection while removing some disposable shoe covers [7]. The hand hygiene system should be strictly implemented according to the newly developed Five Moments for Hand Hygiene included in the WHO Guidelines on Hand Hygiene in Health Care (Advanced Draft) [8].

Clinical triage system needs to be established to assess all patients at admission, allow for early recognition of possible COVID-19 cases and immediate isolation of patients with suspected disease in an area separate from other patients (source control). The number of family members and visitors who are in contact with suspected or confirmed COVID-19 patients should be limited or visiting should be prohibited altogether. The proper disposal of clinical waste should be ensured [9].

Health care workers need to self-monitor for signs of illness and self-isolate. If illness occurs, they should report it to managers and stay at home. A sensible diet, proper rest, and adequate exercise are advised to maintain physical and psychological health. Health care workers should familiarize themselves with related working procedures so as to avoid mistakes [10].

Rationale It is suggested to adjust measures according to the differing conditions so as to set the ICU ward rationally. Contaminated areas, potentially contaminated area and clean areas need to be strictly divided. The buffer zone should be set between every two areas. Posting eye-catching logos on each area is required to prevent straying into the wrong place. Different points of access should be set for medical staff and patients, making sure they do not get crossed. For ICU, tertiary class protection should be correctly performed in each area, which is of great importance for precaution of COVID-19 [11]. The use of negative pressure rooms with natural ventilation is recommended by the WHO guidance to prevent the spread of airborne pathogens among rooms [7, 12].

First-aid materials and medicine such as oxygen tank, electrocardiogram (ECG) monitor, defibrillator, injection pump, infusion pump, endotracheal intubation supplies, portable vacuum extractor, noninvasive ventilator, invasive ventilator, hemofiltration equipment, extracorporeal membrane oxygenation (ECMO) equipment and so on should be prepared. Other equipment, including air disinfecting machine and air cleaner, as well as medical gas systems including oxygen, compressed air, special gas, and vacuum suction systems, need to be assured too.

It is of particular importance to implement effective measures to prevent the spread of COVID-19 in ICU. Disinfection includes concomitant disinfection and terminal disinfection. Concomitant disinfection must be conducted immediately for the materials and environment contaminated by the excretion of the suspected and confirmed patients. Following the end of 1 day’s work in ICU, or the patients’ recovery or death in the isolation ward, terminal disinfection needs to be done carefully. Key disinfection objects include patients’ living supplies such as clothes and quilt, medical supplies, ground and wall space of ICU wards, the surface of desks and bed tables, as well as air [11, 13].

Current evidence indicates that COVID-19 is mainly transmitted from person to person through droplets, contact, and even high concentrations of aerosols [6]. Large amounts of droplets and aerosol are generated by sputum suction in the airway, specimen collection, tracheal intubation, fiber bronchoscopy, tracheotomy, etc. Accordingly, surgeons are at a great risk of contamination. In order to avoid occupational exposure, recommendations during the aerosol-generating procedures in COVID-19 patients are the following:

Rationale Negative pressure rooms are aimed to decrease the concentration of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pathogens. In view of that, the risk of contamination would be decreased during the aerosol-generating procedures in such a setting. During the severe acute respiratory syndrome (SARS) epidemic, it was reported that negative pressure settings were effective in preventing cross-contamination and protecting the staff and patients inside the room [14]. According to WHO recommendations for COVID-19 patients, such locations should be with a minimum of 12 air changes per hour or at least 160 L/second/patient with natural ventilation [3].

Rationale An observational study reported that among 138 hospitalized patients diagnosed with confirmed COVID-19 in Zhongnan Hospital in Wuhan in January, 2020, 40 were healthcare workers [15]. Till March 15, 2020, it has been reported that over 3000 health workers were confirmed with COVID-19, among whom 14 died. The memory of what has happened during the 2003 SARS outbreak is still fresh. A systematic review showed that the healthcare workers who performed aerosol-generating procedures, including endotracheal intubation (odds ratio, 6.6), noninvasive ventilation (odds ratio, 3.1), tracheotomy (odds ratio, 4.2), and manual ventilation before intubation (odds ratio, 2.8) were at higher risk of suffering from SARS infection compared with the non-performers [16].

Most of the infections among healthcare workers occurred at the early stage of this outbreak when the self-protective directive has not yet been established and reinforced. After confirmation of human to human transmission of SARS-CoV-2, the self-protection for healthcare workers was subsequently established and reinforced from the end of January 2020. Level III biosafety protection is mandatory for intubation according to the guidance of the General Office of the National Health Committee [17].

PPE donning process should be strictly followed during high-risk operation: disposable hair cover, fit-tested N95 respirator or equivalent, fluid-resistant gown, two layers of gloves, goggle and face shield, and fluid-resistant shoe covers. The main operator should use portable air-purifying respirator. All the donning processes should be supervised by a professional nurse or assistant.

Doffing process of PPE after high-risk exposure should also be followed: hand hygiene, face shield and goggle removal, fluid-resistant gown removal, outer glove removal, shoe cover removal, inner glove removal, hand hygiene, N95 respirator or equivalent removal, and hair cover removal. The doffing process seems to be of greater importance. All the processes should also be supervised so as to reduce the risk of contamination [18].

Rationale Large amounts of aerosols generated by incubation can increase the risk of transmission and nosocomial infection [16]. Thus, visual devices are recommended to facilitate the procedure, limit operation time [19] and ensure the distance between operator and patient. Routine fiber bronchoscopy operations are not suggested for COVID-19 patients. Meanwhile, most COVID-19 patients have few airway secretions [4] so that the indication of bronchoscopy should be strictly minimized. According to the recommendations by the Centers for Disease Control and Prevention (CDC) [20] and WHO [9], disposable medical equipment should be used for patient care if possible.

Rationale Severe COVID-19 patients with artificial airway tend to suffer from severe hypoxemia [15]. The patient’s secretions, droplets, and aerosols can be widely spread during the operation. Patients should be intubated within 60 s [18].

The procedure of fiber bronchoscopy should be performed gently with great caution in severe COVID-19 patients.

During bronchoscopy, following procedures should be followed to avoid aerosols spreading: artificial airway should be connected with a disposable three-way connector to a ventilator, then (a) ventilator needs to be set to standby mode, (b) the artificial airway needs to be briefly clamped, (c) the bronchoscopy should be quickly inserted into the connector, (d) the clamp should be opened, (e) ventilation should be restored [21].

For the patients requiring mechanical ventilation, it is not advisable to disconnect patients from the ventilator.

Even though some clinical experts insisted that antiviral therapy is unnecessary for seriously ill patients with COVID-19 since the course of disease in severe types is longer than 2 weeks, multiple virus particles have been found at the lung lesions following histopathological examination. Up to date, there is no specific antiviral drug that has been testified and globally recognized effective for treating COVID-19. In China, several antiviral drugs such as ribavirin, ganciclovir, oseltamivir, arbidol, alpha-interferon, chloroquine, lopinavir–ritonavir, and remdesivir have been used in clinical settings for the treatment of COVID-19. Among them, oseltamivir and arbidol hydrochloride are the most commonly utilized; however, these antiviral drugs were originally designed for influenza, and their efficacy and safety for COVID-19 need to be further investigated.

Rationale Ribavirin is a broad-spectrum antiviral drug. Clinical observations have suggested that early use of this drug is efficacious in containing COVID-19. To avoid possible aerosol transmission, we do not recommend alpha-interferon nebulization for COVID-19 infected patients. According to a very recently published clinical study from France, hydroxychloroquine can significantly reduce viral load in COVID-19 patients, and azithromycin can further enhance this effect [22]. In this study, combination use of hydroxychloroquine (HCQ) and azithromycin for at least 3 days at an early stage could rapidly reduce the nasopharyngeal viral load and decrease the length of hospital stay for infected patients. It should be noted that treatment with higher chloroquine diphosphate (CQ) dosage (600 mg CQ twice daily) is not recommended for severe COVID-19 due to its potential safety hazards, especially when taken concurrently with azithromycin and oseltamivir [23]. Nonetheless, a randomized controlled trial (RCT) trial conducted by Cao et al. suggested monotherapy of lopinavir-ritonavir did not bring about any clinical benefits for severe COVID-19 patients compared with standard supportive care, which may be partly caused by the higher throat viral loads in lopinavir–ritonavir group, delayed treatment initiation [24]. Of note, these clinical studies were limited by relatively small sample sizes. More large-scale and well-designed clinical trials are needed to confirm their potential therapeutic effects. Arbidol monotherapy might be better than lopinavir–ritonavir in reducing viral load in COVID-19 patients [25]. A clinical study from Gilead Sciences showed that remdesivir could improve clinical conditions in critically ill patients with COVID-19, and stop patient from receiving invasive mechanical ventilation or ECMO [26]. However, a recent multicentre study published in the Lancet found no benefit of remdesivir in improvement of clinical outcomes for severe COVID-19 [27]. One recent study published in N Engl J Med showed that compassionate use of remdesivir improved clinical outcomes in a subset of severe COVID-19 patients [28]. However, the absence of control groups precludes a final conclusion. The definite therapeutic effectiveness of remdesivir in the treatment of severe COVID-19 needs to be further verified. Remdesivir has been approved as a potential treatment for severe COVID-19 patients by the Japanese Ministry of Health, Labour and Welfare (MHLW) on May 7, 2020 due to the COVID-19 pandemic [29].

The main side-effects of these antivirals include QT interval elongation, bradycardia, hepatic injury, and obvious gastrointestinal reactions such as serious diarrhea and vomiting which may contributed to disease deterioration. Clinical trials testing remdisivir for the treatment of severe COVID-19 patients are underway (NCT04365725, NCT04292899).

Convalescent plasma therapy belongs to passive immunization, which is used for the treatment of virus infections when specific drugs and vaccines are unavailable. Convalescent plasma, which has been used for more than one hundred years, can provide specific antibodies to neutralize and eradicate the viruses from the blood circulation. Up to date, there is no particular treatment for COVID-19. In 2014, the WHO recommended the use of convalescent plasma collected from patients who recovered from the Ebola virus infection as an empirical treatment during the outbreak [30]. During the COVID-19 epidemic period, this method was also recommended by the National Health Commission of China for the treatment of severe and critical patients [6].

Rationale Convalescent plasma has been testified to suppress viremia, shorten the hospital stay, and reduce mortality during several virus epidemics. In 1918 during a Spanish influenza pandemic, convalescent plasma reduced the mortality rate by > 50% in severe patients [31]. Since then, it was also used for prophylaxis or as a treatment for several virus infections such as measles, Argentine hemorrhagic fever, influenza, chickenpox, and infection by cytomegalovirus. Over the past two decades, its efficacy and safety were confirmed during pandemics of SARS, MERS, H1N1 and H5N1 avian flu. During the SARS pandemic in 2003, eighty patients received convalescent plasma at Prince of Wales Hospital, Hong Kong. By the 22nd day, a higher discharge rate was observed in patients (n = 48) given convalescent plasma before day 14 than that given plasma after day 14 (58.3% vs. 15.6%; P < 0.001) [32]. A prospective cohort study conducted by Hung et al. showed that convalescent plasma therapy (n = 20) significantly reduced mortality compared to the control group (n = 73) (20.0% vs. 54.8%; P < 0.01). Meanwhile, plasma treatment lowered the upper respiratory tract virus load and decreased serum cytokines levels in patients with severe pandemic (H1N1) 2009 virus infection [33]. These studies verified the efficacy of convalescent plasma in patients with virus infections. It has been reported that among three severe MERS patients who received convalescent plasma infusion, just two showed neutralizing activity [34]. Among five critically ill patients with COVID-19 receiving mechanical ventilation convalescent plasma infusion, 3 patients were discharged, while 2 clinically ill patients improved and maintained the stable condition till the day 37 after transfusion [35]. A study performed in 10 severe COVID-19 patients found that convalescent plasma treatment could improve clinical outcomes, improve immune function, and promote absorption of lung lesions [36]. Nonetheless, just like any other treatment, convalescent plasma has its limitations. The main limitation refers to the reported studies, which are not randomized trials, but just prospective cohort studies or case series studies. Therefore, it was not possible to eliminate the influence of baseline severity and other treatments when evaluating the effects of convalescent plasma therapy. Other limitations include the risk of transmitting infections to transfusion service personnel, the need for adequate selection of donors with high neutralizing antibody titers, and the risk of other transfusion-transmitted infections [37]. However, regardless of these limitations, since there are still no specific etiological treatments for COVID-19, and convalescent plasma is available, it is reasonable to use it in the treatment of COVID-19 patients.

Indication for HFNC and NIV.

Rationale Noninvasive ventilation support (NIV) and HFNC are important treatments for COVID-19-induced mild and moderate ARDS. The mechanisms of the two treatments are positive end-expiratory pressure, decreased respiratory workload, decreased incidence of intubation, ease of use, and higher comfort. In a randomized trial of adult patients admitted to the ICU for acute hypoxemic, nonhypercapnic respiratory insufficiency, continuous positive airway pressure (CPAP) delivered by face mask was associated with an early improvement in oxygenation; however, it was not associated with a reduced need for intubation or with improved outcomes [38]. A trial that compared HFNC oxygen, standard oxygen via face mask and face mask NIV in 310 patients with acute hypoxemic respiratory failure, reported that the intubation rate was significantly lower with HFNC oxygen than with standard oxygen or NIV among patients with PaO₂/FiO₂ ≤ 200 mmHg at enrollment and, for the whole group (patients with PaO₂/FiO₂ ≤ 300 mmHg), patients managed with HFNC had improved survival. There were no differences in outcomes between NIV and standard oxygen [39]. A sub-study examined the practice of NIV use in ARDS of LUNGSAFE STUDY reporting that NIV was associated with higher ICU mortality in patients with a PaO₂/FiO₂ < 150 mmHg [40]. For COVID-19, there is no sufficient evidence to prove that HFNC is superior to NIV.

Rationale For all cases with noninvasive support, patients should be closely monitored, as deterioration can abruptly occur [41]. In China, some patients presented with hypoxemia, later named “silence hypoxemia”, since these patients were without corresponding clinical manifestations, e.g., no high respiratory rates, high heart rate, respiratory distress, and other hypoxia symptoms. These patients have a high risk of sudden death and should be closely monitored and timely provided with oxygen therapy.

Positive responses are usually evident soon after the initiation of NIV and HFNC. If there is no substantial improvement in gas exchange and respiratory rate within a few hours, invasive mechanical ventilation should be started without delay. Failure to recognize a lack of improvement during noninvasive support may result in further respiratory deterioration and/or cardiac arrest, often with devastating consequences.

Delayed intubation increases ARDS mortality; therefore, early recognition of ARDS severity could avoid delayed intubation. If the use of HFNC fails, endotracheal intubation is unavoidable even with the use of rescue NIV [42].

The indications for HFNC and NIV intubation are a higher level of severity (SAPS II score > 34), hypoxemia (PaO₂/FiO₂ ≤ 150 mmHg), hypoxemia that is not improved following NIV treatment for 1 h, and strong spontaneous breathing (tidal volume with NIV > 12 mL/kg PBW) [43]. ROX index can be used to predict HFNC failure and intubation for patients with respiratory failure; > 4.88, suggests a high chance of success, < 3.85 suggests a high risk of failure, and intubating the patient should be discussed; index between 3.85 and 4.88, suggests the patient should be monitored very closely and intubation delays should be avoided [44].

Nine RCTs including a total of 1629 patients have compared mechanical ventilation strategies that limit tidal volumes (4–8 mL/kg predicted body weight [PBW]: males = 50 + 0.91 × [height (cm) − 152.4] kg and females = 45.5 + 0.91 × [height (cm) − 152.4] kg) and inspiratory pressures with traditional strategies (with tidal volumes 10–15 mL/kg PBW) [45, 46]. Another trial that employed a multilevel mediation analysis to analyze individual data from 3562 patients with ARDS, who were also included in nine previously reported randomized trials, identified driving pressure as the ventilation variable that best-stratified risk. Decreases in driving pressure owing to changes in ventilator settings were strongly associated with increased survival [47].

Low tidal volume (6–8 mL/kg PBW), limited plateau pressure (< 30 cmH2O), and driving pressure (< 15 cm H₂O) could decrease ARDS mortality.

Rationale Alveolar collapse is mainly generated by inflammatory lung edema, impairment of chest wall movement, and surfactant deficiency. Some reports have shown different effects of recruitment maneuvers in ARDS patients due to lung recruitability [48]. From our experience in Wuhan, most of the COVID-19 patients had low lung recruitability [49].

Due to the infectiousness of COVID-19, CT, and the other necessary equipment cannot always be used to evaluate lung recruitability. However, some bedside measurements, such as the pressure–volume curve, recruitment to inflation ratio, and clinical parameters, can be measured by a ventilator and used to evaluate lung recruitability [50].

Rationale Use of positive end-expiratory pressure (PEEP) usually improves gas exchange and helps reduce the need for high FiO2. In addition, appropriate levels may limit VILI by maintaining lung recruitment and improving lung homogeneity [51]. When applied with a constant Pplat, PEEP reduces the driving pressure and keeps the lung recruited.

Because of the lack of resources, PEEP selection criteria may include lung recruitability, PEEP/FiO2 table, respiratory system compliance, optimal oxygenation, and driving pressure [46, 47, 52]. Based on the available data, all PEEP values represent a compromise between the extent of recruitment and overdistension, and hemodynamics.

Rationale In China, hypercapnia has been commonly found in COVID-19-induced ARDS. The mechanisms are related to lung injury inhomogeneity and an increase in dead space. Firstly, optimization of ventilator setting is important; secondly, the prone position could decrease dead space and improve hypercapnia [53]; thirdly, tracheal gas inflation (TGI), which influences sputum drainage, could increase alveolar ventilation and CO₂ removal [54]; fourthly, extracorporeal life support or CO₂ removal equipment could improve hypercapnia.

Rationale Prone positioning has a beneficial effect on oxygenation, lung recruitment, and stress distribution. The physiological effects of prone positioning include redistribution of lung densities, often with the recruitment of well-perfused dorsal regions. Although prone positioning increases chest wall elastance, this change is usually accompanied by improved lung recruitment, a reduction in alveolar shunt and improved ventilation/perfusion ratio, subsequent improvement in oxygenation and CO₂ clearance, a more homogeneous distribution of ventilation and a reduced VILI risk [55, 56].

Indications for prone positioning include moderate-to-severe ARDS (PaO₂/FiO₂ < 150 mmHg), and/or hypercapnia. Duration of prone positioning should be more than 12 h, and the termination of prone positioning should be based on the response of oxygenation, lung mechanics, and hemodynamics.

Because prone positioning could improve lung inhomogeneity, early prone positioning should be provided for COVID-19 infected patients with/without respiratory failure [57, 58] since it could prevent respiratory failure.

Since COVID-19 is highly infectious, implementation of the prone positioning might require more manpower, thus further increasing the workload of medical personnel. Pressure injury of the skin and mucous, facial edema, corneal edema, displacement of the catheter, and airway obstruction must be avoided when placing patients in the prone position.

Most of the COVID-19 patients presented with mild symptoms; however, about 14% of patients developed into severe cases, 5% of them were critically ill with mortality estimates of 2.3−3.83% [59‐61]. Mechanical ventilation alone may not be enough to resolve refractory hypoxemia and hypercapnia in these patients. ECMO could be initiated to maintain oxygenation and avoid ventilator-induced lung injury. A cross-sectional study found that 14 (6.2%) patients treated with ECMO [62].

Rationale The appropriate timing of ECMO in COVID-19 patients might be challenging due to enormous demand and uncertainty related to the reversibility of impaired lungs. To guarantee the reversibility of compromised lungs, ECMO should be launched before injurious mechanical ventilation, which is common in critically ill patients with COVID-19 [63, 64]. The primary purpose of ECMO is the maintenance of sufficient oxygenation, removal of CO₂, avoidance of high respiratory drive, and sequencing of ventilator-induced lung injury. The following traditional indications for ECMO may be suitable for COVID-19 patients: PaO₂/FiO₂ < 50 for over 3 h; PaO₂/FiO₂ < 80 for over 6 h; Irreversible pH < 7.2 for over 6 h.

Rationale The WHO guidance released a statement, in which they suggest referring patients with refractory hypoxemia despite lung-protective ventilation to those settings with expertise in ECMO [3]. The latest guidance document issued by ELSO also suggested that ECMO should be considered according to the standard management algorithm for ARDS in patients with viral lower respiratory tract infections [65]. However, in reality, numerous patients who met the criteria for ECMO were admitted over a short period, which was beyond the capacity of the medical resource, including workforce and equipment. In this context, the priority of the ECMO supply should be balanced between the available medical resources and disease reversibility.

Younger patients with minor or no comorbidities should be given the highest priority when resources are limited. Despite standard contradictions, patients who fit the criteria below may be excluded: (1) patients with significant comorbidities; (2) elderly patients with worsening prognosis; (3) patients on mechanical ventilation for more than 7 days.

Rationale Ventilation with the prone position, which is currently recommended by the guidelines, can improve lung heterogeneity as well as oxygenation [66]. It should be considered in the early stages of the disease rather than as a delayed attempt [58]. Prone position ventilation is currently widely applied for severe COVID-19 patients in China [67]. Even if an ultraprotective ventilation strategy is implemented with the aid of ECMO, prone ventilation is considered to benefit the recovery of the lung.

Elevated myocardial enzymes, such as cardiac troponin T (cTnT), creatine kinase (CK), creatine kinase-MB isoenzyme (CK-MB), have been widely observed in critically ill patients with the COVID-19, indicating potential myocardial injury. A significant elevation of myocardial enzymes often indicates a poor prognosis. Most patients with elevated myocardial enzymes do not present compromised left ventricular systolic function (reduced ejection fraction) or abnormal electrocardiogram. Left ventricular diastolic dysfunction or mild-to-moderate pulmonary arterial hypertension is common in some COVID-19 patients.

Rationale While SARS-CoV-2 and MERS-CoV share similar pathogenicity, it has been shown that MERS-CoV can induce acute myocarditis and heart failure [68]. Elevation of biomarkers of cardiac injury is common among critically ill patients with COVID-19 and associated with a higher risk of in-hospital mortality [63, 69]. Reversible subclinical diastolic dysfunction without systolic impairment was observed in SARS [70]. Comparable to SARS, most COVID-19 patients with elevated myocardial enzymes do not present compromised left ventricular systolic function. Left ventricular diastolic dysfunction or mild-to-moderate pulmonary arterial hypertension have been commonly found in COVID-19 patients. From our experience, tachycardia such as sinus tachycardia and atrial fibrillation were also common, while compensatory tachycardia was absent, even in patients with severe hypoxia or hemodynamic collapse.

The exact mechanism of myocardial injury in COVID-19 remains unknown. It has been suggested that direct myocardial injury is mediated via angiotensin converting enzyme 2 (ACE2). ACE2-dependent myocardial infection was observed in the murine model infected with SARS-CoV [71]. One study published in N Engl J Med provides evidence that angiotensin-converting enzyme inhibitors (ACEI)/angiotensin receptor blockers (ARB) medications in COVID-19 patients did not show any association with increasing susceptibility to SARS-CoV-2 [72].

In patients with hemodynamic instability, non-invasive or invasive monitoring, such as echocardiography or thermodilution methods, should probably be used to guide fluid therapy or administration of vasoactive agents. In patients with life-threatening cardiac dysfunction, extracorporeal life support might be salvage therapy.

Rationale The use of vasoactive drugs revealed that the incidence of shock in critically COVID-19 patients was 35%, and 50% in non-survivor population [5]. The shock could be the result of hypovolemia, cardiac injury, and sepsis. Fever and mouth breathing could cause large amounts of fluid loss in critical COVID-19 patients, while decreased water intake, acute gastrointestinal injury, depression, intubation, and sedation could exacerbate hypovolemia. Previous studies reported on the relationship between dehydration and mortality in severe H1N1 patients [73]. Moreover, older age, comorbidities (especially diabetes and cardiovascular disease), lower lymphocyte count, and higher D-dimer levels were identified as risk factors associated with shock [5, 74]. Cardiac injury was found in 23% critical COVID-19 patients [5], which meant poor fluid responsiveness and the risk of pulmonary edema.

For these reasons, the patients’ volumetric status, as well as the fluid responsiveness, should be dynamically assessed. One meta-analysis of 13 RCTs showed that dynamic assessment of fluid responsiveness could improve the clinically relevant outcomes in ICU, such as mortality reduction, reduced duration of ICU length of stay, and mechanical ventilation [75]. Considering the limited clinical resources in the COVID-19 pandemic, we recommend using simple bedside assessments, such as passive leg raising (PLR), lactate clearance, pulse pressure variation (PPV), and inferior vena cava (IVC) collapsibility or distensibility. A recent meta-analysis determined that the PLR induced changes in cardiac output, with a pooled sensitivity of 0.85 and a pooled specificity of 0.91 [76]. PPV also accurately predicted fluid responsiveness in critical patients. In a meta-analysis including 22 studies and 807 patients, PPV predicted fluid responsiveness with the pooled sensitivity of 0.88 and a pooled specificity of 0.89 [77]. IVC collapsibility resulted as a simple, non-invasive bedside predictor of fluid responsiveness with a sensitivity of 0.84 and a specificity of 0.90 [78]. Early lactate clearance-directed therapy was associated with reduced in-hospital mortality, shorter duration of mechanical ventilation, and shorter ICU-stay [79]. A recent observational study showed higher serum lactate levels in COVID-19 non-survivors (2.9 vs. 1.4 mm/L) [5]. Besides, additional attention should also be paid to mental states, degree of thirst, oliguria, skin temperature, and prolonged capillary refilling time as well.

Rationale Even though fluid management in COVID-19 remains unknown, it could be assumed that these patients would respond to fluid therapy in the same way as other ARDS patients. Previous studies have shown that higher cumulative fluid balance is related to the higher mortality of critically ill patients, especially in cases of ARDS [80] and/or septic shock [81]. Due to pulmonary edema in critical COVID-19 patients [82], excessive fluid therapy could increase extravascular lung water and affect gas exchange, resulting in a poor prognosis. One clinical trial found that the conservative fluid strategy improved lung function, shortened the ICU-stay length and duration of mechanical ventilation compared with a liberal strategy in patients with acute lung injury [83]. Another study reported that more than half of critically COVID-19 patients were older than 60 years [5]. When older patients develop cardiac injury and pulmonary edema, they tend to be less responsive to fluid intake [74]. Conservative fluid strategies could reduce the occurrence of positive fluid balance while ensuring tissue perfusion [83]. Although it has been reported that conservative fluid strategy and liberal strategy have a similar incidence of AKI and the requirement for renal replacement therapy (RRT) [83], it is still necessary to closely monitor the renal function of patients. At the same time, attention should be paid to maintaining electrolyte balance and acid–base balance.

Rationale To date, there are still no studies on fluid types in COVID-19 patients; thus, our observations are based on relevant studies of critically ill patients in general. A systematic review of 69 studies that included 30,020 participants revealed that using colloids (such as starches, dextrans, albumin or fresh frozen plasma, or gelatins) had no difference in mortality in critically ill patients compared to crystalloids [84]. Considering the price and accessibility, fluid resuscitation with crystalloids should probably be used for critically ill patients.

One single-center research reported that low serum albumin (36.62 ± 6.60 g/L) was associated with the progression of COVID-19 pneumonia [85], while another study found no significant differences between the non-aggravation and aggravation patients in the early stage of the disease [86]. Serum albumin level < 30 g/L was identified as an independent risk factor for the 30-day mortality in patients with community-onset pneumonia [87]. Based on the previous evidence and our clinical observations, hypoproteinemia is present in most COVID-19 patients; thus, albumin supplement should probably be used for patients with serum albumin levels below 30 g/L.

Rationale Glucocorticoid use in ARDS remains a controversial topic. It is well known that corticoids are beneficial in the treatment of ARDS since they can alleviate inflammatory response and delay fibrosis [115]. A retrospective study conducted in Guangzhou revealed that proper use of corticosteroids in confirmed critical SARS patients led to lower mortality and shorter hospitalization stay and was not associated with significant secondary lower respiratory infections or any other complications [116].

However, there are some inconsistencies in the existing studies. A study involving 197 patients with ARDS, showed improved oxygenation and lung injury score in less than 12 h but no change in 28-day mortality [117]. Another study found no differences in overall mortality, while mortality was increased when steroids were started after day 14 [118]. As for viral pneumonia, a few studies have found that the administration of corticosteroids in patients with influenza pneumonia is associated with increased ICU mortality [119, 120]. WHO does not recommend routine use of corticoids in the treatment of COVID-19, while treatment with methylprednisolone may be beneficial for patients who develop ARDS, as was shown by a retrospective cohort study of 201 patients with confirmed COVID-19 pneumonia admitted to Wuhan Jinyintan Hospital in China [121]. Given the inconclusive evidence and urgent clinical demand, the guidance published by China National Health Commission on March 4, 2020, suggested the use of glucocorticoids over the short time period (3 to 5 days) for patients with progressive deterioration of oxygenation indicators, rapid imaging progress, and excessive activation of inflammatory response. The dosage of methylprednisolone should not exceed 1–2 mg/kg/day. It should be noted that large doses of glucocorticoid might delay the removal of coronavirus due to immunosuppressive effects.

Thymosin is a peptide originally isolated from thymic tissue, which was initially selected for its ability to restore immune function to thymectomized mice. Thymosin may act on precursor T cells to increase the number of activated T helper cells and expression of Th1-type cytokines such as interleukin-2 and interferon-alpha. The activated DCS and Th1 cells then kill bacterial, fungal, or viral infections and lead to the stimulation of differentiation of specific B cells to antibody-producing plasma cells and an improvement in response to vaccines by stimulation of antibody production [122]. The use of thymosin alpha 1 therapy in combination with conventional medical therapies may be effective in improving clinical outcomes in a targeted population of severe sepsis [123]. Also, it has been observed that lower lymphocytes in COVID-19 patients indicate worse prognosis [114]. Thus, thymosin may theoretically have an effect on COVID-19, which needs to be further investigated.

Immunoglobulin may regulate the host’s immune response in a variety of ways, but it had no effect on mortality in previous sepsis studies. At present, it is not recommended in the treatment of COVID-19.

A study performed in 20 severe or critical COVID-19 patients showed that tocilizumab treatment could improve clinical outcomes, promote absorption of lung lesions, improve immune function, and reduce inflammatory response [124]. However, IL-6 inhibitor sarilumab was shown to be ineffective in the treatment of severe COVID-19, leading to early termination of this clinical trial [125]. Large sample size studies using prospective cohort designs are required to verify the therapeutic effect of IL-6 inhibitors for severe COVID-19.