Intensive care management of patients with COVID-19: a practical approach

SARS-CoV-2, the causative agent of coronavirus disease 2019 (COVID-19), is responsible for the largest pandemic facing humanity since the Spanish flu pandemic in the early twentieth century. Since there is no specific antiviral treatment, optimized support is the most relevant factor in the patient's prognosis. In the hospital setting, the identification of high-risk patients for clinical deterioration is essential to ensure access to intensive treatment of severe conditions in a timely manner. The initial management of hypoxemia includes conventional oxygen therapy, high-flow nasal canula oxygen, and non-invasive ventilation. For patients requiring invasive mechanical ventilation, lung-protective ventilation with low tidal volumes and plateau pressure is recommended. Cardiovascular complications are frequent and include myocardial injury, thrombotic events, myocarditis, and cardiogenic shock. Acute renal failure is a common complication and is a marker of poor prognosis, with significant impact in costs and resources allocation. Regarding promising therapies for COVID-19, the most promising drugs until now are remdesivir and corticosteroids although further studies may be needed to confirm their effectiveness. Other therapies such as, tocilizumab, anakinra, other anti-cytokine drugs, and heparin are being tested in clinical trials. Thousands of physicians are living a scenario that none of us have ever seen: demand for hospital exceed capacity in most countries. Until now, the certainty we have is that we should try to decrease the number of infected patients and that an optimized critical care support is the best strategy to improve patient’s survival.


Introduction
Since December 31, 2019, when China reported a series of cases of acute respiratory failure caused by a new species of coronavirus, SARS-CoV-2, more than 50 million new cases and almost 1,260,000 deaths have been confirmed worldwide. In Brazil, 5,664,115 cases were reported with 162,397 deaths from the disease by November 8th, 2020 [1]. Its rapid spread and high lethality, especially in the most fragile groups such as the elderly and those with comorbidities, make this pandemic a new challenge faced by modern medicine.
This publication aims to provide a specialist consensus on specific management of COVID-19 in intensive care, covering from the admission criteria in the intensive neutrophilia, lymphocytopenia, elevated markers of hepatic and renal failure (aspartate aminotransferase, alanine aminotransferase, creatinine, and urea), elevated acute-phase proteins as markers of inflammation (highsensitivity C-reactive protein, procalcitonin, and serum ferritin), and elevated coagulation function-related indicators (prothrombin time, fibrinogen, and D-dimer) [1,10].
Admission criteria include oxygen requirements equal or superior to 6-8 l/min to reach a peripheral oxygen saturation ≥ 90-92%, respiratory failure, shock, acute organ dysfunction, and patients at high risk for clinical deterioration. However, in many countries, due to the shortage of ICU beds, usually only patients requiring intubation and invasive mechanical ventilation were admitted to ICU.

Pulmonary impairment, physiopathology, and ventilation strategies
The pathophysiology of COVID-19-induced ARDS involves characteristic properties which make it different from other causes of ARDS: patients present an intense endothelial dysfunction with a thromboinflammatory state. Multiple mechanisms of dysregulation in the pulmonary perfusion exist in COVID-19: the abolition of hypoxic pulmonary vasoconstriction, excessive pulmonary vasoconstriction; and microthrombosis or macrothrombosis, leading to increased dead space [13,14]. Pulmonary microthrombosis and endothelial damage that result in V/Q (ventilation/perfusion) mismatch, hypoxemia, and vasodilation ( Fig. 1) [15][16][17]. Increased inflammatory and thrombotic biomarkers are associated with severe clinical presentation and mortality in COVID-19 patients. High levels of D-dimer, IL-6, C-reactive protein, procalcitonin, troponin, LDH, and ferritin are detected in severely ill patients [18].
Hospitalization trends vary with age and may reach around 20% of COVID-19 patients [19]. In hospitalized patients, ventilatory support may vary from the need for (See figure on next page.) Fig. 1 The pathophysiology of SARS-CoV-2 infection. SARS-CoV-2, via its surface spike protein, binds to the human ACE2 receptor after spike protein activation by TMPRSS2. This results in down-regulation of ACE2 and increased angiotensin II levels and consequently increased plasminogen activator inhibitor C-1 expression and reduced fibrinolysis. The disease it causes is associated with an increase in inflammatory cytokines and coagulation disorders, with predisposition to thrombus formation. Mononuclear cells interact with activated platelets and the coagulation cascade, which activate 1 inflammatory cells by binding thrombin and tissue factor with specific protease activated receptors and by binding fibrin to Toll-like receptor 4. The activation of inflammatory cells results in the release of pro-inflammatory cytokines, leading to impairment of the natural coagulation pathways and shut down of fibrinolysis. This state of hyper inflammation and hypercoagulability leads to multiple organ dysfunction, most commonly affecting the lungs, heart and kidneys. ACE2 angiotensin-converting enzyme-2, aPTT activated partial thromboplastin time, ARDS acute respiratory distress syndrome, COVID-19 coronavirus disease 2019, HFpEF heart failure preserved ejection fraction, HFrEF, heart failure reduced ejection fraction, IL interleukin, PAR protease-activated receptor, PT prothrombin time, SARS-COV-2severe acute respiratory syndrome coronavirus 2, TMPRSS2 transmembrane protease serine, TLR4 Toll-like receptor 4, TNFα tumor necrosis factor-α O 2 supplementation through a nasal catheter to invasive mechanical ventilation or extracorporeal membrane oxygenation (venovenous ECMO) in patients with the most severe forms of ARDS. In general, patients must be maintained with the minimum amount of supplemental O 2 for a SpO 2 between 92 and 96%. Without BGA, the degree of hypoxemia can be estimated by SpO 2 /FiO 2 ratios, with values ≤ 315 suggesting ARDS [20].
Non-invasive ventilation: Health services have been and are predicted to be overloaded in most large centers as a result of the spreading COVID-19 pandemic, leading mainly to the lack of ICU beds and insufficient number of mechanical ventilators for patients who need them [21]. Many hospitals have adopted non-invasive ventilation (NIV) as an attempt to prevent respiratory failure from evolving into severe forms that require invasive ventilatory support.
Both the European Society for Intensive Care Medicine (ESICM) with the international Surviving Sepsis Campaign: Guidelines on the management of critically ill adults with coronavirus disease 2019 (COVID- 19), and the National Health Service of England (NHS-England) recommend the use of NIV as an initial measure for respiratory failure in patients with COVID-19 (weak recommendation, low-quality evidence) [22,23]. It is still an ongoing debate about which would be the most recommended non-invasive interface and technique: NIV devices such as BIPAP, or the use of high-flow nasal cannulas (HFNC).
The prone position in non-intubated patients were tested in some initial studies [24][25][26]. Patients who tolerated more than 3 h in prone position present a substantially improve in oxygenation from supine to prone positioning. However, only about half of the patients maintain the benefit after resupination [25,26]. Further studies are needed to reinforce the real benefit of this intervention.

Invasive mechanical ventilation
ARDS is defined as a form of inflammatory pulmonary edema of non-cardiogenic etiology, with a reduction in the areas of normoventilated lung and consequent reduction in respiratory compliance and shunt effect. The Berlin definition proposed categories of ARDS based on degree of hypoxemia: mild (200 mmHg < PaO 2 / FIO 2 ≤ 300 mmHg), moderate (100 mmHg < PaO 2 / FIO 2 ≤ 200 mmHg), and severe (PaO 2 /FIO 2 ≤ 100 mmHg) and variables for severe ARDS: radiographic severity, respiratory system compliance (≤ 40 mL/cm H 2 O), positive end-expiratory pressure (≥ 10 cm H 2 O), and corrected expired volume per minute (≥ 10 L/min) [27].
To manage these patients, maneuvers that lead to recruitment of collapsed areas are usually applied, such as increased positive end-expiratory pressure (PEEP), alveolar recruitment maneuvers, and prone position, leading to a reduction in elastance and increased compliance [28]. Prone positioning presents the potential benefit of a relieve of severe hypoxemia due to reduction of overinflated lung areas, promoting alveolar recruitment and decreasing ventilation/perfusion mismatch. This intervention might be considered in patients with PO2/ FiO2 < 150, in the absence of contraindications [29,30]. The main objective of mechanical ventilation in these patients is to maintain a lung-protective strategy for all patients with ARDS, defined as targeting a tidal volume of 4 to 8 mL/kg predicted body weight (PBW) and a plateau pressure of less than 30 cmH 2 O [28].
A group of experts hypothesize that in COVID, there may be two phenotypes of ARDS [21]. Patients often exhibit normal compliance even in the presence of severe hypoxemia, with normal or even increased minute ventilation, and more than half of these patients do not appear dyspneic. Radiologically, such patients have ground-glass tomographic lesions indicative of interstitial and nonalveolar edema, and these infiltrates are relatively limited in extent at this stage. These patients are called "type L" ("low elastance"), with additional main characteristics of high compliance, low response to PEEP, and low lung weight estimated by chest computed tomography (CT) [21]. Patients may evolve with progressive clinical improvement or, whether due to individual predisposing factors or inadequate management, evolve with a more severe form closer to the classic ARDS. This is named as "type H" (from "high elastance"), showing also low compliance, high response to PEEP, and high lung weight estimated on chest CT [31]. It should be highlighted that this division is conceptual, to facilitate the understanding of the respiratory condition, with types "H" and "L" representing the ends of a spectrum that frequently overlap [31].

Mechanical ventilation strategy according to patient phenotypes ("type L" or "type H")
In severe cases of respiratory failure, as frequently seen in SARS-CoV-2-related ARDS, severe hypoxemia can lead to a persistent increase in respiratory effort, with consequent self-induced lung injury (P-SILI). In addition, other factors such as fluid overload or SARS-1 CoV-2-induced myocardial injury may also play important roles in worsening of the condition through pulmonary congestion [32]. Thus, a mechanical ventilation strategy must take into account the multiple mechanisms of lung injury and the different presentations of the disease-conventional form of ventilation in ARDS will not always be the most appropriate, as described below [31].
• Type L: it is suggested to ventilate "type L" patients, typically patients with good lung compliance, higher tidal volumes (VT) (around 7-8 mL/ kg of ideal body weight). Higher VT helps to avoid reabsorption atelectasis and hypercapnia due to limited VT-induced hypoventilation. The rationale behind this strategy is as follows: the initial feature of these patients is the vasoregulation defect in the pulmonary capillaries-the reflex vasoconstriction that normally occurs in response to hypoxemia is not found in these patients due to endothelial changes and microthrombosis. Elevation of FiO 2 may be sufficient in most patients not experiencing excessive respiratory effort, with maintenance of NIV with BIPAP or HFNC leading to slow and progressive improvement of hypoxemia and reversal of ARDS. However, if the inflammatory condition progresses, or if the patient's ventilatory effort is excessive, secondary pulmonary tissue stress may lead to P-SILI, with severe deterioration of lung function. At this point, intubation with adequate sedation/paralysis can interrupt the vicious cycle. These patients should be ventilated with lower PEEP (between 8 and 10 cmH 2 O) to avoid redirection of blood flow away from the aerated pulmonary capillaries, which would increase the shunt effect. As capillary hypoperfusion can also suffer a gravity-dependent effect, the prone position could be used as a strategy to minimize it and increase oxygenation. • Type H: with disease progression and worsening of inflammatory edema, the patient may progress to "type H". The pathophysiology of this progression is probably the result of a combination of factors: in addition to self-induced lesion (P-SILI), the viral lesion itself leads to uncontrolled inflammation and edema, with local and generalized thrombogenesis, intense release of cytokines, and right ventricular overload. The resulting pulmonary edema is close to classic ARDS presentation, with collapsed alveoli and extensive normoperfused and hypoaerated areas. In these more advanced cases, a mechanical ventilation strategy should be more traditional: elevated PEEP, VT < 6 mL/kg, driving pressure < 14 cmH 2 O, prone position, and alveolar recruitment maneuvers in refractory cases.
As previously stated, categorization in two different profiles facilitates clinical management by indicating the need for different ventilatory approaches. However, due to the frequent overlap of the two types, individualization of ventilatory management is essential. In either case, patients with COVID-19 who undergo mechanical ventilation have an average recovery time of 1-3 weeks [33,34]. The progress toward improvement is characteristically slow; therefore, prolonged sedation is often unavoidable. In most severe cases of ARDS and also in cases of non-protective ventilation or in the occurrence of asynchrony, neuromuscular blockage is useful, and complications such as polyneuropathy of the critically ill patient are usually diagnosed. Figure 2 shows the main principles of the management of COVID-19 respiratory failure.
The main goal of shock and hemodynamic instability management of patients with COVID-19 severe illness is to restore arterial pressure and to optimize cardiac output with the ultimate goal to improve or preserve organ perfusion. Severe lung injury and mechanical ventilation also contribute to deleterious hemodynamic effects. The extensive viral infection and ARDS induced by COVID-19 results in diffuse lung inflammation, consolidation, marked microvascular thrombosis, endothelial dysfunction, and vasoconstriction [41]. In addition, hypoxemia and dead space lead to a rise in pulmonary vascular pressure and resistance and right-ventricular (RV) afterload increase [42].
The increase in RV afterload and preload may result in RV dilatation, in a septal shift toward the left ventricle (LV) and then in a decrease in LV filling that eventually result in low cardiac output and hemodynamic deterioration [42].
Acute Cor Pulmonale (ACP) is a relatively common occurrence in severe COVID-19 patients. In some patients, ACP can be the result of acute pulmonary embolism, a frequent event in the course of COVID-19 [43], which should prompt a specific therapy [44].
A conservative fluid strategy is usually recommended in patients with ARDS. Aggressive fluid administration and hypervolemia is associated with longer ICU stay, prolonged ventilator dependence, and higher mortality [22,42,45,46]. However, hypovolemia can occur in COVID-19, especially in the early phase, or even in later phases due to intense associated sepsis (increased capillary leakage and increased venous capacitance). Uncorrected hypovolemia may lead to peripheral organ hypoperfusion, facilitate thrombi formation in the context of severe COVID-19 coagulopathy, and might even aggravate hypoxemia due to a low PvO 2 effect in case of high degree of pulmonary shunt fraction. It is thus important to assess the benefit/risk ratio of fluid administration. The expected benefit could be assessed by performance of fluid responsiveness tests, administering bolus of hypotonic crystalloids (lactated`s Ringer), and evaluating dynamic variables such as cardiac index and velocitytime integral [47]. Use of tidal volume challenge assessing the changes in pulse pressure variation during a transient increase in tidal volume (e.g., from 6 to 8 mL/kg) [48] is an excellent option in COVID-19 ARDS patients often deeply sedated. If available, advanced hemodynamic monitoring technologies 1 such as transpulmonary thermodilution may help assessing the fluid infusion benefit/ risk ratio.
Additionally, COVID-19 myocardial depression may develop at any phase of the disease, sometimes with fulminant myocarditis which might occur in about 1% of hospitalized patients [49,50]. Early detection of myocardial involvement through the measurement of troponin and natriuretic peptide concentrations and echocardiography is recommended [4,11]. A recent paper showed that even in less severe forms of disease, myocardial inflammation may persist in almost 60% of patients 70 days after disease [51]. Point-of-care echocardiography might help in the diagnosis of shock and in the non-invasive assessment of preload responsiveness [47,[52][53][54].
Norepinephrine is the first-line vasopressor in patients with hemodynamic instability and COVID-19 [22,45,55]. In patients with ARDS, norepinephrine also can improve RV function by restoring mean arterial pressure and thus RV blood supply [42]. If norepinephrine is unable to achieve adequate mean arterial pressure, vasopressin could be added as a second-line vasoactive agent to reach the target blood pressure [22]. Vasopressin might be used as first vasopressor, mainly in cases of atrial fibrillation, pulmonary hypertension, and acute renal failure [56]. Angiotensin 2 (Ang-2) has also been successfully used in COVID-19 patients in a few cases [57]. Although there are no definite trials to support Ang-2′s superiority over conventional vasopressors in COVID patients with vasodilatory shock, there is a physiologic rationale for using the drug. Dobutamine is the inotropic agent the most used in case of acute heart failure. It is indicated in the presence of cardiac dysfunction and in the occurrence of tissue hypoxia after fluid status adjustment and norepinephrine administration.
Prone positioning and inhaled selective pulmonary vasodilators have been used for patients with refractory hypoxemia; moreover, they may have a beneficial hemodynamic effect in particular by decreasing RV afterload and restoring RV function [58,59]. Finally, in cases of refractory ARDS combined or not with refractory cardiogenic shock requiring high vasoactive doses, venovenous or veno-arterial ECMO might be considered to restore blood oxygenation and provide hemodynamic stability [60]. Although many new centers have been created and ECMO services are now available worldwide, access to ECMO is restricted, especially in low-income countries. Establishing models of care, in which severely ill patients who are eligible for ECMO are evaluated by systems that can be used to rapidly transfer and group high numbers of critically ill patients would be important to provide high-quality ECMO care during the pandemics. Another issue of interest in ECMO management is that the SARS-CoV-2-induced infection may be associated with higher rates of thrombotic events of the extracorporeal system during V-V ECMO therapy [61].

Renal impairment, physiopathology, and treatment strategies
Although respiratory failure is the main dysfunction caused by SARS-CoV-2, other organs can also be affected, with cardiac and renal failures as the most relevant disorders [62]. Acute kidney injury (AKI) was present in 6.7% with a mortality rate as high as 91.7% in SARS-CoV (causative agent of SARS) [63]. Among 99 ICU COVID-19 patients, Fominsky et al. found that 72 (75.0%) developed AKI and 17 (17.7%) received continuous renal replacement therapy (CRRT) [64].
In a prospective cohort study that included 701 patients diagnosed with COVID-19, the assessment of renal function at admission showed that serum creatinine was elevated in 14.4% and urea in 13.1% of patients [65]. Abnormalities in the coagulation pathway, including prolonged partial thromboplastin time and high D-dimer, were more common in patients with elevated baseline serum creatinine. The risk factors associated with mortality were: proteinuria in any degree, hematuria, elevated basal creatinine, and renal failure AKIN (Acute Kidney Injury Net) 2 or more. There was a correlation between severity of kidney injury (AKIN stages) and death, with a fourfold higher risk of mortality among those with stage 3 AKIN [65].

COVID-19 nephropathy
Findings like proteinuria and hematuria can occur after COVID-19 infection, with some individuals showing signs and symptoms of AKI. It has been demonstrated that RNA viruses are present in urine and renal tissue, indicating that the kidney may also be a target of COVID-19 infection 1 through direct viral invasion in the tubules and renal interstitium [66,67].
Renal histopathology was examined in a series of autopsies of 26 patients who died of respiratory failure secondary to COVID-19. All patients had evidence of acute tubular injury of varying severity, and a number of other histopathological findings including clusters of erythrocytes and hemosiderin pigments were also present. Of the nine samples tested for the intracellular virus, coronavirus-like particles were identified in seven [66].
Renal failure due to COVID-19 has a multifactorial etiology, with three main mechanisms involved: cytokine injury, organ crosstalk, and systemic effects of infection [68].

Cytokine injury
Cytokine storm through release of IL-6/JAK2/STAT3/ SOCS3 and NF-κB (p65)/IL-18 could work together to induce AKI and increase overall renal-related diagnostic markers [69]. McElvaney OJ et al. showed that IL-1β, IL-6, IL-8, and sTNFR1 were all increased in patients with COVID-19. COVID ICU patients could be clearly differentiated from COVID stable patients, and demonstrated higher levels of IL-1β, IL-6, and sTNFR1 [70]. The contribution of increased vascular permeability and volume depletion, as well as cardiomyopathy-which can lead to type I cardiorenal syndrome-in addition to cytokine activation, is yet to be established. Removing cytokines with extracorporeal therapies, often studied as a promising approach in patients with sepsis and AKI, has been proposed in patients with COVID-19 who develop acute renal failure [7,71].

Organ crosstalk
Connection between alveolar injury and tubular injury has been proven. A retrospective study that included 357 patients with ARDS without kidney disease or AKI at presentation reported that 68% of patients developed AKI. Positive fluid balance, greater disease severity, older patients, and diabetes were independently associated with the development of AKI [72].

Systemic effects
Hemodynamic instability associated with rhabdomyolysis, metabolic acidosis, and hyperkalemia can also occur in COVID-19 patients and contribute to AKI.

Treatment
The indications for renal replacement therapy (RRT) for AKI in critically ill patients are well stablished, regardless of patients' COVID-19 status. However, in an intensive care services overload scenario, providing RRT to an increasing number of patients may exceed the capacity 1 of available machines, supplies, and specialized staff.
In patients without indication of RRT, conservative treatment includes: appropriate dose-loop diuretics (oral or intravenous) for fluid overload and active management of hyperkalemia and metabolic acidosis with potassium binders and sodium bicarbonate. For patients that do not respond to conservative treatment, RRT is indicated.
The use of CRRT remains preferred among critically ill patients with AKI. Even among hemodynamically stable patients who can tolerate intermittent hemodialysis, CRRT, or sustained low-efficiency dialysis (SLED) are preferred, depending on machinery and staff availability and expertise. CRRT or SLED can be managed without 1:1 nursing support, which could potentially help in minimizing the waste of personal protective equipment and limit exposure among nurses on hemodialysis [73].
Regional citrate is the most used anticoagulation strategy during hemofiltration or dialysis. However, some case reports suggest that circuit thrombosis during RRT occurs more frequently in patients with COVID-19 than in other patients, and in these cases, the addition of therapeutic anticoagulation with nonfractioned heparin might be considered [74].

Metabolic impairment and treatment strategies
Data suggest that diabetes, hypertension, and cardiovascular diseases (CVD) are the most common comorbidities related to COVID-19, although prevalence rates vary among different studies. In a pooled data from 10 Chinese studies, prevalence of hypertension, diabetes, and CVD was 21, 11, and 7%, respectively [75].
Recent data suggest that diabetic patients with COVID-19 are more often associated with most severe forms of the disease, varying between 14 and 32% in different studies, with an odds ratio of 2.34 for ARDS compared with patients without diabetes [10].
A study from Wuhan with 161 patients with COVID-19 demonstrated a delayed viral clearance in patients with diabetes. It has been proposed that, in addition to the usual mechanisms (impaired neutrophil chemotaxis, and phagocytosis) by which diabetes predisposes to infections 1 in general, other specific factors related to SARS-CoV-2 can have roles in the increased risk and severity of the disease in diabetes, as following [76]: Increased expression of angiotensin-converting enzyme-2 (ACE2): acute-phase hyperglycemia results in increased ACE2 expression, which can facilitate the entry of viral cells; however, chronic hyperglycemia reduces expression of ACE2, leading to an increased vulnerability to the inflammatory and harmful effects of the virus [77]. In addition, in pancreatic islets, the effect of SARS-CoV on ACE2 blood glucose receptors can lead to hyperglycemia, even in patients without pre-existing diabetes. In patients with SARS-CoV hyperglycemia persisted for up to 3 years after recovery, indicating transient damage to beta cells [78]. Increased furin: the amount of furin, a membranelinked protease belonging to the subtilisin/cexinfamily proprotein convertase (PCSK), increases in diabetic individuals. Its role in entry of viruses into the cell, acting as a facilitator for viral replication, has been demonstrated [79]. Impaired T-cell function: lymphocytopenia was observed in patients with COVID-19 and was correlated with a worse prognosis [80].
Increased interleukin-6 (IL-6): IL-6 levels are higher in patients with diabetes. Moreover, it is one of the most relevant cytokines activated in cytokine storm in COVID-19 patients. Thus, it may play a more deleterious role in SARS-CoV-2 infection.
Another potential pathway that may explain the correlation between COVID-19 and diabetes involves the enzyme dipeptidyl peptidase-4 (DPP-4), one of the main targets of pharmacological treatment in patients with type 2 diabetes. DPP-4 works as a functional receptor for MERS-CoV in vitro. Despite direct participation of DPP-4 in glucose and insulin metabolism in type 2 diabetes, it has been shown to also increase inflammation. However, a possible role in SARS-CoV-2 infection and whether treatment of diabetes with DPP-4 inhibitors could alter the course of COVID-19 infection are not known yet [81].

Glycemic control
Monitoring blood glucose levels is an important factor in acute stage and follow-up, especially in those receiving corticosteroid therapy. To date, limited data are available on the association of blood 1 glucose levels and COVID-19; however, data from SARS and H1N1 infections have shown that poor glycemic control increases the risk of complications and death [82].
Thereby, the recommendations for critically ill patients with COVID-19 are: • Monitor blood glucose in infected patients; • Glycemic control in patients already known to be diabetic: collection of plasma glucose, electrolytes, and pH; • Liberal indication for early use of intravenous insulin in severe cases (ARDS and shock), avoiding subcutaneous use; • Therapeutic objectives: Blood glucose between 72 and 144 mg/dL or 4-16 mmol/L; In elderly patients (> 70 years) or fragile: blood glucose minimum 90 mg/dL or 5 mmol/L [83].

Specific treatment
Several drugs have been studied for the treatment of the SARS-CoV-2. Most studied antivirals in this scenario were the combination of lopinavir-ritonavir and remdesivir [84,85]. Currently, the antiviral therapy that appears most promising is remdesivir. This is a prodrug of a nucleotide analogue that is intracellularly metabolized to an analogue of adenosine triphosphate that inhibits viral RNA polymerases. In previous studies, remdesivir was shown to have in vitro activity against Ebola and several coronaviruses, showing a prophylactic and therapeutic efficacy in nonclinical models [86,87].
Remdesivir has been used recently on a compassionate basis, due to a lack of proven efficacy drugs. A multicenter RCT including 1063 patients receiving remdesivir or placebo showed that remdesivir use led to a significantly shorter duration of hospital stay (11 vs. 15 days), and lower mortality (8% vs. 11.6%) [88]. The FDA has approved this drug for urgent use in COVID-19, it is prescribed intravenously (200 mg IV day 1, and 100 mg IV from day 2 to day 10). A paper published by the same group of authors showed that 5 days of therapy were as effective as 10 days [89].
The antivirals lopinavir/ritonavir, ribavirin, atazanavir, and favipiravir are being tested in the context of COVID-19. Previous in vitro studies suggested that lopinavir presents inhibitory activity against SARS-CoV and MERS-CoV [90][91][92]. Cao B et al. demonstrated in an RCT with 199 hospitalized patients with respiratory failure that lopinavir/ritonavir did not result in any clinical benefit beyond standard care [84]. A recent multicenter trial tested the combination of lopinavir/ ritonavir with interferon beta-1b and ribavirin, and confirmed that in comparison with standard care, the combination of drugs resulted in shorter duration of viral shedding and hospital stay and in clinical improvement [93].
As adjuvant therapies in the treatment of COVID-19, chloroquine and hydroxychloroquine have been evaluated in experimental and clinical studies. These drugs have the ability to increase the endosomal pH of cells and reduce replication of SARS-CoV-2 in vitro [94,95]. However, initial clinical studies have not revealed any clinical benefit for using these drugs either alone or in combination with azithromycin [96,97]. Cavalcanti et al. in a randomized clinical trial with 667 hospitalized patients with suspected or confirmed COVID-19 who presented with mild-to-moderate manifestation, the use of hydroxychloroquine, alone or with azithromycin, did not improve clinical status at 15 days as compared with standard care [98]. Similarly, the use of hydroxychloroquine has not been beneficial in preventing the development of COVID-19 in patients after high-risk exposure [99]. The routine use of these drugs is not recommended.
In addition, Mercuro et al., in a cohort of 90 patients, showed that hydroxychloroquine in COVID-19 patients was associated with a high risk of QTc prolongation, and concurrent treatment with azithromycin resulted in greater changes in QTc. We recommend that physicians carefully evaluate the benefits and potential risks of this drug [100].
The use of corticosteroids reduces mortality in COVID-19 patients needing respiratory support [101]. The trial showed that dexamethasone 6 mg once a day for 10 days results in lower 28-day mortality among those who were receiving either invasive mechanical ventilation or oxygen. Usual regimens of steroids usually prescribed in ARDS which might be considered in COVID-19 include intravenous methylprednisolone 0.5 mg/Kg twice a day for 5 days, dexamethasone 6 mg once a day for 10 days or dexamethasone 20 mg once a day for 5 days followed by dexamethasone 10 mg once a day for 5 days [102][103][104]. A recent meta-analysis included 678 patients who received steroids during COVID-19 (hydrocortisone, dexamethasone, or methylprednisolone) and showed that the administration of systemic corticosteroids, compared with usual care or placebo, was associated with lower 28-day all-cause mortality [105].
Tocilizumab use has been associated with reduced mechanical ventilation and reduced serum IL-6 in these patients [109,110]. The benefit may be related to the phenotype of intense inflammation, characterized by high levels of IL-6, D-dimer, C-reactive protein, LDH, and ferritin. Its preferential use must be carried out through clinical research protocols [111]; patients admitted early to the ICU and still not intubated are probably who mostly benefit. Two retrospective studies showed efficacy of tocilizumab in COVID-19 [112,113]. However, 4 randomized studies did not confirm the efficacy of tocilizumab in COVID-19 [114][115][116][117].
Convalescent plasma has been used for the treatment of infectious diseases since the early twentieth century with reduced mortality in cases series and case reports during the 1918 influenza, the 2003 SARS, and the 2009 influenza H1N13 pandemics [118]. In COVID-19, several uncontrolled case series of convalescent plasma use were performed, suggesting a survival benefit [119,120].
Li et al. published the first RCT of convalescent plasma in COVID-19, showing no significant benefit in clinical improvement or mortality. However, this study suggests possible benefit in the subgroup of severely ill patients. Further clinical trials are needed to establish the clinical indications for antibody therapies against COVID-19 [98].

Antithrombotic treatment
Progressive respiratory failure is the primary cause of death in the COVID-19 pandemic, followed by cardiovascular complications. Pathological studies performed in COVID-19 patients showed severe endothelial injury, associated with the presence of intracellular virus and disrupted cell membranes. Pulmonary vessels had widespread thrombosis with microangiopathy and alveolar microthrombi [121]. A Brazilian study of 10 minimally invasive autopsies revealed the presence of diffuse alveolar damage in the lung, and epithelial viral cytopathic effects in alveolar and small airway epithelia. A variable number of small fibrinous thrombi in small pulmonary arterioles were found in areas of both damaged and preserved lungs. Signs of bacterial pneumonia were observed in 6 of 10 cases [122].
In patients who died from COVID-19-associated or influenza-associated respiratory failure, the histologic pattern in the peripheral lung was diffuse alveolar damage with T-cell infiltration. However, increased thrombogenesis was 9 times more prevalent in patients with COVID-19 than in influenza patients [121]. COVID-19 has been described as a thromboinflammatory disease [15] with thrombogenesis a consequence of severe endothelial injury, exacerbated inflammation, suppressed fibrinolysis, loss of natural anticoagulants, and activation of platelets and coagulation factors. Because of these pathophysiologic findings, initial studies started to look for deep venous thrombosis, arterial thrombosis, and microthrombosis. The available data on thrombotic risk are quite limited and based largely on case series from China [123], the Netherlands [6], and France [124]. Recent studies have described a high incidence of deep venous thrombosis and pulmonary embolism varying from 35 to 78% in COVID-19 patients [5,125,126].
Nonetheless, most experts agree that the signal for increased thrombotic risk is sufficient to recommend pharmacologic venous thromboembolism (VTE) prophylaxis in all hospitalized COVID-19 patients as long as there is no contraindication [28,127]. What remains to be confirmed is the real role of therapeutic anticoagulation in these patients. There is much controversy about this issue, while RCT results are not available [16,128]. The guidelines specifically mention that the anticoagulation regimens may be modified based on extremes of body weight (50% increase in dose if obese), severe thrombocytopenia, or worsening renal function [129].
One of the difficulties in determining the true incidence of thrombosis is that access to diagnostic testing may be limited. In a report from the Netherlands (where routine VTE prophylaxis is given), high rates of VTE were noted among ICU patients [6]. More than one-third of these patients, however, had a pulmonary embolism limited to subsegmental. The therapeutic dose should be considered for patients with severe COVID-19 and signs of 1 sepsis-induced coagulopathy (SIC) and/or high D-dimer (6 × higher the reference values) in association with other biomarkers of severity, in the absence of contraindication for anticoagulation. This can be considered a therapeutic strategy for SARS-CoV-2 infection, based on experts' opinion and a few retrospective studies [15,16]. Moreover, this strategy requires the use of strict institutional protocols that enable surveillance and rapid intervention if complications occur. Figure 4 shows a proposed algorithm to assess thrombogenesis in patients with COVID-19, as well as a treatment suggestion. However, data are still insufficient to identify critical aspects relevant to therapeutic plans, such as the best drug choice, its dosage, and administration time schedule, as well as the duration of treatment.

Ethical dilemma during the pandemics
The COVID-19 epidemic requires an increased number of resources, reinforcement of the ICU teams with new staff members, reorganization of the unit, and protocol changing. All of these might create vulnerability and loss of control for all the professionals. In many times, intensivists should make their choices based on local politics, structural resources, and team abilities [130].

Conclusions
COVID-19 is one of the most challenging health emergencies we have faced this century. Health professionals are experiencing very difficult times, with limitations of resources and information, which when available needs to be confirmed before integration into clinical practice. To date, there is no proven specific treatment for the management of COVID-19, although studies with promising results have recently emerged. We know already that the lowest rates of mortality are related to better quality of care. Early diagnosis, application of effective therapies and adequate strategies of clinical stratification are needed for better outcomes in COVID-19 patients. We have learned that COVID-19-related respiratory dysfunction has unique characteristics that require individualized management and are aware of the importance of supporting the patients' hemodynamics because of the high risk of cardiovascular and renal complications. The long duration of this disease poses a challenge for the health system and health professionals. Validated protocols 1 of care are essential when dealing with millions of affected people in different countries and in different levels of care. Until an effective vaccine is widely available, the world will need to adapt to the reality of a pandemic that has come to change all paradigms of modern medicine.