Multi-organ point-of-care ultrasound for COVID-19 (PoCUS4COVID): international expert consensus

COVID-19 has caused great devastation in the past year. Multi-organ point-of-care ultrasound (PoCUS) including lung ultrasound (LUS) and focused cardiac ultrasound (FoCUS) as a clinical adjunct has played a significant role in triaging, diagnosis and medical management of COVID-19 patients. The expert panel from 27 countries and 6 continents with considerable experience of direct application of PoCUS on COVID-19 patients presents evidence-based consensus using GRADE methodology for the quality of evidence and an expedited, modified-Delphi process for the strength of expert consensus. The use of ultrasound is suggested in many clinical situations related to respiratory, cardiovascular and thromboembolic aspects of COVID-19, comparing well with other imaging modalities. The limitations due to insufficient data are highlighted as opportunities for future research.

• Although CT has the best diagnostic yield [8], access is limited by patient volume, resources and risk of environmental contamination. • Pre-existing conditions [9], and acute exacerbations of these diseases are common. • Instability may preclude intra-hospital transportation.

Evidence
LUS is more accurate than CXR for diagnosing respiratory conditions [21], including interstitial diseases [22], pneumonia [23] and COVID-19 pneumonia [24]. The diagnostic accuracy of addition of LUS outperforms standard emergency department tests for dyspnea [25,26]. LUS can diagnose COVID-19 pneumonia in patients with normal vital signs [27] and distinguish viral and bacterial pneumonias [28]. LUS findings associated with COVID-19 pneumonia are reported to be similar to previously described viral pneumonias [12,22]. Frequently observed are [Additional files [2][3][4][5]: heterogeneous B-lines clusters, separated or confluent (corresponding to ground glass opacities on CT), large band-like longitudinal artifacts arising from normal pleural line (characterized as "light beam" [12]), pleural line irregularities, subpleural consolidations and areas with decreased lung sliding due to poor ventilation. Large consolidations with air bronchograms may be present, more commonly in patients requiring mechanical ventilation, possibly representing progression to ARDS or superimposed bacterial infection. At presentation, the distribution, although bilateral, is usually asymmetrical and patchy [29][30][31]. Lung involvement may be limited to dorsal/basal areas in milder COVID-19 pneumonia [32]. LUS shows good agreement with CT in recognizing lung pathology and its severity [33,34] thus, identifying patients at higher risk of clinical deterioration, ICU admission, mechanical ventilation and mortality [34][35][36]. B-line count, consolidations and thickened pleural lines are associated with positive RT-PCR tests and clinical severity [37,38]. Coupled with pretest probability, bilateral B-lines [single and/or confluent], irregular pleural line and subpleural consolidations increase the likelihood of diagnosing COVID-19 [39,40], while non-specific, bilateral heterogeneous patterns [Additional file 6], combined with a typical clinical presentation, strongly suggest viral pneumonia. Conversely, if pre-test probability is low [41], a bilateral A-pattern on LUS may exclude COVID-19 pneumonia owing to its high negative predictive value for pneumonia [12,30].
Multi-organ PoCUS yields a better diagnostic performance for causes of respiratory failure than LUS alone [42]. As a rapid, accurate diagnostic approach to acute dyspnea [43][44][45], it outperforms standard tests  [26]. Similar results have been reported in undifferentiated shock [46]. PoCUS is recommended as a first-line diagnostic test for investigating respiratory failure and/ or hypotension [22,47]. PoCUS may raise suspicions of falsely negative RT-PCR and/or alternate diagnoses [48]. Recognition of comorbidities (chronic RV or LV dysfunction) and COVID-19-associated complications (DVT and RV failure) may influence patient disposition, and PoCUS can change their management [40].
We present a conceptual framework for triage of respiratory failure [Additional file 7]. Without more data, triage protocols cannot be developed that are universally applicable.

Recommendations
1 We suggest using PoCUS, and especially LUS (presence of heterogeneous B-line clusters, pleural line irregularities, subpleural consolidations), and appropriately integrate the information with clinical assessment to diagnose COVID-19 pneumonia (LQE II-B, Very Good Agreement).

Limitations and future research
More data are required to establish the accuracy of LUS findings for the diagnosis of COVID-19 pneumonia versus other viral pneumonias. PoCUS use for risk stratification, outcome prediction, and its impact on management of COVID-19 needs study.
• Up to 17% of hospitalized COVID-19 patients sustain acute cardiac injury (ACI) that increases mortality [50,[51][52][53]. Besides the inflammatory and direct cellular injury, other possible mechanisms for ACI include hypoxemia and result in oxygen supply/demand imbalance [54]. A close association of acute and fulminant myocarditis with COVID-19 is not established. However, if present, it will result in low output syndrome or cardio-circulatory collapse [55]. Though high-sensitivity troponin assays allow detection of myocardial injury, no cutoff values reliably distinguish myocardial infarction (MI) from other ACI [56]. Elevation of cardiac biomarkers, ECG changes, LV and RV dysfunction [57,58] have been reported in myocarditis and AMI [55,59]. • It is difficult to distinguish the effects of pneumonia from superimposed congestive heart failure [59]. • Respiratory acidosis, alveolar inflammatory edema and microvascular alterations may increase pulmonary vascular resistance [60], and positive pressure ventilation may further increase RV afterload, precipitating RV failure [61]. • Various cardiac manifestations [62] have been described, and some critically ill COVID-19 patients exhibit shock states [51].

Evidence
Echocardiography and FoCUS are established tools for diagnosing cardiovascular disease [47,63,64]. FoCUS can detect pre-existing cardiac disease [Additional file 8] and acute RV and/or LV dysfunction [47]. Echocardiography [65] and FoCUS are recommended by American and European Echocardiography societies as diagnostic/ monitoring tools in COVID-19 [66,67]. FoCUS can guide decisions on coronary angiography [68] and inotropic/ mechanical circulatory support [59,69,70]. Overt symptoms of myocardial ischemia, raised cardiac biomarkers, ECG changes and new LV regional wall motion abnormalities should be carefully evaluated so that myocardial infarction [Additional file 9] diagnostic/therapeutic pathways are followed expediently [54,67,68]. Low voltage QRS complexes, myocardial hyper-echogenicity, diffuse hypokinesia or regional wall motion abnormalities suggest myocarditis [71] [Additional file 11]. Acute cor-pulmonale can occur in COVID-19 [58,72], and FoCUS can detect RV dilatation, paradoxical septal motion and RV longitudinal dysfunction [47] [Additional file 10]. Thus, FoCUS/echocardiography together with clinical and biochemical indices can enhance management of cardiovascular compromise. Recommendations 7. We suggest FoCUS and/or echocardiography assessment in moderate-severe COVID-19 as it may change clinical management or provide information that could be lifesaving (LQE II-B, Very Good Agreement). 8 We suggest FoCUS and/or echocardiography for assessment of hemodynamic instability in moderate-severe COVID-19 (LQE II-B, Very Good Agreement). 9 We recommend FoCUS and echocardiography to diagnose RV and LV systolic dysfunction and cardiac tamponade as etiology of hemodynamic instability in COVID-19 (LQE II-B, Very Good Agreement). 10 We suggest using FoCUS/echocardiography to guide hemodynamic management in severe COVID-19 (LQE II-B, Very Good Agreement).

Limitations and future research
Whether subtypes of COVID-19 exist with more severe cardiovascular involvement and worse prognosis, requires investigation. Study of diastolic function may be of interest in COVID-19.

Evidence
Ultrasound is the mainstay of DVT diagnosis [83]. Screening is advised, when feasible, in the general management of COVID-19 patients [84]. Many factors limit access to formal duplex venous sonography [85]. Although routine screening is not widely recommended [86], twice weekly ultrasound surveillance can detect DVT, avert PE and reduce mortality in ICU patients [87]. Lower extremity ultrasound is recommended in COVID-19 patients with unexplained RV dysfunction, unexplained/refractory hypoxemia, or in patients with suspected PE who are too unstable for intra-hospital transport [86].

Because critically ill COVID-19 patients have high
risk for VTE, we suggest regular screening for DVT, including central vessels with catheters, independent of oxygenation and coagulation (LQE II-A, Very Good Agreement). 12 In moderate-severe COVID-19 with hemodynamic worsening or sudden instability, we suggest FoCUS for prompt investigation of acute cor-pulmonale (LQE II-B, Very Good Agreement). 13 In moderate-severe COVID-19, we suggest that echocardiographic indices of worsening RV function and/or increased pulmonary artery pressure may indicate PE (LQE II-A, Very Good Agreement).

Limitations and future research
DVT prevalence and its role in risk stratification in mild COVID-19 are not known. Correlation of DVT with different COVID-pneumonia phenotypes needs study.

DOMAIN 6-PoCUS and respiratory support strategies [including mechanical ventilation]
Phenotypes of COVID-19 pneumonia associated with similar degrees of hypoxemia but different lung weight, aerated volume and compliance have been described [88]. These range from "classic" ARDS (Phenotype-H) that responds to higher PEEP, to the better aerated low elastance (Phenotype-L) that often requires lower PEEP [89]. Future studies may clarify whether phenotyping COVID-19 pneumonia can guide respiratory support, mechanical ventilation settings, and minimize ventilator-induced lung injury [89]. "Classic" ARDS commonly involves dependent lung regions [90]; the same areas are typically involved in advanced COVID-19 pneumonia [89,91]. Localizing consolidated lung is important to maximize benefit from prone positioning. Prone positioning is preferable when dorsal consolidation is severe with spared ventral zones [92]. Prone positioning in non-intubated patients may rapidly improve oxygenation [93,94].
Importantly, LUS may facilitate identification of patients with greater hypoxemia than expected for their alveolar lung injury (Fig. 3), in whom the pathophysiology may involve deranged perfusion (PE, micro-thrombosis, loss of pulmonary vasoconstriction, extrapulmonary shunt).
Global LUS score is strongly associated with lung tissue density/aeration measured with CT [97]. Using LUS to guide mechanical ventilation has been recommended [98] (Fig. 4). However, recruitment demonstrated by LUS correlates with recruitment estimated by pressure-volume curves [99], but not CT [97]. Although LUS may not predict oxygenation response to prone positioning, it does predict re-aeration of dorsal zones [100] (Fig. 5). LUS findings also correlate with extravascular lung water in ARDS [101,102] and can monitor changes in aeration [103]. This has also been suggested in COVID-19 [104][105][106].

14.
We suggest multi-organ PoCUS including LUS over no imaging to guide respiratory support in COVID-19 with respiratory failure (i.e. ventilation, prone positioning, PEEP, recruitment maneuvers) (LQE II-A, Good Agreement). 15 In addition to standard respiratory monitoring, we suggest LUS over CXR and equally to CT, to guide clinical decisions on respiratory support in COVID-19 with respiratory failure (LQE II-B, Good Agreement). 16 We suggest multi-organ PoCUS over LUS alone for decisions about respiratory support in COVID-19 with respiratory failure (LQE II-B, Good Agreement).

Limitations and future research
The benefit of LUS in ventilated COVID-19 patients is only theoretical. Studies to predict response to prone positioning, PEEP titration and other interventions are awaited. Role of LUS to decide invasive mechanical ventilation is unknown.

DOMAIN 7-Management of fluid administration in COVID-19 patients
Fluid management is fundamentally important and often challenging in critically ill patients [107]. In COVID-19 patients, fluid overload can exacerbate lung dysfunction. Recent recommendations stress the need for conservative fluid strategies [4].

Evidence
A large international survey found that PoCUS was the most frequently used approach to assess fluid responsiveness in critically ill COVID-19 patients [108]. While FoCUS can detect early signs of severe central hypovolemia [47] [Additional file 12], interpretation of inferior and superior vena cava collapsibility/distensibility indices is difficult when a variety of ventilation modalities are employed [18,109]. Transesophageal echocardiography has inherent risks and limitations related to manpower and infection control [110]. Dynamic indices based on stroke volume variation, passive leg raising and mini-bolus administration techniques are good predictors of fluid responsiveness [111,112] and can be assessed with transthoracic echocardiography.
In non-COVID-19 pneumonia patients, LUS has been shown to provide information on fluid tolerance and detect the consequences on the lung of overzealous fluid (See figure on next page.) Fig. 3 Examples of lung ultrasound cumulative patterns of patients presenting with a similar degree of hypoxemia, but very different degree of aeration and respiratory mechanics characteristics, and recalling the recently proposed COVID-19 pneumonia phenotypes [89]. Patient on upper panel presents a nearly normal respiratory system compliance and LUS evidence of a milder lung involvement, reflected in a total LUS score of 11. This suggests a lung condition matching which has been recently described as "Phenotype L, " based on CT findings, and characterized by low lung elastance and low ventilation/perfusion ratio (explaining the severe hypoxia). Based on this imaging and on respiratory mechanics findings, final PEEP was set at 10 cm H 2 0. Upper panel shows LUS evidence of a more diffuse and severe diffuse sonographic interstitial syndrome (cause of the shunt and the severe hypoxia), yielding a total LUS score of 27. Respiratory mechanics characteristics recall what has been described as "Phenotype H" (COVID-19 pneumonia: high lung elastance, high right-to-left shunt). Based on this imaging and on respiratory mechanics findings, PEEP was set at 14 cm H 2 0 after a stepwise recruiting maneuver. LUS, lung ultrasound resuscitation [113,114]. Resolution of B-lines during hemodialysis has been described [115] and also observed in COVID-19 patients [116,117].

17.
We suggest FoCUS to screen for severe hypovolemia in moderate-severe COVID-19 at presentation, while Doppler-based fluid-responsiveness indices may be used for subsequent management (LQE II-A, Very Good Agreement). 18 We suggest that LUS alone is not sufficient as a screening tool for pulmonary congestion in moderate-severe COVID-19 (LQE III, Very Good Agreement). 19 We suggest that LUS alone is not sufficient to judge the appropriateness of fluid administration in moderate-severe COVID-19 (LQE II-B, Very Good Agreement). 20 In moderate-severe COVID-19, we suggest multiorgan PoCUS to monitor efficacy of fluid removal, by not only LUS findings of reduction of B-pattern areas, but also echocardiographic signs of resolution of volume overload and decreasing LV filling pressures (LQE II-B, Very Good Agreement).

Limitations and future research
In COVID-19 pneumonia, the severity of the bilateral interstitial manifestations may either be due to variations in the inflammatory condition of the lung or changes due to pulmonary congestion. Simplified PoCUS-guided fluid management could be beneficial in resource-limited settings and needs further studies.

DOMAIN 8-Monitoring patients with COVID-19
PoCUS FOR RESPIRATORY MONITORING: COVID-19 pneumonia is characterized by a wide spectrum of clinical presentations, from mild-moderate hypoxia to severe manifestations requiring life-sustaining measures [118]. In situations where large numbers of patients are admitted to areas with limited monitoring and staffing, disease progression may go unrecognized. Moreover, rapid progression to respiratory arrest has been reported [119]. Severe COVID-19 pneumonia is characterized by severe respiratory failure [120], but not necessarily as ARDS.

Evidence
Evolution of LUS findings and their quantification using scoring systems are effective in monitoring progression or resolution of lung injury, especially in terms of variations in aeration and extravascular water content [22,98,103,121,122]. LUS is very sensitive, but is not specific enough to identify all causes of respiratory deterioration [22]. A comprehensive semi-quantitative LUS approach [97] can assess severity of lung injury and distribution patterns.
In patients with COVID-19 pneumonia, progression of LUS findings has been correlated with clinical and radiological deterioration. Thus, it can accurately monitor the evolution throughout its spectrum of severity, from mechanically ventilated [104,105,123] or veno-venous-ECMO patients [106], to milder cases [124,125,126]. LUS has helped in identifying superimposed bacterial infections [127], and the response to antibiotic treatment [128]. LUS Monitoring has reduced use of CT and CXR in critically ill and COVID-19 populations [129,130].

21.
We suggest serial LUS for respiratory monitoring in moderate-severe COVID-19 (LQE II-B, Very Good Agreement). 22 We suggest multi-organ PoCUS integrated with other clinical and biochemical variables, in preference to CXR for investigation of respiratory deterioration in moderate-severe COVID-19 (LQE II-A, Very Good Agreement). 23 We suggest multi-organ PoCUS over LUS alone to detect respiratory deterioration and guide treatment in moderate-severe COVID-19 (LQE II-B, Very Good Agreement).

Limitations and future research
LUS has limitations and requires further research in early identification of patients who are more likely to progress to severe respiratory failure with inflammation, their pneumonia phenotype, and separate them from those with congestion.

DETECTION OF MECHANICAL VENTILATION-RELATED COMPLICATIONS:
Approximately 2.5% of all COVID-19 patients [118] and up 88% of those admitted to ICU [9] require invasive mechanical ventilation, which may often last for weeks. The diagnosis of complications associated with prolonged ventilation requires imaging that may be limited due to risk of exposure to healthcare workers and environmental contamination. Thus, PoCUS, performed at the beside by the treating physician, may provide an accurate alternative.

Evidence
Pneumothorax. LUS has significantly higher sensitivity than CXR for the diagnosis of pneumothorax [79% versus 40%], whereas specificity is equally excellent [131]. However, most of these data are from trauma and postprocedural studies and may overestimate diagnostic performance of LUS in COVID-19. The negative predictive value of LUS for pneumothorax is approximately 100% (if pleural sliding, lung pulse and B or C patterns are observed) [132].
Ventilator-associated pneumonia. In the appropriate context, large consolidations not responsive to recruitment maneuvers or suction [133] are highly suggestive of secondary bacterial infection [127,134].
Diaphragmatic dysfunction, and weaning failure from mechanical ventilation. Ventilation-induced diaphragmatic injury can be reliably assessed with ultrasound [135]. Combining LUS score with the evaluation of LV and diaphragm function may improve the success of weaning trials [136][137][138][139]. Assessment of parasternal intercostal muscles thickening fraction seems promising for predicting weaning failure [140]. Detection and treatment of unresolved pulmonary conditions can facilitate weaning [141,142]. Acute cor-pulmonale. The effects of mechanical ventilation on RV function have been well-described. Acute cor-pulmonale becomes an important factor to be considered in the ventilation strategy [61,143].

24.
We suggest a prompt assessment of clinical deterioration with LUS for a timely and accurate bedside diagnosis of pneumothorax in severe COVID-19 (LQE II-B, Very Good Agreement). 25 We suggest LUS for early identification of ventilatorassociated pneumonia in severe COVID-19 (LQE II-B, Very Good Agreement). 26 We suggest multi-organ PoCUS over CXR and CT to assess readiness for weaning, predict success and diagnose the cause(s) of weaning failure in COVID-19 (LQE II-B, Very Good Agreement).

Limitations and future research
The safety and cost-saving impact of LUS in diagnosing complications of mechanical ventilation is yet to be demonstrated. A decision process based on PoCUS for tracheal extubation vs. tracheostomy mandates validation.

Evidence
FoCUS and echocardiography are recommended for hemodynamic monitoring in critical care [47,63,64]. A recent survey found that ultrasound is the most frequently used monitoring tool to assess cardiac output and pulmonary artery pressures in critical COVID-19 patients [108].

27.
We suggest FoCUS and/or echocardiography for hemodynamic monitoring in moderate-severe COVID-19 (LQE II-A, Very Good Agreement). 28 We suggest integrating PoCUS-derived information with data from other devices used for hemodynamic monitoring in severe COVID-19 (LQE II-B, Very Good Agreement).

Limitations and future research
Validated PoCUS-driven hemodynamic management protocols in COVID-19 are needed.

PoCUS FOR MONITORING OF OTHER ORGANS:
Many critically ill COVID-19 patients develop secondary organ dysfunction, including acute kidney injury (AKI), liver injury, rhabdomyolysis and gastrointestinal complications [118,144]. Hemodynamic factors and viral tropism for tubular cells may contribute to AKI [145]. Gastrointestinal complications may result from sepsis, deranged hemodynamics, or microvascular thrombosis [75]. Neurological complications are also not infrequent in COVID-19 [146].

Evidence
PoCUS can exclude post-and pre-renal causes of AKI (by assessing volume status and hemodynamics). It can detect systemic and renal venous congestion, important factors in AKI [147,148], acute gastrointestinal complications [149,150] including cholestasis and bowel ischemia in COVID-19 patients [151]. The use of PoCUS for the diagnosis and management of neurological conditions is acknowledged [152] and may be applicable in COVID-19.

29.
We suggest PoCUS assessment for pre-renal causes of AKI, including hemodynamics and venous congestion in COVID-19 (LQE II-B, Very Good Agreement).

Limitations and future research
Expertise and data on PoCUS applications to detect organ dysfunction in COVID-19 especially AKI and acute abdomen are limited and need further study.

DOMAIN 9-Infection control, PoCUS technique, technology, and protocols
In the context of COVID-19: • Interest in PoCUS has increased.
• Choice of machines is limited.
• Infection transmission to operators and environmental viral dissemination are serious concerns that may impact the quality of ultrasound examination and the choice of equipment. • A systematic scanning approach is required to avoid missing or misinterpreting important findings.

Evidence
Laptop/tablet/pocket-sized machines provide reasonable compromise between portability and capability [153] ( Fig. 6). Multi-frequency probes may be preferable to visualize both deep and superficial structures. While a single phased-array probe is suitable for FoCUS and LUS [154], a convex probe has been recommended by some experts [22]. Topographic zones and scanning techniques require standardization [12,22,30]. There is also a growing interest in telemedicine technology including robotic examinations [155] for remote guidance of minimally trained operators [156,157] [Additional file 14]. To protect healthcare workers and patients, stringent infection control practices are crucial. Available guidance deals with environmental transmission and spread to personnel [158]. Recommendations on disinfectants [159] and information on SARS-CoV-2 survival on fomites [160] are also available.

Limitations and future research
Information on quality, safety, remote mentoring/monitoring and archiving in COVID-19 is limited. Evidence for safety and efficacy of different disinfectants and methods of cleaning contaminated equipment is needed to make robust infection control policies.

Conclusion
This consensus document based on the available evidence and expert opinion should encourage the use of PoCUS to improve patient outcomes during the current pandemic and development of meaningful protocols and practices to overcome COVID-19 and prepare for future challenges.

Authors' contributions
AH and GV contributed equally as authors in conceiving the contents, gathering the relevant material, preparing the manuscript and chairing the steering committee of the process, LM GRADEd the evidence and supervised the Delphi consensus process as methodologist and reviewed the manuscript, AG conceived the evidence presentation and edited, GT contributed to the cardiovascular and hematological sections, LN contributed to concepts of triaging, TV contributed to LUS, FC and FM contributed to the ventilation section, RH and VN contributed to the manuscript, and YA conceived the idea of this work, provided guidance and edited the manuscript. All the authors participated in the Delphi process, provided input for drafting recommendations and reviewed the manuscript.

Funding
There were no financial disclosures specific to this work*.

Availability of data and materials
The data and other material can be made available to the Journal.

Ethics approval and consent to participate
There was no ethics approval required or applicable for this work.

Consent for publication
No material has been borrowed/reproduced from any other authors or publications.