Ultrasound-Enhancing Agent Safely Enhances Left Ventricular Visualization by Transthoracic Echocardiography in Patients on ECMO Support.

OBJECTIVE: The authors investigated if the use of ultrasound-enhancing agents (UEA) can safely improve left ventricular (LV) image quality by transthoracic echocardiography (TTE) in patients on extracorporeal membrane oxygenation (ECMO). DESIGN AND SETTING: This study was performed in a tertiary cardiothoracic and ECMO center in London, United Kingdom. PARTICIPANTS: The authors included 18 prospectively identified consecutive patients requiring TEE supported on peripherally implanted ECMO. INTERVENTION AND MEASUREMENTS: TTE was performed before and after the UEA administration. The authors assessed the LV image quality using the biplane (apical-4-chamber and apical-2-chamber views) endocardial border definition index (1 = good, 2 = suboptimal, 3 = poor, and 4 = unavailable), as well as the feasibility of LV ejection fraction (LVEF) measurement. The authors also gathered sequential clinical information for the next 24 hours. MAIN RESULTS: The patients' median age was 47 years (35, 65), and 5 (28%) were women. The biplane endocardial border definition index improved from the suboptimal to the good range (2.167 [1.812, 3.042] v 1.500 [1.417, 1.792], p = 0.0004) after the use of UEA. The feasibility of LVEF tripled from 25% (n = 5) to 83% (n = 15) (p = 0.0008) with UEA use. The UEA did not set off the bubble alarm and did not impact clinical or ECMO parameters. CONCLUSION: The use of UEA significantly improved the quality of LV biplane images by transthoracic echocardiography, transforming them from suboptimal to good in patients supported with peripherally implanted ECMO. UEA use tripled the feasibility of measuring LVEF by TTE without affecting clinical and ECMO parameters.

Association Between Right Ventricular Systolic Parameters Measured by Echocardiography and Stroke Volume Derived From Pulmonary Artery Catheter in Ischemic Cardiogenic Shock.

OBJECTIVE: To determine the right ventricular (RV) systolic function echocardiographic parameter best associated with native stroke volume (SV) by thermodilution via a pulmonary artery catheter (PAC) in patients admitted to intensive care with ST elevation myocardial infarction (STEMI) complicated by cardiogenic shock (CS). DESIGN, SETTING, AND PARTICIPANTS: An observational cohort study of 43 prospectively identified patients admitted to a tertiary cardiac intensive care unit in London, United Kingdom. INTERVENTIONS: Simultaneous collection of comprehensive transthoracic echocardiographic, clinical, and PAC-derived hemodynamic data. Seven RV systolic function parameters were correlated with the PAC-derived SV. MEASUREMENTS AND MAIN RESULTS: The median patient age was 61 years (interquartile range [IQR], 52-67 years), and 36 of the 43 patients (84%) were male. The median PAC-derived SV and left ventricular ejection fraction were 57 mL (IQR, 39-70 mL) and 31% (IQR, 22%-35%), respectively. The RV outflow tract velocity time integral (RVOT VTI) and tricuspid plane systolic excursion (TAPSE) correlated significantly with the PAC-derived SV (r = 0.42 [p = 0.007] and r = 0.37 [p = 0.02], respectively). The RVOT VTI was independently associated with and predicted low PAC-derived SV (odds ratio, 1.3; p = 0.03) with a good area under the curve (AUC = 0.71; p = 0.02). An RVOT VTI <12.7 cm predicted low PAC-derived SV with a sensitivity of 66% and specificity of 72%. CONCLUSIONS: RVOT VTI is the echocardiographic RV systolic function parameter that best correlates with PAC-derived native SV in patients with STEMI complicated by CS. This parameter can help guide the hemodynamic management of this cohort.

Real-Time Ultrasound Guidance as Compared With Landmark Technique for Subclavian Central Venous Cannulation: A Systematic Review and Meta-Analysis With Trial Sequential Analysis.

OBJECTIVES: We conducted a systematic review and meta-analysis to assess the effectiveness of real-time dynamic ultrasound-guided subclavian vein cannulation as compared to landmark technique in adult patients. DATA SOURCES: PubMed and EMBASE until June 1, 2022, with the EMBASE search restricted to the last 5 years. STUDY SELECTION: We included randomized controlled trials (RCTs) comparing the two techniques (real-time ultrasound-guided vs landmark) for subclavian vein cannulation. The primary outcomes were overall success rate and complication rate, whereas secondary outcomes included success at first attempt, number of attempts, and access time. DATA EXTRACTION: Independent extraction by two authors according to prespecified criteria. DATA SYNTHESIS: After screening, six RCTs were included. Two further RCTs using a static ultrasound-guided approach and one prospective study were included in the sensitivity analyses. The results are presented in the form of risk ratio (RR) or mean difference (MD) with 95% CI. Real-time ultrasound guidance increased the overall success rate for subclavian vein cannulation as compared to landmark technique (RR = 1.14; [95% CI 1.06-1.23]; p = 0.0007; I2 = 55%; low certainty) and complication rates (RR = 0.32; [95% CI 0.22-0.47]; p < 0.00001; I2 = 0%; low certainty). Furthermore, ultrasound guidance increased the success rate at first attempt (RR = 1.32; [95% CI 1.14-1.54]; p = 0.0003; I2 = 0%; low certainty), reduced the total number of attempts (MD = -0.45 [95% CI -0.57 to -0.34]; p < 0.00001; I2 = 0%; low certainty), and access time (MD = -10.14 s; [95% CI -17.34 to -2.94]; p = 0.006; I2 = 77%; low certainty). The Trial Sequential Analyses on the investigated outcomes showed that the results were robust. The evidence for all outcomes was considered to be of low certainty. CONCLUSIONS: Real-time ultrasound-guided subclavian vein cannulation is safer and more efficient than a landmark approach. The findings seem robust although the evidence of low certainty.

EFSUMB Clinical Practice Guidelines for Point-of-Care Ultrasound: Part One (Common Heart and Pulmonary Applications) SHORT VERSION.

OBJECTIVE: To evaluate the evidence and produce a summary and recommendations for the most common heart and lung point-of-care ultrasound (PoCUS). METHODS: We reviewed 10 clinical domains/questions related to common heart and lung applications of PoCUS. Following review of the evidence, a summary and recommendations were produced, including assigning levels of evidence (LoE) and grading of recommendation, assessment, development, and evaluation (GRADE). 38 international experts, the expert review group (ERG), were invited to review the evidence presented for each question. A level of agreement of over 75 % was required to progress to the next section. The ERG then reviewed and indicated their level of agreement of the summary and recommendation for each question (using a 5-point Likert scale), which was approved in the case of a level of agreement of greater than 75 %. A level of agreement was defined as a summary of "strongly agree" and "agree" on the Likert scale responses. FINDINGS AND RECOMMENDATIONS: One question achieved a strong consensus for an assigned LoE of 3 and a weak GRADE recommendation (question 1), the remaining 9 questions achieved broad agreement with an assigned LoE of 4 and a weak GRADE recommendation (question 2), three achieved an LoE of 3 with a weak GRADE recommendation (questions 3-5), three achieved an LoE of 3 with a strong GRADE recommendation (questions 6-8) and the remaining two were assigned an LoE of 2 with a strong GRADE recommendation (questions 9 and 10). CONCLUSION: These consensus-derived recommendations should aid clinical practice and highlight areas of further research for PoCUS in acute settings.

EFSUMB Clinical Practice Guidelines for Point-of-Care Ultrasound: Part One (Common Heart and Pulmonary Applications) LONG VERSION.

AIMS: To evaluate the evidence and produce a summary and recommendations for the most common heart and lung applications of point-of-care ultrasound (PoCUS). METHODS: We reviewed 10 clinical domains/questions related to common heart and lung applications of PoCUS. Following review of the evidence, a summary and recommendation were produced, including assignment of levels of evidence (LoE) and grading of the recommendation, assessment, development, and evaluation (GRADE). 38 international experts, the expert review group (ERG), were invited to review the evidence presented for each question. A level of agreement of over 75 % was required to progress to the next section. The ERG then reviewed and indicated their level of agreement regarding the summary and recommendation for each question (using a 5-point Likert scale), which was approved if a level of agreement of greater than 75 % was reached. A level of agreement was defined as a summary of "strongly agree" and "agree" on the Likert scale responses. FINDINGS AND RECOMMENDATIONS: One question achieved a strong consensus for an assigned LoE of 3 and a weak GRADE recommendation (question 1). The remaining 9 questions achieved broad agreement with one assigned an LoE of 4 and weak GRADE recommendation (question 2), three achieving an LoE of 3 with a weak GRADE recommendation (questions 3-5), three achieved an LoE of 3 with a strong GRADE recommendation (questions 6-8), and the remaining two were assigned an LoE of 2 with a strong GRADE recommendation (questions 9 and 10). CONCLUSION: These consensus-derived recommendations should aid clinical practice and highlight areas of further research for PoCUS in acute settings.

Left Ventricular Ejection Fraction Correlation With Stroke Volume as Estimated by Doppler Echocardiography in Cardiogenic Shock: A Retrospective Observational Study.

OBJECTIVES: Echocardiography is the main tool for cardiac assessment and helps to guide management in patients admitted to the intensive care unit (ICU) with cardiogenic shock (CS). Left ventricular ejection fraction (LVEF) is a commonly used echocardiographic surrogate for left ventricular (LV) systolic function. In this hypothesis-generating study, the authors investigated the correlation between LVEF and stroke volume (SV)/SV index (SVI) estimated by Doppler echocardiography in patients admitted to the ICU with CS and reduced LVEF. DESIGN AND SETTING: This retrospective analysis was performed in a single tertiary cardiac center in London, United Kingdom. PARTICIPANTS: Patients admitted to the ICU over a 34-month period with ST elevation myocardial infarction (STEMI) complicated by CS and LVEF <40%. INTERVENTIONS: Clinical and echocardiographic data were collected. LVEF (total, by visual estimate and by modified Simpson's method) was correlated with SV and SVI estimated by Doppler echocardiography. MEASUREMENTS AND MAIN RESULTS: One-hundred patients were included. The mean age was 62.6 ± 12.7 years and 78% were male patients. The median LVEF was 29% (20-35) and the mean SV and SVI by Doppler echocardiography were 47 mL ± 16 and 25 mL ± 9, respectively. Analysis revealed a weak yet statistically significant correlation among LVEF and Doppler SV and SVI (r = 0.44, confidence interval [CI] 0.26-0.60, p < 0.0001, r = 0.47, CI 0.28-0.62, p < 0.001, respectively). Visually estimated LVEF (n = 74) correlated weakly with Doppler SV and SVI yet better compared to LVEF by Simpson's method (n = 25) (r = 0.48, CI 0.27-0.65, p < 0.0001 and r = 0.49, CI 0.28-0.66, p < 0.0001, respectively, v r = 0.36, CI 0.05-0.67, p = 0.08 and r = 0.37, CI -0.04-0.67, p 0.07, respectively). CONCLUSIONS: There is a weak correlation between LVEF and SV or SVI estimated by Doppler echocardiography in patients admitted to the ICU with STEMI complicated by CS and reduced LVEF. Visually estimated LVEF correlated slightly better with Doppler SV compared to modified Simpson's LVEF.

Training and Accreditation Pathways in Critical Care and Perioperative Echocardiography.

In recent years, there has been a dramatic rise in the use of echocardiography in perioperative and critical care medicine. It is now recommended widely as a first-line diagnostic tool when approaching patients in shock, due to its ability to identify shock etiology and response to treatment noninvasively. To match the increasing training demand, and to ensure maintenance of high-quality and reproducible scanning, several accreditation pathways have been developed worldwide. Critical care echocardiography training pathways can include both transthoracic and transesophageal scanning and range from basic focused protocols to advanced comprehensive scans. The complexity of each individual training program is reflected in its accreditation process. While basic accreditations may require only course attendance and a relatively brief logbook, comprehensive accreditations often require extensive logbooks and written and practical examinations. Currently, the use of transoesophageal echocardiography remains restricted largely to cardiothoracic anesthesia and critical care; however, its use in the general intensive care setting is growing. This narrative review summarizes the most common training pathways, their curricula, and accreditation requirements. The authors initially focus on echocardiography training in the United Kingdom, and then go on to describe similar international accreditations available in Europe, North America, and Australasia.

Utility of the Venous Excess Ultrasound (VEXUS) score to track dynamic change in volume status in patients undergoing fluid removal during haemodialysis - the ACUVEX study.

BACKGROUND: The use of ultrasound assessment, including the Venous Excess Ultrasound (VEXUS) score, is increasingly being utilised as part of fluid status assessment in clinical practice. We aimed to evaluate the ability of the VEXUS score to track fluid removal during the course of the dialysis session and explore the relationship between traditional measures of fluid status and venous congestion. METHODS: Single-centre, observational study in patients undergoing intermittent haemodialysis, who presented above their target dry weight. Patients had serial assessment using VEXUS, lung ultrasound and selected echocardiographic measures, before, during and after fluid removal. RESULTS: Amongst 33 patients analysed, 5 (15%) had an elevated VEXUS score (> 0). There was no difference in starting weight, dry weight or amount of fluid removed in patients with a normal VEXUS score and those with an elevated VEXUS score. In all patients with elevated VEXUS scores, the degree of venous congestion improved during the course of fluid removal. All patients with an elevated VEXUS score had evidence of both right and left ventricular systolic impairment. CONCLUSION: In patients with ESRF undergoing haemodialysis, the incidence of venous congestion as measured by the VEXUS is low. In patients with elevated VEXUS scores, removal of fluid through haemodialysis improves the venous congestion score. The pattern of LV and RV systolic dysfunction suggests that VEXUS may be a reflection of cardiac failure rather than venous volume status. TRIAL REGISTRATION: Ethical approval was provided by South Central-Berkshire Research and Ethics Committee and registered on clinicaltrials.org (IRAS305720). TRIAL REGISTRATION: ISRCTN14351189 - Retrospectively registered on 30/11/2023.

The use of echocardiography in the management of shock in critical care: a prospective, multi-centre, observational study.

PURPOSE: Echocardiography is recommended as a first-line tool in the assessment of patients with shock. The current provision of echocardiography in critical care is poorly defined. The aims of this work were to evaluate the utilisation of echocardiography in patients presenting to critical care with shock, its impact on decision making, and adherence to governance guidelines. METHODS: We conducted a prospective, multi-centre, observational study in 178 critical care units across the United Kingdom (UK) and Crown Dependencies, led by the UK's Trainee Research in Intensive Care Network. Consecutive adult patients (≥ 18 years) admitted with shock were followed up for 72 h to ascertain whether they received an echocardiogram, the nature of any scan performed, and its effect on critical treatment decision making. RESULTS: 1015 patients with shock were included. An echocardiogram was performed on 545 (54%) patients within 72 h and 436 (43%) within 24 h of admission. Most scans were performed by the critical care team (n = 314, 58%). Echocardiography was reported to either reduce diagnostic uncertainty or change management in 291 (54%) cases. Patients with obstructive or cardiogenic shock had their management altered numerically more often by echocardiography (n = 15 [75%] and n = 100 [58%] respectively). Twenty-five percent of echocardiograms performed adhered to current national governance and image storage guidance. CONCLUSION: Use of echocardiography in the assessment of patients with shock remains heterogenous. When echocardiography is used, it improves diagnostic certainty or changes management in most patients. Future research should explore barriers to increasing use of echocardiography in assessing patients presenting with shock.

Right ventricular echocardiographic parameters and prediction of stroke volume in ischemic cardiogenic shock: A retrospective study.

PURPOSE: This study investigated which commonly used right ventricular (RV) echocardiographic parameter correlates best with stroke volume (SV) estimated by Doppler echocardiography in ischemic cardiogenic shock (CS). MATERIALS AND METHODS: We retrospectively reviewed the records of 100 patients admitted to the ICU over 34 months with CS. Tricuspid annular plane systolic excursion (TAPSE), Tricuspid annulus systolic velocity (RV S'), Tricuspid regurgitation maximum velocity (TR Vmax), and RV outflow tract velocity time integral (RVOT VTI) were correlated to SV. RESULTS: Mean age was 62.6 ± 12.7 years and 78% were male. The mean SV, TAPSE, RV S', TR Vmax, and RVOT VTI were 47 ± 16 ml, 16 ± 5 mm, 11 ± 4 mm/s, 1.97 ± 0.73 m/s, and 12.7 ± 5 cm, respectively. RVOT VTI correlated best to SV (r = 0.39 p = 0.01) compared to TAPSE, RV S', and TR Vmax (r = 0.26 p = 0.01, r = 0.15 p = 0.21, r = 0.03 p = 0.78). RVOT VTI independently predicted SV. Univariate analysis demonstrated that only RVOT VTI predicted SV (OD = 1.18 p = 0.04) and had the best area under the curve (0.70, p = 0.03). CONCLUSION: RVOT VTI correlated better (albeit weakly) to and best predicted SV compared to TAPSE, RV S', and TR Vmax in patients admitted to intensive care with CS. This study suggests that RVOT VTI has the potential as a therapeutic target to optimize SV in CS.

The Evolution of Cardiovascular Ultrasound: A Review of Cardiac Point-of-Care Ultrasound (POCUS) Across Specialties.

The use of cardiac point-of-care ultrasound (POCUS) is now widespread in clinics, emergency departments, and all areas of the hospital. Users include medical trainees, advanced practice practitioners, and attending physicians in many specialties and sub-specialties. Opportunities to learn cardiac POCUS and requirements for training vary across specialties, as does the scope of the cardiac POCUS examination. In this review, we describe both a brief history of how cardiac POCUS emerged from echocardiography and the state of the art across a variety of medical fields.

Right ventricular function is associated with 28-day mortality in myocardial infarction complicated by cardiogenic shock: A retrospective observational study.

Background: Right ventricular (RV) function is increasingly being recognised as an important factor influencing outcomes in ST elevation myocardial infarction (STEMI) complicated by cardiogenic shock (CS). In this study, we investigated RV echocardiographic parameters' association with 28-day mortality in patients admitted to intensive care with STEMI complicated by CS with reduced left ventricle ejection fraction (LVEF). Method: We performed a retrospective analysis of patients admitted to intensive care unit (ICU) in a single tertiary cardiac centre over a 34-month period with STEMI complicated by CS and LVEF < 40%. Clinical and echocardiographic data were collected and correlated with 28-day mortality. Results: One-hundred patients were included with a mean age of 62.6 ±12.7 years and 78% were male. Mortality at 28 days was 37%. Respectively, 85%, 40% and 25% of patients required mechanical ventilation, mechanical circulatory support and renal replacement therapy. Tricuspid annulus peak systolic velocity (RV S') was significantly higher in survivors (12 ± 3.3 v 10 ± 3.5 cm/s, p = 0.03) and was an independent predictor of mortality (odds ratio 1.2, 95% confidence interval 1.1-1.4, p = 0.04). RV S' of 10.5 cm/s exhibited best sensitivity and specificity (64% and 65%, respectively; p = 0.02) for mortality. The Kaplan-Meier curve demonstrated 85% risk of 28-day mortality for RV S' < 10.5 cm/s v 53% for RV S' > 10.5 cm/s (p = 0.02). Conclusion: RV function is associated 28-day mortality in patients admitted to ICU with STEMI complicated by CS with reduced LVEF. RV S' predicted mortality with good sensitivity and specificity.

Impact of unacceptable behaviour between healthcare workers on clinical performance and patient outcomes: a systematic review.

BACKGROUND: Recent studies suggest that displays of unacceptable behaviour, including bullying, discrimination and harassment, between healthcare workers (HCWs) may impair job performance, and in turn, increase the frequency of medical errors, adverse events and healthcare-related complications. The objective of this systematic review was to summarise the current evidence of the impact of unacceptable behaviour occurring between HCWs on clinical performance and patient outcomes. METHODS: We searched MEDLINE, Embase, PsycINFO and CINAHL from 1 January 1990 to 31 March 2021. The search results were screened by two independent reviewers and studies were included if they were original research that assessed the effects of unacceptable behaviour on clinical performance, quality of care, workplace productivity or patient outcomes. Risk of bias was assessed using tools relevant to the study design and the data were synthesised without meta-analysis. RESULTS: From the 2559 screened studies, 36 studies were included: 22 survey-based studies, 4 qualitative studies, 3 mixed-methods studies, 4 simulation-based randomised controlled trials (RCTs) and 3 other study designs. Most survey-based studies were low quality and demonstrated that HCWs perceived a relationship between unacceptable behaviour and worse clinical performance and patient outcomes. This was supported by a smaller number of higher quality retrospective studies and RCTs. Two of four RCTs produced negative results, possibly reflecting inadequate power or study design limitations. No study demonstrated any beneficial effect of unacceptable behaviour on the study outcomes. CONCLUSIONS: Despite the mixed quality of evidence and some inconsistencies in the strengths of associations reported, the overall weight of evidence shows that unacceptable behaviour negatively affects the clinical performance of HCWs, quality of care, workplace productivity and patient outcomes. Future research should focus on the evaluation and implementation of interventions that reduce the frequency of these behaviours.

The use of critical care echocardiography in peri-arrest and cardiac arrest scenarios: Pros, cons and what the future holds.

Echocardiography is being increasingly deployed as a diagnostic and monitoring tool in the critically ill. This rise in popularity has led to its recommendation as a core competence in intensive care, with several training routes available. In the peri-arrest and cardiac arrest population, point of care focused echocardiography has the potential to transform patient care and improve outcomes. Be it via diagnosis of shock aetiology and reversibility or assessing response to treatment and prognostication. This narrative review discusses current and future applications of echocardiography in this patient group and provides a structure with which one can approach such patients.

Staff wellbeing in times of COVID-19.

In the last 10 years, there has been increasing interest into the psychological wellbeing of healthcare providers. Within critical care, increasing attention is being paid to the concept of 'burnout' - a cluster of symptoms that adversely affect the health of critical care providers. Publications and statements from the major critical care societies have all addressed this syndrome and emphasised urgency in tackling it. The current COVID-19 pandemic has fundamentally changed the way we work, communicate and learn. Even before the pandemic, there have been growing concerns and acknowledgement that healthcare practitioners in intensive care are at increased risk of burnout and burnout syndrome. There has never been greater pressure on intensive care or indeed healthcare as a whole to look after so many patients during this pandemic and yet there is global acknowledgement that key to overcoming these challenges is to look after the care providers - both physically and psychologically. In this paper, we review the issue of burnout amongst healthcare practitioners during current pandemic. We present the impact of burnout on the individual and the system as a whole but perhaps most importantly, we provide a review of steps being taken to mitigate against these adverse outcomes in the short and longer term.

The use of social media and online communications in times of pandemic COVID-19.

The use of social media as a tool for professional communication and education in healthcare has been increasing; pros and cons of such platforms were extensively debated in recent years with mixed results. During the COVID-19 pandemic, social media use has accelerated to the point of becoming a ubiquitous part of modern healthcare systems. As with any tool in healthcare, its risks and benefits need to be carefully considered. In this article, we review the use of social media in the current pandemic. Importantly, we will illustrate this using experiences from the perspective of large medical organisations and also identify the common pitfalls.

SARS-CoV-2 and ECMO: early results and experience.

INTRODUCTION: In this paper, we describe our experience and early outcomes with critically unwell severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) patients who required extracorporeal membrane oxygenation (ECMO). We present our standard practices around ECMO decision-making, retrieval, cannulation, ventilation, anticoagulation, tracheostomy, imaging and steroids. METHODS: A retrospective cohort study using data from the hospital notes on all SARS-CoV-2 patients who required extracorporeal support at St Bartholomew's Hospital between 1 March 2020 and 31 July 2020. In total, this included 18 patients over this time period. RESULTS: In total, 18 patients were managed with extracorporeal support and of these 14 survived (78%) with 4 deaths (22%). The mean duration from hospital admission to intubation was 4.1 ± 3.4 days, mean time from intubation to ECMO 2.3 ± 2 days and mean run on ECMO 17.7 ± 9.4 days. Survivor mean days from intubation to extubation was 20.6 ± 9.9 days and survivor mean days from intubation to tracheostomy decannulation 46.6 ± 15.3 days. Time from hospital admission to discharge in survivors was a mean of 57.2 ± 25.8 days. Of the patients requiring extracorporeal support, the initial mode was veno-venous (VV) in 15 (83%), veno-arterial (VA) in 2 (11%) and veno-venous-arterial (VVA) in 1 (6%). On VV extracorporeal support, 2 (11%) required additional VVA. Renal replacement therapy was required in 10 (56%) of the patients. Anticoagulation target anti-Xa of 0.2-0.4 was set, with 10 (56%) patients having a deep vein thrombosis or pulmonary embolism detected and 2 (11%) patients suffering an intracranial haemorrhage. Tracheostomy was performed in 9 (50%) of the patients and high-dose methylprednisolone was given to 7 (39%) of the patients. CONCLUSION: In our cohort of patients with severe SARS-CoV-2 respiratory failure, a long period of invasive ventilation and extracorporeal support was required but achieving good outcomes despite this. There was a significant burden of thromboembolic disease and renal injury. A significant proportion of patients required tracheostomy and steroids to facilitate weaning.