Right Ventricular Dysfunction in Patients with Coronavirus Disease 2019 Supported with Extracorporeal Membrane Oxygenation.

To the Editor:

The development of right ventricular dysfunction (RVD) while on venovenous (VV) extracorporeal membrane oxygenation (ECMO) is not well-understood. Before the pandemic, single-center reports have described RVD in patients with acute respiratory distress syndrome (ARDS) on VV ECMO, while others identified improvement in RV function after ECMO initiation.1, 2, 3 Although early reports of patients with coronavirus disease 2019 (COVID-19) ARDS identified echocardiography data in right ventricular (RV) dilation and RVD in 40% and 27%, respectively, the prevalence and severity of RVD in patients with COVID-19 requiring ECMO therapy are unknown. Given this knowledge gap, we performed a pilot study evaluating RV function in patients with severe ARDS due to COVID-19 supported with conventionally cannulated, femoro femoral VV ECMO.

Following Institutional Review Board approval, we analyzed point-of-care echocardiographic data acquired from adult patients with COVID-19 on ECMO during clinical care in the cardiovascular intensive care unit between April 2020 and October 2020. A point-of-care examination was triggered by hemodynamic instability and/or refractory hypoxemia. Two-dimensional (2-D) examinations were performed in accordance with American Society of Echocardiography guidelines. Tricuspid annular plane systolic excursion (TAPSE), tricuspid lateral annular systolic velocity (S’), and end-diastolic diameter (EDD) measurements were obtained. Images with frame rates >30 frames/s subsequently were exported for offline analysis of free-wall longitudinal strain (FWLS) and fractional area change (FAC) with TomTec 2-D CPA software (TomTec Imaging Systems, Unterschleissheim, Germany). Summary statistics are reported as frequency (percentage) for categorical variables and as mean (standard deviation [SD]) or median [interquartile range] for continuous variables.

A total of 11 consecutive patients treated between April and October 2020 were analyzed. Table 1 presents demographic and clinical variables, and Table 2 presents the echocardiographic findings. The mean (SD) time to initial evaluation occurred at 9.17 (7.44) days of ECMO support, prompted by clinical deterioration. The majority of patients (seven, [63.6%]) had abnormal RV size, with mean (SD) EDD values of 4.47 (0.69) cm and 3.64 (0.83) cm at the base and mid chamber, respectively. TAPSE mean (SD) of 2.15 (0.65) cm and S’ of 13.8 (4.69) cm/s were normal for the majority of examinations. However, FWLS and FAC measurements were abnormal in nine patients (81.8%) and ten patients (90.9%), with a mean (SD) of –16.37% (5.97%) and 22.61% (6.2%), respectively. RVD causing clinical instability requiring inotropic support was present in five patients (45.5%).

Table 1

Demographic and Clinical Variables

Variables	Results (n = 11)
Age, y, mean (SD)	49.55 (8.96)
Sex, n (%)
Male	6 (54.5%)
Female	5 (45.5%)
Race, n (%)
White	4 (36.4%)
Black	2 (18.2%)
Asian	2 (18.2%)
Other	3 (27.3%)
Ethnicity, Hispanic, n (%)	3 (27.3%)
BMI, kg/m2, mean (SD)	35.1 (7.6)
Comorbidities, n (%)
Obesity	9 (81.8%)
Type 2 Diabetes Mellitus	6 (54.5%)
Hypertension	4 (36.4%)
Chronic kidney disease	3 (27.3%)
Neurologic disorder	2 (18.2%)
COPD	1 (9.1%)
Asthma	1 (9.1%)
Connective tissue disease	1 (9.1%)
HFrEF	0 (0%)
Chronic liver disease	0 (0%)
Cancer	0 (0%)
Pulmonary fibrosis	0 (0%)
Length of stay, d, mean (SD)	31.8 (17.3)
Length of ICU stay, d, mean (SD)	18.5 (10.7)
ECMO length, d, mean (SD)	20.92 (9.90)
Average daily respiratory variables per patient from ECMO start to first exam, median [IQR]
PaO2, mmHg	72.5 [61.7-77.0]
FIO2, %	60.0 [48.8-55.0]
PaO2/FiO2 ratio, mmHg	131.1 [109.2-165.0]
Tidal volume, mL	365.3 [284.4-425.0]
Plateau pressure, cm H2O	33 [31.7-35.2]
PEEP, cm H2O	12.8 [12.0-17.1]
Variables on first exam day, median [IQR]
Lowest pH, units	7.34 [7.3-7.4]
Lowest PaO2, mmHg	76.2 [56.6-117.4]
Highest PCO2, mEq/L	54.6 [48.0-59.5]
Highest bicarbonate, mmol/L	28.9 [24-31.2]
Lowest SaO2, %	85.9 [83.5-93.1]
Average variables from ECMO start to first exam date, median [IQR]
ECMO highest flow, L/min, median [IQR]	4.66 [4.33-4.79]
ECMO highest SWEEP, L/min, median [IQR]	4.5 [3.64-5.50]
Length ECMO flow >5 L/min, h	14.88 [7.0-18.44]
SOFA, units	9.0 [5.56-9.67]
D-Dimer, µg/mL	5.13 [3.33-9.83]
Ferritin, ng/mL	754.62 [138.25-1468.7]
Highest peak inspiratory pressure	31.06 [28.67-41.3]
Lowest SaO2, %	91.2 [87.3-91.95]
Complications, n (%)
Delirium/Agitation	11 (100%)
AKI	9 (81.8%)
Bleeding/thrombosis	8 (72.7%)
RV failure requiring inotropic support	5 (45.5%)
RRT requirement	4 (36.4%)
Septic Shock	4 (36.4%)
Liver function abnormality	3 (27.3%)
LV dysfunction	1 (9.1%)
Pulmonary embolism	0 (0%)

SD, standard deviation; BMI, body mass index; COPD, chronic obstructive pulmonary disease; HFrEF, heart failure with reduced ejection fraction; ECMO, extracorporeal membrane oxygenation; IQR, interquartile range; ICU, intensive care unit; SOFA, sequential organ failure assessment; AKI, acute kidney injury; RV, right ventricle; RRT, renal replacement therapy; LV, left ventricle.

Table 2

Point-of-Care Echocardiographic Findings

Variables	N	Results
Number of patients		11
Total number of echocardiographic exams		11
Number of days to exam after ECMO start, mean (SD)		9.17 (7.44)
Echocardiographic exam results, mean (SD)
FWLS (%)	11	–16.37 (5.97)
FAC (%)	11	22.61 (6.2)
EDD basal (cm)	11	4.47 (0.69)
EDD mid chamber (cm)	11	3.64 (0.83)
TAPSE (cm)	8	2.15 (0.65)
S' (cm/s)	8	13.8 (4.69)
Incidence of abnormal measurements during ECMO, n (%)
FWLS > –20%	11	9 (81.8%)
FAC <35%	11	10 (90.9%)
EDD basal > 4.2 cm	11	6 (54.5%)
EDD mid cavity > 3.5 cm	11	7 (63.6%)
TAPSE < 1.7 cm	8	3 (37.5%)
S' < 9.5 cm/s	8	0 (0%)
Number of patients with abnormal results, n (%)
FWLS	11	9 (81.81%)
FAC	11	9 (81.81%)
Size	11	7 (63.6%)
TAPSE	11	3 (27.27%)
S'	11	0 (0%)

ECMO, extracorporeal membrane oxygenation; SD, standard deviation; FAC, fractional area change; FWLS, free-wall longitudinal strain; EDD, end-diastolic diameter; TAPSE, tricuspid annular plane systolic excursion; S', tricuspid lateral annular systolic velocity; RV, right ventricle.

RVD is a well-described complication of ARDS, with incidence varying between 22% and 50%. Current evidence suggests similar rates in COVID-19 ARDS. Patients with ARDS requiring VV ECMO do not appear immune to this complication, with prepandemic studies reporting RVD in 18%-34% in this cohort. , Our data further showed that a substantial number of patients with severe ARDS due to COVID-19 may develop clinically significant RVD while being supported with femoro femoral VV ECMO. RV enlargement, abnormal myocardial free wall strain, and abnormal FAC were found in the majority of patients. Additionally, just under half of our patients experienced clinically significant RVD, defined as hemodynamic instability with echocardiographic stigmata of RVD and negative workup for other etiologies. While an improvement in RV function previously has been noted following the initiation of VV ECMO, our findings suggested this effect may be short-lived. On average, RVD was found nine days into ECMO therapy.

Based on our results, RVD complicating ECMO support in patients with COVID-19 ARDS may be more ubiquitous than previously observed in other ARDS ECMO cohorts. Unique COVID-19 characteristics may be responsible for these findings. Pulmonary microthrombi, endothelial injury, lung consolidation, iatrogenic therapies, and patient-specific characteristics such as obesity are just some factors that may play a role in RVD development.

Our findings have important implications for patient care. Some centers around the U.S. preemptively changed their ECMO cannulation practices during the pandemic to promote RV protection, and results have been encouraging. , Our results provide additional data potentially supporting the use of dual-staged cannulae in this population. Furthermore, our findings suggest that echocardiographic evaluation of RV function in patients with COVID-19 ARDS on ECMO should include FAC and FWLS because TAPSE and S’ may not be accurate.

Our study had significant limitations including retrospective nature and small sample size. In addition, the analytical utility of point-of-care echocardiography in FWLS acquisition remains uncertain, and the lack of echocardiograms before ECMO cannulation may have missed preexisting RVD in this population. Further longitudinal studies are needed to fully elucidate the effects of COVID-19 and ECMO on the right ventricle over time. Moreover, describing regional strain at the free wall base, mid-chamber, and apex would help pinpoint the pathophysiology of the dysfunction. Future studies also could compare different ECMO cannulation strategies in this at-risk population

In summary, RVD may develop in a significant number of patients with COVID-19 ARDS supported on ECMO. These findings have significant clinical implications and add to the body of evidence supporting alternative cannulation modalities and use of FWLS and FAC as descriptors of RV function in this patient population.