2.1. Study Design and Participants

This study was approved by The Northwestern University Institutional Review Board (STU00207250 and STU00213616). Because this is a retrospective study, the need for patient consent for data collection was waived by the Institutional Review Board. This retrospective case series was performed at Northwestern University Medical Center in Chicago, Illinois. All consecutive patients who underwent lung transplantation for COVID-19-associated ARDS were enrolled between June 2020 and June 2022. Patients with a pre- and postoperative transthoracic echocardiogram beyond 30 days were included for analysis. Multiorgan transplant and single-lung transplant recipients were excluded from this study.

2.2. Data Collection

Clinical and laboratory characteristics, treatment, and outcomes data were obtained from electronic medical records by the investigative team and reviewed by thoracic surgeons and transplant pulmonologists. The information recorded included demographic data, medical history, underlying comorbidities, laboratory findings, medical course, treatments administered, pulmonary hemodynamics, and transthoracic echocardiography. In addition, intraoperative procedures and postoperative complications were recorded.

2.3. Lung Transplantation Evaluation and Listing in COVID-19-Associated ARDS

ARDS was defined according to the Berlin definition [14]. All patients with COVID-19-associated ARDS were treated by a multidisciplinary team that included surgeons, infectious disease physicians, pulmonary and critical care physicians, and cardiologists over the entire duration of the illness before being considered for transplantation. The decision to initiate extracorporeal membrane oxygenation (ECMO) was made by a multidisciplinary ECMO team, which included transplant pulmonologists, thoracic surgeons, ECMO specialists, and intensivists. During the pandemic’s peak, most patients were cannulated at various outside hospitals for refractory ARDS and transferred to our institution for inability to wean off ECMO and consideration for a lung transplant. Our practice was to apply veno-venous ECMO with RV support using a dual-stage right atrium to pulmonary artery cannula, the Protek-Duo (CardiacAssist, Inc.), due to severe pulmonary hypertension and RV dysfunction. We used this cannulation strategy because of the significant RV dysfunction in our patients with severe refractory ARDS and the additional potential for RV support, with flow directly to the pulmonary circulation. In addition, the single site cannulation strategy helped us to promote ambulation and rehabilitation in the intensive care unit [15].

A referral for lung transplantation was made when this multidisciplinary team concluded that there was no longitudinal evidence of lung recovery at least six weeks after the onset of COVID-19-associated ARDS, which is consistent with our previous study [7, 9, 10]. Lung transplant evaluation was performed according to the International Society for Heart and Lung Transplantation (ISHLT) guidelines, and all patients were discussed in a multidisciplinary meeting to proceed toward transplant listing [16]. The broad transplant criteria for patients with COVID-19-associated ARDS included people aged 70 years or younger, two consecutive lower respiratory fluid polymerase chain reaction tests negative for SARS-CoV-2, single-organ failure, no evidence of irrecoverable brain damage, and a body mass index less than or equal to 35. Before transplant listing, all patients with COVID-19-associated ARDS received pretransplant rehabilitation during hospitalization, achieving sufficient truncal strength to sit upright and move all four limbs against gravity.

2.4. Definition of Primary Graft Dysfunction

PGD was defined based on the ISHLT guideline [17], and graded by PaO₂/FiO₂ ratio as follows: Grade 1: PaO₂/FiO₂ ratio > 300; Grade 2: PaO₂/FiO₂ ratio is 200–300; Grade 3: PaO₂/FiO₂ ratio < 200. The use of ECMO for bilateral pulmonary edema on chest X-ray was classified as grade 3.

2.5. Echocardiographic Evaluation

Pre- and post-lung transplant echocardiographic (echo) data were reviewed by two independent, experienced level 3 trained cardiologists blinded to the study’s results. All of the traditional left ventricular (LV) and RV measurements were performed according to the conventional American Society of Echocardiography (ASE) guidelines. Two-dimensional speckle-tracking echocardiography was performed on cart and remeasured using an offline, vendor-independent analysis program (TomTec Imaging Systems, Munich, Germany). Measurements of the interventricular septum, LV internal diameter at end-diastole, LV internal diameter at end-systole, and posterior wall were performed from the parasternal long axis images. Left atrial (LA) volume index was calculated from apical 4-chamber and 2-chamber views, respectively, at end-systole when the LA chamber size is at its greatest dimension, taking care to avoid foreshortening. The right atrial (RA) area and volume were measured from the end ventricular end-systole when the RA chamber was at its greatest dimension. From an RV focused view, RV size and function were assessed. The basal RV diameter and the end-systolic and end-diastolic RV areas to assess fractional area change (FAC) were measured. RV wall thickness was measured from a subcostal view zoomed on the RV mid-wall at end-diastole. RV size was measured in RV focused apical 4-chamber view. In most patients, RV size could not be assessed in parasternal long axis view, proximal outflow tract parasternal long and short axis views. RV size was empirically graded based on a four-point scale using the basal RV diameter as follows: 1 = normal (25–40 mm), 2 = mildly dilated (41–45 mm), 3 = moderately dilated (46–50 mm), 4 = severely dilated (> 51 mm), while RV function was also evaluated on a four-point scale using FAC as a 2D surrogate for RV ejection fraction (1 = normal (> 35%), 2 = mildly decreased (30%–34%), 3 = moderately decreased (25%–29%), 4 = severely decreased < 25%). Apart from qualitative grading, RV function was also evaluated by tricuspid annular plane systolic excursion (TAPSE), TDI-derived tricuspid lateral annular systolic velocity (RV S′) measured in the RV-focused view, and RV free wall strain is described as follows: The hemodynamic right-sided assessment included the measurement of RV-RA gradient in calculating RV systolic pressure (RVSP) using a modified Bernoulli equation from the peak tricuspid regurgitant velocity signal. Right atrial pressure (RAP) assessment was performed using the diameter and collapsibility of the inferior vena cava from the subcostal view. The RAP was added to RVSP to obtain pulmonary artery systolic pressure (PASP). Mitral inflow parameters were measured, including E and A waves and medial and lateral e′ velocity. The E/e′ ratio was calculated to assess LV filling pressures.

Left ventricular global longitudinal strain was measured after the acquisition of LV focused views optimizing endocardial borders in the apical two-chamber, three-chamber, and four-chamber views at a frame rate of 60–90 Hz, taking care to minimize heart rate variation of no more than 5 beats between each view. RV free wall longitudinal strain was measured in the RV-focused view.

2.6. Statistical Analysis

The sample size was equal to the number of patients treated during the study period, and no statistical sample size was calculated. Continuous variables were reported as mean ± standard deviation or median with an interquartile range of 1–3. The Kaplan–Meier method was used to estimate survival differences between non-COVID-19- and COVID-19-associated ARDS patients. Statistical analysis was performed using EZR (Saitama Medical Center, Jichi Medical University, Japan), a GUI in R (The R Foundation for Statistical Computing, Vienna, Austria) [18]. Wilcoxon signed-rank test was used to compare pre- and postoperative TTE parameters. Statistical significance was denoted for variables with p < 0.05.

3.1. Study Population and Clinical Characteristics of Lung Transplant Recipient

Of 42 patients who underwent lung transplantation for COVID-19-associated ARDS, 36 were included for analysis. Six patients were excluded due to having either a single-lung transplant (n = 2), lobar transplant (n = 1), or dual-organ transplant (n = 1) or for missing postoperative TTE data (n = 2). Demographics and comorbidities of lung transplant recipients and donors can be found in Table 1. The median age of the recipient cohort was 51.1 ± 11.3 years old, 38.9% female, while the median age of the donor cohort was 30.8 ± 12.2 years old with 27.1% females (Table 1). Comorbidities of the transplant recipients included remote smoking history, hypertension, and diabetes. Three patients (8.3%) required dialysis during their critical illness; however, their renal function recovered and had a glomerular filtration rate > 50 mL/min/1.73 m² at the time of transplant listing. Twenty-four patients (66%) were bridged to transplantation with VV-ECMO cannulation, and the average time on ECMO until lung transplantation was 80.5 days (IQR 41.5–127.0). The remaining 12 patients had severe post-COVID pulmonary fibrosis as a sequelae of COVID-19-associated ARDS and were either unable to be weaned off the ventilator (n = 4) or high-flow oxygen (n = 8) for a meaningful recovery. Pulmonary artery pressures recorded in the operating room at the time of lung transplant with an in-dwelling pulmonary artery catheter showed significant pulmonary hypertension with an average PASP of 64.1 ± 27.1 mmHg, pulmonary artery diastolic pressure of 41.4 ± 22.8 mmHg, and mean pulmonary artery pressure of 49.4 ± 23.2 mmHg.

3.2. Intraoperative and Postoperative Outcomes

All patients in the study cohort underwent bilateral lung transplantation on Veno-Arterial ECMO due to the severity of pulmonary hypertension. The average operative time was 8.3 (7.8–9.6) hours, and the average VA-ECMO time was 3.2 (2.7–3.8) hours. Patients required substantial amounts of transfusion products due to coagulopathy related to adherent fibrotic lungs at the time of explant. Nine of the patients (25%) had PGD grade 3. Weaning off the vent was 20 (13.5–17.5) days. In the immediate postoperative days, 16 patients (45%) were electively kept on VV-ECMO post-transplantation. Eighteen patients developed acute kidney injury (51%), out of which, 8 patients required dialysis (23%). Eight patients (25%) were taken back to the operating room primarily for chest wall bleeding complications. Overall post-transplant ICU stay was 20 (13.5–26.5) days, and post-transplant hospital stay was 29 (17.5–38) days.

3.3. Echocardiographic Parameters

Preoperative echo was evaluated at an average of 15.5 days before surgery (IQR 10.0–34.3), and the postoperative echo was assessed at an average of 140 days postoperatively (IQR 108–201). There were no significant differences in LV linear dimensions or LV ejection fraction (LVEF) following lung transplantation in this cohort. LV size at end diastole, LVEF, and left atrial volume index (LAVi) demonstrated an improvement when comparing pretransplant to post-transplant echos. There were significant changes in parameters such as mitral A wave velocity and lateral e′ velocity from 0.76 ± 0.21 m/s to 0.6 ± 0.16 m/s (p < 0.05), and from 9.4 ± 3.0 m/s to 11.4 ± 3.1 m/s (p < 0.05), respectively (Table 2). There was a significant improvement in RV size and systolic function, as seen in Table 2 and Figure 1. From the pretransplant to post-transplant echo, RV size grade improved from an average of 1.7 ± 0.85 to 1.3 ± 0.6 (p < 0.05), while RV function improved from an average of 2.2 ± 1.2 to 1 ± 1 (p < 0.05). In addition, RVSP was reduced from 46.3 ± 18 mmHg to 30.1 ± 7.8 mmHg (p < 0.05). RV free wall strain was obtained only for 16 patients due to difficulty with image acquisition, given artifacts noted from mechanical ventilation and RV support devices. RV free wall strain showed a statistically significant improvement between the pre- and post-transplant echo, with an initial value of −13.9 ± 6.1% and a follow-up value of −18.5 ± 5.4% (p < 0.05). Similarly, left ventricular global longitudinal strain demonstrated an improvement from −16.6 ± 4.0% to −17.6 ± 2.9% (p < 0.05).

3.3.1. Explant Pathology

Explant histopathology assessment of patients with COVID-19-associated ARDS showed extensive alteration of the lung microstructure. There were several vascular changes consistent with severe pulmonary hypertension with resulting vascular remodeling. We selected a few explant histological slides to highlight the vascular remodeling changes including extensive microthrombi, capillary congestion and septal expansion, interlobular septal thickening, muscular hyperplasia, and hypertrophy (Figure 2). These findings overall signify the extent of vascular damage seen in COVID-19-associated ARDS.

3.4. Survival

The median follow-up period was 738 days (IQR 687–862) in the COVID-19-associated ARDS group vs 724 days (IQR 537–959) in the non-COVID group. As of July 1, 2023, 1-year survival rate of COVID-19-associated ARDS is 88.8% vs non-COVID 86.0% (p = 0.07) per Figure 3.

4.1. Limitations

This is a single-center, retrospective study, and the cohort sample size of 36 is small. There is a selection bias as only patients with interpretable echocardiograms pre- and post-transplant were included. Not all patients had optimal echocardiographic windows, which limited utility of all RV parameters, especially strain to assess for subclinical ventricular dysfunction. In addition, the post-transplant echo follow-up, on an average, was only 140 days. A longer duration of follow-up might demonstrate normalization of RV strain. The lack of validation cohort limits the outcome analysis. These patients were carefully selected and transplanted in a center experienced in managing these complex patients. Therefore, outcomes reported here might differ from national registries.