Extracorporeal life support bridge for pulmonary hypertension: A high-volume single-center experience

ARTICLE IN PRESS

http://www.jhltonline.org

ORIGINAL CLINICAL SCIENCE

Extracorporeal life support bridge for pulmonary hypertension: A high-volume single-center experience Erika B. Rosenzweig, MD,a,b Whitney D. Gannon, MSN, MS,c Purnema Madahar, MD, MS,b Cara Agerstrand, MD,b Darryl Abrams, MD,b Peter Liou, MD,d Daniel Brodie, MD,b,1 and Matthew Bacchetta, MDe,1 From the aDepartment of Pediatrics, Columbia University Medical Center, New York Presbyterian Hospital, New York, New York; bDepartment of Medicine, Columbia University Medical Center, New York Presbyterian Hospital, New York, New York; cDepartment of Medicine, Vanderbilt University Medical Center, Nashville, Tennessee; dDepartment of Surgery, Columbia University Medical Center, New York Presbyterian Hospital, New York, New York; and the eDepartment of Thoracic & Cardiac Surgery, Vanderbilt University Medical Center, Nashville, Tennessee, and Biomedical Engineering, Columbia University Medical Center, New York, New York

KEYWORDS: ECLS; pulmonary hypertension; ECMO; pulmonary arterial hypertension; lung transplantation

1

BACKGROUND: Application of extracorporeal life support (ECLS) for advanced pulmonary hypertension (PH) is evolving and may be deployed as a bridge to transplantation (BTT) or in one of several non-BTT uses, such as bridge to recovery (BTR) to the chronic PH clinical state in the setting of an acute PH trigger, bridge through non-transplant surgery (BTNTS), or bridge post-transplantation (BPT). METHODS: We conducted a retrospective analysis of all adult patients with World Symposium on Pulmonary Hypertension Group 1, 3, 4, or 5 PH who received ECLS at Columbia University Medical Center/New York Presbyterian Hospital between January 1, 2010 and August 18, 2018. We describe patient characteristics, outcomes, and our approach to medical and surgical management of these patients. RESULTS: There were 98 patients with significant PH in the cohort (54 female; median age, 48 years [interquartile range, 32−58]). Of these, 44 (45%) patients with PH received ECLS as non-BTT with intent to recover back to their baseline functional state, optimize therapy, or support through a definitive surgery, including 19 BTR, 17 BTNTS, and 8 BPT, and 54 (55%) patients received ECLS as BTT. In the overall cohort, 67 (68.4%) patients received venoarterial ECLS and 31 (31.6%) received venovenous (VV) ECLS. Out of 83 patients, 52 (63%) were liberated from invasive mechanical ventilation, and 85.2% of BTT patients with PH ambulated while on ECLS. Management of PH medications was individualized, often requiring titration with use of inhaled pulmonary vasodilators increased after cannulation in non-BTT. Overall 30-day survival was 73.5%, survival to ECLS decannulation was 66.3%, and survival to hospital discharge was 54.1%. All 8 BPT patients (100%) survived to hospital discharge, 64.7% of BTNTS patients survived to hospital discharge, and 32 (59.3%) BTT patients survived to lung transplantation. Early-era use of VV-ECLS for BTT had worse survival to discharge than those initially configured with venoarterial ECLS, impacting the overall survival and leading to limited use of VV-ECLS in the current era for BPT, BTNTS, and select BTR cases.

These authors have contributed equally to this work. Reprint requests: Erika B. Rosenzweig, MD, Pulmonary Hypertension Comprehensive Care Center, Columbia University Medical Center, New York Presbyterian, CH-2N, 3959 Broadway, New York, New York, 10032.

Telephone: +1-212-305-4436. Fax: +1-212-342-1443. E-mail address: [email protected]

1053-2498/$ - see front matter Ó 2019 International Society for Heart and Lung Transplantation. All rights reserved. https://doi.org/10.1016/j.healun.2019.09.004

2

ARTICLE IN PRESS

The Journal of Heart and Lung Transplantation, Vol 00, No 00, Month 2019

CONCLUSIONS: ECLS instituted by a specialized, multidisciplinary team has a role in the management of advanced PH as BTT or as non-BTT (including BTR, BTNTS, and BPT). Careful selection of ECLS cannulation configurations, patient-specific optimization of PH medical therapies, and avoidance of endotracheal intubation may be effective strategies in managing these complex patients. J Heart Lung Transplant 000;000:1−11 Ó 2019 International Society for Heart and Lung Transplantation. All rights reserved.

Pulmonary hypertension (PH) is a progressive disease that accompanies a variety of systemic disorders and leads to right heart failure and death if not treated. Despite major advances in medical treatment of pulmonary arterial hypertension (PAH) and improved overall outcomes, there is no medical cure aside from lung transplantation for end-stage PAH and pulmonary thromboendarterectomy (PTE) for chronic thromboembolic pulmonary hypertension (CTEPH). Even with advances in novel targeted medical therapy for PAH, patients with all forms of PH can experience progression of their symptoms, leading to worsening right heart failure gradually or suddenly because of a potentially treatable acute insult such as a pneumonia. Although extracorporeal life support (ECLS) was first described for use in PH in neonates with persistent PH of the newborn,1,2 it was nearly 3 decades before it was considered for other forms of adult PH. A key barrier for use of ECLS in adult patients with PH was that, unlike persistent PH of the newborn, most forms of adult PH are irreversible and successful ECLS cannulation and separation from ECLS were considered improbable. Further, induction with general anesthesia for cannulation of patients with PH was considered high-risk and potentially catastrophic. However, with recent advances in PAH therapies and ECLS technology, the notion of bridging acutely decompensated PH patients back to their baseline clinical state and through non-transplant and lung transplant surgeries has become more feasible.3−6 The ability to avoid intubation during cannulation and prolonged mechanical ventilation in some of these patients by using ECLS was embraced as a viable therapeutic option in these high-risk patients with PH.7,8 Thus, a growing number of patients with PH are being bridged with ECLS before and after lung transplantation, using ECLS configuration strategies designed to optimize durable physiological support to enable patients to be awake and ambulatory.3,6−10 However, ECLS experience and data are limited to relatively small numbers of patients with PH included in previous bridge to transplantation (BTT) studies.11,12 Our diverse and extensive application of ECLS for World Symposium on Pulmonary Hypertension (WSPH) Groups 1, 3, 4, and 5 PH13 also includes non-BTT uses such as bridging to pharmacological optimization of PAH and recovery back to baseline chronic PH functional state from acute illness (bridge to recovery [BTR]), as a bridge through non-transplant surgery (BTNTS), and solely following lung transplantation (bridge post-transplantation [BPT]), as well as BTT for severe medically refractory PH. Despite advances in medical therapies for patients with PH that have significantly improved long-term outcomes, triggers such as anemia, arrhythmia, hyperthyroidism,

infection, pulmonary embolism, and pregnancy may lead to sudden right heart failure and lethal acute decompensation.14 The ideal patient for BTR from an acute medical illness is one with a potentially reversible acute trigger exacerbating their stable chronic PH before the onset of systemic end-organ failure, such as pneumonia, or in the rare instance of a newly diagnosed or undertreated patient, to provide time to initiate targeted PH therapy. The goal of BTR is not curative but supportive until these patients return to their previous stable cardiopulmonary functional status. Another major scenario for patients with PH is BTNTS, including PTE for CTEPH. The goal is to support the patient while treating the acute trigger for decompensation or permitting a patient to undergo needed non-transplant surgery. There is also a growing interest in bridging patients following lung transplantation (BPT) who had a history of severe PH.15 Recent work highlights the potential benefit of ECLS post−lung transplantation in patients with a history of PH while the right ventricle (RV) adapts to the reduction in afterload and the left ventricle (LV) accommodates to the increase in pre-load.15−17 To date, there are only a few small series including our own using BTR, BTNTS, and BPT strategies for patients with a history of PH, and none have included such a diverse application of ECLS for all forms of non−WSPH Group 2 patients with significant PH.3,4,7,8,17,18 There is an emerging role for BTT in patients with PH and right heart failure.6,15,18 One benefit of ECLS is that it can be applied dynamically to address the changing gas exchange and pump failure requirements as these patients await lung transplantation. Although there is worldwide experience in BTT for end-stage lung disease, data on the role specifically for patients with significant PH are limited.

Methods We performed a retrospective chart review of all WSPH Group 1, 3, 4, and 5 adult patients who received ECLS as BTT or non-BTT at Columbia University Medical Center between January 1, 2010 and August, 18, 2018. Included patients with PH were diagnosed and classified by a confirmatory right heart catheterization or, if too unstable, by transthoracic echocardiography. Patients with WSPH Group 2 PH who have predominant post-capillary PH because of left-sided heart diseases were excluded from analysis, as they are typically managed with a different treatment strategy. Baseline clinical and physiological characteristics were collected from the 48-hour period before ECLS initiation and compared with characteristics during the 48-hour period after initial cannulation. Medical and surgical management strategies and outcomes were analyzed in relation to clinical and physiological characteristics at baseline and during ECLS support. This study was approved

Rosenzweig et al.

ARTICLE IN PRESS

ECLS bridges for pulmonary hypertension

by the Columbia University Medical Center Institutional Review Board (IRB# AAAR9292).

Institutional approach to ECLS in PH Establishment of treatment goals A multidisciplinary team consisting of a PH specialist, surgeon, and critical care specialist established the treatment goals for each patient before ECLS cannulation. Five main questions were addressed by the team a priori:

1. Is the patient a good candidate for ECLS bridge (i.e., does bridging have a reasonable chance of leading to a successful outcome)? 2. Is the goal BTT or non-BTT? 3. What are the physiological deficits? 4. What is the ideal configuration to ameliorate those deficits? And 5. What is the optimal timing to initiate ECLS bridge?

BTT BTT was considered for patients with PH (typically WSPH Group 1, 3, or 5) actively listed for lung transplantation who had disease progression or clinical decompensation despite maximal medical targeted PH therapy. Exclusion criteria included the presence of comorbidities that would preclude lung transplantation, such as irreversible renal failure, sepsis, severe deconditioning, or prolonged mechanical ventilation. ECLS was often used before endotracheal intubation given the associated risk of cardiovascular collapse.7,18 Physical rehabilitation, including ambulation when able, was prioritized for all BTT patients to maintain conditioning and optimize transplant candidacy.10 In patients who were deemed ineligible for lung transplantation, ECLS as non-BTT for an acute illness or through surgery was considered only when appropriate. Any patient who was cannulated as BTT was either transplanted or died awaiting transplant. None were delisted and converted to BTR.

Non-BTT: BTR, BTNTS, and BPT A BTR approach was used when there was a reversible acute cause for clinical deterioration (e.g., pneumonia, pulmonary embolism, pregnancy, or thyrotoxicosis) where treatment intent was restoration to the patient’s baseline chronic medical condition or optimization of targeted PAH medications while on ECLS. Introduction of ECLS support in these severely decompensated patients allows them to better tolerate introduction and titration of targeted PH therapies that may otherwise be limited because of hemodynamic instability. The multidisciplinary team assessed the probability for recovery before ECLS cannulation. BTNTS was used for patients with PAH who required nontransplant surgery as a bridge to surgery in patients who were at risk for intraoperative hemodynamic decompensation or in those who required ECLS during or shortly following a non-transplant surgery. For example, this approach was used for temporary support before and after PTE in CTEPH patients for reperfusion injury and in other non-transplant surgeries (e.g., Cesarean section and termination of pregnancy). Some patients with significant PH were supported with ECLS during or immediately after lung transplantation only (BPT). The

3

use of venoarterial (VA) ECLS (VA-ECLS) reduces the RV preload, allowing it to adapt to post-transplant reductions in pulmonary vascular resistance; moderates the risk of reperfusion injury to lung allografts from pressure and volume overload from the hypertrophic RV; and allows the LV to adjust to the increased preload while providing gas exchange support.15 Surgical considerations for ECLS configuration were determined by underlying disease severity and reversibility (i.e., intent to decannulate), respiratory and hemodynamic status, patient anatomy, presence of a systemic to pulmonary shunt, age, frailty, comorbidities, lung allocation score, panel reactive antibodies, anticipated ECLS duration, and most importantly, whether there was a need for both hemodynamic and respiratory support. Our approach to ECLS configuration has been previously described.4 When feasible, an upper body ECLS cannulation strategy was utilized to enable patients to be awake, off mechanical ventilation, and ambulating, principally for the BTT population.3

Timing of ECLS cannulation The goal is to cannulate patients before irreversible end-organ dysfunction. Clinical variations that impact transplant wait time including blood type, antibody sensitization, and lung allocation score were also weighed. Our standard ECLS circuits use a Rotaflow pump and a Quadrox iD oxygenator or Cardiohelp System (Maquet, Rastatt, Germany). All patients receive continuous intravenous heparin infusions with a target activated partial thromboplastin time of 40−60 seconds, unless there is an indication for full therapeutic anti-coagulation.

Medical management considerations Patients were managed medically with targeted PAH therapies by a PH specialist before, during, and after ECLS. As ECLS performed in this manner provides only partial cardiac output support, concomitant use of targeted PAH medications to manage the underlying pulmonary vascular disease and right ventricular dysfunction is essential. Most patients were on targeted PAH therapies at admission and required adjustments before, during, and after ECLS support. Patients were preferentially managed with short-acting PH agents to enable rapid titration. For those with parenchymal lung disease, inhaled agents were preferred over intravenous or oral agents to minimize ventilation-perfusion mismatch and resultant hypoxemia. PAH agents with potential endorgan toxicity or drug−drug interactions were usually held or monitored very closely, and those that cause systemic vasodilation were down-titrated if the patient required vasopressors. The typical medical strategy for BTT was down-titration of PAH therapies while leaving on a maintenance regimen to minimize systemic side effects until transplantation, which include systemic vasodilation, thrombocytopenia, and hepatotoxicity. In BTR and BTNTS, the targeted PH therapies were often down-titrated until the patient’s hemodynamics and gas exchange stabilized, and then uptitrated and optimized to facilitate decannulation. ECLS decannulation in BTR, BTNTS, BPT, and BTT patients was performed when the patient was recovered from the acute illness or surgery including lung transplantation, able to tolerate slow and steady decreases in ECLS flow without compromising end-organ function (e.g., lactate remained normal or a decline in pro−brain natriuretic peptide [BNP]), and demonstrated adequate LV systolic function. There is no specific cut off target for pulmonary arterial pressure as these patients all had PH, but targeted PH therapies were optimized before decannulation. The patients were

4

ARTICLE IN PRESS

The Journal of Heart and Lung Transplantation, Vol 00, No 00, Month 2019

not typically weaned if they still required high-dose vasopressors or mechanical ventilation, so they should have tolerated a wean at least to high-flow nasal cannula. For BTT, decannulation occurred at the time of transplant or within a few days to permit RV remodeling and avoid primary graft dysfunction.

Statistical analysis Descriptive statistics were used to characterize the results. Continuous variables are presented as median (interquartile range [IQR]) and categorical variables are presented as n (percentage). Stratified analyses by non-BTT vs BTT status were performed. Categorical variables were compared using Fisher’s exact test or chi-square test. Continuous variables were compared using two-sample t-test or Kruskal−Wallis test. Pre- and post-cannulation variable changes were compared with paired sample t-test for continuous variables and McNemar’s test for categorical variables. We used logistic regression to examine associations between non-BTT vs BTT patients and 30-day mortality, in-hospital mortality, and survival to ECLS decannulation, adjusted for age, gender, and simplified acute physiology score (SAPS) II. Kaplan−Meier curves and log-rank tests for unadjusted analyses were used to examine the distribution between non-BTT vs BTT patients and 30-day survival, in-hospital survival, and survival to decannulation. Cox proportional hazards regression models were used for adjusted analyses examining predictors of 30-day mortality, in-hospital mortality, and survival to ECLS decannulation, controlling for gender, SAPS II, WSPH group, age, baseline right ventricular systolic pressure (RVSP), ECLS bridging indication, initial ECLS configuration, duration of ECLS, and ambulatory status while on ECLS. These predictors were selected based on statistical significance in univariate models and clinical significance. Testing of the proportional hazards assumption was done using Schoenfeld residuals and the assumption was met. Statistical significance was defined as a p-value ≤ 0.05, using twotailed tests of hypotheses. All analyses were performed in Stata, v14.2 (College Station, TX).

Results

the second most common for BTT (40.7%). Of BTR patients, admission diagnoses included 1 patient with extrapulmonary sepsis, 1 with an asthma exacerbation, 1 for massive hemoptysis, and 1 for massive pulmonary embolism. Patients with CTEPH were either admitted for a scheduled PTE, for progressive shortness of breath because of worsening CTEPH, or for an acute massive pulmonary embolism with chronic underlying thromboembolic disease. Of the 15 patients with a diagnosis of CTEPH, 10 underwent a PTE. Of those who underwent PTE, 5 patients received ECLS with the intent to bridge to PTE, 2 patients received ECLS intraoperatively, and 3 patients received ECLS within 2 weeks post-operatively. The median ratio of partial pressure of oxygen to fraction of inspired oxygen was 61 (IQR 48−84) for all non-surgical patients (n = 73). There were 54 BTT patients. Of these, 18 (33.3%) patients had pre-existing diagnoses of WSPH Group 1 PH, including idiopathic PAH, heritable PAH, connective tissue disease, congenital heart disease, and pulmonary venoocclusive disease. The remaining 36 (66.7%) BTT patients were either WSPH Group 3 PH (n = 34) from a variety of underlying advanced lung diseases or WSPH Group 5 (n = 2) from sarcoidosis (Figure 2). Nearly all patients who received ECLS for BTT were admitted for pneumonia, with radiographic evidence of underlying disease progression. One patient was admitted with massive hemoptysis, 1 with thyrotoxic crisis, and 1 with extrapulmonary sepsis. As highlighted in Table 1, these patients were all very ill with severe pulmonary hypertension and cardiorespiratory failure as evidenced by the median RVSP of 69 mm Hg (IQR, 54−106 mm Hg), ratio of partial pressure of oxygen to fraction of inspired oxygen of 64.5 (IQR, 50−81), BNP of 1,266 pg/ml (IQR, 400−2,652 pg/ml), the use of targeted PH therapies and vasopressors at the time of ECLS cannulation, and the median lung allocation score for the BTT cohort of 90.5 (IQR, 87.9−93.7).

Baseline demographics Clinical characteristics ECLS as bridging therapy was performed on 98 patients with PH. Baseline and in-hospital patient characteristics are shown in Table 1, stratified by bridging type (non-BTT and BTT). Baseline characterization of patients who received BTNTS are shown in Supplementary Table S1, available at www.jhltonline.org. WSPH Group 1 PH was the most common diagnostic classification in the non-BTT group (50%), and WSPH Group 3 was the most common diagnostic classification of PH in the BTT group (63%). Figures 1 and 2 show the proportion of PH patients by WSPH group and diagnosis for non-BTT and BTT, respectively. Within nonBTT, the most common ECLS indication was BTR for an acute illness (19, 43.2%) with an additional 17 (38.6%) for BTNTS (15 PTE, 1 Cesarean section, and 1 termination of pregnancy) and 8 (18.2%) for BPT in patients with PH who received ECLS during or shortly following lung transplantation (4 idiopathic PAH, 3 interstitial lung disease, and 1 sarcoidosis). Pneumonia was the most common admission diagnosis for both non-BTT (34.1%) and BTT (50%) patients, and progression of underlying lung disease was

ECLS configuration strategies Overall ECLS initial configurations and settings are shown stratified by non-BTT vs BTT status in Table 2. The majority of patients in both non-BTT and BTT received VA-ECLS as the initial configuration. Initial configurations included femoral VA-ECLS (39.8%), VAV- ECLS (9.2%) sport model (right internal jugular vein to right subclavian artery, 12.2%), and central sport model (right internal jugular vein to innominate artery, 7.1%). There were 31 initial cannulations that were venovenous (VV) ECLS (VV-ECLS) configurations (17 non-BTT and 14 BTT), including 5 utilizing a pre-existing congenital pulmonary-systemic communication (Eisenmenger syndrome) to create an oxygenated right-to-left shunt, 6 in CTEPH patients (5 post-PTE for reperfusion injury/pneumonia, 1 pre-PTE with atrial septal defect), 3 for BPT for predominant primary graft dysfunction, and the remaining 17 in patients with pneumonia or interstitial lung disease flare. Further detail on VV-ECLS baseline characteristics are shown in

Rosenzweig et al. Table 1

ARTICLE IN PRESS

ECLS bridges for pulmonary hypertension

5

Baseline Patient and In-Hospital Characteristics

Characteristic Age, years Female SAPS II BMI WSPH group 1 3 4 5 Indication for ECLS Acute medical illness (BTR) BTNTS BPT Bridge to lung transplant Medications at time of ECLS cannulation PDE5 inhibitor Endothelin receptor antagonist Inhaled vasodilatora IV/SQ prostanoid Diuretic Inotropy Vasopressor RVSP, mm Hg BNP, pg/ml pH PaCO2, mm Hg PaO2, mm Hg P:F ratio Arterial lactate, mmol/liter Serum creatinine, mg/dl Invasive ventilation during hospitalization Tracheostomy during hospitalization

Non-BTT (n = 44)

BTT (n = 54)

47 (27−54.5) 25 (56.8) 27 (22−33) 24.1 (20.5−28)

49 (35−59) 29 (53.7) 25 (21−33) 24.5 (20−27.8)

22 (50) 5 (11.4) 15 (34) 2 (4.6)

18 (33.3) 34 (63) 0 (0) 2 (3.7)

19 (43.2) 17 (38.6) 8 (18.2) 0 (0)

0 (0) 0 (0) (0) 54 (100)

p-value 0.18 0.76 0.99 0.79 <0.001

<0.001

14 (31.8) 7 (15.9) 22 (50) 18 (41) 17 (38.6) 25 (56.8) 26 (59.1) 79 (62−95) 1,259 (795−3,013) 7.36 (7.22−7.44) 41 (34−54) 54 (41−76) 56 (42−118) 3.2 (1.5−6.7) 1 (0.73−1.6) 42 (95.5) 13 (29.6)

18 (33.3) 12 (22.2) 20 (37) 13 (24.1) 32 (59.3) 15 (27.8) 15 (27.8) 69 (54−106) 1,266 (400−2,652) 7.37 (7.26−7.43) 48 (38−63) 65 (50−80) 64.5(50−81) 2.6 (1.4−5.1) 0.81 (0.62−1.16) 45 (83.3) 14 (25.9)

0.87 0.43 0.19 0.08 0.04 0.004 0.002 0.42 0.64 0.87 0.05 0.07 0.43 0.38 0.06 0.14 0.69

Abbreviations: BMI, body mass index; BNP, brain natriuretic peptide; BPT, bridge post-transplantation; BTNTS, bridge through non-transplant surgery; BTR, bridge to recovery; BTT, bridge to transplantation; ECLS, extracorporeal life support; IV, intravenous; P:F ratio, ratio of partial pressure of oxygen to fraction of inspired oxygen; PaCO2, partial pressure of carbon dioxide; PaO2, partial pressure of oxygen; PDE5, phosphodiesterase 5; RVSP, right ventricular systolic pressure; SAPS, simplified acute physiology score; SQ, subcutaneous; WSPH, World Symposium on Pulmonary Hypertension. Data presented as median (interquartile range) or n (%). Missing values: BNP = 25, pH = 2, PaCO2 = 2, PaO2 = 2, lactate = 1, creatinine=1. P:F ratio included only non-surgical patients (BTR = 2, BTT = 2, BPT = 8, BTNTS = 13) a Inhaled vasodilators included iloprost and nitric oxide.

Supplementary Table S2. Eight (26%) of the 31 initial VVECLS patients were converted to VA configurations. Overall configuration changes occurred in 35.1% of cases (Supplementary Table S3). Approximately 50% of BTT patients who started with femoral cannulation converted to an upper body configuration, and 2 patients with WSPH Group 1 PH on ECLS for BTR received an upper body configuration for prolonged support. ECLS duration was significantly longer for BTT than non-BTT (19.5 vs 9 days; p < 0.001). Approximately 7% of all patients required recannulation during their hospitalization.

Mechanical ventilation and early mobilization Intubation and mechanical ventilation rates and strategies are described in Table 2. ECLS cannulation without

intubation was performed on 38 (39.2%) patients (16 nonBTT and 22 BTT), and 52 (53%) were endotracheal extubated during ECLS. Tracheostomy rates were similar between non-BTT and BTT groups. Fifty (51%) patients ambulated during ECLS, with BTT patients having ambulated more than non-BTT (85.2% vs. 9.1%; p < 0.001).

Markers of right-sided heart failure before and after ECLS cannulation Echocardiographic and laboratory data before and after ECLS cannulation are shown in Supplementary Table S4. For the overall cohort, median RVSP was severely elevated at 73 mm Hg (IQR, 58−100 mm Hg) before ECLS cannulation and 62 mm Hg after ECLS cannulation. This echocardiographic data was similar for non-BTT and BTT groups.

6

ARTICLE IN PRESS

The Journal of Heart and Lung Transplantation, Vol 00, No 00, Month 2019

Figure 1 Etiology of pulmonary hypertension in patients on ECLS for non-BTT (BTR, BTNTS, and BPT). ARDS, acute respiratory distress syndrome; BPT, bridge post-transplantation; BTNTS, bridge through non-transplant surgery; BTR, bridge to recovery; BTT, bridge to transplantation; COPD, chronic obstructive pulmonary disease; CTEPH, chronic thromboembolic pulmonary hypertension; ECLS, extracorporeal life support; HPAH, hereditary pulmonary arterial hypertension; ILD, interstitial lung disease; IPAH, idiopathic pulmonary arterial hypertension; SLE, systemic lupus erythematosus; TGA, post-operative transposition of the great arteries; VSD, ventricular septal defect; WSPH, World Symposium on Pulmonary Hypertension.

Figure 2 Etiology of pulmonary hypertension in patients on ECLS for bridge to lung transplantation. ASD, atrial septal defect; COPD, chronic obstructive pulmonary disease; CTD-ILD, connective tissue disease−interstitial lung disease; ECLS, extracorporeal life support; HP, hypersensitivity pneumonitis; HPAH, hereditary pulmonary arterial hypertension; ILD, interstitial lung disease; IPAH, idiopathic pulmonary arterial hypertension; IPF, idiopathic pulmonary fibrosis; NSIP, non-specific interstitial pneumonitis; PVOD/PCH, pulmonary veno-occlusive disease/pulmonary capillary hemangiomatosis; TGA, post-operative transposition of the great arteries; VSD, ventricular septal defect; WSPH, World Symposium on Pulmonary Hypertension.

Serum BNP levels were also severely elevated at baseline consistent with right heart failure and improved within 48 hours of ECLS cannulation for both BTT (1,266 to

549 pg/ml, p = 0.04) and non-BTT patients (1,259 to 1,108 pg/ml, p = 0.03). Median serum creatinine, basic natriuretic peptide, and arterial blood gas measures

Rosenzweig et al. Table 2

ARTICLE IN PRESS

ECLS bridges for pulmonary hypertension

7

ECLS Characteristics and Initial Settings

Characteristic Initial ECLS configuration Veno-venous Veno-arterial Femoral V-femoral A Veno-arterial-venous Sport model Central sport model Distal perfusion cannula Blood flow, LPM % Calculated cardiac output Sweep gas flow rate, LPM ECLS venous oxygen saturation, % Modification to configuration Initial Second Third Recannulation Ambulated with ECLS Duration of ECLS, days Extubated on ECLSa Intubated for ECLS cannulation

Non-BTT (n = 44) 17 (38.6)

BTT (n = 54) 14 (25.9)

p-value 0.18

19 (43.2) 6 (13.6) 2 (4.6) 0 (0) 5 (11.4) 2.7 (2.2−3.5) 76.5 (59−89.5) 2 (1.1−3.6) 76 (70−81)

20 (37) 3 (5.6) 10 (18.5) 7 (13) 5 (9.3) 2.9 (2.3−3.3) 71.5 (61−92) 2.5 (2−3.3) 74 (64−79)

0.54 0.17 0.04 0.01 0.73 0.87 0.82 0.07 0.18

10 (22.7) 0 (0) 0 (0) 3 (6.8) 4 (9.1) 9 (6−14.5) 16 (42.1) 28 (63.6)

28 (51.9) 3 (5.6) 1 (1.9) 4 (7.4) 46 (85.2) 19.5 (11−34) 36 (80) 32 (59.3)

0.003 0.28 — 0.91 <0.001 <0.001 <0.001 0.66

Abbreviations: BTT, bridge to transplantation; ECLS, extracorporeal life support; Femoral A, femoral artery; Femoral V, femoral vein; LPM, liters per minute; VAV, veno-arterial-venous. Data presented as median (interquartile range) or n (%). a 83 patients were intubated while on ECLS (n = 38 bridge to recovery, n = 45 bridge to transplantation).

improved to variable degrees post-ECLS cannulation in the overall cohort and in all groups.

PH medication management PAH medications (Table 1, Figure 3) were used most often in non-BTT patients and WSPH Group 1 BTT patients. Binary use of these medications did not generally change before and after cannulation, aside from an increase in inhaled pulmonary vasodilators in non-BTT patients (p < 0.01). However, the dosages of these medications in general were decreased for BTT patients and escalated for non-BTT patients once the patients were hemodynamically stabilized.

Clinical outcomes For the overall cohort, the median intensive care unit length of stay was 29 days (IQR, 13−50 days), and median hospital length of stay was 40 days (IQR, 21−73 days; Table 3). Duration of invasive mechanical ventilation was 4.5 days (IQR, 1−16 days) and the duration of ECLS support was 12 days (IQR, 6−21 days). Duration of mechanical ventilation in BTT was less than for non-BTT (1.5 days vs 7 days; p < 0.001). Overall, for BTT and non-BTT, survival to decannulation was 66.3% and survival to hospital discharge was 54.1% (Table 3). Of the non-BTT patients overall, 79.6% survived to decannulation and 56.8% survived to hospital discharge. All patients (100%) who received ECLS as BPT and 65% of the BTNTS patients survived to hospital

discharge. Of the BTT patients, 59.3% survived to lung transplantation and 51.9% survived to hospital discharge. Of the BTT patients who received a lung transplant, 88% survived to hospital discharge. The BTT patients who were initially cannulated with VV-ECLS, most in the early era (2010−2014), had substantially worse survival to discharge vs those who were initially cannulated with VA-ECLS (28.5% vs 60%; p = 0.04). Non-BTT patients were more likely to survive to ECLS decannulation than BTT patients (adjusted odds ratio [OR], 3.73; 95% confidence interval [CI], 1.4−9.9; p = 0.008). Non-BTT patients were less likely to ambulate on ECLS or be extubated while on ECLS than BTT patients (Tables 2 and 4). Ambulation during ECLS was associated with higher survival to discharge (adjusted hazard ratio [AHR], 0.19; 95% CI, 0.06−0.64; p = 0.007) (Table 5). The strongest predictor of in-hospital mortality was an initial ECLS configuration of VV- ECLS (36.4% survival to discharge; AHR, 2.64; 95% CI, 1.15 −6.07; p = 0.02) (Table 5). Of those who were cannulated initially with VV-ECLS, non-BTT were more likely to survive to hospital discharge than those initially cannulated with VV-ECLS as BTT, even after adjusting for age, gender, and SAPS II (BTR: OR, 5.37; 95% CI, 1.12−25.29; p = 0.035; BTT: OR, 0.26; 95% CI, 0.06−1.0; p = 0.05). Within the BTT cohort, those who underwent lung transplantation were more likely to have been extubated on ECLS (92.3% vs 63.2%, p = 0.016) and were less likely to have VV-ECLS as their initial configuration, particularly in the current era (12.5% vs 45.5%, p = 0.007) (Table 6).

ARTICLE IN PRESS

8

The Journal of Heart and Lung Transplantation, Vol 00, No 00, Month 2019

Figure 3 Medication profile pre- and post- cannulation to ECLS for non-BTT and BTT. *Inhaled vasodilators included iloprost and nitric oxide. Pre- to post-cannulation use was significantly increased in the bridge to recovery group (50% vs. 70.5%, p = 0.01) BTT, bridge to transplantation; ECLS, extracorporeal life support; ERA, endothelin receptor antagonist; IV, intravenous; PDE5, phosphodiesterase 5; SQ, subcutaneous.

Discussion ECLS is a relatively new but growing approach in the management of advanced pulmonary hypertension.3,4,8,17,18 Our single institutional experience of using ECLS for WSPH Groups 1, 3, 4, and 5 PH as non-BTT (BTR, BTNTS, and BPT) and BTT highlights both the feasibility and the challenges of this strategy and is, to our knowledge, the first to extensively characterize the various medical-mechanical techniques across such a broad range of patients with severe PH. The patients with PH described were extremely ill and

Table 3

all chosen for ECLS because they were not expected to survive without ECLS as a bridge to recovery or to lung transplantation. Most patients received VA-ECLS (68%); however, VV-ECLS was utilized early on in cases when there was primarily a gas exchange problem with adequate ventricular function (e.g., pneumonia, reperfusion injury post-PTE, or primary graft dysfunction after lung transplantation) or when there was a congenital systemic-pulmonary shunt. However, our early experience suggests that initial VV-ECLS configuration is associated with poor outcomes overall for patients with PH, including higher risk for in-

Clinical Outcomes

Outcome Survival to ECLS decannulation Survival to hospital discharge Survival to hospital discharge stratified by ECLS indication BTR (n = 19) BTNTS (n = 17) Bridge to post-transplant (n = 8) BTT (n = 54) Survival to lung transplantation Time from admission to transplant, days Duration of invasive mechanical ventilation, days Hospital length of stay, days ICU length of stay, days Successful PTE for patients with CTEPHa

Non-BTT (n = 44)

BTT (n = 54)

35 (79.6) 25 (56.8)

30 (55.6) 28 (51.9)

6 (31.6) 11 (64.7) 8 (100) — — — 7 (3-12) 33.5 (18-59.5) 23.5 (13-38.5) 10 (66.7)

— — — 28 (51.9) 32 (59.3) 18 (8-31.5) 1.5 (1-3) 49.5 (26-76) 34 (16-57) —

p-value 0.012 0.62

<0.001 0.08 0.06

Abbreviations: BTNTS, bridge through non-transplant surgery; BTR, medical bridge to recovery; BTT, bridge to transplantation; CTEPH, chronic thromboembolic pulmonary hypertension; ECLS, extracorporeal life support; ICU, intensive care unit; PTE, pulmonary thromboendarterectomy. Data presented as median (interquartile range) or n (%). a 15 patients with diagnosis of CTEPH.

Rosenzweig et al.

ARTICLE IN PRESS

ECLS bridges for pulmonary hypertension

Table 4 Odds Ratios for Clinical Outcomes Comparing nonBTT to BTT Patients

Outcome

95% confidence intervals

Odds ratioa

Survival to hospital discharge Unadjusted model 1.22 Adjusted modelb 1.22 Survival to ECLS decannulation Unadjusted model 3.11 Adjusted modelb 3.73 Ambulation while on ECLS Unadjusted model 0.02 Adjusted modelb 0.01 Extubated while on ECLS Unadjusted model 0.18 Adjusted modelb 0.19

p-value

0.55−2.72 0.53−2.78

0.62 0.64

1.25−7.71 1.4−9.9

0.014 0.008

0.005−0.06 0.003−0.05

<0.001 <0.001

0.07−0.48 0.07−0.52

0.001 0.001

Abbreviations: BTT, bridge to transplantation; ECLS= extracorporeal life support. a Odds ratio comparing non-BTT patients to BTT patients. b Adjusted for age, gender, and simplified acute physiology score II.

hospital mortality (AHR, 4.67; 95% CI, 1.64−13.3; p = 0.004). Further, in patients undergoing BTT, VV-ECLS support, even if converted to VA-ECLS, was associated with a significantly worse outcome for patients undergoing BTT than initial VA support (28.6% vs 60% survival to hospital discharge; p = 0.04). In our experience, VA-ECLS is the optimal configuration for BTT patients with pre-existing significant PH or RV dysfunction, even in cases of WSPH Group 3 disease, which is reflected in our substantial decline in usage after 2014 of VV-ECLS as BTT for patients with significant PH and RV dysfunction. Initial configuration was VV-ECLS as BTT (50%) in the early era (n = 11 of 22 BTT patients; 2010−2014) vs VV-ECLS as BTT (9%) in

Table 5

Predictors of In-Hospital Mortality in All Patients In-hospital mortality a

95% CI

p-value

PredictorI

AHR

Male gender Age SAPS II WSPH group Baseline RVSP ECLS for non-BTT Initial configuration of venovenous ECLS Duration of ECLS, days Ambulation while on ECLS

0.51 1.00 1.03 0.88 0.99 0.21 2.64

0.22−1.18 0.97−1.03 0.99−1.08 0.64−1.21 0.98−1.01 0.06−0.71 1.15−6.07

0.11 0.93 0.13 0.43 0.28 0.01 0.02

0.99 0.19

0.97−1.01 0.06−0.64

0.25 0.007

Abbreviations: AHR, adjusted hazard ratio; CI, confidence interval; SAPSII, simplified acute physiology score II; WSPH, world symposium pulmonary hypertension; RVSP, right ventricular systolic pressure; BNP, brain natriuretic peptide; ECLS, extracorporeal life support; BTT, bridge to lung transplantation. a AHR adjusted for gender, age, SAPS II, WSPH group, baseline RVSP, ECLS for non-bridge to transplantation, initial ECLS configuration, duration of ECLS and ambulatory status while on ECLS.

9

current era (n = 3 of 32 BTT patients; 2015−2018). Conversely, patients with PH with VV-ECLS for non-BTT did relatively well in the current era, with 87.5% survival to hospital discharge (Table 6). We concluded that VV-ECLS should be reserved for select non-BTT patients with PH who are bridging through a definitive surgery (BPT or BTNTS) or who have a congenital systemic-pulmonary shunt, which can be utilized to direct oxygenated blood from pulmonary to systemic circulation, and in rare cases of preserved RV function and primary gas exchange deficit, such as pneumonia. Circuit modification was common, occurring in 35% of patients. Typically, this was a change from VV to VA to add hemodynamic support, or from a lower body to an upper body VA configuration to improve upper body oxygenation (avoiding Harlequin syndrome) and to facilitate mobility. Unfortunately, for BTT, even a change in configuration did not improve outcome if initial configuration was VV-ECLS. In some patients with femoral VA-ECLS (n = 8), an additional venous limb was added to an internal jugular vein to optimize cerebral and coronary perfusion and decrease the differential between upper and lower body oxygenation in the setting of preserved left ventricular function and impaired native gas exchange. Of the 98 patients, 38% were successfully cannulated without intubation, including 53% of all WSPH Group 1 patients. Given the potential for severe hemodynamic compromise during endotracheal intubation in WSPH Group 1 patients, this represents a potentially safer approach to support decompensated PH patients. One strategy that emerged was to use femoral VA cannulation while awake to stabilize the PH patient and avoid anesthesia with conversion to upper body durable VA support if needed once the patient was more stable. An important goal for BTT patients is maintenance of physical conditioning while awaiting lung transplantation, particularly because transplant wait times can be very long— weeks to months. Ambulation, which was associated with lower in-hospital mortality, was achieved in 83% of our BTT patients and was facilitated by use of an upper body cannulation strategy. In non-BTT patients, where the goal for these patients was to optimize and decannulate as soon as possible (ECLS duration, 9 vs 19.5 days; non-BTT vs BTT), there is less emphasis on mobilization, and conversion to upper body configurations was less common. The BTNTS and BPT patients were included even though they underwent a definitive surgical cure because of their history of severe PH, reactive pulmonary vascular bed in the case of CTEPH, and the impact of the RV and LV during recovery. These patients tended to be easier to decannulate than a medical BTR patient because of the significant reduction of pulmonary arterial pressure at the time of surgery similar to a BTT patient who undergoes lung transplantation. Our experience has shown that PH patients usually only require partial support, as left ventricular systolic function is typically still preserved. Therefore, concomitant use of targeted PAH medical therapies remains an important component of management because of the substantial amount

ARTICLE IN PRESS

10

The Journal of Heart and Lung Transplantation, Vol 00, No 00, Month 2019

Table 6

VV-ECLS Cannulation and Outcomes BTT n = 14

Non-BTT n = 17

Outcome

Early era (n = 11)

Current era (n = 3)

Early era (n = 9)

Current era (n = 8)

Survive to decannulation Survive to hospital discharge

36.4% 36.4%

0% 0%

77.8% 66.7%

100% 87.5%

Abbreviations: BTT, bridge to transplantation; VV-ECLS, venovenous extracorporeal life support. Early era, 2010−2014; current era, 2015−2018.

of native pulmonary blood flow. PAH medications were titrated with several strategies as part of the mechanicalmedical management of these patients. Inhaled agents (e.g., iloprost and nitric oxide) were preferred over oral or intravenous agents to minimize systemic side effects, including systemic hypotension and worsening ventilation-perfusion mismatch. In addition, short-acting agents were preferred over long-acting agents until patients were stabilized off ECLS. In BTT, medications were typically weaned to lower dosages during ECLS to minimize systemic effects, which can be exacerbated on VA extracorporeal membrane oxygenation with diversion of medication to the arterial circulation. In contrast, non-BTT often involved down-titration during the acute post-cannulation phase to optimize hemodynamic stability, followed by up-titration of PAH medications to optimize patients when they were ready to wean from ECLS. For patients on intravenous prostanoid therapy undergoing VA-ECLS cannulation, we have observed a phenomenon similar to what may occur after atrial septostomy, where the intravenous prostanoid is delivered more directly into the systemic circulation, potentially leading to severe systemic hypotension. Therefore, intravenous prostanoid doses are typically reduced by approximately 10% at the time of ECLS cannulation, with additional down-titration as needed. Oral PAH medications are typically held until the patient is stabilized on ECLS. For BTR, as the patient recovers from their acute illness, the targeted PAH medications can be slowly reintroduced and up-titrated and may require escalation above pre-ECLS dosages, as there has often been a transient clinical worsening. We have also observed specifically that the short-acting phosphodiesterase5 inhibitors should be very slowly reintroduced only when the patient is off vasopressors to avoid systemic hypotension. For patients who are cannulated in the setting of suboptimal PH treatment or treatment-naive status, ECLS offers the opportunity to safely initiate or up-titrate PAH therapy. A multidisciplinary team including a PH specialist should assess the patient before, during, and after ECLS to assist with the complex management of targeted PAH medications. Limitations of this study include the heterogeneous nature of PH, the type of ECLS bridging required, and the change in practice over time from lessons learned. However, the authors believe that our experience illustrates the breadth of potential applications of ECLS in adults with severe PH in a broad variety of scenarios. This series

reflects the lessons learned at a high-volume center for both PH and ECLS and may not be generalizable to centers with fewer resources and less experience. ECLS has an expanding role in the management of advanced WSPH Group 1, 3, 4, and 5 PH as both non-BTT (BTR, BTNTS, and BPT) and BTT. This single-center experience demonstrates that a multidisciplinary team approach involving appropriate selection of patients, cannulation configurations, patient-specific optimization of PH medical therapies, and avoidance of endotracheal intubation is effective in managing these complex patients. With technological advances and increasing surgical experience, more novel and durable cannulation approaches have not only facilitated ambulation during ECLS, but also allowed for prolonged durations of support often needed for BTT. This study suggests potential opportunities to further expand the role of ECLS in patients with PH with acute and chronic decompensation.

Disclosure statement E.B. Rosenzweig, W. Gannon, P. Madahar, C. Agerstrand, D. Abrams, P. Liou, and M. Bacchetta have no disclosures related to this work. D. Brodie receives research support from ALung Technologies; he was previously on their medical advisory board. He has been or anticipates being on the medical advisory boards for Baxter and BREETHE, Inc. The authors would like to thank the Richard Bartlett Memorial Foundation for its support of this work through the Pulmonary Hypertension Comprehensive Care Center at Columbia University Medical Center—New York Presbyterian Hospital. There was no funding for this research other than by the Richard Bartlett Memorial Foundation for statistical support for this project.

Supplementary materials Supplementary material associated with this article can be found in the online version at https://doi.org/10.1016/j.hea lun.2019.09.004.

References 1. Bartlett RH, Gazzaniga AB, Jefferies MR, Huxtable RF, Haiduc NJ, Fong SW. Extracorporeal membrane oxygenation (ECMO)

Rosenzweig et al.

2.

3.

4.

5.

6.

7.

8.

9.

ARTICLE IN PRESS

ECLS bridges for pulmonary hypertension

cardiopulmonary support in infancy. Trans Am Soc Artif Intern Organs 1976;22:80-93. Bartlett RH, Roloff DW, Cornell RG, Andrews AF, Dillon PW, Zwischenberger JB. Extracorporeal circulation in neonatal respiratory failure: a prospective randomized study. Pediatrics 1985;76:479-87. Abrams DC, Brodie D, Rosenzweig EB, Burkart KM, Agerstrand CL, Bacchetta MD. Upper-body extracorporeal membrane oxygenation as a strategy in decompensated pulmonary arterial hypertension. Pulm Circ 2013;3:432-5. Rosenzweig EB, Brodie D, Abrams DC, Agerstrand CL, Bacchetta M. Extracorporeal membrane oxygenation as a novel bridging strategy for acute right heart failure in group 1 pulmonary arterial hypertension. ASAIO J 2014;60:129-33. Javidfar J, Brodie D, Iribarne A, et al. Extracorporeal membrane oxygenation as a bridge to lung transplantation and recovery. J Thorac Cardiovasc Surg 2012;144:716-21. Chicotka S, Pedroso FE, Agerstrand CL, et al. Increasing opportunity for lung transplant in interstitial lung disease with pulmonary hypertension. Ann Thorac Surg 2018;106:1812-9. Biscotti M, Vail E, Cook KE, Kachulis B, Rosenzweig EB, Bacchetta M. Extracorporeal membrane oxygenation with subclavian artery cannulation in awake patients with pulmonary hypertension. ASAIO J 2014;60:748-50. Olsson KM, Simon A, Strueber M, et al. Extracorporeal membrane oxygenation in nonintubated patients as bridge to lung transplantation. Am J Transplant 2010;10:2173-8. Chicotka S, Rosenzweig EB, Brodie D, Bacchetta M. The “central sport model”: extracorporeal membrane oxygenation using the innominate artery for smaller patients as bridge to lung transplantation. ASAIO J 2017;63:e39-44.

11

10. Abrams D, Javidfar J, Farrand E, et al. Early mobilization of patients receiving extracorporeal membrane oxygenation: a retrospective cohort study. Crit Care 2014;18:R38. 11. Hoetzenecker K, Donahoe L, Yeung JC, et al. Extracorporeal life support as a bridge to lung transplantation-experience of a high-volume transplant center. J Thorac Cardiovasc Surg 2018;155:1316-1328.e1. 12. Toyoda Y, Bhama JK, Shigemura N, et al. Efficacy of extracorporeal membrane oxygenation as a bridge to lung transplantation. J Thorac Cardiovasc Surg 2013;145:1065-71. 13. Simonneau G, Montani D, Celermajer DS, et al. Haemodynamic definitions and updated clinical classification of pulmonary hypertension. Eur Respir J 2019;53. 14. Savale L, Weatherald J, Jaı¨s X, et al. Acute decompensated pulmonary hypertension. Eur Respir Rev 2017;26. 15. Pereszlenyi A, Lang G, Steltzer H, et al. Bilateral lung transplantation with intra- and postoperatively prolonged ECMO support in patients with pulmonary hypertension. Eur J Cardiothorac Surg 2002;21:85863. 16. Tsai MT, Hsu CH, Luo CY, Hu YN, Roan JN. Bridge-to-recovery strategy using extracorporeal membrane oxygenation for critical pulmonary hypertension complicated with cardiogenic shock. Interact Cardiovasc Thorac Surg 2015;21:55-61. 17. Tudorache I, Sommer W, K€uhn C, et al. Lung transplantation for severe pulmonary hypertension—awake extracorporeal membrane oxygenation for postoperative left ventricular remodelling. Transplantation 2015;99:451-8. 18. de Perrot M, Granton JT, McRae K, et al. Impact of extracorporeal life support on outcome in patients with idiopathic pulmonary arterial hypertension awaiting lung transplantation. J Heart Lung Transplant 2011;30:997-1002.