ECMO in severe hypoxemia post liver transplant for hepatopulmonary syndrome

Case report

A 59-year-old man with a history of alcoholic cirrhosis, portal hypertension and HPS with severe hypoxemia requiring supplemental oxygen. His Model for End-Stage Liver Disease (MELD) score was 12 points, and pre-transplant arterial oxygen partial pressure (PaO2) was 57 mmHg without supplemental oxygen, with a right-to-left shunt of 13.61%, as determined by Tc99m.MMA lung perfusion scan. The patient underwent orthotopic liver transplantation using the piggy-back technique and a temporary portocaval shunt.

Following transplantation, the patient was successfully extubated within 4 h of admission to the intensive care unit (ICU) and supported with high-flow nasal cannula (HFNC) delivering 60 litres per minute (lpm) with an inspired oxygen fraction (FiO2) of 0.7, achieving a PaO2 of 86 mmHg. Despite initial stability, the patient’s reliance on HFNC persisted, and bilateral B-lines were seen on lung ultrasound after 2 days, showing ongoing respiratory compromise. Diuretics and inhaled iloprost were administered, resulting in improved oxygen saturation (SpO2). However, due to worsening dyspnoea, non-invasive ventilation (NIV) was started on the third day. Subsequently, the patient experienced a seizure in the context of hypoxemia and required reintubation. Imaging studies did not reveal any significant intracranial abnormalities.

Despite aggressive respiratory support with mechanical ventilation and adjunctive therapies including inhaled nitric oxide (iNO) and iloprost, the patient’s hypoxemia persisted with a PaO2 below 60 mmHg, needing further interventions. On the seventh day of ICU admission, veno-venous extracorporeal membrane oxygenation (VV-ECMO) was initiated via a right femoro-jugular approach, with a Biomedicus^® 25-Fr femoral cannula and Biomedicus^® 19-Fr jugular cannula, guided by trans-esophageal echocardiogram, utilising a Cardiohelp device (Medtronic^®). Initial ECMO settings included a swept gas flow of 3 L/min, FiO2 at 1 and a pump speed of 2400 rpm, resulting in a flow of 3.2 L/min. The patient’s cardiac output was 6 L/min throughout ECMO support, facilitating a reduction in ventilator FiO2 while preserving acceptable oxygenation parameters. Despite only transient improvements in SpO2 with inhaled iloprost, intravenous epoprostenol was introduced for sustained effects and potential antiplatelet benefits. This decision aligned with our strategy of avoiding full anticoagulation, given the patient consistently maintained activated clotting time of around 150–175 s.

Over the next days, the patient’s respiratory status gradually improved and serial transthoracic echocardiograms showed a progressive decrease in shunt using an agitated saline bubble test, allowing for a reduction in sedation and awakening to a more responsive state. As the patient’s condition stabilised, ECMO support was gradually weaned, and successful decannulation was achieved on the eighth day of ECMO implantation. Subsequent trials of spontaneous breathing and respiratory support were successful, leading to the patient’s extubation on the 16th day from ICU admission.

During ICU admission, the patient encountered more challenges, including acute graft rejection requiring corticosteroid therapy, Clostridium difficile infection treated with fidaxomicin and paroxysmal atrial fibrillation managed with beta-blockers. Despite these complications, the patient showed steady improvement and was discharged from the ICU to the general ward on the 21st day of admission.

In the ward, the patient’s recovery continued, with gradual reduction in supplemental oxygen requirements. By the 28th day post-transplant, the patient no longer needed supplemental oxygen and was discharged home. Follow-up evaluations showed sustained improvement, with the patient achieving satisfactory oxygenation levels without supplemental oxygen six-month post-transplant.

Analysing this scenery, the scarcity of cases involving severe hypoxemia post-liver transplant due to hepatopulmonary syndrome results in a plethora of treatment options with anecdotal evidence. One possibility is ECMO ¹ , although the rationale for supporting oxygenation while the pathophysiological mechanisms of hepatopulmonary syndrome improve warrants consideration. This case report, among 17 others globally involving both children and adults, ² proves the feasibility of using ECMO in this challenging scenario.

The primary indication for ECMO in this patient was based on two factors: a PaO2/FiO2 ratio below 80 mmHg despite maximal respiratory support measures known for hepatopulmonary syndrome, and the reversibility of the condition, utilising ECMO as a bridge to recovery.

The understanding of the timing for improvement of hypoxemia in this setting remains limited. ³ Nonetheless, like other case reports, ECMO bought enough time for our patient to achieve an acceptable PaO2, allowing successful weaning from vital support.

Considering that patients with hepatopulmonary syndrome typically have low PaO2 levels, we aim to wean off ECMO once patient was awake, managing low levels of pressure support ventilation and maintained a PaO2 above 60 mmHg with an FiO2 of 0.4. Furthermore, by monitoring the shunt daily with the bubble test, we detected a reduction in the shunt. This reduction, along with PaO2 levels, were key factors in deciding when to begin the weaning process from ECMO.

A recent systematic review ² comprehensively outlined cases from the literature, revealing a shared chronological sequence that closely resembled our case. The patient underwent extubation, faced reintubation on day 3 and received ECMO support on day 7. However, we diverged from this pattern with 8 days of ECMO compared to the reported 13 days, early discharge from the ICU at day 21 versus day 40 and home discharge at day 28 versus day 60.

These variations seem linked to a less complicated postoperative period, where, aside from respiratory failure, the patient did not need reoperation or experience other organ failures. Prompt control of Clostridium difficile infection and successful resolution of acute graft rejection contributed to this favourable trajectory.

Upon isolating the eight adult cases in the literature (see Table 1), significant heterogeneity was noted, precluding generalised contextual and outcome analysis.^{4 –10}

OPEN IN VIEWER

Table 1.

ECMO cases reported for hepatopulmonary syndrome in older than 18 years.

Feature	Auzinger et al. 4	Braun et al. 5	Braun et al. 5	Goussous et al. 6	Herden et al. 7	Hogen et al. 8	Monsel et al. 9	Sharma et al. 10	Current case
Aetiology	Alcoholic cirrhosis	Alcoholic cirrhosis	Non-cirrhotic portal hypertension	HCV	Cryptogenic cirrhosis	N/D	Alcoholic cirrhosis	NASH	Alcoholic cirrhosis
MELD	N/D	31	25	28	12	N/D	N/D	22	12
Age (years) and sex	44	50	28	52	62	42	51	60	59
	N/D	Male	Male	Male	Male	Male	Male	Female	Male
HPS diagnosis and follow-up	Previous TTE + Bubble test	Previous	Previous	Previous	Admission	Previous	Admission TTE + Bubble test	Admission TTE + Bubble test	Previous TTE + Bubble test
ECMO configuration	V-VDouble Lumen Bicaval Jugular	V-V	V-V	V-VFemoro-Jugular	N/D	V-A-VTri femoral	V-VJugulo-Femoral	V-VDouble Lumen Bicaval Jugular	V-VFemoro-Jugular
Time TO ECMO	+13 POD	+12 POD	+5 POD	+1 POD	+7 POD	+0 POD	−5 POD	+11 POD	+7 POD
Time ON ECMO (days)	21	47	10	10	5	10	5	13	8
ECMO complications	Cannula migration with haemorrhagic venous infarction of liver graft	None reported	None reported	None reported	None reported	Harlequin Syndrome	None reported	Left Haemothorax + Patent foramen ovale	None
ECMO Anticoagulation target	ACT 180 s	N/D	N/D	aPTT 40–60 s	N/D	No Anticoagulation	Only cannulation Bolus	ACT 140–160 s	ACT 140–160 s
Status	Alive ICU	Dead ICU + 61 POD	Dead ICU + 58 POD	Alive ICU	Alive ICU	Alive ICU	Alive ICU	Alive ICU	Alive 1 year follow up

HCV: hepatitis C virus; N/D: no data; NASH: non-alcoholic steato-hepatitis; TTE: trans-thoracic echocardiography; V-V: veno-venous; V-A-V: veno-arterial-venous; POD: postoperative day; ECMO: extracorporeal membrane oxygenation; ACT: activated clotting time; aPTT: activated partial thromboplastin time; ICU: intensive care unit; HPS: hepato-pulmonary syndrome.

The natural recovery process ensued, and our patient was free from the need for supplemental oxygen at 6 months, consistent with literature reporting a recovery period of about 6–12 months and resolution rate about 94% of cases. ³

Without ECMO support and considering that other treatments did not improve hypoxemia severe enough to cause seizures, this patient might have succumbed to respiratory failure. To date, survival reported from the limited data available in the literature is around 82%. ² This encouraged us to try with ECMO support, and the patient survived.

In conclusion ECMO is a valuable therapeutic possibility in liver transplant recipients experiencing severe hypoxemia secondary to HPS. Careful patient selection, timing of intervention and ongoing monitoring are essential to improve outcomes in this challenging patient population.