Implementation and results of a new ECMO program for lung transplantation and acute respiratory distress.

OBJECTIVE: The development of the extracorporeal membrane oxygenation in Latin America represents a challenge in this specialty field. The objective of this article was to describe the results of a new extracorporeal membrane oxygenation program in an intensive care unit.
METHODS: This retrospective cohort study included 22 patients who required extracorporeal membrane oxygenation and were treated from January 2011 to June 2014. The baseline characteristics, indications, duration of the condition, days on mechanical ventilation, days in the intensive care unit, complications, and hospital mortality were evaluated.
RESULTS: Fifteen patients required extracorporeal membrane oxygenation after lung transplantation, and seven patients required oxygenation due to acute respiratory distress. All transplanted patients were weaned from extracorporeal membrane oxygenation with a median duration of 3 days (Interquartile range - IQR: 2 - 5), were on mechanical ventilation for a median of 15.5 days (IQR: 3 - 25), and had an intensive care unit stay of 31.5 days (IQR: 19 - 53) and a median hospital stay of 60 days (IQR: 36 - 89) with 20% mortality. Patients with acute respiratory distress had a median oxygenation membrane duration of 9 days (IQR: 3 - 14), median mechanical ventilation time of 25 days (IQR: 13 - 37), a 31 day stay in therapy (IQR: 11 - 38), a 32 day stay in the hospital (IQR: 11 - 41), and 57% mortality. The main complications were infections (80%), acute kidney failure (43%), bleeding at the surgical site and at the site of cannula placement (22%), plateletopenia (60%), and coagulopathy (30%).
CONCLUSION: In spite of the steep learning curve, we considered this experience to be satisfactory, with results and complications comparable to those reported in the literature.

INTRODUCTION

The utility and safety of extracorporeal membrane oxygenation (ECMO) in patients with acute respiratory distress syndrome (ARDS) has been the subject of debate over the past 50 years. Although the mortality and incidence of ARDS have decreased in recent decades,( in patients with refractory hypoxemia who are supported by adequate mechanical ventilation (MV), the mortality rate can surpass 50%.( One objective of ECMO support is to guarantee adequate oxygenation. In addition, this support allows for MV to have very low tidal volumes and, consequently, lower inspiration end pressures to prevent pulmonary lesions associated with mechanical ventilation.( The positive experience with ECMO during the 2009 H1N1 pandemic( and its increasing use in Great Britain and the U.S.A. could justify the implementation of ECMO as a vital support measure during severe ARDS.(

In lung transplantation patients, ECMO is a vital support tool during primary graft failure (PGF) and/or in peritransplant circulatory collapse.(

For these reasons, in the year 2011, an ECMO program was launched in the adult intensive care unit (ICU) of the Hospital Italiano de Buenos Aires (HIBA). Because the majority of the published experiences with this type of technology have originated from developed countries, we considered the launch of this type of program to be very important for this region. This work analyzes the characteristics, complications, and results of the first patients to receive ECMO in our adult intensive care unit. The main objective of this study is to describe the baseline characteristics of the population that required ECMO in our center (age, sex, comorbidity load using the Charlson score,( indications ECMO), the duration of the intensive care stay, the duration of mechanical ventilation and mortality. As secondary objectives, we seek to describe and compare the ventilation strategy utilized before and during ECMO and the complications that were presented. These complications were classified into clinical complications and ECMO-associated complications and were extracted from patients’ compiled electronic medical records.

METHODS

A retrospective observational study of 22 patients requiring ECMO in the UCI from the start of the program in January 2011 until June 2014 was conducted. Data were obtained from patient electronic hospital medical records. This study was approved by the Hospital Italiano de Buenos Aires Research Protocol Ethics Committee (Protocol no. 2410). The patients included or their next of kin signed the corresponding consent for their confidential data to be used for investigative purposes.

Our ICU is polyvalent and has 38 beds equipped for highly complex situations. The Italian Hospital has been a lung transplantation center since 1993. The adult intensive care ECMO program was developed based on the recommendations from the ELSO( and NICE guidelines.( For our program, we utilized the 1994 American-European Consensus definition of ARDS,( and after 2012, we utilized the Berlin classification.(

Following our critical care unit’s ARDS protocol, the protective ventilation strategy consisted of low tidal volumes (TV = 6 - 7ml/kg theoretical weight) and plateau pressures (Ppl) < 30mmHg.( After a recruitment maneuver, positive end-expiratory pressure (PEEP) titration occurred in all patients following a PEEP/deflation complacency curve.( Sedation was optimized to reach a Richmond agitation-sedation score (RASS) of -4,( and in patients with a Pa/FiO2 < 150 neuromuscular blocking agents (NBA) were used in continuous infusion.( During the last year, ventilation in a prone position in patients with ARDS and a Pa/FiO2 < 150 were incorporated into the distress protocol despite adequate MV.( Those patients who continued to exhibit refractory hypoxemia (Pa/FiO2 < 100, FiO2 > 0.6), respiratory acidosis with a pH < 7.2, and/or evidence of barotrauma after 12 to 24 hours of these ventilatory strategies( were evaluated for veno-venous (VV) ECMO. Patients who also presented hemodynamic instability secondary to cardiogenic shock were candidates for veno-arterial (VA) ECMO. In our study, this group included lung transplantation patients with acute heart failure in the intraoperative or early postoperative period following severe pulmonary hypertension (due to PGF or during pulmonary artery clamping).

Contraindications for ECMO included the presence of multiple organ failure, recent central nervous system bleeding, active hemorrhaging, or the presence of an anticoagulation contraindication, acute or sub-acute pulmonary fibrosis without the possibility of transplant, and cardiogenic shock secondary to an aortic valve insufficiency.

An ECMO system (Permanent life support system - PLS, Jostra - Quadrox D, Maquet Cardiopulmonary, Hirrlingen, Germany) with a centrifugal pump (Rotaflow, Jostra, Maquet Cardiopulmonary, Hirrlingen, Germany) was used. The PLS utilizes a poly-methylpentene membrane. Medtronic cannulas (17 Fr to 25 Fr) were used. The cardiovascular surgery team placed the vascular access cannulae peripherally using the modified Seldinger technique. For VA ECMO, the afferent cannula was placed through the femoral vein up to the union of the inferior vena cava with the right atrium, and the efferent cannula was placed in the femoral artery. For VV ECMO, the afferent cannula was placed in the right femoral vein, and the efferent cannula was placed in the superior vena cava through the internal jugular vein. Central cannulation of the VA ECMO was performed in only one case. The correct position of the cannulae was controlled with a transesophageal echocardiogram.

All patients were additionally evaluated and monitored by the nutritional support (NS) team following the ASPEN/ESPEN guidelines recommendations( and the guidelines of the institution implementing the nutritional treatment.

The patients were cared for by a multidisciplinary team of doctors, respiratory kinesiologists, nurses, and pulmonologist who received continuing training.

The continuous variables were reported as medians and IQRs after normality evaluation with the Kolmogorov-Smirnov test. The categorical variables were presented as absolute values and frequencies. The matched samples were analyzed with the Wilcoxon signed rank test. We utilized IBM SPSS Statistics 21.0 to conduct the statistical analysis. Results with values of p < 0.05 were considered statistically significant.

RESULTS

For the analysis, patients were divided into two groups according to baseline disease: Group I (15 perioperative lung transplantation patients) and Group II (7 patients with ARSD that did not undergo a lung transplant). The baseline characteristics and indications of both groups are described in tables 1 and 2.

Table 1

Group I characteristics

Patient	Age (years)*	Sex	Diagnosis	Charlson Score*	BT	Indication	Type of ECMO	Days of ECMO*	Days on MV*	Days in ICU*	Days of hospital stay*
1	56	F	PGF	2	No	Scheduled for surgery	VA/VV	4	35	41	72
2	57	F	COPD	4	No	PGF	VV	8	18	18	18
3	28	F	CF	1	Yes	Scheduled for surgery	VV	2	51	63	63
4	19	M	CF	1	Yes	PGF	VV	2	4	19	97
5	48	F	PF	1	No	Cardiogenic shock	VA	10	52	79	101
6	67	M	PF	7	No	Cardiogenic shock	VA	1	13	53	100
7	36	M	Bronchiectasis	1	Yes	Cardiogenic shock	VA	2	11	28	43
8	24	F	CF	1	Yes	Scheduled for surgery	VV	1	1	15	45
9	23	F	CF	1	Yes	PGF	VV	4	34	52	63
10	57	F	COPD	2	No	PGF	VV	5	15	35	57
11	35	M	CF	1	Yes	Scheduled for surgery	VA/VV	3	37	37	37
12	52	F	COPD	2	No	Scheduled for surgery	VA/VV	1	2	10	36
13	19	F	CF	1	Yes	Scheduled for surgery	VA/VV	2	3	26	31
14	25	F	CF	1	Yes	Scheduled for surgery	VA/VV	3	6	20	26
15	50	F	COPD	2	No	Cardiogenic shock	VA	5	46	58	89
Me (IQR)	36 (24.5 - 54)			2 (1 - 2)				3 (2 - 4.3)	15.5 (3 - 35)	31.5 (19 - 53)	60 (36 - 89)

BT - bilateral lung transplant; ECMO - extracorporeal oxygenation membrane; MV - mechanical ventilation; ICU - intensive care unit; F - female; M - male; PGF - primary graft failure; COPD - chronic obstructive pulmonary disease; CF - cystic fibrosis; PF - pulmonary fibrosis; VA - veno-arterial; VV - veno-venous.

Results are expressed as the median (25% - 75%).

Table 2

Group II characteristics

Patient	Age (years)*	Sex	Diagnosis	Charlson Score*	Indication	Type of ECMO	Days of ECMO*	Days on MV*	Days in ICU*	Days of hospital stay*
1	23	M	Multiple trauma	0	2nd ARDS	VV	2	12	6	6
2	23	M	POP thoracic surgery	6	1st ARDS	VV	15	17	18	18
3	58	F	CAP	3	1st ARDS	VV	7	32	38	42
4	45	F	Alveolar proteinosis	1	Hypoxemia	VV	2	15	25	27
5	56	F	CAP	1	1st ARDS	VV	6	8	8	8
6	21	M	Multiple trauma	0	1st ARDS	VV	11	37	37	37
7	44	M	POP esophagectomy	2	1st ARDS	VV	15	56	58	72
Me (IQR)	44 (23 - 50.5)			1 (0.5 - 2.5)			9 (3 - 14)	25 (13 - 37)	31 (11 - 38)	32 (11 - 41)

ECMO - extracorporeal membrane oxygenation; MV - mechanical ventilation; ICU - intensive care unit; F - female; M - male; POP - postoperative; CAP- severe community-acquired pneumonia; ARDS - acute respiratory distress syndrome; VV - veno-venous.

Results are expressed as the median (25% - 75%).

In Group I, veno-venous ECMO was indicated as a respiratory support in four patients with PGF. In seven patients with severe pulmonary hypertension (PHT) VA ECMO was utilized as a surgical strategy during the pulmonary transplant, and the VV modality was continued in the immediate postoperative period. In four patients, VA ECMO was indicated due to a circulatory collapse secondary to severe PHT and heart failure, of which three patients continued with VV ECMO until recovery. All of these patients were intubated and put on ventilation when the anesthesia for the transplant was initiated. Ventilation with nitric oxide (NO) was conducted during the first 48 hours as part of the pulmonary transplant protocol.( In 13 patients, cannulation was performed in the operating room, and in two patients, VV and VA ECMO were indicated in the ICU due to PGF and cardiogenic shock, respectively. In all cases, the initiation of ECMO was early (< 48 hour post-transplant).

In Group II, two patients suffered from ARDS secondary to multiple injuries (one patient required ECMO on two occasions; the second occasion was due to MV-associated pneumonia), two patients had postoperative ARDS, two patients had severe community-acquired pneumonias, and one patient was found to have alveolar proteinosis. Due to the presence of contraindications, only three patients were ventilated in the prone position prior to ECMO (hemodynamic instability, anterior thorax tube, and external fixation related to unstable pelvic fracture). Those who were ventilated in the prone position did not show improvement with this strategy. In all of the Group II cases, ECMO was indicated due to refractory hypoxemia, and the VV method was utilized. The median Pa/FiO2 before the start of ECMO was 82 (IQR: 58 to 109), and the median oxygenation index( was 22 (IQR: 19 to 30.2).

For patients in Group II, we were able to compare the mechanical ventilation data before and after ECMO, which showed that it was possible to reduce TV from a median of 6.3 ml/k theoretical weight (ktw) to 4.6ml/ktw (p = 0.01). In the same manner, the FiO2 could be reduced from a median of 1 to 0.4 (p = 0.007), and the gasometric criteria could also be improved with an increase in the median arterial pH from 7.29 to 7.43 (p = 0.01) and a reduction of PaCO2 from 58mmHg to 38mmHg (p = 0.01). Statistical significance was not achieved for the decrease in median plateau pressure from 26cmH2O to 24.5cmH2O (p = 0.07). This comparison was not possible in lung transplantation patients because in all cases except two, the ECMO placement occurred in the operating room almost simultaneously with the start of invasive mechanical ventilation. In these patients, mechanical ventilation during ECMO also conserved the protective strategy, with a median TV of 5.8ml/kpt (IQR: 4.3 to 7.2ml/kpt), median PEEP of 9.2cmH2O (IQR: 8 to 10cmH2O), median FiO2 of 0.4 (IQR: 0.3 to 0.4), median Ppl of 21.8cmH2O (IQR: 17.8 to 25.8cmH2O), and a median static complacency of 26.4ml/cmH2O (IQR: 17 to 35.8ml/cmH2O).

Sodium heparin in continuous infusion was used in 21% of patients with no evidence of complications associated with the absence of anticoagulation. The control condition for anticoagulation was conducted with aPTT.

Weaning from ECMO was initiated when the patients presented signs of improvement in pulmonary complacency, arterial oxygenation, and chest x-rays. The ventilation parameters were adjusted to a protective standard ventilation, and the flow of oxygen was gradually decreased. For anticoagulation, patient blood flow was reduced during a period of 2 - 3 hours. In patients without anticoagulation, the contribution of gases was gradually decreased while maintaining a constant blood flow to avoid thrombotic complications during the process.(

All patients in Group I were successfully weaned from ECMO. Three of the 15 patients died during hospitalization and had been previously removed from ECMO support due to infectious complications and multiple organ failure (MOF). In two of these patients, PGF was the cause of ECMO initiation.

In Group II, only three patients (42%) were successfully weaned from ECMO. Four patients died during ECMO, three due to MOF and one patient due to central nervous system bleeding.

All patients presented with complications during hospitalization. This information is shown in table 3. With regard to hemorrhagic and coagulopathic complications, a high variability in the blood platelet recount was observed, with an overall median of 180,000 platelets/mm3 (IQR: 69,625 - 260,950 platelets/mm3). All patients required transfusion of at least one blood product during the ECMO session; the median number of platelet units (U) transfused per patient was 8 U (IQR: 0 - 21.5 U); the median number of fresh frozen plasma units per patient was 2.5 U (IQR: 0 - 6 U); the median number of red blood cell units transfused per patient was 4 U (IQR: 1.5 - 8 U). In many cases, patients required platelet transfusion in spite of adequate platelet levels (> 50000U/mm3). Although the presence of bleeding in the cannula insertion site and at the surgical site was relatively frequent, major bleedings occurred in only three patients, one at central nervous system, leading to the patient’s death, and two at the surgical site requiring re-intervention in the operating room (one in a lung transplantation patient with central VA cannulation and another in a patient with a history of thoracotomy). One patient presented with complications of femoral artery cannulation and limb ischemia with compartment syndrome and subsequently underwent a suprapatellar leg amputation.

Table 3

Complications and results

	Group I	Group II
	(N = 15)	(N = 7)
Complications related to ECMO
Oxygenator failure	0	1 (14)
Bleeding
Cannula insertion site	3 (20)	2 (28)
Cerebral	0	1 (14)
Surgical site	3 (20)	1 (14)
Lung	2 (13)	1 (14)
Epistaxis	1 (6)	1 (14)
Thrombocytopenia	9 (60)	4 (57)
Coagulopathy	5 (33)	1 (14)
Limb ischemia	2 (13)	0
Complications not related to ECMO
Infections
VAP	6 (40)	4 (57)
CAI	3 (20)	1 (14)
Soft tissue	2 (13)	1 (14)
ARF + RRT	3 (20)	5 (71)
Pneumothorax	5 (33)	3 (42)
Collapsed lung	1 (6)	2 (28)
Arrhythmia	4 (26)	2 (28)
Mortality	3 (20)	4 (57)

ECMO - extracorporeal oxygenation membrane; VAP - ventilator-associated pneumonia; CAI - catheter-associated infection; ARF - acute renal failure; RRT - renal replacement therapy. Results are expressed as numbers (percentages).

As for NS, an average delay of six days to reach the total calories required and an average delay of nine days to reach protein requirements was observed. This result was related to the fact that 91% of patients presented with abdominal distension, 81.8% presented with constipation, 45.4% presented with gastroparesis, and 27.3% presented with diarrhea. In addition, 53.4% of patients had at least two episodes of hyperglycemia (> 200mg/dL). In 20% of the patients, it was necessary to implement parenteral nutrition.

DISCUSSION

The present ECMO program in a polyvalent adult intensive care unit was developed with the objective of guaranteeing adequate support for patients with hypoxemic respiratory failure after conventional treatment in cases of both ARDS and perioperative pulmonary transplantation.

The execution of ECMO is not free of complications, and although we consider its utilization in carefully selected patients to be safe, we must still consider that within our study, there was a central nervous system bleeding event that led the death of the patient and a lower limb ischemia event that required amputation. The frequency and type of complications did not differ from those published in the literature.( We identified the most important complications according to their frequency and impact as acute kidney failure requiring renal replacement therapy, infectious complications, and bleeding either at the site of cannula insertion or at the surgical site.

Multiple complications associated with anticoagulation existed during ECMO,( and there are various factors to consider when choosing anticoagulation, including the possibility of reversal in the case of hemorrhages, the bioavailability in cases of organ failure, the detection of clots in the circuit, and the presence of thrombocytopenia.( In our case, the population consisted of surgical patients with ARDS and lung transplantations with postoperative complications in which the initiation and maintenance of systemic anticoagulation was difficult due to hemorrhagic complications. Despite the reduced use of anticoagulation agents, we did not witness any thrombotic complications. Another point of controversy is when and how much to transfuse this type of patient. We used the presence of hemorrhages with a > 5 point drop in the hematocrit level or Hb < 7g/dL without the presence of bleeding as a criterion. Although patients generally exhibited increased transfusion requirements during ECMO, these requirements did not differ from those reported in prior studies.(

Regarding mechanical ventilation, the presence of ECMO allowed patients with lower pressures and volumes to be ventilated without complications due to acidosis. This is perhaps one of the most important functions of ECMO in patients who have heterogeneous lung lesions where ventilation strategies could potentially be harmful.(

It is necessary to fully understand the impact of ECMO on the alimentary canal. The evidence suggests that early enteral nutrition can be well tolerated if the possible barriers to its implementation are controlled.( In our experience, the appropriate combination of nutrients and calories was difficult and took some time to achieve. This result is probably related to the high rate of complications associated with enteral nutrition (above all abdominal distension). For that reason, we believe that emphasis should be placed on optimizing NS, possibly by adjusting protocols in this subgroup of patients to assure adequate nutrition in the most critical period of this disease.

Finally, although the use of ECMO in the lung transplantation postoperative period does not yet have well-defined indications, this type of technology is a very useful support tool for the most serious patients during their most critical period. From this experience, we obtained an overall survival of 80% at discharge, with a survival of 50% in the particular case of PGF. Both of these results are similar to or higher than those reported in the literature.(

CONCLUSION

Despite the controversy concerning the patients who would most likely benefit from extracorporeal membrane oxygenation, we believe that the current evidence should compel university centers and other institutions with a high volume of critical care patients to develop a program that offers this therapeutic option to a select population. We consider the results to be satisfactory despite the steep learning curve that we experienced.