Acute respiratory distress syndrome (ARDS) is diagnosed in approximately 5% of mechanically ventilated patients in intensive care units (ICUs) [1]. Severe ARDS may dynamically deteriorate respiratory failure. The mortality in patients with ARDS is still very high, and reaches 40% of mechanically ventilated individuals [2]. For many severely ill patients, the last chance therapy to improve the outcome is implementation of veno-venous extracorporeal membrane oxygenation (V-V ECMO) [3, 4]. The main aim of ECMO therapy implementation is to support exchange of blood gases by removal of carbon dioxide and provision of oxygen directly to the patients’ bloodstream [5]. Undergoing V-V ECMO therapy may be associated with a number of complications, some of which potentially fatal. Two main sources of adverse events are thrombotic and hemorrhagic complications [6]. In order to prevent complications, units that conduct extracorporeal oxygenation require adequate expertise and constant training. Moreover, the experience of the medical team increases patient survival rate [7, 8]. Therefore, V-V ECMO-dedicated centers have been introduced in many countries [9]. Additionally, many patients suffering from severe ARDS, who may require ECMO, are distributed in regional ICUs. Thus, a dedicated experienced retrieval team is required in order to implement ECMO therapy and transfer the patient safely to the ECMO center [10]. Transport of a patient with severe ARDS on ECMO remains challenging and many adverse events have been reported [6, 7].
Few articles can be found regarding Polish experience with ECMO in the literature. The most important initiative is the program “Extracorporeal Membrane Oxygenation for Greater Poland” [10]. It shows an important role of medical simulation in skills testing and creating new, necessary algorithms and non-existing procedures [11]. However, this program is still in progress.
The goal of our study was to assess the safety and feasibility in patients transported on ECMO into our retrieval ECMO center.
METHODS
Population included
This was a retrospective, single-center, case-series study. We extracted data from the hospital’s ECMO database from March 2016 to June 2019, including 38 cases (average number of patients treated per annum is approximately 15 cases). Our center is mainly focused on treating patients with severe respiratory failure and we perform only V-V ECMO support due to lack of a cardiosurgical unit. Only patients who were retrieved to our hospital from regional ICUs were analyzed (31 out of 38 cases – 81.6%).
Ethical approval for this study (permit number KE-0254/37/2018) was provided by the Medical University of Lublin Ethics Committee.
Analyzed parameters
We recorded patients’ diagnosis at admission, baseline demographics, the Sequential Organ Failure Assessment (SOFA) and the Respiratory ECMO Survival Prediction (RESP) scoring systems, laboratory parameters at admission, duration of ECMO therapy and mechanical ventilation time, and the patient survival rate until the ICU discharge.
Goals
The primary aim of our study was to present complications during transport of patients on V-V ECMO from regional hospitals. Our secondary goal was to compare initial laboratory and demographic data in survivors and non-survivors of ECMO therapy.
Patient qualification
The decision to initiate ECMO support was performed by two intensivists according to Extracorporeal Life Support Organization (ELSO) guidelines [12] and recommendations of the Polish National Consultant in the field of Anaesthesiology and Intensive Therapy[13] (Tables 1 and 2).
TABLE 1
Extracorporeal membrane oxygenation inclusion criteria according to Extracorporeal Life Support Organization and Polish National Consultant in the field of Anaesthesiology and Intensive Therapy [11, 12]
TABLE 2
Extracorporeal Life Support Organization exclusion criteria [19]
During a telephone interview the inclusion and exclusion criteria were considered. The decision to implement the therapy was determined up to 24 hours. All vital equipment was collected and checked according to our center checklist presented below (Table 3).
TABLE 3
Equipment check-list applied in our center
Implementation
For cannulation, single lumen cannulae were used (Maquet 15–25 Fr). At the bedside, ultrasound (USG) guided percutaneous cannulation was performed. We preferred the femoral vein as the collecting line, and the internal jugular or the opposite femoral vein as a returning cannula. Apart from one case, V-V ECMO support was implemented before the transfer to our center. The cannulation sites and gauges are presented in Table 4.
Transport
The transport team included a specialist in anesthesia and intensive care, a resident in training and two paramedics. During the transport, all patients, apart from one, were both mechanically ventilated and oxygenated by an ECMO machine. We monitored vital signs: blood oxygen saturation (SpO2), electrocardiography (ECG), respiratory rate (RR) and invasive blood pressure (IBP). In every case, sedatives and neuromuscular blocking agents were used. Circulation was supported with catecholamine infusion as required.
Ventilation during ECMO support
Mechanical ventilation during ECMO was adjusted to 10–15 cm H<sub>2</sub>O value of positive end-expiratory pressure (PEEP) and fraction of inspired oxygen (FiO2) was set to 0.6. A PEEP trial (to find optimal settings) and measurement of static and dynamic compliance were performed at least twice daily. If required, muscle relaxation was continued for 48 hours. Anticoagulation management included the use of low molecular weight heparin only.
RESULTS
We assessed 38 patients from the hospital’s ECMO database for eligibility. Thirty-one patients (81.5%) met the inclusion criteria. Patient characteristics, admission diagnosis, RESP and SOFA scores, some laboratory parameters, and ECMO and mechanical ventilation time are presented in Table 5.
TABLE 5
Baseline characteristics of patients transported on veno-venous extracorporeal membrane oxygenation. Data obtained prior to initiation of therapy
The ECMO retrieval team has not encountered any significant problems in the referring hospitals, besides one case of an unsuccessful cannulation attempt – the patient was transported to our center without ECMO support, using a standard ventilator with FiO2 of 1.0 (Table 7).
TABLE 6
Primary etiology of severe respiratory failure in patients transported on veno-venous extracorporeal membrane oxygenation
TABLE 7
Characteristics of ground transport and complications on veno-venous extracorporeal membrane oxygenation
The median distance and ECMO transport time were 100 km and 70 min, respectively. All transports were made by an ambulance. The ambulance was rented from an ambulance service company, ready to use in up to two hours from a call. An important part of the ambulance equipment was an electric generator. All other necessary devices were prepared before the transport and checked according to a specific checklist (Table 3).
The survival rate until the patient discharge was 64.51% (20 patients). The median RESP score and SOFA score at the admission were 3 and 11 points. The predicted survival rate according to these scoring systems were 65% and 60%, respectively.
We identified 10 (32.25%) complications during the transport, as shown in Table 7, but no major complications occurred. In one case, ECMO was not initiated at the site due to prolonged cannulation. None of these complications affected patient mortality.
The mean length of mechanical ventilation before ECMO implementation was 2.5 days (Table 5), while the mean ECMO support time was 6.56 days (Table 7).
We found that higher body mass index (BMI) (33.5 vs. 26.5; P = 0.00251) and lower serum lactate level (1.25 vs. 1.6; P = 0.0058) on the day of ECMO implementation were a positive predictors of survival until ICU discharge (Table 5).
DISCUSSION
The findings presented in our study showed that the transport of patients on ECMO was relatively safe. None of the minor complications during the transportation affected the patient mortality (Table 7). According to the available data a standard ambulance, without sophisticated equipment, is adequate for this purpose [13]. Higher patient BMI and lower serum lactate level on the admission day significantly decreased patient mortality. We found an association between higher BMI, lower admission serum lactate level and increased survival rate in our population of V-V ECMO patients (Table 5).
Our results regarding the transportation are consistent with the body of literature – transport of patients on V-V ECMO is a relatively safe procedure. Nevertheless, inter-hospital transport of critically ill patients is a big challenge with possible adverse events. The incidence of severe complications during transport of high-risk patient without ECMO is 30% [15, 16]. In the literature, the number of adverse events during ECMO transport varies widely, from 0% up to 42% [17–19]. In most cases, complications occurring during the transport had no adverse effect on patient outcome [7]. The complications could be divided according to 4-grade severity risk categories by Fletcher-Sandersjöö et al. [7]. Alternatively, these complications can be categorized as related to equipment, human error, patient, transport vehicle, or environment. Foregoing data suggest that transport of patients on ECMO to specialized ECMO centers is safe and effective [7, 20, 21]. Importantly, specialized retrieval teams are the main reason for reduction of life-threatening complications due to adequate training and equipment [9, 12, 14].
ELSO highlighted that the best outcome is achieved when V-V ECMO is instituted as quickly as possible [12]. Nonetheless, our study shows no difference in mortality between patients who were mechanically ventilated for 2 or 4 days (P = 0.078) (Table 5).
According to the literature, lactate level is shown to be a useful prognostic tool in the population of V-V ECMO patients [22]. Bonizzoli et al. observed a statistically significant difference in the initial lactate level between survivors and non-survivors (2.68 mmol/L vs. 4.95 mmol/L; P = 0.002), which corresponded to our results (1.25 mmol/L vs. 1.6 mmol/L; P = 0.0058 respectively). Our observation regarding a significant difference in BMI between the group of survivors and non-survivors (33.5 vs. 26.5; P = 0.025) is supported by the data from the mentioned study (26.7 vs. 24.6; P = 0.004 respectively) (Table 5).
Unexpectedly, we found out that patients with the median BMI 26.5 kg m-2 have a higher mortality rate in comparison to individuals with median BMI 33.5 kg m-2 (P = 0.0251) (Table 5). The efficacy of V-V ECMO treatment in obese patients with severe ARDS has already been shown [23]. Nonetheless, there is still a lack of data which support the thesis that a higher BMI can improve outcome in patients on V-V ECMO.
LIMITATIONS
Our study has multiple limitations. Due to the retrospective design of the study not all data were obtainable. There was a lack of data on the number of patients rejected from ECMO and no implementation of monitoring protocols during the transport. Finally, the relatively small number of patients included in the study may limit the overall generalisability of the study findings.
