Feasibility and safety of watershed detection by contrast-enhanced ultrasound in patients receiving peripheral venoarterial extracorporeal membrane oxygenation: a prospective observational study

Background In bifemoral venoarterial extracorporeal membrane oxygenation (VA ECMO), the transition point at which the antegrade pulsatile output from the left ventricle and the retrograde non-pulsatile ECMO output collide is referred to as watershed [1]. Currently, no standard method is available to determine its location. Occasionally, contrast-enhanced computed tomography (CT) or angiography has been used [1–3]. Both techniques, however, bear disadvantages including radiation exposure and use of iodinated contrast media. We assessed the feasibility and safety of contrast-enhanced ultrasound (CEUS) to detect the watershed at the bedside in patients on bifemoral VA ECMO at three ICUs of a European tertiary care facility.


Background
In bifemoral venoarterial extracorporeal membrane oxygenation (VA ECMO), the transition point at which the antegrade pulsatile output from the left ventricle and the retrograde non-pulsatile ECMO output collide is referred to as watershed [1]. Currently, no standard method is available to determine its location. Occasionally, contrast-enhanced computed tomography (CT) or angiography has been used [1][2][3]. Both techniques, however, bear disadvantages including radiation exposure and use of iodinated contrast media. We assessed the feasibility and safety of contrast-enhanced ultrasound (CEUS) to detect the watershed at the bedside in patients on bifemoral VA ECMO at three ICUs of a European tertiary care facility.

Methods
CEUS was performed as soon as possible after ECMOinitiation (Cardiohelp, Maquet, Germany) using Sono-Vue contrast media (Bracco, Italy). Transesophageal echocardiography (x7-2t probe) and transabdominal sonography (3-5 MHz curvilinear probe) were performed concomitantly to display mid-esophageal aortic valve, ascending, descending aorta, and upper esophageal aortic arch long-axis views as well as longitudinal views of the proximal (below diaphragm), mid (level of renal arteries), and distal (above iliac bifurcation) abdominal aorta. The mechanical index was set to 0.05-0.10 field of view. Prior to CEUS, the arterial bubble sensor activating zero-flow mode was disabled. The acoustic alarm was kept active. The presence or absence of pulsatility in the left radial artery was documented.
One milliliter of SonoVue was administered via the venous drainage cannula, followed by a flush of 10 ml normal saline. The obtained images were evaluated qualitatively. If a watershed area was not able to be visualized, contrast-enhanced blood flow was classified into "pulsatile" or "continuous" to discriminate between cardiac and ECMO blood flow. The feasibility of CEUS was assessed based on qualitative image evaluation, the amount of contrast media administered, and the rate of bubble detection. Secondary outcomes were safety and frequency of radial arterial pulsatility. Safety variables included ECMO settings, hemodynamics, and neurologic assessment and were obtained over a 6-h period after CEUS.
The variables are presented as absolute values (n), relative frequencies (%), and median (25-75% IQR). We used random-effects general linear regression models to estimate mean changes for each safety variable (mean ± SD).

Results
Between August 2018 and April 2019, ten patients were enrolled (Table 1). Qualitative detection of watershed location by CEUS was feasible using 1 ml contrast media. In five patients, the watershed could be clearly shown in the abdominal aorta, seconds after contrast media administration (Fig. 1). In the remaining five patients, contrast-enhanced continuous blood flow was visible throughout the abdominal and thoracic aorta indicating watershed location close to the aortic root. The pulsatility of the left radial arterial waveform and opening of the aortic valve was present in all patients. Acoustic bubble detection occurred in all patients after CEUS. No changes in the safety variables related to CEUS occurred (Table 1). CT imaging of the brain (8/10 patients) showed no cerebral lesions suggesting particle embolism.

Discussion
This study assessed the feasibility of CEUS for watershed detection at the bedside in patients on bifemoral VA ECMO. CEUS was apparently safe and provided realtime assessment of the watershed or contrast-enhanced continuous blood flow in the aorta. Increasing evidence indicates that CEUS is safe in critically ill patients, and application areas are ever-expanding [4][5][6]. In bifemoral VA ECMO, CEUS may help to identify patients at risk for differential hypoxia, given that left radial arterial pulsatility was present in all study patients, including those in whom the watershed was located near the aortic root.

Limitations
Transthoracic suprasternal echocardiography may be useful to localize the watershed in the aortic arch but has not been tested. Furthermore, no reference imaging technique has been used to assess the performance of CEUS, because no standard method for the detection of the watershed is available, and no repeated measurements were performed.
Abbreviations CEUS: Contrast-enhanced ultrasound; CT: Computed tomography; VA ECMO: Venoarterial extracorporeal membrane oxygenation Fig. 1 Visualization of contrast-enhanced retrograde non-pulsatile ECMO blood flow. The watershed is marked with an arrow and located distal to the superior mesenteric artery