Contradictory responses of the microcirculation to changes in extracorporeal membrane oxygenation pump flow in patients on venoarterial extracorporeal membrane oxygenation: a prospective observational study

Background: Venoarterial extracorporeal membrane oxygenation (VA-ECMO) pump flow is crucial for maintaining tissue and organ perfusion in patients with cardiogenic shock. Determining the optimal VA-ECMO pump flow rate to maintain adequate tissue perfusion remains a clinical challenge. Poor microcirculation is associated with higher mortality in patients using VA-ECMO. Change in VA-ECMO pump flow is expected to result in a coherent change in microcirculatory blood flow, but this has not been investigated. This study investigated the effect of altering VA-ECMO pump flow rate on sublingual microcirculation. Methods: Sublingual microcirculation images were recorded using an incident dark-field hand held vital microscope at two time points, within 24 h (T1) and at 24-48 h (T2) after VA-ECMO placement. Microcirculation was measured before and 5 minutes after the changes in VA-ECMO pump flow rate at each time point. Events of changing VA-ECMO pump flow rate at T1 and T2 were divided into four groups according to changes in perfused vessel density (PVD): Group A, increased pump flow rate and increased or sustained PVD; Group B, increased pump flow rate and decreased PVD; Group C, decreased pump flow rate and increased or sustained PVD; and Group D, decreased pump flow rate and decreased PVD. The microcirculatory parameters, clinical parameters, and prognosis of each subgroup were recorded. Clinical parameters on 14-day and 28-day survivors and non-survivors were also compared. Results: A total of 25 patients were enrolled, and 38 events with good-quality images at T1 and T2 were categorized. Opposing response of microcirculation was observed in 43.5% of events when VA-ECMO pump flow rate was increased. Microcirculation was decreased in 33.4% of events when VA-ECMO pump flow rate was reduced. No predictive values in microcirculatory or macrocirculatory parameters before changing VA-ECMO pump flow rate were identified.


Study design and patient selection
This prospective observational study was approved by the Research Ethics Committee of National Taiwan University Hospital (approval number: 201703011RINA) and registered on the ClinicalTrials.gov protocol registration system (ID: NCT03210818). It was conducted at National Taiwan University Hospital between November 2017 and December 2018.
Participants were selected from patients receiving ECMO support on the basis of eligibility screening conducted within 12 hours following ECMO placement. Patients who received VA-ECMO support and were between the ages of 20 and 90 years were included. Patients were excluded if they declined to participate, had received re-implantation of ECMO, died within 12 hours, or had circumstance that prevented sublingual microcirculation from being measured within 24 hours after initiating VA-ECMO, such as those in which placement occurred in the evening or the research nurse was on leave. Informed consent was obtained from patients' legally authorized representatives before study enrollment.

VA-ECMO components and placement
For all enrolled patients, the placement and principal components of the VA-ECMO were the same as described in our previous study [9]. To avoid possible malperfusion of the distal limb, an antegrade distal perfusion catheter was used when the mean pressure of the superficial femoral artery was below 50 mm Hg [12]. All patients received standard VA-ECMO management and routine intensive care unit (ICU) care. Heparin was continuously administered to maintain an activated clotting time of 160-180 s if no active bleeding or other complications were observed. The following data were recorded: age, gender, height, body weight, sequential organ failure assessment (SOFA) score [13], indications for VA-ECMO, VA-ECMO pump flow rate, heart rate, mean arterial pressure (MAP), lactate level, activated clotting time, hemoglobin level, fluid balance, and inotropic score. The inotropic score was calculated as 100 × epinephrine dose (mcg/kg/min) + 100 × norepinephrine dose (mcg/kg/min) + dopamine dose (mcg/kg/min) + dobutamine dose (mcg/kg/min) [14]. The simultaneous use of intra-aortic balloon pump (IABP) with VA-ECMO support was recorded. The length of ICU and hospital stay, as well as survival status at 28 days were also recorded.

Change in VA-ECMO pump flow rate and recording of sublingual microcirculation images
The ECMO team adjusted the VA-ECMO pump flow rate to maintain an MAP > 60 mm Hg, central venous oxygen saturation >70%, central venous pressure <15 mmHg, and lactate level of <3 mmol/L, and to avoid a urine output of <0.5 mL/kg/h, pulse pressure < 10 mm Hg, and ECMO-induced hemolysis, or to wean the patients off VA-ECMO support. Sublingual microcirculation images were recorded using an incident dark-field video microscope (CytoCam, Braedius Medical, Huizen, Netherlands) [15]. The images were recorded at two time points: within 24 hours (T1) and at 24 to 48 hours (T2) after VA-ECMO placement. At T1 and T2, images were recorded before changing VA-ECMO pump flow rate. After recording, the VA-ECMO pump flow rate was increased or reduced by the ECMO team member according to the treatment plan. The change in pump flow rate was recorded, and images were recorded at 5 minutes after the change.

Measurements of sublingual microcirculation
At each time point, five video sequences (length: 6 s each) were recorded at different sublingual sites and were digitally stored with code numbers to ensure the anonymity of patient information. Subsequent offline analyses were performed by a single observer blinded to patient information according to the international consensus guidelines for performing sublingual microcirculation by a Task Force of the European Society for Intensive Care Medicine [16]. Two or three sequences with appropriate image quality were selected for analysis using the semi-automated analysis software package Automated Vascular Analysis 3.0 [17].
In accordance with the afore mentioned consensus guidelines [16], the following parameters were investigated: (a) total vessel density (TVD; vessels less than 20 μm), (b) perfused vessel density (PVD), (c) proportion of perfused vessels (PPV), and (d) microvascular flow index (MFI) score. The software was used to automatically calculate TVD. The calculation of PVD was semiautomated using the procedure described in our previous study [9]. The MFI scores were semiquantitatively calculated according to the suggestions made at the roundtable conference [18].

Grouping events of changes in VA-ECMO pump flow rate and changes in PVD
Events of changing VA-ECMO pump flow rate at T1 and T2 were divided into the following four groups according to the changes in PVD. Group A included events of increased or sustained PVD after increasing the ECMO pump flow rate; Group B included events of reduced PVD after increasing the ECMO pump flow rate; Group C included events of increased or sustained PVD after decreasing ECMO pump flow rate; and Group D included events of reduced PVD after decreasing ECMO pump flow rate.

Statistical analysis
All statistical analyses were performed using SPSS version 20 (IBM, Armonk, NY, USA).
Normally distributed numerical data were expressed as means (standard deviation) and compared using t-tests. Nonnormal distribution of numerical data were expressed as medians (interquartile range) and compared using Mann-Whitney tests. Categorical variables were described as percentages and were compared using chi-square tests or Fisher's exact tests as appropriate. A p value of < 0.05 indicated a significant difference.

Patient distribution and characteristics
A total of 70 patients receiving VA-ECMO were considered for inclusion in this trial ( Figure 1). A total of 45 patients were excluded. The enrolled 25 patients had 50 events of changing VA-ECMO pump flow rate at T1 and T2. A total of 38 events with good quality of microcirculation images were enrolled, 19 events at T1 and 19 events at T2 (Figure 1).
Detailed indications of individual patients regarding VA-ECMO applications, key variables, 14-and 28-day survival, survival to discharge, and clinical outcomes are listed in Table 1.
Values for basic patient characteristics, indication for VA-ECMO application, IABP use, fluid status at each time point, ECMO pump flow rate, inotropic score, and T1 SOFA score of the 14-and 28-day survivors and nonsurvivors are presented in Table 2. The T1 inotropic score was lower for the 14-day survivors than for the nonsurvivors. The T1 fluid balance, T2 ECMO initial pump flow rate, and T1 SOFA score were significantly lower for the 14-and 28-day survivors than for the nonsurvivors.

Hemodynamic parameters and lab data at different time points
Values for the hemodynamic parameters and lab data at T1 and T2 are presented in Table   3. At T1, central venous oxygen saturation level was lower in Group A than in Group B [73 (4) % vs. 85 (7) %, p = 0.012], and changes in MAP is higher in Group A than in Group B [5 (4) vs. 0 (4) mm Hg, p = 0.048]; but other hemodynamic parameters did not differ significantly between these groups. No statistical differences of hemodynamic parameters and lab data could be determined between Group C and D at T1. At T2, MFI score before the changes in VA-ECMO pump flow rate was lower in Group A than in Group B [1.8 (0.6) vs. 2.6 (0.1), p = 0.020], but the observed MAP, inotropic score, lactate level, and central venous oxygen saturation level did not differ significantly between Group A and B.
Hemodynamic parameters and lab data between Group C and D showed no significant differences at T2.  Table 4. No significant differences in microcirculatory parameters before changes in VA-ECMO pump flow rate were observed between Group A and B or between Group C and D.

Discussion
We found divergent effects of changing the VA-ECMO pump flow rate on microcirculatory convective capacity as measured by PVD. Our results indicated that microcirculation was decreased in 43.5% of events when the VA-ECMO pump flow rate was increased and in 33.3% of events when the VA-ECMO pump flow rate was reduced. In Group B, PVD and MFI decreased after the VA-ECMO pump flow rate was increased, suggesting that an inappropriate VA-ECMO pump flow support impaired microcirculation. In addition, our results revealed that initial ScvO 2 and changes in MAP were different in Group A and B within 24 hours after VA-ECMO placement. We suggest that further studies are required to investigate the predictability of initial values and delta changes of macrocirculatory and microcirculatory variables on changes in microcirculation following a change in VA-ECMO pump flow rate.
Our main finding was that in a significant portion of patients increasing VA-EMCO pump flow rate instead of increasing tissue perfusion as would be expected, the converse occurred and PVD decreased. Thus, increasing VA-ECMO pump flow may cause a paradoxical decrease in tissue perfusion. Two mechanisms may cause this condition. First, decreased cardiac preload caused by increased venous drainage reduces the venous return of blood flow to the right atrium. At the same time, cardiac afterload and vascular resistance could increase as a result of increased VA-ECMO arterial blood flow [19,20].
The decreased preload and increased afterload may reduce the patient's own cardiac output and sublingual microcirculation. Second, the blood mixing zone moves because of alteration in the proportion of VA-ECMO blood supply. According to the results of a simulation study conducted by Stevens et al. [21], brain perfusion is determined according to the balance between VA-ECMO pump flow rate and the patient's own cardiac output when the mixing zone is proximal to the aortic arch [21,22]. If the patient's residual cardiac function fails to overcome the upcoming pressure work when the VA-ECMO pump flow rate increases, then the patient's own cardiac output will be offset. When the mixing zone moves closer to the aortic arch, the brain blood flow rate decreases. Brain hypoperfusion leads to reduced oxygen delivery and can significantly impair brain function and cause neurological complications. In a study conducted by Nasr et al., 4% of patients treated with VA-ECMO had an ischemic stroke, which led to a higher rate of discharge to long-term facilities, longer length of stay, and higher hospitalization costs [23]. Subsequent neurological complications, associated with poor outcomes and higher mortality, has become some of the main concerns during VA-ECMO support [24].
Measurement of sublingual microcirculation may provide an effective measure of whether this condition occurs. An animal study conducted by Boucek et al. revealed moderate correlations between cerebral regional oxygen saturation (rSO 2 ) and changes in microcirculatory parameters during cardiopulmonary resuscitation [25]. Moreover, brain rSO 2 correlated well with global cerebral perfusion [25]. Tamosuitis et al. found that MFI scores were significantly lower for brain and sublingual mucosa of brain-dead patients compared with those of healthy controls [26]. Therefore, our results suggest that sublingual microcirculation may reflect brain perfusion when the VA-ECMO pump flow rate is being altered.
Another finding of our study is that microcirculation improved when the VA-ECMO pump flow rate was reduced. This may indicate that the patient's own cardiac output is sufficient and that the demand for VA-ECMO support is lower. Two possible mechanisms may contribute to this finding. First, a higher VA-ECMO pump flow rate causes a larger reduction in cardiac preload, stroke volume, and pulse pressure, and a larger increase in cardiac afterload [20]. This reminds us that a VA-ECMO pump flow rate that is too high may contribute to decreased cardiac output and left ventricular overload [27][28][29]. The second possible mechanism is recovery of the patient's cardiac function and the suitability of weaning from VA-ECMO. Patients in this condition who undergo prolonged VA-ECMO treatment are exposed to unnecessary risks of vascular complications, neurological complications, renal injury, bleeding, and infection [30]. Sublingual microcirculation monitoring may help to prevent excessively high VA-ECMO pump blood flow rate and accelerate the weaning process of VA-ECMO. In Group D, microcirculation decreased     Tables.docx   Tables.docx