Effect of Early Mobilization from Bed to Wheel Chair on Regional Ventilation Distribution Assessed by Electrical Impedance Tomography in Respiratory Failure Patients

Backgrounds: There was limited knowledge about the effect of early mobilization on regional lung ventilation in patients with respiratory failure. The aim of the study was to examine whether electrical impedance tomography (EIT) could help to predict the improvement in ventilation distribution due to mobilization. Methods: Forty-one patients with respiratory failure, who had weaned from ventilator and received early mobilization were prospectively enrolled in this study. EIT was used to assess regional lung ventilation distributions at 4 timepoints during the early mobilization from bed to wheelchair (T base : baseline, supine position at the bed, T 30min : sitting position on the wheelchair after 30min, T 60min : sitting position on the wheelchair after 60min, T return : return to supine position on the bed after early mobilization). The EIT-based global inhomogeneity (GI) and center of ventilation (CoV) indices were calculated. EIT images were equally divided into four ventral-to-dorsal horizontal regions of interest (ROIs 1-4). Depending on the improvement of ventilation distribution in dependent regions at T 60min (threshold set to 15%), patients were divided into recruited (DR) and non-recruited (Non-DR) groups. Results: From the bed to the wheelchair, a signicant and continuous increase of dependent regional ventilation distribution (ROI 3+4: baseline vs. T 30min , vs. T 60min : 45.9±12.1 vs. 48.7±11.6 vs. 49.9±12.6, p=0.015) and COV (COV baseline vs. T 30min , vs. T 60min : 48.2±10.1 vs. 50.1±9.2 vs. 50.5±9.6, p=0.003). Besides, there was a signicant decrease of GI at T 60min . Patients in the DR group (n=18) had signicantly higher oxygenation than the Non-DR group (n=23) after early mobilization. ROI4 Tbase was signicantly negatively correlated to ΔSpO 2 (R=0.72, p (cid:0) 0.001). Using a cut-off value of 6.5%, ROI4 Tbase had a 79.2% specicity

Besides, there was a signi cant decrease of GI at T 60min . Patients in the DR group (n=18) had signi cantly higher oxygenation than the Non-DR group (n=23) after early mobilization. ROI4 Tbase was signi cantly negatively correlated to ΔSpO 2 (R=0.72, p 0.001). Using a cut-off value of 6.5%, ROI4 Tbase had a 79.2% speci city and 58.8% sensitivity to predict response of dependent region recruitment due to early mobilization. The corresponding area under curve was 0.806 (95%CI, 0.677-0.936).
Conclusions: EIT may be a promising tool to predict the ventilation improvement resulted from early mobilization.
Some research reported that early mobilization can increase ventilator-free days and prevent postoperative respiratory complications [6][7][8]. However, other study reported early mobilization did not show an expected outcome of decreasing mechanical ventilation days [9][10][11]. Considering the human resources, it is controversial whether all patients with mechanical ventilation need early mobilization. A part from improvements in strength and functional status, to promote the recovery of lung function is also a main purpose of early mobilization. On one hand, early mobilization can induce body position changing so as to in uence the movement of diaphragm [12]. On the other hand, it is still not clear whether early mobilization can really promote lung recruitment in dependent area or not.
Electrical impedance tomography (EIT) is a noninvasive, non-ionizing monitoring system, which can obtain real-time imaging of regional lung ventilation at bedside by monitoring electrical impedance changes of the underlying tissue [13,14]. EIT has already been used in Positive End Expiratory Pressure(PEEP) titration of acute respiratory distress syndrome (ARDS) patients [15,16]. To our best knowledge, no one has investigated the change of regional lung ventilation during the process of early mobilization. The aim of the study was to examine whether EIT could identify the impact of early mobilization on pulmonary ventilation distribution, and screen out those patients who bene t more from early mobilization, in line with individualized treatment.

Patients
This study for human subjects was approved by The Institutional Research and Ethics Committee of the Peking Union Medical College Hospital (JS-1896). Written informed consent was obtained from all patients before enrollment.
Adult critically ill patients with respiratory failure who received early mobilization were enrolled in this study. Inclusion criteria were the following: (1) presentation of respiratory failure-Partial pressure of oxygen/Fraction of inspiration oxygen(PaO 2 /FiO 2 ) 300, oxygen saturation(SpO 2 )< 92%, or respiration rate(RR) > 25 bpm. (2) Weaned from mechanical ventilation 24h. (3) need to carry out early mobilization according to patient's condition (e.g. ICU duration≥1 week, high risk of respiratory complication). Patients were excluded from the study if they were aged < 18 years, were pregnant, had a body mass index (BMI) over 50 kg/m 2 , had ribcage malformation, had any contraindication against using EIT monitoring (automatic implantable cardioverter de brillator, chest skin injury, etc.), or had any contraindication against early mobilization (hemodynamic instability, FiO 2 60%, arrhythmia, etc).

Physiological measurements
We collected baseline data at enrollment, including age, sex, ICU duration, days of intubation, type of surgery. Respiratory parameters like RR, SpO 2 , FiO 2 , Tidal volume, and hemodynamic parameters, including heart rate (HR), mean arterial pressure (MAP) were obtained at different time points.

Experimental protocol
Participants went through early mobilization as sitting on the wheel chair. The patient's condition before early mobilization was de ned as baseline, with supine position on the bed. Then patients were moved from bed to wheelchair and sitting for 1 hour. After that, patients will return to supine position on the bed. During this movement, patients had EIT belt xed on their chest and monitoring regional lung ventilation at 4 time points (T base : baseline, supine position before early mobilization, T 30min : sitting position on the wheelchair for 30min, T 60min : sitting position on the wheelchair for 60min, T return : return to supine position on the bed after early mobilization.) Moreover, patients were spontaneous breathing with xed FiO 2 during the procedure.
EIT measurements EIT measurements were performed with PulmoVista500 (Dräger Medical, Lübeck, Germany). During the protocol, a silicone EIT belt with 16 electrodes were attached on the surface of patient's chest in one transverse plane corresponding to the 4th intercostal parasternal space and was then connected to the EIT monitor for bedside visualization. At different positions, we make sure the 16-electrode array was attached to the same level on the chest surface, so that the lung ventilation is comparable. EIT evaluates lung ventilation by collecting potential differences and the known excitation currents in a 32*32 pixel matrix [17]. Each scan pixel demonstrates the instantaneous relative local impedance change compared with a reference state of local impedance. The stimulation frequency and amplitude were adjusted automatically by the EIT device to minimize the in uence of background noise. EIT measurements were continuously performed at 20 Hz. In addition, the data were digitally ltered using a low-pass lter with a cutoff frequency of 0.67 Hz to eliminate cardiac-related impedance changes. EIT data were analyzed by o ine computer program.

Analysis of EIT data
The EIT data of 2 minutes at baseline and under different time points were acquired. Tidal images were calculated as the difference images between end-inspiration and end-expiration. An average of 2-min tidal images was calculated at each position to minimize the in uence of spontaneous breathing. Mean tidal images were divided into four symmetrical, non-overlapping ventral-to-dorsal horizontal regions of interest (ROIs), ranging from the gravity-independent area to the gravity-dependent area, namely, the ventral (ROI1), mid-ventral (ROI2), mid-dorsal (ROI3) and dorsal (ROI4) regions.
The global inhomogeneity (GI) index can indicate the overall change in inhomogeneity of ventilation and the status of local lung distribution [18]. In brief, the sum of differences between individual pixels to average value was calculated. This sum value was normalized to the amplitude of impedance tidal variation. A lower GI index value is related to a better ventilation homogeneity.

Page 5/18
The center of ventilation (CoV) describes the weighted geometrical center of the ventilation distribution [19][20][21]. When most of the tidal ventilation distributes to the dependent lung region, this results in a higher CoV value.
Patients were divided into two groups: (1)  Normal distribution was assessed with Kolmogorov-Smirnov normality test. Normally distributed results are presented as mean ± SD whereas non-normally distributed results are presented as median (25th-75th percentile). The Mann-Whitney test was used for comparisons of groups (recruitment group VS notrecruitment group). Comparisons of the trends of the related parameters according to the different positions were performed using a General Linear Model Repeated Measures (GLMRM) [22]. This model is an extension of the classical ANOVA, which allows handling both xed effect (different position) and random effect (patient). GLMRM takes into account the correlation between multiple measurements on one patient and thus the estimated marginal means were adjusted for the covariates and the trends of related EIT parameters corresponding to the different position. When Mauchly's test of sphericity is not demanded (p <0.05), Epsilon (Greenhouse-Geisser) was used for the corrected test. All statistics were twotailed, and a p value of less than 0.05 was considered to be signi cant.

Results
From Jan 2019 to Jan 2020, this study has enrolled 41 patients. The main characteristics of the patients are summarized in Table 1. The mean patient age was 64.3 ±14.1, and 33/41(80%) were men. 9/41(22%) of patients had tracheotomy. The mean days of ICU duration was 14.3 ± 13.3. Data are given as mean ± SD Effects of early mobilization on regional lung ventilation distribution The trend of ROI4 (dorsal region) signi cantly increased (p=0.01) after early mobilization, but ROI 1-3 showed no signi cantly difference during study phases. Moreover, GI signi cantly decreased(p=0.038) and COV signi cantly increased(p=0.047) at T 60min compared to baseline (Fig. 1).

Effects of early mobilization on respiratory and hemodynamic parameters
As for respiratory parameters, a signi cant increase was observed in SpO 2 (p=0.045) and RR(p=0.001).
Besides, Heart rates has signi cantly increased after early mobilization(p=0.011). But MAP remained stable over all study phases. (Table 2) Difference between DR and Non-DR groups 18 patients could be categorized to the DR group, and 23 patients to the Non-DR group. Baseline characteristics of two groups are compared in Table 3. There were no signi cant differences in age, PaO 2 /FiO 2 , HR, RR and days of ICU duration between these two groups at baseline.
Changes of respiratory and hemodynamic parameters in these two groups during study phases are shown in Table 4. There is a signi cant increase of SpO 2 in DR group, while there is no signi cant difference in Non-DR group after early mobilization.   Using ROI 4 (base) to distinguish DR patients If we use ROI 4 at baseline as a predictor to distinguish DR group from Non-DR group, ROC curve of ROI 4 (base) diagnostic value was showed in Figure 3. The best cut-off for ROI 4 (base) is 6.5%, which means that in patients with an initial ROI 4 (base) 6.5%,early mobilization may improve oxygenation and prefer to be performed. The speci city is 79.2%, while the sensitivity only 58.8%. Area under curve is 0.806(95%CI, 0.677-0.936).

Discussion
In this study, we focused on body position change from supine to sitting on the wheelchair. We found that 1) early mobilization can induce lung recruitment in some but not all patients. 2) EIT is feasible in monitoring regional ventilation distribution change during early mobilization. 3) ROI 4 (base) was negatively related to ΔSpO 2 and lung recruitment in the dependent regions, which suggested that baseline ventilation distribution in the most dorsal regions might predict the recruitment responsiveness of early mobilization.

Effects of early mobilization on lung recruitment
As shown in previous study, early mobilization may introduce lung recruitment to prevent postoperative respiratory complications and increase ventilator-free days [6][7][8]. But controversial results were found in other studies [9,10]. In our study, early mobilization can induce lung recruitment in gravity-dependent lung area in some, but not in all patients, which explained the controversial results in previous studies.
Furthermore, we divided all patients into two types:(a) in DR group, patients showed recruitment in the dorsal region and better oxygenation after early mobilization. (b) in Non-DR group, patients didn't take advantage from early mobilization with no difference in oxygenation and recruitment. Thus, patients classi ed as type DR may have a greater potential for recruitment, which indicates that early mobilization therapy may bene t these patients and is recommended. Patients classi ed as type Non-DR may have limited potential for recruitment and should be used cautiously. EIT on regional lung ventilation Early mobilization in uences regional lung ventilation by body position changing. Previous researches have used EIT to monitor ventilation improvement during body position changing [23][24][25][26][27] and proved EIT's feasibility. Hough et al [24] con rmed that body position changes might in uence regional ventilation distribution in infants, whereas Pauline S et al [25] reported no difference between lateral and supine positioning. Riedel et al [26] investigating supine and prone position by EIT and found ventilation in dependent lung region increased in the prone position.
In this study, we used EIT to monitor regional ventilation change from supine to sitting position. The results show that ROI4, CoV and GI improved in sitting position, which means more gas into dorsal region and ventilation became more homogenous. This result is in accord with previous study [28,29]. One study showed that FEV1 increased when healthy volunteers changed from supine to sitting position [28].
Another study found that sitting position constantly and signi cantly relieved expiratory ow limitation and auto-positive end-expiratory pressure resulting in a dramatic drop in alveolar pressures in critically ill obese patients under mechanical ventilation [29]. Due to the change of gravity direction and the shift of diaphragm, it was possible that EIT measurement included partially different lung tissue in supine and sitting positions. Nevertheless, the improvement from T 30min to T 60min con rmed the improvement of ventilation distribution in sitting position.

ROI 4 (base) , SpO2 and lung recruitment
A previous study showed that early mobilization could improve SpO 2 [30], but may not be applicable to all of ICU patients. In the present study, we found SpO 2 improvement in 78% patients in the DR group (14/18) and only 26% patients have improved SpO 2 in Non-DR group (6/23). In addition, we found that ΔSpO 2 and ΔROI(3+4)were positively related and ROI 4 (base) was negatively related to ΔSpO 2 , which means patients with lower baseline ROI 4 proportion may have higher possibility to get lung recruitment and improved oxygenation after early mobilization. In clinical practice due to limited manpower, position change to sitting isn't easy to carry out for ICU patients. So it will be bene cial to distinguish speci c group of patients who would bene t the most from early mobilization. The responsiveness to early mobilization could be partially predicted with ROI 4 base (Fig. 3).

Limitations
This study also has some limitations. First, we only monitored one-time of early mobilization for 1 hour and the study period may be not long enough to monitor regional ventilation changing. Second, the sample size might be considered not large enough for subgroup analysis. Third, during the 4 time points, patients need to be moved from supine position from bed to sitting position on wheelchair, although we had conducted it with cautious, the position of 16-electrode belt of EIT might be still slightly shifted. So we did not compare the end expiration lung impedance level. Fourth, the threshold used to distinguish recruited and non-recruited groups (15%) was arbitrary and its clinical value needs to be further studied.

Conclusion
Early mobilization can induce lung recruitment in dorsal lung regions and improve oxygenation in some but not all patients. The baseline ventilation distribution in the most dorsal regions might predict the recruitment responsiveness of early mobilization. EIT may be useful to identify patients who could bene t most from early mobilization.

Figure 3
The ROC curve of using ROI 4(base) to distinguish DR patients(AUC 0.806 )