Can bronchoconstriction and bronchodilatation in horses be detected using electrical impedance tomography?

Abstract Background Electrical impedance tomography (EIT) generates images of the lungs based on impedance change and was able to detect changes in airflow after histamine challenge in horses. Objectives To confirm that EIT can detect histamine‐provoked changes in airflow and subsequent drug‐induced bronchodilatation. Novel EIT flow variables were developed and examined for changes in airflow. Methods Bronchoconstriction was induced using stepwise histamine bronchoprovocation in 17 healthy sedated horses. The EIT variables were recorded at baseline, after saline nebulization (control), at the histamine concentration causing bronchoconstriction (Cmax) and 2 and 10 minutes after albuterol (salbutamol) administration. Peak global inspiratory (PIFEIT) and peak expiratory EIT (PEFEIT) flow, slope of the global expiratory flow‐volume curve (FVslope), steepest FVslope over all pixels in the lung field, total impedance change (surrogate for tidal volume; VTEIT) and intercept on the expiratory FV curve normalized to VTEIT (FVintercept/VTEIT) were indexed to baseline and analyzed for a difference from the control, at Cmax, 2 and 10 minutes after albuterol. Multiple linear regression explored the explanation of the variance of Δflow, a validated variable to evaluate bronchoconstriction using all EIT variables. Results At Cmax, PIFEIT, PEFEIT, and FVslope significantly increased whereas FVintercept/VT decreased. All variables returned to baseline 10 minutes after albuterol. The VTEIT did not change. Multivariable investigation suggested 51% of Δflow variance was explained by a combination of PIFEIT and PEFEIT. Conclusions and Clinical Importance Changes in airflow during histamine challenge and subsequent albuterol administration could be detected by various EIT flow volume variables.


| INTRODUCTION
Mild to moderate asthma in horses commonly is diagnosed based on clinical signs, increased tracheal mucus viewed on endoscopy, and bronchoalveolar lavage cytology. 1,2 It is unclear which diagnostic tool is most appropriate because a clinically applicable objective lung function test is not readily available. The lack of alignment in diagnostic approach between practitioners and international consensus presents a barrier to advancing evidence-based practice 2,3 and there is a need for a sensitive diagnostic tool, ideally accessible and noninvasive, that can provide an objective measurement of lung function in the field.
Such devices would best detect subtle to marked alteration in lung mechanics, which occurs across the range of manifestations seen in the stages of asthma in horses. 1 Electrical impedance tomography (EIT) is a noninvasive, radiationfree functional imaging modality, which allows assessment of ventilation of conscious and anesthetized horses. [4][5][6] The fundamental principle of EIT is Ohm's law which describes the resistance of tissue to electrical current from measurements of voltage. Alternating currents are used instead of direct current making impedance the main outcome measure for EIT as the analogue to resistance. Lung tissue impedance changes during breathing as a function of air content. Electrodes equidistantly mounted on a belt are placed around the thorax. Weak alternating currents are applied and voltages measured between pairs of electrodes. From these measurements, changes in impedance within the thorax can be calculated and cross-sectional images of the distribution of impedance changes within the thorax are displayed 50 times per second. Based on the reported linear relationship between EIT impedance change and changes in lung air volume, it can be assumed that impedance change correlates with gas volume, and total impedance change over a single breath can be used as a surrogate for tidal volume. [7][8][9] Therefore, peak inspiratory (PIF EIT ) and peak expiratory (PEF EIT ) global flow can be evaluated by calculating the first derivative of the global EIT volume signal. 10 Electrical impedance tomography derived flow signals have been shown to detect changes in airflow during histamine bronchoprovocation in horses. 11 In that pilot study, flowmetric plethysmography (FP) described by the variable delta flow (Δflow) was compared against regional and global peak inspiratory and expiratory flow variables calculated from the EIT measurements. Although EIT flow variables and Δflow represent different measurement principles of flow, EIT flow variables were shown to follow the same pattern of change as Δflow over the duration of the histamine challenge. The authors concluded that EIT has potential as a noninvasive, objective diagnostic tool to monitor airflow changes occurring in horses with asthma.
Our aim was to confirm the observation noted for EIT flow variables in the pilot study by examining histamine-provoked changes in airflow in a larger population of horses and to determine if novel EIT variables could provide additional information on changes in breathing pattern during bronchoconstriction, as well as evaluate subsequent airflow changes after drug-induced bronchodilatation. A secondary aim was to compare EIT flow variables with the corresponding flow variables measured using a pneumotachograph and the validated FP variable Δflow.
It was hypothesized that all EIT-derived variables would change during histamine challenge and return to preprovocation levels after drug-induced bronchodilatation. Secondarily, it was hypothesized that EIT flow variables would follow the same pattern as standard flow variables and Δflow, and finally that the variance in Δflow could be best explained by the EIT flow variables.

| Animals
Twenty horses (12 Standardbreds and 8 Thoroughbreds) from the institutional teaching herd were included. The horses included 3 mares and 17 geldings, with a mean (SD) age of 11.7 (7.9) years and body weight of 511 (64) kg. The median (range) body condition score was 6 of 9 (range, 5-7/9). 12 Horses were considered clinically healthy based on clinical examination and long-term history. Before the study, the horses were housed in an irrigated paddock and received supplemental feed of meadow hay.

| Preparation
Each horse was walked into stocks and had an 18G catheter (Surflo, Terumo, Philippines) aseptically placed in the right jugular vein. The horse then was sedated with a loading dose of 5 μg/kg of detomidine (Detomo Vet, Ceva, Glenorie, NSW, Australia) IV before immediately receiving a detomidine constant rate infusion (12.5 μg/kg/hour) administered through a fluid pump. The hair over the thorax directly caudal to the scapula (5th-6th intercostal space) was moistened with water and electrically nonconductive ultrasound gel was applied circumferentially in this region. Thereafter, a stretchable neoprene EIT belt with 32 equidistantly-mounted stainless steel electrodes was placed under slight tension around the thorax on top of the gel.
Contact with the skin was maintained by application of a cohesive stretchable bandage placed over the EIT belt. The belt was connected to the EIT device (BBvet, SenTec AG, Landquart, Switzerland). Sufficient contact of the electrodes with the skin was verified by the EIT software and visual inspection of adequate electrode-skin resistance values during the study period. The EIT measurements were performed with 47 impedance tomograms per second.
A modified Graz consensus reconstruction algorithm for EIT (GREIT) 13 adapted for horse anatomy was used to generate EIT images for each horse, representing breathing-related regional changes of impedance.
Additional details on the use of EIT technology and image reconstruction are reported in the literature. [13][14][15][16] Each horse was concurrently fitted with FP hardware to measure the effect of histamine bronchoprovocation on Δflow (Open Plethysmography, Ambulatory Monitoring, Ardsley, New York). A facemask, pneumotachograph (No. 5 Fleisch), and abdominal and thoracic respiratory inductance plethysmography bands (RIP), which make up the FP unit, were fitted to the horse and calibrated according to the manufacturer's instructions. 17 Head carriage was maintained on a stand in a consistent horizontal position to avoid nasal edema. 18 17 This time point was referred to as C max .

| Histamine bronchoprovocation
After reaching 1 of the 2 described end points and data collection for C max , albuterol sulfate (Ventolin, GlaxoSmithKline, Victoria, Australia) was administered into the face mask using a metered dose inhaler (100 μg/actuation to ideally dose at 2 μg/kg). One puff was administered via the nebulizer port of the face mask at the beginning of each inspiration until the predefined dose was reached. Data collection for 3 minutes started 2 minutes (A 2min ) and 10 minutes (A 10min ) after the last puff was administered. Once data collection for A 2min and A 10min was complete, all instruments and devices were removed, the detomidine CRI stopped, and the horse was observed for an additional 10 minutes in the stocks before being returned to the paddock.

| Electrical impedance tomography variables
A sequence of at least 8 and a maximum of 22 breaths free of any movement artifacts was preselected for each time point depending on the respiratory rate and cooperation of the horse during the 3-minute measurement period using a dedicated software (Ibex, SenTec AG, Landquart, Switzerland). 19 Breaths with obvious movement artifacts were removed manually. The EIT signal of each preselected sequence was exported and analyzed using custom software, 20 which is available as part of EIDORS. 21 The following variables were derived from the EIT signal:

| Gas flow variables and EIT-derived tidal volume
The global impedance changes, representing the EIT volume signal in arbitrary units (AU), was calculated as the sum of all pixels within the entire EIT image. The tidal volume-related EIT factor VT EIT was calculated by subtracting the impedance distribution at start of inspiration from the impedance distribution at end of inspiration ( Figure 1A). [7][8][9] From the EIT volume signal, the global EIT gas flow was calculated as the first derivative of the global volume over each breath ( Figure 1A). 10 From the flow signal, global peak inspiratory (PIF EIT ) and global peak expiratory (PEF EIT ) flow were calculated for the whole lung for each breath. 11

| Spirometry and flowmetric plethysmography variables
Peak inspiratory (PIF Spiro ), peak expiratory (PEF Spiro ) flows and tidal volume (VT Spiro ) were measured using a calibrated Fleisch pneumotachograph. This flow factor measures the maximum difference between the flow derived from the pneumotachograph and the calibrated flow from the respiratory inductance bands. 17 Delta flow has been validated as a variable that can detect narrowing in airway diameter after histamine bronchoprovocation. 18,22 The same software was used to record the spirometry data measured by the pneumotachograph.  in some horses that sweated excessively (an example of mild interference is seen in Figure 1A,B in breath 3 of the saline measurement). In 1 horse, interference made EIT data analysis unusable. Both EIT and FP data sets were available in 17 horses.

| Data analysis
Horses showed significantly higher indexed EIT flow variables (PIF EIT : P < .001, PEF EIT : P < .001), indexed spirometry variables (PIF Spiro : P = .01, PEF Spiro : P = .02) and FP flow index, Δflow (P = .002) as well as respiratory rate (P < .001) at C max compared to the diluent control nebulization (saline;  The FV slope at C max was significantly higher (P = .03; Table 1) than at control nebulization but was not significantly different from control at A 2min and A 10min ( Figure 5). Although FV slope max followed the same F I G U R E 2 Graphical illustration of flow volume loops generated from 6 breaths during the measurements at baseline (A) and C max (B) in the same horse as in Figure 1 and determination of the intercept in the FV loop (FV intercept /VT EIT ). The volume change during inspiration and expiration are indicated by arrows. The expiratory portion of the curve is depicted by the portion of the loop with positive flow. The volume increase during inspiration is plotted toward the left on the x-axis. The computing of the FV intercept slopes and lines are depicted in green and brown. The FV intercept is indicated by a red line and is the percentage of the expiratory portion of the flow volume loop where flow is constant pattern, the change between the control nebulization and C max was not significant. The FV intercept /VT EIT was significantly lower (P < .001) at C max , but not significantly different from the control nebulization at A 2min and A 10min ( Figure 5). The variables VT EIT and VT Spiro were not significantly different from the control nebulization values at any measurement point (Table 1).
Further analysis of all 5 EIT variables showed the variance in Δflow was best described by the variance in both flow variables, PIF EIT and PEF EIT , with an adjusted R 2 of 0.51. When FV slope was included, the adjusted R 2 increased to 0.53.

| DISCUSSION
The aim of our study was to investigate whether histamine-provoked changes in airflow and subsequent drug-induced bronchodilatation can be detected using noninvasive EIT measurements, motivated by promising findings in a pilot study. 11 Results showed global EIT peak flows can detect airflow changes during histamine challenge, and that EIT could detect a reversal of airflow changes induced by the administration of albuterol after histamine challenge. Additionally, novel variables (FV slope and FV intercept /VT EIT ) derived from the flow-volume curve of the EIT signal were identified as also detecting airflow changes. Similar to the peak flows measured by spirometry and the primary FP variable, Δflow, these EIT-derived flow indices significantly changed after histamine administration and returned to control values with subsequent albuterol administration.
Inhaled histamine has been shown to induce bronchoconstriction through activation of H1 receptors, 23 and peak flows increase within the airways. 11,24 This is true in both histamine-induced bronchoconstriction and the natural disease state of severe asthma in horses. [25][26][27][28] Both devices, the EIT and the pneumotachograph, registered significantly higher measurements of inspiratory and expiratory peak flows during C max (Figure 4). This result shows that global EIT flow variables, based on the assumption that the measured impedance change corresponds with inhaled and exhaled volume, follow the  Our study showed that EIT also could detect airflow changes associated with bronchodilatation induced by albuterol. Albuterol has been shown to be effective in decreasing histamine induced contraction of equine bronchial muscle in vitro, 29,30 as well as being effective in inducing bronchodilatation in the natural disease state of severe asthma in horses. [31][32][33][34] This finding supports the use of EIT and its variables as a possible monitoring tool to assess treatment success in horses with asthma. Future studies on naturally occurring asthma in horses are necessary to explore this application.
To our knowledge, this study describes for the first time the computing of a FV loop from the EIT signal in veterinary medicine.
The shape of the FV loop our study suggests a biphasic inspiratory and expiratory breathing pattern in most horses before the administration of histamine. This biphasic pattern also has been described in horses at rest using spirometry, 25,26,35 and in combination with electromyography. 36 During histamine challenge at C max , the biphasic pattern of the EIT-derived FV loop was lost in all but 2 horses during inspiration, whereas a distinct biphasic pattern still was observed during expiration in all but 1 horse. Loss of the biphasic pattern of the FV loop in both inspiration and expiration has been described in horses with severe asthma. 25,26 The shape of FV loops with a high peak expiratory flow early in expiration followed by a low flow rate has been described in horses during severe airway obstruction. 26 To quantify the expiratory flow pattern observed in the EIT-derived FV loop, a global variable that described the shape of the global FV loop (FV slope ) and 1 regional variable that identified the pixel with the steepest slope within the ROI and quantifies the steepness (FV slope max) were designed ( Figure 1C).
At C max , the FV slope was significantly higher than the control and subsequently returned to the control values after albuterol administration, suggesting that this variable can be used to describe changes in airflow in horses. The FV slope max followed the same pattern as FV slope , but the change was not significant. One reason for lack of significance was that both FV slope and FV slope max displayed marked variability among horses. In some horses, the RIP bands of the FP unit, which applies a current traveling through the thoracic and abdominal bands, resulted in interference with the measured EIT signal, which influenced the plotting of the FV loop and made the computing of these 2 variables challenging for the software. The clinical usefulness of both EIT slope variables should be further explored in the absence of FP hardware.
The RIP bands used in FP have been used to assess changes in the breathing pattern of horses during airflow changes. 25 Our intention was to create an EIT variable that identified these changes in The primary FP variable, Δflow, which is a combined variable of airflow and thoracic volume changes, 18 also was higher at C max , which confirmed the findings of the pilot study. 11 One aim of our study was to evaluate which EIT variables best explain Δflow. 18,22 Multivariable analysis suggested that 51% of the variance in Δflow was explained by a combination of PIF EIT and PEF EIT . This finding was anticipated because Δflow is calculated in part using airway flow measured by the pneumotachograph of the FP unit. 22 The FV slope also made a small contribution to the variance in Δflow.
As expected tidal volume did not change significantly with a change in flow variables with either measuring device during the study period. Previous studies have identified unchanging tidal volumes during bronchoconstriction in horses. 11,25,26 The respiratory rate increased with histamine and then decreased with albuterol, whereas tidal volume did not change significantly, indicating an increase and then subsequent decrease in minute ventilation. The observed increase in respiratory rate results in a decrease in inspiratory 24,25 and expiratory time, 24 which leads to the measured increase in flow variables at stable tidal volume. It has been postulated that the increase in minute ventilation could occur in response to mild hypoxemia as a consequence of airway narrowing. 28 It remains challenging to assess pulmonary mechanics using existing pulmonary function testing modalities immediately after exercise in the horse. The minimal portable EIT instrumentation which can be powered by a battery, is noninvasive and does not require a face mask is well suited to clinical situations. It is possible that EIT instrumentation can be undertaken in the field, allowing airway status to be investigated before and immediately after exercise in horses suffering from naturally-occurring asthma. 37  Another limitation is that albuterol was administered as puffs into the mask, which makes precise dosing of the drug impossible. The aim of albuterol administration was to show a gradual improvement of EIT values back to baseline, which made early measurement after drug administration necessary. To meet this aim, this method of application for the albuterol was chosen because it was essential to leave the mask in place to be able to start data recording immediately. Finally, healthy horses were used in our study, further studies to validate EIT using a naturally-occurring airway obstruction model against other methods of pulmonary function testing will be required to support the use of this modality as a diagnostic tool for horses with asthma.

| CONCLUSION
The EIT variables, PIF EIT and PEF EIT , and the variables derived from the FV loops, FV slope and FV intercept /VT EIT , detected changes in airflow after histamine bronchoprovocation in horses and subsequent albuterol administration. Global EIT flow variables showed the same pattern of change as did flow variables measured by spirometry and best explained Δflow. This modality could be useful to detect changes in airflow in horses with asthma and further work is required to determine if a combination of the variables measured by the EIT may be useful to detect subtle changes in pulmonary function in horses with asthma.

ACKNOWLEDGMENT
No funding was received for this study.

CONFLICT OF INTEREST DECLARATION
Andy Adler and Martina Mosing are developing a commercial product that uses technology developed in this paper. All other authors have declared that their contribution to the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

OFF-LABEL ANTIMICROBIAL DECLARATION
Authors declare no off-label use of antimicrobials.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION
This experimental in vivo study was performed on experimental horses and was approved by Murdoch University Institutional Animal Ethics committee (permit number R2895/17).

HUMAN ETHICS APPROVAL DECLARATION
Authors declare human ethics approval was not needed for this study.