Effect of position on transdiaphragmatic pressure and hemodynamic variables in anesthetized horses

Abstract

Recumbency affects respiratory mechanics and oxygenation in anesthetized horses. Changes in pleural and abdominal pressures that can impair ventilation have not been described in all recumbencies. The objective of this study was to determine the effects of patient positioning on transdiaphragmatic pressure and selected hemodynamic variables. Horses were maintained under total intravenous general anesthesia with nasal oxygen supplementation. Transnasal balloon catheters in the stomach and thoracic esophagus were used to measure intrathoracic and gastric pressures in standing horses and in anesthetized horses positioned in right and left lateral recumbency, dorsal recumbency, reverse Trendelenburg position, and Trendelenburg position. Transdiaphragmatic pressure was calculated as the difference between gastric and intrathoracic pressures. Measurements of oxygen saturation (SpO₂), heart rate, systolic, diastolic and mean arterial pressures, and respiratory rate were obtained every 5 minutes. When compared to dorsal recumbency, gastric expiratory pressure is decreased in the standing position. Thoracic expiratory pressure is decreased in standing and reverse Trendelenburg. Transdiaphragmatic expiratory pressure and SpO₂ are decreased in Trendelenburg. Heart rate is increased in reverse Trendelenburg. Systolic, diastolic, and mean arterial pressures are decreased in reverse Trendelenburg and increased in left lateral and right lateral recumbency. We found that there is wide variation in respiratory pressures between horses and positions and they are not predictive of associated changes in hemodynamic variables.

Introduction

Large differences in arterial and alveolar oxygen tensions have been demonstrated in recumbent horses due to ventilation perfusion mismatch (1–4). A major source of this ventilation perfusion mismatch is the development of atelectasis due to decreased functional residual capacity (1,2,5–7). During recumbency, the cephalad shift of the diaphragm caused by pressure from the abdominal contents may lead to the development of atelectasis in the dependent lung (7,8). The transmission of intra-abdominal pressure to the thoracic cavity has been demonstrated to impair respiratory mechanics in multiple species (9–11). Intra-abdominal pressure has also been shown to change with patient positioning, body mass, disease processes, and surgical technique (12–16).

Placing patients in a reverse Trendelenburg position has been used in humans and horses successfully in an attempt to improve ventilation, by decreasing the pressure from the abdominal cavity on the pleural cavity and lungs (17–19). Alternatively, the Trendelenburg position is used in some equine surgical procedures such as urogenital laparoscopy. Placing horses in this position with abdominal insufflation results in decreased PaO₂ and pH as well as increased PaCO₂ and mean arterial pressure by increasing the pressure transmission from the abdomen (20).

Transdiaphragmatic pressure, or the difference between abdominal pressure and pleural pressure, is commonly used to evaluate diaphragmatic contractility and work of breathing and has been evaluated in laterally recumbent and exercising horses (21,22). However, these values have not been evaluated together in horses in different recumbencies. Therefore, the purpose of this study was to establish baseline pleural and abdominal pressures, and subsequently, transdiaphragmatic pressure in various recumbent positions, to better understand the effects of patient positioning.

Materials and methods

Animals

Ten horses free of cardiopulmonary disease were used for a randomized crossover study. The experimental protocol was approved by the Kansas State University Institutional Animal Care and Use Committee.

Instrumentation

A 14-gauge intravenous (IV) catheter was placed in the left jugular vein. Horses were sedated with 0.4 mg/kg body weight (BW) IV xylazine (Anased LA; MWI Animal Health, Boise, Idaho, USA). Two custom made balloon catheters were fastened together with the balloons 40 cm apart as previously described for standardization (21). Balloons were filled with 5 mL of air. The catheters were placed transnasally such that one balloon was in the gastric lumen and the other was in the thoracic esophagus caudal to the heart (Figure 1). Placement of the catheters was confirmed by observation of characteristic pressure changes with the respiratory cycle using aneroid manometers. After catheters were placed, they were connected to pressure transducers. Signal processing was used to determine differential pressures (Omega PX26-005 DV; Omega Engineering, Stamford, Connecticut, USA) coupled with stages of amplification. The amplified signal was digitized and logged to a computer using the NI myDAQ platform and LabVIEW software (National Instruments, Austin, Texas, USA).

Figure 1.

Diagram depicting the position of the balloon catheters in reference to the diaphragm with the catheters connected at a standardized 40 cm distance and connected to a differential pressure transducer coupled to two stages of amplification. The amplified signal was digitized and logged to a computer using the NI myDAQ platform and LabVIEW software.

Anesthesia

General anesthesia was induced with 1.1 mg/kg BW, IV xylazine (Anased LA; MWI Animal Health), 2.2 mg/kg BW, IV ketamine (Ketaset; Zoetis, Parsippany, New Jersey, USA), and 0.05 mg/kg BW, IV midazolam (Akorn, Lake Forest, Illinois, USA). Horses were maintained on a constant rate infusion of GKX [50 g/L guaifenesin (Medisca, Irving, Texas, USA), 4.4 mg/kg/L ketamine, 1.1 mg/kg/L xylazine] at 1 L/hr. Horses were supplemented with nasal insufflation of oxygen at 15 L/min. Instrumentation was performed with the horses positioned on a surgery table in dorsal recumbency. Horses were instrumented with a facial arterial catheter for continuous arterial pressure monitoring with the pressure transducer placed at the level of the right atrium, base apex electrocardiogram leads, and lingual pulse oximetry. Once horses were instrumented, body position was changed based on random assignment.

Data collection

Measurements of intrathoracic and gastric pressures were collected in standing horses and in anesthetized horses positioned in left lateral recumbency, right lateral recumbency, dorsal recumbency, dorsal recumbency with the table tilted head down 15°, and dorsal recumbency with the table tilted head up 15°. The order was determined by random assignment. Two minute intervals followed repositioning before measurements were made in the next position.

Three consecutive breaths were chosen from a time pressure waveform that represented the normal breathing pattern in each position, excluding breath holds or sighs (Figure 2). Peak inspiratory and expiratory pressures were measured from the gastric and intrathoracic waveforms for the 3 breaths. Values were averaged to determine mean intrathoracic inspiratory, intrathoracic expiratory, gastric inspiratory, and gastric expiratory pressure. Transdiaphragmatic pressure was calculated for inspiration and expiration as the difference between mean gastric pressure and mean intrathoracic pressure. Measurements of oxygen saturation (SpO₂), heart rate, systolic, diastolic and mean pressure, and respiratory rate were collected every 5 min after induction of general anesthesia.

Figure 2.

Pressure in centimeters of water was graphed over time for the gastric (dashed line) and thoracic (solid line) balloons. Peak inspiratory and expiratory pressures for each balloon were identified for three consecutive breaths and averaged.

Data analysis

Respiratory pressures and hemodynamic variables were analyzed by a mixed linear regression model. Effect of position on respiratory pressures and hemodynamic variables and effect of respiratory pressures on hemodynamic variables were determined with a Wald test. Level of significance was set at P < 0.05.

Results

Animals

Ten horses were included in the analysis including 3 mares, 6 geldings, and 1 stallion. There were 8 Quarter horses, 1 Tennessee walker, and 1 Thoroughbred. Age ranged from 4 to 25 y (mean = 13.4 y). Weight ranged from 425 to 580 kg (mean = 496.4 kg). The amount of total GKX received by each horse was dependent on weight and time under anesthesia and ranged from 596 to 1630 mL (mean = 832 mL). This resulted in an adequate plane of anesthesia for all horses for the duration of the experiment.

Effect of position on respiratory pressures and hemodynamic variables

Hemodynamic variables and respiratory pressures are reported by position as means and 95% confidence intervals (Figure 3 and Table I). When position was found to be a significant factor, each position was compared to dorsal recumbency as the reference position. In reverse Trendelenburg, thoracic expiratory pressure and systolic, diastolic, and mean arterial pressures were decreased and heart rate was increased. In Trendelenburg position, transdiaphragmatic expiratory pressure and SpO₂ were decreased. Systolic, diastolic, and mean arterial pressures were increased in left and right lateral recumbencies. In the standing position, gastric and thoracic expiratory pressures were decreased. Position did not significantly impact gastric, thoracic, or transdiaphragmatic inspiratory pressures or respiratory rate.

Figure 3.

Mean +/− 95% confidence interval values for gastric, thoracic, and transdiaphragmatic (PDI) pressures in cmH₂O for both inspiration and expiration are reported by position. Values significantly different from dorsal recumbency are designated with an asterisk.

Table I.

Mean ± 95% confidence interval values for oxygen saturation (SpO₂), heart rate, arterial blood pressure (BP), and respiratory rate, reported by position.

Position	SpO2 (%)	Heart rate (bpm)	Systolic BP (mmHg)	Diastolic BP (mmHg)	Mean BP (mmHg)	Respiratory rate (bpm)
Dorsal	93.3 ± 3.33	34.3 ± 3.01	119.1 ± 14.11	73.1 ± 11.85	87.8 ± 13.68	13.4 ± 3.79
Dorsal RT	88.71 ± 4.12	40.75 ± 5.58*	64.25 ± 19.5*	35 ± 15.29*	45.63 ± 16.23*	12.38 ± 4.64
Dorsal T	87.63 ± 4.17*	33.36 ± 2.13	126.54 ± 18.47	81.81 ± 15.9	96.45 ± 17.61	13.54 ± 3.43
Left lateral	92.5 ± 2.88	34.6 ± 3.11	136.8 ± 15.49*	89.8 ± 13.68*	106 ± 15.56*	13.4 ± 3.51
Right lateral	91.2 ± 1.29	33.7 ± 0.85	137.9 ± 3.17*	90.1 ± 3.36*	105.8 ± 3.37*	13.6 ± 1.27

^*Values significantly different from dorsal recumbency.

bpm — beats per min; RT — reverse Trendelenburg; T — Trendelenburg.

Effect of respiratory pressures on hemodynamic variables

SpO₂ was positively correlated with gastric and transdiaphragmatic inspiratory pressures. Systolic arterial pressure was positively correlated with gastric and thoracic expiratory pressures. Respiratory rate was negatively correlated with expiratory transdiaphragmatic pressure.

Discussion

Changes in gastric and thoracic respiratory pressures with change in position were variable between horses but followed an expected pattern. For the majority of horses and positions, gastric and thoracic inspiratory pressures tended to be negative or subatmospheric, and expiratory pressures tended to be positive. Transdiaphragmatic pressure, however, did not display a trend among horses or positions and seemed to be more variable in both the magnitude and direction of deflection.

Pleural, abdominal, and transdiaphragmatic pressures can be used to measure compliance, energy expenditure of breathing, and respiratory muscle function (22,23). Distending pressure of the lung is the difference between atmospheric pressure and pleural pressure and determines the lung volume and ventilation (23). The distending pressure is impacted by pressure from the abdominal viscera on the diaphragm (23). In this study, our interest was in establishing normal measurements of the relative pressure being placed across the diaphragm in horses that could be used to illuminate the effect of patient positioning on the distending pressure of the lung and not necessarily the work of breathing. The increase in gastric and thoracic expiratory pressures seen in recumbency compared to standing may be responsible for the decrease in functional residual capacity and airway closure that has been reported in recumbent horses (2,6,7). The decrease in thoracic expiratory pressure in reverse Trendelenburg position would theoretically reduce airway closure. It is possible this is responsible for the improvement in oxygenation seen in horses and humans in this position in previous studies, however no benefit was seen in SpO₂ in the current study (18,19,24). This could be due to a lack of mechanical ventilation, a lower inspired oxygen concentration, and a relatively short amount of time interval spent in this position during this study. An absence of benefit seen in our study with this position could also be a consequence of random position assignment, as Binetti (18) showed that reverse Trendelenburg was able to prevent but not reverse gas exchange impairment from atelectasis formation.

The decision to perform this experiment under total intravenous anesthesia (TIVA) without mechanical ventilation was to provide a more accurate assessment of baseline pressures without additional confounding factors. Unlike Kowalczyk et al (21), we found TIVA to produce more predictable and regular respiratory rates and breathing patterns than did inhaled anesthesia with isoflurane during our pilot study.

The use of ketamine has been shown to preserve intercostal function but depress diaphragmatic contractility (25,26). This results in changes in transdiaphragmatic pressure when breathing against a closed airway; however, because we were more interested in evaluating changes in pressure across the diaphragm and not diaphragmatic strength, it likely did not have a profound effect on our findings like it did in other studies (21). The decrease in diaphragmatic contractility could have, however, decreased the ability of the diaphragm to resist pressure from abdominal viscera, particularly in the Trendelenburg position, resulting in a greater effect than would be seen with inhaled anesthetics.

The total amount of injectable anesthetics delivered was dependent upon the weight of the horse as well as the duration of the experiment. Duration of each experiment was affected by the sequence of changes in position with some repositioning taking longer than other sequences. As the administration of GKX was delivered as a constant rate infusion based on body weight, this should not have impacted our findings. All horses were found to be at an adequate anesthetic depth throughout the experiment and the stimulation of changing positions did not result in observed changes in anesthetic depth or breathing character and rate. Horses placed in Trendelenburg position showed an increase in respiratory effort compared to the other positions after approximately 2 min, indicating that the changes seen with position were not due to physical manipulation from the change in position but likely the pressure exerted on the thoracic cavity from the abdominal viscera.

Changes in respiratory pressures were not predictive of the associated hemodynamic changes seen with change in position. This indicates that factors other than respiratory pressure are responsible for hemodynamic changes in different positions. Arterial blood pressure was decreased in the reverse Trendelenburg position, similar to the effects of this position seen on cardiovascular parameters in humans, cattle, and swine (27–29). In humans, hypotension can occur due to a decrease in preload from blood pooling in the lower extremities (27). Schauvliege et al (19) showed no difference in arterial blood pressure between dorsal and 7° reverse Trendelenburg; in fact, arterial pressure increased with time in their study. This was likely due to the use of dobutamine, which was also increased over time. To a lesser degree, table tilt may have also limited blood pooling and adverse cardiovascular side effects in our study and in humans. In this study, we elected a table tilt of 15° because the 30° tilt that is often used clinically in humans was unnecessary to surgically access the caudal abdomen; moreover, decreased tilt in horses and cattle prevents adverse cardiovascular effects (17,18,24,29). The improvement in arterial pressure seen in both lateral recumbencies compared to dorsal recumbency is also consistent with previous findings and is likely due to the decreased pressure on the caudal vena cava from the abdominal viscera, resulting in an improved preload (30). The decrease in SpO₂ in Trendelenburg is consistent with previous studies and is likely due to the weight of the abdominal viscera on the diaphragm, resulting in atelectasis of the caudodorsal lung fields and therefore, an increased ventilation/perfusion mismatch (20).

A limitation of this study is the ability of the gastric and esophageal balloons to accurately measure global pleural, abdominal, and transdiaphragmatic pressures in all positions. A previous evaluation of the use of esophageal balloons to estimate pleural pressure shows that despite the pleural pressure gradient increasing from dorsal to ventral, balloons in the middle and caudal thoracic esophagus were not significantly different from any of the direct measurements (31). This was evaluated in standing ponies; therefore, we do not know if the same accuracy is maintained in different recumbencies in horses (31). Measurement of abdominal pressure differs with measurement location and body position (15,22). The use of gastric balloons for the measurement of intra-abdominal pressure has been shown to be poorly correlated with direct measurement in horses; however, direct measurements also differ greatly depending on the location of measurement (15,32,33). Therefore, the relative cranial position of the gastric balloon in the abdomen would subject it to forces by the caudal viscera and may not provide a completely accurate representation of the global intra-abdominal pressure as positions changed. The measure of gastric pressure in the cranial abdomen, however, is likely more applicable to our goal of identifying pressure exerted on the pleural cavity by the abdominal cavity.

In summary, inspiratory pressures in horses tend to be subatmospheric and expiratory pressures are more positive. The change in these pressures with changes in position are as one would expect with relative pressure exerted from the abdominal viscera. Transdiaphragmatic pressure does not appear to change as intuitively with position, nor does it predict physiologic changes seen in different positions.