Adequate oxygenation of arterial blood depends on the matching of ventilation to perfusion at the alveolar level. Development of VA/Q mismatch, where alveoli may be well ventilated but poorly perfused or vice versa, decreases the likelihood that blood passing through the lungs will become fully oxygenated. Anesthetized recumbent horses develop atelectasis in the dependent regions of the lungs,1,2 resulting in Qs/Qt in which the atelectatic regions of the lungs receive perfusion but not ventilation, which in turn allows blood to pass through the lungs without being oxygenated. Both VA/Q mismatch and Qs/Qt allow the addition of poorly oxygenated blood, called venous admixture, to enter the arterial circulation, resulting in hypoxemia (Pao2 < 80 mm Hg), which is a common complication in anesthetized horses.1–7 Standard treatment for hypoxemia in anesthetized horses is primarily aimed at improving alveolar ventilation, but poorly ventilated and nonventilated alveoli are difficult to reinflate, and hypoxemia secondary to VA/Q mismatch and Qs/Qt is not always relieved by these treatments.2,3,8–10 In humans, hypoxemia secondary to a variety of intrapulmonary conditions is often treated by changing perfusion rather than alveolar ventilation, and this can be achieved by delivery of NO, a potent vasodilator, in the inhaled gases.11 Inhaled NO preferentially reaches only areas of adequate alveolar ventilation, and the subsequent vasodilation redistributes perfusion to these ventilated alveoli, thereby improving VA/Q matching and Pao2.
Nitric oxide pulsed into the early phase of inhalation has been used to improve Pao2 and VA/Q matching while decreasing Qs/Qt in recumbent isoflurane-anesthetized horses,12–15 but the intrapulmonary location of redistribution of perfusion induced by PiNO administration is not known. Redistribution could occur at individual alveoli scattered throughout the lungs or by mass movement of blood flow away from discrete dependent atelectatic areas. Although Qs/Qt in the dependent atelectatic regions of the lungs is considered to be the primary contributor to hypoxemia in anesthetized horses,1,2 redistribution induced by PiNO administration would require perfusion to move away from atelectatic areas in the dependent regions of the lungs to aerated regions in the nondependent regions of the lungs, meaning that a large volume of blood would need to be moved en masse against gravity. Conversely, it could be that diffuse areas of VA/Q mismatch are important contributors to hypoxemia and that redistribution induced by PiNO administration occurs at individual alveoli scattered throughout the lungs. However, for redistribution to occur in individual alveoli, alveoli with low ventilation would have to receive a concentration of iNO adequate to change perfusion, but low-ventilation alveoli open intermittently and do not consistently participate in gas exchange, which makes delivery of effective concentrations of NO unlikely. It is also possible that regardless of the location of redistribution, iNO-mediated vasodilation may not be the only factor contributing to the change in perfusion, and there is evidence that iNO may cause a concomitant increase in concentration of the potent vasoconstrictor endothelin-1 in unventilated alveoli distant to the ventilated alveoli receiving iNO.16,17 Vasoconstriction in alveoli with low ventilation would limit perfusion to those hypoxemic regions of the lungs, aiding the iNO in redistributing perfusion to areas of high ventilation.
The matching of VA/Q and magnitude of Qs/Qt throughout the lungs can be determined by a MIGET,18–20 which uses the solubility coefficients of a variety of inert gases dissolved in blood. However, the MIGET does not provide spatial perfusion data.18 The actual spatial distribution of blood flow in the lungs can be visualized by use of 2-D perfusion scintigraphy with IV injected human macroaggregated albumin labeled with the standard radioactive tracer technetium Tc 99m.
The aims of the study reported here were to image the spatial distribution of pulmonary blood flow by means of scintigraphy, evaluate VA/Q matching and Qs/Qt by means of the MIGET, and measure arterial oxygenation and plasma endothelin-1 concentrations before, during, and after PiNO administration in dorsally recumbent isoflurane-anesthetized horses. Our hypothesis was that PiNO administration would result in redistribution of blood flow from the atelectatic, dependent regions of the lungs to the aerated, nondependent regions of the lungs, thereby decreasing Qs/Qt and improving Pao2 and VA/Q matching.
Materials and Methods
Horses—Three university-owned healthy Standardbreds (1 mare and 2 geldings) with a mean body weight of 513 kg (range, 459 to 562 kg) and mean age of 9 years (range, 8 to 9 years) were used for the study. The study protocol was approved by the local ethics committee for animal experiments.
Anesthesia—Food but not water was withheld for 12 hours prior to anesthesia, and horses were deemed healthy on the basis of physical examination findings. Twenty minutes after IM administration of acepromazine (0.03 mg/kg), 7.5% guaifenesin was infused IV until each horse developed ataxia (approx 100 mg of guaifenesin/kg), at which point anesthesia was induced with an IV bolus of thiopentone sodium (4 to 5 mg/kg). Horses were intubated with a 24-mm cuffed endotracheal tube and placed in dorsal recumbency on a padded table. Anesthesia was maintained with spontaneous breathing of oxygen (> 90%) and isoflurane delivered in a standard circle breathing system by an agent-specific precision vaporizer mounted on a large animal anesthesia machine. Throughout the study, end-tidal isoflurane concentrations of 1.5% to 1.7% (approx 1.2 to 1.25 times the minimum alveolar concentration of isoflurane in horses) were maintained. The gas monitor was calibrated before each experimental period by use of commercially prepared calibration gas.
Study protocol and PiNO administration—After induction of anesthesia, horses were instrumented for data collection, and inert gases for assessment by use of the MIGET were infused for 60 minutes. Baseline data were collected at the end of the inert gas infusion; then PiNO administration with a proprietary devicea was begun. The device was activated by negative pressure and delivered a volumetric dose into the endotracheal tube at the onset of inspiration. The delivery device was connected to a cylinder supply of NO in nitrogen (2,000 μg/mL), and NO was delivered in the first 45% of the inspiratory phase. This dose was determined in previous dose titration studies12–15 performed by our laboratory group.
Data were again collected after 30 minutes of PiNO administration. Administration of PiNO was then discontinued, and data were collected 30 minutes later.
Instrumentation—Once anesthesia was induced, ECG electrodes were placed on the thorax in a lead II configuration, and a catheter was inserted percutaneously into a facial artery. An introducer kit was used to place a 7F thermodilution catheter via a jugular vein into the pulmonary artery for measurement of pulmonary arterial blood pressure and for collection of mixed-venous blood samples for blood gas analysis. A pigtailed, multiport catheter was introduced into a jugular vein, advanced into the right ventricle, and retracted into the right atrium for injection of saline (0.9% NaCl) solution for Qt determination. Catheters were positioned by use of pressure-tracing guidance and simultaneous ECG monitoring. Blood pressures (SAP, DAP, MAP, SPAP, DPAP, and MPAP) were measured by use of pressure transducers zeroed to atmospheric pressure and positioned at the level of the shoulder joint, which was considered to correspond to the level of the right atrium. Determination of Qt was made by use of a thermodilution technique in which a bolus of 20 mL of cold (0°C) saline solution was rapidly manually injected through the pigtailed catheter. A minimum of 3 injections were made, and means of data from 3 injections with values within 10% of each other were calculated at each time period. Values of Qt, systemic and pulmonary arterial blood pressures, heart rate, respiratory rate, tidal volume, end-tidal carbon dioxide fraction, and end-tidal isoflurane fraction were recorded from a standard anesthesia monitor. Arterial and central (mixed) venous blood samples were obtained for assessment of arterial blood pH, mixed-venous blood pH, Pao2, Po2, Paco2, and partial pressure of mixed-venous carbon dioxide by use of a standard electrode technique. Arterial oxygen saturation and So2 were measured with a standard electrode technique, and blood hemoglobin concentrations were measured spectrophotometrically. Samples for blood gas analysis were stored on ice and analyzed within 30 minutes after collection. Blood gas values were corrected for atmospheric pressure but not body temperature. Mixed-venous blood was collected in chilled tubes containing EDTA (final concentration, 10mM) and centrifuged at 4°C for 10 minutes to separate the plasma, which was subsequently analyzed for endothelin-1 concentration. Acid ethanol was added to precipitate the protein. The precipitate was analyzed for endothelin-1–like immune reactivity by use of a radioimmunoassay involving antiserum raised against endothelin-1 in rabbits.
MIGET—Distributions of alveolar ventilation and perfusion were estimated by use of the MIGET.18,19 Six inert gases (sulfur hexafluoride, ethane, cyclopropane, enflurane, ether, and acetone) were dissolved in isotonic sodium chloride solution and infused into a jugular vein at a dose of 30 mL/min for 60 minutes beginning immediately after induction of anesthesia so that the gases could reach equilibrium prior to data collection. At each collection time, arterial and mixed-venous blood samples were aspirated from the respective catheters, and mixed-expired gases were collected from a heated mixing box connected to the expiratory limb of the large animal circle breathing system. Gas concentrations in the blood samples and expirate were measured with a gas chromatograph.b On the basis of the blood data and mixed-expired gas data, VA/Q ratio was calculated in accordance with the original technique.18 Data measured or calculated from the MIGET included blood flow (Qt) and SD of its logarithmic distribution; tidal volume and SD of its logarithmic distribution; right-to-left vascular shunt (percentage of shunted perfusion; perfusion of lung regions with VA/Q < 0.005); dead space (ventilation of lung regions including apparatus dead space with VA/Q > 100); and the residual sums of square of VA/Q distributions, which represented the minimized sum of squared differences between the calculated and measured retentions of the 6 inert gases and was used to determine the goodness of fit of the model to the data.
Scintigraphy—Perfusion scintigrams of lung fields caudal to the heart were used to determine spatial distribution of pulmonary blood flow. Scintigrams were obtained with a gamma camera that had a large field of view and was fitted with a general-purpose low-energy collimator.c An aliquot of 500 MBq of technetium Tc 99m–macroaggregated human serum albumin was prepared in accordance with standard methods21 and injected slowly over 1 minute, which included 3 breaths in 2 horses and 8 to 10 breaths in the third horse, to ensure even distribution throughout the lungs. Images were acquired and processed with the aid of a nuclear medicine programd controlling the gamma camera. Each scintigram was obtained as a dynamic acquisition at a rate of 2 s/frame for 1 minute, with a resolution of 256 × 256 pixels. Pulmonary perfusion was evaluated in each horse before, during, and after PiNO administration. Neither the camera head nor the horses were moved between image acquisitions.
Each 2-second frame had a poor signal-to-noise ratio because of relatively few scintillation points, so the edges of the lungs were poorly defined. All frames were filtered with a medium-resolution Metz (fast Fourier transform) filter to remove high-frequency components (noise) and yield definitive margins of the lungs. To eliminate motion artifact caused by breathing, 20 images obtained at the expiratory pause (smallest-sized lung field) were randomly chosen and summed so that each image was a single static 10-second image. This process was performed for each horse before, during, and after PiNO administration. To compensate for decay of the tracer technetium Tc 99m, all frames were corrected for loss of radioactivity during the time from their acquisition to the time of the acquisition of images collected during PiNO administration. Because images acquired during PiNO administration and 30 minutes after administration included background radioactivity from the preceding injections, this background activity was subtracted from the images acquired during PiNO administration and 30 minutes after administration.
For each examination, activity in each pixel in the corrected summed image during PiNO administration was subtracted from pixel activity in the baseline image to yield images that showed the redistribution of blood flow (Figure 1). The resulting functional image identified the areas that increased and decreased in activity as a result of NO inhalation. The border between areas of increased and decreased activity was a well-defined narrow zone where perfusion did not change. Two ROIs (a dependent and a nondependent ROI) were drawn manually around the 2 areas of changing activity. The ROIs were copied and pasted onto all images of the examination, and total activity in each ROI was recorded. All ROIs were drawn and measured 4 times. Mean counts for each ROI in the images at administration of PiNO and 30 minutes after administration were plotted as the percentage change from the baseline value.
The frames with the largest lungs (at maximal inspiration) were extracted from each scintigram and saved as a summed static image. Regions of interest around the entire lung field of each of these static images were drawn automatically as an isocount line at a threshold of 15% of the maximum counts/pixel. The areas (ie, number of pixels in an ROI) were recorded at peak inspiration and at expiration, and the mean difference was calculated for each horse. This number yielded an indication of the relative tidal volumes.
Data collection—All data, including heart rate, SAP, MAP, DAP, SPAP, MPAP, DPAP, respiratory rate, tidal volume, arterial blood pH, mixed-venous blood pH, Pao2, Po2, Sao2, So2, Qt, and plasma endothelin-1 concentration, were measured or calculated before, during, and 30 minutes after PiNO administration. Data obtained by use of the MIGET and scintigraphy were collected at the same times. Following data collection after PiNO administration, the horses were moved to a padded stall and allowed to recover from anesthesia.
Data analysis—Because of the small number of horses, individual variables were listed as descriptive data, and no statistical analysis was performed for between-horse variability. However, mean data at each time point were analyzed for variability over time. Normal distribution was confirmed by examining the residuals from the model with a Shapiro-Wilk test. Repeated-measures ANOVA with a Bonferroni post hoc test was used to compare data obtained by use of the MIGET and physiologic data.e For scintigraphic data, the ratio of dependent ROI activity to the sum of dependent ROI activity and nondependent ROI activity was calculated and compared by means of a mixed-effects ANOVA, with horse as a random factor and event as a fixed factor, by means of a Bonferroni post hoc test.f Residuals from the model were tested with the Shapiro-Wilk test and found to be normally distributed. Values of P < 0.05 were considered significant.
Results
MIGET—The mean ± SD Pao2 (126 ± 58 mm Hg vs 50 ± 7 mm Hg) and Sao2 (98 ± 4% vs 85 ± 5%) increased significantly, whereas Qs/Qt (51 ± 7% vs 32 ± 7%) decreased significantly during PiNO administration, compared with the baseline value (Figures 2 and 3; Table 1). Values of VA/Q matching (as determined by the SD of the perfusion logarithmic distribution) improved in all horses during PiNO administration, compared with the baseline value (0.83 ± 0.17 vs 1.00 ± 0.17). Indices decreased toward the baseline value at 30 minutes after PiNO administration. Changes in Po2 and So2 paralleled those of Pao2 and Sao2. Changes in oxygenation, Qs/Qt, and VA/Q were less in 1 horse than in the other 2 horses. Residual sums of square of VA/Q distributions were low in all evaluations, which indicated a good fit of the model to the data.
Mean ± SD data for each of 3 dorsally recumbent isoflurane-anesthetized horses before, during, and 30 minutes after PiNO administration.
Variable | Before | During | After |
---|---|---|---|
Qs/Qt (%) | 51 ± 7 | 32 ± 7* | 51 ± 11 |
Pao2 (mm Hg) | 50 ± 7 | 126 ± 58* | 67 ± 18 |
Sao2 (%) | 85 ± 5 | 98 ± 4* | 91 ± 6 |
Po2 (mm Hg) | 30 ± 4 | 41 ± 9* | 37 ± 7 |
So2 (%) | 58 ± 9 | 73 ± 12* | 68 ± 12 |
Paco2 (mm Hg) | 62 ± 5 | 74 ± 9* | 62 ± 14 |
SD of perfusion logarithmic distribution | 1.02 ± 0.15 | 0.78 ± 0.10* | 1.03 ± 0.04 |
SD of tidal volume logarithmic distribution | 0.90 ± 0.11 | 1.07 ± 0.55 | 0.99 ± 0.13 |
Residual sums of square of VA/Q distributions | 0.58 ± 0.47 | 0.47 ± 0.63 | 0.07 ± 0.32 |
Heart rate (beats/min) | 37 ± 1 | 39 ± 2 | 43 ± 2 |
Qt (L/min) | 27 ± 1 | 25 ± 1 | 25 ± 2 |
MAP (mm Hg) | 53 ± 6 | 73 ± 12* | 73 ± 11 |
MPAP (mm Hg) | 15 ± 3 | 14 ± 3 | 17 ± 5 |
Respiratory rate (breaths/min) | 7 ± 4 | 6 ± 3 | 7 ± 4 |
Tidal volume (L) | 7 ± 3 | 7 ± 3 | 8 ± 4 |
Endothelin-1 (pmol/mL) | 5.59 ± 0.51 | 7.65 ± 1.91 | 5.49 ± 2.43 |
Value differs significantly (P < 0.05) from value before PiNO administration.
Scintigraphy—Percentage blood flow decreased to the dependent ROIs and increased to the nondependent ROIs in all horses during PiNO administration (Figure 4) and returned to baseline values at 30 minutes after PiNO administration. The differences between baseline values and values during PiNO administration, and between values during PiNO administration and values at 30 minutes after administration, were significant (P < 0.001). No difference was found between baseline values and values at 30 minutes after PiNO administration. The change in the lung area between end-expiration and peak inspiration was small in 1 of the 3 horses (Figure 5).
Other physiologic data—Total anesthesia time was 2 hours and 15 minutes. During anesthesia, all horses had various degrees of respiratory depression as indicated by an increase in Paco2 (Table 1). Two horses were hypercapnic (Paco2 > 45 to 50 mm Hg) at all times, and the other horse was hypercapnic during PiNO administration and 30 minutes after administration. Despite the hypercarbia, there were no significant differences in arterial or mixed-venous blood pH. Compared with results for the other 2 horses, 1 horse had a higher respiratory rate and lower tidal volume (Figure 4), but there were no differences in horses over time. The SAP, MAP, and DAP were higher during PiNO administration and 30 minutes after administration, compared with the baseline values. Two horses were hypotensive (MAP < 70 mm Hg) at baseline, and the other horse was hypotensive at all times. Observations for DAP and SAP were the same as those for MAP. The Po2 and So2 were higher during PiNO administration than at the other time points. Heart rate was higher during PiNO administration than the baseline value. Values of Qt, SPAP, MPAP, DPAP, and plasma endothelin-1 concentration did not change significantly during the experiments. Although recovery from anesthesia was not specifically evaluated, recovery was uncomplicated in all horses, as expected on the basis of results for a previous study.22
Discussion
Analysis of spatial scintigraphic images and data for the present study indicated that the pulmonary blood flow was redistributed en masse against gravity from dependent, presumably atelectatic lung fields to nondependent, presumably aerated lung fields during PiNO administration in dorsally recumbent isoflurane-anesthetized horses. A limitation of planar scintigraphy is that it provides a 2-D image and lungs are 3-D organs. Thus, we could not discern the magnitude of the redistribution across the lungs, nor could we quantify flow, but the results of the redistribution were readily measurable as increased Pao2, improved matching of VA/Q, and decreased Qs/Qt.
Although continuous delivery of iNO over the entire inspiratory phase of ventilation did not improve arterial oxygenation in a previous study23 of halothane-anesthetized horses, PiNO administration during the early phase of inspiration did counteract hypoxemia and Qs/Qt formation in dorsally recumbent13,15 and laterally recumbent11,12 isoflurane-anesthetized horses. It appears that pulsing iNO in the first 30% to 45% of the inspiratory phase, as was done in the present study, is the most effective method to improve oxygenation in anesthetized horses.12,13,23 However, the pulse dose can be too small to be effective, and PiNO administration in only 20% of the inspiratory phase can result in minimal or no improvement.12 Administration of PiNO at 45% of the inspiratory phase in a horse with a small tidal volume, as for 1 horse in this study, can result in less improvement than when a full dose is delivered. This could be caused by the fact that a small volume of iNO reaches the conducting airways but delivery to the alveoli (ie, area of gas exchange) is minimal. In addition to the lower tidal volume, that same horse in this study also received a decreased volume of iNO, compared with the volume of iNO in the other horses, because of a limitation in the capability of the prototype PiNO delivery unit. The unit delivers the expected volume of iNO in each breath when the breath duration is ≥ 7 seconds (ie, maximum 8 breaths/min). Breaths with a shorter duration do not allow the unit to return to normal delivery mode between breaths, and the unit will fail to activate with each breath. That horse breathed 12 times/min, which resulted in iNO delivery in some, but not all, breaths. However, even with the lower PiNO administration, changes in Pao2, VA/Q, and Qs/Qt were measurable.
In the study reported here, the horses were hypercapnic at all (2 horses) or some (1 horse) time points. Hypercarbia, which can alter gas exchange by decreasing the amount of oxygen available in the alveoli, is generally treated by means of positive pressure ventilation. However, this technique can decrease total blood flow and alter VA/Q matching and Qs/Qt2; thus, hypercarbia was not treated in this study. In a clinical situation, intermittent positive pressure ventilation would have been used to treat hypercarbia in anesthetized horses. Also, all horses were hypotensive at baseline, and 1 horse was hypotensive at all times. Hypotension, which can alter gas exchange by causing a decrease in pulmonary blood flo w, is generally treated in anesthetized horses by the administration of positive inotropic drugs such as dobutamine. We did not treat hypotension in this study because dobutamine can increase blood flow to all regions of the lungs, including atelectatic regions,24 thereby increasing Qs/Qt. In a clinical situation, dobutamine would have been administered to treat hypotension in anesthetized horses.
Because of the consistent changes in all 3 horses, the simultaneous data collection via the MIGET and scintigraphy, and the fact that both techniques are sensitive and specific for the indices measured, we believe that the results reported here are robust despite the low number of animals. Scintigraphy and the MIGET were performed simultaneously, allowing us to determine that the distribution of the blood flow visualized with scintigraphy changed at the precise time that the Qs/Qt was decreased and the VA/Q more evenly matched during PiNO administration (Figures 3 and 4). Furthermore, all major variables of interest (Pao2, Sao2, Qs/Qt, and pulmonary blood flow) had significant changes in the same direction in all 3 horses, including in the horse that received less iNO. However, the low number of animals may have made it difficult to detect changes in variables that have high inherent variability, such as plasma endothelin-1 concentration. However, it is also possible that endothelin-1 is not the controlling factor in alterations of pulmonary blood flow during administration of iNO or that endothelin-1 is released in localized lung regions, which makes systemic measurement of changes difficult or impossible.
In the present study, improved VA/Q matching and decreased Qs/Qt that occur during PiNO administration resulted from a redistribution of pulmonary blood flo w, against gravity, from dependent to nondependent lung regions. Addition of PiNO to inhaled gases could be used clinically to alleviate hypoxemia in anesthetized recumbent horses because hypoxemia is commonly caused by atelectasis.1,2
ABBREVIATIONS
DAP | Diastolic arterial blood pressure |
DPAP | Diastolic pulmonary arterial blood pressure |
iNO | Inhaled nitric oxide |
MAP | Mean arterial blood pressure |
MIGET | Multiple inert gas elimination technique |
MPAP | Mean pulmonary arterial blood pressure |
NO | Nitric oxide |
PiNO | Pulse-delivered inhaled nitric oxide |
Po2 | Mixed-venous partial pressure of oxygen |
Qs/Qt | Pulmonary blood shunting |
Qt | Total perfusion |
ROI | Region of interest |
SAP | Systolic arterial blood pressure |
SPAP | Systolic pulmonary arterial blood pressure |
So2 | Mixed-venous oxygen saturation |
VA/Q | Ventilation-perfusion |
Datex-Ohmeda Research Unit, Helsinki, Finland.
5890 series II, Hewlett Packard, Atlanta, Ga.
Picker SX 300, Picker International Inc, Cleveland, Ohio.
Hermes Medical Solutions, Hägersten, Sweden.
Prism, version 5, GraphPad Software Inc, La Jolla, Calif.
SAS Data Management, SAS Institute Inc, Cary, NC.
References
1. Nyman G, Hedenstierna G. Ventilation-perfusion relationships in the anaesthetised horse. Equine Vet J 1989; 21:274–281.
2. Nyman G, Funkquist B & Kvart C, et al. Atelectasis causes gas exchange impairment in the anaesthetised horse. Equine Vet J 1990; 22:317–324.
3. Day TK, Gaynor JS & Muir WW, et al. Blood gas values during intermittent positive pressure ventilation and spontaneous ventilation in 160 anesthetized horses positioned in lateral or dorsal recumbency. Vet Surg 1995; 24:266–276.
4. Dodam JR, Moon RE & Olson NC, et al. Effects of clenbuterol hydrochloride on pulmonary gas exchange and hemodynamics in anesthetized horses. Am J Vet Res 1993; 54:776–782.
5. Keegan RD, Gleed RD & Sanders EA, et al. Treatment of low arterial oxygen tension in anesthetized horses with clenbuterol. Vet Surg 1991; 20:148–152.
6. Gleed RD, Dobson A. Improvement in arterial oxygen tension with change in posture in anaesthetised horses. Res Vet Sci 1988; 44:255–259.
7. Whitehair KJ, Willits NH. Predictors of arterial oxygen tension in anesthetized horses: 1,610 cases (1992–1994). J Am Vet Med Assoc 1999; 215:978–981.
8. Shawley RV, Mandsager RE. Clinical use of positive-pressure ventilation in the horse. Vet Clin North Am Equine Pract 1990; 6:575–585.
9. Nyman G, Frostell C & Hedenstierna G, et al. Selective mechanical ventilation of dependent lung regions in the anaesthetized horse in dorsal recumbency. Br J Anaesth 1987; 59:1027–1034.
10. Moens Y, Lagerweij E & Gootjes P, et al. Differential artificial ventilation in anesthetized horses positioned in lateral recumbency. Am J Vet Res 1994; 55:1319–1326.
11. Creagh-Brown BC, Griffiths MJ, Evans TW. Bench-to-bedside review: inhaled nitric oxide therapy in adults. Crit Care 2009; 13:221.
12. Grubb T, Frendin JH & Edner A, et al. The effects of pulse-delivered inhaled nitric oxide on arterial oxygenation, ventilation-perfusion distribution and plasma endothelin-1 concentration in laterally recumbent isoflurane-anaesthetized horses. Vet Anaesth Analg 2013; 40:e19–e30.
13. Nyman G, Grubb TL & Heinonen E, et al. Pulsed delivery of inhaled nitric oxide counteracts hypoxaemia during 2.5 hours of inhalation anaesthesia in dorsally recumbent horses. Vet Anaesth Analg 2012; 39:480–487.
14. Heinonen E, Nyman G & Merilainen P, et al. Effect of different pulses of nitric oxide on venous admixture in the anaesthetized horse. Br J Anaesth 2002; 88:394–398.
15. Heinonen E, Hedenstierna G & Meriläinen P, et al. Pulsed delivery of nitric oxide counteracts hypoxaemia in the anaesthetized horse. Vet Anaesth Analg 2001; 28:3–11.
16. Hambraeus-Jonzon K, Bindslev L & Frostell C, et al. Individual lung blood flow during unilateral hypoxia: effects of inhaled nitric oxide. Eur Respir J 1998; 11:565–570.
17. Hambraeus-Jonzon K, Chen L & Freden P, et al. Pulmonary vasoconstriction during regional nitric oxide inhalation: evidence of a blood-borne regulator of nitric oxide synthase activity. Anesthesiology 2001; 95:102–112.
18. Wagner PD. The multiple inert gas elimination technique (MIGET). Intensive Care Med 2008; 34:994–1001.
19. Hedenstierna G, Nyman G & Kvart C, et al. Ventilation-perfusion relationships in the standing horse: an inert gas elimination study. Equine Vet J 1987; 19:514–519.
20. Wagner PD, Saltzman HA, West JB. Measurement of continuous distributions of ventilation-perfusion ratios: theory. J Appl Physiol 1974; 36:588–599.
21. Votion D-M, LeKeux P. Pulmonary scintigraphy. In: Dyson SJ, Pilsworth RC, Twardock AR, et al, eds. Equine scintigraphy. Newmarket, Suffolk, England: Equine Veterinary Journal Ltd, 2003; 263–278.
22. Grubb T, Edner A & Frendin JH, et al. Oxygenation and plasma endothelin-1 concentrations in healthy horses recovering from isoflurane anaesthesia administered with or without pulse-delivered inhaled nitric oxide. Vet Anaesth Analg 2013; 40:e9–e18.
23. Young LE, Marlin DJ & McMurphy RM, et al. Effects of inhaled nitric oxide 10 ppm in spontaneously breathing horses anaesthetized with halothane. Br J Anaesth 1999; 83:321–324.
24. Swanson CR, Muir WW III. Dobutamine-induced augmentation of cardiac output does not enhance respiratory gas exchange in anesthetized recumbent healthy horses. Am J Vet Res 1986; 47:1573–1576.