An Impedance-Based Model for the Assessment of Cardiopulmonary Function in Rabbits

Abstract

Improving the quality of physiologic data collected from research animals is most easily accomplished by collecting as much information as possible from a single subject, thereby reducing animal use and error associated with satellite groups. We investigated the feasibility of using a large-animal implantable telemetry device in New Zealand white rabbits (n = 6). The first task was to develop an implantation technique that yielded calibrated tidal volume (Vt) measurements that were within 10% of those obtained simultaneously from a pneumotachograph, a low-noise electrocardiogram, and stable blood pressure. The second task was to challenge implanted rabbits with the respiratory stimulant doxapram to assess linearity of the calibration across a range of Vt. Of the 3 electrode placements attempted, only one resulted in calibrations consistently below 10% error. Optimal electrode placement resulted in calibrated Vt measurements within 1.7% ± 0.3% of those obtained from a pneumotachograph during normal tidal breathing, 7.3% ± 0.7% of those after saline injection, and 6.0% ± 0.5% of those after doxapram injection. The Vt range was 9 to 15 mL for normal tidal breathing and saline injection and 25 to 30 mL after doxapram injection. Increases in mean arterial pressure of 25.0 ± 6.82 mm Hg and decreases in heart rate of 56.3 ± 6.82 bpm were associated with doxapram injection only. Our findings represent the first time that multiple cardiopulmonary endpoints have been assessed by telemetry in conscious, restrained rabbits. Whether animal position affects calibration accuracy warrants investigation.

Two of the greatest challenges in performing inhalation toxicology studies are appropriately determining the dose delivered to test subjects and monitoring the function of major organ systems during and after exposure. Most often, aerosol concentrations are presented in units of milligrams of material per cubic meter of air in the exposure atmosphere (mg/m³), and dosages are presented as the aerosol concentration multiplied by the duration of exposure (mg/m³ × min). When performing in vivo inhalation toxicity studies, determining the actual dosimetry is complicated by the difficulty in knowing the deposition ratios of the aerosol and the differences in respiratory rates and tidal volumes (Vt) between individual animals in an exposure group. To assist in determining accurate dosimetry, accurate measures of the respiratory rate and tidal volume during the exposure period are necessary. These parameters are easily obtained in small animal models, such as rats and mice, by using a nose-only or head-only exposure system coupled to a plethysmograph.

Although the use of a plethysmograph will enable respiratory parameters to be monitored during the exposure period, the assessment of respiratory function after exposure requires repeatedly replacing animals in the plethysmograph at various time points to collect data. Doing so might induce artifacts in the data due to restraint-associated stress. Furthermore, the use of a plethysmograph in large animals may not always be compatible with the exposure methodology. Finally, whole-body inhalation exposures are often conducted for compounds that exhibit both inhalational and percutanous hazards, thus eliminating the use of all monitoring options but whole-body plethysmographs, which are considered less sensitive than other direct measures.¹³ Some of these issues can be resolved through the use of implantable telemetric devices.

Recent advancements in implantable telemetric devices have enabled the assessment of cardiopulmonary function in animal species larger than rodents, such as cynomolgus monkeys, beagle dogs, and miniature swine.^2,3,8 These transmitters combine an impedance pneumography technique with existing blood pressure and ECG monitoring capabilities, thus achieving continuous collection of data pertaining to respiratory rate and volume, blood pressure, ECG, body temperature, and overall animal activity.

At our facility, rabbits are often used as a nonrodent animal model for risk assessment purposes due to their specific advantages over large animal models for inhalation toxicity studies. Rabbits are small enough to fit within the whole-body and nose-only exposure systems used in our facility but they have a relatively large body mass, as compared with rodent species. Their additional body mass allows for more accurate extrapolation of data to generate human estimates for the toxicity of chemical warfare agents. Furthermore, rabbits have been used recently in cardiovascular and respiratory studies and appear to be increasing in use within the scientific community.^4,11 The present study investigates the feasibility of applying existing impedance pneumography techniques to the rabbit model, with the specific goal of applying this technology to inhalation toxicology studies.

Materials and Methods

The present study was performed in an AAALAC-accredited facility. All animal use complied with the Guide for the Care and Use of Laboratory Animals and the Animal Welfare Act¹ and adhered to an animal-use protocol approved by the local IACUC at the Edgewood Chemical Biological Center.

Selection of optimal electrode positioning.

In rats, electrode position influences the magnitude of the impedance signal.⁴ Our lab previously determined that placing the impedance electrodes on the chest of rabbits results in an unsuitable magnitude of impedance (very little lateral thoracic expansion is associated with normal respiration in rabbits). Therefore, several electrode configurations were investigated to determine optimal placement to achieve increased magnitude (Figure 1 A through C). The first configuration consisted of 4-cm loops, with negative electrodes on the diaphragm at the 9- and 12-o'clock positions, as viewed in the transverse plane caudocranially. Positive electrodes were placed on the left side of the chest, with one electrode caudal and one electrode cranial to a line along the transverse plane at mid chest (just above xiphoid process). Another configuration consisted of a change in the location of the chest electrodes, in that they were placed with one on the left side of the chest and the other on the right side of the chest at the level of the manubrium. In the final configuration, the electrode placement on the diaphragm was oriented such that the electrodes spanned the entire width of the diaphragm. This placement required increasing the size of the exposed electrode loop to 8 cm to span the diaphragm, with the same chest electrode configuration as in the previous configuration but with 8-cm instead of 4-cm loops.

Figure 1.

Schematic of potential electrode configurations. (A) This cofiguration has 4-cm loops with negative electrodes on the diaphragm in the 9- and 12-o'clock positions and positive electrodes on right side of the chest. (B) This configuration has 4-cm loops, with chest electrodes placed bilaterally at the level of the manubrium. (C) Here, 8-cm loops span the diaphragm and are placed bilaterally on the chest at the level of the manubrium.

We used 3 male New Zealand White rabbits (Millbrook Labs, Amherst, MA) to determine the optimal electrode configuration. Each animal underwent doxapram challenge only, allowing for evaluation of the overall error associated with each configuration. This information was used to select the configuration associated with the lowest error (Figure 1 C) for final evaluation. A separate group of 6 rabbits then was used to evaluate the accuracy of the optimal electrode configuration. Each rabbit underwent a normal tidal breathing experiment, a saline negative-control experiment, and a doxapram-challenge experiment, with a minimum of 24 h between each experiment, as described later.

Final surgical approach.

Male New Zealand White rabbits (n = 6; weight, 2.5 to 3.5 kg) were implanted with a telemetric transmitter (model TL11M3-D70-PCTR, Data Sciences International, St. Paul, MN) outfitted with a small-animal pressure sensor. This device is capable of measuring ECG, blood pressure, body temperature, animal activity, and impedance-based respiratory rate and volume. All rabbits were sedated by using a ketamine (45 mg/kg IM)–xylazine (4 mg/kg IM) cocktail. After sedation, the entire ventral surface was clipped, from approximately 4 cm above the manubrium of the chest extending caudally to the inguinal region. In addition, rabbits received buprenorphine (0.05 mg/kg IM), glycopyrrolate (0.1 mg/kg IM), and cefazolin (20 mg/kg IV). Rabbits were transferred to a heated surgical table, placed in dorsal recumbancy, and maintained on isoflurane anesthesia (1.5% to 2.5%, 1.0 L/min O₂). A midline incision was made along the linea alba, beginning at the xiphoid process and extending caudally to the inguinal region. The telemetry unit was tacked to the inner abdominal wall by using 4-0 suture (Ethibond Excel, Ethicon, Somerville, NJ). The left iliac artery was isolated, and the blood pressure catheter was introduced, advanced approximately 5 cm into the abdominal aorta, and secured in place by using tissue glue (Vetbond, 3M, St Paul, MN). The liver was retracted, the fascia connecting the liver to the diaphragm was cut down to its junction with the vena caval hiatus, and the negative impedance electrodes were secured to the diaphragm by using 5-0 suture (Ethibond Excel, Ethicon). The positive ECG electrode was sutured to the inner abdominal wall under the last rib on the rabbit's left side as in lead II configuration by using 4-0 suture (Ethibond Excel, Ethicon). A 1- to 2-cm transverse incision was made at the manubrium, and the positive impedance electrodes and negative ECG electrodes were tunneled subcutaneously from the abdominal incision to the manubrium. Subcutaneous pockets were formed on both sides of the manubrium, and the positive impedance electrodes were placed in these pockets and secured in place by using 5-0 suture (Ethibond Excel, Ethicon). The negative ECG electrode was placed in a subcutaneous pocket just cranial to the impedance electrode on the animal's right chest and secured in place with 5-0 suture (Ethibond Excel, Ethicon). The position of the impedance electrode was adjusted during surgery to yield optimal signal quality and to develop a standard procedure for this technique. A representative radiograph of the electrode placement is provided in Figure 2. All skin and muscle incisions were closed with 5-0 polydiaxone suture (PDS II, Ethicon). During recovery, the area between the scapulae was clipped, and a transdermal fentanyl patch (25 μg/h) was placed; the patch was left in place for 48 to 72 h as indicated according to the results from a pain-scoring sheet. Meloxicam (0.2 mg/kg SC) was administered daily for 5 d after surgery. Weight loss and reduced gastrointestinal motility were common among all rabbits.

Figure 2.

Representative radiographs (A, ventral view; B, left lateral view) of a rabbit implanted with a telemetric transmitter. A solid-tipped, intravenous, negative ECG electrode was used in this animal. White, negative ECG electrode; red, positive ECG electrode; blue, positive impedance electrodes; purple, negative impedance electrodes.

All rabbits were allowed to recover for at least 10 d prior to the beginning of any data collection. In cases where weight loss that was greater than expected, testing was delayed until rabbits displayed a return to a regular rate of weight gain and normal behaviors (freely moving about the cage, normal intake and output).

Calibration and data acquisition.

Calibration of the impedance signal was performed through the simultaneous collection of respiratory rate and volume from both the telemetric transmitter and a facemask pneumotachograph in restrained animals. Conscious rabbits were restrained in a normal standing posture, and the pneumotachograph (10 L/min; series 8300, Hans Rudolph, Shawnee, KS) was attached to an acrylic facemask that used a heavy latex seal around the nose and mouth of the animal. The signals from the pneumotachograph were detected by a differential pressure transducer (DP45, Validyne Engineering, Northridge, CA), amplified through an analog-to-digital signal conditioner (ACQ 7700, Data Sciences International), and recorded by using a data exchange matrix (Data Sciences International). Signals from the pneumotachograph and the telemetric transmitter were interfaced with the Ponemah Physiology Platform version 5.0 by using the Pulmonary Air Flow and Pulmonary Volume analysis modules (Data Sciences International).

Study design.

Prior to baseline studies, animals were acclimated to a restraint device (model 502-AR, Plas Labs, Lansing, MI) incrementally over 3 sessions of 5, 10, and 15 min in duration. The accuracy of the impedance pneumography method was evaluated under 3 conditions: baseline tidal breathing, tidal breathing after saline administration, and tidal breathing after doxapram administration. Doxapram HCl (20 mg/mL, Wedgewood Pharmacy, Swedesboro, NJ) was administered as a single IV bolus (5 mg/kg) through the marginal ear vein. Data for baseline tidal breathing were recorded for 10 min on the day after the completion of acclimation procedures. Data for tidal breathing after saline or doxapram administration were recorded for 15 min (5 min baseline, 10 min after challenge) on different days after the baseline recording days. Doses of saline (equivolumetric to the doxapram dose) were administered to each animal on 2 separate study days prior to doxapram challenge. The purpose of saline administration was to determine the component of the Vt increase that was due to handling and injection and to assess whether increases in error were an effect of doxapram or otherwise associated with some aspect of administering the injection.

Doxapram challenge was chosen to assess the linearity of the Vt calibration across a range of tidal volumes in a conscious rabbit. Each doxapram dosing session was repeated 2 times per animal, once a week, for 2 consecutive weeks. Data from the pneumotachograph and telemetric transmitter were collected simultaneously throughout the dosing session. These data were analyzed by using the Pulmonary Volume analysis module in the Ponemah Physiology Platform (Data Sciences International).

Data analysis.

Direct Vt data are presented as the mean of all 6 rabbits averaged every 20 s for the duration of the recording. Individual calibration was always performed within the first 5 min of the baseline data. The Student t-test (α = 0.05) was used to test for statistical difference between Vt data derived from the pneumotachograph compared with the telemetric transmitter. When postchallenge data were compared with baseline data, a single mean baseline value was calculated (average of the 5-min baseline period), and 1-min postchallenge means were compared with the baseline mean by using one-way ANOVA (α = 0.05) and a Tukey posttest. All average data are reported as means ± SEM.

Results

Calibration during conscious tidal breathing.

After acclimation each animal was placed in restraint and baseline data were collected for 10 min. The mean Vt derived from the pneumotachograph was 11.9 ± 0.2 mL (mean ± SEM), as was that from the calibrated impedance signal. Data shown in Figure 3 A are 20-s averages of Vt for both the pneumotachography- and impedance-derived values. The Vt over the course of the 10-min baseline session ranged from 14.5 to 10.6 mL and stabilized after the rabbits had been in the restraint for about 5 min. The mean error between the pneumotachography- and impedance-derived Vt for the duration of the recording period was 1.7% ± 0.3%.

Figure 3.

Baseline cardiopulmonary measures. (A) Baseline tracing representing 20-s averages of Vt acquired through pneumotachograph (solid circles) and telemetric transmitter (open circles; error bars represent SEM). (B) Baseline heart rate in restrained rabbit (error bars represent 1 SD). (C) Baseline MAP in restrained rabbit (error bars represent 1 SD).

Cardiovascular response during calibration displayed similar trends to Vt. The average heart rate decreased from 220 bpm to 199 bpm; this trend did not stabilize over the course of the 10-min calibration session (Figure 3 B). The overall average heart rate for the entire duration was 209 ± 6 bpm. The overall average mean arterial blood pressure (MAP) for the entire duration was 89.9 ± 2.1 mm Hg; the average MAP increased from a low of 85.1 mm Hg to a maximum of 93.0 mm Hg and stabilized around 90 mm Hg after 6 min in restraint (Figure 3 C).

Evaluation of response to negative control.

After saline challenge (negative control), the average peak Vt derived from the pneumotachograph was 12.8 ± 1.0 mL, and that from the calibrated impedance signal was 10.8 ± 0.4 mL. These values do not represent a statistically significant change from baseline Vt after administration of saline, nor are they statistically different from one another (Figure 4 A). The average error between the pneumotachography- and impedance-derived Vt over the course of the recording period was 7.3% ± 0.7%, and the maximum was 17.7%.

Figure 4.

Cardiopulmonary response to saline challenge. (A) Tracing representing 20-s averages of Vt acquired through pneumotachograph (solid circles) and telemetric transmitter (open circles) during saline challenge (error bars represent SEM; vertical line represents point of saline injection). (B) Effect of saline challenge on heart rate in restrained rabbit (error bars represent 1 SD). (C) Effect of saline challenge on MAP in restrained rabbit (error bars represent 1 SD).

The cardiovascular response during saline challenge was similarly unremarkable. No significant trend was observed in heart rate as a result of saline administration (Figure 4 B). The average heart rate for the 10-min postsaline period was 203 ± 4 bpm, and the average MAP was 88.9 ± 1.9 mm Hg. The average MAP increased from a low of 86.0 mm Hg to a maximum of 92.1 mm Hg, but this value was prior to saline injection. MAP returned to values around 87 mm Hg by the end of the recording period (Figure 4 C).

Evaluation of response to positive-control drug.

The respiratory response to doxapram challenge (positive control) was characterized by a rapid, transient increase in tidal volume. The average peak Vt derived from the pneumotachograph was 25.6 ± 1.9 mL, and that from the calibrated impedance signal was 23.6 ± 1.4 mL. This difference represents approximately a 100% increase in volume over baseline. Over the course of the entire recording period, only one data point differed significantly between the pneumotachography- and impedance-derived Vt values (Figure 5 A). Furthermore, the average error between the pneumotachography- and impedance-derived Vt over the course of the recording period was 6.0% ± 0.5%, and the maximal error was 13.4%.

Figure 5.

Cardiopulmonary response to doxapram challenge. (A) Tracing representing 20-s averages of Vt acquired through pneumotachograph (solid circles) and telemetric transmitter (open circles) during doxapram challenge (error bars represent SEM; vertical line represents point of doxapram injection; *, P < 0.05). (B) Effect of doxapram challenge on heart rate in restrained rabbit (error bars represent 1 SD;*, P < 0.05). (C) Effect of doxapram challenge on MAP in restrained rabbit (error bars represent 1 SD;*, P < 0.05).

Like the Vt response, the cardiovascular response to doxapram challenge was marked. Heart rate decreased from a baseline of 215 ± 1 bpm to a minimum of 189 ± 5 bpm, representing a 12.3% decrease in heart rate (Figure 5 B). Mean arterial blood pressure rapidly increased from an average baseline of 92.2 ± 0.5 mm Hg to a maximum of 113.0 ± 2.2 mm Hg, representing a 22.7% increase in MAP. This increase gradually declined but did not return to baseline by the end of the recording period (Figure 5 C).

Discussion

Impedance pneumography has been available for human clinical use since the 1960s.^5,6,9,10 Since that time, this technique has come to play a role in several nonclinical basic medical research laboratories. External electrodes have been used to measure and accurately calibrate Vt in rats.⁵ Furthermore, the technique of telemetric impedance pneumography used in the present study has been successfully applied in cynomolgus monkeys, beagle dogs, and miniature swine.^2,3,8 However, in all of these studies, the electrodes were placed bilaterally on the chest to measure transthoracic impedance as a function of lateral thoracic expansion. In rabbits, initial experiments confirmed that bilateral electrode placement on the chest yields an inadequate calibration of the transthoracic impedance signal. This outcome is likely due to the respiratory mechanics of rabbits, in that they are obligate diaphragmatic breathers as compared with obligate chest breathers. Therefore, the present study set out to identify an electrode placement that yielded an accurately calibrated impedance signal in rabbits. Figure 1 illustrates the placement of all electrodes for this model.

The initial electrode placement for rabbits called for negative electrodes on the diaphragm in the 9- and 12-o'clock positions as viewed in the transverse plane, caudocranially. Positive electrodes were placed on the left side of the chest, with one electrode caudal and one electrode cranial to an imaginary line along the transverse plane at the manubrium (Figure 1 A). The error between the impedance and pneumotachograph Vt values from this placement were routinely 25% to 40%, especially at higher Vt associated with doxapram challenge. An improvement in error was achieved when the chest electrodes were placed with one on the left side of the chest and the other on the right side of the chest at the level of the manubrium (Figure 1 B), but the error with this configuration remained in the 25% range. Finally, the electrode placement on the diaphragm was oriented such that the electrodes spanned the entire width of the diaphragm (Figure 1 C). We theorized that this configuration would result in increased accuracy at higher volumes, given that the previous placement at the 9- and 12-o'clock positions missed capturing the point of maximal diaphragmatic deflection. The revised placement required increasing the size of the exposed electrode to 8 cm to span the diaphragm. The result was an optimized electrode placement that resulted in error of less than 10% on a routine basis during normal tidal breathing. With electrode placement optimized, we evaluated the accuracy and calibration requirements.

The method of calibration used in this study was an individual calibration on the day of study. Throughout the acclimation period, calibration values (slope and intercept of linear regression curve provided in Pulmonary Volume analysis module) were monitored for variability. After the collection of these calibration values, we attempted to use a group-average calibration and an individual average calibration value. The data in Table 1 represents the individual slope and intercept values of the calibrated regression curves taken on 5 nonconsecutive days from each rabbit. For any animal, averaging these values and then applying that average as the calibration results in error that exceeds the 10% threshold. In addition, choosing just one of these values from any given day and applying it to all days of collection results in error in excess of the 10% threshold. In Beagle dogs, using individual mean or group mean calibration values appears to provide acceptable accuracy.⁸ As expected, individual calibration remains the most accurate method in rabbits.

Table 1.

Individual slope and intercept values on nonconsecutive study days

	Rabbit no.
	1		2		3		4		5		6
Day	Slope	Intercept	Slope	Intercept	Slope	Intercept	Slope	Intercept	Slope	Intercept	Slope	Intercept
1	3.2126	–125.2	4.3680	–172.2	3.6288	–134.3	3.0802	–128.0	3.1948	–202.4	2.8455	–119.2
2	2.8017	–132.8	4.2581	–154.1	3.0018	–116.2	3.1829	–138.3	2.4697	–168.6	2.6779	–112.2
3	2.7509	–145.7	3.6831	–160.6	3.0304	–110.4	3.7604	–160.7	2.7689	–184.3	3.7134	–126.2
4	2.3627	–108.3	3.2297	–123.9	2.6592	–106.0	4.2899	–169.4	3.3026	–214.5	6.1028	–127.9
5	2.0207	–90.1	3.6815	–142.8	3.1673	–118.8	2.7152	–141.1	3.1797	–195.0	6.0235	–131.0

Data shown are the absolute slope and intercept from the regression analysis used to calibrate the impedance waveform with pneumotachography-derived volumes.

The average values for Vt, MAP, and heart rate (11.9 mL, 89.9 mmHg, and 209 bpm, respectively) acquired during baseline recordings of rabbits fell within normal physiologic limits. Cardiopulmonary parameters for New Zealand white rabbits are: resting heart rate, 200 to 300 bpm; normotensive MAP, 85 to 110 mm Hg, and resting tidal volume, 4 to 6 mL/kg.^4,7,12 The average error associated with the Vt measurement during normal tidal breathing was 2%, indicating a high level of accuracy between the pneumotachograph and telemetry implant. In comparison, in studies using beagle dogs, individual calibration factors resulted in approximately 5 to 6% error over the course of a doxapram dosing regimen.⁸ This finding indicates that the rabbit model is well within the limits of accuracy observed in other species while at rest and is a reasonable starting point to evaluate the accuracy of calibration across a range of tidal volumes.

The saline-control treatment served 2 purposes for this study. The first purpose was to determine the component of the Vt increase seen during doxapram administration that might be due to animal handling and injection alone. The peak impedance-derived Vt after saline injection was 10.7 mL, which represents a 1.4% increase over baseline and was not statistically significant from baseline tidal volumes. This result indicates that all significant increases in Vt can be attributed to doxapram and that they are not artifacts of vehicle injection. This finding contrasts with that in beagle dogs, in which saline administration is associated with a 76% increase in Vt over baseline.⁸ The second purpose of the saline control in rabbits was to assess whether increases in error after doxapram injection were due to the volume of injection or to an effect of doxapram. The peak error between the pneumotachograph and impedance Vt after saline injection was 17.7% whereas that after doxapram injection was 13.4%. Although the error after saline injection is slightly greater than that after doxapram, it is important to note that separation between the pneumotachography and impedance data after injection (Figures 3 A and 4 A) shows a similar trend regarding both the saline and doxapram injections. These data indicate the increase in error after injection may be due to the volume being injected or to a change in mechanics due to handling, rather than to an effect of doxapram itself. These aspects do not fall under the scope of the current study, and with a larger study population, these errors might decrease. As expected, cardiovascular parameters remained in the normal range throughout the saline recording period.

The doxapram treatment was designed to evaluate whether the calibration performed on the baseline data prior to challenge maintained linearity when Vt was increased in a conscious model. The Vt increase after doxapram administration was at least 23 to 25 mL on average, representing approximately a 100% increase in Vt over baseline. The overall average error throughout the entire recording period was 6.0%, with a peak error of 13.4% occurring just after the maximal volume. In fact, error increased as Vt increased in all cases of doxapram challenge. This effect can be explained by a previously described nonlinear viscoelastic model.¹⁴ That model describes the nonlinear relationship between pressure and lung volume that occurs at moderate to high lung volumes in rabbits. At moderate lung volumes (above approximately 23 mL in 2.5- to 3.5-kg rabbits), an exponentially greater pressure is required to achieve higher lung volumes. This situation is due to the fact that the thoracic cage reaches a point of maximal expansion, after which the lungs are still able to be inflated. The functional outcome of this phenomenon in impedance-based pneumography is that as lung volume increases, the relationship between impedance and volume becomes similarly nonlinear due to decreased thoracic expansion. Therefore, although the impedance signal diminishes as thoracic expansion reaches a maximum, the volume measured at the pneumotachograph continues to increase because the lung tissue is still compliant to additional volume. The current study was not designed to investigate the factors involved in this nonlinear relationship; furthermore, the software-based calibration is built on a simple linear regression, which loses validity as lung volumes reach the upper physiologic limits. In addition, compensatory increases in MAP of as much as 22.6% above baseline and decreases in heart rate down to 12.2% below baseline occurred after doxapram administration.

Finally it should be noted that most rabbits experienced significant weight loss after surgery. The average weight loss across the group was 12.1% ± 0.9%. The average number of days required to achieve daily weight gain was 7 ± 3 d, and the average number of days required to return to the presurgical weight was 28 ± 10 d. In addition, because our New Zealand white rabbits had an intense inflammatory response to the polyglycolic acid suture we used, we recommend using polydioxanone suture, which seems to minimize this response.

In conclusion, this study demonstrated the feasibility of using telemetry-based impedance pneumography in rabbits and revealed possible limitations and areas for improvement in this model. Although the accuracy of this method may be limited at higher lung capacities, impedance pneumography is an accurate method for measuring Vt under normal respiratory volumes in conscious restrained rabbits. Additional research is required to determine whether error increases due to animal posture or as a function of respiratory mechanics. The data generated during the doxapram challenge demonstrate the utility of monitoring cardiopulmonary parameters in a single animal, the primary advantages of which are time- and dose-matched data and reduction of overall animal use. Finally, this model provides a method for gathering data that previously were inaccessible in whole-body inhalation studies.