Pediatric Pulmonology 10:208-213 (1991)

Therapeutic Methods -

Aerosol Delivery in Neonatal Ventilator Circuits: A Rabbit Lung Model

Duncan Cameron,' Rosemary Arnot,* Michelle Clay,' and Michael Silverman'

Summary. The benefits of inhaled therapy in ventilated neonates are recognized, but the reliability of drug delivery in nebulizer-ventilator circuits is uncertain. We quantified the effect of changing variables. Twenty-three freshly killed rabbits (1.15-1.9 kg) were ventilated via a tracheostomy by a pressure-limited, time-cycled ventilator (Neovent). A radioaerosol of g g T ~ m pertechnetate from an Ultravent nebulizer (Mallinkrodt) was fed into the proximal ventilator tubing. Two 3-minute nebulizations at "standard settings" were followed by 2 at altered pressure, frequency, gas flow, I:E ratio, or position of the nebulizer in the circuit. Each nebulization was followed by a 3-minute gamma camera image and total deposited radioactivity was measured in excised lungs and trachea. Images demonstrated good peripheral aerosol deposition. At standard settings, lung deposition averaged 2.8% of the aerosol released. This was decreased markedly by reducing tidal volume (ventilator pressures) and residence time of aerosol (I:E ratio). Reduced gas flow decreased deposition slightly, presumably by increased particle size and marginally reduced tidal volume. Deposition did not change with increased frequency; increased minute ventilation was offset by decreased residence time of the aerosol. We conclude that the Ultravent nebulizer can be used to nebulize drugs in a standard neonatal circuit, although the dose delivered is small. Tidal volume and aerosol residence time are important determinants of aerosol delivery. Pediatr Pulmonol 1991 ; 10:20&213.

Key words: Jet nebulizer; pressure-limited, time-cycled neonatal ventilator; 9 9 T ~ m radio- aerosol; gamma-camera image; tissue radioactivity; effects of ventilator settings.

INTRODUCTION

Increasing interest has been shown in the use of nebulizers to deliver aerosol therapy to ventilated new- born babies. This has centered around the use of bron- chodilators in infants with bronchopulmonary dys- plasia,'?' but other classes of drugs are also under preliminary investigation including sodium cromoglycate and topical corticosteroids. However, such studies have been performed with little idea of the actual dose of drug received. The inhaled route of therapy offers the twin advantages of a high local concentration of drug to the lung without systemic side-effects. Ventilated subjects, however, pose particular problems in effective adminis- tration of aerosol, as large amounts of aerosol are deposited in ventilator tubing and proximal airways without reaching peripheral a i r ~ a y s . ~ Quantification of aerosol reaching the lungs has proved difficult, but some work with volume-cycled ventilation of rabbits has already suggested that different inspiratory-expiratory ratios will affect lung deposition of aerosol^.^ This is in 0 1991 Wiley-Liss, Inc.

keeping with observations on radioaerosol deposition in spontaneously breathing adults where breathing patterns also influence deposition.3

For purposes of drug administration, it is obviously important to identify factors that may influence delivery. Our study aimed to investigate the effect of different ventilation parameters, namely pressure, inspiratory ex- piratory (1:E) ratios, breathing rate, and gas flow rate, upon aerosol deposition. For this purpose a ventilated rabbit-lung model was developed. Sequential gamma-

From the Departments of Pediatrics and Neonatal Medicine' and Medical Physics,' Royal Postgraduate Medical School, Hammersmith Hospital, London, England.

Received July 20, 1989; (revision) accepted for publication November 14. 1990.

Address correspondence and reprint requests to Dr. M. Silverman, Department of Pediatrics and Neonatal Medicine, Royal Postgrad- uate Medical School, Hammersmith Hospital, London W 12 OHS, England.

Aerosol Delivery in Neonatal Ventilator Circuits 209

An Ultravent jet nebulizer (Mallinkrodt Ltd., United Kingdom) was inserted into the inspiratory limb of each circuit via a T-piece connector 24 cm from the endotra- cheal tube. This nebulizer produces a high-output, fine- particle-size aerosol and is widely used for diagnostic lung imaging; its advanta es in ventilator circuit have already been highlighted. The entire gas flow to the circuit was provided through the nebulizer, thus produc- ing a retrograde gas flow through the inspiratory line and obviating the need for a humidifier in the circuit proximal to the nebulizer.

The aerosol mass median diameter (AMMD) was measured with a Malvern 2600 laser particle sizer (Malvern Instruments, United Kingdom) at different flow rates. Particle sizing was performed with four different Ultravent nebulizers, each studied on three occasions. Particle size was measured as an aerosol of normal saline emerged from an endotracheal tube attached to a connec- tor, tubing and nebulizer in an arrangement similar to that used for ventilating the rabbits.

?

camera imaging followed by assessment of tissue radio- activity was used to measure lung deposition of a 99Tcm pertechnetate radioaerosol which was introduced by a nebulizer into the ventilator circuit at different ventilator settings. Clearance of pertechnetate from the lung is rapid with a half-time of about 12 minutes in humans,6 and for this reason rabbits were studied in the immediate post mortem period to avoid clearance by the bronchial and pul rnonary circulation.

MATERIALS AND METHODS

Twenty-three Dutch rabbits (weight range, 1.15-1.95 kg; mean, 1.5 kg) were studied after being killed with intravenous pentobarbitone (400 mg). The data from only 20 rabbits will be reported here, because 3 developed pneumothoraces during the study period and their results were thus invalid. Although filters had been used in the circuit of the rabbits who developed pneumothoraces, malfunction of the ventilator pressure relief valve was due to crystalline material (aminophylline) deposited from aerosols during a completely different set of exper- iments when no filters had been used. This highlighted the importance of using filters in nebulizer-ventilator circuits.

Immediately following death, each rabbit was venti- lated through a standard-length, shouldered size 3 mm endotracheal tube (Portex, UK) via tracheostomy, using a pressure-limited, time-cycled ventilator (Neovent 90, Vickers Medical, United Kingdom), smooth-walled rub- ber ventilator tubing, and a standard right-angled con- nector. Low resistance filters (Pall Filters Ltd, UK) were inserted into both inspiratory and expiratory lines (Fig. 1).

I

Experimental Protocol 9 9 T ~ m pertechnetate (400 MBq in 4 mL normal saline)

was loaded into the nebulizer. Each rabbit received 4 periods of mechanical ventilation each lasting 3 minutes, during which the gas flow through the nebulizer intro- duced radioaerosol into the circuit. Between periods of ventilation the lungs were held inflated at a pressure of 10 cm H,O by CPAP allowing a 3-minute, gamma-camera image to be taken. To this end, all rabbits were ventilated supine beneath the pin-hole collimator of a gamma camera (Toshiba) connected to a computer on which anterior images were recorded in a 128 X 128 matrix for

0 i i

T N- Nebuliser

F- Filter GC- Gamma camera P-Plethysmograph

R- Rabbit

M- Monitor \Computer

T- Transducer CR-Chart Recorder

Fig. 1. Experimental setup.

210 Cameron et al.

subsequent analysis. The sequence of repeated ventila- tion and imaging was performed without interruption, and in each rabbit the total period of study was less than 45 minutes. At the end of each experiment the rabbit lungs and trachea were excised for measurement of the total radioactivity deposited.

Four control rabbits, Group A, received all periods of ventilation at standard settings of the ventilator, namely at a pressure of 30/2 cm H20, a frequency of 30 breaths per minute, an I:E ratio of 1:1, and a gas flow rate of 8 L/min. The other 16 rabbits received 2 initial ventilation periods at standard settings, followed by 2 periods during which one of the ventilator settings was altered. These altered settings comprised:

Group B. I:E ratio, 0.5:1.5 (4 rabbits) Group C. Pressure 1612 cm H,O (4 rabbits) Group D. Frequency of ventilation 60 bpm (4 rab-

bits) Group E. Flow rate of L/min (4 rabbits)

In one rabbit the distribution of radioaerosol within the lung was compared with the image produced by a 81Krm gas ventilation scan to compare penetration of the aerosol with that of a gas.

To assess any change in lung compliance in the post-mortem state during the period of study, 8 of the rabbits were ventilated inside a specially constructed 2.3 L Perspex plethysmograph. The signal from a 25- cm-range pressure transducer (SE Labs, United King- dom) was amplified (SE Labs EMMA) and plotted on a UV chart recorder (SE Labs). Calibration of pressure change was achieved by injecting known volumes of air into the rabbit through the endotracheal tube, at the same frequency as the mechanical ventilator. The respiratory system compliance was measured at the beginning and end of each set of experiments, as the tidal volume divided by the airway opening pressure difference (i.e., peak pressure minus PEEP) during ventilation at standard settings, after ensuring that end-inspiratory and end- expiratory plateaux were achieved at these settings.

Image Analysis and Radioactivity Measurements

For each rabbit, a region of interest was chosen to include both lungs and used on all images to obtain the number of counts accumulated during each period of nebulization. The relative amount of radioactivity depos- ited in each period was calculated.

The nebulizers were counted with a 9 9 T ~ m standard before and after the experiment using a scintillation counter and the fraction of radioactivity that had been nebulized was calculated. The fraction of solution neb- ulized (without allowance for evaporation) was obtained by weighing.

Excised tissues were measured in a high pressure ionization chamber and readings compared with that of a standard of known relation to the initial nebulizer con- tent. Tissue radioactivity was calculated as a fraction both of the initial nebulizer content and of the amount of radioactivity nebulized. Using these results and the data from image analysis, the absolute amount of radioactivity deposited during each nebulization was obtained. Cor- rections for background and for radioactive decay were made in all calculations.

Statistical Analysis

The change in lung radioaerosol deposition between the first and second pair of nebulizations for each rabbit was calculated. For the rabbits where the second pair of nebulizations entailed altered settings, the mean changes for each of Groups B-E was compared with the mean change of group A using the Mann-Whitney test (Fig. 2).

RESULTS

Of the 20 rabbits, 4 were studied at standard settings throughout. 20% of the initial 400 MBq nebulizer charge was aerosolized in the 12-minute study period. The dose deposited in the lungs increased slightly with time, such

Group A Standard settmgs

throughout (4 rabbits)

0.151 Group 6 t I E ratio for nebs

3+4 (shaded) (4 r a b t s )

p=O 03

Group C 1 Rate tor nebs

3+4 (shadedX4 r a b t s )

p=o 88

Group D I Pressure for nebs rn

0.05 3+4 (shadedX4 rabbits)

p=0.03 0

Group E I Flow tor nebs 0.10

0.05 3+4 (shadedI(4 rabbits)

n p=O 04 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4

Nebulisations

Fig. 2. Lung deposition of radioaerosol per 3 minute nebuliza- tion per kg rabbit as a percentage of initial nebulizer charge for each rabbit during each experiment. Shaded columns represent results of changed ventilator settings.

Aerosol Delivery in Neonatal Ventilator Circuits 21 1

icant increase in dose delivered to the rabbit lungs during successive periods of ventilation at standard settings (Table 1) was attributed to this phenomenon. This does not affect the validity of the significant changes found when altered settings were used in the later nebulization periods, since in all 3 cases the observed change was a decrease in deposition.

The mass median diameter (and geometric SD) of the aerosol emerging from the endotracheal tube was 1.36(2.8) at a flow of 8 Limin and 4.67 (2.8) p at a flow of 6 L1min. These values are larger than those previously published for this nebulizer, due to the interposition of T-piece, connector tubing and endotracheal tube between nebulizer and particle sizer.

that there was a marginally higher deposition in the second pair of nebulizations than by the first pair (Table 1, Fig. 2). This was attributed to a concentrating effect in the nebulizer residue over time. Over the 12-minute period at standard settings, 2.8% of the radioaerosol produced (2.24 MBq) was deposited in the lungs. As a proportion of lung aerosol deposition, the amount depos- ited in the trachea averaged 10% (0.224 MBq).

All other rabbits had 2 ventilation periods at standard settings followed by 2 periods at altered settings (Table 1, Fig. 2). A significant fall in deposition was observed when I:E ratio was reduced from 1:l to 0.5:1.5. Plethys- mography revealed that this maneuver was not associated with a fall in tidal volume. Reducing the pressure from 3012 to 1612, which halved the tidal volume, also produced a significant fall in deposition. Reduction in the nebulizer flow rate had a smaller effect. Increasing the ventilator frequency did not have a significant effect on aerosol deposition in the lungs.

Each ventilation image showed a satisfactory wide- spread deposition of aerosol to the periphery of the lung fields. This impression was confirmed in one rabbit by comparison of the technetium images with a "Krm gas image (Fig. 3).

In the 8 rabbits studied by plethysmography, there was a relationship between inflation pressure (10, 20, 30 cm H,O) and tidal volume. The compliance of the respira- tory system did not alter during the experimental period, from the mean (SD) initial compliance of 0.95 (20.17) to the final compliance of 1 .O (20.18) mL X cmH,O- '.

During the study period, output from the nebulizers averaged 0.12 mL1min. However, there was a disparity between the change in weight and the change in radio- activity of the nebulizer during the study period signify- ing an increase in concentration. The small-but-signif-

DISCUSSION

Several authors have addressed the problems of deliv- ering therapeutic aerosols to neonates on ventilators. Ahrens et al.4 have shown in a lung model that most of a conventionally generated aerosol is likely to be depos- ited in the endotracheal tube and central airways before reaching the lung periphery.

Others7,* have studied aerosol deposition in the rabbit lung using a volume-cycled infant ventilator, and found that nebulizers producing small particle size improved lung deposition. No studies have examined the effect of changing ventilator settings of a pressure limited time- cycled circuit, the generally preferred mode for neonates, where most aerosol is lost in the continuous gas flow. Our interest was in the effect of ventilation parameters upon deposition in such a circuit.

Particles deposit by three main mechanisms, largely relating to their Larger particles (3-10 p) deposit mainly by impaction; therefore they tend to deposit in

TABLE 1-Mean (+ SD) Lung Deposition Per 3 Min Nebulization Per kg Rabbit as a Percentage of Initial Nebulizer Charae

Difference Significance of difference between between standard and

Standard Altered standard and altered settings in Ventilator settingsa settingsb altered settings relation to Group A

Group settings n (periods I + 2) (periods 3 + 4) (% standard) (Mann- Whitney test)

A Standard settings 4 0.09 (0.02) 0.10 (0.02) + 1 1

€3 I:E ratio 0.5:l.S 4 0.10 (0.03) 0.06 (0.03) - 40 P < 0.01 C pressure 1612 cm H20 4 0.09 (0.01) 0.05 (0.02) - 44 P < 0.01 D frequency 60/min 4 0.10 (0.02) 0.1 1 (0.02) + 10 N.S. E flow 6 L/min 4 0.08 (0.01) 0.06 (0.02) - 25 P < 0.05

aStandard settings were: pressure, 30/2 cm H2O; frequency, 30 breaths per min; I:E, 1:l; flow rate, 8 L/min.

throughout Altered settings:

ventilation parameter changed as in column 1 .

212 Cameron et al.

Fig. 3. Distribution of activity in a single rabbit produced by ventilation with krypton gas (left) and technitiurn aerosol (right). The artefact in the aerosol scan, relates to deposition at the endotracheal tube connection.

ventilator tubing and central airways where their velocity is greatest before reaching the more peripheral airways. Particles between 1 and 5 p. may deposit by sedimenta- tion, a gravitational effect greatest in regions of low flow. A large proportion of 0.5-1 p. particles are not deposited and reappear in the exhaled air. Particles smaller than 0.5 p tend to behave like gas in the lungs, diffusing freely. Prolonging residence time of aerosol or increasing tidal volume in the lungs will favour the deposition of particles.

In our study, lowering ventilator pressure caused a halving of tidal volume and a significant fall in deposi- tion. Lowering the proportion of inspiratory time (i.e., 1:E ratio) also brought about a significant fall in deposi- tion, presumably due to the decreased residence time of the aerosol in the lung, as tidal volume was unchanged. Increasing the ventilator frequency increases the minute ventilation and should increase lung deposition but with a fixed I:E ratio; the decreased residence time would tend to counteract this effect. As no change in deposition was observed in this maneuver it appears that these two effects cancelled each other out in our model. Decreasing the flow rate produced larger aerosol particles which would have had an increased tendency to deposit in ventilator tubing. This effect may be countered to an extent by the lower turbulence and impaction associated with lower flow. Hence a relatively small reduction in aerosol deposition was observed.

The use of freshly killed rabbits in our study enabled us to quantify deposition without problems of pulmonary

clearance of aerosol, and enabled us to have each rabbit serve as its own control by the use of sequential images. This method thereby circumvents some of the maneuvers necessary to quantify and compare deposition adopted by Flavin et al.’

Extrapolation from a dead rabbit to a live human neonate needs care and there are many differences between our model and a mechanically ventilated human baby. Changes in lung compliance during our period of study were minimal, but presumably would have been significant if the study period had been prolonged. Rabbit lungs differ anatomically from human lungs in both their lobar arrangement and branching patterns.’ Deposition patterns and quantities in a healthy rabbit lung may differ significantly from that in a diseased neonatal lung with areas of air trapping, atelectasis and the presence of secretions.” Our model is more of a lung preparation than an animal model, but despite the limitations men- tioned above, it does offer useful information.

The standard ventilator settings which we used pro- duced a tidal volume of about 18 mL/kg in the rabbits. This gave a lung deposition of 2.8% of generated aerosol. The altered pressure setting of 16/2 cmH,O gave a tidal volume of about 9 mWkg, closer to the tidal volume of a mechanically ventilated newborn baby. Lung deposi- tion was approximately halved. Thus a realistic figure for lung deposition in a normal neonate using a similar circuit would be about 1.4% of generated aerosol.

Our ventilator circuit was similar to most neonatal circuits apart from the absence of a humidifier. A

Aerosol Delivery in Neonatal Ventilator Circuits 213

Mallinkrodt provided the Ultravent nebulizers and Pro- fessor Peter Lavender provided the facilities for the experiments. We are grateful to Dr. Malcolm Kensett and Mrs. Diana Bates of the UK Medical Research Council Cyclotron Unit, Hammersmith Hospital, for the use of its high pressure ionization chamber.

nebulizer will provide some humidification of its own, and since the gas flow in our circuit was supplied solely through the nebulizer, for the purpose of our study, the absence of the humidifier was unimportant. Temperature fall in gas presented to infants due to aerosol generation might have an effect on lung function' ' and water content from any humidified gas mixing with aerosol would tend to increase particle size.

The risk of crystalline deposition of aerosol in venti- lator pressure relief valves was clearly shown in the 3 rabbits that suffered pneumothoraces. When the majority of the flow is supplied through the nebulizer, retrograde gas flow will occur through the inspiratory limb, hence it is of extreme importance to insert filters not just in the expiratory line,I2 but also in the inspiratory limb.13

Various authors have demonstrated the efficacy of bronchodilators in bronchopulmonary dysplasia, and their use is increasing in ventilated neonates. Other inhaled drugs may also be of use in this neonatal setting, including sodium cromoglycate and topical corticoster- oids.

The efficiency of nebulization may be increased by using a triggering mechanism which coordinates nebuli- zation and inspiration. l4 A comparison of various nebu- lizers in ventilator circuits would also be helpful to determine the relative merits of different models, which may vary considerably in terms of output, particle size and convenience of use.

In this study we have confirmed that it is possible to deliver a fine aerosol with good penetration to lung peripheries, in a rabbit lung model ventilated with a neonatal circuit. We have shown that different ventilator settings may affect total lung deposition of aerosol. We have also encountered a major potential problem of crystalline deposition of the solute drug in ventilator valves, which may be avoided by the appropriate use of filters.

Further studies in human neonates, are required to delineate factors determining deposition of aerosol in diseased lungs, either by use of radioaerosols, or by use of an aerosol of a non-toxic marker substance which could then be measured in blood or urine.

ACKNOWLEDGMENTS We thank Action Research for the Crippled Child and

Fisons Pharmaceuticals for generous financial support.

REFERENCES

I . Gomez-Del Rio M, Gerhardt T, Hehre D, Feller R, Banclari E. Effect of a beta-agonist nebulisation on lung function in neonates with increased pulmonary resistance. Pediatr Pulmonol. 1986; 2:287-291.

2. Motoyama EK, Fort MD, Klesh KW, Mutich RL, Guthrie RD. Early onset of airway reactivity in premature infants with bron- chopulmonary dysplasia. Am Rev Respir Dis. 1987; 13650-57.

3. Brain JD, Valberg PA. Deposition of aerosol in the respiratory tract. Am Rev Respir Dis 1979; 120:1325-1373.

4. Ahrens RC, Ries RA, Popendorf W, Wiese JA. The delivery of therapeutic aerosols through endotracheal tubes. Pediatr Pul- monol. 1986; 2:19-26.

5 . Dahlback M, Karlberg IB, Nerbrink 0, Prytz M, Wollmer P. Deposition of tracer aerosols in the rabbit respiratory tract. J Aerosol Sci. 1987;18:733-736.

6. Arnot RN, Burch WM, Orfanidon DG, et al. Distributions of an ultrafine 99Tcm aerosol and 8 lKrm gas in human lungs compared using a gamma camera. Clin Phys Physiol Meas. 1986; 7:345- 359.

7. Flavin M, MacDonald M, Dolovich M, Coates G, H O'Brodovich. Aerosol delivery to the rabbit lung with an infant ventilator. Pediatr Pulmonol, 1986;2:35-39.

8. Clarke SW, Pavia D. Aerosols and the Lung. London, England, Buttenvorths, 1984.

9. Schlesinger RB. Comparative deposition of inhaled aerosols in experimental animals and humans: A Review. J Toxicol Environ Health. 1985; 15:197-214.

10. Kim CS, Eldridge MA, Aerosol deposition in the airway model with excessive mucus secretions. J Appl Physiol. 1985; 59:176& 1772.

1 1 . Cameron D, Clay M, Silverman M. Evaluation of nebulizers for use in neonatal ventilator circuits. Crit Case Med. 1990; 18:86& 870.

12. Demers RR, Parker J , Frankel L, Smith SW. Administration of ribavirin to neonatal and paediatric patients during mechanical ventilation. Respir Care. 1986;31:1188-1195.

13. Cameron D, Clay MM, Silverman M. Ribavirin and acute bronchiolitis in infancy. BMJ. 1988; 297: 1472.

14. Dahlback M, Wollmer P, Drefeldt B, Jonson B. Controlled aerosol delivery during mechanical ventilation. J Aerosol Med. 1989; 2:339-347.