-
Archives ofDisease in Childhood 1992; 67: 25-30
Delivery of therapeutic aerosols to intubated babies
Jonathan Grigg, Shmuel Arnon, Tracey Jones, Andrew Clarke,
Michael Silverman
AbstractDelivery of drug aerosols to the lungs ofventilated
neonates by metered dose inhalerand spacer (Aerochamber) and
ultrasonicnebuliser (Pentasonic) was assessed usingsodium
cromoglycate.The mean proportion of a known intra-
tracheal dose ofsodium cromoglycate excretedin the urine of four
intubated infants was37-5%. After assuming that 38% of the
sodiumcromoglycate aerosol reaching the neonatallung will be
excreted in the urine, three puffs(15 mg) delivered by metered dose
inhaler andspacer resulted in a pulmonary dose of 258 [ig(1-7%,
n=7V. A dose of 20 mg (4 ml) sodiumcromoglycate ultrasonically
nebulised overfive minutes into the inspiratory limb of astandard
ventilator circuit produced a pul-monary dose of 257 [tg (1-3%,
n=7).Of two in vitro lung models assessed, a
combination of filter and neonatal test lungwas superior to a
multistage impactor inestimating the in vivo pulmonary
sodiumcromoglycate dose delivered by metered doseinhaler and spacer
(243 tg v 1740 ig).
Department ofPaediatrics andNeonatal Medicine,Royal
PostgraduateMedical School,Hammersmith Hospital,London W12
ONNJonathan GriggShmuel ArnonMichael SilvermanFisons
plc,Pharmaceutical Division,LoughbroughTracey JonesAndrew
ClarkeCorrespondence to:Dr Silverman.Accepted 15 September 1991
Therapeutic aerosols are often used in intubatedneonates despite
a paucity of information on thepulmonary dose or the most efficient
deliverysystem. A number of devices are available forthe production
of therapeutic aerosols.
Jet nebulisation is widely used and broncho-dilators
administered by this method have beenshown to improve lung function
in intubatedinfants with bronchopulmonary dysplasia.'There are,
however, problems with jet nebulis-ation including gas cooling,2
reduction inhumidity,3 and a requirement for an auxiliaryhigh
pressure gas supply. It is also very in-efficient, with 0-2% of the
drug dose in thenebuliser reservoir reaching the neonatal
lung.4
Table I Subject details
Subject Gestation Weight Diagnosts Sodium cromoglycateNo (weeks)
(g) delivery method
1 35 2100 Myopathy IT+Pentasonic2 35 4000 Persistent fetal
circulation IT+Pentasonic3 41 4200 Myopathy IT+Pentasonic4 31 1700
RDS IT+Pentasonic+Aerochamber5 24 800 RDS Pentasonic+Aerochamber6
30 1400 RDS Aerochamber7 38 3500 Gastroschisis Aerochamber8 26 1000
RDS Aerochamber9 25 1100 RDS Aerochamber10 27 500 RDS Aerochamber11
30 1400 Apnoea Pentasonic12 38 4700 RDS (diabetic mother)
Pentasonic13 36 1400 Congenital abnormalities IT14 26 1000 RDS
IT
RDS=respiratory distress syndrome; IT=intratracheal sodium
cromoglycate dose.
Ultrasonic nebulisation and metered doseaerosol delivered by
spacer are alternativedelivery methods which have little effect
onventilator gas humidity or temperature. Neitherhas been assessed
in intubated neonates.The Pentasonic ultrasonic nebuliser (De-
Vilbiss) nebulises drugs using 2-25 MHz pizo-electric crystal.
It weights 160 g and is portable.The Aerochamber with 15 mm
connection(Trudell Medical) is a modification of an infantspacer
device accepting standard metered doseinhalers (MDI) and is
designed to be insertedinto the inspiratory limb of an adult
ventilatorcircuit. This 11 x4 1 cm spacer has an approxi-mate
volume of 145 ml and allows aerosol to begenerated within the
centre of the ventilator gasflow. Both the Pentasonic and
Aerochamber canbe used in a neonatal ventilator circuit
afterminimal adaptation.
This study aimed to assess the efficiency ofaerosol delivery to
intubated infants lungs byusing sodium cromoglycate as a marker of
drugdelivery. Sodium cromoglycate (Intal, Fisonsplc) is non-toxic
and is excreted unchanged inthe bile and urine.5 We first estimated
thefraction of a known intratracheal dose of sodiumcromoglycate
excreted in the urine of ventilatedneonates over 24 hours. By
combining thisinformation with 24 hour urinary sodiumcromoglycate
excretion after an ultrasonic orMDI dose, the total dose deposited
in the lungwas estimated. Pulmonary sodium cromoglycatedeposition
by MDI in vivo was then comparedwith that estimated in vitro by
either a test lungwith filter or a multistage impactor.
MethodsIN VIVO ASSESSMENTInfants studied were receiving
intermittentpositive pressure ventilation via a pressurelimited,
time cycled ventilator (Sechrist) for arange of conditions. These
included respiratorydistress syndrome (n=8), myopathy
(n=2),persistent fetal circulation (n= 1), surgery (n= 1),and
multiple congenital abnormalities (n= 1)(table 1). All infants were
intubated by ashouldered endotracheal tube (Portex) and hadblood
urea and creatinine concentrations withinthe normal range.This
study was approved by the ethics
committee of the Royal Postgraduate MedicalSchool. Infants were
studied after writtenparental consent.
Intratracheal doseIsotonic sodium cromoglycate (1400 ,ug, 0-2
ml)
25
on July 10, 2021 by guest. Protected by copyright.
http://adc.bmj.com
/A
rch Dis C
hild: first published as 10.1136/adc.67.1_Spec_N
o.25 on 1 January 1992. Dow
nloaded from
http://adc.bmj.com/
-
Grigg, Arnon,3tones, Clarke, Silverman
was instilled through a size 5 French gaugesuction catheter
wedged in the right mainbronchus of intubated neonates. Urine
wascollected over a 24 hour period and assayed forsodium
cromoglycate by radioimmunoassay.6 Ifleakage from the collecting
bag occurred, the 24
Figure I Pentasonic nebuliser inserted into the inspiratory limb
ofa ventilator circuit.
Figure 2 Ventilation through an Aerochamber spacer.
E
D
F
G
A B
JH
inspiratory line
hour collection was discarded. Sodium cromo-glycate removed from
the lung by suction wasmeasured in the total pool of
endotrachealaspirates collected over 24 hours after
instillation.The actual intratracheal dose available to theinfant
was the difference between the intra-tracheal dose delivered and
the amount removedby endotracheal suction. The proportion of
theintratracheal dose excreted in the urine over a24 hour period
was then calculated.The urinary elimination half life of sodium
cromoglycate was calculated from a plot of themidtime point for
each urine collection (min)against the log excretion rate
([ig/min).7
Assessment of nebuliser and MDIThe Pentasonic ultrasonic
nebuliser was inserteddirectly into the inspiratory limb of the
constantflow ventilator circuit, 10 cm from the rightangled
endotracheal tube connector (fig 1).Before insertion the one way
inspiratory valvewas removed from the nebuliser chamber.Humidified
inspiratory gas entered the nebuliserthrough the upper part of the
nebuliser andexited through a side port. The chamber wasloaded with
20 mg sodium cromoglycate dis-solved in 4 ml of normal saline and
the nebuliserrun for five minutes. The 24 hour urinaryexcretion of
sodium cromoglycate was measuredby radioimmunoassay on a single
pooled sample.The Aerochamber was inserted directly onto
the endotracheal tube after disconnection fromthe ventilator
circuit. Infants were then venti-lated by a thumb occluded
'Nottingham puffer'set at a flow of 8 I/min and an appropriate
blowoff pressure and inserted into the end of thespacer (fig
2).Each infant received three puffs of 5 mg
sodium cromoglycate via MDI. The MDI wasactuated at end
expiration and each dose wasseparated by five manual breaths. After
remov-ing the spacer, infants were reconnected to theventilator
circuit and 24 hour pooled urinarysodium cromoglycate excretion
measured byradioimmunoassay.
IN VITRO MODELAerochamber assessmentThe neonatal lung was
modelled by a test lung(Draeger) inserted onto a 3-5 mm
shoulderedendotracheal tube. MDI aerosol passing throughthe
endotracheal tube was collected by a 0 25[im filter (Aerosol
Medical) inserted betweenthe tube tip and test lung (fig 3). The
dose ofsodium cromoglycate and method of adminis-tration were
similar to that used in vivo.Inspiratory pressure was set at 20 cm
H20giving a tidal volume to the test lung of 15 ml.After three MDI
doses (15 mg), sodium cromo-glycate was eluted from the filter with
20 mlwater and measured by radioimmunoassay.
Figure 3 In vitro models ofpulmonary sodium cromoglycate
deposition by metered doseinhaler and spacer. A= test lung, B
=filter, C=multistage impactor, D=3-5 mm endotrachealtube,
E=Aerochamber, F=sodium cromoglycate metered dose inhaler, G=thumb
occlusionvalve, H=pressure release valve.
Particle sizeThe mass median aerodynamic diameter(MMAD)
ofparticles generated by the ultrasonicnebuliser was measured by
laser diffraction.Ultrasonic aerosol was generated within a
con-
26
on July 10, 2021 by guest. Protected by copyright.
http://adc.bmj.com
/A
rch Dis C
hild: first published as 10.1136/adc.67.1_Spec_N
o.25 on 1 January 1992. Dow
nloaded from
http://adc.bmj.com/
-
Delivery oftherapeutic aerosols to intubated babies
tinuous 8 I/min airflow which passed through 10cm of ventilator
tubing and 3 5 mm Portexendotracheal tube. Particle size was
measuredby a laser particle sizer (2600 Malvern Instru-ments) 2 cm
from the endotracheal tube tip.The MMAD of particles generated by
MDI
was measured by multistage impactor. Aerosolgenerated by MDI was
carried through theAerochamber and endotracheal tube by a gasflow
of 8 1/min. This was then sucked into amultistage impactor
operating at 60 1/min andseparated into three size fractions (fig
3).Sodium cromoglycate deposited in each fractionafter 20 doses
(100 mg) was assayed by aspectrophotometer that had been
previouslycalibrated using sodium cromoglycate solutionsof known
concentrations. The total dosedeposited within the multistage
impactor and
100)
C1.°
0.- 0
co° .01c:
.001
0 0
0 0
.
0 200 400 600 800 100012001400Time (min)
Figure 4 Urinary excretion ofsodium cromoglycate fromsubjects 13
and 14.
Table 2 In vivo urinary excretion and estimated
pulmonarydeposition after sodium cromoglycate delivered
byAerochamber and metered dose inhaler
Subject 24 Hour urinary Estimated pulmonaryNo excretion of dose
sodium
sodium aomoglycatetcromoglycate* (jig)(mg)
4 69 1815 68 1796 109 2877 118 3108 104 2739 143 37610 75
197
Mean (SEM) 98 (11) 258 (28)
*After 3x5 mg doses of sodium cromoglycate.tAssuming urinary
excretion 38% of pulmonary dose.
MMAD of the aerosol was calculated from theamount eluted from
each size fraction. At theend of the impactor experiment, sodium
cromo-glycate deposited in the endotracheal tube andAerochamber was
also eluted and assayed.
ResultsIN VIVOFour infants (subjects 1-4, table 1) had
trachealaspirates saved after a 1400 tig intratrachealdose enabling
an estimation of the actual intra-tracheal dose available for
systemic absorption.Mean urinary excretion as a percentage of
theestimated intratracheal dose was 37-5% (range22-3-5917%). In two
infants (subjects 13 and14, table 1) several timed urine samples
afterthe intratracheal dose were available, althoughtheir
endotracheal aspirates were not saved.Their urinary excretion half
lives were 100 and310 minutes (fig 4).
Seven infants received a dose of sodiumcromoglycate via the MDI
and seven via theultrasonic nebuliser. Mean (SEM) 24 hoururinary
sodium cromoglycate excretion after a15 mg dose via Aerochamber and
MDI was 98(11) ,g (table 2). Mean 24 hour urinary
sodiumcromoglycate excretion after five minutes ofultrasonic
nebulisation with a nebuliser fill of 20mg was 98 (18) ,ug (table
3). Endotrachealsuction was not performed until at least fourhours
after aerosolised sodium cromoglycate.However, endotracheal
aspirates were not savedfor radioimmunoassay.
Assuming that 38% of an inhaled dose ofsodium cromoglycate is
excreted in the urine,the Aerochamber and MDI delivered 258 (28),ug
to the lung (1-7% of the 15 mg aerosoliseddose). The Pentasonic
delivered 257 (47) ,tg(1-3% of the initial 20 mg nebuliser
fill).
IN VITRO
Mean deposition of sodium cromoglycate onto afilter situated
between the endotracheal tubeand test lung after 3 x 5 mg puffs
delivered byMDI with Aerochamber was 243 pg (n=3).Total deposition
within the multistage impactorafter 20 puffs into the Aerochamber
was 11-6mg or 1740 pg per three MDI puffs (table 4). Ofthe total
dose aerosolised by the MDI, 81-4%was deposited within the spacer
and 7-8% in theendotracheal tube.
Table 3 In vivo urinary excretion and estimated pulmonary
deposition after ultrasonic nebulisation ofa sodium
cromoglycatesolution for 5 minutes
Subject Ventilator Ventilator Inspiratory 24 Hour EstimatedNo
gas flow rate time urinary excretion pulmonary dose
(I/min) (per min) (sec) sodium cromoglycate* sodium
cromoglycatet(mg) (mg)
1 8 30 0 5 104 2742 8 30 0-5 50 1323 8 30 0-5 170 4474 8 30 0 5
138 3635 12 75 0 4 47 123
11 8 20 0 3 63 15812 8 30 0 7 115 303
Mean (SEM) 98 (18) 257 (47)
*Nebuliser fill of 20 mg sodium cromoglycate in 4 ml.tAssuming
urinary excretion 38% of pulmonary dose.
27
on July 10, 2021 by guest. Protected by copyright.
http://adc.bmj.com
/A
rch Dis C
hild: first published as 10.1136/adc.67.1_Spec_N
o.25 on 1 January 1992. Dow
nloaded from
http://adc.bmj.com/
-
Grigg, Arnon,Jones, Clarke, Silverman
Table 4 Comparison of in vitro and in vivo aerosol delivery of
sodium cromoglycate
Initial nebuliser Mean pulmonary MMADfill or dose dose in vitro
or (Psm)generated by in vivo (ig)MDI (mg) (% of initial dose)
Aerochamber+MDI in vitro;multistage impactor* 15 1740 (11-6)
90
Aerochamber+MDI in vitro;filter+test lung (n=3) 15 243
(1-62)
Aerochamber+MDI in vivo (n=7) 15 258t (1-72)Pentasonic in vivo
(n=7) 20 257t (1-28) 3-4t*Calculated from 100 mg initial
dose.tEstimated assuming urinary excretion 38% of pulmonary
dose.1Measured in vitro using Malvern laser particle sizer.
The MMAD of particles generated by MDI,after passing through the
Aerochamber andendotracheal tube, was 9 0 ,um. MMAD
ofultrasonically nebulised sodium cromoglycateafter passing through
an endotracheal tube was2-8 ttm (table 4).
DiscussionThe Pentasonic nebuliser and Aerochamberwith MDI
delivered a detectable dose of sodiumcromoglycate to ventilated
neonates. A signifi-cant proportion of the sodium
cromoglycateaerosol generated by both delivery systems wasin
droplets small enough to penetrate the distalairways. The pulmonary
dose of sodium cromo-glycate aerosolised by MDI, estimated from
theurinary excretion of a known intratracheal dose,was similar to
that deposited on a filter modelusing a neonatal test lung.
However, the dosedeposited in a multistage impactor
significantlyoverestimated the in vivo dose.
Estimation of the pulmonary dose in vivodepends on the
proportion excreted in theurine. Most published data is from adult
sub-jects. Sodium cromoglycate is not metabolisedin the lung or
liver and is excreted unchanged inthe bile and urine.5 Some 70-90%
of a 1 mgdose of sodium cromoglycate instilled into theadult lung
by fibreoptic bronchoscope will beabsorbed into the circulation and
of this,3346% is excreted in the urine.8 9 The plasmahalf life of
either a bronchoscopically instilled orinhaled dose is 64-165
minutes after an initialrapid half life of 1-9 minutes, with
lungabsorption being the rate limiting step.8 Thepulmonary dose
calculated from urinary excre-tion is similar to that obtained
using plasmadata.'0 The wide variability in the fraction
ofintratracheal sodium cromoglycate excreted inthe urine of
ventilated neonates in this study isnot surprising. Pulmonary
atelectasis, mucus,and epithelial damage may have
significantlyaltered sodium cromoglycate bioavailabilitybetween
infants. When combined with theclinical problems of collecting
endotrachealaspirates and urine samples, the fractionalurinary
excretion must be regarded as only arough approximation. However,
if the urinaryexcretion half life is less than seven hours,
acollection period of 24 hours should havecontained most of the
dose excreted in theurine.
Combined in vitro and in vivo assessment ofaerosolised drug
delivery in ventilated neonates
has previously been reported. Watterberg et al,using a
multistage impinger model, detected19% of the sodium cromoglycate
dose placed inthe reservoir of a jet nebuliser emerging from
anendotracheal tube. However, when the samenebuliser and loading
dose (20 mg) were used invivo, less than 0-1% (20 ,tg) was excreted
in theurine of intubated infants.4 A similar over-estimate of the
in vivo pulmonary dose wasobtained by the multistage impactor in
ourstudy and may result from incomplete model-ling of aerosol
behaviour. In a conventionalpressure limited, time cycled
ventilator, inspir-ation is a result of the build up of
pressurewithin the circuit from obstructing gas exitwhile fresh gas
in-flow continues. The require-ment for a continuous gas flow
during the restof the cycle, means that a significant proportionofa
constantly nebulised drug will be unavailableto the infant. If all
the flow through the circuitis directed into a multistage impactor,
drugwasted in the expiratory phase in vivo will becollected. There
is a similar reason for theoverestimation by multistage impactor of
thedose delivered by Aerochamber with MDI.Here, the source of error
is that the multistageimpactor cannot model inspiratory and
expira-tory flow as aerosol needs to be continuouslydrawn through
the impactor system. The com-bination of filter and neonatal test
lung doesmodel inspiration and expiration and the agree-ment
between the in vivo pulmonary dose (258,tg) and sodium cromoglycate
deposited on thefilter (243 ,ug) after a 15 mg MDI dose
reflectsthis improved modelling.The filter model also has the
advantage that,
where aerosol is generated within the continuousgas flow,
wastage of drug during expiration ismodelled. Filter/test lung
models seem topredict accurately jet nebulised aerosol deliveryinto
neonatal ventilator circuits in vivo. Usingaminophylline as a drug
marker the Ultraventjet nebuliser delivers 0-22% of the
reservoirdose to a test lung." A figure of 0 2% has beenreported
after jet nebulisation in vivo.4 Despitethe errors in estimating
the pulmonary dose ofsodium cromoglycate aerosol, urinary
excretionof sodium cromoglycate by the infants in thisstudy
reflects the absolute minimum amount ofdrug delivered to the lung.
About 0-7% of theMDI dose and 0-5% of the ultrasonicallynebulised
dose was excreted in the urine over 24hours. Although we did not
measure it, thedeposition of aerosol into the lungs by
ultrasonicnebuliser is likely to be dependent on thepattern of
mechanical ventilation, as it is for ajet nebuliser inserted into a
ventilator circuit. 12The Pentasonic has not previously been
assessed in vivo although the Aerochamber hasbeen used to
deliver bronchodilators to intubatedadults receiving volume cycled
ventilation. AnAerochamber and MDI inserted into theinspiratory
limb of a ventilator circuit willdeliver 5 7% of the original dose
to the adultlung.'3 In adult ventilator circuits, there is
nocontinuous flow and most of the aerosoliseddrug is delivered to
the patient's endotrachealtube. By placing the Aerochamber
directlyonto the neonate's endotracheal tube the problemof
continuous flow is overcome and delivery
28
on July 10, 2021 by guest. Protected by copyright.
http://adc.bmj.com
/A
rch Dis C
hild: first published as 10.1136/adc.67.1_Spec_N
o.25 on 1 January 1992. Dow
nloaded from
http://adc.bmj.com/
-
Delivery oftherapeutic aerosols to intubated babies 29
partially mimics that of the volume cycledventilator. It may
also be possible to enhanceultrasonic drug deposition by placing
thenebuliser directly onto the endotracheal tubeand ventilating
through the chamber.The therapeutic effectiveness of an aerosol
depends not only on the quantity of drugreaching the lung but
also on particle size, bestdescribed by the MMAD. MMAD is
thediameter above and below which 50% of themass of the particles
is found. Particles less than5 [tm are more likely to deposit in
the distalairways whereas larger particles impact on thelarger
diameterairways. 14Although theoptimumaerosol size for neonates is
unknown, theoreticalmodels suggest a similar patterm of
depositionto adults.'5 In the past, ultrasonic nebulisershave been
bulky and have produced a highlyvariable output containing
significant numbersof large droplets. 6 In this study the
combinationof both the Pentasonic and Aerochamber withan
endotracheal tube produced high qualityaerosol. The size
distribution of aerosolgenerated by the Pentasonic was similar to
thatproduced by jet nebulisation. "l The MMAD ofparticles generated
within the Aerochamber waslarger (9 ,tm), as measured by the
imperfectmultidose impactor model, but a significantproportion of
the aerosol was below 5 [lm.The Pentasonic and the Aerochamber
were
clinically easy to use. The Aerochamber fitteddirectly onto the
standard Portex endotrachealtube connector without adaption and
there wereno problems in ventilating infants through thespacer. The
Pentasonic could be inserted intothe inspiratory line with minimal
adaption.There was no change in ventilator pressuresduring
ultrasonic nebulisation and no drop ingas temperature, as measured
by a thermistersited in the endotracheal tube connector. It
wasimportant to remove the one way valve from thePentasonic to
allow any build up of pressure inthe circuit to be vented through
the emergencyinspiratory blow off valve. The 145 ml Aero-chamber
temporarily increased the ventilatordead space, and for this
reason, we preferred toinsert the device briefly, and not to try
tomaintain mechanical ventilation with the spacerin place. However,
spacers are important whenusing a MDI in ventilator circuits.
Particles areejected from the MDI canister with high velocityand
mass.'7 18 As the propellants evaporateparticles decrease in size
toward the respirablerange (
-
30 Grigg, Arnon,Jones, Clarke, Silverman
of DeVilbiss jet and ultrasonic nebulizers. Chest
1987;92:991-4.
17 Dhand R, Malik SK, Balakrishnan M, Verma SR. Highspeed
photographic analysis of aerosols produced bymetered dose inhalers.
J Pharn Pharmacol 1988;40:429-30.
18 Clarke SW, Newman SP. Differences between pressurizedaerosol
and stable dust particles. Chest 1981;80:907-17.
19 Newman SP, Moren F, Pavia D, Little F, Clarke SW.
Depositon of pressurized suspension aerosols inhaledthrough
extension devices. Am Rev Respir Dis 1981;124:317-20.
20 Silverman M. Aerosol therapy in the newborn. Arch Dis
Child1990;65:906-8.
21 Salmon B, Wilson NM, Silverman M. How much aerosolreaches the
lungs of wheezy infants and toddlers? Arch DisChild
1990;65:401-3.
on July 10, 2021 by guest. Protected by copyright.
http://adc.bmj.com
/A
rch Dis C
hild: first published as 10.1136/adc.67.1_Spec_N
o.25 on 1 January 1992. Dow
nloaded from
http://adc.bmj.com/