341
Dilution of nebulised aerosols by air entrainment
in children
Guidelines for use of aerosolised drugs in children
are inconsistent. In a study of 14 infants, 22
children, and 4 adults inspired nebulised aerosols
were diluted more for large than for small subjects,
because of air entrainment which occurred when
inspiratory flow exceeded nebuliser flow. Infants
under 6 months of age did not entrain air and
would receive undiluted aerosols. All other
subjects entrained air, which caused up to a 5-fold
dilution in inspired aerosol concentration as
subject size increased. In subjects who entrained
air, the ratio of inspired nebuliser output versus
total nebuliser output was relatively constant, and
was related to the respiratory pattern. For a given
nebuliser solution concentration, infants who do
not entrain will inspire more concentrated
aerosols than older children. Once entrainment
occurs, the mass of drug inspired is largely
independent of size. Regimens for nebulised drug
delivery in children may require revision.
Introduction
Nebulised aerosols are widely used in the treatment and
investigation of respiratory disorders in children,12 but there
is no consistent approach to calculation of the dose of agents
delivered by aerosol. Recommendations are made for some
drugs to be delivered with the same concentration of drug in
the nebuliser solution for children of all ages,’ and for others
to be administered according to a weight-corrected
concentration.1.2
Another component of nebulised drug delivery, hitherto
unrecognised, relates to air entrainment-which occurs
when inspiratory flow rate exceeds nebuliser output. Tidal
inspiratory flow rates in most children are likely to exceed
nebuliser flow rates, so that the larger the child, the greater
the tendency for aerosols to be diluted in terms of mass of
aerosolised agent per unit volume of inspired gas. We
investigated the effect of body size on both the dilution of
aerosols by air entrainment and on the amount of aerosol
inspired. Such infprmation could lead to more logical
aerosol drug dosage regimens for children.
Subjects and methods
40 subjects were studied: 14 infants aged 1-12 months, 22 children
aged 3-16 years, and 4 adults aged 20-23 years. The infants, none of
whom had any respiratory symptoms or signs, were participants
(with full informed parental consent) in a concurrent, longitudinal
study of airway responsiveness in early childhood. The trial
protocol had been approved by hospital and university ethics
committees. The children comprised 11with no history of or
current respiratory disease, 6 with asthma, and 5 with cystic fibrosis;
none were ill at the time of investigation. All adults had no history of
or current respiratory problems. Age, height, weight, experience
with useof nebulisers, and time since last medication, where
appropriate, were determined for all subjects.
Infants were induced to sleep with 80 mg/kg chloral hydrate.
Tidal breathing was then recorded through a ’Fleisch l’
pneumotachograph (P. K. Morgan, Chatham, UK) with an output
linear to 1 1/s, attached to a sealed face maskA bias airflow of 6
1/min was maintained into the top of the mask during data
collection, to negate the effect of breathing through the dead space
of the pneumotachograph. An aerosol was not administered because
it was thought unlikely that sleeping infants would alter their
breathing pattern in response to a saline aerosol-a view confirmed
subsequently (unpublished observations). All other subjects
breathed through a standard mouthpiece for 6 min, and inspired
nebulised isotonic saline generated by a nebuliser (Airlife,
Montclair, California, USA) operated from a compressed air supply
that delivered 6 1/min. Flow was measured with a heated Fleisch 1
pneumotachograph attached to the mouthpiece and nebuliser via a
T-piece connector.
Fig 1-Diagrammatic representation of entrainment with
typical flow-time traces from an adult. a young child, and an
infant under 6 months of age.
Nebuliser flow subtracted so that zero flow relates to subject not
pneumotachograph Arrow indicates nebuliser flow rate of 6 I/mln
T=mspratory time for single breath, Tot=total time for single
respiratory cycle.
g =volume of nebuliser output inspired per breath (V n) ... (which
includes all )=totat nebuliser output during single respiratory cycle
(T’Olx nebuliser flow rate)
Tidal volume (VJ=area bounded by inspiratory flow loop and zero
flow axis
Flow signals obtained during tidal breathing were amplified and
relayed via an analogue-digital converter to a computer software
package for respiratory data analysis (ANADAT, from Dr J. Bates,
McGill University, Montreal, Canada). Bias or nebuliser flow was
electronically subtracted from the flow signal before digital
integration to determine tidal volume (Vt). The inspiratory flow
under 100 ml/s (equivalent to a nebuliser output of 6 1/min) was
integrated to obtain the volume of nebuliser output inspired per
breath (Vn) (fig 1), and mean inspiratory time (T,) and duration of a
single respiratory cycle (TlOt) were determined over a mean of 9
consecutive respiratory cycles (range 5-19).
ADDRESS Department of Respiratory Medicine, Princess
Margaret Hospital for Children, Perth, WA 6001, Australia
(G G Collis, BSc, C H Cole, MB, P N Le Souef, MD) Correspondence
to Dr P N Le Souef, University Department of Paediatrics, Princess
Margaret Hospital for Children, GPO Box D184, Perth, WA 6001,
Australia
 342

Fig 2-Dilution of aerosol by air entrainment (V n/Vt) against

subject height.
Fig 3-Ratio of nebuliser output inspired to total nebuliser
V n/V! = concentration of inspired aerosol with respect to concentration
of aerosol produced by nebuliser flow of 6 I/min output (Rn1) against height at6 I/min nebuliser flow rate.
Curve shown represents best fit based on polynomial regression
Symbols and curve as for fig 2
analysis. 0=over 3 years old, X=6-12 months (with entrainment),
. under 6 months (no entrainment).
=

To measure the effects of air entrainment on the amount of

aerosol inspired; we calculated two novel ratios. The first, V/V to
reflects the degree of aerosol dilution due to entrainment. When
expressed as a percentage, a value of 100% means that air
entrainment does not occur and that the entire inspiratory volume is
comprised of nebulised aerosol; a progressive fall in the Vn/Vt ratio
would result from a progressive increase in inspiratory flow rate
above nebuliser output (6 l/min) with a proportional increase in
entrainment and dilution of the aerosol. (For example, a ratio of
20% would mean that nebulised aerosol constitutes 20%, and
entrained air 80%, of the tidal volume).
The second ratio, 1’ represents the proportion of the nebulised
drug inspired, and is equivalent to the nebuliser output inspired
divided by the total nebuliser output (ie, Rni =if the entire
nebuliser output is inspired). For a single respiratory cycle,
Fig 4-Relation between breathing pattern (as T1/Ttot) and at
Pn.= Vn 6 I/min nebuliser flow rate.

Symbols as for fig 2, - - -represents line of identity

Trot x nebuliser flow rate.
Linear regression analysis of V/V against anthropometric data
was initially applied only to children and adults; infants who did not
Rni was closely related to Ti/Ttot in children and adults, all
entrain air were not included in these analyses because V n/Vt would of whom entrained air (r 091, p < 0’001) and in all subjects
=

always be 100% and would not vary with height, weight, or age.
However, data from all subjects were used to assess relations
between both Vn/V, and R, against anthropometric parameters.
Polynomial curves were plotted with the best-fit polynomial
regression equation and the significance of this relation assessed by
Spearman rank correlation analysis. Student’s t test was used to
evaluate the significance of differences in R, between subjects who
entrained air and those who did not.
Results
At a nebuliser flow rate of 6 llmin, height correlated with
Vn/Vt and hence the degree of entrainment for the combined
data from the 22 children and 4 adults (r 0-52, p < 001),
= -
older than 6 months, all of whom entrained air (r = 0-86,
p < 0’001). However, the relation in younger infants who did
not entrain air was poor (fig 4). Rni shows increasing
discordance with T;/Ttot as entrainment falls: values for
children and adults are on or immediately below the line of
identity, fall below this line for infants who entrain air, and
fall well below the line for younger infants who did not
entrain air.
No significant differences were found in Vn/Vt or Rm
between children with asthma, cystic fibrosis, or no
respiratory disease.

but age (r= -0.36, p=0’07) and weight (r= -0’37,

p=007) did not. Height was therefore used to examine
relations between subject size and Vn/Vt and Rni’ The 5
infants who were below 6 months of age did not entrain air
and received undiluted aerosols (Vn/Vt= 100%). All other
subjects inhaled aerosols that were diluted due to air
entrainment (fig 2), with a minimum Vn/Vt of 19-2%. The
Spearman rank correlation coefficient for the relation
between height and Vn/Vt for the whole group was -0,81
(p < 0-001).
In older infants and other subjects, all of whom entrained
air, the calculated mass of aerosol inspired at a nebuliser flow
rate of 61/min was independent of size with a mean (SD) Rni
of 0-37 (0-04) (fig 3). By comparison, the infants who did not
entrain air had a mean (SD) Rni of 0-22 (0-04) (p < 0-001).
Discussion
Dosage regimens for therapeutic and diagnostic drug
delivery by aerosol to children of different ages seem to
overlook basic aspects of inhalation mechanics related to air
entrainment, which affects both aerosol concentration and
the quantity of aerosol inspired. Our data indicate that the
relation between the concentration of an agent in a nebuliser
solution (Cs) and concentration of that agent per unit
volume of inspired gas (Ci) will not be constant. Ci can be
five times greater in a young infant than an older child, for a
given Cs. We also found that the quantity of a nebulised
agent inhaled at a nebuliser flow rate of 6 1/min may be
independent of size after early infancy. This inconsistent
(but predictable) relation between Cs and Q in children
means that delivered doses of aerosolised agents for children

of different sizes should not be based on the mistaken
assumption that small children take small breaths and thus
receive appropriately small doses. For dose to be regulated
by breath size, with a single direct relation between Cs and
C1, would require a nebuliser output above the tidal
inspiratory flow rate of the largest children; this flow rate
may be 500 ml/s, which would require an impractically high
nebuliser output of 30 1/min or more.
The variable relation between Cs and Ci also calls into
question studies of airway responsiveness in children in
which tidal breathing is used to deliver the aerosolised
agent.3-5 Indeed, we recently speculated that infants may
have bronchial hyperresponsiveness compared with older
children,3 on the basis that a low concentration of histamine
in the nebuliser solution could produce an airway response.
Yet if corrected for the increased inspired aerosol
concentration for a given Cs that was almost certainly
present in these infants, their airway responsiveness is
unlikely to be different from older children. Clearly, such
studies should assess response to an agonist in terms of mass
of agent per unit volume of inspired gas, rather than the
nebuliser solution drug concentration.
Our finding that the quantity of aerosol inspired is largely
independent of size after 6 months of age indicates that a
single Cs will be inadequate to deliver appropriate doses of
aerosolised drugs to children of all ages. Quite small children
would receive the same dose as large adults, so weight-
corrected drug delivery could vary by over 1000%. This
phenomenon may be the reason for recent observations of
atropine poisoning in young children who receive repeated
nebulised doses of ipratropium bromide (D. M. Cooper,
personal communication). This problem could be solved by
the availability of nebuliser solutions that contain different
drug concentrations for children of different sizes. However,
further research will be needed to determine appropriate
concentrations and size corrections-and, indeed, a basis
upon which to make size corrections (eg, body weight,
height, absolute volume, airway surface area, and so on).
Most aerosolised agents that are given to children on a
size-related basis use the child’s weight as a guide.1.2
Although the origins of this practice are not clear, this
approach appears reasonable until more information on
other options is available. Few trials have evaluated
alternative methods,6.7 but at least one centre has recently
replaced a weight-corrected regimen by one with a single
nebuliser solution concentration.8
Infants under 6 months of age do not entrain air and thus
will tend to size-correct their own doses. Without
entrainment, the smaller the infant, the smaller will be the
minute volume and the dose of drug delivered. Thus the
lowest Cs administered could reasonably be the size-
corrected Cs for an infant whose respiratory flow rate
matches the nebuliser flow rate. (Smaller infants may have a
slightly higher minute volume because of their relatively
higher metabolic rate, but this effect should be small.)
Our finding that the dose received is directly related to the
inspiratory pattern (specifically Ti/Ttot ) is not surprising:
most of the nebuliser output is inspired during inspiration,
and none during expiration, so Rn, and T1/Ttot will be nearly
the same in anyone who entrains air, regardless of size.
However, as peak inspiratory flow rates decrease, as
observed in younger children, more of the early and late part
of inspiration is at a flow rate below the nebuliser output (fig
1), so that Rni falls for a given value of TdTtot (fig 4). The
relation between Rni and Ti/Ttot disappears once
343
entrainment This relation may have clinical
ceases.
implications: for example, if Ti/Ttot is reduced by airway
obstruction, a patient would receive a lower dose of
aerosolised agent.
Calculation of doses of aerosolised drugs given to children
should no longer be based on the simple assumption that
nebuliser output determines the dose delivered over a given
inhalation time.9 1’ Furthermore, the sole use of a single
concentration or a weight-corrected concentration of an
active agent in the nebuliser solution will not suffice for all
children. We therefore recommend that such agents be
administered in a weight-corrected concentration for
children over 6 months of age (for a nebuliser flow of around
6 1/min), with the weight-corrected dose appropriate for a
6-month-old used for younger infants; that dose-response
studies of nebulised agents in children should account for
either the inspired aerosol concentration or the inspired
dose; and that studies of childhood airway responsiveness in
which nebulised agonists are delivered during tidal
breathing should make inter-subject comparisons according
to inspired concentration of agonist rather than nebuliser
solution concentration. And even should this more rigorous
approach be adopted, it only takes account of the dose of
aerosolised agents delivered to the nose and mouth: other
important variables (such as aerosol droplet size, aerosol
deposition, airway geometry, airway area, and site of
activity), all of which are likely to influence response to an
aerosolised agent,12 may need to be taken into account.
We thank Dr Igor Gonda and Mr Paul Phipps for discussions which helped
lead to this work; Prof Louis Landau for review of the manuscript; Miss
Debra Turner for technical help; Mr Ralph Baker for medical illustrations;
and Dr Steven Stick and Miss Sally Young for assistance with infant data
collection.
This work was supported by a Telethon research grant.
REFERENCES
1. Canny GJ, Levison H. Aerosols&mdash;therapeutic use and delivery in
childhood asthma. Ann Allergy 1988; 60: 11-19.
2. Phelan PD, Landau LI, Olinsky A. Respiratory illness in children.
Oxford: Blackwell, 1982.
3. Le Souef PN, Geelhoed GC, Turner DJ, Morgan SEG, Landau LI.
Response of normal infants to histamine. Am Rev Respir Dis 1989; 139:
62-66.
4. Tepper RS. Airway reactivity in infants: a positive response to
methacholine and metaproterenol. J Appl Physiol 1987; 62: 1155-59.
5. Cockcroft DW, Killian DN, Mellon JJA, Hargreave FE. Bronchial
reactivity to inhaled histamine: a method and clinical survey. Clin
Allergy 1977; 7: 235-43.
6. Ritchie D, Erben A, McLennan L, Landau LI, Phelan PD. Dose of
terbutaline respirator solution in children with asthma. NZ Med J
1979; 89: 332-34.
7. Perera BJC. A dosage schedule for salbutamol nebulisation in childhood
asthma. Ceylon Med J 1987; 32: 105-08.
8. Anonymous. Respiratory conditions. In: Royal Children’s Hospital,
Melbourne, Paediatric Handbook. Melbourne, Australia: Royal
Children’s Hospital, 1989: 115-22.
9. Ryan G, Dolovich MB, Obminski G, et al. Standardization of inhalation
provocation tests: influence of nebuliser output, particle size, and
method of inhalation. J Allergy Clin Immunol 1981; 67: 156-61.
10. Cockcroft DW, Berscheid BA. Standardization of inhalation provocation
tests: dose vs concentration of histamine. Chest 1982; 82: 572-75.
11. Hargreave FE, Sterk P, Adelroth EC, Ramsdale EH. Airway
responsiveness to histamine and methacholine: advances in
measurement and interpretation. Respiration 1986; 50 (suppl 2): 72-76.
12. Brain JD, Valberg PA. State of the art. Deposition of aerosol in the
respiratory tract. Am Rev Respir Dis 1979; 120: 1325-73.