Thorax 1993;48:154-159

Pulmonary deposition of a nebulised aerosolduring mechanical ventilation

Simon H L Thomas, Michael J O'Doherty, Helen M Fidler, Christopher J Page,David F Treacher, Thomas 0 Nunan

AbstractBackground There is increasing use oftherapeutic aerosols in patients under-going mechanical ventilation. Few stud-ies have measured aerosol delivery to thelungs under these conditions withadequate experimental methods. Hencethis study was performed to measurepulmonary aerosol deposition and todetermine the reproducibility of themethod ofmeasurement during mechan-ical ventilation.Methods Nine male patients were stud-ied during mechanical ventilation afteropen heart surgery and two experimentswere performed in each to determine thereproducibility of the method. A solutionof technetium-99m labelled humanserum albumin (Q""Tc HSA (50 ,ug);activity in experiment 1, 74 MBq; inexperiment 2, 185 MBq) in 3 ml salinewas administered with a Siemens Servo945 nebuliser system (high setting) and aSystem 22 Acorn nebuliser unit.Pulmonary deposition was quantified bymeans of a gamma camera and correc-tions derived from lung phantom studies.Results Pulmonary aerosol depositionwas completed in 22 (SD 4) minutes.Total pulmonary deposition (% nebuliserdose (SD)) was 2*2 (0.8)% with 1*5% and0 7% depositing in the right and leftlungs respectively; 0.90%o of the nebuliseractivity was detected in the endotrachealtube or trachea and 51% was retainedwithin the nebuliser unit. Considerablevariability between subjects was foundfor total deposition (coefficient of varia-tion (CV) 46%), but within subject repro-ducibility was good (CV 15%).Conclusions Administration of aerosolin this way is inefficient and furtherresearch is needed to find more effectivealternatives in patients who requiremechanical respiratory support. Thismethod of measurement seems suitablefor the assessment of new methods ofaerosol delivery in these patients.

(Thorax 1993;48:154-159)

Nebulised drug aerosols are administered tothe lungs via the airways with the aim of pro-ducing therapeutic benefit while minimisingsystemic adverse effects. Use of this methodis increasing in intensive therapy units andcurrently about a third of patients undergoing

mechanical ventilation receive drugs throughthis route.' The use of nebulised bronchodila-tors is well established; more recently othertypes of drug, particularly antimicrobialagents, have been administered in this way.-4Although deposition of nebulised aerosols hasbeen studied in detail during normal breath-ing, little is known about the amount ofaerosol that reaches the lungs during mechan-ical ventilation.

In vitro studies and animal studies per-formed during mechanical ventilation5 6 havesuggested that considerable deposition ofaerosol occurs in the endotracheal tube withlittle of the drug dose reaching the lungs. Thelimited human data available suggest thatonly 1-2%-2-9% of the nebuliser dose reachesthe lungs.'-7 These studies have used 99mTcdiethylenetriaminepenta-acetate and 99mTcsulphur colloid; however, both these tracersseemed unsuitable for measuring pulmonaryaerosol deposition.8 '0 Colloidal 99mTc humanserum albumin (99mTc HSA) appeared moreappropriate9-12 and the validity and repro-ducibility of measurements made with thismarker during spontaneous respiration havebeen reported.'3The aims of our study were to apply the

99mTc HSA marker technique to the measure-ment of pulmonary deposition of aerosol dur-ing mechanical ventilation in a group ofpatients with apparently healthy lungs afteropen heart surgery. The study had two mainarms: firstly, to measure the deposition effi-ciency of an inspiratory phase activated jetnebuliser system with equipment that has notbeen: studied before though it is commonlyused in intensive care units; secondly, todetermine the immediate reproducibility ofthese measurements of aerosol deposition.

MethodsThe study was approved by the ethics com-mittee of West Lambeth Health District.Patients about to undergo elective open heartsurgery were invited to participate and allthose taking part gave informed written con-sent. Those patients with pre-existing respira-tory symptoms or disease were excluded asthe intention of the study was to measureaerosol deposition in patients with normallungs. Lung function (forced expiratory vol-ume in one second (FEVI), forced vitalcapacity (FVC), peak expiratory flow (PEF))was measured before surgery. Experimentalstudies were performed within a few hours ofthe open heart surgery in those patients who

Division ofPharmacologicalSciences andToxicology, UnitedMedical and DentalSchool (St Thomas'sCampus)S H L ThomasDepartment ofNuclear MedicineM J O'DohertyC J PageT 0 NunanIntensive TherapyUnit, St Thomas'sHospital, LondonSE1 7EHH M FidlerD F TreacherReprint requests to:Dr S H L Thomas, WolfsonDepartment of ClinicalPharmacology, Newcastleupon Tyne NE 1 7RUReceived 6 April 1992Returned to authors11 June 1992Revised version received15 July 1992

Accepted for publication22 July 1992

154

on Septem

ber 28, 2020 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.48.2.154 on 1 February 1993. D

ownloaded from

Pulmonary deposition of a nebulised aerosol during mechanical ventilation

still required mechanical ventilatiortime. The decision on the duration otion was made on clinical grounds b3staff who were not involved in the sonly a few of those who gave informten consent were eventually studied.

All the patients were intubatedventilated with a Siemens Servmechanical ventilator in volumemode. The position of the endotrac]was verified from a chest x rVentilator settings were made on th4the patient's clinical requirements .not altered during the study. Artergas analysis was performed in allimmediately before the study.

Pulmonary aerosol deposition Msured by a modification of a tdesigned for spontaneously breathingand described in detail elsewhere."3-administration of aerosol a xenon- 13a breath hold image was obtainedthe lung edges and regions. T1achieved by injection of 10 ml (200the gas into a port on the catheteconnection at the proximal end of ttracheal tube (fig 1) immediately beftion of the lungs. Ventilation was stc10 seconds as soon as the lungs wand an anterior ventilation imacquired over this period with a gamera linked to a computer. Ventilathen resumed. Expired 133Xe was coa Douglas bag from the expirationthe ventilator until washout was comr

Administration of aerosols producnebulisers to patients undergoingcycled mechanical ventilation requthe jet nebuliser is driven by gas duinflation only (inspiratory phaseactivation). The volume of the gasthis must be taken into account N

tidal volume is set. In these studiesachieved with a Siemens 945 Servodriver (high flow setting) linked to22 Acorn nebuliser (Medic-Aid Lthat was connected to the catheter mendotracheal tube (fig 1). This nebiver pumps gas at the set oxygen ction into the ventilator circuit thrinebuliser. The resulting increase iIrespiratory minute volume is detecsensor in the ventilator and allowsate adjustment of the preset r(minute volume.Two studies of aerosol deposit

VENTILATORCathe

Figure 1 Ventilator circuit and nebuliser apparatus. ET-endotracheal tube.

n at thatf ventila-y medical;tudy andned writ-

and werego900Ccontrol

:heal tube

performed on each patient. The first study(experiment 1) was performed with a nebu-liser solution of 74 MBq 99mTc HSA (50 ,ug)in 3 ml saline. During the 30 minute periodof administration of aerosol the appearance of99mTc HSA in the lungs was monitored con-tinuously (anterior projection) with a portablegamma camera (small field of view withdiverging collimator) in 15 second countingframes. After aerosol administration 300 sec-

ray film. ond static scans were taken of the lungs (pos-e basis of terior and anterior projections) and fromand were these the geometric mean intrapulmonaryrial blood activity was calculated.'3 The posterior viewI patients was obtained by scanning the lungs from

underneath through the bed. Activity in thevas mea- nebuliser was measured with the gammatechnique camera before and after nebulisation.g subjects Experiment 2 was performed immediately15 Before after completion of experiment 1. Identical33 ('33Xe) nebulisation equipment, ventilator settings,to define and experimental methods were used except[his was that on this occasion the nebuliser solutionMBq) of contained 185 MBq 99mTc HSA activity.er mount Activity detected in the chest after experimentthe endo- 1 was subtracted from that measured afterfore infla- experiment 2.opped for Absolute pulmonary deposition of the neb-vere filled ulised aerosol, expressed as a percentage ofage was the initial nebuliser dose, was estimated by aima cam- modification of the lung phantom methodtion was described by Newman et al.'6 Preliminaryllected in experiments were performed on three patientsoutlet of to measure their anteroposterior chest tissueiplete. attenuation by means of a 30 cmced by jet diameter 57Co flood source as previouslyvolume described.'3 These experiments indicated that

ires that their tissue attenuation was similar to thatring lung produced by a 16 cm thickness of mix-D tis-nebuliser sue equivalent material. This figure was simi-used for lar to that found in a larger group of healthywhen the subjects.'3 To correct for tissue attenuation athis was piece of absorbent paper uniformly soakednebuliser with 74 MBq 99mTc was imaged directly ona System the collimator face and then anterior and pos-_td, UK) terior (from beneath the bed) scans of thisiount and phantom were taken with an 8 cm thicknessuliser dri- of mix-D (half of the tissue attenuation esti-,oncentra- mated in the preliminary studies) placedtough the between it and the camera for each view. Then expired same 99mTc activity was also placed within acted by a nebuliser and measured on the camera. Fromappropri- the ratio of counts detected in the nebuliserespiratory and counts detected in the anterior and pos-

terior phantom views it is possible to calculatetion were a correction factor for each view that relates

activity detected in the phantom to activitywithin the nebuliser unit. This same correc-tion factor was applied for each patient in thestudy. This was considered appropriate

ter mount because it was impractical to make individual-q measurements in these patients and becauseI} ET tube our previous study had shown that the use of

individually measured correction factors onlyATIENT reduced the intersubject variability in pul-

monary deposition measurements by a smalland statistically non-significant amount.'3The distribution of deposited aerosol withinthe lung was studied by calculating the ratiosof peripheral to central and upper to lower

155

on Septem

ber 28, 2020 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.48.2.154 on 1 February 1993. D

ownloaded from

6Thomas, O'Doherty, Fidler, Page, Treacher, Nunan

Table I Details of the nine patients studied

Patient details

Mean (SD) Range

Preoperative assessment:Age (y) 57 (7) (46-66)Height (m) 1-76 (0.04) (1-70-1-80)Weight (kg) 89 (11) (68-103FEV, (1) 3-11 (0-88) (2-03-4-15)FVC (1) 3-43 (1-05) (2-08-4-15)PEF (I min) 526 (100) (422-622)

Postoperative state:FiO2 (%) 59 (14) (30-80)pH 7-45 (0-06) (7 35-7 54)P02 (kPa) 18-0 (5-7) (116-27-6)Pco, (kPa) 5-2 (0-6) (4 4-5 9)Ventilation rate (breaths/min) 13 (1) (12-15)Respiratory minute volume (I min ') 9-1 (1-3) (6-8-11-2)Inspiratory time (%) 30 (4) (25-33)Pause time (%) 17 (5) (10-20)

FEVy-forced expiratory volume in one second; FVC-forced vital capacity; PEF-peakexpiratory flow; FiO,-fractional inspired oxygen; P02-partial pressure of oxygen;Pco2-partial pressure of carbon dioxide.

zone 99mTc activity. Each ratio was then divi-ded by a similar ratio obtained from the '33Xescan to correct for regional ventilation. Valuesless than unity indicate preferential deposi-tion in the central or lower lung zones.The reproducibility of the measurements

of aerosol deposition was determined by atwo way (subjects and experiments) repeatedmeasures analysis of variance and calculationof the coefficient of variation (CV) (withinand between subjects) as the square root ofthe appropriate mean square value divided bythe mean deposition for both experimentscombined. The method of Bland and Altmanwas also used.'7

ResultsStudies were performed on nine men. Table

Table 2 Radioaerosol distribution (n = 9)

Deposition (%) CV (%)

Experiment 1 Experiment 2 Between WithinMean (SD) Mean (SD) subjects subjects

Pulmonary depositionBoth lungs 2-28 (0 84) 2-12 (0-69) 46 15Right lung 1-54 (0 70) 1-39 (0 69) 63 20

Central 0-31 (0-17) 0-27 (015) 75 32Peripheral 1-08 (0 50) 0-84 (0 35) 52 17Upper 0 30 (0-16) 0-22 (0-15) 78 66Lower 0 49 (0-26) 0 37 (0-13) 49 57

Left lung 0 75 (0 30) 0 73 (0-32) 57 5Central 0 10 (0-10) 0-14 (0 20) 143 72Peripheral 0-56 (0-21) 0-51 (0-23) 57 20Upper 0-16 (0 06) 0-18 (0-12) 73 23Lower 0-24 (0-13) 0-22 (0-13) 75 25

Extrapulmonary depositionTracheal/endotracheal tube 0-88 (0-51) 1-12 (0 79) 74 50Exhalation filter 11 1 (2-7) 11-8 (1-8) 26 11Nebuliser retention 51-5 (8-1) 47-8 (7-4) 19 15Unaccounted (tubing) 34-1 (8 9) 37-2 (9 0) 33 24

Deposition ratios ('33Xe corrected)*Peripheral:central 0-54 (0-19) 0-51 (0 23) 53 12Upper:lower 1-40 (0 58) 1-22 (0-68) 60 30

Time to plateau (min) 23-4 (7 4) 22-4 (4-5) 24 9

*Values greater than 1 indicate preferential deposition in the peripheral or upper lungzones. All differences between experiments were non-significant by paired t tests.

c .0 04

<l D. 08

0 1E 024.C 0)

2

CLa 0 10 20 30

Time (min)

Figure 2 Dynamic deposition of radioaerosol in rightperipheral lung during nebulisation. Open circlesexperiment 1; closed circles experiment 2. Results are meanand SE values.

1 shows their clinical details, including theirpreoperative lung function and ventilationconditions after operation. Three were smok-ers and six were ex-smokers. All had under-gone coronary artery bypass graft surgery andwere intubated with 8 or 9 mm cuffed endo-tracheal tubes.

Figure 2 shows the dynamic aerosol depo-sition. For each experiment deposition wascompleted within the 30 minute period ofstudy and the mean time for completion(time to plateau) was 23 minutes for experi-ment 1 and 24 minutes for experiment 2.

Figure 3 presents examples of representa-tive I"Xe ventilation and 99mTc HSA deposi-tion scans. These show poor ventilation anddeposition of aerosol in the left lung. Thispattern was found in six of the patients; inone patient ventilation and deposition in theright lung was poor. Substantial deposition ofaerosol was found around the lower part ofthe endotracheal tube.

Quantification of aerosol depositionshowed that only 3-2% of the initial nebuliserdose reached the patient's respiratory tract.Of this, 1% was deposited in the endotrachealtube or trachea and only 2-2% reached thelungs (table 2). Most of the nebuliser solutionwas retained within the nebuliser unit or wasdeposited in the nebuliser connection orcatheter mount; 11-5% of the aerosol wasdeposited in the exhalation port filter; and32% of the initial nebuliser activity was notaccounted for, presumably because of deposi-tion within the ventilator circuit. It was notpossible to verify this directly because of thesize of the circuit-several metres of semirigidtubing, two traps, and a large heated waterbath-and because of the difficulty in makinga reliable correction for the geometric distri-bution of counts.

Deposition of aerosol in various lungregions (expressed as corrected counts) anddeposition ratios of peripheral:central lungand upper:lower lung (each corrected forregional lung ventilation from similar ratioscollected from the '33Xe scan) indicated pref-erential deposition of aerosol in the centraland upper lung regions (table 2).

Table 2 shows the repeatability of the

156

on Septem

ber 28, 2020 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.48.2.154 on 1 February 1993. D

ownloaded from

Pulmonary deposition of a nebulised aerosol during mechanical ventilation

95% limits of agreement of 0-65 and+ 1-68-that is, 95% of experiment 1 valueswere from 0-65 to 1-68 times the correspond-ing experiment 2 values. Corresponding lim-its of agreement for nebuliser output were

0-70 and 1-22 with a mean bias of 0-92.

DiscussionThe effectiveness of nebulised drug treatmentduring mechanical ventilation is expected todepend in part on the efficiency of drug de-position in the lungs. Inconsistent drug deliv-ery may explain why some studies have foundnebulised drugs to be effective under theseconditions'8-20 whereas others have not.' 7 Thetwo previous attempts to measure aerosoldeposition during mechanical ventilation havebeen inadequate because they have usedinappropriate radionuclide tracers and subop-timal imaging techniques. Fuller et al used a99mTc sulphur colloid tracer and estimatedpulmonary deposition from a Bennett twin jetnebuliser to be 1-2%. A large proportion ofthis tracer becomes attached, however, to theplastic tubing of the ventilator circuit and theamount of a therapeutic drug reaching thelung may be underestimated by as much as

three times with this method.'0 MacIntyre et

4-

CN

0

0.

xw

Figure 3 Examples of representative '33Xe breath hold (A) and 99*Tc deposition (B)scans from the same subject. In scan B the mediastinum and endotracheal tube areas havebeen masked to improve definition of the lungs.

experiments, obtained by comparison ofresults of experiment 1 and experiment 2 foreach patient and expressed as within subjectCV. Figure 4 presents individual results fortotal pulmonary deposition. Although therewas considerable variability between subjects,the within subject variability was comparablewith the variability in nebuliser radioaerosoloutput for deposition in the lungs as a wholeand for the peripheral lung regions. Withinsubject variability was greater for depositionin the smaller lung regions and for the tra-cheal region. Analysis of variability'7 was per-formed for total pulmonary deposition (fig 4).There was no significant bias between experi-ments (experiment 1 - experiment 2,d=0-16%, 95% confidence interval, (95%CI) - 0-25 to +0-58%) and the 95% limits ofagreement (with their 95% CIs) were - 0-94(-1-91 to 0-03) and 1-26 (0-29 to 2-23)%.As the discrepancy may increase in relation tothe mean value for the two experiments (fig4, middle panel) a similar analysis was per-formed on log transformed data (fig 4, lowerpanel). This showed a mean bias of 1-05 with

3-

2

1

14

-1

0.0

0 1 2

Experiment 1

I .

3 4

E ------------------------- d+2s

x 0

------ ----------------------------d0~~~~~@~ dE

xi -1- ------___------.--------- d -2s0 1 2 3 4 5

(Experiment 1 + experiment 2)/20

X05---------------d+2sg -050-0

-------------0------------ d

o -05

E._x-.1.o0 -05 0o0 05 1.0 1.5

(In experiment 1 + In experiment 2)/2

Figure 4 Comparison of total pulmonar-v deposition (%initial nebuliser dose) after experiments I and 2. Top panelsimple comparison; middle patnel Bland-Altman analysis ofagreement;" lower panel Bland-Altmnan analysis of logtransformed data.

0lffwwx

157

1)

on Septem

ber 28, 2020 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.48.2.154 on 1 February 1993. D

ownloaded from

Thomas, O'Doherty, Fidler, Page, Treacher, Nunan

al I estimated deposition to be 2-8% from a99mTc diethylenetriaminepenta-acetate mar-

ker. This marker may also underestimate lungdeposition as it is rapidly removed from thelungs into the bloodstream. Furthermoresome of the activity may have been in the pul-monary blood pool rather than in the air pas-sages, introducing further error. In neitherstudy were both anteroposterior and pos-teroanterior scans taken for the determinationof geometric mean counts. This will lead tofurther inaccuracy if the front to back distri-bution of pulmonary activity is uneven.The marker used in our experiments is

more suitable for deposition work because itis not removed from the lungs by passage intothe bloodstream and remains at its site ofdeposition for several hours before removalby mucociliary clearance. This process is slowand no corrections are required to take it intoaccount. The behaviour of the tracer in theaerosol seems to mimic that of therapeuticdrugs9-1 and its addition does not alter par-ticle sizes significantly.""1 We have used thegeometric mean count to obtain estimates ofpulmonary deposition. Our lung phantom,used to correct deposition data for tissueattenuation and geometric errors, was shownin preliminary studies to be associated withattenuation similar to that in our patients andthis phantom method of correction estimatesactual intrapulmonary activity to within 6%.21With these methods we have confirmed

that aerosol delivery is poor during mechani-cal ventilation. The lung deposition of 2-2%found here is much less than the 5% deposi-tion measured with both the same methodsand nebuliser during spontaneous breathing,but is consistent with the delivery of 4%-5%through an endotracheal tube in in vitromodels.22 2' Deposition varied widely betweenpatients, as has been found in subjectsbreathing normally,'3 but immediate repro-ducibility within the same patient was good.The distribution of deposition within thelungs was similar to that previously found insupine subjects breathing spontaneously;increased aerosol was deposited in the upperlungs.'4 The ratio of central to peripheraldeposition was reduced compared withresults obtained during spontaneous respira-tion. It is possible that larger aerosol particlesimpact on the ventilator tubing or endotra-cheal tube and do not reach their usual site ofdeposition in the larger bronchi. Increasedcentral deposition is seen in patients with air-flow obstruction24 and in view of their needfor inhaled bronchodilators further studies inthis group would be of interest. It is not clearwhy the ventilation and deposition in the leftlung were so poor after open heart surgery. Itis possible that this was due to manipulationof the lung during surgery. Alternatively theinspiratory air flow may have been preferen-tially directed into the right main bronchus bythe endotracheal tube, although its positionwas confirmed radiologically in each case.Studies after other types of surgery would beof interest.The output of this nebuliser was about

50% during mechanical ventilation, similar tothe output in spontaneously breathing sub-jects."3 Thus a reduction in nebuliser outputis not the explanation for the poor deposition.Instead it seems that a substantial proportionof the aerosol is deposited on the ventilatortubing and endotracheal tube, as described byothers. 1 5 6 25 Inspiratory phased nebuliseractuation as used here might be expected toincrease aerosol delivery to the lungs byreducing the inevitable wastage of the aerosolduring expiration.26 Observation of the nebu-liser suggests that this does not happenbecause there is a delay between its actuationand the production of aerosol. As a resultmost of the aerosol is produced too late in theinspiratory period to be carried to the lungs.Instead it is lost down the expiratory limb ofthe ventilator circuit during the subsequentexpiratory phase. Improved delivery isobtained by continuous nebulisation27 butthis is incompatible with volume cycled venti-lation.

Several ways of improving aerosol deliveryduring mechanical ventilation have been sug-gested. These include use of nebulisers pro-ducing submicronic particles, which are lesslikely to deposit on the tubing56; increasingthe volume of nebuliser solution and thus thedrug output of the nebuliserl" 23; changing theventilator settings to increase the inspiratory23or pause times; altering the position of thenebuliser in the ventilator circuit2326 or thelength of tubing between nebuliser andpatient25; or making use of large capacityultrasonic nebulisers23 or pressurised metereddose inhalers, with27 or without' spacerattachments. None of these techniques hasyet been compared in vivo by adequateexperimental methods. The methoddescribed here seems suitable for such studiesbecause it uses an adequate tracer andmethod of estimating intrapulmonary activityand because paired measurements can bemade in the same patient within a short timewith acceptable repeatability. This is impor-tant as it allows patients to be studied withdifferent nebuliser arrangements before theirclinical state changes. In view of the pooraerosol delivery found with current methodsof aerosol administration under these condi-tions, such studies are urgently required.

We thank Valerie Arnold and Heather Bowen for technicalassistance and the staff of Mead ward, St Thomas's Hospital,for their help and cooperation with the study. We are particu-larly grateful to Mr B T Williams, Mr G Venn, Dr C Apps,Dr T Hunt, and Dr R Linton for allowing us to study theirpatients and for their cooperation with the research. Thiswork was supported by a grant from the Intensive CareSociety.

1 Fuller HD, Dolovich MB, Posmituck G, Wong Pack W,Newhouse MT. Pressurised aerosol versus jet aerosoldelivery to mechanically ventilated patients.Comparison of dose to the lungs. Am Rev Respir Dis1990;141:440-4.

2 Stoutenbeek CP, van Saene HK, Miranda DR, ZandstraDF, Langrehr D. Nosocomial gram-negative pneumo-nia in critically ill patients. A 3 year experience with anovel therapeutic regimen. Intensive Care Med1 986;12:4 19-23.

3 Frankel LR, Wilson CW, Demers RR, Parker JR.Lewiston NJ, Stevenson DK, et al. A technique for thf

158

on Septem

ber 28, 2020 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.48.2.154 on 1 February 1993. D

ownloaded from

Pulmonary deposition of a nebulised aerosol duning mechanical ventilation

administration of ribavirin to mechanically ventilatedinfants with severe respiratory syncytial virus infection.Crit Care Med 1987;15: 1051-4.

4 Girard PM, Clair B, Certain A, Bidault R, Matheron S,Regnier B, et al. Comparison of plasma concentrationsof aerosolised pentamidine in nonventilated and venti-lated patients with Pneumocystosis. Am Rev Respir Dis1989;140:1607- 10.

5 Ahrens RC, Ries RA, Popendorf W, Wiese JA. The deliv-erv of therapeutic aerosols through endotracheal tubes.Pediatrpulmonol 1986;2:19-26.

6 Flavin M, MacDonald M, Dolovich M, Coates G,O'Brodovich H. Aerosol delivery to the rabbit lung withan infant ventilator. Pediatr Pulmonol 1986;2:35-9.

7 Maclntyre NR, Silver RM, Miller CW, Schuler F,Coleman RE. Aerosol delivery in intubated, mechani-cally ventilated patients. Crit Care Med 1985;13:81-4.

8 Thomas SHL, Langford JA, George RDG, Geddes DM.Aerosol deposition in the human lung: effect of high-fre-quency oscillation on the deposition characteristics of aninhaled nebulised aerosol. Clin Sci 1988;75:535-42.

9 Smaldone GC, Perry RJ, Deutsch DG. Characteristics ofnebulizers used in the treatment of AIDS relatedPneumocystis carinii pneumonia. Journal of AerosolMedicine 1988;1:113-26.

10 O'Riordan TG, Greco MJ, Perry RJ, Smaldone GC.Nebulizer function during mechanical ventilation. AmRev Respir Dis 1992;145:1117-22.

11 O'Doherty MJ, Thomas SHL, Page C, Clarke AR,Mitchell D, Heduan E, et al. Does "9-Tc human serumalbumin alter the characteristics of nebulised pentami-dine isethionate? Nucl Med Commun 1989;10:523-9.

12 Perry RJ, O'Riordan TG, Smaldone GC. Inaccuracies inmeasurements of nebulized drug delivery. Am RevRespir Dis 199 1;143:A708.

13 Thomas SHL, O'Doherty MJ, Page CJ, Nunan TO.Variability in the measurement of nebulised aerosoldeposition in man. Clin Sci 1991;12:767-75.

14 O'Doherty MJ, Thomas SHL, Page CJ, Bradbeer C,Nunan TO, Bateman NT. Does inhalation of pentami-dine in the supine position increase deposition in theupper part of the lung? Chest 1990;97:1343-8.

15 O'Doherty M, Thomas S, Page C, Bradbeer CS, NunanT, Bateman N. Pulmonary deposition of nebulised pen-tamidine. The effect of nebuliser type, dose and volume

of fill. Thorax 1990;45:460-4.16 Newman SP, Woodman G, Clarke SW. Deposition of

carbenicillin aerosols in cystic fibrosis: effects of nebu-liser system and breathing pattern. Thorax 1988;43:318-22.

17 Bland JM, Altman DG. Statistical methods for assessingagreement between two methods of clinical measure-ment. Lancet 1986;i:307- 10.

18 Gay PC, Rodarte JR, Tayyab M, Hubmayr RD.Evaluation of bronchodilator responsiveness in mechan-ically ventilated patients. Am Rev Respir Dis1987;136:880-5.

19 Wegener T, Wretman S, Sandhagen B, Nystrom SO.Effect of ipratropium bromide aerosol on respiratoryfunction in patients under ventilator treatment. ActaAnaesthesiol Scand 1987;31:652-4.

20 Legare M, Petrof B, Simkovitz P, Goldberg P, GottfriedS. Aerosolised ipratopium bromide in mechanically ven-tilated COPD patients (abstract). Am Rev Respir Dis1988;137:60.

21 Forge NI, Mountford PJ, O'Doherty MJ, Coakley AJ. Acomparison of quantification techniques in planarradionuclide lung imaging (abstract). Nucl Med Commun1991;12:296.

22 Fraser I, DuVall A, Dolovich M, Newhouse MT.Therapeutic aerosol delivery in ventilator systems. AmRev Respir Dis 1981;123:107.

23 O'Doherty MJ, Thomas SHL, Fidler HM, Page CJ,Treacher DF, Nunan TO. Delivery of a nebulisedaerosol to a lung model during mechanical ventilation.Effect of ventilator settings and nebuliser type, position,and volume of fill. Am Rev Respir Dis 1992;146:383-8.

24 Vezina W, Chamberlain M, Vinitski S, King M,Nicholson R, Morgan WK. Radioaerosol ventilationimaging in ventilator-dependent patients. Technicalconsiderations. Clin Nucl Med 1985;10:759-66.

25 Kim CS, Eldridge MA, Sackner MA. Delivery efficiencyof aerosols in intubated subjects. Am Rev Respir Dis1984;129:A1 10.

26 Hughes JM, Saez BS. Effects of nebulizer mode and posi-tion in a mechanical ventilator circuit on dose efficiency.Respiratory Care 1987;32:1131-5.

27 Simonds AK, Newman SP, Cox D, Clarke SW. Aerosoldelivery during mechanical ventilation: nebuliser ornebuhaler (abstract). Thorax 1990;45:315P.

159

on Septem

ber 28, 2020 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.48.2.154 on 1 February 1993. D

ownloaded from