Archives of Disease in Childhood 1996;75:F213-F218

CURRENT TOPIC

Graphic analysis of pulmonary mechanics in

Graphic analysis of pulmonary mechanics inneonates receiving assisted ventilation

Sunil K Sinha, Joanne J Nicks, Steven M Donn

With the increasing complexity of neonatalrespiratory care, pulmonary mechanics testingis emerging as a valuable tool to aid clinicaldecision making in the management of venti-lated infants. Although there are as yet no pub-lished randomised controlled clinical trials tosuggest that pulmonary mechanics testingreduces mortality or morbidity, it has-in con-junction with clinical, radiological, and bloodgas monitoring-changed neonatal ventilationfrom "good judgment" to "informed judg-ment."' It is not surprising that pulmonarymechanics testing is increasingly becoming anessential element in the assessment of patientstatus, therapeutic evaluation, and manage-ment guidance of infants with ventilatordependence. A working knowledge of pulmo-nary mechanics also improves understandingof pulmonary physiology and pathophysiologyand their responses to mechanical ventilatorysupport.2 Unfortunately, these graphical datacontinue to be additional information that theaverage bedside clinician is currently unsurehow to interpret.The purpose of this article is to disseminate

the information on this technique based on ourexperience using the Bird Graphic Monitor(Bird Products Corporation, Palm Springs,California, USA), which in conjunction withits Partner IIi Monitor and VIP Bird Infant/Pediatric Ventilator, provides a continuous dis-play ofgraphic representation ofpressure, flow,or volume waveforms, or relations betweenpressure-volume and flow-volume in the formof loops. It also gives numerical values for pres-sure, tidal volume, and flow, and can be usedfor establishing trends up to 24 hours. All theabove graphics can be obtained as printoutswhen used with a standard computer printer.

BackgroundPulmonary mechanics monitoring consists ofmeasurements of several variables which definedifferent aspects of lung function. Although ithas become popular to refer to this type ofanalysis as "pulmonary function testing,"measurements of functional residual capacity(FRC) are not included. Devices are becomingavailable to determine FRC at the bedside,using either nitrogen washout or heliumdilution techniques, but at present themethodology is extremely expensive and im-practical. Many different pulmonary mechan-

ics tests can be performed on ventilatedinfants; only a few are commonly used in clini-cal practice. Specifically, clinicians are inter-ested in the pressure necessary to cause aflow ofgas to enter the airway and increase the volumeof the lungs. From these variables, several othermeasures of pulmonary mechanics can also bederived, such as pulmonary compliance andresistance, and resistive work of breathing(energetics). Compliance is the term used todescribe the relation between a change involume and the pressure required to producethat change: C=AV/P. It gives informationabout the elasticity of the lungs. Resistance is aresult of friction of gas flow against the air con-ducting system. It is roughly measured as thechange in transpulmonary pressure divided bychange in air flow, and is an indicator of airwayfunction. The product of compliance andresistance determines the time constant, whichis a measure of how quickly the lungs caninhale or exhale, or how long it takes for thealveoli and proximal airway pressure to equili-brate. From these mechanics, which can bedisplayed as numerical values or as graphic sig-nals, useful information can be obtained andused for diagnosing specific lung pathology,evaluating disease progression, and determin-ing therapeutic interventions.Although the facilities to monitor airway

pressure, tidal volume, and gas flow inventilated infants have existed for some time,their application has been mostly limited toresearch purposes. However, recent advancesin microprocessor technology for on-lineanalysis of pulmonary mechanics have madesuch evaluations easily available for bedsideclinical application."4 Indeed, most of the newventilators either come with, or have an optionfor, a graphical display, and this has become anessential feature of the newer ventilators avail-able for infant/paediatric use. These on-linesystems obviate the need for interruptingventilation. In addition, because of technologi-cal advantages, the sensors are very light inweight and add minimal dead space to the ven-tilatory circuit. This permits their applicationto even the smallest preterm infants.

Putative advantagesThe rationale of pulmonary mechanics testingin ventilated infants is based on the assumptionthat early identification of pulmonary prob-

Division ofWomenand Children,Directorate ofNeonatology,South ClevelandHospital,MiddlesbroughClevelandTS4 3BWS Sinha

Department ofPediatric RespiratoryCareJJ Nicks

Department ofPediatrics (Division ofNeonatal-PerinatalMedicine),University ofMichiganMedical Center,Ann Arbor, Michigan,USASM Donn

Correspondence to:Dr Sunil Sinha.

Accepted 20 May 1996

F213

on July 4, 2020 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.75.3.F

213 on 1 Novem

ber 1996. Dow

nloaded from

Sinha, Nicks, Donn

Paw17.50cmH20

-1

VT4ml

VT4ml

10

5

-E0.1

Reference TarB cursor cur

5

Inspiratoryflow

s 2 4 6

SEl ., ^ ss.1'IJ- ' 1 -'S'IV '1 Ivl>'

-5 X2.90 Seconds

5r- I

0.41pm

-5

'r *rr- T rExpiratoryflow

Paw 19.00 cm H2O

rgetrsor

8

25-3-96 11:44

Figure 1 Scalar waveforms for airway pressure (PAW") and tidal volume (VT) (A), anc

tidal volume and airway flow ( (B). Time intervals can be measured using a referencicursor (dotted line). A target cursor can be moved to any given point on a wave to obtai:numerical values ofpressure, volume, orflow at that point.

Fixed inspiratory time Termination sensitivity onTermination sensitivity off Flow terminated bre

/

-10

Figure 2 Flow and volume waveforms showing an initial inspiratory time which is to

long, but which is subsequently shortened by use of termination sensitivity mode.

lems, either inherent or iatrogenic, and institu-tion of appropriate therapeutic or ventilatoryadjustments will improve the dysfunctionand/or reduce the incidence of acute and

: chronic lung injuries.56 Besides assessment ofacute respiratory distress and evaluation ofmechanical ventilation, potential benefits of

q on-line pulmonary graphics include assess-ment of suitability for weaning, monitoring of

10 complex treatments such as extracorporealmembrane oxygen (ECMO) or nitric oxide,and follow up of chronic lung disease.

Methods-I The two forms of respiratory graphics most

widely utilised in clinical practice are scalarwaveforms and loops, both showing simultane-ous relations between pressure, volume, and airflow. Although the software program whichmonitors and plots these graphs can also givenumerical values for each variable and calcu-lated parameter, in practice, most of the usefulinformation can be derived by the visualinspection of their morphology. This varies,according to specific respiratory functionstatus, provided allowance has been made forerrors arising from inaccurate calibration andartefacts.

It is beyond the scope of this paper to detailthe specific methods through which the basicsignals of flow, volume, and pressure undergotransduction to graphic or numeric displays,and several references are available for suchpurpose.78 However, the same microprocessorbased technology used to provide patienttriggered ventilation is used in conjunctionwith highly sensitive transducers generallyplaced close to the endotracheal tube connector.

Respiratory waveformsFigure 1 shows a simultaneous representationof pressure and volume (A) and volume and

d flow (B) scalar waveforms during trigger venti-e lation in time cycled pressure limited mode.n The magnitude of individual parameters and

their time interval can be measured by use ofthe reference and target cursors. The basicinterpretation and the clinical application ofthese waveforms are described below.

athFLOW WAVEFORMSFlow may be thought of as the volume of gasdelivered per unit of time. Inspiratory flow isplotted above the abscissa (positive) andexpiratory flow is plotted below the abscissa

t~ (negative). The intersection of the abscissa andordinate occurs at zero flow. The duration andshape of the inspiratory and expiratory flowwaveform are affected by many factors, includ-

10 ing type of ventilation, inspiratory:expiratorytime ratio, impedance to gas flow during expi-ration, and effect of therapeutics, such as bron-chodilators.The major advantage of looking at flow

waveforms is in assessing the amount of flow- that occurs during the inspiratory phase to see

if the set inspiratory time is inappropriatelylong. Figure 2 shows a flow waveform with

0 fixed inspiratory time. This may be needed toachieve a high mean airway pressure and better

I a i ii - i . i i i i

F214

on July 4, 2020 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.75.3.F

213 on 1 Novem

ber 1996. Dow

nloaded from

Pulmonary mechanics in neonates receiving assisted ventilation

10 _

-10

10

;-, II

Lower resistance

F 1LJXDIJX fNA,Higher resistance

-10

Figure 3 Flow waveforms. Upper tracing shows lower expiratory resistance. Increasedexpiratory resistance in bottom tracing results in flattened slope and slow return to baseline.

oxygenation during the acute stages of respira-tory distress syndrome with stiff lungs andshort time constants, but once the lungsbecome compliant and if baby is breathing at afast rate especially while on patient trigger ven-tilation mode, fixed inspiratory time may proveto be inappropriately long. This is likely (math-ematically) to result in insufficient expiratorytime and contribute towards air trapping. Inthese situations, inspiratory flow can be limitedby either shortening the inspiratory time orinstituting a mode called "termination sensitiv-ity" (in VIP Bird Ventilators), which limit theinspiratory flow to a preset value between75-95% of peak inspiratory flow and then trig-ger the expiratory phase. This permits com-plete synchronisation. The duration and shapeof expiratory flow waveforms depend both on

End expiratoryflow

rArrn A nf lrAG :1p ~LFigure 4 Flow waveform shows air trapping. Expiratoryflow (below baseline) neverreaches zero before next breath is initiated.

the resistance and the compliance, but becauseof the small diameter endotracheal tubes usedin neonates, resistance is a more importantdeterminant for its configuration. The expira-tory waveform can be used to observe changesin expiratory resistance (fig 3) and the presenceof gas trapping (fig 4). Normally, expiratoryflow reaches a zero flow state before the nextbreath begins. If the expiratory time is tooshort, the expiratory flow fails to reach a zeroflow state before the next breath starts, thussuggesting that gas trapping may be occurring.A similar appearance as shown in fig 4 canhappen due to a large leak around the endotra-cheal tube, but this can be differentiated eitherby using the pressure-volume loops which failto close, or by monitoring the differencebetween inspiratory and exhaled tidal volume.

VOLUME AND PRESSURE WAVEFORMSMeasurement of tidal volume is becomingincreasingly important in ventilatory manage-ment. The desired inspiratory tidal volume fora ventilated breath ranges between 5.2-7.2ml/kg9 while a tidal volume of 3.5-5 ml/kg forspontaneous breaths generally indicates suit-ability of the infant for weaning, provided thisis matched by satisfactory arterial blood gaseswhich remain the best index of ventilatoryadequacy. Minute ventilation (tidal volume xrespiratory rate) is also currently being evalu-ated as one of the predictor of weaning(250-400 ml/kg/minute),'0 especially in smallinfants who have short inspiratory times and ahigher respiratory frequency. Unlike flowwaveforms, both the inspiratory and expiratoryphase of volume and pressure waveforms arepositive and are therefore plotted above theabscissa. Breath-to-breath variability andlonger term trends are useful in selecting themode of ventilation and individualised adjust-ments to customise ventilator parameters foreach patient. Figure 5 shows differences intidal volume delivery between mechanical andspontaneous breaths during synchronised in-termittent mandatory ventilation (SIMV). Thetop waveform indicates that spontaneous tidalvolumes inbetween mechanically deliveredbreaths are ineffective. On the bottom graph,the infant has much higher spontaneous tidalvolumes suggesting readiness to wean.

Mechanical breathSpontaneous breath

ia -\ 1/

X4

Figure 5 Volume waveform. Top tracing shows ineffective spontaneous breaths; lowertracing shows improved tidal volume delivery during spontaneous breathing.

Pulmonary mechanics loopsThe loops are commonly used to demonstratecorrelations between airway pressure andvolume, and airflow and lung volume (fig 6).Recently, attempts have also been made to pro-duce a flow-pressure loop, but it is not yetavailable in clinical practice. At present, there isno agreed convention regarding the manner inwhich the loops are generated. Some devicesdraw them clockwise, others anticlockwise.Some label upward deflections as negative,others positive. The clinician must be familiarwith these differences to be able accurately tointerpret the presented data.

PRESSURE-VOLUME LOOPSThese loops graphically depict the correlationsbetween ventilator inspiratory pressure and the

10

E

0.

-10

40

-1040

CD

-10

- f Iefe ( i rI I i I

OV * '. Is

No %I -- ilk 3"JXP1ki1x / \-j

F215

/\n ~~~~~~~~~~~~~~~~~~~~~~~I

-/"' I

on July 4, 2020 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.75.3.F

213 on 1 Novem

ber 1996. Dow

nloaded from

Sinha, Nicks, Donn

Expiration -100

20 Vt (ml)

Inspiration 100

E

Paw (cm H20)

Figure 6 Pulmonary mechanics loops. Left graph displays pressure-volume relations. Ninspiratory and expiratory limbs, drawn counterclockwise on the Bird Graphic Monitor.the right is theflow-volume relation drawn clockwise, in which airflow to the patient(inspiration) is plotted positively, airflow awayfrom the patient (expiration) is plottednegatively.

457

25 -10

Paw (cm H20)-15 L

Paw (cm H20)

Figure 7 Pressure-volume relation. On the left, the compliance line twuld have an axisabout 30° above horizontal, indicating poor compliance. On the right, the compliance linhas improved to about 45°. This scenario might be seen in an infant with respiratorydistress syndrome after surfactant administration.

Over disl

35

Paw (cm H20)

-tension

-10

-20Paw (cm H20)

Figure 8 Pressure-volume relation. Graph on left demonstrates overdistension. Noteminimal change in volume over the last 20% of the breath. A reduction in ventilatorypressure (right graph) normalises the loop.

45r

E

-10

-15

~~~~35Critical opening

-15 pressure

Paw (cm H20)

Paw (cm H20)

Figure 9 Demonstration of effect ofPEEP on pressure-vlume relation. Left: inadequaiPEEP results in poor volume ddlivery as pressure increases. Right: higher PEEP results iappropriate volume delivery with increasing pressure.

volume of gas entering the lungs. The horiztal axis represents the degree of posiipressure exerted by the ventilator and the N

tical axis represents the gas volume changefthe lungs (fig 7). The lower portion of the kishows the dynamic relation between voluchanges and pressure changes on inspiratithe upper portion of the loop shows volu

changes during expiration. A line connectingthe points of zero flow-changeover from expi-ration to inspiration-is the compliance line,and the slope of that line reflects pulmonarycompliance. Compliance can be calculated as

5 the change in volume divided by the change inpressure; in practice an estimate of compliancecan be made by looking at the slope of thecompliance line, which is normally 450 fromthe horizontal axis. If this slope is more towardsthe vertical axis, compliance is improving, and

On conversely is decreasing if this line movesnearer to the horizontal axis. It must berealised that slope of the compliance line alsodepends on the calibrations and can bemisleading if not adjusted properly. Devicessuch as VIP Bird have an automatic scalingfunction which serves to reference the curvesuch that normal compliance line should beabout 45°. Over-distension is implied when lit-tle or no change in volume occurs as increasingpressure is delivered. One way to recognise thisis to look at the terminal (upper) portion of the

25 pressure-volume loop and see how much tidalvolume is being delivered per unit of pressurechange. This has been schematically demon-strated in fig 8 created from a test lung forinstructive purposes. A useful parameter in thiscase is the "C20."" This can be estimated by

ze calculating total compliance of the lung in rela-tion to the compliance of the last 20% of thebreath (by dividing tidal volume by peak pres-sure). If the compliance of the last 20% is lessthan 80% of the total compliance, it is sugges-tive of over-distension, and the ventilatorshould be adjusted to reach a C20 value higherthan 80%. This may be accomplished bydecreasing peak inspiratory pressure, inspira-tory time, and in some instances, positive end

L expiratory pressure.35 Pulmonary compliance is a measurement of

the distensibility ofthe lung and is in a sense anindicator of the function of the lung paren-chyma. Disease processes that make the lungstiffer and thus decrease pulmonary compli-ance in neonates include surfactant deficiencystates, pneumonia, pulmonary oedema, andpulmonary hypoplasia. Increased resistance,such as that seen in bronchopulmonarydysplasia, causes a "bowing" or hysteresisaround the compliance line, and is an indirectmeasure of the work ofbreathing (by either theinfant or the mechanical ventilator) to move

j the lungs. This resistive work of breathing35 increases as pulmonary mechanics worsen,

reflecting the increased amount of energy spentby the infant or the ventilator to breathe. It canbe roughly estimated by the total area covered

~te by the pressure-volume loop and the area adja-cent to its deflation limb and the ordinate.There are, however, major problems in theinterpretation of the resistive work ofbreathing

on- in infants on ventilation, as altering the circuittive flow alone can have considerable effects even inver- the absence of changes in lung mechanics.s in The pressure-volume loop may also be usedDOp to evaluate optimal peak expiratory endime pressure (PEEP) (fig 9). On the left, the loopion, shows a delay in volume delivery despitemine increasing pressure, indicating insufficient

25

-5

45r-

I

0

-15

70r

-10

-20

I .-

F216

_-

on July 4, 2020 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.75.3.F

213 on 1 Novem

ber 1996. Dow

nloaded from

Pulmonary mechanics in neonates receiving assisted ventilation

-1 50

Vt (ml)150

e'

E

-150

20

150

-E

Figure 10 Flow-volume loop. Left graph indicates high resistance, demonstrated by loweflow at a given volume. Right graph indicates lower resistance (such as after bronchodilatreatment) with higherflow at a given volume.

25 _

-10

15

E4-

25

Paw (cm H20)

Figure 11 Endotracheal tube leak. Volume waveform fails to reach baseline atend-expiration. Pressure volume loop fails to close.

30 2

PIP Reintubation28

cm

H20

2 hour trends

ET suction

13:21

VT4ml

11:20 11:40 12:00 12:20 12:40 13:00 13:21

Figure 12 Trend data, pressure-targeted ventilation. Even though peak inspiratorypressure is held constant, tidal volume delivery is variable and decreasing.

opening pressure. An increase in PEEP, shovon the right, results in an immediate risevolume as pressure is delivered.

FLOW-VOLUME LOOPAnother valuable measurement in the pulmnary mechanics profile is the resistance of gflow through the airways. This can be assessby observing the flow-volume loop, whi(graphically displays the relation betweechange in volume and change in airflow. Higairway resistance results in lower flow forgiven volume. Conditions causing high resisance in ventilated babies include airw;obstruction from endotracheal tube kinking (

blockage with secretions, meconium aspirationsyndrome, and bronchopulmonary dysplasia.

Figure 10 shows flow-volume loops, whichshould be rounded and smooth around thehorizontal axis like an egg on its end. The lowerportion of the loop represents inspiratory data;

5 the upper limb describes the expiratory corre-lation. (Because there is no agreed conventionon direction of inspiratory and expiratory limbamong various manufacturers, it is importantto become familiar with the direction of the

er flow-volume loop on individual machines.) Ator flow-volume loop depicts changes in inspira-

tory and expiratory flow against volumechanges, and is useful in detecting flow restric-tion from airway abnormalities. Under condi-tions ofhigh airway resistance, flow is lower fora given volume. Flow will be lower on inspira-tion with high inspiratory resistance, or loweron expiration with high expiratory resistance.

Z There is a fixed inspiratory and expiratoryresistance associated with an endotrachealtube. The example in fig 10 shows increasedinspiratory and expiratory flow from left toright, which might be seen after bronchodilatortreatment. This could also happen as a result ofhigh ventilatory pressure or improvement inlung function after an improvement in respira-tory distress syndrome. The ability to identifyand quantify endotracheal tube leaks may befacilitated through graphic monitoring. Leakscan be identified both on the volume waveform as well as pressure-volume loops (fig 11).Increased resistance and interruption to gasflow from excessive condensation arising in theventilatory circuit or accumulation of secre-tions in the airway can be identified by theappearance of jaggedness on graphics.There are a number of other useful clinical

situations, including the assessment of differ-ent ventilatory modalities (such as pressuresupport ventilation) on pulmonary function'2and the assessment of patient-ventilator inter-action (synchronous versus asynchronous),'3There is also an advantage in collecting andevaluating the trends of specific variables suchas tidal volume, minute ventilation, andfrequency for spontaneous and mechanicalbreaths as well as peak pressure, inspiratorytime, and mean airway pressure. For example,the trend depicted in fig 12 is consistent withan infant on pressure ventilation where despitepeak inspiratory pressure (PIPO remainingconstant, tidal volume has gradually decreased.This situation (as in secondary lung complica-

vn tions during respiratory distress syndrome)in demands adjustment of ventilatory parameters

by the operator. Conversely, tidal volumedelivery may automatically increase after im-provement in lung compliance despite PIP

0- being the same (as after surfactant replace-as ment), and cause iatrogenic complications ifed ventilatory adjustments are not made in time.c-hen Limitationsgh Despite the major advantages of the currenta system providing pulmonary mechanics graph-

It- ics, it should be realised that the informationay provided about the function of the lungs onlyor complements (and does not substitute) infor

i

F217

on July 4, 2020 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.75.3.F

213 on 1 Novem

ber 1996. Dow

nloaded from

Sinha, Nicks, Donn

mation gained by other means of patient moni-toring, including clinical signs and blood gasexamination. Like radiographs, graphicsshould only be taken as suggestive of acondition rather than being definitive. Pulmo-nary graphic waveforms can be misshapenbecause of inherent inaccuracy of themeasurement system (from calibration differ-ences between inspiration and expiration), orfrom temporary artefacts arising from patientposition or impedance to the gas flow bycondensation in the ventilatory circuit. It isimperative to correct these pitfalls beforeaccepting the findings of graphics as a guid-ance to change ventilatory settings or assess thepatient's clinical condition.'4 In this particularrespect, the "trends" over a period of timeseem to be of more value than individualbreath analysis. Although there are few datashowing a beneficial or more cost effectiveapproach to neonatal ventilation by the con-tinuous use of pulmonary mechanics monitor-ing, the development and implementation ofthis technology should make some studies fea-sible in the near future.

ConclusionOn-line pulmonary graphic analysis representsanother major advance in respiratory technol-ogy which promises to improve the safety andefficacy of neonatal mechanical ventilation.Clinicians are now afforded a breath-to-breathview of pulmonary mechanics and respiratorywaveforms. This permits constant surveillanceof conditions such as air trapping before theybecome clinically obvious, and the fine tuning

of ventilator settings to customise managementaccording to the problems and responses of theindividual patient. Clinicians should availthemselves of this window of opportunity.We thank Bird Products Corporation for permission to use thefigures.

1 MacDonald KD, Wirtschafler DD. Continuous neonatalpulmonary mechanics with the BICORE CP-100 monitor.Neo Intensive Care 1992; 5:55-61.

2 MacIntyre NR, Hagus CK. Graphical Analysis ofFlow, Pres-sure and Volume During Mechanical Ventilation. Riverside,CA: Bear Medical Systems, 1987.

3 Goldstein M. Continuous in-line pulmonary mechanics.Neo Intensive Care 1993;5:42-5.

4 Lloyd JS, Cvetnic WG. Continuous in-line respiratorymonitoring in the critically ill preterm infant. Neo IntensiveCare 1994;7: 14-16.

5 Fisher JB, Mammel MC, Coleman JM, Bing JR, Boros SJ.Identifying lung overdistension during mechanical ventila-tion by using volume-pressure loops. Pediatr Pulmonol1988; 5:10-14.

6 Rosen WC, Mammel MC, Bing DR, Boros SJ. Bedside pul-monary function testing reduces pneumothoraces duringinfant mechanical ventilation. Pediatr Res 1989; 4:324.

7 Cunningham MD. Monitoring pulmonary function. In:Goldsmith JP, Karotkin EH, Barker S, eds. Assisted Ventila-tion of the Neonate. Philadelphia: WB Saunders Co., 1988:233-44.

8 Donn SM. Neonatal and Pediatric Pulmonary Graphic Analy-sis: Principles and Clinical Applications. Armonk, NY:Futura Publishing Co., 1997.

9 Bhutani V, Abbasi S. Evaluation of pulmonary function inthe neonate. In: Polin RA, Fox WW, eds. Fetal and Neona-tal Physiology. Philadelphia: WB Saunders, 1992: 866.

10 Spitzer AR. Mechanical ventilation. In: Spitzer AR,ed.Intensive Care of the Fetus and Neonate. St Louis, MI:Mosby Year Book, 1996: 558.

11 Comroe JH, Forster RD, Dubois AB, Briscoe WA, CarlsenE. The Lung: Clinical Physiology and Pulmonary FunctionTests. Chicago: Year Book Mecial Publishing, 1962.

12 Nicks IJ, Becker MA, Donn SM. Bronchopulmonarydysplasia: response to pressure support ventilation.J Perinatol 1994; 14:495-7.

13 Servant G, Nicks JJ, Donn SM, Bandy KP, Lathrop C,Dechert RE, et al. Feasibility of applying flow-synchronized ventilation to very low birthweight infants.Respiratory Care 1992; 37:249-53.

14 Gerhardt TO. Measurement ofpulmonary mechanics in theNICU: limitations to its usefulness. Neonatal RespiratoryDiseases 1995; 5:1-10.

F218

on July 4, 2020 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.75.3.F

213 on 1 Novem

ber 1996. Dow

nloaded from