Ov

Srcam

R�

vi

Cw

Mtwiitttmvimpwa

ccttdcw

ACN

3

MECHANICAL VENTILATION OF VERY LOW BIRTH WEIGHT INFANTS: ISVOLUME OR PRESSURE A BETTER TARGET VARIABLE?

JAIDEEP SINGH, MD, SUNIL K. SINHA, MD, PHD, PAUL CLARKE, MB, FRCPCH, STEVE BYRNE, MD, PHD, AND STEVEN M. DONN, MD

bjective To compare the efficacy and safety of volume-controlled (VC) ventilation to time-cycled pressure-limited (TCPL)entilation in very low birth weight infants with respiratory distress syndrome (RDS).

tudy design Newborns weighing between 600 and 1500 g and with a gestational age of 24 to 31 weeks who had RDS wereandomized to receive either VC or TCPL ventilation and treated with a standardized protocol. The 2 modalities wereompared by determining the time required to achieve a predetermined success criterion, on the basis of either thelveolar-arterial oxygen gradient <100 mm Hg or the mean airway pressure <8 cm H2O. Secondary outcomes includedortality, duration of mechanical ventilation, and complications commonly associated with ventilation.

esults The mean time to reach the success criterion was 23 hours in the VC group versus 33 hours in the TCPL group (P.15). This difference was more striking in babies weighing <1000g (21 versus 58 hours; P � .03). Mean duration of

entilation with VC was 255 hours versus 327 hours with TCPL (P � .60). There were 5 deaths in the VC group and 10 deathsn the TCPL group (P � .10). The incidence of other complications was similar.

onclusion VC ventilation is safe and efficacious in very low birth weight infants and may have advantages when comparedith TCPL, especially in smaller infants. (J Pediatr 2006;149:308-13)

echanical ventilation of the newborn has been traditionally accomplished by using time-cycled pressure-limited(TCPL) ventilation. This form of ventilation is designed to deliver a volume of gas with a preset peak inspiratorypressure during a defined cycle time. As a result, the peak pressure at the proximal airway remains constant; however,

he tidal volume delivered to the lungs is variable depending on both the underlying pulmonary and chest wall mechanics. Thus,hen the lungs are stiff, tidal volume delivery is lower at the same peak pressure than when the lungs are more compliant. This

nconsistency in tidal volume delivery may be undesirable, especially in very pre-termnfants, because both overexpansion (volutrauma) and under expansion/collapse (atelec-otrauma) are thought to contribute to the pulmonary injury sequence.1-4 This has led tohe introduction of a number of newer forms of ventilation, which aim to deliver a desiredidal volume automatically, irrespective of the underlying lung mechanics.5 One suchodality is volume-controlled (VC) ventilation, also referred to as volume-targeted or

olume-limited ventilation, in which the primary gas delivery target is tidal volume, andnspiratory pressure is automatically adjusted from breath-to-breath depending on pul-

onary compliance. Thus, in conditions with low lung compliance (stiff lungs); moreressure is generated to deliver the desired tidal volume. As lung compliance improvesith resolution of the underlying pulmonary condition, the pressures generated are

utomatically reduced, which is sometimes referred to as auto-weaning.5,6

In a previously published randomized trial, infants treated with VC ventilationould be weaned faster and had a significant reduction in the duration of ventilationompared with their counterparts who were treated with TCPL ventilation.7 At the timehis study was performed, infants weighing �1200 grams were not included because ofechnological limitations. Since then, improvements in ventilator technology enabling theelivery of smaller tidal volumes allow the use of VC ventilation in smaller babies. Weompared the safety and efficacy of this new modality of ventilation in very low birtheight infants who had respiratory failure at birth and required mechanical ventilation.

aDO2 alveolar-arterial oxygen gradientPAP continuous positive airway pressure

RDS respiratory distress syndromeTCPL time-cycled pressure-limited

See editorial, p 290

From James Cook University Hospital,Middlesbrough, United Kingdom; HopeHospital, Salford, United Kingdom; and theDepartment of Pediatrics, Division of Neo-natal-Perinatal Medicine, University ofMichigan Health System, C.S. Mott Chil-dren’s Hospital, Ann Arbor, Michigan.Presented at the Annual Meeting of thePediatric Academic Societies, Washington,DC, May 16, 2005.

Submitted for publication Aug 18, 2005;last revision received Dec 16, 2005; ac-cepted Jan 23, 2006.

Reprint requests: Prof Sunil K. Sinha, Pro-fessor of Paediatrics & Neonatal Medicine,University of Durham & James Cook Uni-versity Hospital, Marton Road, Middles-brough TS4 3BW, UK. E-mail: sunil.sinha@stees.nhs.uk.

0022-3476/$ - see front matter

Copyright © 2006 Mosby Inc. All rightsreserved.

ICU neonatal intensive care unit VC Volume-controlled

08

10.1016/j.jpeds.2006.01.044

S

saTwrbNa

I

bbdTotiaamabimo

P

taHmibiirscradtigacn

R

taarsaptocrottrHat

V

BSpvtbjwbTP(srAib

svnpwawLdCdfci

M

METHODS

tudy SiteThe study was conducted at the tertiary neonatal inten-

ive care units (NICUs) at James Cook University Hospitalnd Hope Hospital between January 2002 and May 2004.he institutional review board approved the protocol, andritten informed parental consent was obtained. However,

ecruitment was discontinued earlier at the second centerecause of a relatively high number of protocol violations.onetheless, all patients have been included in the main

nalysis on an intention-to-treat basis.

nclusion CriteriaBabies were eligible for the study when they were born

etween 24 and 31 weeks’ completed gestation, weighedetween 600 and 1500 grams at birth, and had respiratoryistress syndrome (RDS) requiring mechanical ventilation.he diagnosis of RDS was made on the basis of a compositef clinical (respiratory distress) and radiographic features (re-iculogranular appearance with air bronchograms and dimin-shed lung volume) appearing within the first 24 hours of lifend evidence of respiratory insufficiency by using blood gasnalysis. All the infants in the trial received surfactant replace-ent therapy (poractant alfa, �100mg/kg), which was given

ccording to the unit practice (routinely within 30 minutes ofirth in babies <28 weeks, and as treatment after intubationsn babies �28 weeks). Infants with severe congenital malfor-

ations that adversely affect life expectancy or respiratoryutcome were not included in the study.

rimary OutcomeThe primary outcome criterion for this study was the

ime from entry into the study until achievement of either anlveolar-arterial oxygen gradient (AaDO2) <13 kPa (100 mmg), or a mean airway pressure <8 cm H2O, which had to beaintained for at least 12 hours. These measures were chosen

n preference to the subjective outcome measure of extubationecause they represent the level of support from which mostnfants can be extubated. However, when extubation occurrednadvertently before reaching these parameters and the babyemained extubated for at least 24 hours, this was also con-idered a successful outcome. Other outcome measures in-luded the total duration of mechanical ventilation, the du-ation of artificial respiratory support (mechanical ventilationnd continuous positive airway pressure [CPAP]), survival toischarge from the NICU, and the frequency of complica-ions commonly associated with premature birth and mechan-cal ventilation, such as chronic lung disease (defined as oxy-en dependency or needing CPAP at 36 weeks postmenstrualge); intraventricular hemorrhage, periventricular leucomala-ia, or both; patent ductus arteriosus requiring treatment; and

ecrotizing enterocolitis (Bell stage II or greater). u

echanical Ventilation Of Very Low Birth Weight Infants: Is Volume Or

andomizationPatients were randomized within 6 hours of the initia-

ion of mechanical ventilation. Eligible babies were stratifiedpriori into 2 groups according to birth weight (600-1000 g

nd 1001-1500 g). Assignments were generated by using aandom permutated block algorithm for each birth weighttratum irrespective of place of birth with block sizes of 2, 4,nd 6. The randomization sequence was kept hidden from theeople actually caring for the babies, and they were unawarehat a block randomization design had been used. Assignmentf a ventilation modality was made according to instructionsontained in sealed opaque envelopes. Once the infants wereandomized to the assigned modality of ventilation, no cross-ver was allowed. When infants were deemed to be failing onhe assigned modality of ventilation, high-frequency oscilla-ory ventilation was an option, if they developed severe respi-atory failure characterized by an AaDO2 �80 kPa (600 mm

g) or an oxygenation index �25, or an intractable thoracicir leak. Analyses of data, however, were done on an inten-ion-to-treat basis.

entilation StrategiesAll infants in this study were treated with the VIP

IRD Gold ventilator (Viasys Healthcare Systems, Palmprings, CA) by using a standardized ventilatory managementrotocol designed for this study. In both groups, ventilatorariables were set to target an exhaled tidal volume (VTe) of 4o 6 mL/kg. This was monitored and adjusted on an hourlyasis. In the VC group, the delivered tidal volume was ad-

usted, and in the TCPL group, the peak inspiratory pressureas adjusted. During the acute phase of illness, infants inoth study groups were placed in the assist/control mode.argeted blood gas indices, including a pH 7.25 to 7.40,aCO2 4.5 to 6.5 kPa (35-49 mm Hg), and PaO2 7 to 10 kPa

50-75 mm Hg) were used in both groups during the initialtage of ventilation. Subsequently, PaCO2 was permitted toise up to 8 kPa (60 mm Hg) if the pH remained >7.20.ssessments of blood gases were done within 1 hour of entry

nto the study and subsequently at least every 4 to 6 hours inoth groups.

Once the infants were recovering from their acute re-piratory illness (PIP �16 cm H2O and FiO2 �0.3), theentilatory mode was changed from assist/control to synchro-ized intermittent mandatory ventilation with pressure sup-ort ventilation. Caffeine citrate (20 mg/kg) was administeredhen the infants were stable for 6 to12 hours while weaning

nd receiving minimal support. After extubation, the infantsere placed on nasal CPAP (Infant Flow Driver, EMEimited, Brighton, UK), using pressures of 4 to 6 cm H2Oelivered through short binasal prongs. The duration ofPAP therapy was not specified by the trial protocol and wasetermined by the attending clinicians. Infants were weanedrom CPAP to nasal cannula oxygen or room air. In cases oflinical deterioration after extubation, the decision to re-ntubate and ventilate was made on the basis of a standard

nit protocol, including hypercapnia (PaCO2 �8 kPa or 60

Pressure A Better Target Variable? 309

maWts

S

ts(daavnwCoc

cuhCttimaW

anfi

gawea

aaf29rear

Argmh(k((

tpC16aiTweF

F

3

m Hg) associated with acidosis (pH �7.20), recurrentpnoea (�3 episodes/hour), or the need for resuscitation.

hen this happened, the initially assigned modality of ven-ilation was used. The remainder of the clinical care wasimilar for both groups.

tatistical AnalysesOn the basis of the results of the previously published

rial, which had shown that the time taken to achieve theuccess criterion was significantly lower with VC ventilation65.6 hours versus 125.8 hours.), a sample size calculationetermined that 45 babies per group would be needed to show33% difference in the 2 groups, with a 2-sided alpha of 0.05

nd 80% power. Normally distributed continuous outcomeariables were compared with the unpaired Student t test, andon-parametric continuous outcome variables were analyzedith the Wilcoxon rank-sum test or Mann Whitney U test.ategorical variables were compared with the chi-square testr Fisher exact test when appropriate. All analyses wereonducted with 2-tailed tests.

Time-based analyses of the interval to achieve the suc-ess criterion and the duration of ventilation were done bysing the Kaplan-Meier survival curves with Cox proportionalazards estimates to obtain unadjusted hazard ratios with 95%Is. Cox regression or logistic regression was used to inves-

igate treatment effects. Baseline variables with the potentialo be important prognostic factors were identified and werencluded in the univariate analysis. They were included in the

ultivariate model only when a clinically important imbal-nce was observed. The statistical software used was SPSS for

indows version 12 (SPSS, Chicago, IL).

RESULTSDuring the study period, 221 eligible patients were

dmitted to the NICU and 110 were enrolled. The reasons foron-inclusion are shown in Figure 1. Analyses were per-ormed on 57 infants assigned to VC ventilation and 52nfants assigned to TCPL ventilation.

igure 1. Flow chart showing recruitment of subjects to the study

There were no significant differences in the demo- i

10 Singh et al

raphic characteristics of both groups at entry into the study,lthough a higher proportion of infants randomized to TCPLere outborn (Table I). The severity of respiratory illness at

nrollment, assessed by using mean airway pressure, AaDO2,nd oxygenation index, was also similar.

There was no significant difference in the time taken tochieve the primary outcome measure in the 2 groups inggregate (Figure 2). However, there was a trend towardaster weaning in infants assigned to VC ventilation (mean,3 hours; 95% CI, 18- 28) versus TCPL (mean, 33 hours;5% CI, 22-44; Cox proportional hazards estimate hazardatio, 1.3; 95% CI, .9-1.9; P � .15). Five of 19 infantsnrolled at Hope hospital were protocol violations. A protocolnalysis excluding these five infants gave essentially the sameesults as the intention-to-treat analysis.

The 2 criteria for the primary end point (mean Paw oraDO2) were also assessed independently among infants

ecruited at the main center (n � 90). Infants in the VCroup achieved a mean Paw �8 for 12 consecutive hours in aean time of 79.8 hours (range, 38.5-121 hours) versus 80.5

ours (range, 23-138 hours) in the infants in the TCPL groupP � .79). The mean time taken to achieve an AaDO2 �13Pa (100 mm Hg) for 12 consecutive hours was 24 hoursrange, 18.5-29 hours) in the VC group versus 39 hoursrange, 25-53 hours) in the TCPL group (P � .08).

In the cohort of infants weighing �1000 g (n � 50),here was no significant difference in the time taken to reachrimary outcome measure (27 hours in the VC group [95%I, 17-38] and 32 hours in the TCPL group [95% CI,4-59], P � .96). However, in the cohort of infants weighing00 to 1000 g (n � 59), infants assigned to VC ventilationchieved the primary end point faster than their counterpartsn the TCPL arm (VC: mean, 21 hours [95% CI, 17-24];CPL: mean, 58 hours [95% CI, 42-74]). This differenceas statistically significant with the Cox proportional hazards

stimate (hazard ratio, 1.83; 95% CI, 1.04-3.20; P � .03;igure 3).

A posthoc analysis on the basis of the severity of illness

ncluding only those infants (n � 41) who had moderate to

The Journal of Pediatrics • September 2006

sma3

9P

Va1dtira(htWaHtp1n

erbpdI

Fca

Fpc

T

B

G

MCIICA

A

A

R

O

M

evere respiratory failure at entry into the trial (AaDO2 �100m Hg) also demonstrated that infants in the VC group

chieved faster weaning [(mean, 27 hours; 95% CI, 19-6)]compared with infants in TCPL group (mean, 59 hours;

igure 2. Kaplan-Meier plot showing time taken to achieve successriterion. The vertical axis shows cumulative survival, and the horizontal

able I. Baseline characteristics of study infants

PARAMETERSVC

(N � 57)TCPL

(N � 52)

irthweight in grams, mean(SD)

985 (232) 976 (248)

600–1000 grams n � 29 n � 291001–1500 grams n � 28 n � 23estational age in weeks,

median (IQR)27.1 (25.5, 28.3) 27.2 (25.5, 28.8)

ale sex (%) 33 (57.8%) 32 (61.5%)aucasian (%) 52 (91.2%) 48 (92.3%)

nborn (%) 51 (89.4%) 43 (82.6%)ntubated at birth (%) 48 (84.2%) 42 (80.7%)horioamnionitis 14 (24.5%) 8 (15.3%)ntenatal steroidsOne dose 10 (17%) 11 (21%)Two doses 45 (78.9%) 38 (73.0%)

pgar score at 1 minute,median (IQR)

6 (4–8) 7 (5–9)

pgar scores at 5 minutes,median (IQR)

9 (8–10) 9 (8–9)

espiratory status at entryto trial, mean (SD)

AaDO2 19.0 (18) kPa 16.3 (14.2) kPaMean paw (cm H2O) 9.4 (1.7) 9.1 (1.4)OI 6.6 (5.3) 5.5 (3.1)

I, oxygenation index.

pxis shows time in hours.

echanical Ventilation Of Very Low Birth Weight Infants: Is Volume Or

5% CI, 32-87; hazard ratio, 2.14 [95% CI, 1.07-4.29];� .03).

The mean duration of ventilation was 255 hours in theC arm (95% CI, 160-349) versus 327 hours in the TCPL

rm (95% CI, 214-441; hazard ratio, 1.05; [95% CI, 0.67-.63]; P � .6). The mean duration of oxygen therapy was 44ays in the TCPL group (95% CI, 33-54) versus 45 days inhe VC group (95% CI, 36-53; P � .91). The proportion ofnfants alive and not receiving any form of mechanical respi-atory support (assisted ventilation or CPAP) or oxygen ther-py with time was not significantly different in the 2 groupsFigure 4). There were 10 deaths before discharge from theospital in the TCPL arm, compared with only 5 deaths inhe VC arm (odds ratio, 0.27; 95% CI, 0.06-1.07; P � .1).

ith multivariate analysis, there was no significant effect ofny particular modality of ventilation per se on survival.owever, more infants in the TCPL group required rescue

reatment with high-frequency oscillatory ventilation com-ared with their counterparts in the VC group (21% versus4%; P � .26). The frequency of measured complications waso different in the 2 groups (Table II).

DISCUSSIONThis was a pragmatic study designed to assess the

fficacy and safety of VC ventilation in the management ofespiratory failure in very low and extremely low birth weightabies. Although a truly masked study would have beenreferable, the practicalities of blinding were felt to be tooifficult to accomplish. Instead, a rigorous protocol was used.t was postulated that the study would provide general com-

igure 3. Kaplan-Meier plot showing time taken to achieve primary endoint in infants with birth weight 600 to 1000 g. The vertical axis showsumulative survival, and the horizontal axis shows time in hours.

arative information for traditional TCPL and VC ventila-

Pressure A Better Target Variable? 311

trrntataaatca

tcmssistrpmcp(

dr

cdd

r2tvssvfioi

avlae

telDsteltbdmfscw

ssida

pstfacavsuathit

Fat

3

ion, which has recently become available for the treatment ofespiratory failure in very preterm newborns. Although weealized that a randomized controlled trial of this size wouldot definitively prove whether one method is superior, we felthat it was imperative to demonstrate the safety, feasibility,nd efficacy of this new technique before conducting a largerrial. The chosen success criteria provided a standard measuregainst which the speed of weaning could be objectivelyssessed. Short-term criteria, such as the speed of weaningnd the total duration of mechanical ventilation and exposureo supplemental oxygen, are also important and relevant tolinical perspectives and to the determination of resource usend cost effectiveness.

Infants randomized to receive VC ventilation had arend toward achieving the success criterion faster than theirounterparts treated with TCPL ventilation. This was evenore marked and statistically significant in the subgroup of

maller infants who weighed �1000 g. These findings areimilar to those of Sinha et al.7 In that study, babies random-zed to receive VC ventilation achieved the success criteriaignificantly faster and required a significantly shorter dura-ion of ventilation than their counterparts randomized toeceive TCPL ventilation. Infants in this study achieved therimary end point faster than in the study by Sinha et al. Thisay have resulted from an increased use of antenatal corti-

osteroids in this study (95% versus 44%) and a higher pro-ortion of infants being ventilated immediately after delivery83% versus 33%).

Additionally, all infants in this study received animal-erived surfactant, whereas half the infants in the earlier studyeceived synthetic, non-protein-containing surfactant.

There was no significant difference in the incidence ofomplications in the 2 groups in this study. However, alleaths in the first week of life were related to respiratory

igure 4. Kaplan-Meier curves showing ages at which infants were alivend weaned off any form of mechanical respiratory support or oxygenherapy.

isease and occurred exclusively in infants randomized to l

12 Singh et al

eceive TCPL ventilation. This was unexpected, because thestudy groups were closely matched for the factors affecting

he severity of RDS (Table I). Although the modality ofentilation per se did not show an independent effect onurvival on multivariate analysis, the odds ratio almost reachedignificance favoring survival among infants treated with VCentilation (odds ratio, 0.5; 95% CI, 0.2-1.2; P � .10). Thesendings, however, should be interpreted with caution becausef the small size of this study. This will need to be addressedn subsequent larger trials.

The ventilator used in this trial monitors both inspirednd exhaled tidal volume at the proximal airway, using aariable orifice pneumotachograph. One of the recognizedimitations of volume ventilation in the newborn is gas leakround the endotracheal tube. For this reason, we chose to usexhaled tidal volume for making ventilator adjustments.

Because the targeted tidal volume was the same in bothhe VC and the TCPL groups, an explanation for the appar-nt benefit of VC needs to be considered. One answer mightie in the way in which flow (and hence volume) is delivered.

uring TCPL, there is rapid flow delivery, resulting in aharp rise in airway pressure and delivery of volume early inhe inspiratory phase. Theoretically, this should favor thexpansion of the more compliant areas of the lung, possiblyeading to non-homogeneous gas delivery. In VC ventilation,here is a “square wave” flow delivery, resulting in a slower riseut more sustained inspiratory pressure, with peak volumeelivery occurring at end-inspiration. This might result inore uniform filling of the lung and less atelectotrauma. A

urther benefit might accrue from auto weaning. Although aimilar tidal volume target was selected for both groups,hanges in the TCPL group required a clinical decision,hich may not have been performed as rapidly.

An explanation of the differences demonstrated in themaller birth weight cohort probably reflects the overall sickertatus of these infants. The duration of mechanical ventilationn the larger birth weight cohort may have been too short toemonstrate a short-term advantage of VC, which was morepparent in the more premature infants.

Evidence for the importance of volutrauma in theathogenesis of lung injury comes from both animal trials andtudies in human adults.2,8,9 It would appear that the consis-ency of tidal volume delivery during VC ventilation in theace of varying lung compliance and the auto weaning ofirway pressure may be clinically advantageous, especially inonditions in which lung compliance can change rapidly, suchs after surfactant treatment of RDS. In addition to VCentilation, other volume-targeted modalities, such as pres-ure-regulated volume control, volume guarantee, and vol-me-assured pressure support, have also become avail-ble.10-12 Despite different ways to provide a pre-specifiedidal volume, all these modalities, which use different devices,ave in common an objective to control tidal volume delivery

n an attempt to provide optimal lung inflation. Stability ofidal volume delivery may be beneficial, especially in extremely

ow birth weight infants, who are at increased risk of sustain-

The Journal of Pediatrics • September 2006

iTrgtp

btRtsstft

1p2i3v4

Bc5n6C7vi18e9tdN1rdC1eim1a1l

T

M

DCSPASPSPN )

M

ng complications associated with mechanical ventilation.his is further supported by the findings of a recent Cochrane

eview showing a significant beneficial effect of volume-tar-eted ventilation in the newborns, although this was restrictedo only short-term outcomes, such as duration of ventilation,neumothorax, and intraventricular hemorrhage.13

In summary, in this study VC ventilation was found toe both a safe and effective method of ventilating very pre-erm, extremely and very low birth weight newborns withDS. There were trends toward faster weaning, reduction in

he duration of respiratory support, and an improvement inurvival. These findings were even more noticeable in theubgroup of infants who weighed �1000 g at birth. These arehe babies who are most at risk for sustaining complicationsrom mechanical ventilation and may benefit from VC ven-ilation with less ventilator-induced lung injury.

REFERENCES. Attar MA, Donn SM. Mechanisms of ventilator-induced lung injury inremature infants. Semin Neonatol 2002;7:353-60.. Dreyfuss D, Saumon G. Barotrauma is volutrauma, but which volumes the one responsible? Intensive Care Med 1992;18:139-41.. Clark RH, Slutsky AS, Gerstmann DR. Lung protective strategies ofentilation in the neonate: what are they? Pediatrics 2000;105:112-4.

able II. Incidence of complications in the study gro

Outcome measuresVC

(N � 57

ortalityOverall 5600–1000 grams 41001–1500 grams 1uration of ventilation in hrs, mean (95% CI) 255 (160–3LD at 36 weeks PCA 16 (28.1%urvived without CLD at 36 weeks PCA 36 (63.1%neumothorax 2 (3.5%)ny IVH 28 (49.1%evere IVH 5 (8.8%)VL 2 (3.5%)evere IVH or PVL 7 (12.3%DA needing treatment 17 (29.8%EC (�Bell stage II) 6 (10.5%

. Clark RH, Gerstmann DR, Jobe AH, Moffitt ST, Slutsky AS, Yoder S

echanical Ventilation Of Very Low Birth Weight Infants: Is Volume Or

A. Lung injury in neonates: causes, strategies for prevention, and long-termonsequences. J Pediatr 2001;139:478-86.. Donn SM, Sinha SK. Newer modes of mechanical ventilation for theeonate. Curr Opin Pediatr 2001;13:99-103.. Sinha S, Donn S. Volume-controlled ventilation: variations on a theme.lin Perinatol 2001;28:547-60.

. Sinha SK, Donn SM, Gavey J, McCarty M. Randomised trial ofolume controlled versus time cycled, pressure limited ventilation in pretermnfants with respiratory distress syndrome. Arch Dis Child Fetal Neonatal Ed997;77:F202-5.. Dreyfuss D, Saumon G. Ventilator-induced lung injury: lessons fromxperimental studies. Am J Respir Crit Care Med 1998;157:294-323.. Anonymous. Ventilation with lower tidal volumes as compared withraditional tidal volumes for acute lung injury and the acute respiratoryistress syndrome. The Acute Respiratory Distress Syndrome Network.

Engl J Med 2000;342:1301-8.0. Piotrowski A, Sobala W, Kawczynski P. Patient-initiated, pressure-egulated, volume-controlled ventilation compared with intermittent man-atory ventilation in neonates: a prospective, randomised study. Intensiveare Med 1997;23:975-81.

1. Lista G, Colnaghi M, Castoldi F, Condo V, Reali R, Compagnoni G,t al. Impact of targeted-volume ventilation on lung inflammatory responcen preterm infants with respiratory distress syndrome (RDS). Pediatr Pul-

onol 2004;37:510-4.2. Keszler M, Abubakar K. Volume guarantee: stability of tidal volumend incidence of hypocarbia. Pediatr Pulmonol 2004;38:240-5.3. McCallion N, Davis P, Morley C. Volume-targeted versus pressure-imited ventilation in the neonate [review]. The Cochrane Database of

TCPL(N � 52)

Odds ratio(95% CI) P value

10 .5 (0.2–1.2) .191

327 (214–441) 1.1 (0.7–1.6) .617 (32.6%) .9 (0.5–1.5) .625 (48.1%) 1.3 (0.9–1.9) .14 (7.6%) .4 (0.1–2.0) .4

22 (42.3%) 1.2 (0.8–1.8) .55 (9.6%) .9 (0.3–2.8) .95 (9.6%) .4 (0.1–1.6) .2

10 (19.2%) .6 (0.3–1.5) .315 (28.8%) 1.0 (0.6–1.9) .98 (15.4%) .7 (0.3–1.8) .4

ups

)

49)))

)

))

ystematic Reviews 2005, issue 3.

Pressure A Better Target Variable? 313