19 COM TRADU+ç+âO

8/13/2019 19 COM TRADU+ç+âO

http://slidepdf.com/reader/full/19-com-traducao 1/7

Am J Respir Crit Care Med Vol 163. pp 1444–1450, 2001Internet address: www.atsjournals.org

Computed tomography (CT) assessment of positive end-expiratorypressure (PEEP)-induced alveolar recruitment is classically achieved

by quantifying the decrease in nonaerated lung parenchyma on a

single juxtadiaphragmatic section (Gattinoni’s method). This ap-

proach ignores the alveolar recruitment occurring in poorly aer-ated lung areas and may not reflect the alveolar recruitment of the

entire lung. This study describes a new CT method in which PEEP-

induced alveolar recruitment is computed as the volume of gas pen-

etrating in poorly and nonaerated lung regions following PEEP. In

16 patients with acute respiratory distress syndrome a thoracic spi-

ral CT scan was performed in ZEEP and PEEP 15 cm H2O. Accord-ing to the new method, PEEP induced a 119% increase in func-

tional residual capacity (FRC). PEEP-induced alveolar recruitment

was 499 279 ml whereas distension and overdistension of previ-

ously aerated lung areas were 395 382 ml and 28 6 ml, respec-tively. The alveolar recruitment according to Gattinoni’s method

was 26 24 g and no correlation was found between both meth-

ods. A significant correlation was found between PEEP-induced al-

veolar recruitment and increase in Pa

O2

only when recruitment was

assessed by the new method (Rho

0.76, p

0.003), suggestingthat it may be more accurate than Gattinoni’s method.

Thoracic computed tomography (CT) enables an accurateevaluation of the volume of gas and tissue present in the lungsof patients with acute respiratory distress syndrome (ARDS)

(1). Using CT, it is possible to assess the pulmonary distribu-tion of the increase in gas volume resulting from tidal ventila-tion and positive end-expiratory pressure (PEEP) (2) and toseparate PEEP-induced lung overdistension from alveolar re-cruitment (3, 4). In the mid-1990s, Gattinoni and coworkersmeasured PEEP-induced alveolar recruitment on a single jux-tadiaphragmatic CT section by quantifying the decrease in

nonaerated lung parenchyma characterized by CT attenua-tions ranging between

100 and

100 Hounsfield units (HU)(2, 5, 6). This approach ignores the alveolar recruitment occur-ring in poorly aerated lung areas characterized by CT attenua-tions ranging between

100 HU and

500 HU and, therefore,tends to underestimate PEEP-induced alveolar recruitment.In addition, by considering a single CT section, it may under-estimate or overestimate alveolar recruitment occurring in theentire lung. In fact, PEEP-induced alveolar recruitment may

transform nonaerated lung areas into poorly aerated lung ar-eas and poorly aerated lung areas into normally aerated lungareas. Because of these complex gas transfers between the dif-ferent lung compartments, it appears quite difficult to quantifyalveolar recruitment using a global approach. Clinically, PEEP-induced alveolar recruitment can be defined as the volume of gas penetrating in poorly and nonaerated alveolar structures fol-lowing the administration of PEEP, whereas PEEP-inducedalveolar distension can be defined as the volume of gas pene-trating in previously normally aerated alveolar structures. Theaim of the present study was to describe a new CT method forassessing PEEP-induced alveolar recruitment based on this sim-ple definition.

METHODS

Inclusion Criteria

Sixteen consecutive hypoxemic patients with ARDS defined accord-ing to criteria proposed by the American-European Consensus Con-ference (7) were prospectively studied within the first 10 d of theiracute lung disease. Inclusion criteria were Pa

O

2

200 mm Hg at fractionof inspired oxygen (F

IO

2

) 1 and zero end-expiratory pressure (ZEEP)and absence of left ventricular failure defined as a pulmonary capillarywedge pressure (PCWP)

18 mm Hg and/or a left ventricular ejectionfraction

50% as estimated by transesophageal echocardiography.Exclusion criteria were untreated pneumathorax and bronchopleuralfistula at inclusion in the study. Informed consent was obtained fromthe patients’ next of kin.

Cardiorespiratory Measurements and Study Design

During the study period, all patients were sedated and paralyzed witha continuous intravenous infusion of fentanyl, midazolam, and vecu-ronium. Patients were ventilated using controlled mechanical ventila-tion (César Ventilator, Taema, France). For each patient, tidal volumewas limited to values

8 ml/kg and respiratory rate was increased bythe physician in charge in order to achieve Pa

CO

2

values between 30and 50 mm Hg without generating auto-PEEP (8). An inspiratorytime of 33% and F

IO

2

1 were maintained throughout the study period.All patients were monitored using a fiberoptic thermodilution pulmo-nary artery catheter (Baxter Healthcare Co., Irvine, CA) and radial orfemoral arterial catheters.

Cardiorespiratory parameters were measured and recorded as re-cently described (9). Pressure–volume curves of the respiratory sys-tem were measured in ZEEP conditions according to a recently de-scribed technique (10). Inflation, starting and quasistatic respiratory

(

Received in original form April 27, 2000 and in revised form October 17, 2000

)

Dr. Malbouisson was the recipient of a scholarship provided by the French Minis-tery of Foreign Affairs (Ref. 23344471).

The following members of the CT scan ARDS Study Group participated in thisstudy: I. Goldstein, E. Mourgeon, and P. Coriat, Unité de Réanimation Chirurgi-

cale Pierre Viars, Department of Anesthesiology, Hôpital de La Pitié-Salpêtrière,Paris, France; Maria T. Bugalho, Department of Critical Care Medicine, Hospitalde Santarem, Santarem, Portugal; P. Grenier and P. Cluzel, Department of Radi-ology, Hôpital de la Pitié-Salpêtrière, Paris, France; J. Richecoeur, General ICU,Pontoise Hospital, Pontoise, France; P. Gusman, Department of Anesthesiology,Santa Casa, de Misericordia, Juiz de Fora, MG, Brasil; F. Préteux and C. Fetita, In-stitut National des Télécommunications, Evry, France.

Correspondence and requests for reprints should be addressed to J. J . Rouby, Ré-animation Chirurgicale Pierre Viars, Department of Anesthesiology, La Pitié-Salpêtrière Hospital, 47-83 boulevard de l’Hôpital, 75013 Paris, France. E-mail:[email protected]

This article has an online data supplement, which is accessible from this issue’stable of contents online at www.atsjournals.org

Computed Tomography Assessment of PositiveEnd-expiratory Pressure-induced Alveolar Recruitmentin Patients with Acute Respiratory Distress Syndrome

LUIZ M. MALBOUISSON, JEAN-CHARLES MULLER, JEAN-MICHEL CONSTANTIN, QIN LU, LOUIS PUYBASSET,JEAN-JACQUES ROUBY, and the CT Scan ARDS Study Group

Réanimation Chirurgicale Pierre Viars, Department of Anesthesiology, Hôpital de la Pitié-Salpêtrière, University of Paris VI, Paris, France;Department of Anesthesiology, Hospital das Clínicas, Universidade de São Paulo, São Paulo, Brazil; and Department of Anesthesiology,Hospital Gabriel Montpied, Clermont Ferrand, France



Malbouisson, Muller, Constantin, et al.

: Assessment of PEEP-induced Alveolar Recruitment 1445

compliances, and the lower inflection point were determined accord-ing to Gattinoni and coworkers (11). In each patient, a thoracic CT

scan was obtained at end expiration in ZEEP conditions and at aPEEP of 15 cm H

2

O. In our institution, the thoracic CT scan is consid-ered an integral part of the respiratory management of patients withARDS. Hemodynamic and respiratory parameters were measured inZEEP and PEEP conditions within a few hours of the thoracic CTscan in 5 patients and at the same moment in 11 patients.

Computed Tomography Measurement of Lung Volumes

Acquisition of the CT sections

. Each patient was transported to theDepartment of Radiology (Thoracic Division) by two intensivists.Lung scanning was performed from the apex to the diaphragm using aspiral Tomoscan SR 7000 (Philips, Eindhoven, The Netherlands) aspreviously described (1). Acquisition of spiral sections was obtainedat end expiration—at a lung volume equal to functional residual ca-pacity (FRC)—in ZEEP, the patient being disconnected from theventilator (FRC-PEEP), and 15 min after applying a PEEP of 15 cm

H

2

O, the connecting piece between the Y piece and the endotrachealtube being clamped at end expiration (FRC-PEEP). Airway pressurewas monitored during the CT scan acquisition in PEEP to ensure thata pressure of 15 cm H

2

O was actually applied. All CT sections wererecorded on an optical disk for later computerized analysis.

Measurement of lung, gas, and tissue volumes

. The lung volumewas computed as the total number of voxels present in a given region

of interest times the volume of the voxel. The lung is composed of gasand tissue. The respective volumes of gas and tissue were measuredaccording to a previously described analysis (4), based on the tightcorrelation existing between the CT attenuation and the physical den-sity (12). Total lung volume, the volume of gas and tissue, the fractionof gas, and the weight of tissue were measured using a specifically de-signed Software (Lungview) according to a method previously de-scribed (13, 14).

Computed Tomography Measurement of Alveolar Recruitment

Gattinoni’s method

. As recommended by Gattinoni and coworkers(2), alveolar recruitment was assessed on a single CT section located1 cm above the diaphragmatic cupola and computed as the decrease inthe weight of nonaerated lung parenchyma between ZEEP and PEEP(Figure 1). PEEP-induced alveolar recruitment (REC

ALV

) was ex-

pressed as percentage of variation of the weight of the nonaeratedlung parenchyma: REC

ALV

(%)

(W

ZEEP

W

PEEP

)/W

ZEEP

, whereW

ZEEP

is the weight of the nonaerated lung parenchyma in ZEEPconditions and W

PEEP

is the weight of the nonaerated lung paren-chyma in PEEP conditions.

Figure 1. Computed tomography assessment of alveolar recruit-ment according to Gattinoni and coworkers (2). The left upper panels show the juxtadiaphragmatic CT section of a patient with ARDS obtained in ZEEP conditions. The left lower panels show thesame CT section after implementation of PEEP 15 cm H2O. In bothCT sections, lung parenchyma is manually delineated from chestwall and mediastinal structures (dashed lines ). The middle panels show the color encoding analysis applied to the lung parenchymain ZEEP and PEEP conditions. In the color encoding analysis, nor-mally aerated lung parenchyma appears in dark gray, poorly aer-ated lung parenchyma in light gray, and nonaerated lung paren-chyma in white. The right upper and lower panels represent thelung weight distribution according to the CT attenuation in ZEEPand PEEP conditions. The gray areas represents the weight of thenonaerated lung parenchyma. PEEP-induced alveolar recruitmentis computed as the difference of the weight of nonaerated lungparenchyma between ZEEP and PEEP.

Figure 2. Computed tomography assessment of alveolar recruit-ment according to the new method. Image 1 shows the manualdelineation (dashed line ) of the left lung parenchyma on a CT sec-tion in ZEEP. On the delineated left lung parenchyma, a color en-coding analysis is applied (image 2). Normally aerated lung paren-chyma characterized by CT attenuations ranging between 500and 900 Hounsfield units (HU) appears in dark gray, poorly aer-ated lung parenchyma in light gray, and nonaerated lung paren-chyma in white. Image 3 shows the manual delineation (dashed line ) of the poorly and nonaerated lung regions following limits is-sued from the color encoding. Image 4 shows the manual delinea-tion (dashed line ) of the left lung on the same CT section obtainedat a PEEP of 15 cm H2O. Image 5 shows the manual delineationperformed on the CT section obtained in PEEP according to the de-lineation performed in ZEEP (image 3) and referring to anatomicallandmarks such as pulmonary vessels (white arrows ) or segmentalbronchi (black arrows ). During the procedures, both images 3 and 5are displayed on the screen of the computer in order to facilitatethe “transposition” of the delineation performed in ZEEP on the CTsection obtained in PEEP. PEEP-induced alveolar recruitment is defined as the increase in gas volume within the delineated zone be-tween ZEEP (image 3) and PEEP conditions (image 5).



1446

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 163 2001

The new method

. PEEP-induced alveolar recruitment was assessedon the entire lung and defined as the increase in the volume of gaspenetrating in nonaerated and poorly aerated lung regions followingPEEP administration. The CT analysis was performed on each CTsection from the apex to the diaphragm.

In a first step, the following analysis was performed on each CTsection obtained in ZEEP conditions. On the screen of the computeron which the CT image was displayed, the left and right lung bound-aries were manually delineated. Then, normally, poorly and nonaer-ated lung areas were visualized on the screen of the computer using acolor encoding system integrated in the Lungview software. Eachnonaerated voxel characterized by a CT attenuation ranging between

100 and

100 HU was colored in white; each poorly aerated voxelcharacterized by a CT attenuation ranging between

500 and

100HU was colored in light gray; and each normally aerated voxel char-acterized by a CT attenuation ranging between

500 and

900 HUwas colored in dark gray. As shown in Figure 1, the color encodingserved to separate two regions of interest on each CT section: nor-mally aerated lung regions and poorly or nonaerated lung regions. Asshown in Figure 2, the roller ball of the computer was used to delin-eate with a dashed line poorly and nonaerated lung regions by follow-ing the limits given by the color encoding.

In a second step, left and right lung boundaries were manually de-lineated on each CT section obtained in PEEP conditions. Then thetwo CT sections obtained in ZEEP and PEEP conditions and corre-sponding to the same anatomical level were simultaneously displayedon the screen of the computer. By referring to anatomical landmarkssuch as pulmonary vessels, fissures, and segmental bronchi, the limitbetween the two regions of interest was manually redrawn on the CT

section in PEEP according to the previous delineation performed inZEEP (Figure 2). As shown in Figure 3, in patients with diffuse CThyperattenuations the delineation of poorly and nonaerated lung re-gions coincided with the delineation of lung boundaries because of thelack of any normally aerated lung areas in ZEEP conditions. In eachof the two regions of interest delineated in ZEEP and PEEP—nor-mally aerated lung region and poorly and nonaerated lung regions—the volumes of gas and tissue were measured.

PEEP-induced alveolar recruitment was computed as the increase ingas volume within the poorly and nonaerated lung regions (V

Gas

PEEP

V

Gas

ZEEP

) following PEEP administration (alveolar recruitment inabsolute values) divided by the FRC measured in ZEEP conditions(FRC

ZEEP

): REC

ALV

(%)

(V

Gas PEEP

V

Gas ZEEP

)/FRC

ZEEP

. Alveo-lar recruitment was normalized to FRC in order to take into accountthe degree of loss of aeration before PEEP administration. It is obviousthat for a given PEEP-induced alveolar recruitment expressed in ab-

solute values, the physiological impact on gas exchange varies mark-edly according to the initial reduction of FRC. The efficiency of PEEP-induced alveolar recruitment (EF

REC

) expressed as a percentage wascomputed as the ratio between the increase in gas volume withinpoorly and nonaerated lung regions and the increase in FRC on thewhole lung between PEEP and ZEEP: EF

REC

V

Gas PEEP

V

Gas ZEEP

)/(FRC

PEEP

FRC

ZEEP

), or EF

REC

REC

ALV

/(REC

ALV

DIST

ALV

);where V

Gas

ZEEP

and V

Gas

PEEP

are the volumes of gas present withinpoorly and nonaerated lung regions in ZEEP and PEEP conditions,FRC

ZEEP

and FRC

PEEP

are the end-expiratory volumes of gas presentin the entire lung in ZEEP and PEEP conditions, REC is alveolar re-cruitment (in absolute values), and DIST

ALV

is PEEP-induced alveolardistension.

PEEP-induced alveolar distension and overdistension were de-fined as the increase in gas volume within normally aerated lung re-

gions. PEEP-induced alveolar overdistension was defined as the in-crease in lung volume characterized by a CT attentuation rangingbetween

1000 and

900 HU (3). PEEP-induced alveolar distensionwas defined as the increase in lung volume characterized by a CT at-tenuation ranging between

900 and

500 HU.

Statistical Analysis

Because of their normal distribution, respiratory and hemodynamicparameters were compared before and after the administration of PEEP using a paired Student’s t

test. In contrast, the effects of PEEPon lung volumes of gas and tissue were assessed using a Wilcoxon testbecause the distribution of these parameters was not normal. All cor-relations were made by means of Spearman correlation rank analysis.

The correlations between two variables were accepted when the non-parametrical correlation coefficient (Rho) was 0.5. Agreement be-tween both methods aimed at measuring alveolar recruitment wastested using the Bland and Altman method (15). The statistical analy-sis was performed using Statview 5.0 (SAS Institute Inc., Cary, NC).All data are expressed as mean

SD unless otherwise specified. Thesignificance level was fixed at 0.05.

RESULTS

Study Population

Sixteen patients with primary ARDS (4 females and 12 males,mean age 49

18 yr) were studied. ARDS was related topostoperative pulmonary infection (n

7), bronchopulmo-nary aspiration in the postoperative period (n

5), and lung

contusion (n

4). The mean delay between the onset of ARDS and the inclusion in the study was 5

3 d. The lung injury severity score (16) was 3

0.7, the quasistatic respiratorycompliance measured in ZEEP conditions was 39

10 ml/cmH

2

O, the starting compliance was 33

16 ml/cm H

2

O, and theinflation compliance was 46

17 ml/cm H

2

O. A lower inflec-tion point was present in 15 patients, with a mean value of 6.2

3.2 cm H

2

O (extremes 2 to 14 cm H

2

O). Twelve patients hadseptic shock, defined according to reference criteria (17). Allpatients with septic shock were treated with norepinephrine ata mean dose of 3.3

2.9 mg/h. Cardiorespiratory parametersin ZEEP conditions are summarized in Table 1. The overallmortality rate was 37.5%.

PEEP-induced Alveolar Recruitment and Distension

As shown in Table 2, PEEP induced a 119% increase in FRC(p 0.001) without modifying the volume of lung tissue andthe lung weight. The increase in FRC had two components:PEEP-induced alveolar recruitment in previously poorly andnonaerated lung regions and distension or overdistension of previously normally aerated lung regions. PEEP-induced alve-olar recruitment was equivalent to a 56% increase in FRC-ZEEP and represented half of the increase in FRC, the otherhalf being a distension of previously normally aerated lung ar-eas. As shown in Table 3, PEEP induced a 10% decrease inthe tissue volume of the juxtadiaphragmatic CT section.

PEEP-induced alveolar recruitment was assessed either onthe entire lung (new method) or on a single juxtadiaphrag-

Figure 3. Computed tomography assessment of PEEP-induced al-veolar recruitment using the new method in a patient with diffuseloss of aeration. Image 1 shows the manual delineation (dashed line ) of the left lung on a CT section obtained in ZEEP. Accordingto the color encoding analysis (image 2), the left lung is entirelycomposed of poorly aerated (light gray colored areas) andnonaerated lung regions (white colored area). Because of the lackof normally aerated lung regions in ZEEP, PEEP-induced alveolar recruitment can be defined as the increase in gas volume presentin the lung following PEEP (image 3).



Malbouisson, Muller, Constantin, et al.: Assessment of PEEP-induced Alveolar Recruitment 1447

matic CT section (Gattinoni’s method). According to the newmethod, Vgas PEEP Vgas ZEEP was 499 279 ml, representinga 56 20% increase in FRCZEEP, whereas the alveolar recruit-ment was 26 24 g according to Gattinoni’s method, repre-senting a mean decrease of 42 41% in nonaerated tissueweight. As shown in Figure 4, there was neither correlation(Rho 0.273, p 0.18) nor agreement between the twomethods. This result held true when alveolar recruitment was

expressed in absolute values. PEEP-induced lung distensionwas 395 382 ml, representing an increase of 95 40% in thevolume of gas present in normally aerated lung regions inZEEP. As shown in Table 2, PEEP-induced overdistension of the normally aerated parenchyma was 23 67 ml.

Relationships between PEEP-induced Alveolar Recruitment,Arterial Oxygenation, and Pulmonary Hemodynamics

Significant relationships were found between PaO2 and S/ T

measured in ZEEP conditions and FRCZEEP

and more tightlywith the volume of total lung tissue (VTISSUE): PaO2

(mmHg) 0.7 FRCZEEP (L) 77 (Rho 0.5, p 0.05) and S/ T

(%) 0.09 FRCZEEP (L) 49 (Rho 0.5, p 0.05); PaO2

(mm Hg) 0.1 VTISSUE (L) 271 (Rho 0.8, p 0.02)and S/ T (%) 0.02 VTISSUE (L) 18 (Rho 0.6, p 0.03).As shown in Table 1, PEEP induced a significant increase inPaO2 and a significant decrease in S/ T. As shown in Figure5, a significant correlation was found between PEEP-inducedalveolar recruitment and increase in PaO2 only when alveolarrecruitment was assessed using the new method.

In ZEEP, a significant relationship was found between meanpulmonary artery pressure (MPAP) and the volume of nonaer-ated lung parenchyma to overall lung volume/ratio (Rho 0.64,p 0.002). No correlation was found between MPAP and

FRCZEEP or the volume of lung tissue. As shown in Table 1,PEEP significantly increased the pulmonary vascular resis-tance index (PVRI) and central venous pressure and signifi-cantly decreased cardiac index (CI), mean arterial pressure,and oxygen delivery index. No correlations were found betweenPEEP-induced alveolar distension and increase in MPAP orPVRI and decrease in CI.

Factors Influencing the Efficiency of PEEP-inducedAlveolar Recruitment

As shown in Figure 6 the efficiency of PEEP-induced alveolarrecruitment, defined as the ratio between alveolar recruitmentand increase in FRC induced by PEEP, was highly correlatedwith the proportion of poorly and nonaerated lung paren-

chyma in ZEEP conditions (Rho

0.95, p

0.0001). It wasnegatively correlated to static respiratory compliance (Crs)(Rho 0.67, p 0.002) and positively correlated to the valueof the lower inflection point (Rho 0.575, p 0.009).

DISCUSSION

In this study, a new CT method is described for assessingPEEP-induced alveolar recruitment in patients with ARDSand compared with the method proposed by Gattinoni and co-

·Q

·Q

·Q

·Q

·Q

·Q

·Q

·Q

TABLE 2. EFFECTS OF PEEP ON LUNG VOLUMES OF GAS ANDTISSUE AS EVALUATED ON ALL CT SECTIONS

ZEEP PEEP p Value*

Global Analysis Performed on the Entire LungsLung volume, gas tissue, ml 2227 559 3117 818 0.001Functional residual capacity, ml 733 390 1553 740 0.001

Volume of lung tissue, ml 1498 447 1492 463 NSLung weight, g 1513 453 1502 464 NSOverdistended lung volume, ml 1 2 24 64 0.005

Normally aerated lung volume, ml 749 588 1809 1056 0.001Poorly aerated lung volume, ml 704 340 761 555 0.001Nonaerated lung volume, ml 772 375 532 250 0.001

Regional Analysis Performed on Poorly and Nonaerated Lung RegionsLung volume, gas tissue, ml 1490 731 1939 886 0.001Gas volume, ml 262 167 672 370 0.001 Volume of lung tissue, ml 1225 586 1185 588 NS

Lung weight, g 1242 593 1196 624 NS Alveolar recruitment, % 56 20

Regional Analysis Performed on Normally Aerated Lung Regions

Lung volume, gas tissue, ml 737 720 1177 1129 0.003Gas volume, ml 471 466 881 853 0.003 Volume of lung tissue, ml 272 260 306 290 0.03

Lung weight, g 138 134 157 158 0.03

Definition of abbreviations : PEEP positive end-expiratory pressure; ZEEP zero end-expiratory pressure.

* Data are expressed as mean standard deviation. Parameters were compared bymeans of a Wilcoxon test.

TABLE 3. EFFECTS OF PEEP ON LUNG VOLUMES OF GAS ANDTISSUE AS EVALUATED ON A SINGLE JUXTADIAPHRAGMATICCT SECTION

ZEEP PEEP p Value*

Lung volume, gas tissue, ml 196 37 222 55 0.005Gas volume, ml 61 24 104 46 0.001 Volume of lung tissue, ml 135 35 119 35 0.002

Lung weight, g 136 35 126 39 0.02Overdistended lung volume, ml 0 0 6 20 0.01Normally aerated lung volume, ml 60 39 110 61 0.001

Poorly aerated lung volume, ml 64 33 58 42 NSNonaerated lung volume, ml 71 29 48 23 0.001Nonaerated lung weight, g 69 31 45 26 0.005

Definition of abbreviations : PEEP positive end-expiratory pressure; ZEEP zero end-expiratory pressure.

* Data are expressed as mean standard deviation. The variables were compared bymeans of a Wilcoxon test.

TABLE 1. RESPIRATORY AND HEMODYNAMIC PARAMETERS OFPATIENTS MEASURED IN ZEEP AND IN PEEP CONDITIONS

ZEEP PEEP p Value*

PaO2, mm Hg 125 54 224 75 0.001

S/ T, % 42 10 32 8 0.001PaCO2

, mm Hg 44 11 43 9 NSETCO2

, mm Hg 27 7 28 6 NS

DO2I, ml/min/m2 524 176 434 122 0.001 V O2I, ml/min/m2 139 54 118 27 0.03EaO2, % 27 6 28 7 NS

MAP, mm Hg 94 29 78 11 0.03SVRI, dyn/s/cm5/m2 1449 530 1660 593 0.02MPAP, mm Hg 27 7 28 5 NSPVRI, dyn/s/cm5/m2 288 134 351 109 0.01

CVP, mm Hg 9 3 11 3 0.003PCWP, mm Hg 12 3 13 3 NSCI, L/min/m2 4.6 1.8 3.5 1.1 0.005

HR, bpm 102 24 98 23 NSLVSWI, g/min/m2 46 14 38 10 NSRVSWI, g/min/m2 11 6 10 5 NS

Definitions of abbreviations : CI cardiac index; CVP central venous pressure; DO2I oxygen delivery index; EaO2 extraction oxygen ratio; ETCO2 end-tidal carbon diox-ide concentration; HR heart rate; LVSWI left ventricular stroke work index; MAP mean systemic arterial pressure; MPAP mean pulmonary arterial pressure; NS non-significant; PCWP pulmonary capillary wedge pressure; PEEP positive end-expira-tory pressure; PVRI pulmonary vascular resistance index; S/ T shunt fraction;RVSWI right ventricular stroke work index; SVRI systemic vascular resistance index; V O2I oxygen consumption index; ZEEP zero end-expiratory pressure.

* Data are expressed as mean standard deviation. The variables were compared bymeans of a paired Student’s t test.

Q·

Q·

· ·



1448 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 163 2001

workers (2). This new method is based on a regional CT anal-ysis that quantifies the volume of gas penetrating in poorlyand nonaerated lung regions following the administration of PEEP. It requires a CT analysis of the entire lung and also en-ables an assessment of PEEP-induced alveolar distension andoverdistension occurring in normally aerated lung regions. Al-veolar recruitment as assessed by this new method is well cor-

related with PEEP-induced improvement in PaO2.

Comparison of the Two Methods

Neither correlation nor agreement was found between the newmethod and the one proposed by Gattinoni and coworkers in1995 (2). Gattinoni’s method is based on two fundamental as-sumptions: (1) the alveolar recruitment assessed on a single juxtadiaphragmatic CT section is representative of the overallalveolar recruitment occurring in the entire lung; and ( 2) thealveolar recruitment occurring in nonaerated lung areas is rep-resentative of the alveolar recruitment occurring in poorly aer-ated lung regions. Recently, it was shown that in patients withARDS, PEEP-induced alveolar recruitment as assessed byGattinoni’s method (extended to all CT sections) decreases

along a cephalocaudal axis and that in the caudal lung regionsalveolar recruitment is minimum (13). In another study, it waseven shown that an alveolar derecruitment could be observedin diaphragmatic lung regions (1). As a consequence, it ishighly unlikely that PEEP-induced alveolar recruitment as-sessed on a single juxtadiaphragmatic CT section can be repre-sentative of the alveolar recruitment of the entire lung.

Another source of error associated with Gattinoni’s methodresults from the PEEP-induced increase in the cephalocaudal

dimension of the lung. As previously reported (1, 13) the over-all amount of lung tissue was not modified by PEEP in thepresent study whereas a 10% reduction of lung tissue was ob-served in the juxtadiaphragmatic CT section. This paradoxicalresult may be explained by a PEEP-induced cephalocaudalexpansion of the lungs: if the thickness of each CT section re-mains constant, then the amount of tissue per CT section auto-

matically decreases according to the increase in lung size. Inthe present study, to analyze the entire lung, two or three ad-ditional CT sections had to be performed in PEEP as com-pared with ZEEP. As a consequence, a decrease in the weightof the nonaerated lung parenchyma detected on a single juxta-diaphragmatic CT section following PEEP administration doesnot always reflect a true alveolar recruitment. The amount of tissue present on a single basal CT section may also be influ-enced by the redistribution of pulmonary blood flow (18). Ithas recently been reported that PEEP may overdistend upperlobes while recruiting lower lobes (4). As a consequence, pul-monary blood flow may be diverted from upper lobes towardlower lobes resulting in a regional increase in lung tissue thatmay conceal alveolar recruitment-induced decrease in the

weight of nonaerated lung parenchyma.The systematic underevaluation of alveolar recruitment inthe poorly aerated lung regions represents another limitationof the method proposed by Gattinoni. Very likely, PEEP-in-duced alveolar recruitment transforms nonaerated lung re-gions into poorly aerated lung regions and poorly aerated lungregions into normally aerated lung regions. This may explainwhy the overall volume of poorly aerated lung parenchyma re-mained unchanged in the present study confirming the resultsof several previous studies (3, 4, 6). In contrast with Gatti-noni’s method, the new method takes into account the alveo-

Figure 4. Comparison between PEEP-induced alveolar re-cruitment assessed as the increase in gas volume in poorlyand nonaerated lung regions in the entire lung (newmethod) and PEEP-induced alveolar recruitment assessed asthe decrease in the weight of nonaerated lung regions in asingle juxtadiaphragmatic CT section (Gattinoni’s method).Both methods are expressed as percentage of variation fromcontrol values in ZEEP. In the right panel , the agreement be-tween the two methods is represented according to the analysis of Bland and Altman (15).

Figure 5. PEEP-induced changes in PaO2 as a function of PEEP-inducedalveolar recruitment assessed using the new method ( left side ) andGattinoni’s method (right side ). All parameters are expressed as per-centage of variation from control values in ZEEP.

Figure 6. Efficiency of PEEP-induced alveolar

recruitment as a func-tion of the proportionof poorly and nonaer-ated lung parenchymain ZEEP. The efficiencyof PEEP-induced alveo-lar recruitment was as-sessed as the ratio be-tween increase in gasvolume within poorlyand nonaerated lungregions and the in-crease in FRC on thewhole lung betweenPEEP and ZEEP.



Malbouisson, Muller, Constantin, et al.: Assessment of PEEP-induced Alveolar Recruitment 1449

lar recruitment occurring in poorly aerated lung regions. Withthe color encoding system included in Lungview, the lung pa-renchyma can be easily partitioned into two compartments: apotentially recruitable compartment composed of poorly andnonaerated lung regions and an already recruited compart-ment composed of normally aerated lung regions. This re-gional analysis enables PEEP-induced alveolar recruitmentoccurring in the recruitable compartment to be differentiatedfrom PEEP-induced distention occurring in the normally aer-ated compartment.

Limitations of the New Method

The new method is based on a regional analysis of PEEP-induced reaeration of insufficiently aerated lung regions. InZEEP conditions, using a color encoding software it is easy toseparate lung regions characterized by a loss of aeration fromnormally aerated lung parenchyma. By referring to anatomi-cal landmarks such as bronchi and vessels that are visible onlyin the parts of the lung that are normally aerated, the samelung regions can be delineated in PEEP conditions allowingthe assessment of reaeration of insufficiently aerated lung re-gions. However, this analysis cannot be performed separatelyon poorly and nonaerated lung regions because anatomicallandmarks are not visible in the absence of normal pulmonary

aeration. As a consequence, the new method does not permitus to separate PEEP-induced alveolar recruitment occurringin poorly and nonaerated lung regions.

In a limited number of patients, the very heterogeneousdistribution of the loss of aeration within the lungs does notpermit appropriate delineation between normally and insuffi-ciently aerated lung areas in PEEP conditions. As shown inFigure 7, multiple islets of normally aerated lung parenchymacan be surrounded by large areas of poorly aerated lung pa-renchyma rendering the delineation of the same zones of in-terest in ZEEP and PEEP conditions very difficult. This mor-phological pattern, although rarely observed in patients withARDS, precludes any attempt of assessing PEEP-induced al-veolar recruitment using the new method.

Along these lines, another potential limitation of the newmethod is related to the respective dimensions of voxels andalveolar structures (alveolar ducts, alveolar sacs, and alveoli).Recent studies in humans (19) and animals (20) with ARDShave shown that alveolar structures remaining normally aer-ated are distended by mechanical ventilation delivering tidal

volumes 10 ml/kg. The effect of reducing tidal volume on al-veolar dimensions is not known. If the administration of lowtidal volumes does not prevent the distension of alveolarstructures remaining aerated, then the mean alveolar volumeshould be around 0.11 mm3 assuming that a single distendedalveolar structure is a sphere with an internal diameter of 0.6 mm(19). Because in the present study the volume of the voxel wasranging between 1.7 and 4.7 mm3, a single voxel could there-fore contain one to five distended alveolar structures. If theadministration of low tidal volumes— 8 ml/kg in the presentstudy—prevents airspace enlargement, then the voxel couldcontain between 120 and 350 alveolar structures, assumingthat the mean diameter of a normally sized alveolar structureis around 0.3 mm (21–23). As a consequence, a voxel with aCT attenuation of300 HU could contain either 105 normallyand 245 nonaerated alveolar structures or 350 poorly aeratedalveolar structures. In the first situation, the gas penetratingthe voxel following PEEP could distend normally aeratedalveolar structures without recruiting nonaerated alveolarstructures, resulting in an overestimation of PEEP-induced al-veolar recruitment. Although theoretically possible, this possi-bility appears unlikely. Histologically, ARDS is made of largeareas of nonaerated and poorly aerated lung parenchyma—corresponding to alveolar structures totally or partially filled

with edema and/or inflammatory cells—coexisting with nor-mally aerated lung regions. The limit between normally andinsufficiently aerated lung areas is most often well defined andthe histological pattern of multiple islets of normally aeratedalveolar structures scattered within large areas of insufficientlyaerated lung parenchyma is exceptionally observed. As a con-sequence, the error related to the respective dimensions of voxels and alveolar structures concerns mainly the zone of de-lineation between normally and poorly or nonaerated lung pa-renchyma, that is less than 1% of the lung parenchyma ana-lyzed. To increase the accuracy of PEEP-induced alveolarrecruitment assessed by the new method, the size of the voxelshould be reduced as much as possible. Voxels as small as 0.2mm3 can be obtained by using ultrafast spiral CT scanners of

the last generation, which provide a matrix of 1024 1024.

Gas Exchange, Pulmonary Hemodynamics, andAlveolar Recruitment

In the present study, FIO2 1 was used. As a consequence, im-

pairment in arterial oxygenation measured in ZEEP condi-

Figure 7. Representative CT sectionof the patient that was excluded from the study because the newmethod did not allow the quantifi-

cation of PEEP-induced alveolar re-cruitment. A “mosaic” distributionof normally, poorly, and nonaer-ated lung regions characterized theleft lung as seen on the color en-coding analysis (right figure ). As aconsequence the manual delinationperformed in ZEEP could not betransposed in PEEP and PEEP-in-duced alveolar recruitment couldnot be measured.



1450 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 163 2001

tions and improvement in oxygenation parameters observedin PEEP were exclusively dependent on pulmonary shunt(nonaerated lung areas) and not on venous admixture (poorlyaerated lung areas).

In ZEEP conditions, the degree of arterial oxygenation im-pairment correlated with the loss of lung aeration and with thevolume of total lung tissue. With Gattinoni’s method, no cor-relation was found between PEEP-induced alveolar recruit-ment and improvement in arterial oxygenation. Very likely,this lack of correlation was related to the poor accuracy of themethod for assessing PEEP-induced alveolar recruitment oc-curring in shunt regions. In contrast, a significant correlationwas found between PEEP-induced improvement in arterialoxygenation and alveolar recruitment as assessed using thenew method. This result appears logical. In ARDS, arterialhypoxemia is essentially related to the presence of nonaeratedlung regions remaining perfused. PEEP basically improvesarterial oxygenation by reestablishing some degree of aerationin these shunt regions. By founding the new method on theassessment of the reaeration of insufficiently aerated lung re-gions, it is not surprising to have found a significant correla-tion between PEEP-induced alveolar recruitment and improve-ment in arterial oxygenation.

In this study, a good correlation was found between the ini-

tial loss of pulmonary aeration and the efficiency of PEEP-induced alveolar recruitment. In fact, it has been shown thatthe distribution of PEEP-induced increase in gas volumewithin the lung parenchyma is markedly dependent on dispar-ities in regional lung compliances (13), which, in turn, are de-termined by the regional loss of gas (1). Therefore, it appearslogical that the greater the extension of poorly and nonaer-ated lung regions in ZEEP, the lower the disparities in re-gional lung compliances and the greater the efficiency of PEEP-induced alveolar recruitment. In addition, the efficiencyof PEEP-induced alveolar recruitment was positively corre-lated to the lower inflection point. These results agree with theprevious findings of Vieira and coworkers (4). According tothese authors, the patients, whose ARDS is characterized by

diffuse lung CT attenuations, a marked lower inflection pointon the lung P–V curve, and a low respiratory compliance, re-spond to increasing PEEP levels by a progressive and sus-tained alveolar recruitment. In contrast, the patients, whoseARDS is characterized by lobar lung CT attenuations involv-ing predominantly the lower lobes, the absence of a lower in-flection point on the lung P–V curve and a moderate decreasein respiratory compliance respond to increasing PEEP levelsby an overdistension of previously aerated lung regions.

In conclusion, the alveolar recruitment defined as the pen-etration of gas within poorly and nonaerated lung regions fol-lowing the administration of PEEP can be quantified using aCT analysis of the entire lung. In contrast with previous CTmethods, it takes into account the alveolar recruitment occur-ring in poorly and nonaerated lung regions, quantifies PEEP-induced alveolar distension occurring in normally aerated lungregions, and is correlated with PEEP-induced improvement inarterial oxygenation. Further studies are required to deter-mine whether this CT method is correlated with methodsbased on the determination of alveolar recruitment from P–Vcurves obtained in ZEEP and PEEP conditions.

References

1. Puybasset L, Cluzel P, Chao N, Slutsky AS, Coriat P, Rouby JJ, and theCT Scan ARDS Study Group. A computed tomography scan assess-ment of regional lung volume in acute lung injury. Am J Respir Crit

Care Med 1998;158:1644–1655.2. Gattinoni L, Pelosi P, Crotti S, Valenza F. Effects of positive end-expira-

tory pressure on regional distribution of tidal volume and recruitmentin adult respiratory distress syndrome. Am J Respir Crit Care Med

1995;151:1807–1814.3. Vieira SR, Puybasset L, Richecoeur J, Lu Q, Cluzel P, Gusman PB, Co-

riat P, Rouby JJ. A lung computed tomographic assessment of posi-tive end-expiratory pressure-induced lung overdistension. Am J

Respir Crit Care Med 1998;158:1571–1577.4. Vieira SR, Puybasset L, Lu Q, Richecoeur J, Cluzel P, Coriat P, Rouby

JJ. A scanographic assessment of pulmonary morphology in acutelung injury: significance of the lower inflection point detected on thelung pressure-volume curve. Am J Respir Crit Care Med 1999;159:

1612–1623.5. Gattinoni L, D’Andrea L, Pelosi P, Vitale G, Pesenti A, Fumagalli R. Re-

gional effects and mechanism of positive end-expiratory pressure inearly adult respiratory distress syndrome. JAMA 1993;269:2122–2127.

6. Gattinoni L, Pesenti A, Bombino M, Baglioni S, Rivolta M, Rossi F,Rossi G, Fumagalli R, Marcolin R, Mascheroni D, Torresin A. Rela-tionships between lung computed tomographic density, gas exchange,and PEEP in acute respiratory failure. Anesthesiology 1988;69:824–832.

7. Bernard GR, Artigas A, Brigham KL, Carlet J, Falke K, Hudson L,Lamy M, Legall JR, Morris A, Spragg R. The American-EuropeanConsensus Conference on ARDS: definitions, mechanisms, relevantoutcomes, and clinical trial coordination. Am J Respir Crit Care Med

1994;149:818–824.8. Richecoeur J, Lu Q, Vieira SR, Puybasset L, Kalfon P, Coriat P, Rouby

JJ. Expiratory washout versus optimization of mechanical ventilationduring permissive hypercapnia in patients with severe acute respira-tory distress syndrome. Am J Respir Crit Care Med 1999;160:77–85.

9. Rouby JJ, Puybasset L, Cluzel P, Richecoeur J, Lu Q, Grenier P and theCT Scan ARDS Study Group. Regional distribution of gas and tissuein acute respiratory distress syndrome—part 2: Physiological correla-tions and definition of an ARDS severity score. Intensive Care Med

2000;26:1046–1056.10. Lu Q, Vieira SR, Richecoeur J, Puybasset L, Kalfon P, Coriat P, Rouby

JJ. A simple automated method for measuring pressure-volume curvesduring mechanical ventilation. Am J Respir Crit Care Med 1999;159:275–282.

11. Gattinoni L, Pesenti A, Avalli L, Rossi F, Bombino M. Pressure-volumecurve of total respiratory system in acute respiratory failure: com-puted tomographic scan study. Am Rev Respir Dis 1987;136:730–736.

12. Mull RT. Mass estimates by computed tomography: physical densityfrom CT numbers. Am J Padiol 1984;143:1101–1104.

13. Puybasset L, Muller JC, Cluzel P, Coriat P, Rouby JJ, and the CT ScanARDS Study Group. Regional distribution of gas and tissue in acute res-piratory distress syndrome—part 3: consequences on the effects of posi-tive end-expiratory pressure. Intensive Care Med 2000;26:1215–1227.

14. Malbouisson LM, Busch CJ, Puybasset L, Lu Q, Cluzel P, Rouby JJ, andthe CT Scan ARDS Study Group. Role of the heart in the loss of aer-ation characterizing lower lobes in acute respiratory distress syn-drome. Am J Respir Crit Care Med 2000;161:2005–2012.

15. Bland JM, Altman DG. Statistical methods for assessing agreement be-tween two methods of clinical measurement. Lancet 1986;1:307–310.

16. Murray JF, Matthay MA, Luce JM, Flick MR. An expanded definitionof the adult respiratory distress syndrome. Am Rev Respir Dis 1988;138:720–723.

17. American College of Chest Physicians/Society of Critical Care MedicineConsensus Conference. Definitions for sepsis and organ failure andguidelines for the use of innovative therapies in sepsis. Crit Care Med

1992;20:864–874.18. Schuster DP, Howard DK. The effect of positive end-expiratory pres-

sure on regional pulmonary perfusion during acute lung injury. J Crit

Care 1994;9:100–110.19. Rouby JJ, Lherm T, Martin de Lassale E, Poete P, Bodin L, Finet JF,Callard P, Viars P. Histologic aspects of pulmonary barotrauma incritically ill patients with acute respiratory failure [see comments]. In-

tensive Care Med 1993;19:383–389.20. Goldstein I, Bughalo MT, Marquette CH, Lenaour G, Lu Q, Rouby JJ

and the Experimental ICU Study Group. Mechanical ventilation-inducedair-space enlargement during experimental pneumonia in piglets. Am

J Crit Care Med 2001;163:958–964.21. Thurlbeck WM. The internal surface area of nonemphysematous lungs.

Am Rev Respir Dis 1967;95:765–773.22. Weibel ER. Morphometry of the human lung. Berlin: Springer-Verlag;

1963.23. Gehr P, Bachofen M, Weibel ER. The normal human lung: ultrastruc-

tural and morphometric estimation of diffusion capacity. Respir Phys-

iol 1978;32:121–129.

Documents

19 COM TRADU+ç+âO