TIDAL HYPERINFLATION DURING LOW TIDAL
VOLUME VENTILATION IN ACUTE RESPIRATORY
Pier Paolo Terragni 1, Giulio Rosboch 1, Andrea Tealdi 1, Eleonora Corno 1,
Eleonora Menaldo 1, Ottavio Davini 2, Giovanni Gandini 2, Peter Herrmann 3,
Luciana Mascia 1, Michel Quintel 3, Arthur S. Slutsky 4, Luciano Gattinoni 5,
and V. Marco Ranieri 1
Word count: 2,985
Subject category list: 2. ARDS and ALI: patient studies
Running head: Hyperinflation and “ARDSnet” strategy
Address for correspondence and reprint requests:
V. Marco Ranieri, MD
Università di Torino
Dipartimento di Anestesiologia e Rianimazione
Ospedale S. Giovanni Battista-Molinette
Corso Dogliotti 14, 10126 Torino
Supported by Ministero Università e Ricerca (grant # PR60ANRA04 ) and Regione Piemonte (grant #
Dipartimento di Anestesiologia e Rianimazione, Università di Torino, Ospedale S. Giovanni Battista-
Molinette, Torino, Italy.
Dipartimento di Radiologia, Università di Torino, Ospedale S. Giovanni Battista-Molinette, Torino, Italy.
Department of Anesthesiology, University of Gottingen, Gottingen, Germany
Interdepartmental Division of Critical Care, Division of Respiratory Medicine, University of Toronto, St.
Michael’s Hospital, Toronto, ON, Canada
Istituto di Anestesia e Rianimazione, Fondazione IRCCS – “Ospedale Maggiore Policlinico, Mangiagalli,
Regina Elena” di Milano; Università degli Studi di Milano, Italy
This article has online data supplement, which is accessible from this issue’s table of content online at
AJRCCM Articles in Press. Published on October 12, 2006 as doi:10.1164/rccm.200607-915OC
Copyright (C) 2006 by the American Thoracic Society.
Abstract (worlds count: 240)
Rationale: Tidal volume and plateau pressure limitation decreases mortality in ARDS.
Computed tomography demonstrated a small normally aerated compartment on the top
of poorly aerated, and non-aerated compartments that may be hyperinflated by tidal
Objectives: We hypothesized that despite tidal volume and plateau pressure limitation,
patients with larger non-aerated compartment are exposed to tidal hyperinflation of the
normally aerated compartment.
Measurements and Main Results: Pulmonary computed tomography at end-expiration
and end-inspiration was obtained in 30 patients ventilated with a low tidal volume (6
ml/kg predicted body weight). Cluster analysis identified 20 patients where tidal
inflation largely occurred in the normally aerated compartment (69.9 ± 6.9 %; “more
protected”), and 10 patients where tidal inflation largely occurred within the
hyperinflated compartments (63.0 ± 12.7 %; “less protected” ). The non-aerated
compartment was smaller and the normally aerated compartment larger in the “more
protected” than in “less protected” (P = 0.01). Pulmonary cytokines were lower in the
“more protected” than in the “less protected” (P < 0.05). Ventilator free days were 7 ± 8
and 1 ± 2 in the “more protected” and “less protected”, respectively (P = 0.01). Plateau
pressure in “more protected” ranged between 25 and 26 cm H2O and in “less protected”
between 28 and 30 cm H2O (P = 0.006).
Conclusions: Limiting tidal volume to 6 ml/kg PBW and plateau pressure to 30 cm H2O
may not be sufficient in patients characterized by a larger non-aerated compartment.
Keywords: Acute Lung Injury, inflammatory response, mechanical ventilation, VILI
The acute respiratory distress syndrome (ARDS) is the inflammatory response of the
lungs to direct or indirect insults. It is clinically characterized by sudden onset, severe
hypoxemia, radiographic evidence of bilateral pulmonary infiltration, and absence of left-heart
failure (1). Mechanical ventilation, the main supportive therapy which is used to maintain
adequate oxygenation, may lead to the activation of inflammatory processes and may augment
or produce a pulmonary damage that is indistinguishable from that caused by the underlying
disease process (ventilator induced lung injury: VILI) (2). A multi-center, randomized clinical
trial conducted by the ARDS Network (“ARDSnet”) demonstrated that a ventilatory strategy
using a tidal volume (VT) of 6 mL/kg predicted body weight (PBW) decreased mortality by
22% compared to a strategy using a VT of 12 mL/kg PBW (3). A recent observational study
confirmed that use of a VT higher than 6 mL/kg PBW was independently associated with a
worse outcome from ARDS (4).
Analysis of computed tomography (CT) images of patients with ARDS has demonstrated
a non-homogeneous distribution of pulmonary alterations grouped into 4 patterns:
hyperinflated, normally aerated, poorly aerated, and non-aerated compartments interspersed
and/or distributed along the ventral-dorsal axis (5-7). The normally aerated compartment is
relatively small but receives the largest part of the tidal volume (5, 6) and may therefore be
exposed to excessive alveolar wall tension and stress failure (8); the non-aerated compartment
can be re-aerated during ventilation and the tidal re-aeration of alveoli adjacent to fully
expanded and consolidated regions may therefore cause shear stress (8). Nieszkowska and
coworkers (9) found that in 14 out of 32 ARDS patients, prevention of expiratory
derecruitment with 15 cm H2O of positive end-expiratory pressure (PEEP) was obtained at the
price of the hyperinflation of the normally aerated compartment. More recently, Gattinoni and
coworkers provided direct visual evidence that patients with greater non-aerated and smaller
normally aerated compartments had a worse outcome than patients with smaller non-aerated
and greater normally aerated compartments (10).
The present study set out to examine the hypothesis that patients characterized by a CT
scan distribution of pulmonary lesions with a large dependent non-aerated compartment and a
small non-dependent normally aerated lung compartment may be exposed to tidal
hyperinflation despite the use of the “ARDSnet” protective ventilatory strategy.
Some of the results of these studies have been previously reported in the form of an
Methods (words count: 586)
Inclusion criteria were: age ≥ 18 years; diagnosis of ARDS (3, 12). Exclusion criteria
were: >3 days had elapsed since ARDS criteria were met and mechanical ventilation was
initiated; pulmonary artery occlusion pressure > 18 mm Hg, if measured; history of ventricular
fibrillation or tachyarrhythmia, unstable angina or myocardial infarction within preceding
month; pre-existing chronic obstructive pulmonary disease (3); major chest wall abnormalities
(2); chest tube with persistent air leak; abdominal distension (2); body mass index >30;
pregnancy; known intracranial abnormality; enrollment in another interventional study (3, 12).
The institutional review board approved the study (10).
Patients were ventilated according to the "ARDSNet" protective ventilatory strategy (3,
12). As soon as targets of the ventilatory protocols were reached and physiologic parameters
were stable (10), patients were transferred to the CT scan facility. Lung scanning was
performed from apex to base at end-expiratory and end-inspiratory occlusions (8, 10). During
the transport and the exam, ventilator settings and the ventilator itself were the one used for the
clinical management; particular attention was paid to avoid ventilator disconnection.
The CT scanner was set as previously described (8, 10); the non-aerated (+100 and –100
Hounsfield units), poorly aerated (–101 and –500 Hounsfield units), normally aerated (–501
and –900 Hounsfield units), and hyperinflated (–901 and –1000 Hounsfield units) lung
compartments were identified (6, 13). The volume of each compartment (i.e. the sum of gas
plus tissue volume) for each slice, as well as the volume of the entire lung was measured at
end-expiration and end-inspiration (8, 10).
“Protected tidal inflation” and “tidal hyperinflation” were defined as the volume of the
normally aerated and hyperinflated compartment at end-inspiration minus the volume of the
normally aerated and hyperinflated compartment at end-expiration, respectively. “Tidal
recruitment of the non-aerated compartment” and “tidal recruitment of the poorly aerated
compartment” were defined as the volume of the non-aerated and of the poorly aerated
compartments at end-expiration minus the volume at end-inspiration. All were expressed as %
of the total tidal inflation-related change in CT lung volume (8).
Weight of the entire lung and of each compartment at end-inspiration was measured (8,
Five-ten min after CT measurements, a broncho-alveolar lavage was collected in the CT
suite and stored as previously described (2). Tumor necrosis factor-α soluble (TNF-αsR55 and
TNF-αsR75), interleukin 6 (IL-6), interleukin 8 (IL-8) and interleukin 1β (IL-1β) and IL-1
receptor antagonist (IL-1Ra) were measured (2).
Number of ventilator free days and the number of patients alive at 28 days immediately
after study entry was calculated (3, 12).
Values are given as mean ± SD. Cluster analysis and the cubic clustering criterion were
used to identify the maximal degree of association between patients and “ protected tidal
inflation”, “tidal hyperinflation”, “tidal recruitment of the non-aerated compartment” and
“tidal recruitment of the poorly aerated compartment”. Cluster analysis entails grouping
similar objects into distinct, mutually exclusive subsets referred to as clusters; elements within
a cluster share a high degree of “natural association,” whereas the clusters are relatively
distinct from one another (14-16). Two-tailed t test, Mann-Whitney U test, Χ2 test, and Fisher
exact test were used. Multivariate stepwise regression analysis with backward elimination was
used to explain whether differences in cytokines concentration and number of ventilator free
days between clusters were related to disease severity and/or to VILI. Dependent variables
[clinical (4) and CT (6, 13) markers of disease severity and CT markers of VILI (8)] were
entered in the regression if statistically different between clusters (4) (SAS Institute, Cary,
Clinical characteristics of the study subjects are shown in Table 1. Ventilator free days
and mortality rate at 28 days were 5 ± 7 days, and 33%, respectively. The time between the
onset of ALI/ARDS and the study varied from 1 to 3 days.
Amount of “protected tidal inflation” and “tidal hyperinflation” best discriminated two
clusters of patients (R2 = 0.73). In a cluster of 20 patients “ protected tidal inflation” and
“tidal hyperinflation” represented the 69.9 ± 6.9 and the 8.1 ± 5.4 % of the total tidal-inflation
associated change in CT lung compartments, respectively (”more protected”). In a second
cluster of 10 patients, “protected tidal inflation” and “tidal hyperinflation” were the 23.1 ±
14.4 and the 63.0 ± 12.7 % of the total tidal-inflation associated change in CT lung
compartments, respectively (”less protected”) (Figure 1). “Tidal recruitment of the poorly
aerated compartment” and “tidal recruitment of the non-aerated compartment” were the
12.6 ± 4.7 and 9.3 ± 5.7 and the 6.7 ± 4.3 and 7.1 ± 6.1 % of the total tidal change in CT lung
compartments in the ”more protected” and ”less protected”, respectively.
Representative CT slices of the lung obtained 2 cm above the dome of the diaphragm at
end-expiration and end-inspiration from a “more protected” (Figure 2 A, left) and a “less
protected” patient (Figure 2 B, left) are shown. Lung density histograms of tidal inflation-
related changes in CT lung compartments in “more protected” (Figure 2 A, right) showed an
increase of volume in the normally aerated compartment with a peak at -810 HU. In “less
protected”, tidal inflation reduced volume in the normally aerated compartment, with an
increased volume of the hyperinflated compartment, with a peak at -910 HU (Figure 2 B,
With the exception of IL-1Ra, BAL concentration of Il-6, IL-1β, IL-8 and of both TNF-α
receptors were lower in “more protected” than in “less protected” (P < 0.05) (Figure 3).
Number of ventilator free days in “more protected” was higher (P = 0.01) than in “less
protected” (7 ± 8 vs. 1 ± 2, respectively). Mortality rates at 28 days from admission were 30 %
and 40 % in “more protected” and “less protected”, respectively (P = 0.21). The amount of
“tidal hyperinflation” correlated with the pulmonary concentration of all inflammatory
cytokines (P < 0.01).
Clinical characteristics of “more protected” and “less protected” are shown in Table 1.
Age, gender, SAPS II and underlying diseases responsible for ARDS did not differ between the
two groups of patients; PaO2/FiO2 ratio was higher in “more protected” than in “less
protected” (P = 0.009). PPLAT in “more protected” ranged between 25 and 26 cm H2O and in
“less protected” between 28 and 30 cm H2O (P = 0.006).
End-inspiratory weight and volume of the total lung and of the different CT lung
compartments in the overall population and in “more protected” and “less protected” are
shown in Table 2. Lungs were heavier in “less protected” than in “more protected” (P =
0.008); weight and volume of the hyperinflated and non-aerated CT lung compartment were
higher, and of normally aerated lower in “less protected” than in “more protected” (P <
An amount of “tidal hyperinflation” > 40% of the tidal inflation-associated change in CT
lung compartment identified the ”less protected” patients and corresponded to a PPLAT value ≥
28 cm H2O (Figure 4).
Dependent variables entered into the multivariate stepwise regression analysis were
weight of the entire lung (6, 13), PaO2/FiO2 ratio, PPLAT (4) amount of “protected tidal
inflation” and “tidal hyperinflation” (8). Tidal recruitment of non-aerated and poorly aerated
compartments were not included in the model because they were not selected as differentiating
characteristics between clusters. Since total lung weight and PPLAT correlate with disease
severity consistently to weights of non-aerated, normally aerated and hyperinflated
compartments (6, 13) and to minute ventilation (4) respectively, the latter were not included in
the regression analysis although they differ in the two groups of patients.
“Tidal hyperinflation” was the only variable independently associated with concentration
of IL-6 (P = 0.001), IL-1β (P = 0.0025), IL-8 (P = 0.001), TNF-αsR55 (P = 0.007) and TNF-
αsR75 (P = 0.001) and number of ventilator free days (P = 0.005).
The present study demonstrates that the “ARDSnet” strategy may not be protective in all
patients with ARDS since: (a) 1/3 of patients experienced substantial “tidal hyperinflation”
with tidal volumes of 6 ml/kg PBW and PPLAT lower than 30 cm H2O; in these patients
concentration of inflammatory mediators was higher and number of ventilator free days was
lower than in the 2/3 of patients that experienced less (although not zero) “tidal
hyperinflation”. (b) Values of PPLAT lower than 28 cm H2O are associated with less tidal
hyperinflation than values of PPLAT ranging between 28 and 30 cm H2O. (3) Although our data
do not indicate that a “safe” limit of PPLAT exists, values less than 28 cm H2O seem to be
associated to the “more protective” ventilatory settings.
Crucial for interpretation of these results is to clarify whether the higher concentration of
inflammatory mediators and the lower number of ventilator free days seen in “less protected”
patients may be simply due to more severe underlying lung injury. Multivariate stepwise linear
regression analysis with backward elimination showed that the amount of “tidal
hyperinflation” was the only variable associated with cytokines concentration and number of
ventilator free days. We may therefore speculate that in patients with heavier lungs, a larger
dependent non-aerated compartment and a smaller non-dependent normally aerated
compartment [i.e. lungs characterized by a high “potential for recruitment” (10) and a small
“baby lung” (17)] the “ARDSnet” protective ventilatory strategy does not fully protect the
lungs from VILI since hyperinflation of the small “normal” lung may occur despite lowering
VT to 6 ml/kg PBW and limiting PPLAT to 30 cm H2O.
Before discussion of these results, some considerations are required. First, although our
analysis separated a first cluster of patients characterized by a predominant “protected tidal
inflation” from a second cluster of patients characterized by a predominant “tidal
hyperinflation” (Figure 1), these two clusters represent different ranges of a continuum since
most ARDS patients experience tidal overdistention in some regions while tidal recruitment
and increased normal aeration occur simultaneously in other regions (13, 18). Second, the lack
of a CT scan at zero end-expiratory pressure does not allow to identify patients at risk of tidal
hyperinflation since at the PEEP levels used in the present study, lung morphology is
influenced by factors not related to the kind of lung injury, the most important being the
potential for recruitment (10). Third, the CT scan thickness used in the present study was 5
mm. Such spatial resolution may expose to a significant underestimation of tidal hyperinflation
since Viera and coworkers recently demonstrated that an higher spatial resolution (2 mm) may
have provided a more accurate measurement (19). Fourth, patients’ age ranged between 49
and 82 yrs. As a consequence it is likely that some degree of hyperinflation was present in
some patients since it is well known that lung emphysema is related to age (20). Fifth, tidal
recruitment of non-aerated and poorly aerated compartments was relatively small in both
clusters. This might indicate that with tidal volumes of 6 ml/kg PBW, pressure and volume
excursions are small enough that tidal recruitment/derecruitment is not significant i.e. that low
tidal volume ventilation, intended to reduce tidal hyperinflation, may also minimize tidal
ARDS is morphologically characterized by the distribution of the loss of lung aeration
along the vertical axis with a small number of normal alveoli located in the non-dependent lung
and a large consolidated, non-aerated region located in the dependent lung (6, 10, 21-23).
Analysis of pulmonary CT images of patients (10, 18, 24) and animals (8, 25, 26) with ARDS
during mechanical ventilation have demonstrated that the normally aerated compartment may
receive the largest part of each breath and may therefore be hyperinflated and exposed to
excessive alveolar wall tension and stress failure. Insufficient levels of PEEP may cause tidal
recruitment de-recruitment of parts of the consolidated region and may therefore expose these
regions to shear stress (8). These events may lead to worsening of the pulmonary and systemic
inflammatory response, distal organ dysfunction, and ultimately organ failure (27). The
“ARDSnet” study demonstrated that a 22% reduction in mortality could be obtained using a VT
of 6 mL/kg PBW instead of 12 mL/kg PBW. In that study, the mean PPLAT on the first day was
25 ± 7 cm H2O in the 6 mL/kg group vs. 33 ± 9 cm H2O in the 12 mL/kg group (3).
Controversy exists regarding the extent to which VT and inspiratory airway pressures
should be reduced to protect the lungs from VILI (28-31). Some investigators have
recommended that inspiratory plateau pressures lower than 30 to 35 cm H2O may be
considered safe and that further reductions in VT and PPLAT are without benefit (28, 29, 31).
Hager and coworkers recently examined the data collected in the “ARDSnet” study and found
that mortality decreases as PPLAT declined from high to low levels at all levels of as PPLAT (32).
These data suggest that patients in the higher VT group would have benefited from VT
reduction even if they already had PPLAT < 30 cm H2O (32). In the present study all patients
were ventilated according to the low VT arm of the “ARDSnet” study; pulmonary concentration
of inflammatory cytokines was higher and number of ventilator free days was smaller in
patients who had a PPLAT ≥ 28 cm H2O than in patients who had a PPLAT ≤ 26 cm H2O.
Prospective clinical trials are required to prove that ventilation at the lower PPLAT would have
improved outcomes in those patients ventilated with PPLAT ranging between 28 and 30 cm H2O.
Lung hyperinflation has been previously reported as resulting from mechanical
ventilation with PEEP (7, 9, 19, 21). Nieszkowska and coworkers (9) found that in 32 ARDS
patients, expiratory de-recruitment was prevented by maintaining a PEEP of 15 cm H2O.
However, the “price” of this beneficial effect of PEEP in 1/3 of the patients was hyperinflation
of the non-dependent lung regions. These patients had a CT scan distribution of pulmonary
lesions characterized by a large amount of non-aerated and poorly aerated lung distributed in
the dependent regions and a small amount of normally aerated lung distributed in the non-
dependent regions (9), similar to those observed in our “less protected” patients. Under these
circumstances is very likely that what has been repeatedly reported for PEEP-induced
hyperinflation is also true for tidal inflation-related hyperinflation: patients with a focal loss of
lung aeration at zero end-expiratory pressure are at higher risk of hyperinflation than patients
with a diffuse loss of lung aeration (7, 9, 19, 21).
Information regarding the effects of tidal inflation on hyperinflation of lung regions in
patients with ARDS is limited. Crotti and coworkers (33) quantified the amount of
hyperinflated lung in 5 ARDS patients ventilated in pressure control mode at PPLAT of 30 cm
H2O and at different levels of PEEP and found that hyperinflated lung tissue ranged between 1
and 5 % of the whole lung tissue. In our study, the end-inspiratory weight of hyperinflated lung
tissue ranged between 0.1 and 3.9 % of the total lung weight in “more protected” and between
the 1.5 and 8.7 % of the total lung weight in “less protected” pattern (P = 0.01).
In conclusion, the present results may confirm the notion that the best ventilatory strategy
should be ideally adapted to the size of the aerated lung. The “ARDSnet” protocol limiting VT
to 6 ml/kg PBW and limiting PPLAT to 30 cm H2O may therefore not be sufficient to minimize
VILI in patients with ARDS whose disease process is characterized by a distribution of
pulmonary lesions with a small non-dependent normally aerated compartment and a large
dependent non-aerated compartment.
16 Download full-text
1. Ware, L. B., and M. A. Matthay. 2000. The acute respiratory distress syndrome. N Engl
J Med 342(18):1334-49.
2. Ranieri, V. M., P. M. Suter, C. Tortorella, R. De Tullio, J. M. Dayer, A. Brienza, F.
Bruno, and A. S. Slutsky. 1999. Effect of mechanical ventilation on inflammatory
mediators in patients with acute respiratory distress syndrome: a randomized controlled
trial. Jama 282(1):54-61.
3. 2000. Ventilation with lower tidal volumes as compared with traditional tidal volumes
for acute lung injury and the acute respiratory distress syndrome. The Acute
Respiratory Distress Syndrome Network. N Engl J Med 342(18):1301-8.
4. Sakr, Y., J. L. Vincent, K. Reinhart, J. Groeneveld, A. Michalopoulos, C. L. Sprung, A.
Artigas, and V. M. Ranieri. 2005. High tidal volume and positive fluid balance are
associated with worse outcome in acute lung injury. Chest 128(5):3098-108.
5. Vieira, S. R., L. Puybasset, Q. Lu, J. Richecoeur, P. Cluzel, P. Coriat, and J. J. Rouby.
1999. A scanographic assessment of pulmonary morphology in acute lung injury.
Significance of the lower inflection point detected on the lung pressure-volume curve.
Am J Respir Crit Care Med 159(5 Pt 1):1612-23.