JOBNAME: cocc 11#1 2005 PAGE: 1 OUTPUT: Wed December 29 20:44:40 2004

lww/cocc/91165/COCC11110
Prod#: COCC11110

How to ventilate patients with acute lung injury

and acute respiratory distress syndrome
Luciano Gattinoni, Pietro Caironi and Eleonora Carlesso

Purpose of review Abbreviations

The purpose of this paper is to review the mechanisms of ALI acute lung injury
ventilator-induced lung injury as a basis for providing the ARDS acute respiratory distress syndrome
bw body weight
less damaging mechanical ventilation in patients with acute EELV end-expiratory lung volume
respiratory failure. IL interleukin
NF-kB nuclear factor k-B
Recent findings LPS lipopolysaccharide
In normal lungs, high tidal volume causes an immediate NIH National Institutes of Health
PEEP positive end-expiratory pressure
gene upregulation and downregulation. Although the VILI ventilator-induced lung injury
importance of alveolar inflammatory reaction is well known, VT tidal volume

recent findings suggest the potential role of airway

distension in causing ventilator-induced lung injury. The
initial activation has been shown to occur in the airways,
accounting for the damages induced by high peak flow. The
healthier lung regions are more exposed to the injury, since
they may be subjected to strain. Challenge with endotoxin
enhances in a synergistic manner the pulmonary
inflammation induced by mechanical ventilation. However,
mechanical strain and endotoxin seem to trigger lung
inflammation through two different pathways. Despite
convincing experimental and clinical evidences of lung
injury, the clinical implementation of low tidal volume
ventilation is still limited and has not yet become part of
standard clinical practice. Setting positive end-expiratory
pressure remains an open problem because the ALVEOLI
study did not provide any exhaustive answers, likely because
of methodologic problems and, unphysiologic design.
Summary
Gentle lung ventilation must be standard practice. Because
stress and strain are the triggers of ventilator-induced lung
injury, their clinical equivalents should be measured
(transpulmonary pressure and the ratio between tidal
volume and end-expiratory lung volume). For a rational
application of positive end-expiratory pressure, the potential
for recruitment in any single patient should be estimated.
Keywords
ventilator-induced lung injury, mechanical ventilation, acute
lung injury, acute respiratory distress syndrome, cytokines,
lung recruitment
Curr Opin Crit Care 11:69–76. ª 2005 Lippincott Williams & Wilkins.
Istituto di Anestesia e Rianimazione, Ospedale Maggiore di Milano – IRCCS,
Università degli Studi di Milano, Milano, Italy.
Correspondence to Luciano Gattinoni, Istituto di Anestesia e Rianimazione,
Ospedale Maggiore Policlinico-IRCCS, Via Francesco Sforza, 35 20122 Milano, Italy
Tel: +39 02 55033232/3231; fax: +39 02 55033230;
e-mail: gattinon@policlinico.mi.it
Current Opinion in Critical Care 2005, 11:69–76
ª 2005 Lippincott Williams & Wilkins.
1070-5295
Introduction
Mechanical ventilation is the most common therapy in in-
tensive care units. However, since its introduction in clin-
ical practice, its potential harm has been progressively
recognized. Indeed, many studies have been devoted to
the ventilator-lung interaction, involving epidemiologists,
intensivists, biochemists, and molecular biologists. In
this report, we will review what we consider to be the
most significant contributions in the field produced by
our scientific community from August 2003 to September
2004.
Danger of mechanical ventilation
We recently reviewed this topic and will summarize here
the general framework [1•].
It is well recognized that the first trigger of injury is me-
chanical. In the diseased lung, the distribution of stress
(ie, the tension of the lung skeleton fiber system) and
the strain (ie, the elongation of the fibers) is not homoge-
neous. Consequently, the healthy regions of lung paren-
chyma bear more stress and strain than the collapsed
and consolidated lung regions, which are somewhat pro-
tected because they bear stress but do not strain. The pul-
monary endothelial and epithelial cells are anchored to the
fiber skeleton and accommodate their surface to this
unphysiologic strain. The cell deformation increases the
permeability of the endothelial layer through a variety
of mechanisms, including increased size of endothelial
fenestrations, active contraction of the endothelial cells,
and cytokine production. The macrophages and the epithe-
lial cells react, via stretch-activated sensors, by producing
interleukin (IL)-8 and metalloproteinases. IL-8, in partic-
ular, seems to play a pivotal role by recruiting and activat-
ing neutrophils. Neutrophils, in turn, produce a variety of
69
JOBNAME: cocc 11#1 2005 PAGE: 2 OUTPUT: Wed December 29 20:44:41 2004
lww/cocc/91165/COCC11110
70 Respiratory system
proinflammatory mediators, leading to full-blown pulmo-
nary inflammation. Indeed, the balance between proin-
flammatory and antiinflammatory mediators, which are
also increased during ventilator-induced lung injury (VILI),
is in favor of the former.
In this context, we will discuss whether the recent studies
have added new knowledge, confirming or contradicting
the previous results. Our belief is that a clear understand-
ing of the biology underlying VILI is essential for a rational
approach to mechanical ventilation in patients with acute
lung injury (ALI) and acute respiratory distress syndrome
(ARDS).
Mechanisms of ventilator-induced lung injury
Because the biology of VILI has been investigated in var-
ious experimental settings, it is more convenient to ana-
lyze the problem by making a distinction between
experiments based on one-hit model (effects of mechan-
ical ventilation in healthy lung) and those in which a two-
hit model was used (effects of mechanical ventilation in
already injured lung). Here we will discuss first the
strength of the noxious stimulus and then the biologic
response.
Strength of the stimulus
In these studies, different experimental settings were
used, such as cell cultures, isolated perfused or unper-
fused lung models, or in vivo experiments. Similarly, differ-
ent species were used, such as mice, rats, rabbits, dogs,
and pigs. Despite these differences, most of the studies
had a common denominator: perturbation of the system
by high or low tidal volume (VT), which represents the
strength of the stimulus. High VT ranged from 12 to 44
mL/kg of body weight (bw), whereas low VT ranged be-
tween 6 and 8 mL/kg bw. The use of VT according to body
weight represents an attempt to normalize the strength of
the stimulus. Because the VILI is triggered by strain on
the lung structure, it is important to evaluate the relation
between VT per kilogram of body weight and a putative
strain within different species. As shown in Figure 1,
6 mL/kg bw corresponds to an alveolar diameter change
of approximately 7% in humans, whereas the same VT
per kilogram of body weight corresponds to an alveolar
diameter change of approximately 16% in rats and ap-
proximately 25% in mice. In humans, the same alveolar
distension would require, respectively, a VT of approxi-
mately 15 mL/kg bw and approximately 23 mL/kg bw
[1•,2]. Therefore, to produce an injury in healthy human
lungs, a very large VT per kilogram of body weight would
be necessary. However, it is questionable whether a VT
equal to 6 to 7 mL/kg bw is really a low VT in rats, mice,
and rabbits, considering the resulting alveolar distension
(Fig. 1). Finally, the strength of the mechanical stimulus
necessary to produce damage in already injured lungs may
be substantially different. However, despite these limita-
tions, the injurious mechanical ventilation is a useful model
for studying the possible mechanisms determining VILI,
although the results cannot be transferred as such to hu-
man beings.
Biologic response
One-hit model
An interesting report by Copland et al. [3] highlights the
importance of the lung/body size when mechanical venti-
lation is used to cause VILI. In fact, the same ventilatory
pattern was overall more harmful in adults than in new-
born rats. Copland et al. [4] also reported that 30 minutes
of injurious ventilation—VT 25 mL/kg bw, equivalent to
a VT greater than 50 mL/kg bw in humans (Fig. 1)—were
sufficient to upregulate especially 10 genes encoding for
transcription factors, stress proteins, and inflammatory
mediators and to downregulate especially 12 genes mainly
encoding for metabolic enzymes. Such an early activation
was present in the absence of any histologic or mechanical
signs of injury. This type of response is characteristic of
a chaotic system—that is, redundancy of the answer to
a given stimulus. Moreover, it emphasizes the importance
of time. In that study, the gene expression was assessed
after 30 minutes, and we do not know whether the acti-
vation might have occurred in an even shorter time. If that
is the case, even interventions of very short duration, such
as recruitment maneuvers, could be potentially dangerous.
The pivotal role of IL-8 as an early mediator of the inflam-
matory cascade has been confirmed in vitro by stretching
alveolar cell cultures and by several studies performed in
in vivo models [5–7]. The injurious ventilation causes
macrophage activation and the early presence of inflam-
matory mediators such as IL-6 and IL-8 (or its rodent
equivalent, MIP-2), which in turn determine neutrophil
infiltration [8–10].
If the excessive unphysiologic strain is the trigger of VILI
through the activation of stretch-activated sensors of the
epithelial and endothelial cells anchored to the extracel-
lular matrix, it follows that the lung regions that are more
exposed to the injury should be the healthier ones, be-
cause the permanently collapsed or consolidated regions
cannot strain and are relatively protected. There is grow-
ing evidence that this is the case. Caruso et al. [11] found
that in an in vivo rat model, mechanical ventilation with VT
equal to both 6 and 24 mL/kg bw caused an inflammatory
reaction. However, the expression of inflammatory cytokines
was higher in the nondependent lung regions, which were
likely more strained during mechanical ventilation.
Haitsma et al. [10] showed in an in vivo adult rat model
that even noninjurious ventilation, or moderately injurious
ventilation, may induce inflammatory cytokines even in
normal lungs. Additionally, Kurahashi et al. [12], in an
in vivo rat model, reported that high VT ventilation (12
JOBNAME: cocc 11#1 2005 PAGE: 3 OUTPUT: Wed December 29 20:44:41 2004
lww/cocc/91165/COCC11110

Ventilation of patients with acute lung injury Gattinoni et al. 71

Figure 1. Alveolar diameter changes as a function of VT/kg in different species

Alveolar diameter changes relative to baseline diameter [(actual diameter ! baseline diameter)/baseline diameter]. Changes are expressed as
a function of VT/kg in different species (men, mice, rabbits, and rats). Baseline diameter dimensions were derived from Mercer et al. [2]. Alveolar
diameter changes due to volume changes were computed assuming the same relation existing in a sphere, ie, % D diameter = !1 + (1 + % D
Volume)1/3. Dotted arrows indicate different intraspecies putative strains caused by low or high VT/kg bw (6 and 12 mL/kg bw, respectively) and
interspecies differences due to different respiratory system characteristics. Modified with permission [1].

mL/kg bw, equivalent to a VT of 30 mL/kg in humans) was
more harmful in comparison with low VT ventilation only
in the healthy lung, in a model in which the other lung was
injured by Pseudomonas aeruginosa. Moreover, in an ex vivo
rat model, atelectatic lungs (not strained) were found
to be more protected than lungs undergoing cyclic open-
ing and closing [13]. Resolution of atelectasis, by contrast,
seems to attenuate bacterial growth and translocation dur-
ing experimental pneumonia in piglets [14].
form is smoothed. Interestingly, it has been shown in rats
that early gene activation of the inflammatory cascade is
primarily located in the bronchoalveolar airway epithelium
and that leukocyte recruitment occurs primarily in the rat
trachea after the airway distension induced by positive
end-expiratory pressure (PEEP) [4,18]. Taken together,
these findings lead to a reevaluation of the possible impor-
tance of airway distension as a cause of VILI.

Taken together, these experiments raise an important is- Two-hit model

sue in the era of the open lung approach. If the closed lung
regions remain unopened throughout the entire respiratory
cycle and consequently seem to be more protected, why
should we try to open them? Wouldn’t it be better to keep
them always closed until the resolution of the underlying
disease? Further investigations of this topic are warranted.
Recent studies have suggested that not only VT, but the
pattern by which it is delivered, may be important for
the pathogenesis of VILI. In fact, it has been found in
an in vivo rabbit model that increased inspiratory time
increases lung damage [15]. Furthermore, increased peak
inspiratory flow at the same VT may increase lung injury in
rabbits and may decrease arterial oxygenation in piglets
[16,17]. We may speculate that the possible damages in-
duced by the peak flow should be primarily located in
the airways, because at the alveolar level the flow wave-
Although the sequence of events leading to the injury is
more complex because two noxious stimuli are operating
(eg, mechanical strain and endotoxin/lung lavage/acid aspi-
ration), the two-hit model is more similar to the clinical
scenario, wherein the mechanical ventilation is applied
to the already injured lung. In general, the effects of
the two noxious stimuli may be either additive or syner-
gistic. In an ex vivo rat model, Whitehead et al. [19] com-
pared low, high, and medium VT ventilation (7, 40, and 15
mL/kg bw, respectively) after the intratracheal instillation
of lipopolysaccharide (LPS). The results were in part sur-
prising because the inflammatory response, as reflected
by tumor necrosis factor-a and MIP-2 levels in bronchoal-
veolar lavage fluid, was markedly lower in lungs ventilated
at a high VT than in lungs ventilated at a VT of 7 or
15 mL/kg bw. Of note, after high VT ventilation, alveolar
macrophages were mainly detected in the interstitial
JOBNAME: cocc 11#1 2005 PAGE: 4 OUTPUT: Wed December 29 20:44:47 2004
lww/cocc/91165/COCC11110
72 Respiratory system
space outside the alveolar compartment. The accompany-
ing editorial by Dreyfuss and Rouby [20], somewhat ven-
omous (very likely as a consequence of the long-standing
debate between the two groups), emphasizes some con-
tradictory findings between this study and previous find-
ings reported by the same group [19,21]. Nevertheless,
this study suggests that low VT ventilation increases the
inflammatory response when the lung has been previously
challenged with LPS. This has been more clearly shown
by Altemeier et al. [22•], who, in an in vivo rabbit model,
found that a VT equal to 15 mL/kg bw alone resulted in
a minimal pulmonary cytokine expression but markedly
enhanced the inflammatory reaction when associated with
intravenous administration of LPS. The authors con-
cluded in favor of a synergistic effect of systemic endo-
toxin and mechanical ventilation on the increased
cytokine response.
The possible mechanisms underlying the synergy be-
tween mechanical ventilation and endotoxin were ele-
gantly investigated by Uhlig et al. [23••]. In an isolated
and perfused mouse lung model, they found that overven-
tilation caused the nuclear translocation of the transcrip-
tion factor nuclear factor-kB (NF-kB) followed by an
inflammatory reaction involving alveolar macrophages and
epithelial type II cells. The nuclear translocation of NF-
kB was inhibited by infusion of Ly294002, a specific inhib-
itor of the phosphoinositide 3-OH kinase. By contrast,
when the stimulus consisted of LPS challenge, Ly294002
did not prevent the nuclear NF-kB translocation. These
results were also confirmed in an in vivo rat model. There-
fore, it is conceivable that two different pathways are re-
sponsible for the inflammatory process consequent to
different stimuli, and this may account for a synergistic
action of high VT mechanical ventilation and LPS on pul-
monary inflammation and cytokine expression.
Mechanical ventilation of patients with
acute lung injury and acute respiratory
distress syndrome
Since the report of Amato et al. [24], who compared the
high PEEP—low VT and the low PEEP—high VT ventila-
tory strategies for lung protection, several outcome stud-
ies have been performed, testing the effects of mechanical
ventilation with different VT in ALI/ARDS. Of note,
whereas Amato et al. [24] compared strategies that in-
cluded two different variables (PEEP and VT), all the
other studies compared different levels of VT only. Al-
though three studies were inconclusive, the largest Na-
tional Institutes of Health (NIH) trial, comparing VT of
6 versus 12 mL/kg bw, was definitely in favor of low VT ven-
tilation [25–28].
Despite this success, the implementation of a low VT
strategy in clinical practice is still limited, among both
the NIH study participants and the nonparticipants
[29,30]. Several reasons, including the patient’s discom-
fort, have been advocated, but they are not yet well de-
fined [31,32]. One possible explanation may be the
concern that has been raised about the NIH protocol. It
has been suggested that the different outcomes in the
NIH study were not caused by the beneficial effects of
low VT but rather by the detrimental effects of 12 mL/kg
bw ventilation [33]. This view was obviously challenged
by most of the authors of the outcome studies [34]. How-
ever, a recent cohort study demonstrated that the use
of VT of 12 mL/kg bw or higher is an independent
risk factor for the development of ALI/ARDS [35•]. In
our opinion, independently of the specific value assigned
to the harmful VT, it is evident that over the past
25 years the decreased mortality rate in ALI/ARDS
has been associated with a progressive decrease in the
VT used [32].
At this point, it is important to emphasize that tailoring
the VT to the body weight (ideal or actual) may be mis-
leading. What matters, in fact, is the actual lung strain,
for which the ratio between VT and the end-expiratory
lung volume (EELV) is a more appropriate surrogate
[36]. It is obvious that the same VT per kilogram of body
weight may result in completely different VT/EELV ratios
(ie, strain) depending on the severity of the disease. In
fact, the strain is defined as the elongation of the lung struc-
ture compared with its resting position. Unfortunately,
EELV measurement is not a routine practice in inten-
sive care units. In conclusion, the best strategy we can
recommend now is the use of gentle lung ventilation with
low VT. This strategy is generally recognized as beneficial,
as stated by several recent reviews, despite some caveats
[36–39].
One of the most important studies for its potential conse-
quences is the recently published ALVEOLI trial [40••].
The patients were randomized to high or low PEEP strat-
egy, according to a predefined FiO2/PEEP scale. In both
groups, the VT was set as 6 mL/kg bw, and the plateau
pressure was limited to 30 cm H2O. The trial was inter-
rupted after the enrollment of 549 patients, according
to a prespecified ‘‘futility stopping rule.’’ The conclusions
were that ‘‘in ALI/ARDS patients, treated with 6 ml/kg
ideal body weight VT, limiting the end inspiratory plateau
pressure to 30 cm H2O, the clinical outcomes were similar
whether lower or higher PEEP levels were used.’’ This
study, however, had several problems outlined in the ac-
companying editorial [41]. The main concerns were the
imbalance between the two groups at the randomization
and the protocol amendment made during the study. In
fact, the patients randomized to the high PEEP group
were significantly older and had lower PaO2/FiO2 ratios.
After the enrollment of 171 patients, the PEEP/FiO2 scale
was modified to increase the PEEP difference between
JOBNAME: cocc 11#1 2005 PAGE: 5 OUTPUT: Wed December 29 20:44:48 2004
lww/cocc/91165/COCC11110
the two groups. The authors discussed these limitations,
but despite several statistical arrangements they were not
able to detect differences between the two groups in any
of the predefined outcomes, including mortality. The po-
tential consequence of this study, which suggests that
PEEP level is indifferent, may be devastating and sharply
contrasts with the daily clinical practice in patients with
severe ALI/ARDS.
In our opinion, the limits of this study go well beyond
the protocol amendment or the slight group differences
at randomization, if we consider the study design under
a physiologic perspective. The rationale by which PEEP
could reduce mortality during ALI/ARDS is the decrease
of VILI: high PEEP should make the mechanical ventila-
tion less dangerous than low PEEP. The theory behind
the beneficial effects of PEEP is that it may prevent
the cyclic opening and closing of the pulmonary units,
keeping the lung open throughout the respiratory cycle.
In an open lung, according to the model of Mead et al.
[42], the stress and strain should be more homogeneously
distributed throughout the lung parenchyma, preventing
the trigger of the inflammatory reaction. Accordingly,
PEEP may be beneficial only if a given ALI/ARDS lung
has a sufficient potential for recruitment. Because recruit-
ment is an inspiratory phenomenon, sufficient plateau
pressure must be provided to open the lung, and sufficient
PEEP has to be applied to keep the lung open. In the
sponge lung model, more typical of extrapulmonary
ARDS, the alveolar opening pressures may be as high as
45 cm H2O, and the PEEP adequate to prevent the col-
lapse due to gravitational forces may range between 15
and 20 cm H2O—that is, the order of magnitude of the
hydrostatic superimposed pressure in the heavy ARDS
lung [43–46]. However, if the potential for recruitment
is low, as is frequently observed in pulmonary ARDS from
diffuse pneumonia, where consolidated lung regions pre-
vail over the collapsed lung regions, the PEEP may be
harmful because it just increases the alveolar stress and
strain, enhancing VILI [47]. Accordingly, high PEEP
should be reserved for patients with a high potential for
recruitment, and low PEEP for patients with a low poten-
tial for recruitment. There is a growing suggestion that
the potential for recruitment may largely vary among
the ALI/ARDS population [47–49]. Indeed, the definition
of the potential for recruitment is a key issue in exploring
the putative beneficial effects of PEEP or in preventing its
harmful effects. Unfortunately, the PEEP setting is gen-
erally based on the improvement of oxygenation, and
we have known for more than 20 years (and unfortunately
we forget) that PEEP may increase PaO2 without any lung
recruitment, simply because of a decrease in and/or a dif-
ferent distribution of pulmonary perfusion [50]. The
PaCO2 decrease, for the same minute ventilation, seems
to be a more reliable indicator of recruitment of perfused
lung regions [51].
Ventilation of patients with acute lung injury Gattinoni et al. 73
We may now comment on the ALVEOLI study from
a physiologic perspective. Both groups were treated with
a plateau pressure at 30 cm H2O. Because this limit pre-
vents opening of the lung regions with an opening pres-
sure higher than 30 cm H2O, it is questionable whether
the ALVEOLI study tested the open lung approach, as
did the study by Amato et al. [24]. The recruitment
maneuvers, in which 45 cm H2O was applied, were aban-
doned after 80 patients because only small and transient
benefits in oxygenation were observed [52]. However, giv-
en the fact that oxygenation changes are a poor index of
recruitment, the recruitment maneuver, in our opinion,
should be used in the context of the open lung approach.
Moreover, the recruitment maneuver per se, even if suc-
cessful in opening the lung, does not succeed if it is
not followed by a PEEP level sufficient to keep open
the newly recruited lung regions. More important,
patients with different potentials for recruitment were
likely distributed similarly among the two groups. As
the authors recognize in the discussion, the possible ben-
efit of high PEEP in patients with a high potential for re-
cruitment may have been offset by its harmful effects in
patients with a low potential for recruitment. Finally, the
PEEP was set according to an arbitrary FiO2/PEEP scale to
reach the same oxygenation between the two groups. As
previously discussed, the oxygenation changes are not
necessarily related to lung recruitment. Furthermore,
the FiO2/PEEP scales, curiously, are not parallel but seem
to converge at high FiO2 levels (Fig. 2). Indeed, in our
opinion, the study design was inappropriate to show a
PEEP effect on prevention of VILI and improvement in
survival, mainly because it was not based on a physiologic
rationale. The importance of setting the correct PEEP
level is still an open and hot question.
How do we ventilate these patients?
Without denying the importance of randomized trials, in
daily clinical practice we have to face a single patient in-
stead of a population of patients [36]. Because the popu-
lation is heterogeneous, our first approach is to define the
characteristics of the single patient.
1. First, we determine the history and clinical presenta-
tion: the underlying pathologic changes of ARDS oc-
curring with pneumonia are likely different from
ARDS occurring with peritonitis.
2. We define, in each patient, the potential for recruit-
ment by CT scan taken at 5 cm H2O PEEP and at
45 cm H2O plateau pressure at end inspiration (3-slice
technique). Then the patient undergoes a PEEP trial
at 5 and 15 cm H2O PEEP, in volume-controlled ven-
tilation, and we consider together the variations of
PaO2, PaCO2, and respiratory system compliance. The
patients in whom only PaO2 improves are usually those
with a low potential for recruitment. In this case, we
will use a PEEP between 5 and 10 cm H2O, whereas
JOBNAME: cocc 11#1 2005 PAGE: 6 OUTPUT: Wed December 29 20:44:48 2004
lww/cocc/91165/COCC11110
74 Respiratory system
Figure 2. Graphic representation of positive end-expiratory pressure (PEEP) and fractional inspired oxygen (FiO2) combinations
designed for the ALVEOLI trial study
Closed circles, lower-PEEP strategy combinations. Open circles, higher-PEEP group strategy combinations. Gray circles, steps excluded from the
higher-PEEP strategy (PEEP < 12 cm H2O) after 171 patients had already been enrolled in the trial. Dash, mean FiO2 values during day 1 treatment:
0.54 ± 0.18 (lower-PEEP group) 0.44 ± 0.17 (higher-PEEP group). Published with permission [40••].
we will use a PEEP greater than 15 cm H2O in patients
in whom the overall gas exchange and the CT scan
show a high potential for recruitment. An ongoing
multicenter study (Gattinoni L, unpublished data)
seems to provide a solid physiologic basis to justify
this approach.
3. We then use a low VT rather than using a VT according
to the body weight (mL/kg), and we try to tailor VT
value according to the EELV as measured by helium.
We believe that the VT/EELV ratio is the greatest in-
dicator of strain clinically available.
4. As for pressure, we measure the transpulmonary pres-
sure, which is the real indicator of stress.
All this is done at the beginning, as soon as the patient is
admitted to the intensive care unit, and it is repeated if
any discrepancy between the clinical scenario and the ob-
jective data, such as gas exchange or lung imaging, is ob-
served. Overall, our ventilatory approach is based on the
distinction between patient recruiters and nonrecruiters,
according to the results obtained from different challenge
tests. From a physiologic perspective, the differentiation
between recruiters and nonrecruiters is the key issue in
tailoring the safest mechanical ventilation in ALI/ARDS
patients.
Conclusion
Further advances in the past 12 months have been made in
our understanding of the mechanism of VILI. Recent data
emphasize the speed of the beginning of the inflammatory
reaction cascade, involving an immediate upregulation
and downregulation of numerous clusters of genes. Mac-
rophage activation and IL-8 production seem to maintain
their pivotal role in the inflammatory reaction induced by
high VT. The airway distension caused by PEEP or peak
inspiratory flow may also play a role in VILI. The healthier
lung regions may be the main target of the injury because
the collapsed and consolidated regions are relatively pro-
tected. Despite the experimental and clinical evidence
that high VT ventilation is harmful, there is still some re-
luctance to use VT of 6 mL/kg bw. Unquestionably, how-
ever, high VT should never be used. The ALVEOLI study
was not able to show any influence of PEEP (high or low)
on outcome. Setting PEEP, however, is still an open issue
because the methodologic problems and the unphysio-
logic design of this study do not allow any conclusion to
be drawn. Our believe is that mechanical ventilation
should be tailored to different variables, such as transpul-
monary pressure and EELV. Unfortunately, despite its
physiologic rationale, this approach is far from being used
and proved.
References and recommended reading
Papers of particular interest, published within the annual period of review, have
been highlighted as:
• of special interest
•• of outstanding interest
1 Gattinoni L, Carlesso E, Cadringher P, et al. Physical and biological triggers of
• ventilator-induced lung injury and its prevention. Eur Respir J Suppl 2003;
47:15s–25s.
JOBNAME: cocc 11#1 2005 PAGE: 7 OUTPUT: Wed December 29 20:44:53 2004
lww/cocc/91165/COCC11110
This review and discussion of VILI (triggers and mechanisms) focused on
mechanical force distribution in the diseased lung and the biologic inflammatory
response.
2 Mercer RR, Russell ML, Crapo JD: Alveolar septal structure in different spe-
cies. J Appl Physiol 1994; 77:1060–1066.
3 Copland IB, Martinez F, Kavanagh BP, et al. High tidal volume ventilation
causes different inflammatory responses in newborn versus adult lung. Am
J Respir Crit Care Med 2004; 169:739–748.
4 Copland IB, Kavanagh BP, Engelberts D, et al. Early changes in lung gene
expression due to high tidal volume. Am J Respir Crit Care Med 2003;
168:1051–1059.
5 Jafari B, Ouyang B, Li LF, et al. Intracellular glutathione in stretch-induced
cytokine release from alveolar type-2 like cells. Respirology 2004; 9:43–53.
6 Dos Santos CC, Han B, Andrade CF, et al. DNA microarray analysis of
gene expression in alveolar epithelial cells in response to TNF-{alpha}, LPS
and cyclic stretch. Physiol Genomics 2004; Sep 18 (Epub ahead of
print).
7 Mascarenhas MM, Day RM, Ochoa CD, et al. Low molecular weight hyalur-
onan from stretched lung enhances interleukin-8 expression. Am J Respir Cell
Mol Biol 2004; 30:51–60.
8 Li LF, Yu L, Quinn DA: Ventilation-induced neutrophil infiltration depends
on c-Jun N-terminal kinase. Am J Respir Crit Care Med 2004; 169:518–524.
9 Kotani M, Kotani T, Ishizaka A, et al. Neutrophil depletion attenuates interleu-
kin-8 production in mild-overstretch ventilated normal rabbit lung. Crit Care
Med 2004; 32:514–519.
10 Haitsma JJ, Uhlig S, Verbrugge SJ, et al. Injurious ventilation strategies cause
systemic release of IL-6 and MIP-2 in rats in vivo. Clin Physiol Funct Imaging
2003; 23:349–353.
11 Caruso P, Meireles SI, Reis LF, et al. Low tidal volume ventilation induces
proinflammatory and profibrogenic response in lungs of rats. Intensive Care
Med 2003; 29:1808–1811.
12 Kurahashi K, Ota S, Nakamura K, et al. Effect of lung-protective ventilation on
severe Pseudomonas aeruginosa pneumonia and sepsis in rats. Am J Physiol
Lung Cell Mol Physiol 2004; 287:L402–L410.
13 Chu EK, Whitehead T, Slutsky AS: Effects of cyclic opening and closing at
low- and high-volume ventilation on bronchoalveolar lavage cytokines. Crit
Care Med 2004; 32:168–174.
14 van Kaam AH, Lachmann RA, Herting E, et al. Reducing atelectasis attenu-
ates bacterial growth and translocation in experimental pneumonia. Am J Re-
spir Crit Care Med 2004; 169:1046–1053.
15 Simonson DA, Adams AB, Wright LA, et al. Effects of ventilatory pattern on
experimental lung injury caused by high airway pressure. Crit Care Med 2004;
32:781–786.
16 Maeda Y, Fujino Y, Uchiyama A, et al. Effects of peak inspiratory flow on de-
velopment of ventilator-induced lung injury in rabbits. Anesthesiology 2004;
101:722–728.
17 Dembinski R, Henzler D, Bensberg R, et al. Ventilation-perfusion distribution
related to different inspiratory flow patterns in experimental lung injury. Anesth
Analg 2004; 98:211–219.
18 Lim LH, Wagner EM: Airway distension promotes leukocyte recruitment in rat
tracheal circulation. Am J Respir Crit Care Med 2003; 168:1068–1074.
19 Whitehead TC, Zhang H, Mullen B, et al. Effect of mechanical ventilation on
cytokine response to intratracheal lipopolysaccharide. Anesthesiology 2004;
101:52–58.
20 Dreyfuss D, Rouby JJ: Mechanical ventilation-induced lung release of cyto-
kines: a key for the future or Pandora’s box? Anesthesiology 2004;
101:1–3.
21 Uhlig S, Ranieri M, Slutsky AS: Biotrauma hypothesis of ventilator-induced
lung injury. Am J Respir Crit Care Med 2004; 169:314–315.
22 Altemeier WA, Matute-Bello G, Frevert CW, et al. Mechanical ventilation with
• moderate tidal volumes synergistically increases lung cytokine response
to systemic endotoxin. Am J Physiol Lung Cell Mol Physiol 2004; 287:
L533–L542.
This in vivo rat model of VILI showed the synergism between mechanical ventila-
tion and endotoxin action on proinflammatory mediator release.
23 Uhlig U, Fehrenbach H, Lachmann RA, et al. Phosphoinositide 3-OH kinase
•• inhibition prevents ventilation-induced lung cell activation. Am J Respir Crit
Care Med 2004; 169:201–208.
This paper reports an elegant investigation, in an in vivo and ex vivo animal model,
of different mechanisms determining inflammatory cellular activation and cytokine
production after endotoxin challenge and high inflammatory VT mechanical venti-
lation.
Ventilation of patients with acute lung injury Gattinoni et al. 75
24 Amato MB, Barbas CS, Medeiros DM, et al. Effect of a protective-ventilation
strategy on mortality in the acute respiratory distress syndrome. N Engl J Med
1998; 338:347–354.
25 Stewart TE, Meade MO, Cook DJ, et al. Evaluation of a ventilation strategy to
prevent barotrauma in patients at high risk for acute respiratory distress syn-
drome. Pressure- and Volume-Limited Ventilation Strategy Group. N Engl J
Med 1998; 338:355–361.
26 Brochard L, Roudot-Thoraval F, Roupie E, et al. Tidal volume reduction for
prevention of ventilator-induced lung injury in acute respiratory distress syn-
drome. The Multicenter Trail Group on Tidal Volume reduction in ARDS. Am J
Respir Crit Care Med 1998; 158:1831–1838.
27 Brower RG, Shanholtz CB, Fessler HE, et al. Prospective, randomized, con-
trolled clinical trial comparing traditional versus reduced tidal volume ventila-
tion in acute respiratory distress syndrome patients. Crit Care Med 1999;
27:1492–1498.
28 The Acute Respiratory Distress Syndrome Network: Ventilation with lower
tidal volumes as compared with traditional tidal volumes for acute lung injury
and the acute respiratory distress syndrome. N Engl J Med 2000; 342:
1301–1308.
29 Weinert CR, Gross CR, Marinelli WA: Impact of randomized trial results on
acute lung injury ventilator therapy in teaching hospitals. Am J Respir Crit
Care Med 2003; 167:1304–1309.
30 Young MP, Manning HL, Wilson DL, et al. Ventilation of patients with acute
lung injury and acute respiratory distress syndrome: has new evidence
changed clinical practice? Crit Care Med 2004; 32:1260–1265.
31 Rubenfeld GD, Cooper C, Carter G, et al. Barriers to providing lung-protec-
tive ventilation to patients with acute lung injury. Crit Care Med 2004; 32:
1289–1293.
32 Villar J, Kacmarek RM, Hedenstierna G: From ventilator-induced lung injury to
physician-induced lung injury: why the reluctance to use small tidal volumes?
Acta Anaesthesiol Scand 2004; 48:267–271.
33 Eichacker PQ, Gerstenberger EP, Banks SM, et al. Meta-analysis of acute
lung injury and acute respiratory distress syndrome trials testing low tidal vol-
umes. Am J Respir Crit Care Med 2002; 166:1510–1514.
34 Amato M, Brochard L, Stewart T, et al. Metaanalysis of tidal volume in ARDS.
Am J Respir Crit Care Med 2003; 168:612–613.
35 Gajic O, Dara SI, Mendez JL, et al. Ventilator-associated lung injury in patients
• without acute lung injury at the onset of mechanical ventilation. Crit Care Med
2004; 32:1817–1824.
This interesting retrospective cohort study suggests that high VT ventilation is one
of the most important and independent risk factors for the development of ALI.
36 Marini JJ, Gattinoni L: Ventilatory management of acute respiratory distress
syndrome: a consensus of two. Crit Care Med 2004; 32:250–255.
37 Rouby JJ, Constantin JM, Roberto De AG, et al. Mechanical ventilation in pa-
tients with acute respiratory distress syndrome. Anesthesiology 2004;
101:228–234.
38 Moloney ED, Griffiths MJ: Protective ventilation of patients with acute respi-
ratory distress syndrome. Br J Anaesth 2004; 92:261–270.
39 Petrucci N, Iacovelli W: Ventilation with smaller tidal volumes: a quantitative
systematic review of randomized controlled trials. Anesth Analg 2004; 99:
193–200.
40 Brower RG, Lanken PN, MacIntyre N, et al. Higher versus lower positive end-
•• expiratory pressures in patients with the acute respiratory distress syndrome.
N Engl J Med 2004; 351:327–336.
This important prospective, randomized, controlled trial evaluated, during a low
VT ventilation strategy, the possible benefits of the application of different PEEP
levels.
41 Levy MM: PEEP in ARDS: how much is enough? N Engl J Med 2004; 351:
389–391.
42 Mead J, Takishima T, Leith D: Stress distribution in lungs: a model of pulmo-
nary elasticity. J Appl Physiol 1970; 28:596–608.
43 Bone RC: The ARDS lung: new insights from computed tomography. JAMA
1993; 269:2134–2135.
44 Gattinoni L, D’Andrea L, Pelosi P, et al. Regional effects and mechanism of
positive end-expiratory pressure in early adult respiratory distress syndrome.
JAMA 1993; 269:2122–2127.
45 Crotti S, Mascheroni D, Caironi P, et al. Recruitment and derecruitment dur-
ing acute respiratory failure: a clinical study. Am J Respir Crit Care Med 2001;
164:131–140.
46 Gattinoni L, Pelosi P, Crotti S, et al. Effects of positive end-expiratory
pressure on regional distribution of tidal volume and recruitment in adult
JOBNAME: cocc 11#1 2005 PAGE: 8 OUTPUT: Wed December 29 20:44:54 2004
lww/cocc/91165/COCC11110
76 Respiratory system
respiratory distress syndrome. Am J Respir Crit Care Med 1995; 151:1807–
1814.
47 Gattinoni L, Pelosi P, Suter PM, et al. Acute respiratory distress syndrome
caused by pulmonary and extrapulmonary disease: different syndromes?
Am J Respir Crit Care Med 1998; 158:3–11.
48 Gainnier M, Michelet P, Thirion X, et al. Prone position and positive end-ex-
piratory pressure in acute respiratory distress syndrome. Crit Care Med
2003; 31:2719–2726.
49 Broccard A: Positive end-expiratory pressure or prone position: is that the
question? Crit Care Med 2003; 31:2802–2803.
50 Dantzker DR, Lynch JP, Weg JG: Depression of cardiac output is a mecha-
nism of shunt reduction in the therapy of acute respiratory failure. Chest
1980; 77:636–642.
51 Gattinoni L, Vagginelli F, Carlesso E, et al. Decrease in PaCO2 with prone
position is predictive of improved outcome in acute respiratory distress syn-
drome. Crit Care Med 2003; 31:2727–2733.
52 Brower RG, Morris A, MacIntyre N, et al. Effects of recruitment maneuvers in
patients with acute lung injury and acute respiratory distress syndrome ven-
tilated with high positive end-expiratory pressure. Crit Care Med 2003; 31:
2592–2597.