You are on page 1of 10

Topographic Distribution of Tidal Ventilation in Acute Respiratory Distress Syndrome: Effects of Positive End-Expiratory Pressure and Pressure

Tommaso Mauri, MD1,2; Giacomo Bellani, MD, PhD1,2; Andrea Confalonieri, MD2; Paola Tagliabue, MD2; Marta Turella, MD1,2; Andrea Coppadoro, MD1; Giuseppe Citerio, MD2; Nicolo’ Patroniti, MD1,2; Antonio Pesenti, MD1,2

Objective: Acute respiratory distress syndrome is characterized by collapse of gravitationally dependent lung regions that usually diverts tidal ventilation toward nondependent regions. We hypothesized that higher positive end-expiratory pressure and enhanced spontaneous breathing may increase the proportion of tidal ventilation reaching dependent lung regions in patients with acute respiratory distress syndrome undergoing pressure support ventilation. Design: Prospective, randomized, cross-over study. Setting: General and neurosurgical ICUs of a single universityaffiliated hospital. Patients: We enrolled ten intubated patients recovering from acute respiratory distress syndrome, after clinical switch from controlled ventilation to pressure support ventilation. Interventions: We compared, at the same pressure support ventilation level, a lower positive end-expiratory pressure (i.e., clinical positive end-expiratory pressure = 7 ± 2 cm H2O) with a higher one, obtained by adding 5 cm H2O (12 ± 2 cm H2O). Furthermore, a pressure support ventilation level associated with increased respiratory drive (3 ± 2 cm H2O) was tested against resting pressure support ventilation (12 ± 3 cm H2O), at clinical positive end-expiratory pressure.
*See also p. 1811. 1 Department of Health Sciences, University of Milan-Bicocca, Monza, Italy. 2 Department of Perioperative Medicine and Intensive Care, San Gerardo  Hospital, Monza, Italy. This study has been performed in the General and Neurosurgical Intensive Care Units of the university-affiliated San Gerardo Hospital, University of Milan-Bicocca, Monza, Italy. Supported, in part, by Institutional and FSE Regione Lombardia, Milan, Italy. Dräger Medical GmbH, Lübeck, Germany, provided free of charge electrical impedance tomography monitoring device and healthy controls data. Dr. Coppadoro received grant support from Draeger. The remaining authors have disclosed that they do not have any potential conflicts of interest. For information regarding this article, E-mail: Copyright © 2013 by the Society of Critical Care Medicine and Lippincott Williams & Wilkins DOI: 10.1097/CCM.0b013e318287f6e7

Measurements and Main Results: During all study phases, we measured, by electrical impedance tomography, the proportion of tidal ventilation reaching dependent and nondependent lung regions (Vt%dep and Vt%nondep), regional tidal volumes (Vtdep and Vtnondep), and antero-posterior ventilation homogeneity (Vt%nondep/ Vt%dep). We also collected ventilation variables and arterial blood gases. Application of higher positive end-expiratory pressure levels increased Vt%dep and Vtdep values and decreased Vt%nondep/Vt%dep ratio, as compared with lower positive end-expiratory pressure (p < 0.01). Similarly, during lower pressure support ventilation, Vt%dep increased, Vtnondep decreased, and Vtdep did not change, likely indicating a higher efficiency of posterior diaphragm that led to decreased Vt%nondep/Vt%dep (p < 0.01). Finally, Pao2/Fio2 ratios correlated with Vt%dep during all study phases (p < 0.05). Conclusions: In patients with acute respiratory distress syndrome undergoing pressure support ventilation, higher positive end-expiratory pressure and lower support levels increase the fraction of tidal ventilation reaching dependent lung regions, yielding more homogeneous ventilation and, possibly, better ventilation/perfusion coupling. (Crit Care Med 2013; 41:1664–1673) Key Words: acute respiratory distress syndrome; electrical impedance tomography; lung collapse; mechanical ventilation; outcome; positive end-expiratory pressure


cute respiratory distress syndrome (ARDS) is characterized by bilateral lung inflammation, edema, and infiltration of inflammatory cells (1–3). CT studies showed that, in patients with ARDS, inflammatory edema increases lung weight and causes bilateral collapse and loss of aeration in gravitationally dependent lung regions (4–6). Gravitational loss of lung aeration decreases the fraction of lung receiving tidal ventilation (i.e., the “baby lung”), thus increasing the risk of “barotrauma” and ventilator-induced lung injury (VILI) (6). Furthermore, regional loss of ventilation is a main determinant of hypoxia in ARDS, as dependent regions receive the
July 2013 • Volume 41 • Number 7


as per clinical deciance changes (i. Diaphragm dysfunction portion of tidal ventilation reaching dependent lung regions in was assessed by ultrasounds (20) when clinically (e. the focus of investigators has moved toward the clinical meaning of EIT-derived parameters (15). EIT dedicated belt. results. presence of surgical wounds dressing).ccmjournal.g. admitted to the general and neurosurgical ICU of the uniFigure 1. contraindication end-expiratory pressure (PEEP) levels have been shown to proto EIT use (e. containing 16 equally spaced Critical Care Medicine www. might help optimize PSV in patients with ARDS. predicted body weight. study phases. increased positive end-expiratory pressure (PEEP) (top two images) and decreased PSV (bottom two images) induced redistribution of tidal impedMV to PSV. and might promote VILI by increasing plateau pressure (11). High positive were age younger than 18 years old.and severe cardiovascular instability. SimPEEP and lower support levels (leading to more intense diaplified Acute Physiology Score II values. active diaphragm contraction increases the pro.. elevated PEEP negatively impacts hemodynamics (10) the right position (e. study. Electrical impedance tomography image reconstruction of regional distribution of changes in impedversity-affiliated San Gerardo ance between end-inspiration and end-expiration (black = no change. Exclusion criteria most relevant fraction of lung perfusion (7.g. 15–17). thus and Lung Injury Score (LIS) (21) were recorded at enrollment. Sepsis-Related Organ phragm contraction) might increase the fraction of tidal venFailure Assessment score. age. 13) and with supra-elevation) suspected. Alternatively. and informed consent was obtained. after they sentative patient recovering from acute respiratory distress syndrome undergoing pressure support ventilation were switched from controlled (PSV) during four different study phases (see text for details). tilation reaching gravitationally dependent lung regions.e.g. 8). EIT Monitoring We assessed regional ventilation distribution by electrical Patients were positioned in semirecumbent position during all impedance tomography (EIT): a relatively new. EIT measures the impedance offered by different imaged lung areas to passage of low-voltage alternate electrical currents (i.. altered diaphragm function (e. As shown. Monza.. combined with other clinical data. white = max change) from one repreHospital.. but recently. MATERIALS AND METHODS Study Population We enrolled ten consecutive intubated patients with ARDS (19). presence of pacemaker or automatic implantable mote ventilation of dependent lung regions in ARDS (9). low maxispontaneously breathing healthy subjects in comparison with mal inspiratory pressure) or radiologically (e. hemidiaphragm paralysis). In this approved the study. patients recovering from ARDS)..g. EIT technology has been studied for a long time (18). howcardioverter defibrillator). days spent on MV. we assessed the effects of different PEEP and support settings on regional ventilation in a group of patients recoverDemographic Data Collection ing from ARDS undergoing PSV (14). EIT allows continuous monitoring of the regional distribution of tidal ventilation (17). As air offers the highest impedance within lungs.e... pregnancy. We reasoned that higher 1665 .. sion (i. hemidiaphragm controlled mechanical ventilation (MV) (12. bedside lung imaging method (13.g. noninvasive.e. Our We also recorded in-hospital mortality.Clinical Investigations radiation-free. ARDS etiology. regional impedance). of tidal ventilation) from nondependent (nondep) to dependent (dep) lung regions. causing more homogeneous distribution of ventilation. The institutional ethical committee high-assist pressure support ventilation (PSV) (13). body mass index (BMI). impossibility to place the EIT belt in ever. relative changes in regional lung impedance should reflect changes in air content of that particular region: thus. Italy.

PEEPclin + 5 cm H2O) measurements versus ten PEEPlow (i. and ten PSVhigh measurements versus ten PSVlow. if p0. PEEP was maintained at PEEPclin level in both cases. Cstrs = static respiratory system compliance. then PSVclin was regarded as PSVhigh. while leaving Fio2. PEEPclin). 1). 4. Germany). EIT Data We performed offline analyses of raw EIT data recorded at the end of each study phase and immediately prior to the next. 2.1 ≥ 2 cm H2O). – = descriptive qualitative variable that cannot be summarized by a number. Homogeneity of the antero-posterior distribution of tidal volume: we calculated a regional index of ventilation homogeneity as the ratio of Vt%nondep and Vt%dep (that is Main Characteristics Gender Male Female Female Male Male Female Female Female Male Male 5 males/ 5 females Age (yr) 56 75 44 54 53 68 62 66 72 53 60 ± 10 Body Mass Simplified Acute Sequential Organ Index Physiology Failure Assessment (kg/m2) Score II Score 26 41 27 27 23 33 23 25 23 28 27 (23–29) 40 54 36 46 64 33 89 53 40 82 54 ± 19 7 5 2 5 8 6 5 5 5 19 5 (5–7) Acute Respiratory Distress Syndrome Etiology Trauma Postoperative respiratory failure Pneumonia Postoperative respiratory failure Septic shock Trauma Pneumonia Pneumonia Postoperative respiratory failure Pneumonia PEEPclin = clinically set positive end-expiratory pressure (corresponding to study PEEPlow). respectively). performed at PEEPclin. Study Protocol In each patient. performed at the same PSVclin. patients were switched to volume assist/control ventilation to determine respiratory system compliance Table 1. registered at 20 Hz. 2.1 ≥ 2 cm H2O. we could compare ten PEEPhigh (i.1 < 2  cm H2O) versus a low PSV level (associated with p0. Thus.e. was placed around each patient’s thorax at the fifth or sixth intercostal space and connected to a commercial EIT monitor (PulmoVista 500.. We aimed to compare a high PSV level (defined as a PSV level associated with p0.e.. Regional distribution of tidal volume: we calculated an estimate of the absolute value of tidal volume reaching dependent and nondependent lung regions by multiplication of expiratory Vt measured by the ventilator × Vt%dep and × Vt%nondep (Vtdep and Vtnondep.ccmjournal.  Patients’ Patient No. Clinically selected PSV (PSVclin) and PEEP (PEEPclin) levels. Data are summarized as mean ± sd or by median (interquartile range) for nonnormally distributed variables. We measured the following: 1. 3. as previously described (15–17). and inspiratory and expiratory triggers unchanged: 1. and a new PSVlow was selected during this phase. and stored for offline analysis. EIT data were generated by application of small alternate electrical currents rotating around patient’s thorax.1 ≥ 2 cm H2O. 1 2 3 4 5 6 7 8 9 10 Mean ± sd or median (interquartile range) (Cstrs) by means of end-expiratory and end-inspiratory occlusions. At the end. PSV inspiratory ramp. at least 4 cm H2O lower than PSVclin and set to obtain p0. Lübeck.Mauri et al electrodes. We measured the percentage of tidal ventilation ventilating each ROI as the proportion of total tidal impedance variation ventilating each ROI (Vt%nondep and Vt%dep. 1666 www. each lasting 20 minutes. To this end. At the opposite.. then PSVclin was regarded as PSVlow and a new PSVhigh was selected during this step. 3. 1): gravitationally nondependent (nondep) ROI from halfway to the top of the imaging field and dependent (dep) ROI from halfway to the bottom (Fig. respectively). PSVclin and PEEPhigh ( July 2013 • Volume 41 • Number 7 . at least 4 cm H2O higher than PSVclin to achieve p0. if PSVclin was associated with p0. PEEPclin + 5 cm H2O). Dräger Medical GmbH.1 < 2 cm H2O.1 during PSVclin was < 2 cm H2O. we tested three different ventilator settings in random order. During all study phases.e. Relative distribution of tidal ventilation: we identified two contiguous regions of interest (ROI) (Fig.

Nonnormally distributed variables were compared by a Mann-Whitney U test (two independent samples) or Wilcoxon test (two paired samples). as appropriate. Nine of ten patients were enrolled within a week from intubation. Quadrants Involved 3 0 3 2 1 0 1 2 2 1 2 ± 1 Lung Injury Score 2 1 2.75 1 (1–1. Normally distributed variables.. Association between variables was assessed by a Spearman rho coefficient. and postoperative respiratory failure (three patients). All patients had ARDS diagnosed at intubation and were recovering from it.05 (two tailed) was considered statistically significant.ccmjournal. PEEPhigh). Normally distributed data are indicated as mean ± sd. LIS values did not exceed mild-to-moderate severity range. We assessed normal distribution of each variable by Lilliefors test.0 (Systat Software. Days on Mechanical Ventilation 1 2 20 1 2 4 2 2 4 3 2 (2–4) Pao2/Fio2 (mm Hg) 174 253 250 278 244 271 263 253 280 140 253 (226–273) PEEPclin (cm H2O) 9 5 12 5 5 5 5 9 5 8 7 ± 2 Cstrs (mL/cm H2O) 80 31 23 73 52 38 48 59 70 51 52 ± 18 No. we also collected ventilator settings and arterial blood gas analysis. A level of p value of less than 0. ΔEELIgl values should reflect changes in patient’s end-expiratory lung volume. respectively). BMI 27 ± 4 kg/m2) during spontaneous quiet breathing.5 1 1 1 1 1. RESULTS Patient Characteristics Patients’ main characteristics are reported in Table 1. being already switched by the attending physician from controlled to assisted MV. NY). San Jose. Five patients still fulfilled ARDS criteria when the study was performed. Armonk. 5. 4. ARDS etiology was trauma (two patients).. CA) and by IBM SPSS Statistics 19 (International Business Machines Corp. while median and interquartile range (IQR) are used to report nonnormally distributed variables. Statistical analyses were performed by SigmaPlot 11. Healthy Controls By means of the same EIT technique. pneumonia (four patients). baseline) and another (i. Physiological Data At the same time point of EIT data analysis. Patients were 60  ±  10 years old and 5 (50%) were women. Pixel-level regional ventilation heterogeneity: pixel-level heterogeneity index (H) was calculated as the standard deviation of the proportion of total tidal impedance variation distributed to each pixel in each ROI (Hnondep and Hdep.75) Outcome Survivor Survivor Survivor Survivor Nonsurvivor Survivor Survivor Survivor Survivor Nonsurvivor 8 survivors/2 nonsurvivors Critical Care Medicine www. aged 45  ±  9 yr old. Statistical Analysis We choose the sample size based on previous similar studies (22).75 1 1. instead.Clinical Investigations equal also to Vtnondep/Vtdep ratio). Lower levels indicate more homogenous distribution of tidal ventilation. Two patients died before hospital discharge.e. PEEPlow vs. Vt%nondep and Vt%dep values were collected from 15 semirecumbent nonintubated healthy adults (four women and 11 1667 .  Relative changes of global end-expiratory lung impedance (ΔEELIgl): defined as the sum of percentage change (at single pixel level) in absolute EELIi value between one study phase (i. Thus. were analyzed by independent samples or paired t test and by one sample z test. septic shock (one patient).e..

01) when PEEPlow was considered baseline (Table 3). Gattinoni et al (9) also showed that tidal ventilation redistribution followed recruitment of previously collapsed lung regions.06 0. Vt%dep and Vt%nondep during PEEPhigh did not differ from healthy controls.01). respectively.7–5. MVe = expired minute ventilation. Correlation Between Regional Ventilation and Oxygenation Pao2/Fio2 ratios correlated with Vt%dep during all PEEP and PSV phases (Spearman rho = 0.1 = pressure generated during the first 100 ms of inspiration.1 (Fig. increasing PEEP from 7  ±  2 cm H2O to 12 ± 2 cm H2O significantly increased relative and absolute ventilation of dependent lung regions and antero-posterior ventilation homogeneity to values that became more similar to healthy controls. Hnondep decreased. a p < 0. Vt%dep. whereas respiratory rate and p0. leaving Fio2 value. only Vtnondep decreased (p = 0.42 ± 0. During PSVlow. while Hdep significantly increased (p < 0. as compared with PEEPlow (p < 0. Data are expressed as mean ± sd or median (interquartile range) for nonnormally distributed variables.Mauri et al Table 2. in comparison with PSVhigh. In this study.06 0.05 for all comparisons) (Fig. inspiratory ramp. PEEPhigh was associated with increased Vt%dep and Vtdep (p < 0. Vt%nondep/Vt%dep ratios were lower in survivors (1. 2. and Vt%nondep/Vt%dep during PSVlow did not differ from healthy controls’ values. Vt%nondep.629 and 0. during PEEPlow study phase). Gattinoni et al (9) already showed that increasing PEEP from 0 to 20  cm H2O improved the percentage of tidal ventilation reaching dependent lung areas in eight sedated and paralyzed patients with ARDS undergoing two single-slice CT scans at each PEEP level. in our study. EELIgl increased at higher PEEP levels and paralleled dependent redistribution July 2013 • Volume 41 • Number 7 .e.01).06 Paco2 (mm Hg) 39 ± 5 41 ± 5 42 ± 6 39 ± 4 PSV = pressure support ventilation.4 [1.05 for all) (Fig. while Vtdep did not change (Table 3. PEEP = positive end-expiratory pressure. while Hdep did not change (Table 3).06 0. while during PEEPlow.05 vs PSVlow (paired t test). Ppeak = peak inspiratory pressure. they were significantly different from controls’ values (p < 0.01) (Table 3). Table 3). Vt%dep and Vtdep values were significantly higher in survivors in comparison with nonsurvivors (median 40% IQR [28–50]% vs 17 [16–18]% and 3.01). PSVhigh.01 vs PEEPhigh (Wilcoxon’s test).4 [2. PEEPhigh significantly improved oxygenation and increased peak inspiratory pressure. Regional ventilation distribudep tion was more homogenous during PEEPhigh.6] vs 4.01).5 [1. and PEEPhigh were performed randomly in each patient for 20 min.42 ± 0. despite lower mechanical ventilatory support. p0. Finally. Furthermore.487. PEEP Effects on Regional Ventilation As expected. 5).e. 4. 3) and respiratory rate increased (Table 2). while p0..05 for both).01).9] mL/kg vs 1.05 for both) (Fig. more homogenous lung ventilation) (p < 0. PSVlow. PEEPlow.42 ± 0..1 were not affected by higher PEEP levels (Table 2). c p < 0. and inspiratory and expiratory triggers unchanged.05).01 vs PEEPhigh (paired t test). During PSVlow. Association Between Regional Ventilation Distribution and Outcome When measured at clinically set PSV and PEEP levels (i. whereas at PSVhigh.  Effects of Different Positive End-Expiratory Pressure and Support Levels on Patient Respiratory Parameters PSV Level (cm H2O) 8 ± 5 8 ± 5 3 ± 3 12 ± 3 c Study Phasea PEEPlow (n = 10) PEEPhigh (n = 10) PSVlow (n = 10) PSVhigh (n = 10) PEEP (cm H2O) 7 ± 2a 12 ± 2 7 ± 2 7 ± 2 Pao2/Fio2 (mm Hg) 262 (231–283)b 289 (239–309) 264 (250–290) 257 (225–297) Fio2 0. PSVlow was associated with increased Vt%dep (p < 0. 2. 4). 1668 www. as testified by decreased Vt%nondep/Vt%dep values (p < 0. p < 0. and with lower Vt%nondep/ Vt%dep ratios (i. Table 3). Vt and peak inspiratory pressure were lower. Hnondep decreased. DISCUSSION The main findings of this study can be summarized as follows: in a population of patients recovering from ARDS undergoing PSV.3–1. Pressure Support Effects on Regional Ventilation PSVlow did not affect gas exchange. During PSVlow. Similarly.42 ± 0.01) (Table 3).0]. p < 0. as compared with PSVhigh (p < 0.01 vs PSVlow (paired t test). both of those two EIT parameters were significantly different from controls (p < 0.2–4. 4.01 for both) and with decreased Vt%nondep and Vtnon(p < 0.0–2.01) (Fig. we showed that in patients with ARDS clinically switched to PSV. with decreased Vt%nondep (p < 0.ccmjournal. ΔEELIgl did not vary significantly between the two PSV phases (Table 3). as compared with PSVhigh (Fig. b p < 0.01).5] mL/kg. d p < 0. During PEEPhigh. higher PEEP and lower support levels increase the proportion of tidal ventilation reaching dependent regions and induce more homogenous antero-posterior distribution of ventilation. PEEPhigh was associated with significant changes in ΔEELIgl (p < 0. p < 0. Table 3).9 [4. Table 3).

EELIgl increase may reflect both recruitment and overdistension of nondependent regions (20).40–7. respectively) increased (C and D).ccmjournal. Thus. impedance threshold levels able to discriminate normal lung inflation versus overdistension in human subjects are not yet clearly identified. Critical Care Medicine 1669 . *p < 0.0 ± 6.2 ± 3. HC = healthy controls.4 ± 4.42 (7.05. as shown by lower p0. of tidal ventilation. our data cannot exclude that increased plateau pressure during PEEPhigh might have caused nondependent overdistension.44) 7.1 c Respiratory Rate (b/min) 16 ± 5 17 ± 5 19 ± 5 15 ± 4 d MVe (L/min) 8.6c Figure 2.01 and †p < 0.5 ± 3 6. worsening regional compliance and redirecting tidal ventilation to more dependent lung portions (23). To date.4 8.1 18.8 8. at variance with what is observed during spontaneous ventilation (12.1 values) versus lower PSV (associated with www.8 ± 1. We assessed ventilation of dependent lung regions in patients with ARDS during higher resting PSV (associated with minimal active diaphragmatic contribution to ventilation.7 ± 3 8. and our hypotheses should be addressed in formal validation studies.0 ± 1.2 12.6 1.7 3.8 ± 5.0 ± 2.7 ± 2.40–7.43 (7.41–7.42 (7.9 c Ppeak (cm H2O) 16.8 ± 1.0 ± 0.1a 21. 13).44) 7.1 (cm H2O) 2.0 ± 0.46) Tidal Volume (mL/kg) 8. whereas relative and absolute ventilation of nondependent regions (Vt%nondep and Vtnondep.4 p0.7 9.2 ± 3. At higher positive end-expiratory pressure (PEEP) levels.1 ± 6.40–7.42 (7.Clinical Investigations pH 7.7 7. most diaphragm displacement occurs in nondependent areas. respectively) decreased (A and B).42) 7.7 1. the proportion and the absolute value of tidal volume reaching dependent lung regions (Vt%dep and Vtdep. During controlled or highassist MV. in acute respiratory distress syndrome patients undergoing pressure support ventilation.1 ± 2. with minimal movement at the most dependent level.

u.5 cm H2O).6 (0. The fact that absolute tidal volume reaching dependent lung regions did not vary between the two PSV levels might seem in contrast with ventilation redistribution results.0–3.. absolute tidal volume decreased. regional compliance unlikely changed between the two PSV phases as they lasted 20 minutes and were performed at the same PEEP level. regional changes in Vt must have reflected changes in regional driving pressure.6) 0.9 3. substantial diaphragmatic contribution to ventilation and higher p0. his or her p0. Vt%dep = end-inspiratory fraction of tidal ventilation reaching dependent lung region.9 ± 1.00043 0. Hnondep = ventilation heterogeneity in nondependent lung region. a p < 0. is a function of regional driving pressure and lung compliance (24). Regional driving pressure during PSV is the sum of mechanical ventilatory support (which significantly decreased during PSVlow) plus patient’s regional inspiratory effort.f (n = 10) 55 ± 8 Healthy controls (n = 15) 36 ± 15a.00035 1.0c — 3.9 ± 1.f 1.00181 ± 0.1 < 2 cm H2O. Thus.00157 ± 0. Vtdep = tidal volume reaching dependent lung region. In our study. e p < 0.4 (1.00031 0. PSVlow.e.4–3.00121 ± 0. f p < 0. Another possible explanation is that during PSVhigh. indeed. d p < 0. Hdep = ventilation heterogeneity in dependent lung region. with 1670 www.05 vs healthy controls (Mann-Whitney U test).  Effects of Different Positive End-Expiratory Pressure and Support Levels on Patient Electrical Impedance Tomography Data Study Phasea Vt%nondep Vt%dep (%) (%) Vtnondep Vtdep (mL/kg) (mL/kg) Vt%nondep/ Vt%dep Hnondep Hdep ΔEELIgl (a.0 ± 1. PSVhigh.8a 41 ± 16 42 ± 15 32 ± 16c.1 (1. Data are expressed as mean ± sd or median (interquartile range) for nonnormally distributed variables.1).8) 0.1 ± 1.00117 ± 0.1 (1. Pressure support ventilation (PSV)low was chosen to obtain pressure generated during the first 100 ms of inspiration (p0.7a 2.ccmjournal. At the opposite.1 6.f (n = 10) PEEPhigh 59 ± 16 (n = 10) PSVlow 58 ± 9 (n = 10) PSVhigh 68 ± 10c.f 45 ± 8 5.1 = 2 cm H2O during PSVhigh. PEEP = positive end-expiratory pressure.6 ± 1.8–2.00037c 0. instead.1) values (i.5 (1.00176 ± 0.00032 2.1 ± 2. patients’ respiratory muscles could have been over-assisted and their July 2013 • Volume 41 • Number 7 . We observed that lower PSV levels are associated with increased proportion of tidal ventilation reaching gravitationally dependent lung regions. and inspiratory and expiratory triggers unchanged. whereas PSVhigh was selected as a resting condition with p0.0–2.00034 3.0 2.0) 0. in our study. a Figure 3.01 vs PEEPhigh (Wilcoxon’s test). obtained by multiplication of Vt%dep × expiratory Vt measured by the ventilator. and PEEPhigh (Table 2) were performed randomly in each patient for 20 min leaving Fio2 value.3)d. inspiratory ramp. c p < 0. obtained by multiplication of Vt%nondep × expiratory Vt measured by the ventilator.01 vs PEEPhigh (paired t test).00039c PSV = pressure support ventilation.) Baseline 1485 ± 1755e –115 ± 1268 Baseline PEEPlow 65 ± 14a.Mauri et al Table 3. – = variables are not available for this study group.0–1. this result is in line with others described in this study. PEEPlow.00157 ± 0. This result is also in line with diaphragm characteristics long known by respiratory physiologists: more favorable anatomical shape and longer muscular fibers length of dependent regions of the diaphragm make their contraction stronger than nondependent (12).00040a 0. One patient had p0.f 5. Vt%nondep = end-inspiratory fraction of tidal ventilation reaching nondependent lung region. Regional tidal volume. unchanged tidal volume to dependent areas and with more homogeneous antero-posterior tidal distribution to values more similar to healthy volunteers. unchanged values of tidal volume reaching dependent regions likely indicate higher inspiratory force generated by dependent zones of the diaphragm in comparison with nondependent.9 ± 1.00128 ± 0.3)b 0.00103 ± 0. a measure of patient’s respiratory drive) ≥ 2 cm H2O. however. ΔEELIgl = relative changes of global end-expiratory lung impedance when PEEPlow or PSVhigh was considered baseline (see text for details). where.1 was one of the highest during PSVlow (3.4 — 1.01 vs PSVlow (paired t test).5 ± 1.01 vs PEEPlow (one sample z test).9 2. Vtnondep = tidal volume reaching nondependent lung region. b p < 0.01 vs PSVlow (Wilcoxon’s test).

thus.Clinical Investigations PSVlow. Gattinoni et al (26) found that ICU mortality was greater among patients with During PSVlow. As we did not perform direct in comparison with that observed among patients with a lower proportion of nonaerated lung tissue. we cannot draw any definitive conclusion on underlying mechanisms. while it did not change during PEEPhigh. and.e.. ultimately. Published studies study phases. even in the presence of a similar proportion of regional ventilation. in this study. It was developed in the early 1910s and heterogeneity of dependent lung regions increased during then adapted in the late 1970s as a medical technique (18). During lower PSV. it seems rational that higher ventilation fractions in dependent regions shall be associated with oxygenation. The correlation between oxygenation and the measured fraction of ventilation reaching dependent lung regions is not unexpected. as in Figure 4. VILI. Ventilation heterogeneity of nondependent lung regions tilation reaching collapsed dependent lung regions and lower decreased both during higher PEEP and lower PSV. regional tidal volume absolute values must be dent lung regions. *p < 0.. Critical Care Medicine www.ccmjournal. CT scan studies regional tidal volume (Vtdep) (C and D). relative redistribution was likely caused by higher force generated by dependent diaphragm regions that might have caused increased regional mechanical stress and heterogeneity (12). respectively) decreased (A and B) (see text for explanation details). A possible explanation measures of regional diaphragm contractility. which increases regional lung mechanics homogeneity (15). diaphragm contraction could have increased globally and along all respiratory cycles. During these antero-posterior ventilation homogeneity.05. the observed Vt redistribution. inflammation. Vt read by the ventilator) with a regional one (i. end-organ dysfunction. by the way. 9). Dependent lung regions receive the highest fraction of lung perfusion (7.e. are the most prone bution of impedance within a body from measurements on to develop barotrauma and VILI (25). instead. HC = healthy controls. and the inference that the behavior of all lung areas showed that nonsurvivors had lower proportions of tidal venis equal to the single slice imaged by EIT may be misleading. ventilation its surface (14–17). death (3. we Vt%dep). which. decreased Hnondep). This might be related to different mechanisms involved in tidal redistribution during PEEPhigh and PSVlow: PEEP-associated tidal ventilation redistribution might have followed increased EELIgl and alveolar recruitment. patients with ARDS undergoing lung CT scans. At variance. for this finding is that in patients with larger collapse of depenFurthermore. tidal volume redistribution reduced the amount and our results might suggest that MV strategies that increase of air reaching the baby lung (6) and might have avoided ventilation of dependent lung regions may be beneficial in excessive overdistension (barotrauma) (23) and decreased patients with ARDS (9). MV is mainly distributed to the nondepentaken with caution as they combine a “global” respiratory dent aerated lung that becomes at higher risk of developing variable (i.e. instead. whereas relative and absolute ventilation of nondependent regions showed that the magnitude of (Vt%nondep and Vtnondep. lower pressure support ventilation (PSV) our results. leading to a higher proportion of collapsed lung tissue at baseline PEEP.01 and †p < 1671 . 8). In keeping with these results. In patients with acute respiratory distress syndrome . heterogeneity of regional mechanics in nondependent areas EIT uses alternate electrical currents to measure the distri(i.. lung aeration loss is related to ARDS severity: in a study on 68 diaphragm might have only been triggering the ventilator. was associated with increased ventilation fraction reaching dependent lung regions (Vt%dep) and unchanged Finally.

for monitoring of regional ventilation in spontaneously breathing patients). probably because of better ventilationperfusion matching.g.g.86–0. surgical. ACKNOWLEDGMENTS We thank all the patients who participated in the study and their families. EIT also has many important limitations: it lacks standardization. However. recruitment at higher PEEP levels and increased diaphragm activity at lower PSV).92 (range 0. the number of publications on EIT in international scientific journals is rapidly growing: researchers are 1672 CONCLUSIONS In patients with ARDS. 5) Although interesting. and since then. We thank all the San Gerardo Hospital General and Neurosurgical ICU staff for their work and support. and the assumption that EIT data represent the entire lung may be clinically misleading. we should have added one extra study phase (PSV level tested during study phase 3 and PEEPhigh). and it can only track changes of lung air content and not of lung tissue content. for daily assessment of regional lung mechanics) or feasible (e. We thank Eckhard Teschner of Dräger Medical GmbH.. EIT is not a substitute for CT scan.5 hrs). 2) EIT offers data only from a single biconvex slice of the lungs. and these characteristics enhanced EIT use as a monitoring technique for intensive care.e. Our study has a few important limitations: 1) We studied patients with ARDS when switched from controlled MV to PSV as per clinical decision (i.Mauri et al Figure 5. the behavior of all other lung areas can only be inferred. To do so.e. higher force generated by diaphragm vs. and relatively simple to use. researchers observed that EIT was particularly suited to image the lungs. after switching to assisted MV. 16). with this study design. As a matter of fact. July 2013 • Volume 41 • Number 7 . and emergency patients (14. 10–15 cm thick (15–17).ccmjournal. PEEP = positive end-expiratory pressure. lower pressure support and higher PEEP levels induce redistribution of tidal ventilation fractions from nondependent to dependent lung regions and increase antero-posterior ventilation homogeneity by different mechanisms (i. The association between ARDS patients’ outcome and ventilator settings that induce dependent redistribution of tidal ventilation remains to be established. the time spent by each patient with the EIT electrodes belt was already close to that suggested by manufacturer to avoid skin lesions (1–1. PSV = pressure support ventilation. spatial resolution is low. EIT lung imaging is reliable. Hinz et al (27) reported a highly significant linear correlation between regional ventilation measured by EIT and scintigraphy scanning with R2 of 0. Pao2/Fio2 ratios were significantly correlated in acute respiratory distress syndrome patients with ventilation of dependent lung regions (Vt%dep) during all study phases. but we used reasonable surrogates (i.1).e. particularly interested in EIT’s role in titration and personalization of lung protective MV (15). during both controlled and assisted MV. induced ARDS in 12 pigs. In conclusion. rapid. In that study. by oleic acid and measured tidal ventilation distribution by EIT and by ventilation scintigraphy using 99mTc-labeled carbon particles. who deserve to be formally studied in the future. intubated and mechanically ventilated. the above mentioned correlation between ventilation distribution and mortality needs further validation because of the small sample size. 4 ± 5 days after diagnosis) and studied a population with mild-to-moderate lung injury severity. Studies adopting two or more EIT monitoring devices simultaneously applied at different chest levels may yield more accurate and global results..e. ΔEELIgl and p0. our results may not apply to patients with more severe ARDS and with larger and thicker collapse of dependent regions. As a consequence. 3) Our study design did not allow us to test the interaction between PEEP and pressure support on regional distribution of tidal ventilation. regional alveolar recruitment). Lübeck.. 4) We did not directly measure the mechanisms at the basis of the observed redistribution phenomena (i. but results from the present and previous studies (15) seem to indicate that EIT might represent an alternative to CT scans when they are either not appropriate (e.. EIT has been validated in several preclinical lung imaging studies: Hinz et al (27).. Thus. Therefore. noninvasive.. for example. Soon after.97).

et al: Tidal recruitment assessed by electrical impedance tomography and computed tomography in a porcine model of lung injury. Cronan M. et al: Lung recruitment in patients with the acute respiratory distress syndrome. Crit Care Med 2012. 39:2173–2176 11. Pelosi P. where do we want to go and what do we need to get there? Physiol Meas 2012. Crit Care Med 2011. Anesthesiology 2012. 40:903–911 17. Solà J. et al: Relationships between lung computed tomographic density. et al: Whither lung EIT: Where are we. Froese AB. Anesthesiology 1988. Chest 2003. 20:597–604 21. Kaczka DW. Mauri T. Radke OC. Vitale G. Anesthesiology 1974. et al: Lung regional metabolic activity and gas volume changes induced by tidal ventilation in patients with acute lung injury. Zinserling J. Romand JA. et al: An expanded definition of the adult respiratory distress syndrome. West JB. et al: Analysis of regional mechanics in canine lung injury using forced oscillations and 3D image registration. measured with radioactive carbon dioxide. JAMA 1993. D’Andrea L. 269:2122–2127 5. J Appl Physiol 1960. Am J Respir Crit Care Med 1995. Murray JF. 307:2526–2533 20. Crit Care Med 2008. Christensen GE. and Emilio Carmignotto of Dräger Medical Italia SpA. Coppadoro A. Gattinoni L. et al: Vertical gradient of regional lung inflation in adult respiratory distress syndrome. Gerscovich EO. 342:1334–1349 2. Am J Respir Crit Care Med 1994. Am J Respir Crit Care Med 2008. Cereda M. Pelosi P. The ARDS Definition Task Force: Acute respiratory distress syndrome. Amato MB. Gattinoni L. Vidal Melo MF. Dudykevych T. Ann Biomed Eng 2011. Elke 1673 . 69:824–832 6. 149:8–13 12. 116:1227–1234 14. Boyd TK. 31:S85–S95 24. 38:1300–1307 23. Musch G. Crit Care Med 2010. et al: Regional overdistension identified with electrical impedance tomography in the perflubron-treated lung. et al: Pressure support ventilation in patients with acute lung injury. JAMA 2012. 18: 29–34 16. Muders T. Bellani G. Gattinoni L. Bellani G. et al: Relation between shunt. Marcora B. Grychtol B. Pelosi P. The Berlin definition. Bellani G. Pulletz S.ccmjournal. 138:720–723 22. Dollery CT: Distribution of blood flow and ventilation-perfusion ratio in the lung. Cressoni M. Martin JB. Crit Care Med 2002. 30:1747–1752 26. McGahan JP. Gattinoni L. Intensive Care Med 2005. Cortinovis B. 183:1193–1199 4. et al: Effects of positive end-expiratory pressure on regional distribution of tidal volume and recruitment in adult respiratory distress syndrome. for their technical support. 39:1112–1124 25. Schneider T. Italy. N Engl J Med 2006. J Ultrasound Med 2001. et al: Role of absolute lung volume to assess alveolar recruitment in acute respiratory distress syndrome patients. Guerra L. et al: Air cysts and bronchiectasis prevail in nondependent areas in severe acute respiratory distress syndrome: A computed tomographic study of ventilator-associated changes. et al: Pentraxin 3 in acute respiratory distress syndrome: An early marker of severity. aeration. 15:405–410 8. Maisch S. Physiol Meas 2010. Caironi P. Ware LB. Curr Opin Crit Care 2012. Bryan AC: Effects of anesthesia and paralysis on diaphragmatic mechanics in man. Matthay MA. Gattinoni L. Pesenti A. Pesenti A: Imaging in acute lung injury and acute respiratory distress syndrome. REFERENCES 1. et al: Regional effects and mechanism of positive end-expiratory pressure in early adult respiratory distress syndrome. Am J Respir Crit Care Med 2011. D’Andrea L. et al: Regional ventilation by electrical impedance tomography: A comparison with ventilation scintigraphy in pigs. Arnold JH. Cao K. Respirology 2011. et al: Ultrasonographic evaluation of diaphragmatic motion. Neumann P. Corsico (MI). Musch G. Crit Care Med 2000. Wolf GK. Crotti S. 354:1775–1786 27. 33:679–694 19. Luce JM. Hinz J. and PEEP in acute respiratory failure. et al: Heart-lung interactions measured by electrical impedance tomography.Clinical Investigations Germany. 16:523–531 18. gas exchange. Bellani G. 151:1807–1814 10. Reifferscheid F. Patroniti N. Matthay MA: The acute respiratory distress syndrome. Treggiari MM. Bohm SH. Mauri T. 31:776–784 7. N Engl J Med 2000. 177:292–300 9. et al: Regional ventilation distribution determined by electrical impedance tomography: Reproducibility and effects of posture and chest plane. 124:314–322 Critical Care Medicine www. Adler A. Heller AR. et al: Spontaneous breathing during general anesthesia prevents the ventral redistribution of ventilation as detected by electrical impedance tomography: A randomized trial. Luepschen H. Am Rev Respir Dis 1988. Bellani G. 41:242–255 13. 28:1269–1275 15. and perfusion in experimental acute lung injury. 36:2302–2308 3. Bombino M. Foti G. Pesenti A: The concept of “baby lung”.