Contents lists available at SciVerse ScienceDirect
Respiratory Physiology & Neurobiology
jou rn al h omepa g e: www.elsev ier .com/ locate / resphys io l
ssessment of extravascular lung water by quantitative ultrasound and CTn isolated bovine lung
rancesco Corradia, Lorenzo Ballb, Claudia Brusascob, Anna Maria Riccioc, Michele Baroffioc,iulio Boviod, Paolo Pelosib, Vito Brusascoc,∗
Dipartimento Cardio-Nefro-Polmonare, Azienda Ospedaliera Universitaria di Parma, 43126 Parma, ItalyDipartimento di Scienze Chirurgiche e Diagnostiche Integrate, Università di Genova, 16132 Genoa, ItalyDipartimento di Medicina Interna, Università di Genova, 16132 Genoa, ItalyDipartimento dei Servizi, IRCCS AUO San Martino – IST, 16132 Genoa, Italy
a r t i c l e i n f o
rticle history:ccepted 4 April 2013
eywords:ray-scale analysisisual scoring
a b s t r a c t
Lung ultrasonography (LUS) and computed tomography (CT) were compared for quantitative assessmentof extravascular lung water (EVLW) in 10 isolated bovine lung lobes. LUS and CT were obtained at differentinflation pressures before and after instillation with known amounts of hypotonic saline. A video-basedquantitative LUS analysis was superior to both single-frame quantitative analysis and visual scoring inthe assessment of EVLW. Video-based mean LUS intensity was strongly correlated with EVLW density
2 2 2
uantitative imagingomputed tomographyulmonary edemahysical density
(r = 0.87) but weakly correlated with mean CT attenuation (r = 0.49) and physical density (r = 0.49).Mean CT attenuation was weakly correlated with EVLW density (r2 = 0.62) but strongly correlated withphysical density (r2 = 0.99). When the effect of physical density was removed by partial correlation analy-sis, EVLW density was significantly correlated with video-based LUS intensity (r2 = 0.75) but not mean CTattenuation (r2 = 0.007). In conclusion, these findings suggest that quantitative LUS by video gray-scale
mor
-lines
analysis can assess EVLW
. Introduction
Increase in extravascular lung water (EVLW) is a hallmark ofcute lung injury of various etiologies (Martin et al., 2005; Warend Matthay, 2005). Monitoring of EVLW in critically ill patientsas prognostic value (Wiedemann et al., 2006) and is useful foruid control in acute lung injury (Sakka et al., 2002), acute respira-ory distress syndrome (Wiedemann, 2008), and acute heart failureSwedberg et al., 2005). Therefore, reliable quantification of EVLW
ight provide a valuable therapeutic guidance in intensive care.The methods currently used for measuring EVLW are either
naccurate, e.g., chest X-ray (Collins et al., 2006; Halperin et al.,985; Lichtenstein et al., 2004; Mant et al., 2009; Sivak et al., 1983),r invasive, e.g., thermo-dilution methods (Isakow and Schuster,006). Thus, new techniques able to quantify EVLW non-invasivelynd in real time are warranted. Lung ultrasonography (LUS) haseen proposed as a useful and simple method for the assessment
f EVLW based on the appearance of the so-called comet-tailsAgricola et al., 2005), more recently named B-lines (Volpicelli et al.,012). Indeed, B-lines visual score correlated in humans with chest
∗ Corresponding author at: Dipartimento di Medicina Interna, Università dienova, Viale Benedetto XV, 6, 16132 Genoa, Italy. Tel.: +39 0105554894;
X-ray visual scores (Jambrik et al., 2004; Lichtenstein et al., 1997),lung weight and density determined by quantitative computedtomography (CT) (Baldi et al., 2012), invasively determined EVLW(Picano et al., 2006), and also with lung weight directly measuredin experimental animals (Jambrik et al., 2010).
Visual analysis of LUS can be however difficult because lungshave often an inhomogeneous speckled appearance in criticallyill patients, reflection intensity varies depending on site of EVLWaccumulation and body position, and intra- or inter-observer vari-ability may be non-negligible. Computer-based techniques forinterpretation of ultrasonography have been applied to variousorgans (Doi, 2007) but not yet to lung.
Therefore, we planned the present study in isolated bovine lungto test the hypothesis that (1) the amount of EVLW can be reli-ably estimated by LUS echo-intensity, quantitatively determinedby computer-assisted gray-scale analysis and (2) quantitative LUSresults are comparable to those obtainable by quantitative CT andsuperior to visual scoring.
2. Materials and methods
2.1. Tissue preparation and procedure
Five bovine lungs were obtained from the local abattoir. Twolobes of each lung were isolated and intubated with a Rüschelit
uffed 4½ tube (Rüsch, Kernen, Germany) within 2 h from the deathf the animal. Lobes were ventilated (Savequick III ventilator, SIAREngineering International Group, Italy) at constant positive airwayressure (CPAP). After removal of unventilated or leaking regions,
obes were disconnected from the ventilator and weighed by a pre-ision balance (Sartorius BP3105-2 Scale, Göttingen, Germany) tobtain initial mass, excluding tube weight. Lobes were then ven-ilated in intermittent positive pressure mode (20 cmH2O plateauressure) and instilled with 0.45% NaCl solution. Three lobes were
nstilled with 0.5 ml/g of tissue, three with 1 ml/g, three with ml/g, and one was not instilled to serve as a control. Tubes werelamped and lobes left for 10 h at 4 ◦C, manually rotated everyour to promote homogeneous water distribution, then unclampednd aspirated to eliminate fluid from large airways. Finally, lobesere weighted to obtain instilled mass, excluding tube weight. The
mount of retained EVLW was calculated as the difference betweennstilled and initial mass.
Lobes were mechanically ventilated with 50% FiO2 and CPAPode. After a recruitment maneuver at constant airway pressure
f 40 cmH2O for 30 s, pressure was decreased in a stepwise mannero 30, 20, and 10 cm H2O. At each pressure, a CT helical scan and
LUS scan were obtained in rapid succession, as described in theollowing paragraphs.
.2. CT scanning
Helical CT scans were obtained by a Somatom 6 scannerSiemens AG, Erlangen, Germany) set at 130 kVp, 200 mAs, 6 × 1.0-
m collimation, 1.50 pitch factor, and 50-cm data collectioniameter. Reconstruction parameters were 5.0-mm slice thick-ess and medium smooth convolution kernel (B41s). Images weretored as uncompressed DICOM files at standard 512 × 512 pixelesolution.
Quantitative analysis was performed by Maluna® software (Uni-ersity of Gottingen, Germany) to determine lobe volume and meanttenuation in Hounsfield Units (HU), after manual segmentationrocedure and exclusion of airways with diameter >3 mm. Physicalensity of each lobe at the three inflation pressures was calculateds instilled mass divided by volume.
.3. Ultrasonography
Lung lobes were scanned by a MylabTM Five (Esaote, Ansaldo,enova, Italy) equipped with a 40-mm LA332 linear probe oper-ting at 10 MHz with the following settings: 28%-gain, 50%-timeain compensation, maximum emission power, removal of 2ndarmonic and automatic post-processing to avoid artifacts attenua-ion, and focus setting at 6-cm depth to maximize ultrasound beamollimation. A chloroprene water-filled bag with 1.5-cm diame-er was interposed to outdistance the transducer from the lobes,hus simulating the presence of chest wall (Soldati et al., 2011).ltrasonic probe was held perpendicularly to the lobe surface andressure was kept at the minimum required for full adherenceetween lobe, chloroprene bag, and transducer. Both probe andhloroprene bag were covered with a thin layer of ultrasound gel.hree faces of each lobe were scanned, at each inflation pressure,s single-frame or 80 video-recorded images over the mid surface.mages were stored as uncompressed DICOM files, as 8-bit graynits (GU) ranging from 0 (black) to 255 (white) in an array of74 × 470 pixels.
Three methods were used for the analysis of LUS (Fig. 1),.e., the visual scoring method proposed by Jambrik et al. (2010)
nd two quantitative methods based on computer-assisted gray-cale analysis (Adobe Photoshop Extended CS6 graphic package).n a still-frame quantitative method, the whole lung surface wascanned and a single image was randomly chosen. Within each
10 107 221 114 107
a Non-instilled control lobe.
ultrasound image, a region of interest was selected beginning fromthe chloroprene–lung interface and the first clearly visible horizon-tal reflection artifact. Images showing signs of incomplete adhesionbetween the probe and lung surface were discarded. The mean echointensity of each region of interest was determined and the meanof three lobe faces was retained for analysis. In a video-based quan-titative method, the whole lung surface was scanned moving theprobe at a perpendicular angle and constant velocity. The proce-dure was repeated for each face of the lobe. Afterwards, 80 frameswere stacked in a single image by a pixel-per-pixel mean algorithm,to obtain a mean value representative of the entire lobe side. Themean echo intensity of three lobe faces was retained for analysis.
2.4. EVLW density
EVLW density was defined, for each lobe and pressure, as theratio between EVLW and CT-determined volume and expressed asg/ml. This was taken as a variable for correlation analyses becausethe water-to-air ratio is more relevant than the absolute wateramount as a physical determinant of both CT and LUS imaging.
2.5. Statistical analysis
Mean LUS intensities, CT attenuation, visual scores, EVLW den-sity, and physical density were normally distributed (p > 0.10 byD’Agostino-Pearson’s Omnibus test), thus parametric statisticswere used. The possibility of applying single regression analysisto pooled repeated-measures from different lobes was verified byan extra sum-of-squares F test. The relationships between vari-ables were tested for significance by Pearson’s correlation analysis.Partial correlation coefficients were also calculated between quan-titative LUS, CT, and EVLW density to eliminate the influence ofphysical density. All tests were two-tailed and statistical signifi-cance was set at p < 0.05. All statistical analyses were performedwith SPSS, version 20.0 (IBM Corp.).
3. Results
The amount of EVLW retained by saline-instilled lobes rangedbetween 16 and 107% of the initial weight (Table 1).
Control lobe had a quantitative LUS < 50 GU in both single-frame and video-based methods and a visual score of 0 at allpressures, confirming that LUS cannot visualize normally aeratedlung parenchyma. The control lobe was thus considered as outlierand excluded from statistical analysis. As compared with controllobe, saline-instilled lobes showed consistent increments of CTmean attenuation and LUS intensity (Fig. 2).
Among LUS methods, the video-based one provided data thatwere most closely correlated with either EVLW or physical density(Table 2). The linear regression of mean video-based LUS inten-sity vs. mean EVLW density was not significantly different among
246 F. Corradi et al. / Respiratory Physiology & Neurobiology 187 (2013) 244– 249
Fig. 1. Methods used for lung ultrasonography. (a) Visual score method: count of vertical artifacts (arrows); (b) still-frame quantitative method with manual segmentationof single image (upper bordered panel) and gray-scale frequency distribution (lower borbordered panel) and gray-scale frequency distribution (lower bordered panel).
Table 2Correlations between lung ultrasonography (LUS) data and EVLW or physicaldensity.
LUS method
Visual score Still frame, GU Video, GU
EVLW density, g/ml 0.34** 0.38** 0.87***
Physical density, g/ml 0.08 0.22* 0.49***
GU, gray-units; values are r2.* p < 0.05.
** p < 0.01.
icbd
mb(wpdsp
4
tnv(i
4.2. Comments on results
*** p < 0.001.
ndividual lobes for slope (p = 0.68), or intercept (p = 0.45), or theirombination (p = 0.08). Therefore, the relationship between video-ased quantitative LUS intensity and EVLW density was adequatelyescribed by a single linear regression.
The correlation between mean video-based LUS intensity and CTean attenuation was significant but weak (Fig. 3). Mean video-
ased LUS intensity was more strongly correlated with EVLWr2 = 0.87) than physical density (r2 = 0.49), whereas the oppositeas the case for CT mean attenuation (Fig. 4). Once the effect ofhysical density was removed by partial correlation analysis, EVLWensity was significantly correlated with video-based LUS inten-ity (r2 = 0.75, p < 0.001) but not mean CT attenuation (r2 = 0.007,
= 0.668).
. Discussion
The major findings of this study in isolated bovine lung arehat quantitative LUS intensity, determined by a video-based tech-ique, (1) was correlated with EVLW density more closely thanisual score, or single-frame quantitative LUS, or quantitative CT
2) unlike quantitative CT, provided an estimate of EVLW that wasndependent of physical density.
dered panel); (c) video quantitative method obtained by staking 80 frames (upper
4.1. Comments on methodology
In this model, pulmonary edema was induced in isolated bovinelung by instillation of hypotonic saline into the airways. This isopposite to what happens in natural conditions, where fluid is mov-ing from blood vessels to interstitial space and, from there, intothe alveoli. We used hypotonic saline in an attempt to have partof instilled fluid moving into the interstitial space. Comparison ofultramicroscopic images of control lobe and one saline-instilledlobe (Fig. 2) suggests that this might have happened, althoughquantitative evaluation of interstitial thickness was not possible.Our LUS findings were characterized by appearance of B-lines com-parable to those described in pulmonary disorders characterizedby an increase of EVLW, such as pulmonary edema (Agricola et al.,2005) or acute respiratory distress syndrome (Copetti et al., 2008;Lichtenstein et al., 2004), and in experimental models of acute lunginjury (Gargani et al., 2007; Rajan, 2007). The similarity betweenour findings in isolated lung and those of studies in different speciesand conditions may provide a rationale for creating suitable modelsfor translational studies.
LUS does not provide a direct measurement of EVLW but onlydescribes a phenomenon, i.e., the artifact produced by ultrasoundsat the water–air interface, as recently demonstrated even in non-biological specimens (Soldati et al., 2011). Thus, quantification ofEVLW by gray-scale analysis is based on the assumption that thenumber of water–air interfaces is proportional to the amount ofwater, independent of its physical density.
A limitation of the present study is that, due to the small sizeof lobes, EVLW was not preferentially distributed to dependentlung regions as it occurs in congestive pulmonary edema. Thus,no comparisons were possible between LUS and CT on regionalbasis.
In this study the LUS intensity was significantly correlated withquantitative CT data, thus one could have concluded that both
F. Corradi et al. / Respiratory Physiology & Neurobiology 187 (2013) 244– 249 247
Fig. 2. Images of control lobe (left) and one saline-instilled lobe (right). Upper panels: computed tomography with attenuation distribution. Middle panels: ultrasonographywith gray-scale distribution by video-based quantitative method. Lower panels: scanning electron microscopy (StereoScan 360 microscope, Leica Cambridge Instruments,U
mnattcdtoetb(vti
K) showing interstitial thickening in saline-instilled lobe.
ethods are similarly reliable for quantifying EVLW. However sig-ificant, the correlation between mean quantitative LUS intensitynd CT mean attenuation was rather weak (r2 = 0.49), suggestinghat the two methods are in large part sensitive to different fea-ures of lung pathology. Indeed, mean LUS intensity was stronglyorrelated with EVLW density but weakly correlated with physicalensity, whereas the opposite was the case for CT mean attenua-ion. There are physical reasons for this discrepancy. The intensityf echographic gray-scale image is proportional to the amount ofchogenic interfaces, with white indicating the maximum. Quanti-ative LUS is based on the analysis of B-lines, which are generatedy the presence of a very large number of water–air interfaces
Copetti et al., 2008; Soldati et al., 2009). By contrast, CT pro-ides a measure of physical density (Mull, 1984), thus detectinghe presence of water not because of its nature but because ofts weight, in addition of lung tissue weight. This interpretation
is supported by the results of partial correlation analysis show-ing that CT mean attenuation was not significantly correlated withEVLW density when the effect of physical density was removed.These data suggest that comparison of quantitative LUS and quan-titative CT have the potential of distinguishing increased watercontent from increased tissue content in lung disorders where pul-monary edema and parenchymal abnormalities may co-exist. Ina very recent study (Barskova et al., 2013) B-lines were observedin patients with early systemic sclerosis in the presence butalso, though less frequently, in the absence of HRCT-documentedinterstitial lung disease. This discrepancy may be because B-lines can be generated by either pulmonary fibrosis or vasculitis,
characteristic features of systemic sclerosis that are not alwayscoexisting. A specifically designed study may reveal whether quan-titative LUS can help distinguish these two pathophysiologicalmechanisms.
Fig. 3. Relationship between mean LUS by video-based quantitative method andmean CT attenuation. GU, gray-units; HU, Hounsfield units. Each triplet of symbolsrefers to one saline-instilled lobe at inflation pressures of 10, 20 and 30 cmH2O.R
sEcssotmd
ing, (2) studies to minimize the effects of rib cage on LUS signals,
Fbi
egression lines and 95% confidence limits for pooled data are shown.
Consistent with our hypothesis, quantitative LUS was clearlyuperior to visual scoring in detecting the presence of amounts ofVLW as small as 16% of initial weight. Until now, LUS has beenompletely dependent on operator’s visual interpretation. Scoringystems based on progression or regression of abnormal patterns,paced B-lines, coalescent B-lines, and consolidation are effectivenly in determining large changes in lung aeration (i.e., >600 ml), as
hose caused by positive end-expiratory pressure used for recruit-
ent maneuvers in ARDS (Bouhemad et al., 2011), but its accuracyiminishes dramatically for milder changes of lung aeration. In a
ig. 4. Upper panels: Relationships between mean LUS intensity by video-based quantitatetween mean CT attenuation and EVLW (left) or physical density (right). GU, gray-units
nflation pressures of 10, 20 and 30 cmH2O. Regression lines and 95% confidence limits fo
Neurobiology 187 (2013) 244– 249
recent study in vivo (Baldi et al., 2012) the correlation between LUSvisual score and CT-determined lung physical density was found tobe slightly better than the correlation between quantitative LUSintensity and physical density of the present study. Differencesbetween their and our study are that their LUS and CT measure-ments included vascular volume and lung weight was not directlymeasured. Moreover, quantitative LUS turns images into numbers,which is expected to make more objective the assessment of overallchanges of air-to-fluid ratio and lung aeration than visual scor-ing. Between the quantitative LUS methods used in the presentstudy, the video-based was superior to the single-frame method,likely because the latter cannot be representative of the wholelung.
From a technical point of view, quantitative LUS requires onlya computer and dedicated software with all system settings con-stant in every measurement. Moreover, the transducer used forquantitative LUS must offer a compromise between tissue penetra-tion and spatial resolution. In this study a 10-MHz transducer wasused because this frequency offers sufficient tissue penetration.From a clinical point of view, the close linear relationship betweenvideo-based LUS mean intensity and EVLW density suggests thatthe latter can be calculated from the former once all echogra-phy settings are adequately standardized. However, further studiesare necessary before this method is proposed for clinical applica-tion. These include (1) studies where pulmonary edema is inducedin a way that more closely simulates what happens in disease,such as increasing pulmonary capillary pressure by fluid load-
and (3) studies comparing quantitative LUS with dilution meth-ods in the presence of non-homogeneously distributed pulmonaryedema.
ive method and EVLW (left) or physical density (right). Lower panels: Relationships; HU, Hounsfield units. Each triplet of symbols refer to one saline-instilled lobe atr pooled data are shown.
In conclusion, the present study shows that quantitative LUSnalysis can detect the presence of even small amounts of EVLWore reliably that visual scoring or CT scanning. Unlike quanti-
ative CT scanning, quantitative LUS is not dependent of physicalensity. As compared with existing methods for EVLW assessment,uantitative LUS has the advantage of being objective, inexpensive,on-invasive, and radiation-free. Interventional human studies areeeded to assess the usefulness of quantitative LUS in combinationith quantitative CT for determining the relative roles of EVLW andarenchymal density in disease.
onflict of interest
The authors declare no conflict of interest.
eferences
gricola, E., Bove, T., Oppizzi, M., Marino, G., Zangrillo, A., Margonato, A., Picano, E.,2005. Ultrasound comet-tail images”: a marker of pulmonary edema: a com-parative study with wedge pressure and extravascular lung water. Chest 127,1690–1695.
aldi, G., Gargani, L., Abramo, A., D‘Errico, L., Caramella, D., Picano, E., Giunta, F.,Forfori, F., 2012. Lung water assessment by lung ultrasonography in intensivecare: a pilot study. Intensive Care Medicine.
arskova, T., Gargani, L., Guiducci, S., Randone, S.B., Bruni, C., Carnesecchi, G., Con-forti, M.L., Porta, F., Pignone, A., Caramella, D., Picano, E., Cerinic, M.M., 2013.Lung ultrasound for the screening of interstitial lung disease in very early
systemic sclerosis. Annals of the Rheumatic Diseases 72, 390–395.ouhemad, B., Brisson, H., Le-Guen, M., Arbelot, C., Lu, Q., Rouby, J.J., 2011. Bed-
side ultrasound assessment of positive end-expiratory pressure-induced lungrecruitment. American Journal of Respiratory and Critical Care Medicine 183,341–347.
ollins, S.P., Lindsell, C.J., Storrow, A.B., Abraham, W.T., 2006. Prevalence of negativechest radiography results in the emergency department patient with decompen-sated heart failure. Annals of Emergency Medicine 47, 13–18.
opetti, R., Soldati, G., Copetti, P., 2008. Chest sonography: a useful tool to dif-ferentiate acute cardiogenic pulmonary edema from acute respiratory distresssyndrome. Cardiovascular Ultrasound 6, 16.
oi, K., 2007. Computer-aided diagnosis in medical imaging: historical review, cur-rent status and future potential. Computerized Medical Imaging and Graphics31, 198–211.
argani, L., Lionetti, V., Di Cristofano, C., Bevilacqua, G., Recchia, F.A., Picano, E.,2007. Early detection of acute lung injury uncoupled to hypoxemia in pigs usingultrasound lung comets. Critical Care Medicine 35, 2769–2774.
alperin, B.D., Feeley, T.W., Mihm, F.G., Chiles, C., Guthaner, D.F., Blank, N.E., 1985.Evaluation of the portable chest roentgenogram for quantitating extravascular
lung water in critically ill adults. Chest 88, 649–652.
sakow, W., Schuster, D.P., 2006. Extravascular lung water measurements and hemo-dynamic monitoring in the critically ill: bedside alternatives to the pulmonaryartery catheter. American Journal of Physiology Lung Cellular and MolecularPhysiology 291, L1118–L1131.
Neurobiology 187 (2013) 244– 249 249
Jambrik, Z., Gargani, L., Adamicza, A., Kaszaki, J., Varga, A., Forster, T., Boros, M.,Picano, E., 2010. B-lines quantify the lung water content: a lung ultrasoundversus lung gravimetry study in acute lung injury. Ultrasound in Medicine andBiology 36, 2004–2010.
Jambrik, Z., Monti, S., Coppola, V., Agricola, E., Mottola, G., Miniati, M., Picano, E., 2004.Usefulness of ultrasound lung comets as a nonradiologic sign of extravascular
lung water. American Journal of Cardiology 93, 1265–1270.Lichtenstein, D., Goldstein, I., Mourgeon, E., Cluzel, P., Grenier, P., Rouby, J.J., 2004.
Comparative diagnostic performances of auscultation, chest radiography, andlung ultrasonography in acute respiratory distress syndrome. Anesthesiology100, 9–15.
Lichtenstein, D., Meziere, G., Biderman, P., Gepner, A., Barre, O., 1997. The comet-tailartifact. An ultrasound sign of alveolar-interstitial syndrome. American Journalof Respiratory and Critical Care Medicine 156, 1640–1646.
Mant, J., Doust, J., Roalfe, A., Barton, P., Cowie, M.R., Glasziou, P., Mant, D., McManus,R.J., Holder, R., Deeks, J., Fletcher, K., Qume, M., Sohanpal, S., Sanders, S., Hobbs,F.D., 2009. Systematic review and individual patient data meta-analysis of diag-nosis of heart failure, with modelling of implications of different diagnosticstrategies in primary care. Health Technology Assessment 13, 1–207, iii.
Martin, G.S., Eaton, S., Mealer, M., Moss, M., 2005. Extravascular lung water inpatients with severe sepsis: a prospective cohort study. Critical Care 9, R74–R82.
Mull, R.T., 1984. Mass estimates by computed tomography: physical density fromCT numbers. AJR. American Journal of Roentgenology 143, 1101–1104.
Picano, E., Frassi, F., Agricola, E., Gligorova, S., Gargani, L., Mottola, G., 2006. Ultra-sound lung comets: a clinically useful sign of extravascular lung water. Journalof the American Society of Echocardiography 19, 356–363.
Rajan, G.R., 2007. Ultrasound lung comets: a clinically useful sign in acute respira-tory distress syndrome/acute lung injury. Critical Care Medicine 35, 2869–2870.
Sakka, S.G., Klein, M., Reinhart, K., Meier-Hellmann, A., 2002. Prognostic value ofextravascular lung water in critically ill patients. Chest 122, 2080–2086.
Sivak, E.D., Richmond, B.J., O‘Donavan, P.B., Borkowski, G.P., 1983. Value of extravas-cular lung water measurement vs portable chest X-ray in the management ofpulmonary edema. Critical Care Medicine 11, 498–501.
Soldati, G., Copetti, R., Sher, S., 2009. Sonographic interstitial syndrome: the soundof lung water. Journal of Ultrasound in Medicine 28, 163–174.
Soldati, G., Giunta, V., Sher, S., Melosi, F., Dini, C., 2011. Synthetic” comets:a new look at lung sonography. Ultrasound in Medicine and Biology 37,1762–1770.
Swedberg, K., Cleland, J., Dargie, H., Drexler, H., Follath, F., Komajda, M., Tavazzi, L.,Smiseth, O.A., Gavazzi, A., Haverich, A., Hoes, A., Jaarsma, T., Korewicki, J., Levy,S., Linde, C., Lopez-Sendon, J.L., Nieminen, M.S., Pierard, L., Remme, W.J., 2005.Guidelines for the diagnosis and treatment of chronic heart failure: executive
summary (update 2005). Revista Espanola de Cardiologia 58, 1062–1092.Volpicelli, G., Elbarbary, M., Blaivas, M., Lichtenstein, D.A., Mathis, G., Kirkpatrick,
A.W., Melniker, L., Gargani, L., Noble, V.E., Via, G., Dean, A., Tsung, J.W., Soldati,G., Copetti, R., Bouhemad, B., Reissig, A., Agricola, E., Rouby, J.J., Arbelot, C., Liteplo,A., Sargsyan, A., Silva, F., Hoppmann, R., Breitkreutz, R., Seibel, A., Neri, L., Storti,E., Petrovic, T., 2012. International evidence-based recommendations for point-of-care lung ultrasound. Intensive Care Medicine 38, 577–591.
Ware, L.B., Matthay, M.A., 2005. Clinical practice. Acute pulmonary edema. The NewEngland journal of medicine 353, 2788–2796.
Wiedemann, H.P., 2008. A perspective on the fluids and catheters treatment trial(FACTT). Fluid restriction is superior in acute lung injury and ARDS. Cleveland
Wiedemann, H.P., Wheeler, A.P., Bernard, G.R., Thompson, B.T., Hayden, D., deBois-blanc, B., Connors Jr., A.F., Hite, R.D., Harabin, A.L., 2006. Comparison of twofluid-management strategies in acute lung injury. The New England Journal ofMedicine 354, 2564–2575.
