Densities in dependent lung regions worsen oxygenation in patients with acute respiratory distress syndrome. Identification of these densities requires examination using computed tomography (CT). In this study, the authors evaluated the use of transesophageal echocardiography (TEE) to estimate densities in the dependent lung.
Forty consecutive patients with acute lung injury or acute respiratory distress syndrome who underwent CT and TEE examination were included in this study. Densities in the lower left lung area were detected through the descending aorta by TEE. Density areas observed by TEE were compared with those obtained by CT. The effect of positive end-expiratory pressure (PEEP) application on density area was also evaluated.
Density areas in the dependent lung region measured by TEE were 12.0+/-6.1 cm2 (mean +/- SD) at mid esophageal position. Density areas evaluated using TEE in the left lung correlated significantly with those estimated with CT in the left and right lungs (P < 0.01 in both lungs). In addition, the authors observed a significant correlation between PaO2/FIO2 and density areas estimated using TEE (P < 0.05). During positive end-expiratory pressure application, the area of density estimated with TEE decreased and PaO2 improved.
The authors clearly demonstrated that it is possible to estimate the density area of the dependent left lung regions in patients with acute lung injury or acute respiratory distress syndrome using TEE. It is also possible to observe the changes of density areas during application of positive end-expiratory pressure.
Click on the links below to access all the ArticlePlus for this article.
Please note that ArticlePlus files may launch a viewer application outside of your web browser.
THE lungs of patients with acute respiratory distress syndrome (ARDS) appear as a patchwork of normal and dense areas separated by well-defined boundaries when viewed using computed tomography (CT) in many cases. Most pulmonary densities are found in dependent regions. 1–3These densities are known to reduce oxygenation in patients with ARDS and can be detected using CT but not clearly by conventional chest radiography.
It is not possible to observe intact lungs containing air using echocardiography. The aorta acts as an acoustic window on the left side. When a density area in the left lung increases and becomes adjacent to the aorta, it is possible to detect and estimate the density of the lesion and the lesion’s size via the descending aorta.
Transesophageal echocardiography (TEE) has been used to estimate hemodynamic variables in the critically ill. Acute lung injury (ALI) and ARDS are also major concerns in such patients. In this study, we evaluated whether it is possible to observe densities in the dependent lung region using TEE in clinical situations. In addition, positive end-expiratory pressure (PEEP) or prone position has been recommended to patients with ALI and ARDS. We also estimated how PEEP works on the dependent lung density area.
Materials and Methods
With local ethics committee approval (Ethical Review Committee, Hirosaki University School of Medicine, Hirosaki, Japan) and with informed consent by each patient, 40 consecutive patients with ALI and ARDS were enrolled in the study. Diagnoses of ALI and ARDS were made on the basis of the guidelines of the American–European Consensus Conference on ARDS. 4All patients underwent TEE investigation as part of the evaluation of their hemodynamic status. By means of TEE, the following numbers of diagnoses were detected: 10 hypovolemia, 2 endocarditis, 3 intracardiac cavity thrombi, and 2 pericardial effusions. During the investigation, the dependent area in the left lung was observed through the descending aorta.
Thirty men and 10 women patients were involved in this study. Patient characteristics were as follows: 55.9 ± 19.5 yr of age, 59.9 ± 13.0 kg, 163.0 ± 8.0 cm, and Pao2/Fio2of 199.8 ± 115.6. Sixteen patients presented after having pneumonia and 10 after having sepsis, 8 were postoperative, 4 were postcardiopulmonary resuscitation, and 2 were admitted after trauma. Bilateral infiltrations were seen on frontal chest radiograms and varied from moderate, mild, and severe in all patients. We could not find differences in the chest radiograms between the patients with dependent lung density and ones with no density. Ten patients died during intensive care unit stay.
The trachea of each patient was intubated and the lungs were mechanically ventilated before and during examination. Arterial pressure, central venous pressure, end-tidal carbon dioxide tension, and arterial oxygen saturation were monitored continuously. Morphine or midazolam, or both, were administrated for sedation. No PEEP was applied during the examination. Patient ventilation was supported by a Servo900C machine (Siemens, Solna, Sweden) or a Bear 1000 machine (IMI, Riverside, CA) using synchronized intermittent mandatory ventilation or pressure support ventilation. Fio2ranged from 0.4 to 1.0, depending on the gas analysis result. Time from tracheal intubation to CT examination was 90 ± 56 h (mean ± SD). TEE examination was performed within 24 h of CT scanning. A diagnosis of densities in dependent lung regions was made by radiologists in the hospital using CT by the definition of Gattinoni et al . 1
Echography was performed using an ultrasound system (Hewlett Packard SONOS 1500; Andover, MA) equipped with a 5-MHz 64-element transesophageal multiplane echoprobe and recorded on 0.5-in video tape. The lower left lung area was observed via the descending aorta using TEE. A homogenous condensed area in the left lung was considered to be density in the dependent lung region. Each lesion was observed at the ascending aorta and the mid esophageal and lower esophageal positions. The position was decided according to the patient echocardiography view. The mid esophageal position was determined when the four-chamber view was observed. The lower esophageal position was defined just before the transgastric position. Then, we rotated the probe 90° counterclockwise, keeping the same level, and observed the left lung. The image from the video was traced manually on the screen, and the area was calculated using a program equipped in an ultrasound system.
For CT examination a General Electric high light advantage scanner was used (General Electric, Waukesha, WI). Exposures were taken at 120 KV and 170 mA. Slice thickness was 10 mm, and the dimension of the pixel reconstruction matrix was 1.5 × 1.5 mm. Lung scanning was performed from the apex to the diaphragm. All images were observed and photographed at a window width of 1.600 Hounsfield units and level of −700 HU. After obtaining a frontal scan covering the chest, images at the same position as the TEE examination were evaluated. CT image was traced manually and the area was calculated using customized soft ware (J-MAC system; Fineworks, Sapporo, Japan). CT and TEE images were considered to be in the identical position when they showed the same anatomic cardiac structure.
In eight patients, PEEP of 5, 10, and 15 cm H2O was applied to improve oxygenation. Each level of PEEP was maintained for at least 15 min. Patients underwent ventilation with controlled mechanical ventilation using a tidal volume of 8 ml/kg and a respiratory rate of 20 breaths/min. The left dependent lung region was observed using TEE at the mid esophageal position during PEEP. At the end of each PEEP application, changes in the density area and gas analysis were evaluated.
All data are presented as mean ± SD and analyzed using the unpaired Student t test or one-way analysis of variance. The Dunnett posttest was performed if results of one-way analysis of variance were significant. The correlation coefficient between two variables was tested by linear regression analysis. A value of P < 0.05 was considered to be significant.
Interobserver and intraobserver variability were evaluated in 60 randomly selected video images. For comparisons, an investigator was blind to the results of the previous investigation. The deviation of each measurement from the mean of the two measurements was calculated for the observer variation. 5,6Bias and precision were also calculated.
Of the 40 patients enrolled, 14 did not have densities in the left dependent lung region as seen during CT scanning. In addition, no density was observed using TEE in these 14 patients. However, a reticular pattern and ground glass opacification were seen in these patients’ lungs via CT. 7Three dependent lung densities, two pleural effusions, and one bulla were found in the right lungs of these 14 patients. Twenty-six patients had a obvious area of density in the left dependent lung as seen during CT scanning. In all but two, this was confirmed using TEE. The density areas of the two patients were small, 2.4 cm2and 2.5 cm2, respectively. All patients who had a density area in the left lung also had density in the right lung. There was no difference in Pao2–Fio2between patients who had density and those with no density (179.1 ± 63.5 mmHg in patients with density, 192.9 ± 80.2 mmHg in patients with no density).
Figure 1depicts TEE and CT data in the same patient in whom density in the dependent lung lesion was seen. The density was outlined and the structures were labeled. The areas of densities in the dependent lung region measured 12.0 ± 6.1 cm2, 17.9 ± 8.4 cm2, and 19.2 ± 8.2 cm2(mean ± SD, n=24) at the ascending aorta and the mid esophageal and lower esophageal positions, respectively (fig. 2). As the densities approached the diaphragm, their area increased (P < 0.01). There was a significant correlation between the areas estimated using CT and TEE in both lungs (P < 0.01) (figs. 3 and 4). There was also a significant correlation between Pao2/Fio2and density area estimated using TEE at the mid esophageal position (P < 0.05) (fig. 5).
During PEEP, the dependent lung density area decreased in accordance with the increase of PEEP (figs. 6 and 7). Pao2increased as density area decreased (fig. 7). There was significant correlation between the rate of change in the dependent lung area and Pao2(fig. 8). Ultrasound image analysis showed an interobserver variability, bias, and precision as −0.01 ± 0.53 cm2, and 0.54, and 0.096, respectively. Intraobserver variability, bias, and precision were 0.27 ± 0.56 cm2, 0.53, and 0.098, respectively.
The pathologic characteristics of ARDS may change during the course of a disease, from edema to fibrosis and disruption of the alveolar septa. 7In patients with ARDS, densities may account for up to 70–80% of the lung field, depending on the severity of respiratory failure. 1Lung densities are concentrated primarily in the dependent regions. 8Pelosi et al . 9reported that increased lung weight caused by edema results in a collapse of lung regions along the vertical axis via transmission of hydrostatic forces. Thus, lesions appear primarily in the dependent regions of the lung. Such lesions can be detected using CT. However, CT scanning cannot discriminate between atelectasis, edema, and consolidation or a combination of these three different density-producing lesions. 10
The density areas in the left lung were observed at the aortic, mid esophageal, and lower esophageal positions. Area increased as the observing position approached the diaphragm (P < 0.01). Puybasset et al . 11reported that nonaerated areas were predominantly found in the lower esophageal position (juxtadiaphragmatic regions) using CT. Distribution of poorly aerated lung increased along the cephalocaudal axis for the overall lung. Puybasset et al . 11suggested that in addition to the decrease in transpulmonary pressure caused by an increased abdominal pressure, an anatomic factor is implicated in the cephalocaudal distribution of lung hyperdensities. 11
We found that the left lung density area estimated using TEE correlated significantly with the left lung density evaluated using CT. It is possible to observe the left lung through the descending aorta if the density is large enough and adjacent to the aorta. We could not detect density using TEE in two patients, and their density areas were not large enough to be adjacent to the aorta in CT. In addition, when the lesion is too large to observe through the descending aorta, the upper portion of the lesion cannot be estimated. Therefore, TEE may underestimate the area of extremely large lesions.
In this technique, only the left lung density area can be observed because of the left-sided deviation of the descending aorta. However, we found a significant correlation between left lung density area estimated using TEE and right lung density area using CT. ALI included injury to both the lung endothelium and the lung epithelium. Originally, in patients with ALI or ARDS, a bilateral infiltrate shadow must be seen on the chest radiograph. According to CT study data by Puybasset et al ., 11there is no difference in the mean value of nonaerated volume in both lungs in patients with ALI.
Fourteen patients in whom no density areas were observed in the left lung also showed low Pao2/Fio2.In these patients, we found reticular or ground-glass opacified lesions, or both, instead of a density area. Desai et al . 12reported that ground-glass opacification is one of the major CT patterns in ARDS patients and is more frequently seen in the anterior quadrants. Owens et al . 13reported that reticular patterns seen on follow-up CT scans in ARDS patients were most pronounced in areas that were densely consolidated previously. In the 40 patients in the current study, density areas in the dependent lung regions were major CT findings in patients with ALI and ARDS.
In this study, there was a significant correlation between Pao2/Fio2and left lung density area estimated with TEE in those patients with a density area. Gattinoni et al . 14reported that the areas of density estimated using CT correlated significantly with Pao2. They also demonstrated a significant correlation between shunt ratio and uninflated lung tissue, as seen on CT scans. Pelosi et al . 9suggested that the shunt fraction, because of the perfusion of the noninflated tissue, is probably the main cause of severe hypoxia in ARDS patients.
When the density areas exist in the dependent lung region, PEEP and placement of the patient in the prone position are recommended to improve oxygenation in these patients. 15–17Increasing PEEP caused progressive clearing of radiographic densities and increased the mass of normally inflated tissue in ARDS patients. 18Brunet et al . 19reported that PEEP increased “normally aerated” areas, decreased “nonaerated” areas, and did not change “poorly aerated” zones. In this study, we found that there was a decrease in density in the left lung lesions and oxygenation improved.
In conclusion, it is possible to observe densities in dependent left lung regions using TEE. This procedure showed the effects of PEEP on the density area in the left lung. The clinical usefulness of this method remains to be fully evaluated.
The authors thank Dr. Yasu Oka, Emeritus Professor of Anesthesia, Department of Anesthesiology, Albert Einstein College of Medicine, New York, New York, for her comments and encouragement.