With the probe indicator pointing cephalad and the orientation marker on the upper left corner of the US display screen, superior structures will be oriented screen-left. The same superficial structures seen in pneumothorax evaluation (discussed below) should be seen in pleural effusion analysis, but due to the deeper field of view, they will often appear closer to the top of the screen (more superficially) and appear smaller in comparison. These include subcutaneous tissue, ribs and rib shadows, sliding lung (apposed parietal and visceral pleura), and B-lines.

A key anatomic landmark in the analysis of an effusion is the hemidiaphragm on the side of interest. When imaging the lower thoracic area, the diaphragm appears as a curved, band-like, smooth-bordered structure originating superficially and inferiorly and arching deep and superiorly (Fig. 10.3). Respirophasic motion of this structure, particularly in the spontaneously breathing patient, should aid in differentiating the diaphragm from aerated lung. Identifying the diaphragm with confidence will help the proceduralist locate an appropriate site for thoracentesis or tube thoracostomy. As free fluid (e.g., ascites and cystic fluid collections) can also be located below the diaphragm, identifying anechoic fluid while scanning the lower thorax is insufficient to diagnose a pleural effusion. The hemi-diaphragm must first be identified to ensure that the apparent fluid identified is cephalad to the hemidiaphragm, thus preventing nondiagnostic or erroneous sub-diaphragmatic punctures with potentially serious complications.

On the left side, concurrent pericardial effusion can sometimes be seen (Fig. 10.2). When imaging the thorax from the mid-axillary line, Atkinson et al. [21] reported that the ultrasonographic signature of the thoracic spine – a densely echoic line extending cephalad from the deep portion of the diaphragm – is only visible when pleural fluid has displaced the normally interposed lung. They argue that visualization of the deep and posterior spine (termed V-line) is strongly supportive of the presence of pleural fluid (Fig. 10.3).

Heterogeneously echoic structures below the diaphragm are the spleen (left side) or the liver (right side). Ultrasonographic characteristics of pleural fluid can lend clues to the etiology of the fluid itself. Pleural fluid will generally appear anechoic, though the presence of cells, septations, clot, or air in the pleural space can cause a heterogeneous appearance (Figs. 10.4 and 10.5). Free-flowing fluid usually appears homogeneously anechoic. Complex fluid due to hemopneumothorax, parapneumonic, or malignant effusions can assume a range of ultrasonographic appearances. However, for reasons discussed below (see Evidence Review and Evidence-Based Use), use of the ultrasound appearance of pleural fluid alone to guide decision-making regarding whether diagnostic sampling or other procedures should occur should generally be avoided.

The principle goals for point of care US in the evaluation of pleural effusion should be to identify the presence or absence of pleural fluid and to determine the feasibility of diagnostic or therapeutic thoracentesis. The initial ultrasonographic approach to the patient should consider patient positioning and identification of key anatomic landmarks. If a patient can sit up in bed and lean forward over a table, a detailed examination of the lower posterior thorax can be performed to more thoroughly assess dependent fluid collections.

The area from the mid-axillary line to approximately mid-scapular line should be systematically scanned with a curvilinear or cardiac probe, and the diaphragm should be clearly identified on the right side (inferior) of the screen. The probe should be applied firmly to the skin, with adequate interposed gel, and oriented perpendicularly to the skin at all times. Inspiration, either patient or ventilator-driven, can help to identify the diaphragm through its downward motion. Inferior to the diaphragm, the spleen or liver should be identified. We often mark with a sterile marker the level of the diaphragm, keeping in mind that diaphragm motion will change that level slightly during the respiratory cycle. Scanning above the level of the diaphragm can then be performed to identify relatively anechoic collections consistent with pleural effusion. Observation of atelectatic or consolidated lung in the far field is helpful in confirming the presence of interposed pleural fluid between the parietal and visceral pleura (Fig. 10.2). If pleural fluid is identified, the depth of the effusion should be measured with the gradations on the screen. The approximate thickness of subcutaneous tissue should be noted and needle length adjusted accordingly if instrumentation is being considered. We then typically scan superiorly, identifying the vertical height of the effusion; the interface where the effusion diminishes in depth and eventually disappears, replaced by the bright reflection of air-filled lung, marks the superior edge of the effusion. This upper border of the effusion can also be marked with a sterile marker for reference.

Since images will only be obtained through intercostal windows, a skin mark may be made superficially just above the superior border of the lower rib of the appropriate intercostal space in order to guide subsequent thoracentesis or tube thoracostomy. The ribs should be palpated, and marking the course of the ribs above and below the procedural mark with a sterile marker can be helpful for orientation. We usually identify a relatively dependent site to maximize ability to extract pleural fluid while maintaining a margin of safety above the dome of the diaphragm. Keep in mind that the diaphragm level may rise during the course of pleural fluid drainage.

Pleural procedures should be carried out immediately following the ultrasound and site marking, and should be performed in the same patient position and directional orientation as the ultrasound (perpendicular to the skin surface). Excessively medial posterior (i.e., towards the spine) puncture sites for thoracentesis should be avoided due to the risk of puncture or laceration of the intercostal artery, which is more likely to be exposed below the rib in this region [22]. The site identified for thoracentesis should be above the diaphragm, in a relatively lateral position, and identified as having an adequate depth of fluid.

For chest tube placement, lateral positions in the “triangle of safety” should be considered, depending on patient circumstance and fluid location. The “triangle of safety” is denoted by the lateral border of the pectoralis major anteriorly, by the lateral border of the latissimus dorsi posteriorly, and by a line extended along the 5th intercostal space inferiorly (Fig. 10.6) [23]. The chosen area should be cleaned, then sterilely prepped and draped after ultrasound. We find that real-time ultrasound using a sterile probe cover is sometimes beneficial but is typically not necessary during most pleural procedures.

The volume of pleural fluid can be estimated with one of the several published techniques. Balik et al. [9] found the maximal distance between the diaphragm and the base of the lung at end-expiration correlates with the volume of effusion according to the following equation: volume of effusion (in mL) = 20 × distance of separation (in mm). Another group [24] found that end-expiratory intrapleural distances > 45 mm may reliably predict larger (around 800 mL) effusions. Remerand et al. [10] calculated pleural fluid volume by multiplying the area of the effusion (as measured in two ultrasonographic planes) in supine patients with free-flowing effusions at the midpoint between caudal and apical limits of the effusion. No one method has been exhaustively validated, but the method by Balik is frequently cited in the literature [4, 9–11].

Different disease processes of the pleura and pleural space may be elucidated sonographically based on varying acoustic profiles [20, 25]. Four patterns have been described to categorize the effusion in a manner that suggests a specific disease process. Anechoic fluid is homogeneously dark and is most consistent with transudative fluid. Complex nonseptated fluid contains more cellular debris, fibrin, blood, and is more likely to be exudative. Complex septated fluid may have organized locules, septations, and pleural adhesions, and is consistent with exudative fluid that likely requires tube thoracostomy [26]. Homogeneously echogenic fluid is generally consistent with empyema. Establishing appropriate gain settings is crucial to the correct interpretation of complex pleural fluid images, since overgain may cause relatively anechoic fluid to appear more echogenic.

Color Doppler has been reported as a means to distinguish small effusions from pleural thickening [27]. In patients suspected of having effusion based on radiography, Doppler analysis can be performed in addition to standard B-mode imaging. When B-mode images suggest the presence of effusion, a positive Doppler signal suggests the presence of fluid (dynamic) rather than pleural thickening (static). M-mode identification of the pleural fluid sinusoid sign can also be helpful in distinguishing small effusions from pleural thickening [28]. As most point-of-care pleural ultrasound in the intensive care unit is done to determine an optimal site for thoracentesis or tube thoracostomy, making a fine distinction based on ultrasound between a very small pleural effusion that is unlikely to be clinically significant and pleural thickening is not likely to be excessively important in the critical care setting.

A growing body of evidence supports the use of US in routine evaluation and management of pleural effusions in the critically ill. As a diagnostic tool, US outperforms conventional chest radiography. Diacon et al. [3] prospectively compared clinical exam and CXR to US for identifying a site of at least 10 mm of fluid for thoracentesis in 255 cases for possible instrumentation. US increased the number of accurate sites detected by 26 % and prevented possible organ puncture in 15 % of clinically and radiographically determined sites. Kocijancic et al. [25] determined that US had a positive predictive value of 92 % for detecting small effusions. In the surgical intensive care unit, a focused thoracic US examination demonstrated 83.6 % sensitivity, 100 % specificity, and 93.6 % accuracy for detecting pleural effusion [29].

In another study of pleural effusion in ICU patients the sensitivity of handheld US and standard radiography was 91 % and 74 %, respectively (not statistically significant). Specificity was 100 % for US versus 73 % for CXR (p = .008). The authors attribute three cases of US false negative to the inability to diagnose a small effusion (<50 mL) in patients with higher body mass indices (BMIs) [30]. Additionally, the ability of US to differentiate effusion from consolidation was demonstrated in a study of 97 ICU patients with findings suggestive of effusion on CXR. Effusion was better predicted by US (R² = 0.74, ROC 0.99) compared to CXR (R² = 0.51, ROC 0.70). The lack of lateral chest radiography in the ICU may weaken the ability of CXR to reliably identify pleural effusion [24].

Quantitation of pleural fluid volume by US has been well studied. Rothlin et al. [31] used portable handheld US in trauma patients to quickly rule in the presence of pleural fluid, which in that setting helped to diagnose traumatic hemothorax. More specific techniques to ascertain the precise volume of pleural fluid have been described by Balik et al. (discussed above) and Vignon et al. [24]. The latter group found that end-expiratory intrapleural distances of > 45 mm on the right and > 50 mm on the left predicted effusions > 800 mL with sensitivity of 94 % and 100 % and specificity of 76 % and 67 %, respectively, for each hemithorax. Their model was less robust for larger effusions. Roch et al. [32] conducted a prospective study of 42 ICU patients on mechanical ventilation who had effusion identified on CXR and in whom drainage was considered. An ultrasound-measured distance of 5 cm between the posterior chest wall and the lung base correlated with > 500 mL effusions with sensitivity 83 %, specificity 90 %, positive predictive value (PPV) 91 %, and negative predictive value (NPV) 82 %.

US has been shown to be a useful but imperfect diagnostic adjunct to clinical exam and standard radiography in the assessment of pleural fluid characteristics prior to thoracentesis. In a study of 118 thoracenteses in febrile medical ICU patients with effusions, 43 % of collections characterized as either complex septated, homogeneously echogenic, or complex nonseptated and relatively hyperechoic turned out to be empyema on initial or follow-up thoracentesis [12]. Ultrasonographic pleural fluid characteristics must be considered in the overall clinical context, however. Hirsch et al. evaluated 50 patients (24 % in the ICU) and, in 46 aspirations, 16 were complex septated, of which one was a transudate. Anechoic effusions were mixed (6 transudates, 5 exudates, 1 hematoma) [33]. Chen et al. [11] reviewed the US findings of 149 Light’s criteria-confirmed transudates and demonstrated that greater than half had a complex nonseptated appearance. Kearney et al. [13] found that US can better identify septations than CT scan, but that US characteristics did not reliably predict the nature of fluid in empyema. The authors found that some stage I exudates (pH > 7.2, lactate dehydrogenase [LDH] < 1,000 IU/l, glucose > 2.2 mmol/l, no organisms) contained septations and that some anechoic collections turned out to be frank pus when aspirated. These studies have considerable variation in the US protocol used, timing of imaging, and sampling. Further study is needed before US can be used as an effective prognostic tool for pleural effusion in ICU patients.

There is strong evidence that US increases the safety and yield of thoracentesis. Safety has been addressed in small retrospective studies of patients on antiplatelet medications (Abouzgheib 2012, n = 24; Dammert 2013, n = 43) who underwent US-guided pigtail catheter placement and suffered no complications [15, 18]. A larger study of 450 thoracenteses (67.8 % performed with US, 32.2 % performed without) revealed a significantly higher pneumothorax and tube thoracostomy rate of 4.1 % in the non-US group compared to 0.7 % in the US-guided group [17]. Lichtenstein et al. [16] prospectively studied 40 mechanically ventilated patients with effusions at least 15 mm and found no complications (pneumothorax or hemoptysis) in 45 procedures. However, in a prospective, randomized study, Kohan et al. [34] found no difference in the rate of complications in 205 patients who underwent thoracentesis with or without chest ultrasound. Most current studies of thoracentesis or tube thoracostomy involve US guidance in their protocols, reflecting that routine use of ultrasonography in this setting is effectively the standard of care.

The sensitivity and specificity of US operated by newly trained providers is variable, however. After an 8.5 h multimodality training session in US-naïve ICU residents, Chalumeau-Lemoine et al. demonstrated in 23 patients with confirmed pleural effusions a sensitivity of 58.3 % and specificity of 70 % [35]. This suggests the need for adequate and protocolized training in the use of point-of-care US. A questionnaire-based study of pulmonary and critical care fellowship directors identified significant interest in nonvascular applications of US (including evaluation of pleural effusion). Insufficient faculty expertise, lack of knowledge of the data supporting its use, and financial considerations were cited as barriers to more widespread use of US in critical care medicine fellowships [36]. There is also evidence that bedside US may facilitate care of patients with pleural effusion in nonindustrialized countries [37].