Is the Lateral Decubitus Radiograph Necessary for the Management of a Parapneumonic Pleural Effusion?^* FREE TO VIEW

Approximately 50% of patients who are admitted to the hospital with community-acquired pneumonia will develop a pleural effusion.¹² While most parapneumonic effusions will resolve without specific therapy, approximately 10% will become complicated or progress to empyema.³ Because delays in the drainage of these clinically significant effusions have been associated with increased morbidity and mortality,⁴ prompt detection and accurate characterization of a parapneumonic effusion are important. Guidelines⁵ based on a consensus conference have comprehensively summarized the literature regarding parapneumonic effusions and have made specific recommendations for their management. Although this document reflects the “state of the art,” as the authors noted, most studies of parapneumonic effusions have been observational, often with a limited numbers of patients.

One of the recommendations in the guidelines was that all patients with a parapneumonic effusion should have a chest radiograph performed with a lateral decubitus view.⁵ This recommendation also was incorporated into the American Thoracic Society guidelines⁶ for the care of adults with community-acquired pneumonia. The rationale is that if the thickness of the effusion on the lateral decubitus view is < 1 cm, the effusion is small enough so that no further intervention is needed.¹ While there is good evidence that small effusions resolve without specific therapy,¹ the suggestion that a lateral decubitus radiograph is required to determine the appropriate course of action with respect to a parapneumonic effusion appears to have been made without sufficient supporting data. The purpose of this article is to review the indications and usage of the lateral decubitus chest radiograph and demonstrate its limited value in the setting of parapneumonic pleural effusions.

Radiographs of the chest made with the patient in the standard erect position usually will not reveal the presence of a pleural effusion < 300 to 500 mL in size, as the fluid usually pools in the posterior costophrenic sulcus.⁷⁸ The varying volumes reported may be due to differences in patient size, as a larger individual probably can “hide” a larger volume of fluid than can a smaller individual. The lateral decubitus view was first described by Merlo Gomez and Heidenreich,⁹ in 1924, as a technique to identify the presence of small pleural effusions. Its first English-language description was by Rigler,⁷ in 1931, who used this technique to confirm the presence of pleural effusions in a small series of patients despite the absence of a visible effusion using the standard erect views. Subsequently, other investigators demonstrated that the sensitivity of this view can be increased by placing a pillow under the patient’s pelvis, so that the thorax slopes downward away from the lung base.¹⁰ By injecting known amounts of fluid into cadavers, in 1973 Moskowitz et al⁸ demonstrated that effusions as small as 5 to 10 mL could be detected using the lateral decubitus view.

While the lateral decubitus view will often reveal very small effusions that have not been identified by erect chest radiographs, there generally is no reason that such a small parapneumonic effusion needs to be diagnosed. Light et al¹ reported that small parapneumonic effusions always resolved without any specific intervention. Difficulty in diagnosing a larger pleural effusion should be rare when a lateral erect radiograph is obtained. Taryle et al² found that 16 of 20 parapneumonic effusions (80%) were identifiable by radiographs obtained with posteroanterior and/or lateral erect views, while 4 effusions were identified on the lateral decubitus view only. One of these was a very small parapneumonic effusion that was probably not clinically significant, although it could be determined from the article how large the other three effusions were. In a multicenter study of empyemas,¹¹ difficulty in identifying the presence of pleural fluid was thought most often to be due to the lack of a lateral erect chest radiograph. Occasionally, a larger pleural effusion may be difficult to identify, if it is in a subpulmonic distribution or if there is a superimposed dense alveolar infiltrate. In such cases, the lateral decubitus radiograph may confirm the presence of a clinically significant effusion, although often a CT scan of the chest will provide more clinically useful information.

While the lateral decubitus view often can identify loculations, they can usually be ascertained by a standard erect view.³ The fundamental question is, “Is it necessary to know whether the fluid is loculated prior to performing a thoracentesis?” Although many loculated effusions resolve without drainage,⁴ the presence of loculations correlates with the development of a complicated pleural effusion or empyema.¹¹¹²¹³ The drainage of loculated effusions is sometimes necessary even in the face of the results of pleural fluid chemistry testing that do not suggest the need for tube thoracostomy.¹² However, since loculations occur more frequently with larger effusions,¹² it is not clear that the presence of loculations provides any more prognostic information than does the semiquantitative estimate of fluid volume based on chest radiographs obtained with the patient in the erect position.⁴ Even if the presence of loculations provides prognostic information, a thoracentesis would ordinarily be performed before deciding on the need for more aggressive therapy. If the thoracentesis is successful in removing all of the radiographically visible fluid, or if the fluid clearly shows the need for tube thoracostomy, the lateral decubitus view will yield no clinically useful information. In the case of an effusion that cannot be completely removed by thoracentesis, the presence of loculations can be presumed, and the subsequent management will be directed by the fluid characteristics. In some cases, the performance of tube thoracostomy may be difficult due to the presence of multiple and/or small loculations, and a CT scan of the chest will yield much more useful information than will radiograph made with a lateral decubitus view. Support for this line of thought (although not for omitting the lateral decubitus view) can be inferred from a published treatment algorithm¹⁴ that recommended the performance of a radiograph with the lateral decubitus view and then thoracentesis, whether or not the fluid is loculated. Interestingly, when one queries house officers about why they ordered a lateral decubitus radiograph, the invariable answer is to “see whether the fluid is loculated, because loculated fluid cannot be tapped without radiographic guidance.” I could find no support in the literature for this idea but have encountered it at several institutions. If a large pleural effusion is identifiable by the standard erect view chest radiographs and a physical examination, there should be little risk in performing a “blind” thoracentesis, whether the fluid is loculated or not. In the specific setting of multiloculated empyema, radiographic guidance by either CT scan or ultrasound can be invaluable in guiding the placement of small-bore drainage tubes.

This is the major reason that the lateral decubitus radiograph is recommended.¹⁵ What data support this? In 1980, Light et al¹ reported that all pleural effusions resulting in a < 1-cm layer seen on a radiograph with a lateral decubitus view resolved with antibiotic therapy only. This knowledge, combined with their experience that an effusion of this size could usually be tapped easily,¹⁵ resulted in a recommendation that a lateral decubitus radiograph be performed in all cases of suspected parapneumonic effusion.¹ However, clinically relevant estimates of the size of pleural effusions can be made without the use of the lateral decubitus radiograph.¹¹ Also, estimated fluid volume based on the lateral decubitus radiograph correlates poorly with the true volume. In a study comparing the accuracy of ultrasound and the lateral decubitus view predicting the volume of a pleural effusion, Eibenberger et al¹⁶ found that the mean prediction error using the lateral decubitus radiograph was 465 mL, compared to 224 mL using ultrasound. The actual fluid volume in three patients with an effusion that was 1.2 cm thick on the lateral decubitus radiograph ranged from approximately 350 to 1,400 mL.

Since empyemas that occupy only 20% of the hemithorax occur commonly,¹¹ it is clear that a lateral decubitus radiograph is not required to determine the need for sampling effusions of this size or larger. For smaller effusions, the work of Light et al¹ provides strong evidence that thoracentesis is not necessary for effusions that are < 1 cm thick on the lateral decubitus view. They also suggested that it may be difficult to perform a blind thoracentesis when the effusion is smaller than this.¹ Therefore, if the smallest effusion that can be easily tapped could be identified by radiographs with the standard posteroanterior and lateral views, then that size will correspond to the size that current recommendations suggest should be tapped.

In my experience, a blind thoracentesis is almost always easily performed if the column of fluid is > 5 cm in height at the posterior costophrenic sulcus. Smaller effusions leave little room for error. This is a more logical measurement to rely on than one obtained from the lateral decubitus radiograph, since it shows exactly how much fluid is present in the location where the needle usually will be placed. Of course, the height of the effusion cannot be measured directly since the inferior recess of the posterior costophrenic sulcus is obscured by the fluid. However, it can be estimated by comparing the height of the fluid above the contralateral posterior costophrenic sulcus. Unlike the domes of the diaphragm, the two sulci are usually at similar levels.

In order to assess the validity of this method, I compared the appearance of effusions imaged by radiographs with both lateral erect and lateral decubitus views in 20 patients (see “^Appendix”). As can be seen by the linear regression (Fig 1 ), the measurements obtained by both methods are closely correlated (p < 0.001; r² = 0.51), albeit with some scatter, presumably due to inexact quantification of the amount of fluid associated with both views. The regression line also demonstrates that 1 cm fluid on the lateral decubitus view correlates with approximately 5 cm fluid on the lateral erect view. While the two methods give similar results, the use of the standard lateral erect view obviates the need for the subsequent lateral decubitus view.

Despite the lack of evidence supporting the necessity of the lateral decubitus radiograph, it could be argued that it is inexpensive and without risk, and therefore should be performed. However, when a pleural effusion is found, performing the lateral decubitus radiograph generally necessitates sending the patient back to the radiology suite. Although therapy with antibiotics should never be delayed in order to perform a thoracentesis, a second trip to radiology could result in such a delay. This could detrimentally affect patient outcomes, since a delay in the start of antibiotic therapy of even a few hours may increase the risk of mortality from pneumonia.¹⁷ The performance of a radiograph with the lateral decubitus view also might delay thoracentesis, which has been associated with worsened outcomes in some patients.⁴¹⁸ While ideally the performance of the lateral decubitus radiograph would delay the thoracentesis only minimally, by the time the results are obtained, the attending physician may have left the hospital or the house officer may have moved onto the next person who has been admitted to the hospital.

Despite having worked at several institutions, I do not recall any pulmonary specialists who routinely have obtained lateral decubitus radiographs. In addition, at least one respected chest radiologist believes that the lateral decubitus chest radiograph is overused.¹⁹ Recommending, without adequate evidence, a practice that is apparently widely disagreed with, can breed skepticism about the entire guideline in question.

In summary, a parapneumonic effusion that is large enough to be potentially clinically significant can almost always be defined by chest radiographs with posteroanterior and lateral erect views. Patients with effusions clearly occupying > 20% of the hemithorax should have a thoracentesis performed, and thoracentesis should be considered strongly for smaller effusions that have a column of fluid of > 5 cm that is visible on the lateral erect radiograph. Since an effusion of this size correlates with an effusion with 1-cm thickness on the lateral decubitus view, an effusion smaller than this should resolve without drainage. If loculations are suspected, based either on the appearance of the standard radiographs or on the inability to completely drain the fluid, then a CT scan will define the pleural anatomy more accurately than will the lateral decubitus radiograph. Occasionally, the lateral decubitus radiograph can confirm the presence of fluid if there is a subpulmonic effusion or if the costophrenic angles are obscured by infiltrate. In most cases, however, it yields no clinically useful information but may unnecessarily delay proper therapy.

The radiographs of all patients with pleural effusions who had radiographs performed with a lateral decubitus view between January 1, 1999, and December 31, 2001, at the University of Connecticut Health Center were reviewed. Patients in whom radiographs with lateral decubitus views did not allow the measurement of the amount of fluid due to poor technical quality were excluded, as were patients with bilateral effusions that obscured all landmarks, those who had radiographs with no lateral erect view performed, those who had near complete filling of the hemithorax, or those in whom the fluid was loculated. In all patients, the two radiographs were performed within ≤ 27 h of each other. The height in centimeters of the pleural fluid above the inferior aspect of the contralateral posterior sulcus was measured on the lateral erect radiograph. These results were compared to the thickness of the pleural fluid layering measured on the lateral decubitus view. Linear regression analysis²⁰ was performed (Excel 97; Microsoft Inc; Redmond, WA). Radiographs were assessed while the reader was blinded to the results of the alternate view of the same patient. This study was approved by the Institutional Review Board of the University of Connecticut Health Center.

The author would like to thank Jonathan Fine, MD, for his critical review of the manuscript, and Jan Tate, MPH, for assistance with the statistical analysis.