Pleural ultrasonography versus chest radiography for the diagnosis of pneumothorax: review of the literature and meta-analysis

Abstract

Introduction: Ultrasonography is being increasingly utilized in acute care settings with expanding applications. Pneumothorax evaluation by ultrasonography is a fast, safe, easy and inexpensive alternative to chest radiographs. In this review, we provide a comprehensive analysis of the current literature comparing ultrasonography and chest radiography for the diagnosis of pneumothorax.

Methods: We searched English-language articles in MEDLINE, EMBASE and Cochrane Library dealing with both ultrasonography and chest radiography for diagnosis of pneumothorax. In eligible studies that met strict inclusion criteria, we conducted a meta-analysis to evaluate the diagnostic accuracy of pleural ultrasonography in comparison with chest radiography for the diagnosis of pneumothorax.

Results: We reviewed 601 articles and selected 25 original research articles for detailed review. Only 13 articles met all of our inclusion criteria and were included in the final analysis. One study used lung sliding sign alone, 12 studies used lung sliding and comet tail signs, and 6 studies searched for lung point in addition to the other two signs. Ultrasonography had a pooled sensitivity of 78.6% (95% CI, 68.1 to 98.1) and a specificity of 98.4% (95% CI, 97.3 to 99.5). Chest radiography had a pooled sensitivity of 39.8% (95% CI, 29.4 to 50.3) and a specificity of 99.3% (95% CI, 98.4 to 100). Our meta-regression and subgroup analyses indicate that consecutive sampling of patients compared to convenience sampling provided higher sensitivity results for both ultrasonography and chest radiography. Consecutive versus nonconsecutive sampling and trauma versus nontrauma settings were significant sources of heterogeneity. In addition, subgroup analysis showed significant variations related to operator and type of probe used.

Conclusions: Our study indicates that ultrasonography is more accurate than chest radiography for detection of pneumothorax. The results support the previous investigations in this field, add new valuable information obtained from subgroup analysis, and provide accurate estimates for the performance parameters of both bedside ultrasonography and chest radiography for pneumothorax evaluation.

Figures

Figure 1

A diagram represents the review process and selection of included studies.

Figure 2

Forest plot for sensitivity, specificity of CXR for the detection of pneumothorax. Inconsistency (I2) describes the percentage heterogeneity across studies that are not due to chance. I2 can be calculated as I2 =100% 3 (Q2df)/Q (see Figure 4 legend for definition of Q). df = degree of freedom = number of studies-1.

Figure 3

Forest plot for sensitivity, specificity of US for detection of pneumothorax. Inconsistency (I2) describes the percentage heterogeneity across studies that is not due to chance. (Refer to Figure 1 legend explanation of statistics).

Figure 4

Forest plot for diagnostic odds ratio (DOR) of US (left) and CXR (right). DOR = positive likelihood ratio/negative likelihood ratio = TP × TN/FN × FP. Inconsistency (I2) describes the percentage heterogeneity across studies that is not due to chance. Tau-squared represents the amount of heterogeneity. Cochran Q is a statistic that represents a ratio of total observed variation to within-study error; it is usually computed by summing the squared deviations of each study’s estimate from the overall meta-analytic estimate.

Figure 5

Summary receiver operative curves for US (left) and CXR (right). AUC, Area under the curve; SE, standard error.