Diagnostic utility of pleural fluid carcinoembryonic antigen and CYFRA 21-1 in patients with pleural effusion: a systematic review and meta-analysis

Pleural effusions (PE) are the most common complications that may be produced by a wide variety of diseases. A large number of studies exploring the role of carcinoembryonic antigen (CEA) and cytokeratin fragment 19 (CYFRA 21-1) marker in differential diagnosis of PE have been published, employing differing methodologies with sometimes conflicting results. A comprehensive systematic review would be useful to synthesize the currently available bulk of information. The objective of this work was to assess and compare the overall value of pleural fluid CEA and CYFRA 21-1 in differential diagnosis of PEs with a meta-analysis. All the English and Chinese published studies for differential diagnosis of PEs by pleural fluid CEA and CYFRA 21-1 were collected. Methodological quality of the included studies was evaluated. Pooled sensitivity and specificity were calculated, the threshold effect and the possible sources of heterogeneity were also analyzed. Summary receiver operating characteristic (SROC) curve analysis was used to compare the differential diagnostic ability of pleural fluid CEA and CYFRA 21-1. A total of 19 studies were included in the meta-analysis, with a total of 3,228 subjects. Pooled sensitivity and specificity of CEA and CYFRA 21-1 were 45.9% (43.2-48.5%) and 97.0% (96.0-97.8%), and 47.3% (44.0-50.6%) and 91.8% (89.5-93.7%), respectively. Both CEA and CYFRA 21-1 have a threshold effect, the main source of heterogeneity was from variable assay methods. The areas under the SROC curve (AUCs) of CEA and CYFRA 21-1 were 0.7691 and 0.8213, respectively. There was no statistical significance between the AUC of CEA and CYFRA 21-1 (P>0.05). Both CEA and CYFRA 21-1 have good performance in the differential diagnosis of PE, when compared with CEA, CYFRA 21-1 has no advantage.