Abstract
BACKGROUND: Pleural fluid cell differentiation is recommended as an initial investigation of undiagnosed unilateral pleural effusions along with routine biochemical, microbiological, and cytological analysis. However, the sensitivity and specificity of cell differentiation from initial pleural aspiration remains unknown. As such, this study aimed to establish the value of pleural cell differentiation in the investigation of undiagnosed pleural effusions. METHODS: We performed a prospective study of 1,023 patients presenting to a UK tertiary pleural service with an undiagnosed unilateral pleural effusion between December 2008 and December 2016. Twenty-six patients were excluded due to incomplete data and unclear diagnoses, leaving 997 for analysis. Two respiratory consultants confirmed the diagnosis at 12 months. Cell differentiation was defined according to the British Thoracic Society (BTS) guidelines. RESULTS: The median age of the cohort was 73 years [95% confidence interval (CI): 71-74 years]. Of the 997 patients, 532 (53.4%) were found to have a malignant aetiology and 465 (46.7%) were non-malignant. On review of the pleural fluid cytology, 423 were mixed (42.4%) with no predominant cell type, 352 were lymphocytic (35.3%), 95 were neutrophilic (9.5%), 72 were eosinophilic (7.2%), and 55 were blood-stained (5.5%). Of the lymphocytic effusions, 53.7% had malignant aetiology. A lymphocyte predominance had a sensitivity of 35.5% (95% CI: 31.5-39.8%) and specificity of 65.0% (95% CI: 60.4-69.3%) for malignancy. The majority (80%) of neutrophilic effusions had infectious aetiology with a positive likelihood ratio of 20.6 (95% CI: 12.8-33.1), however, 14% of cases (13/95) had a final diagnosis of malignancy. Of the eosinophilic effusions, 43.0% were also of malignant aetiology, with a positive likelihood ratio of 0.66 (95% CI: 0.42-1.04). CONCLUSIONS: This study demonstrated that pleural fluid cell differentiation has limited diagnostic value. Although it may guide towards the likely diagnosis, it lacks diagnostic accuracy as a rule-in/out test and should be used cautiously by clinicians.