Diagnostic performance of pleural biomarkers in pleural effusion: Comparative evaluation of CRP, CRP/albumin ratio, albumin, and the serum–pleural albumin gradient

Pleural effusion has heterogeneous etiologies, and the comparative behavior of serum and pleural biomarkers and their derived indices across clinically defined diagnostic groups remains incompletely characterized. This study aimed to comparatively evaluate serum and pleural biomarkers and derived indices in adults with pleural effusion, including serum–pleural biomarker relationships, discriminative performance for lung cancer, and biomarker differences across selected etiological groups. This prospective observational study included 70 consecutive adults who underwent thoracentesis at a single tertiary center. Serum and pleural fluid C-reactive protein, albumin, total protein, lactate dehydrogenase, and glucose were measured, and derived gradients, ratios, and C-reactive protein/albumin indices were calculated. Spearman correlation, Wilcoxon rank-sum tests with Benjamini-Hochberg false discovery rate correction, and receiver operating characteristic curve analysis were performed. Median age was 70 years (interquartile range: 60–81). Diagnostic categories were not mutually exclusive and included pneumonia (37.1%), heart failure (31.4%), malignant pleural effusion (22.9%), and lung cancer (21.4%). Pleural fluid albumin yielded the highest diagnostic accuracy for lung cancer (area under the curve=0.816; 95% confidence interval: 0.679–0.930; p<0.001), with 71.4% sensitivity and 91.7% negative predictive value at a 22.5 g/L cutoff. The serum–pleural albumin gradient followed (area under the curve=0.749; p=0.002), whereas all C-reactive protein–based markers showed no meaningful discriminative performance. All seven classical exudate/transudate parameters differed significantly between heart failure and non–heart failure groups (all corrected p-values <0.001). Serum–pleural correlations were numerically higher in the lung cancer subgroup, although correlation coefficients were not formally compared between etiological groups. In exploratory analyses, pleural fluid albumin and the serum–pleural albumin gradient demonstrated moderate discriminative performance for lung cancer. The observed variation in serum–pleural biomarker relationships across etiological subgroups should be regarded as exploratory and hypothesis-generating.