Objectives: 1) to evaluate the possibility of distinguishing pleural transudates and exudates through the joint determination of 26 biochemical parameters in pleural effusion and in plasma (including the determination of high molecular weight proteins, acute phase reactants, and proinflammatory citokines), and 2) to formulate a logistic regression equation for optimizing the classification efficiency, comparing the equation obtained with Light's criteria.
Patients and methods: All diagnostic thoracocentesis carried out in La Rioja Autonomous Community during a 22-month period were evaluated. The 245 clinical records were evaluated periodically along a minimum of 2 years, after the discharge of the patients. In pleural effusion and in plasma the following were quantified: total proteins, LDH, glucose, amylase, cholesterol, albumin, cholinesterase, phosphatase alkaline, urea, beta2-microglobulin, IgG, IgM, alpha2-macroglobulin, C reactive protein, transferrin, alpha1-antitrypsin, serum amyloid A protein, interleukin 1-beta, interleukin 6, tumoral necrosis factor-alpha, and lysozyme. In addition, the cellularity, polymorphonuclear elastase and adenosine deaminase were evaluated in pleural fluid.
Results: The LDH pleural effusion/plasma ratio was the individual parameter that showed higher area under the receiver operating characteristic curve for the separation of pleural transudates and exudates. Interleukin 6 and tumoral necrosis factor-alpha showed pleural effusion/plasma ratios higher than the unit, which suggests an in situ citokines production. An predictive logistic regression equation was obtained that incorporates only LDH and cholesterol ratios, including the diuretic treatment of the patient at the time of thoracocentesis, which did not modify the protein concentrations in pleural effusion. Except for LDH ratio, the logistic regression equation showed an area under the receiver operating characteristic curve higher than that of all the evaluated individual parameters, with a sensitivity of 95% and a specificity of 85% (70% for the Light's criteria).
Conclusions: LDH ratio is the best individual parameter for distinguishing pleural transudates and exudates. The additional evaluation of cholesterol ratio and of the diuretic state of the patient make possible to improve the clinical efficiency of this classification. The quantification of high molecular weight proteins, acute phase reactants and citokines does not contribute additional significant information.