Useful clinical biological markers in diagnosis of pleural effusions in children.

In a patient with an undiagnosed pleural effusion, the first question to answer is whether the fluid is an exudate or a transudate. This is usually determined by means of Light's criteria, which differentiate transudative effusions from exudative effusions by measuring the levels of total protein and lactate dehydrogenase in the pleural fluid (PF) and serum. In patients under diuretic treatment, Light's criteria misclassify transudates as exudates, but the serum to pleural fluid albumin gradient usually remains above 12 g/L. When tests are done only in PF, protein concentration >30 g/L performs at least as well as the other individual markers. To diagnose tuberculous pleuritis among exudates, PF adenosine deaminase and PF interferon-g exhibit high diagnostic accuracy. When malignancy is suspected the addition of tumour markers to the results of cytologic analysis increases the rate of detection. Other biochemical markers are useful in specific circumstances involving pleural effusion, such as amylase in effusions due to pancreatitis, or oesophageal rupture, and triglycerides in chylothorax. Several PF markers are associated with complicated parapneumonic effusion - e.g. low PF pH and glucose, and high PF LDH activity -- although PF pH appears to be the best biochemical aid in decisions regarding chest tube drainage. Recent reports suggest that neutrophil-derived enzymes (polymorphonuclear elastase and myeloperoxidase) can be useful as early indicators of the need of chest tube insertion; however these findings must be confirmed in large series. This review discusses the clinical usefulness of biochemical markers in the diagnosis and management of pleural effusions. The vast majority of prospective studies in this field have been conducted in adults and, although the mechanisms of pleural effusion production do not differ in children and adults, the prevalence of each etiologic cause does. Therefore it seems advisable to confirm or recalculate the predictive values of each marker in the paediatric population.